OUTSOURCING, TEAMWORK AND BUSINESS MANAGEMENT
No part of this digital document may be reproduced, stored in a retrieval system or transmitted in any form or by any means. The publisher has taken reasonable care in the preparation of this digital document, but makes no expressed or implied warranty of any kind and assumes no responsibility for any errors or omissions. No liability is assumed for incidental or consequential damages in connection with or arising out of information contained herein. This digital document is sold with the clear understanding that the publisher is not engaged in rendering legal, medical or any other professional services.
OUTSOURCING, TEAMWORK AND BUSINESS MANAGEMENT
KARL E. CARETTAS EDITOR
Nova Science Publishers, Inc. New York
Copyright © 2009 by Nova Science Publishers, Inc.
All rights reserved. No part of this book may be reproduced, stored in a retrieval system or transmitted in any form or by any means: electronic, electrostatic, magnetic, tape, mechanical photocopying, recording or otherwise without the written permission of the Publisher. For permission to use material from this book please contact us: Telephone 631-231-7269; Fax 631-231-8175 Web Site: http://www.novapublishers.com NOTICE TO THE READER The Publisher has taken reasonable care in the preparation of this book, but makes no expressed or implied warranty of any kind and assumes no responsibility for any errors or omissions. No liability is assumed for incidental or consequential damages in connection with or arising out of information contained in this book. The Publisher shall not be liable for any special, consequential, or exemplary damages resulting, in whole or in part, from the readers’ use of, or reliance upon, this material. Any parts of this book based on government reports are so indicated and copyright is claimed for those parts to the extent applicable to compilations of such works. Independent verification should be sought for any data, advice or recommendations contained in this book. In addition, no responsibility is assumed by the publisher for any injury and/or damage to persons or property arising from any methods, products, instructions, ideas or otherwise contained in this publication. This publication is designed to provide accurate and authoritative information with regard to the subject matter covered herein. It is sold with the clear understanding that the Publisher is not engaged in rendering legal or any other professional services. If legal or any other expert assistance is required, the services of a competent person should be sought. FROM A DECLARATION OF PARTICIPANTS JOINTLY ADOPTED BY A COMMITTEE OF THE AMERICAN BAR ASSOCIATION AND A COMMITTEE OF PUBLISHERS. LIBRARY OF CONGRESS CATALOGING-IN-PUBLICATION DATA Available upon request.
ISBN 978-1-61470-493-5 (eBook)
Published by Nova Science Publishers, Inc.
New York
CONTENTS Preface
vii
Chapter 1
Offshoring Knowledge Work: How Far Can it Go? Evidence from Drug R and D David Finegold, Niclas Erhardt and Mari Sako
Chapter 2
Multinational Exploration of Acquired R&D Activities Jens Gammelgaard
15
Chapter 3
Management and Control in Service Firms: Bridging the Gap between Organization Studies and Service Management Per Skålén
31
Chapter 4
Activity Awareness and Complex Teamwork John M. Carroll, Mary Beth Rosson, Craig H. Ganoe, Marcela Borge, Jamika D. Burge, Umer Farooq, Gregorio Convertino, Paula M. Bach, Helena Mentis and Hao Jiang
47
Chapter 5
Teamwork in Today’s World Carol Boswell and Sharon Cannon
73
Chapter 6
Teamwork and PBL-Based Teacher Education: A Study on Prospective Science Teachers’ Opinions Laurinda Leite and Esmeralda Esteves
83
Chapter 7
Developing Effective Teams and Protecting the Vulnerable: An Interprofessional Journey Susan Morison and Moira Stewart
99
Chapter 8
The Impact of Engineering Design on Outsourcing Decisions Mahmood Al-Kindi and Ali A. Yassine
115
1
vi
Contents
Chapter 9
Time Allocation and Outsourcing within Households. Differences in Lifestyle between Native Dutch and Immigrants in the Netherlands J.R. Cornelisse-Vermaat, H. Maassen van den Brink, J.A.C. van Ophem and G. Antonides
135
Chapter 10
Outsourcing and Public Sector Efficiency: How Effective Is Outsourcing in Dealing with Impure Public Goods? Argentino Pessoa
167
Chapter 11
Examination of Dedicated Relationships between Automotive Suppliers and Carmakers: Evidence on the Flagship / 5 Partners Model Bart Kamp
191
Chapter 12
Allocating Outsourced Warranty Service Contracts Michelle Opp, Ivo Adan, Vidyadhar G. Kulkarni and Jayashankar M. Swaminathan
213
Chapter 13
The Importance of Context in Determining Consumer Response to Food Safety Events: The Case of Mad Cow Disease Discovery in Canada, Japan and the United States Sayed Saghaian, Leigh Maynard and Michael Reed
235
Index
267
PREFACE This new book is dedicated to the nontrivial problems of organizations trying to accomplish anything. Outsourcing, for all its seemingly attractiveness, contains hidden costs in coordinating and can involve creating negative customer responses. Teamwork is a concept of significance in outsourcing as well as in internal activities all of which are within the realm of business management. Chapter 1 focuses on the reasons why drug companies choose to offshore and/or outsource some R and D tasks and not others, and how they manage the process and the issues that arise once the work is moved offshore. The chapter sets forth an analytical framework for analyzing the offshoring/outsourcing decision that includes the availability and cost of talent, how closely related the knowledge is to the core strategy of the firm, and the nature of work/task itself. The task is not treated as a given, but rather is considered to be alterable through the process of offshoring/outsourcing, either necessitated by distance requiring a new mode of working and/or enhancing modularity (i.e. separability of tasks that enable a job to become ‘impersonal’) that had been possible prior to offshoring. Chapter 2 presents the results of a survey of 54 Danish multinational corporations that have acquired activities abroad. The role of the acquired R&D units was the focus of the survey, particularly with respect to the schism between basic and applied R&D, and the schism between autonomous and network R&D. This paper establishes the connection between a multinational corporation that follows a capability-motivated acquisition strategy and the R&D role new subsidiaries should play in order for the acquired resources to be utilized corporation-wide. Statistical findings reveal the need to follow a combination of basic and network-oriented R&D activities when focusing on capability development. Chapter 3 draws together previous research within the boundaries of organization studies and services management, and then proposes an alternative focus for studying management and control in service firms. The chapter argues that organization studies have contributed to the study of management and control in service firms but without actually focusing on the central issues that services management research has argued should characterize management and control in service firms. By bringing together the two traditions an approach to empirical research that would deepen understanding of management and control in service organizations is outlined. Collaborators must attain and maintain reciprocal awareness of shared activity in order to coordinate effectively (Dourish & Bellotti, 1992). They need to be assured that their partners are ‘there’ in some sense, which is not always evident or simple in computer-mediated
viii
Karl E. Carettas
collaboration. They need to know what tools and resources their counterparts can access, who they know that might know something, or know how to do something that would be critical. They need to know what relevant information their collaborators know, and what they expect, as well as their attitudes and goals. They need to know what criteria their partners will use to evaluate joint outcomes, the moment-to-moment focus of their attention and action during the collaborative work, and how the view of the shared plan and the work actually accomplished evolves over time. Research on collaboration and technology support for collaboration has identified several types of awareness: social awareness, action awareness, workspace awareness, situation awareness (for an excellent review, see Schmidt, 2002). Most investigations of awareness research have focused on synchronous phenomena: awareness of who is participating in an ongoing activity, awareness of what each person is currently doing in that activity context, and awareness of how the team as a whole is performing. Asynchronous awareness phenomena, for example those supported by version control systems, shared calendars, and project management software, have received less attention. Our research has focused on activity awareness, a programmatic concept for the mutual awareness of partners in a shared activity of significant scope and duration. Activity awareness transcends synchronous awareness of where a partner's cursor is pointing, where the partner is looking, etc. It involves monitoring and integrating many different kinds of information at different levels of analysis, such as events, tasks, goals, social interactions and their meanings, group values and norms, and more. It involves monitoring and integrating more-or-less continuingly to learn about developing circumstances and the initiatives, reactions, and sense making of other people with respect to on-going and anticipated courses of action. Activity awareness is not merely a matter of coordinating state information. It is continually negotiated and constructed throughout the course of a collaborative interaction. Thus, it is a process that is constitutive of collaboration. In the balance of Chapter 4 we will first describe fieldwork characterizing routine social practices to establish and maintain activity awareness during complex teamwork in regional emergency management planning, emergency room operations, collaborative education, open source software development, scientific collaborations, and management of nonprofit community groups. We then describe software systems we have developed to support activity awareness in complex teamwork in some of these contexts. We close with some discussion of the challenges of supporting activity awareness. Teamwork in all venues today requires management and staff to utilize a variety of approaches to ensure safety in the workplace. The Institute of Medicine’s report is just one example of an organization identifying the importance of effective collaboration by all employees to ensure the wellbeing of engaged participants. Communication is crucial in the response to the call for improved partnership within the workplace. Communication tools such as SBAR (situation, background, assessment, and recommendation), LDS (Let’s Do Something) leadership style, and Huddles can be readily utilized to facilitate effective and comprehensive delivery of key information essential for teamwork regardless of settings. Chapter 5 will discuss these communication tools through the use of safety issue examples such as safe medication administration. Aspects of teamwork will be defined, delineated, and applied. The application of evidence-based practice guidelines would serve as the foundation for the discussion related to effective and successful implementation of a sound alliance within any workplace setting.
Preface
ix
It is fully accepted that successful science learning depends at least in part on teachers’ teaching competences. In addition, there is some evidence that teachers tend to teach as they were taught. Therefore, in order to develop teachers’ innovative and student-centered teaching competences, methods courses should acknowledge teaching methodologies similar to those that prospective teachers will be required to use in their future as teachers. Problem-Based-Learning (PBL) and teamwork are student-centered teaching approaches that may foster the development of relevant competences for students as citizens. In fact, PBL can promote learning how to learn competences while teamwork can foster the development of social and interpersonal skills. However, to successfully put these teaching approaches into practice teachers need to fully acknowledge big role changes. Hence, in order to prepare innovative teachers, PBL, together with teamwork, should be used in initial science teachers’ education programs. Chapter 6 describes how 38 prospective physical sciences teachers evaluate teamwork and PBL carried out within a methods course to approach a module on Using the lab for physical sciences teaching. Data were collected by means of a questionnaire, a self-evaluation grid and a videotaped discussion focusing on the diverse parts of a PBL sequence: problems formulation from a scenario, problem solving and synthesis and evaluation. Results indicate that prospective teachers valued PBL and teamwork, as they felt that the latter helped them to cope with the new roles that they were required to undertake throughout the PBL sequence. However, the facilitating effect of teamwork seems to be insufficient to lead students to fully overcome their difficulties with more unusual tasks. Nevertheless, it seems that they may be prone to use these teaching approaches when they become science teachers. Chapter 7 examines teamwork in medical care and makes particular reference to the pediatric team. It considers the characteristics of an effective team and the perceived benefits of team working in healthcare. It discusses the important role that education, and in particular interprofessional education, might have in helping to prepare a future workforce capable of effective patient focused team working. The contributory effect of different professional cultures is also examined and arguments are presented that reflect on the meliorating role of appropriate interprofessional education. The theoretical arguments are illustrated with reference to recent highly publicized and significant failures by teams responsible for the health and well-being of children. Recent case studies and judicial reviews from the United Kingdom and the United States are discussed. Many models in the literature examine outsourcing based on product modularity; however, modularity is assumed to be known and exogenous. In reality, modularity is a decision variable defined (i.e. built into the product) during the engineering design phase of the product development (PD) process. This paper bridges the gap between the outsourcing literature and the engineering design literature by incorporating into the outsourcing decision model detailed engineering design information regarding the time spent on various engineering design activities within the PD process (e.g., system design, detailed design, and testing and integration). Chapter 8, a mathematical model is developed to study the impact of outsourcing and time spent in the various engineering design activities on firm’s revenue (represented by a marketing window) for different PD scenarios. These scenarios differ in four major factors: technological capability of the firm and its suppliers, design task size and complexity, nature
x
Karl E. Carettas
of detailed design work (i.e., fraction of rework), and amount of outsourcing. It is shown that this model is a convex optimization problem that admits a global optimum; however, no explicit closed-form solution could be obtained and the problem was solved using optimization software. The optimal solution reveals several interesting managerial insights regarding the impact of the various engineering design decisions on the outsourcing decision. First, spending more time in system design leads to higher outsourcing fraction and vice versa; that is, well defined product architectures lead to higher outsourcing. Second, higher firm capability makes outsourcing less attractive. Third, outsourcing is found to be more attractive at the medium task sizes compared to larger or smaller tasks. Fourth, a product with a complex architecture will lead the firm to spend more time in system design and thus outsource more. Lastly, as the rework fraction of detailed design increases, it is better to spend more time in system design and outsource more. Due to the increased female labour participation in the past decades, households lack time to perform all households and care activities. At present in the Netherlands, households can outsource home cleaning to a cleaning lady/man, cooking to restaurants (or people can eat ready meals or takeaway food), and childcare can be outsourced to day care centres. The increased household income, attributed to higher female labour participation, gives more possibilities to outsource domestic work. Outsourcing could not only be determined by socioeconomic and demographic variables, culture (or ethnicity) can also be of importance in explaining outsourcing within households. Chapter 9 aims to determine the time households spend on domestic tasks and care activities and whether differences in lifestyle and ethnicity are related to outsourcing behaviour. Time spent on household and care activities is estimated with a model including socioeconomic and demographic variables and including some lifestyle determinants. Household expenditures on different types of outsourcing possibilities within households are measured and differences are drawn between native Dutch and non-western immigrants. For the analyses a sample (2001) is used that consists of Dutch, Turks, Surinamese/Antilleans, and Moroccans (N=2551). The analyses show that immigrants spend less time on household and care activities compared to native Dutch. Both Household income and level of education are determinants of the expenditures on outsourcing of domestic tasks and care activities. Native Dutch and Surinamese/Antilleans have comparable expenditures on home cleaning and childcare, whereas Moroccans and Turks spend more on takeaway food and delivery food. The results reveal differences as well as similarities in lifestyle between native Dutch and non-western immigrants. The debate on new public management, together with the shortage of public funds, has had a considerable impact on public administration. Accordingly, many governments have searched positive impacts on the efficiency, equity and quality provision of public services through increasing competition and active participation of the private sector, considering outsourcing as the appropriate instrument to attain such endeavor. However, private involvement in public services provision is controversial. While, on the one hand it is touted as a way to increase efficiency and accountability by turning over choices to individuals in the market place, on the other hand, some argue that it has the potential to produce considerable fraud and corruption if managerial control by the public sector is weak. So, given this context, Chapter 10 aims to assess the private involvement in public services in
Preface
xi
efficiency terms, putting aside ideological considerations. So, after the introduction, we present a definition of public goods and we characterize their different types, with particular emphasis on “impure” public goods. Section 3, focuses on market failures together with equity considerations as the main reasons that configure the role of the public sector in providing impure public goods, as well as on the possibility of government failures. Section 4 deals with the benefits and costs of outsourcing in the public sector. Section 5 describes the most frequent forms of private sector involvement in the provision of impure public goods, as well as the advantages and disadvantages of the different options. Section 6 carries out some comments on the need for regulation. Finally, section 7 concludes. The Flagship / 5 Partners model argues that key suppliers are dedicated exclusively to flagship firms and that flagship firms work on an exclusive basis with key suppliers. As the F/5P model is partly rooted in empirical analyses of the North-American automotive industry, it is interesting to test its external validity on the European car industry. Similarly, it seems relevant to test its claim of exclusive buyer-supplier relationships as there are also scholars that report on non-exclusive b2b practices. Chapter 11 analyzes the component supply relationships for 32 car models to test the exclusivity presumptions of the F/5P model. Results show that industry-wide client bases on behalf of suppliers and multiple sourcing practices by carmakers are a stronger empirical reality than exclusive flagship firm-key supplier relationships. Outcomes also indicate that previously in-house parts of carmakers successfully succeed in establishing client relationships with third party OEMs. Motivated by our interactions with a leading manufacturer of computers, in Chapter 12 we consider static allocation as applied to the problem of minimizing the costs of outsourcing warranty services to repair vendors. Under static allocation, a manufacturer assigns each item to one of several contracted repair vendors; every time a particular item fails, it is sent to its preassigned vendor for repair. In our model, the manufacturer incurs a repair cost each time an item needs repair and also incurs a goodwill cost while items are undergoing repair. We model each service vendor as a finite population multi-server queueing system and formulate the resulting outsourcing problem as an integer-variable resource allocation problem. After establishing convexity results regarding the queue lengths at the repair vendors, we show that marginal allocation is optimal. Through a detailed computational study we compare the optimal algorithm with five static allocation heuristics in terms of time and optimality gap. Our study indicates that the optimal algorithm takes less than a minute to solve industry size problems on average. Further, the commonly used heuristics are far away from the optimal on average, thus emphasizing the benefits of the optimal allocation algorithm. We also compare the optimal static allocation to two simple dynamic allocation heuristics. The results of this study further validate the use of static allocation as a justifiable and easy-to-implement policy. Among other computational insights we show that when the number of items to be allocated is large, a single-server approximation leads to optimal allocations in most of the cases. Chapter 13 consists of three parts that present two complementary statistical analyses. In the first part, we show consumer reactions to BSE in Japan using Directed Acyclic Graphs and historical price and quantity decompositions. The Japanese beef markets faced two subsequent cases of BSE discoveries in 2001, eroding consumer confidence in beef supply channels with huge economic losses to the Japanese beef industry. In the second part, we look at BSE’s impact along the U.S. supply chain using similar contemporary time-series methods. The U.S. beef industry faced BSE in 2003, which led to differential impacts on farm, wholesale, and retail markets. Relative to the U.S., Japanese consumers have a strong
xii
Karl E. Carettas
preference for domestically produced beef, encouraged by retail country-of-origin labeling and BSE media coverage critical of imported beef. Consistent with these differences in preferences, marketing, and information, we observe more negative and more nuanced reactions to BSE in Japan versus the U.S. The third part highlights contextual differences in Canada. A double-hurdle model of Canadian fast food beef purchases shows no significant BSE impacts on the likelihood or quantity of fast food beef item purchases. When applied to Canadian supermarket beef purchases, however, a striking pattern emerges. After the initial BSE event in 2003, when media coverage focused mainly on the plight of ranchers, beef demand increased significantly. Moreover, demand increased the most in Alberta, the center of Canada’s beef industry. Following two later BSE events, beef demand fell significantly. The results illustrate the importance of context along at least five dimensions: the food purchase venue, the geographic proximity of consumers to BSE events, the ordering of BSE events, the role of supplier behavior, and the nature of media coverage.
In: Outsourcing, Teamwork and Business Management ISBN: 978-1-60456-956-8 Editor: Karl E. Carettas, pp. 1-14 © 2009 Nova Science Publishers, Inc.
Chapter 1
OFFSHORING KNOWLEDGE WORK: HOW FAR CAN IT GO? EVIDENCE FROM DRUG R AND D David Finegold1, Niclas Erhardt2 and Mari Sako2 1
Rutgers University USA University of Maine USA 3 Said Business School UK 2
Introduction Offshoring of services is seen as a relatively new form of internationalization, enabled by trade liberalization policies and advances in information and communication technologies. Starting with simple and codified tasks, companies are now offshoring increasingly complex and knowledge-based activities requiring more qualified workers. Consequently, there is growing concern in the US and Europe that high quality jobs are being lost to low wage emerging economies, most notably China an India. Historically, primarily unskilled and semi-skilled manufacturing jobs were considered vulnerable to offshoring. By contrast, the extent of offshoreability of service jobs is not so well correlated with the skill content of those jobs (Blinder, 2006). Some white collar tasks that require relatively little formal education or qualifications may remain ‘personal’ requiring face-to-face contact and judgment-based interaction, but other more highly skilled tasks may be rendered ‘impersonal’, enabling delivery from a distance with little need for such interaction. In addition, the very definition of what counts as highly skilled may vary across cultures and nations – Aron (2008), for example, found that complex quantitative analysis was rated as the most highly skilled positions in the US and UK, while in India and China these tasks were rated as relatively routine, while those requiring very contextdependent personal interaction were seen as the most highly skilled. This chapter analyzes the potential and limits of offshoring knowledge work, examining which tasks are being outsourced and offshored in the most knowledge-intensive portion – Research and
2
David Finegold, Niclas Erhardt and Mari Sako
Development (R and D) – of one of the most knowledge-intensive sectors of the economy: the drug industry. Research and development in the drug industry is one of the areas that has attracted a great deal of public investment in regions across the US and Europe, as governments seek to create and retain new biotech and pharmaceutical companies and the high-wage jobs they generate. The hope was that a number of factors would make these jobs unlikely to relocate: their highly skill content – a higher percentage of PhDs than any other sector, the highly regulated nature of the industry, and the huge risks and investment required to develop new drugs, costing an average of over $1 billion per new drug and over 90 percent of projects never leading to an approved product. Recently, however, R and D in this sector has begun to move to locations outside the advanced industrial economies, a relatively new phenomenon, made possible by the internet and the changes in intellectual property regimes in India and China. The sector therefore provides an ideal setting for studying the evolving nature of the offshoring of knowledge work and the factors that are driving this. The chapter focuses on the reasons why drug companies choose to offshore and/or outsource some R and D tasks and not others, and how they manage the process and the issues that arise once the work is moved offshore. The chapter sets forth an analytical framework for analyzing the offshoring/outsourcing decision that includes the availability and cost of talent, how closely related the knowledge is to the core strategy of the firm, and the nature of work/task itself. The task is not treated as a given, but rather is considered to be alterable through the process of offshoring/outsourcing, either necessitated by distance requiring a new mode of working and/or enhancing modularity (i.e. separability of tasks that enable a job to become ‘impersonal’) that had been possible prior to offshoring.
Off-shoring and Outsourcing Knowledge Work: Analytical Framework Offshoring refers to the sourcing and coordinating of tasks across national borders. Offshoring may include both in-house (i.e. captive) sourcing and outsourced activities that cross the boundary of the firm. Outsourcing, in turn, may occur both domestically (onshore) and abroad (offshore). In this chapter, we focus on the simultaneous shifts in the location of work and the boundary of the firm – i.e. when key aspects of drug discovery and development are shifted from being performed in-house at or near the corporate headquarters to a contractor or partner in India or China. The increase in off-shoring/outsourcing has largely been driven by cost reduction, adding flexibility, access to know-how and facilities that a company may be unable to afford alone to address changes in the market and customer demands (Kumar and Eickhoff, 2006; Ward, 2004). In the drug industry, however, the desire for cost savings in R and D must be balanced by the key role that new knowledge generation plays in building and maintaining a competitive advantage. The traditional view of off-shoring (as well as outsourcing) would suggest that intellectual property (IP) such as competencies, processes and know-how that is core to the business is kept in-house, while non-core knowledge is outsourced (Kumar and Eickhoff, 2006). The logic is that keeping the core-knowledge from “leaking” would prevent it from eventually reaching competitors that might copy the capability that the company relies
Offshoring Knowledge Work: How Far Can it Go? Evidence from Drug R and D
3
ge
1.
le d
2.
Low
no w
Short Supply
3.
K
Abundant Supply
C
High
or e
Talent Availability
on for competitive advantage. This strategy is still practiced in national biotech industries. For example, Norus’s (2005) study on the biotech industry in Copenhagen found that companies tended to in-source core competencies based on collaborative networks fostering flow of knowledge. The growth of the industry was fostered by keeping sticky knowledge (core) within the network and leaking and exporting less sticky knowledge to external supplier and partners. However, intensifying global competition, drying “pipelines”, challenges with finding novel products, and the need to increase the speed of launching products into the marketplace are all trends leading companies to rethink the organization of the R and D function. Producing innovative and quality products is no longer exclusively explored within the boundaries of the firm; an emerging trend suggests that companies are starting to off-shore their R and D function as well. Companies such as Dell, BP and Shell, have enjoyed benefits of outsourcing intellectual capabilities (i.e. R and D) that enabled them to build and maintain their leading positions in their industries (Quin, 1999). Moreover, Toyota’s success has in part been attributed to their inter-firm network among suppliers that enable effective knowledge sharing based on institutional routines (Dyer and Nobeoka, 2000). Kinder (2003) provided evidence on the competitive advantage of having strong external supply network as conduits of value and flow of knowledge to enhance innovation and creativity. Quadros, Consoni and Quintao, (2005) studied the trend of R and D outsourcing in the Brazilian automobile manufacturing, which was larger than previously thought, and is growing based on cooperative research networks. Interestingly, MacPherson, (forthcoming) pointed out that external venders can provide innovative services to complement core competencies of client firm. The evidence suggests that it is not the in-house stock of “core-competencies” per se that seems to generate a competitive edge (as the traditional off-shoring/outsourcing argument would predict), but rather the company’s ability to integrate and execute with speed and rely on external partnerships to compliment and supplement its own R and D capabilities.
Modular (Standardized)
Integrated (Ill-Structured)
Task Modularity
Figure 1. Determining Outsourcing and off-shoring of Work Core Knowledge, Task Modularuty and Talent Availability.
4
David Finegold, Niclas Erhardt and Mari Sako
The research literature identifies three central factors that may determine what types of work in drug discovery and development are likely to be outsourced or off-shored. The cube in Figure 1 provides a framework for integrating these three factors: (a) CORE KNOWLEDGE: control over core capabilities and the protection of core workforce; (b) TALENT AVAILABILITY AND COST: the relative cost and supplyy of talent in off-shored and out-sourced locations and (c) TASK MODULARITY: the extent to which tasks are welldefined and standardized (modular) or ill-structured and require integrated know-how. The first factor affects primarily the outsourcing (make-or-buy) decision but not the location (offshoring) decision; so for example, fear of knowledge leakage may prevent firms from outsourcing to an external provider, but it can be dealt with through captive offshoring. By contrast, the second factor affects primarily the location (offshoring) decision but not the outsourcing decision per se, in a world of imperfectly mobile international labor markets. The third factor, the nature of tasks, however, affects both the decision on the boundary of the firm and that of location. To illustrate how this decision-making framework operates, let us provide examples of drug development work that might fall in some of the key boxes: 1. In house/onshore: for US drug companies, the early stages of drug discovery biology have historically been performed in-house because the supply of talent was greatest, it was difficult to modularize tasks, and there are major concerns about the loss of IP (first quadrant in Figure 1). 2. Outsource/onshore: early stage Phase I clinical trials are often done in partnership with nearby teaching hospitals to gain access to the knowledge and specialized expertise of their key medical thought leaders and facilitate completion of this phase as quickly as possible. These close trust relationships help to foster protection of IP until the firm is ready to release the results. 3. Offshore/outsource: large-scale production of small molecule drugs is increasingly being both outsourced and offshored, since the task is easy to modularize and measure, and the talent to produce the drugs is widely available. This has tended to go to locations (such as Puerto Rico, Ireland, Singapore) which offer both strong protection for IP and tax and other investment incentives that make it a low-cost option for firms. Particularly smaller companies will often use a contract manufacturing organization (CMO) to gain a skill set that is lacking within the organization, to save capital costs, and to leverage economies of scale. Other factors might also affect the R and D location decision. For example, governments may offer incentives to firms to create these jobs in their jurisdiction; we treat such incentives as effectively reducing the cost of labor. We will explore this framework in the context of two case studies of drug discovery work in early-stage US-based drug companies that involve the greatest coordination challenges: shifting work that was previously done in-house at the corporate headquarters (#1) to having the work both outsourced and offshored (#3).
Offshoring Knowledge Work: How Far Can it Go? Evidence from Drug R and D
5
Company Cases Biologics. Biologics1 is a midsize drug company in the northeast region of the USA that focuses on immunological diseases, with drugs that can be administrated via tablet, inhaled or topical forms. In 2007 the company had 150 fulltime employees, and sales of over $20 million. It has several promising clinical compounds in phase I and II clinical development and in order to sustain its growth, they have been collaborating with both large and small external companies. Starting in 2000, Biologics began to contract out work to emerging clinical research organizations (CROs) in both China and India. They started with a firm where they had a personal relationship with the founder, who had extensive experience in the U.S. pharmaceutical industry. Routine chemistry, and later stage clinical development have been the main tasks being contracted to offshore partners, while pharmacokinetics and animal testing have been outsourced to partners in the U.S. Most of the early stage biology, proprietary chemistry, and Phase I trials continues to be kept in-house. Biologics presently pays for 30 full-time equivalents (FTEs) at their partner compared with 55 internal medical chemists. Most of the internal chemists have PhDs or Masters degrees, while a majority of the chemists in their China partner are at the B.S. level, with only a handful of PhDs who manage the work. While India has been the main location for contracting out development work to this point, Biologics believes China may become the preferred location for earlier stage research, as the Chinese government invests heavily in research universities and developing their own know-how and supply of PhDs, and opens up its economy further for western companies and foreign investments. Small Pharma. Small Pharma is a private company founded in the late 1990s that focuses on small-molecule drugs that target the protein transcription process, with application to a wide range of rare genetic diseases. Their current pipeline is promising with products targeting a range of areas. Similar to Biologics, Small Pharma is currently collaborating with large pharmaceutical companies to strengthen their access to development and commercialization capabilities and financial resources. The company has enjoyed rapid growth in the last two years, more than doubling in size to over 150 employees thanks to success with several large grants and deals with large pharma. But the CEO has been determined to grow conservatively, avoiding investment in expensive facilities and to avoid costly mistakes and having to lay off part of its workforce. Small pharma began experimenting with outsourcing/offshoring in 2006. Their first partnership with an external CRO was in China for routine chemistry work. It later shifted its offshoring focus to India, where its research partner now has 25 chemists and five managers. The relationship was viewed as such a success that they decided to expand into biology, both with the original partner and a second firm in India.
Reasons for Offshoring and Outsourcing Our semi-structured interviews with senior executives and project managers at these two companies identified a set of key drivers and barriers to offshoring R and D work. 1
Names are fictitious to protect companies’ identity.
6
David Finegold, Niclas Erhardt and Mari Sako
Cost Reduction The primary driver of offshoring in biopharma R and D, consistent with findings from other sectors, is cost reduction. A fundamental challenge for small biotech companies is to generate funding and to manage cost associated with the discovery and development process – a long and risky endeavor. The need for contracting emerged in the chemistry area, as a way to both save costs and allow existing scientists to be more productive. It freed up time for scientists to focus on interesting and promising work while reducing the cost of labor intensive work. According to both firms, in India and China: The CROs can do the work for roughly one quarter of the labor cost. There are things we simply cannot get done. They help us meet our goals and milestones. Travel expenses are additional, but we try not to travel more than two times per year.”
The labor cost estimate assumes relatively similar productivity from the offshore workforce, which the firms are able to achieve for most workers after an initial start-up training period. Close monitoring of individual output allows them to identify where there are productivity or quality issues and to have their partner take corrective action. This cost-saving estimate, however, does not include the additional burden this places on the firm’s already hard working scientific managers who must oversee the relationship with the Indian or Chinese partner. Said one: “What is not factored in is the extra cost in managerial time. That comes out of the hours of salaried managers on top of other duties.”
Managing Risk and Volatility: Protecting the Core Workforce Contracting out work also provides the firms with a means to minimize risk, by allocating exploratory projects to external contractors since it would be riskier to hire core employees for projects that may not be sustainable. One of the CEO’s goals when he founded Small Pharma was never to have to lay any employees off. This can be a major challenge in a highly volatile and risky sector where only one of 1,000 drug targets makes it from the lab to the market, and typically requires a decade or more of development. To minimize the chance of layoffs, he has tried to grow his firm as conservatively as possible, hiring only individuals for whom there are multiple projects to fully occupy them and several years of funding to support their work. Contracting provides the firm with a buffer workforce for those projects whose future is highly uncertain and providing flexibility necessary in drug discovery and development.
Extending the Drug Portfolio Contracting out work to much lower cost providers allows biopharmaceutical companies to use their scarce internal capital to fund small-scale development projects that would not be feasible with a US workforce. By farming out certain non-mission critical projects, Small Pharma was able to conduct early stage research on focused compounds to “test the waters.” If the research proves promising, these results can then be used to attract additional funding
Offshoring Knowledge Work: How Far Can it Go? Evidence from Drug R and D
7
from government grants, foundations, investors, and/or corporate partners. As one manager explained: We have multiple therapeutics and a number of targets that we are not exploring, we have interest in them, but we are simply not working on them (internally)… It's not about replacing work; with our limited space, we outsource work that was not currently being done here. We did not take jobs away from the US. The economics is such that we could not afford the work if we did it here.
This is how Small Pharma advanced a small anti-bacterial program, which began with some exciting preliminary results identified by one of Small Pharma’s researchers working on an anti-infectant. The project, however, was a lower priority than many of the others in their pipeline and the firm had neither the personnel nor the resources to pursue it internally. The offshoring option, enabled the project to progress to the next stage of development. If the firm is able to use this new data to attract further funding, then the initial offshoring will have generated additional employment both in New Jersey and India, thus contradicting a common perception of offshoring as a zero sum game, where a fixed number of jobs is done in the US or in Asia.
Not Skill Shortages Another rationale that is sometimes cited for reasons for offshoring is the lack of specialized skills available in the local labor market. In the case of these two firms, however, that was not the case. This may be due to the fact that they are based in New Jersey, a state where 75% of the world’s largest pharmaceutical companies have a significant presence. Many large pharmaceutical firms have merged and/or restructured over the last decade, resulting in large layoffs. This experienced workforce, along with the graduates from strong research universities in the area, has resulted in a large supply of bioscience talent available for start-up companies. Indeed, many international firms -- such as Novo Nordisk, Roche, Bayer, and many of the Japanese pharmaceutical companies -- have located their US headquarters in New Jersey to tap into the supply of workers with experience in all stages of bringing a drug to market.
Challenges to Effective R and D Offshoring Although the firms saw a number of potential benefits from moving some R and D work to India and China, they faced a common set of challenges to making these relationships work effectively. The challenges, similar to those found in other sectors (Quin, 1999), ranged from managerial and personnel issues -- such as managing expectations, increased complexity, virtual interaction and cross-cultural differences -- to logistical challenges.
8
David Finegold, Niclas Erhardt and Mari Sako
Managing Virtually Offsetting some of the significant direct labor cost savings, were the indirect costs associated with managing a workforce halfway around the world. The challenge began with identifying local companies that had the necessary competence and facilities to conduct the work. Although the hiring process was the responsibility of the contractors in India and China, Biologics and Small Pharma had to devote significant time to training the offshore workers to develop the specific competencies needed to carry out the required tasks. The physical distance also can create communication issues, particularly in the early stages of the relationship. Without the ability for daily, informal face-to-face interactions that they had taken for granted when all the work was performed in-house, firms were forced to establish more formal work routines, processes, and expectations to manage the relationship. As one Small Pharma manager observed: For the first six weeks we didn’t know how the work was going. It was frustrating. We realized we needed more structure on the report in the beginning. So we said, we need a written report….We reached a point when productivity was low. I talked to the lead PhD. I needed to tell her about our expectations and their performance. The productivity was less than 10% of expected. But it was just for one person. I got good realistic output from the others… We are very clear about our expectations these days.
Cheap and rapid communication -- through the phone, internet and e-mail – are essential enablers of managing virtually. But while such mechanisms are useful for sharing data and transmitting factual information, they are not as useful for hashing through problems. The lack of non-verbal cues needed for reaching agreements and understanding are amplified in cross-cultural settings where terms like “no” and “deadline” may be interpreted quite differently. Trouble-shooting complex problems or sharing “lessons learned”, tends to work best when teams have had enough face-to-face interaction to foster high degrees of trust. One mechanism to try to build such richer communication is video conferencing. However, in this case the 13-15 hour time difference meant that members on one side of the team would have to make calls from home and would not have access to the company’s communication technology in the office. For routine work, this time difference could be an advantage, allowing the Asian partner to advance projects after the US scientists had gone home. But when problems or questions arose that required more immediate action, individuals sometimes became frustrated with the long delays in getting an e-mail response. Even when video conferences could be scheduled they were not viewed as a full substitute for in-person meeting. It was in part for this reason that US managers still traveled several times/year to Asia to spend time with their partner.
Logistics Offshoring work to developing nations makes coordination particularly challenging (Parker and Anderson, 2002). Given the complexity in the R and D process, coordinating the know-how, processes, resources, personnel, timelines, etc has the potential to reduce much of the added value that the partnership was intended to generate. Initially the Asian firms faced delays and added costs in getting all of the materials needed for their experiments, but as the
Offshoring Knowledge Work: How Far Can it Go? Evidence from Drug R and D
9
industry has expanded in these countries, the leading global reagents and material suppliers have all set up operations in these nations. The shipment of chemical and biological materials across national borders, however, can still create obstacles, and has forced the US firms to organize the work process so they are not offshoring any mission critical steps on key projects. This might slow down the work as one informant commented: Shipment of bacterial strains that are drug resistant was a big problem. We had to jump through hoops. It does not matter how well you do the paper work, they will get stuck in customs. One package sat for a month... And now, I requested them to ship back two strains so we can reproduce them for internal testing and they can’t get it out of the country. US customs is more strict on the incoming stuff. We have alternatives. I don’t like it, I’d prefer to work in the original strain, but we can make it work.
Cross-Cultural Differences Cross-cultural managerial issues have long been a part of managing the innovation process in most US bioscience start-ups, since the scientific workforce graduating from US universities is very international. In our relatively small sample firms, for example, over a dozen different nationalities were working side-by-side in the R and D labs. Contracting work to India and China can create additional challenges, such as language barriers and different norms of communicating expectations and progress (Anderson, Davis-Blake and Parker, 2007). A common theme was that the eagerness of the suppliers to get and retain the business created the risk of exaggerating capabilities and failing to notify the partner right away when inevitable problems occurred with experiments. When asked about potential issues, the Indian firm’s first response was typically “we have got it under control,”, which was not always the case as one manager from Small Pharma commented: That's one of the problems, with India, they say they can do everything! But it’s a difference of what they think they can do and what they really can do. You have to think about what do you need help with and make clear you need to know if problems arise.
These communication challenges seem to create a need for increased formalization of procedures, processes and expectations in off-shoring due to the lack of close supervision. This is not surprising since many of the CRO relationships are somewhat new and emerging which implies a great deal of learning required from both sides. For the CRO, it means to learn what is expected, how the work should be done and in what form to report it. And from the company’s side, learning involves understanding how to manage from a distance and establishing and enforcing reasonable expectations.
Intellectual Property The potential loss of intellectual property (IP) has traditionally been a major concern for companies considering outsourcing to India or China (e.g. Carson and John, 2007). This concern would appear to be magnified in the drug industry, where IP is a vital source of competitive advantage. Surprisingly, our firms indicated that for the types of work they were contracting, loss of IP was not a major concern. This was in part because they had proceeded
10
David Finegold, Niclas Erhardt and Mari Sako
cautiously, starting with small assignments while they built trust with their partners. This is reinforced by the social networks within this industry and the strong reputational effects, which create a disincentive for a firm to cheat as one project manager from Small Pharma described: We have had some discussions around IP and are comfortable with our CRO. They now have a huge relationship with a large pharma company that we know as well; they are not going to screw us over, or it would go down in flames…And realistically, it’s the same risk we face if people leave our company; we have as much protection with our internal employees. It’s not much difference there.
A similar response was given by a manager at Biologics: IP? I'm not paranoid about this. The external vendors are very careful to manage their reputations to make sure things don’t leak out. The real issue is not the chemistry (used to produce inputs into the discovery process); the worry is more about knocking off the end product so that they can start producing it.
The firms also used a number of specific techniques to protect their IP including: code words, secure FTP servers for protected file transfers to share large data files inherent in development work, and compartmentalizing the work. For example, while the CRO was working on a compound to screen and identify promising molecular structures, its employees were not provided the knowledge of the molecule’s actual purpose, thus minimizing the risk if the compound was to leak to a competitor.
Enablers for R and D OffShoring and Outsourcing Many of the barriers noted at Small Pharma and Biologics are common to other companies that have offshored knowledge work. Prior research suggests some strategies for coping with these issues which were tried by our two companies. For example, Tarakci, Tang, Moskowitz and Plante, (2006) suggest adopting incentive contracts to minimize coordination problems, by encouraging the contractor to operate in a way that seeks to take proactive actions to correct problems and to create a win-win situation by optimizing total profit for both firms. Another mechanism to address these challenges is to employ personnel dedicated to managing outsourcing relationships, referred to as “boundary spanners”. Boundary spanners are staff whose primary task is to integrate, coordinate, and manage work across functional, company and national boundaries. Anderson et al. (2007a) found boundary spanners were particularly important for effective distributed product development to reduce potential language barriers and cultural differences. In our sample, both parties in the outsourcing relationship used boundary spanners. In the Asian contractors, there were key managers of each project, who were PhDs, often educated in the US or UK, who had strong English skills and spent time at the client getting to know the key people and process so that they could translate it to their workforce back home. Likewise, the US partners assigned a point person, often someone who was himself a native of the contractor country, to act as a point person to work closely with the CRO to oversee the project, to travel to Asia to provide onsite training
Offshoring Knowledge Work: How Far Can it Go? Evidence from Drug R and D
11
and knowledge transfer, and to track statistics to check quality of data and costs. The main focus is not the cost per se, but rather the cost effectiveness to assure that the company really achieves the value added.
Discussion Evidence from our case studies suggests that the framework set out in Figure 1 appears to provide a useful guide for understanding the offshoring and outsourcing R and D decisions of drug companies. Both Biologics and Small Pharma were more likely to shift those aspects of the research process to an offshore partner where the work was not seen as core to the critical development path, where the work could be modularized, and where comparable talent was available with a significant cost advantage in offshore locations.
Cost and Availability of Talent The core driver of the decision to offshore research in our cases was clearly labor cost savings. Unlike past global movement of R and D, where many international firms set up R and D operations in the US to tap into the supply of talent and to gain access to the world’s most lucrative drug market, in this wave of offshoring the goal is to try to make limited R and D dollars go further by spending less for comparable quality work. In chemistry, the prime driver of offshoring was to lower the cost for key labor intensive, yet less time sensitive tasks, while in biology, the much lower cost per project allowed US firms to pursue lower priority projects that otherwise it could not afford, in hopes that some will yield promising results that can attract additional resources.
Task Modularity In order to outsource and offshore work effectively, a degree of modularity must occur in the work process. A central question that we wanted to explore in the study is whether modularization of work preceded outsourcing or was driven by the requirements of outsourcing. In both of our cases, it appears to be the latter, but the way in which work was modularized, differed substantially between the two scientific areas based on the nature of the work. In chemistry, the CRO’s were asked to produce discrete batches of compounds in certain quantities, one piece of a multi-step process. In biology, CRO’s were given whole projects, but ones that were, as noted, small and not seen as core to the success of the company. Contracting out of biology was at a far less mature stage than chemistry. This appeared to be due to all three factors in our framework: the work is highly integrated and uncertain, and thus more difficult to modularize. The integrated nature of the tasks required the contracting researchers to know the target under exploration, thus making it harder to protect IP for core work. And the tacit expertise for commercial discovery biology is much harder to find in India and China at this point, as one of the informants from Biologica commented:
12
David Finegold, Niclas Erhardt and Mari Sako The problem is that they [CRO’s in India and China] lack the experience in the drug discovery. They don’t have the know-how that you need. They can follow instructions to make a drug part of the development process but they don’t have the softer knowledge you need to make decisions along the way. The chemist area is more developed.
In contrast, creating and testing chemical compounds is far more standardized and labor intensive, which makes it easier and more attractive to contract out to take advantage of differences in labor cost. The routine nature of work reduces the need for interdependent collaboration, since clear guidelines can be identified and enforced. Less interaction is necessary to solve problems. The most standardized work, however, has been automated on expensive, high throughput machines, and is generally retained in-house because labor cost advantages are outweighed by transaction cost disadvantages. While certain elements of the development process are thus more prone to be contracted out, the evidence suggests that contracting R and D has evolved gradually, in an opportunistic fashion, rather than being part of a pre-defined sourcing strategy. It was the chance to take advantage of these lower cost suppliers that drove changes in the design of the work process, rather than a decision to modularize the way the work was performed in the US creating the opportunity to outsource. As one manager from Small Pharma noted, they started by asking: What could they do for us? We wanted to see how they performed. We have increased their responsibilities as they showed performance. We also want to develop good working relations with them. Are they just CRO or a collaborator that will impact our work?
Core Knowledge Our case study firms continue to keep the work that is seen as most critical – in terms of protecting core knowledge and time sensitivity – in-house in the US. For the elements of R and D that are moved to India and China, three factors appear to explain why the firms are not overly concerned with the loss of intellectual property: 1) the work that is being performed offshore does not involve core IP: the specific chemicals being synthesized do not reveal the drug targets of interest and the biology projects are not the key ones in the drug pipeline; 2) they are using a series of standard control and security measures to protect information; and 3) perhaps most important, in the closely knit drug industry, reputational effects are vital, particularly for new start-up service companies. The US firms are confident that their partners will do everything possible to protect their IP, because any leakage would become known and jeopardize their ability to attract and/or retain much larger pharmaceutical clients. As the relationship has progressed, these factors were reinforced by a form of implicit institutional trust that evolved between the partners. Mangers made it clear that traveling to the CRO’s periodically was vital for developing this trust and a healthy working relationship.
Conclusion We have offered a framework for understanding what types of drug R and D work is likely to remain in US biotech companies and what is likely to be outsourced and offshored. Our study also helps to explain why this is occurring and likely to increase in the future and provides insight into how firms can manage these relationships effectively. It appears that in
Offshoring Knowledge Work: How Far Can it Go? Evidence from Drug R and D
13
this case, unlike some other sectors of the economy, the desire to access lower cost labor outside the US is leading to the modularization of the work process, rather than modularization occurring as a more effective way to perform bioscience R and D, some of which can then be outsourced or offshored. It is interesting to note that in an industry where IP has historically been seen as essential to creating value, that the loss of knowledge is not perceived as a major barrier for offshoring or outsourcing. While we have attempted to shed some light on the emerging trend of offshoring of R and D in the biotech industry, we clearly do not know the long term impacts of this trend. As the Indian and Chinese firms continue to develop their capabilities, moving from more routine chemistry to discovery biology, it would seem plausible that their current relationship as “service organizations” to US and European biotech companies, could evolve toward becoming more of a collaborative partnership in which drugs are co-developed. This might ultimately promote more cost effective drug development and foster more therapies for neglected diseases of the developing world (Finegold, Shakti, and Shahi, 2005). We already observed some evidence of such win-win relationships, where promising results from exploratory work in India that the firm could not have afforded in the US has enabled a novel antibiotic to move forward in development, generating more jobs in both Asia and the US. As this trend toward global movement of bioscience R and D is likely to grow, it will be important to explore the implications for firms and the workforce. One possibility is that US firms will set up their own research operations in India or China. While this may be more feasible for large pharmaceutical firms, it is a possibility that Small Pharma’s leadership team is already considering, led by a manager in the US who would be willing to return to his native India. While there are major challenges with operating effectively in the Indian market, this would potentially allow better control of more proprietary work, enhance communication, and cut out the costs of the middleman. We may also observe mergers between leading players in the different countries, such as the recent announcement of an offer by DaiichiSankyo, the second-largest Japanese pharmaceutical company to take a controlling interest in one of India’s largest bioscience firms, Ranbaxy (Krauskopf, 2008). If offshoring of R and D continues, and if more offshoring relationships were to become collaborative partnerships, it also has important implications for the education and training and labor market for US scientists. It seems to offer enhanced prospects for the select few who are doing the most cutting-edge research. Different career ladders may be created with new types of jobs for those capable of managing complex global relationships, which will require new ‘boundary spanning’ skill sets. But more prevalent R and D offshoring also suggests the destruction of face-to-face apprenticeship of junior scientists undertaking routine work, and the threat of growing wage pressure and/or loss of routine jobs in the US. It might become increasingly difficult to regard advanced science qualifications as tickets to good jobs.
References Anderson, E. G., Davis-Blake, A. and Parker, G. G. (2007). Managing Outsourced Product Design: The Effectiveness of Alternative Integration Mechanisms. Working Paper. Anderson, E., Davis-Blake, A., Erzurumlu, S., Joglekar, N., and Parker, G. (2007a). “The Effects of Outsourcing, Offshoring, and Distributed Product Development Organization
14
David Finegold, Niclas Erhardt and Mari Sako
on Coordinating the NPD Process,” Chapter in Handbook on New Product Development, Eds. C. Loch, S. Kavadias. Aron, R. (2008), Global sourcing of knowledge-intensive services: Operational risk, complexity and the demand for skilled workers in the U.S., paper for Sloan Industry Studies Annual Conference, Boston, MA, May 1-2. Blinder, A. S. (2006). Offshoring: The Next Industrial Revolution? Foreign Affairs, March/April. Carson, S. j. and John, G. (2007). A Transaction Cost Explanation of Property Rights Sharing in Outsourced Research Development and Engineering Relationships. Working paper. Dahlman, C. J. (2005). China and India: Emerging Technological Powers. Issues in Science and Technology, Spring: 45-53. Dyer, J. H. and Nobeoka, K. (2000). Creating and Managing a High-Performance Knowledge Sharing Network: The Toyota Case. Strategic Management Journal, 21(3): 345-367. Finegold, D., Shakti, D. and Shahi, G. (2005). “A New Variety of Capitalism: Case Study of the Emerging Business Models Indian Bioscience Industry,” paper presented at the EGOS Conference, Berlin, Germany, July. Kinder, T. (2003). Go with the flow – a conceptual framework for supply relations in the era of the extended enterprise. Research Policy, 32: 503-523. Krauskopf, L. (2008). Daiichi-Ranbaxy May Signal Big Pharma-Generic Deals,” Reuters, http://in.reuters.com/article/businessNews/idINIndia-34037120080612. Kumar, S. and Eickhoff, J. H. (2006). Outsourcing: When and how should it be done? Information Knowledge Systems Management, 5: 245-259. Norus, J. (2006). Building Sustainable Competitive Advantage from Knowledge in the Region: The Industrial Enzymes Industry. European Planning Studies, 14(5): 681-696. Parker, G. G. and Anderson, E. G. (2002). From Byer to Integrator: The Transformation of the Supply-Chain Manager in the Vertically Disintegrating Firm. Production and Operations Management, 11(1): 75-91. Quadros, R., Consoni, F. and Quintao, R. (2005). R and D outsourcing to research institutions: a new look into R and D in teh Brazilian automobile industry. Paper presented in the 13th GERPISA International Colloquium. Quinn, J. B. (1999). Strategic Outsourcing: Leveraging Knowledge Capabilities. Sloan Management Review, Summer: 9-21. Tarakci, H., Tang, K., Moskowitz, H. and Plante, R. (2006). Incentive maintenance outsourcing contractss for channel coordination and improvement. IIE Transactions, 38: 671-684. Ward, S. (2004). Outsourcing research: What is your position? Business Information Review, 21(4): 227-239.
In: Outsourcing, Teamwork and Business Management ISBN: 978-1-60456-956-8 Editor: Karl E. Carettas, pp. 15-30 © 2009 Nova Science Publishers, Inc.
Chapter 2
MULTINATIONAL EXPLORATION OF ACQUIRED R&D ACTIVITIES Jens Gammelgaard* Copenhagen Business School, Department of International Economics and Management Porcelænshaven 24, 2000 Frederiksberg, Denmark
Abstract This paper presents the results of a survey of 54 Danish multinational corporations that have acquired activities abroad. The role of the acquired R&D units was the focus of the survey, particularly with respect to the schism between basic and applied R&D, and the schism between autonomous and network R&D. This paper establishes the connection between a multinational corporation that follows a capability-motivated acquisition strategy and the R&D role new subsidiaries should play in order for the acquired resources to be utilized corporation-wide. Statistical findings reveal the need to follow a combination of basic and network-oriented R&D activities when focusing on capability development.
Keywords: Acquisition; Research and Development (R&D), Basic R&D, Applied R&D, Autonomy, Network, Capabilities.
Introduction Is it possible to advance capability-creating processes in a multinational corporation (MNC) through the acquisition of another firm’s R&D activities? The MNC’s awareness of the acquired firm’s R&D activities and subsequent use of integration strategies influence the utilization and exploration of the acquired resources. Awareness reflects an acquiring MNC’s initial intention to explore the acquired firm’s capabilities instead of following other strategic goals, such as market access. Integration indicates the MNC’s choice among possible roles for its new R&D units, including the need to address the schism between autonomy (a situation where the R&D unit does not cooperate with other MNC R&D units) and a network *
E-mail address:
[email protected]
16
Jens Gammelgaard
model (a situation which emphasizes interdependencies among different MNC R&D units). The nature of the R&D activities is also topical in terms of capability development. The strategic possibilities are again two-fold, since the acquired R&D unit can be assigned to applied R&D activities, i.e. modifications of existing products, or it may create new products or processes, a strategy that approaches basic research. The purpose of this paper is, therefore, to analyse whether the initial strategic wish to gain access to another firm’s capabilities, and the subsequent strategies concerning integration and the acquired firm’s R&D role are related. The architecture of the arguments is as follows. The following section briefly treats acquisition motives and emphasizes the importance of access to the acquired firm’s R&D activities as a reason for acquisition. The subsequent section stresses the different roles an acquired R&D unit is allowed to play in an MNC. Discussions of methods and measurements are followed by the presentation of results from a survey that covered 54 Danish acquisitions abroad in the period from 1994 to 1998. Finally, conclusions are presented.
Acquisition Motives Cisco Systems gained access to specific R&D capabilities within the Internet server and communication equipment fields through acquisitions. Corporations like Intel, General Electric and Nestlé all initiated technology-driven acquisitions during the 1990s as a vehicle to develop capabilities (Bower, 2001; Mitchell & Capron, 2002; Ranft & Lord, 2002). However, Gammelgaard (2004) found that earlier surveys of mergers and acquisitions (M&A) motives were restricted to include only resource exploitation strategies: investigating the direct outcome effect such as increased market shares, cost reductions and risk minimization through diversification. Wernerfelt (1984) extended this approach by using the resource-based view of the firm to put an emphasis on the acquired firm’s resources and their exploration. Here, acquiring firms followed a long-term oriented goal of creating value from utilising and improving the resources and capabilities of the acquired firm in a corporate-wide setting. In this respect, resources can be defined as anything tangible or intangible controlled by the firm that enables it to conceive of and implement strategies that strengthen or weaken its ability to create, produce and offer goods and services to a market (Wernerfelt, 1984; Barney, 1991; Sanchez et al, 1996). Christensen (2000) defined capabilities as functional, operational or technical superior capacities that may be further subdivided into specific individual skills or specialised team-based resources. Through acquisitions, the acquiring MNC gains access to the skills of the employed R&D engineers (see Nelson & Winter, 1982) and organisational learning processes, i.e. the social interaction of R&D engineers that result in new knowledge and products (Teece et al, 1997). The M&A literature seldom stressed access to the acquired firm’s R&D activities as a main motive. Different surveys investigating M&A motives clearly pointed to the growth of the firm through the extension of existing markets or the entering of new markets as the dominant motive (Newbould, 1970; Baker et al, 1981; Lindgren, 1982; Hunt et al, 1987; Suverkrup & Hauschildt, 1990; Davis et al, 1993; Norburn & Schoenberg, 1994). Furthermore, Ansoff et al (1972) found the completion of product lines through M&A, which made it possible to offer customers a full line of services, to be momentous. Additionally, Chakrabarti et al (1994) saw cost reductions through scale or scope economies as an important motive. Chakrabarti et al (1994) did highlight the capability perspective as among
Multinational Exploration of Acquired R&D Activities
17
the most important motives. Finally, Belderbos (2003) established that a major motive for Japanese manufacturing MNCs to acquire abroad was to gain access to R&D capabilities at a faster pace. Presumably, firms tend to focus on the capabilities embedded in the acquired firm. According to Serapio et al (2000, p. 2), MNCs now accelerate their direct investments in overseas R&D since “more than 100 multinational companies have acquired multiple laboratories abroad and are increasingly tapping into these laboratories for new sources of technologies”. Weston et al (1999) provided the example of M&A in the global chemical industry. They highlighted the motive of gaining access to key scientists in the acquired firm, who in turn were used for development of particular R&D programs, and pointed to the creation of broader technology platforms at the higher strategic level. However, surveys of acquisition motives have rarely touched on the acquired firm’s R&D activities as a strategic motive. Sometimes, access to R&D activities was only a motive subordinate to a strategic desire for cost reductions through economies of scale (Hughes et al, 1980) or through avoidance of duplicate efforts. Cooke (1986) spoke for the full, or more efficient, utilisation of intangible resources, such as specialists or high-tech equipment. This discussion is often associated with the synergy approach, where combinations of, for example, technical expertise embedded in one firm and manufacturing knowledge in the other create added value (Capron & Mitchell, 1998). In addition, Hagedoorn & Duysters (2002) found that a strategic and organizational fit between companies improved technological performance in general. Other contributions more directly connect acquisition motives and the R&D activities of the acquired firm. Dettmer (1963) focused on this perspective by addressing access to better and complementary unexploited technology in the acquired firm as an acquisition motive. Later, Chen & Su (1997, p. 73) highlighted the motive of “seeking of technological advantages or knowledge capital of a takeover target”, making it top priority on their motive list. More recently, Bower (2001) stated that access to the acquired firm’s R&D activities was one of five acquisition motives emphasized. Acquisitions were, in this context, a substitute for in-house R&D activities, and helped the acquiring firm to quickly build up positions in highly competitive and dynamic markets. Finally, Chakrabarti, et al, (1994) provided a closer look at the technological perspective of acquisition by identifying one cluster of “technological acquirers” that strategically sought new technology and know-how. This aim was achieved through close cooperation between the two R&D departments subsequent to the acquisition, which resulted in the redeploying of R&D resources into more productive uses. Using this line of argumentation, one can assert that a knowledge-seeking acquisition is positively related to integration strategies, based on the interdependencies between the acquired firm, its headquarters, and other subsidiaries.
Characteristics of R&D Activities Ranft (1997) argued that R&D activities were often the driver in a firm’s innovative processes. A major challenge for acquiring firms has, therefore, been to tap into and explore the acquired firm’s R&D resources and capabilities. Håkanson & Nobel (1993) suggested that the acquired unit’s R&D activities be expanded by assigning it a group-wide responsibility within specific areas. Decentralised and loosely coupled networking organisations (Hill et al, 2000), theoretically describable as a ‘heterarchy’ (Hedlund, 1986), a ‘transnational
18
Jens Gammelgaard
organisation’ (Bartlett & Ghoshal, 1989), or a ‘differentiated network’ (Nohria & Ghoshal, 1997), have typically used this strategy. In decentralised MNCs, subsidiaries have acted very independently by being responsible for capability development within specified areas. In the end, these are assessed by other MNC units as highly important (White & Poynter, 1984; Birkinshaw & Hood, 1998; Holm & Pedersen, 2000; Frost et al, 2002). Subsidiaries have played a wide range of roles within MNC’s (Schmid, 2000). This paper focuses on the role of the acquired firm’s R&D unit and therefore emphasizes a unit acting as a centre of gravity that significantly contributes to the entire MNC capability development (Chiesa, 1995; Brockhoff, 1998). In terms of R&D activities, the subsidiary can play different roles that more or less qualify for specific mandates and positions. The R&D unit may, on the one hand, concentrate on pure capability creation processes that do not specifically relate to a certain product. On the other hand, the unit may centre on more product-oriented activities, in which the R&D unit only pays attentions to pure modifications of the headquarters products to fulfil local market demands. Secondly, the R&D unit can be very autonomous in its behaviour when R&D activities take place independent of other R&D activities in the MNC. In contrast, the R&D unit might be fully integrated with other R&D units of the MNC, so that R&D activities come about interdependently with other corporate units. Figure 1 illustrates these two spectres by including the four archetypical roles of a subsidiary’s R&D unit in an MNC.
Figure 1. Roles of an R&D unit in an MNC subsidiary
The purpose of following a combination of an applied and autonomous R&D strategy is to customise MNC developed products to meet local customers’ specific demands. To fulfil this role, the R&D unit modifies headquarters’ or other sub-units’ products without depending on additional resources from other MNC units. Consequently, the final product is often not usable (or saleable) in other MNC units or in their related markets. In the combination of network and applied R&D activities, the subsidiary R&D unit is responsible for modifying existing products in cooperation with headquarters or other sub-units, so the final product design meets global demands. Development of subsidiary-specific capabilities, as in the situation of an autonomous and basic R&D strategy, requires concentrated R&D activity without control or influence from other R&D units in the MNC. The outcome of such a
Multinational Exploration of Acquired R&D Activities
19
strategy is the creation of unique and disparate capabilities, which will be considerably different from the resources and capabilities located in other MNC units. A negative effect here is the substantial degree of tacitness and inimitability in the underlying knowledge of the developed capabilities. Consequently, the capabilities will probably not be utilised in other parts of the MNC. The mixture of network and basic R&D is, therefore, preferable for building MNC relevant capabilities, because the R&D unit embraces the advantage of being inspired by other R&D units in its creation and development of capabilities. In this combination, the developed capabilities will still be unique, while at the same time, they will fulfil the requirements of other MNC units. Literature addressing capability-creating R&D activities in MNC subsidiaries often classifies basic R&D activities at the top of the hierarchy of different subsidiary R&D roles. The “corporate technology unit”, where subsidiary-generated new technology, which is of a long-term or exploratory nature for either headquarters (Ronstad, 1978) or in general (Taggart, 1998), has been an example of a top-hierarchical position. However, Medcoff (1997) and Nobel & Birkinshaw (1998) found that basic R&D was not necessarily usable for other corporate units in the short term or even in the medium run. According to these researchers, the purpose of basic R&D is to discover new platforms of knowledge, which is not specifically associated with particular products. Zander (1999) considered basic R&D to be an exploration strategy focusing on new insights and fields of expertises by developing certain items of knowledge not producible elsewhere in the corporation or, at times, elsewhere in the industry. The end goal of the subsidiary’s R&D unit was, therefore, to develop knowledge considerably different from existing MNC knowledge. Gerybadze & Reger (1999) pointed to a trend of more foreign R&D sites being assigned the role of creator of basic technologies. Conversely, surveys by Papanastassiou and Pearce (1999), and Pearce (1999) revealed that basic R&D activities in subsidiaries were of only minor importance. Further, Florida (1997) demonstrated that applied R&D had a much higher importance when it encompassed the aim of creating commercial concepts. The other aspect investigated in previous research is the division between autonomous and network R&D strategies. Persaud et al (2002, p. 61) defined autonomy as: ”the degree to which a subsidiary R&D lab has control over the strategic decisions affecting its direction and operations”. The authors further claimed that autonomy positively effected innovation, by leading to greater freedom in developing unique relationships with both internal and external partners, although the risk of opportunistic behaviour - the subsidiary R&D unit following its own research goals rather than the goals defined by headquarters - was present. Furthermore, the writers further statistical support for autonomy being dividable into two areas: the freedom to choose with whom one establishes relationships, and the freedom to select which areas the R&D unit was to do research. Reflecting the scope between autonomy and network, Brockhoff (1998) described three standards of R&D roles: the ‘hub’, the ‘competence centre’ and the ‘network model’. In a hub, decision-making was centralised and headquarters co-ordinated all other R&D laboratories. In the competence centre, R&D activities were experimental, isolated and specialised. Finally, the network model was characterised by high intra-organisational integration and intensive subsidiary involvement in the formulation and implementation of strategies. Likewise, Birkinshaw (2002) divided R&D roles into the ‘integrated network’, where R&D centres were tightly, and the ‘loosely-coupled network’, in which laboratories were given autonomy positions and specific roles to fulfil. Birkinshaw (2002) recommended the integrated network solution when the underlying R&D
20
Jens Gammelgaard
knowledge was not easily observed and the loosely-coupled network when assets were characterized by a low degree of mobility (the extent to which the knowledge base could be separated from its physical setting). In general, Chiesa & Manzini (1996) advised that headquarters govern R&D units if the network model is implemented. As suggested by Birkinshaw and Hood (2001), this would allow headquarters to distribute resources by delegating mandates and supporting initial “genius” knowledge creation while, at the same time, enabling the blocking of unprofitable research programmes. The surveys investigating the success rate of network-based R&D activities have not provided a clear picture. Brockhoff & Schmaul (1996) and Ensign et al (2000) spoke of positive success rates, whereas Taggart (1997) and Taggart & Hood (1999) took the opposite position. Several MNCs have used the network-oriented structure, making R&D units dispersed in different countries responsible for certain product or technology areas. For example, Chiesa (2000) described how Nissan, the Japanese car manufacturer, developed a minivan as an outcome of cooperation between headquarters and different US-located R&D centres in California, Michigan and Tennessee. Downey (2003) provided the example of Nokia having 18,000 engineers scattered across 69 sites worldwide. Furthermore, he described the development of the 777 aircraft in which Boeing operated with 238 crossfunctional teams, including customers, operators and line mechanics in collaborative design networks. Birkinshaw’s (2002) case study of Ericsson’s Radio System business reflected on this complexity by portraying how the firm developed its third generation mobile telephony by involving 10,000 engineers located in at least 20 different R&D sites around the world. Finally, Gassmann & Zedtwitz (1999) demonstrated how the Schindler Group, at present a worldwide leader in escalators and elevators, built up its R&D capacity through acquisitions, where acquired firms took specific positions in a highly integrated organisation. Today, the company employs around 500 engineers working in several R&D centres around the world. Through both intraorganisational cooperation, and close collaboration with local science centres and universities, they have developed the complex technology behind the elevator keypads that operate the car based on the number, location and destination of waiting passengers together with aerodynamic influences. The network system makes sure that the development of such unique components is usable worldwide. The question then becomes which strategy the acquiring firm should emphasize when integrating the acquired firm if: (1) it is a capability-based acquisition, and (2) the purpose is to create synergy subsequent to the acquisition. Haspeslagh & Jemison (1991) highlighted the risk of simply absorbing the acquired firm, emphasizing value destruction caused by key employees leaving the firm. Conversely, preserving the acquired firm by leaving it in an entirely autonomous position will not lead to synergy in the long run. Haspeslagh & Jemison (1991) proposed to start the integration process with a preservation strategy, and build a symbiotic approach over time. This becomes a kind of networking organisation where the acquired firm keep elements of autonomy while it is simultaneously absorbed. Gassmann & Zedtwitz (1999) also recommended that the acquired firm reserve some degree of autonomy, but the unit must be forced into a corporate network at the same time, creating the opportunity to improve MNC knowledge. Håkanson & Nobel (2001) advised strong integration, since the acquired firm typically favoured transfers of developed technology to the parent organization - apparently since transfers of tacit and complex knowledge has been interpreted as easier within hierarchies (Kogut & Zander, 1995; Almeida et al, 2002).
Multinational Exploration of Acquired R&D Activities
21
The discussion leaves two dilemmas. One is the choice between autonomous and network-oriented R&D, while the other is the availability of the two alternatives: basic or applied R&D. The challenge is to find the right balance when walking the tightrope between the different strategies. A recommendable strategy is to bring in elements from all four extremities, but still emphasising the network and basic R&D. Using this model, the capability-based acquired firm keeps its superiority in terms of a specific knowledge or technology, but at the same time this capability will be of use elsewhere in the corporation. The model advocates a more cohesive approach to decentralisation and networking, a proposal that corresponds to a recent finding of Gerybadze & Reger (1999) concerning MNC organisational structures. Based on this discussion, I hypothesize: Hypothesis: An acquired capability-based R&D unit focuses its activities on basic and network-oriented R&D.
Methods and Measurements Data for this study was collected through a questionnaire survey undertaken in the spring of 2000. The questionnaire was sent to those Danish industrial firms that acquired a foreign firm in the period from 1994 to 1998, during which 151 Danish MNCs acquired 469 firms abroad. Three mailings by post and a follow-up phone call to non-responding firms resulted in 54 returned questionnaires bringing about a response rate of 35.76 percent. A bias control of the responding acquiring firms compared to non-responding firms, including figures on numbers of acquisitions made in the survey period, investments countries, the year of establishment, numbers of employees at the end of 1993 and the end of 1998, and corporate turnover in 1993 and 1998, showed no bias of significance when comparing mean values using a one-tailed t-test. The questionnaire primarily contained questions concerning factual figures, such as the percentage of R&D cost compared to turnover, and questions to be answered on a 1 to 7 point Likert scale. This section presents the descriptive data and statistical analysis based on a t-test. The purpose of using the t-test was to distinguish the group behaviour of capability-based acquiring firms from an opposing group of acquiring firms that solely follow growth and market-related strategies. The partitioning of observations was based on a non-hierarchical clustering method resting on a random selection of five variables’ seed points covering different aspects of the acquired firm’s capabilities. The purpose of the cluster analysis was to group objects based on the characteristics they possessed, including high internal homogeneity on the one hand and high external heterogeneity on the other. Different clustering tests were run for selecting the procedure leading to the highest degree of external heterogeneity. The clustering of firms was based on recommendations from Hair et al (1998) and Der and Everitt (2002). This paper elucidates basic characteristics of the capability-based acquisition building on the resource-based view framework of Wernerfelt (1984) and Barney (1991). The five variables selected for clustering are: (1) the importance for the acquiring firm of gaining access to the acquired firm’s capabilities; (2) the importance for the acquiring firm of gaining access to the acquired firm’s relations to local science centres; (3) the inimitability of the acquired resources; (4) the non-tradability of the acquired resources; and (5) the uniqueness
22
Jens Gammelgaard
of the acquired resources. The interrelatedness between the five factors was satisfactory with a Crombach Alpha Coefficient of 0.78. Using a non-hierarchical clustering procedure assigned the clusters in regard to a specified number. In this case, the number was two. In this survey, the cluster seeds were randomly selected. Using this technique, one cluster of 22 capability-based acquired firms and another cluster of 32 market-based acquired firms emerged. The differences in means are presented in Table 1. Table 1. Mean differences: Five clustering variables using a non-hierarchical random selection clustering method Variable Access to capabilities Access to scientific centres Degree of inimitability Degree of non-tradability Degree of uniqueness
Mean capabilitiesbased cluster 5.79 3.59 4.06 4.50 5.41
Mean markedbased cluster 3.66 1.76 2.33 2.53 2.14
F – Statistics 26,70*** 17.29*** 30.91*** 29.63*** 71.52***
n = 53 Based on 1-7 Likert scale questions †p <.10; *p <.05; **p < .01; ***p < .001.
Five questions covered the aspect of network versus autonomous-based R&D with regard to R&D activities. The respondents were asked to assess the degree of integration measured on a 1 to 7 point Likert scale where 1=autonomous and 7=integrated. Furthermore, respondents were questioned on the degree to which the acquired firm could make its own strategic decisions concerning R&D. Third, respondents were asked about the degree of cooperation between the acquired firm’s R&D unit and other R&D units in the MNC. Finally, respondents were asked whether R&D-related knowledge transfers to and from other MNC units had taken place. To test the concept of basic versus applied R&D, respondents were asked to mark the importance of four different R&D activities using a 100% scale, where the percentage selected should demonstrate the relative weight. The four strategies were: (1) development of technological capabilities, (2) development of new product or processes, approaching basic R&D, and (3) development of existing products or processes, and (4) adaptation of MNC products to the local market demands, approaching applied R&D.
Results The acquiring Danish firms were typically medium-sized with less than 1000 employees, although a few firms were very large and internationalised, leading to a mean of 4877 total employees and a mean distribution of 1182 and 3695 employees in Denmark and abroad respectively. The typical Danish firm acquired less than one firm per year in the period, but again, some firms acquired more frequently. One firm acquired 74 foreign firms in 18 different countries. The acquired firms were often small, with a mean of 488 employees and turnover averaging US$ 30 million at the time of takeover. The acquisitions normally took
Multinational Exploration of Acquired R&D Activities
23
place in the nearby markets of Sweden, Germany, the Netherlands and the UK. Furthermore, the US was a popular investment country. R&D costs in proportion to turnover, and the number of R&D employees in proportion to the total number of employees in the acquired firm were also tested. The figures, as presented in Table 2, cover the period at the time of takeover (the acquired firms in the sample was acquired between 1994-1998) and 1999. The R&D activity level in proportion to turnover and employees was in the range of 4 to 5% in both periods. A t-test did not show any significant changes in the resources used between the two time periods. Table 2. Acquired firm R&D figures R&D Factor R&D in proportion to turnover R&D employees in proportion to total employees
Time of takeover Mean Variance 3.98 % 26.97 %
Mean 5.02 %
Variance 35.98 %
4.40 %
4.74 %
30.91 %
34.98 %
1999
N = 43
The high variances indicate a great disparity within the cohort of acquired firms showing a wide span going from no activity at all to an R&D cost at 30% of turnover. Likewise 16% of the firms specified that between 10% and 25% of employees were working in the R&D unit. The great variation is related to disparity within the group of participating firms and acquired units in terms of such as aspects as type of industry, target size, and firm age. One aim was to test whether the acquired capability-based R&D unit conducted basic or applied R&D. Table 3 provides an overview of the distribution in percentages in terms of four R&D roles. The development of technological capabilities covered the pure basic R&D, whereas the development of new products and processes included the same aspect in a modified form. Development of existing corporate products and processes, and pure adaptations of corporate products to fulfil local demands, belonged to the applied R&D activities. The figures in Table 3 show the percentage of the total activity. For example, development of technological capabilities counts for 10.91% of the total R&D activity among the capability-based acquired firms. The hypothesis, suggesting that the capability-based group would exhibit a higher degree of basic R&D, was only partly supported. Such activity covered 52% of the total activity in the capability-based group compared to 35% in the market-based group. The degree of basic technology research was almost equal when comparing the two groups. One reason could be that basic R&D processes taking place in firms are often only modified version of ‘true’ basic research in which the relationship to the end product is kept. Instead, if firms need particular elements of ‘basic knowledge’ they typically tap into the research located in local science centres - such as universities (Gulati et al, 2000; Rynes et al, 2001). Consequently, in this sample, responsibility within basic R&D related to direct product or process developments to a much higher degree, since such activities were significantly higher in the capability-based group. Both groups of acquiring firms highly prioritised the development of products in terms of new developments and further development of existing corporate products. The most important role of the market-based R&D units was to customise corporate products in order to
24
Jens Gammelgaard
meet local market demands, in contrast to the capability-based group in which applied R&D activities were only of minor importance. However, in the latter group, pure applied R&D was still more important than the basic R&D. Table 3. T-test: Differences in means concerning the role of the acquired R&D unit R&D factor
Mean Capability-based group
Mean Market-based group
Development of new technological capabilities Development of new products/processes Development of existing corporate products/processes Adaptation of corporate products to local demands
10.91% (10.43) 40.91% (18.36) 28.77% (17.91) 19.41% (12.89)
10.83% (13.62) 23.75% (21.76) 22.50% (18.79) 42.92% (39.97)
T-values 0.02 2.44* 0.96 1,98*
N = For the competence-based group and the outcome-based group respectively: 22,12 Based on a t-test assuming equal variances. The figures in the category of adaptation of corporate products are based on a t-test assuming unequal variances Standard deviation values are set in parentheses †p <.10; *p<.05; **p < .01; ***p < .001.
Table 4.T-test for differences in R&D activity mean values
Nature of R&D activity Degree of integration R&D unit co-operation R&D-related knowledge transfers from the acquired firm to other corporate units R&D-related knowledge transfers from headquarters to the acquired firm Acquired firm responsible for own R&D activities
Mean Capability-based group
Mean Market-based group
4.89 (1.64) 5.45 (2.79) 4.90 (2.09)
4.22 (3.24) 3.62 (4.85) 2.90 (3.19)
5.00 (3.17)
4.86 (3.63)
0.27
3.76 (2.32)
3.19 (1.96)
1.37†
T-values 1.45† 3.61*** 4.45***
N = For the competence-based group and the market-based group respectively: (27,18); (31,21); (31,21); (30,21) and (30,21), all based on t-test assuming equal variances Variance values are set in parentheses First question: 1 = autonomy position, 7 = integrated position Remaining questions: 1= to a low degree, 7= to a high degree †p <.10; *p<.05; **p < .01; ***p < .001.
The other main factor investigated was the degree of integration. The result of the t-test, as presented in Table 4, showed a significant difference between the two groups of acquiring
Multinational Exploration of Acquired R&D Activities
25
firms, though in absolute figures both clusters emphasised the network model. The firms in the capability-based group seemingly preferred an integrated status while still retaining some autonomy. The tendency in the market-based group was similar, but more firms favoured the pure archetypal solution, as shown through the higher variance. The extent of knowledge transfers between the acquired R&D unit and other MNC units further indicated a higher level of integration. The low variance in the capability-based group confirmed that the acquired firms, in general, transferred knowledge to other corporate units, while the market-based group only saw intensive knowledge transfers taking place occasionally. The high degree of knowledge transfers that took place in the capability-based group was associated with the high degree of underlying cooperation between the R&D units, as expressed by the high correlation coefficient between the two factors (0.63; p<0.0001). Regarding R&D collaboration, the difference between the two groups was again significant at a 99.9% level. The figures clearly showed that the ownership of resources on which other units depend creates an incentive for knowledge transfers. Furthermore, the R&D units in both groups of acquired firms were dependent on knowledge from headquarters, but probably for different reasons. In the case of the capability-based acquired firms, the continued exploration of capabilities was essential, whereas in the market-based cases the upgrading of resources turning them into capabilities was more important. In general, headquarters’ R&D units took an active part in the R&D processes of the acquired firm. Finally, the acquired firm’s mandate to make its own strategic decision concerning R&D was tested. Here, a higher level of autonomy was present in the capability-based group, in contrast to the aforementioned higher level of integration. However, possession of capabilities, upon which other units depend, typically paves the way for a subsidiary to win responsibility within the corporation – expressing the dilemma between integration and autonomy (Taggart, 1998). Regarding, the schism between network and autonomous-based R&D, the acquired R&D unit was supposed to be highly integrated while still leaving some scope of autonomous activities. Whether this was evident is hard to say. The t-test revealed a higher degree of integration and cooperation with other MNC units, but simultaneously indicated a higher degree of strategic responsibility and less dependency on headquarters. As in the case of basic versus applied R&D, the figures showed a tendency, but not unambiguous proof. A likely interpretation is that the intention of integrating the acquired firm turned out to be too difficult in practice. Organisational and national cultures, and the historically different trajectories of the two firms are likely to ruin the integration process (Shrivastava, 1986; Brockhoff, 1998). However, the desire to avoid destroying value through failure to properly preserve capabilities could be an alternative explanation (Haspeslagh & Jemison, 1991). Furthermore, the R&D capabilities of the acquired firm will typically be embedded in persons and organisational systems, leading to high transfer and integration costs. These complications might cause a situation where integration costs exceed value creation. Even though the acquiring firms have high expectations for a subsequent integration of R&D capabilities, the cost of doing so might have blocked the process.
Conclusion This chapter reflects on the relationship between the MNC’s strategic initial awareness of possibilities for capability-creation processes in the new subsidiary, and the nature of R&D
26
Jens Gammelgaard
activities in that particular subsidiary. The role of the acquired firm’s R&D unit reflects a strategic choice between independency and interdependency on the one hand, and basic versus applied R&D on the other hand. The question raised here is what impact the acquisition motive of gaining access to capabilities has on the R&D activities in the acquired firm subsequent to the takeover. The strategic motivation can be interpreted as a search for both capabilities and synergistic effects. The purpose, therefore, is to tap into another independent firm’s capabilities and make them usable for the different units of the acquiring MNC. However, the full utilisation of the acquired capabilities is only possible through the use of explorative and resource-combining strategies. The acquired R&D units must be given the needed mandates and resources to further develop new and unique capabilities. A focus on basic oriented research, with applicability in mind, is necessary. To make capabilities relevant, the network R&D strategy should be emphasized, although elements of autonomy should be encompassed, leaving room for uniqueness. This combination of R&D strategies is an emerging perspective and is in contrast to the traditional centralised MNC. A t-test analysis showed a possible relationship between the capability-seeking acquisition motive and a subsequent explorative integration strategy However, emphasis was put on product development rather than knowledge and technologies, which probably was an outcome of the inimitability and non-marketability characteristics of acquired resources. Basic-oriented capability developing processes were of low importance in this connection, and the MNC often accessed some alternative sources. To conclude, the capability-seeking acquiring MNC preferred a modified integration of the target firm’s R&D unit, which was in line with the well-known symbiotic integration strategy as proposed by Haspeslagh & Jemison (1991). This secured a preservation of the autonomous and basic-oriented R&D activities of the acquired firm and, at the same time, enforced the same activities in a corporate research network.
References Almeida, P., Song, J. & Grant, R. M. (2002). Are Firms Superior to Alliances and Markets? An Empirical Test of Cross-Border Knowledge Building. Organiation Science. 13(2), 147-161 Ansoff, B., Portner, & Radosevich, (1972). Twenty Years of Acquisition Behaviour in America - A Comparative Study of Mergers and Acquisitions of US Manufacturing Firms 1946-1965. London Baker, H. K., Miller, T. O. & Ramsperger, B. (1981). A Typology of Merger Motives. Akron Business and Economic Review, 12(4), 24-29 Barney, J. B. (1986). Organizational Culture: Can it be a Source of Sustained Competitive Advantage? Academy of Management Review, 11(3), 656-665 Barney, J. B. (1991). Firm Resources and Sustained Competitive Advantage. Journal of Management, 17(1), 99-120 Bartlett, C. A. & Ghoshal, S. (1989). Managing Across Borders - The Transnational Solution. London: Century Business. Belderbos, R. (2003). Entry Mode, Organizational Learning and R&D in Foreign Affiliates: Evidence from Japanese Firms. Strategic Management Journal, 24(3), 235-259
Multinational Exploration of Acquired R&D Activities
27
Birkinshaw, J. (2002). Managing Internal R&D Networks in Global Firms: What Sort of Knowledge is Involved? Long Range Planning, 35(3), 245-267 Birkinshaw, J. & Hood, N. (1998). Multinational Subsidiary Evolution: Capability and Charter Change in Foreign-owned Subsidiary Companies. Academy of Management Review, 23(4), 773-795 Birkinshaw, J. & Hood, N. (2001). Unleash Innovation in Foreign Subsidiaries. Harvard Business Review, 79(3), 131-137 Bower, J. L. (2001). Not All M&As Are Alike – and That Matters. Harvard Business Review, 79(3), 93-101 Brockhoff, K. (1998). Internationalization of Research and Development. Berlin: Springer. Brockhoff, K. & Schmaul, B. (1996). Organization, Autonomy, and Success of Internationally Dispersed R&D Facilities. IEEE Transactions of Engineering Management, 43(1), 33-40 Capron, L. (1999). The Long-Term Performance of Horizontal Acquisitions. Strategic Management Journal, 20(11), 987-1018 Capron, L. & Mitchell, W. (1998). Bilateral Resource Redeployment and Capabilities Improvement Following Horizontal Acquisitions. Industrial and Corporate Change, 7(3), 453-484 Chakrabarti, A., Hauschildt, J. & Süverkrüp, C. (1994). Does it Pay to Acquire Technological Firms. R&D Management, 24(1), 47-56 Chiesa, V. (1995). Globalizing R&D Around Centres of Excellence. Long Range Planning, 28(6), 19-28 Chiesa, V. (2000). Global R&D Project Management and Organization: A Taxonomy. The Journal of Product Innovation Management, 17(5), 341-359. Chiesa, V. & Manzini, R. (1996). Managing Knowledge Transfer within Multinational Firms. International Journal of Technology Management, 12(4), 462-476. Chen, C. & Su, R. (1997). Do Cross-Border Acquisitions of U.S. Targets Differ from U.S. Domestic Takeover Targets? Global Finance Journal, 8(1), 71-82. Christensen, J. F. (2000). Building Innovative Assets and Dynamic Coherence in MultiTechnology Companies. In Resources, Technology and Strategy – Explorations in the Resource-based Perspective, eds. N. J. Foss & P. L. Robertson. London: Routledge. Cooke, T. E. (1986). Mergers and Acquisitions. Oxford: Basil Blackwell Davis, E., Shore, G. & Thompson, D. (1993). Continental Mergers are Different. In European Mergers and Merger Policy, eds. M. Bishop & J. Kay. Oxford: Oxford University Press Der, G. & Everitt, B. S. (2002). A Handbook of Statistical Analyses using SAS. Boca Raton: Chapman & Hall/CRS Dettmer, R. G. (1963). Reasons for Mergers and Acquisitions. In American Management Association Corporate Growth through Merger and Acquisition. Management Report 75. New York Downey, C., Greenberg, D. & Kapur, V. (2003) Reorienting R&D for a horizontal future. Research Technology Management, 46(5), 22-28 Ensign, P. C., Birkinshaw, J. M. & Frost, T. S. (2000). R&D Centres of Excellence in Canada. In The Emergence and Impact of MNC Centres of Excellence – A Subsidiary Perspective, eds. U. Holm & T. Pedersen. Houndsmill, Bas.: MacMillan Press Florida, R. (1997). The Globalization of R&D: Results of a Survey of Foreign-Affiliated R&D Laboratories in the USA. Research Policy, 26(1), 85-103
28
Jens Gammelgaard
Frost, T. S., Birkinshaw, J. & Ensign, P. C. (2002). Centers of Excellence in Multinational Corporations. Strategic Management Journal, 23, 997-1018. Gammelgaard, J. (2004) Access to Competence – An Emerging Acquisition Motive. European Business Forum, 17(Spring), 44-47 Gassmann, O. & Zedtwitz, M. von (1999). New Concepts and Trends in International R&D Organization. Research Policy, 28(2/3), 231-250 Gerybadze, A. & Reger, G. (1999). Globalization of R&D: Recent Changes in the Management of Innovation in Transnational Corporations. Research Policy, 28(2/3), 251-274. Gulati, R., Nohria, N. & Zaheer, A. (2000). Strategic Networks. Strategic Management Journal, 21(3), 203-215 Hagedoorn, J. & Duysters, G. (2002). The Effect of Mergers and Acquisitions on the Technological Performance of Companies in a High-tech Environment. Technology Analysis & Strategic Management, 14(1), 67-85 Hair, Jr., J. F., Anderson, R. E., Tatham, R. L. & Black, W. C. (1998). Multivariate Data Analysis. Upper Saddle River, New Jersey: Prentice Hall Haspeslagh, P. & Jemison, D. (1991). Managing Acquisitions - Creating Value Through Corporate Renewal. Oxford: The Free Press Hedlund, G. (1986). The Hypermodern MNC – A Heterarchy? Human Resource Management, 25(12), 9-35 Hill, S., Martin, R. & Harris, M. (2000). Decentralization, Integration and the PostBureaucratic Organization: The Case of R&D. Journal of Management Studies, 37(4), 563-585 Holm, U. & Pedersen, T. (2000). The Emergence and Impact of MNC Centres of Excellence – A Subsidiary Perspective. Houndsmill, Bas.: MacMillan Press Howells, J. (1990). The Internationalization of R&D and the Development of Global Research Networks. Regional Studies, 24(6), 495-512 Hughes, A., Mueller, D. C. & Singh, A. (1980). Hypotheses about Mergers. In Determinants and Effects of Mergers. An International Comparison, ed. D. C. Mueller. Cambridge, Mass.: Oelgeschlager, Gunn & Hain, Publishers, Inc. Hunt, J. W., Lees, S., Grumbar, J.J. & Vivian, P. D. (1987). Acquisitions: the human factor. Egon Zehnder International. Håkanson, L. & Nobel, R. (1993). Foreign Research and Development in Swedish Multinationals. Research Policy, 22(5/6), 373-396 Håkanson, L. & Nobel, R. (2001). Organizational Characteristics and Reverse Technology Transfer. Management International Review, 41(4), 395-420. Kogut, B. & Zander, U. (1995) Knowledge and the Speed of the Transfer and Imitation of Organizational Capabilities: An Empirical Test. Organization Science, 6(1), 76-91 Lindgren, U. (1982) Foreign Acquisitions: Management of the Integration Process. Dissertation. Stockholm Business School Medcoff, J. W. (1997). A Taxonomy of Internationally Dispersed Technology Units and its Application to Management Issues. R&D Management, 27(4), 301-318 Mitchell, W. & Capron, L. (2002). Managing Acquisitions to Change and Survive. European Business Forum, 9(Spring), 51-55 Newbould, G. D. (1970). Management & Merger Activity. Liverpool: Guthstead Limited
Multinational Exploration of Acquired R&D Activities
29
Nelson, R. R. & Winter, S. G. (1982). An Evolutionary Theory of Economic Change. Cambridge: The Belknap Press of Harvard University Press Nobel, R. & Birkinshaw, J. (1998). Innovation in Multinational Corporations: Control and Communication Patterns in International R&D Operations. Strategic Management Journal, 19(5), 479-496 Nohria, N. & Ghoshal, S. (1997). The Differentiated Network: Organizing Multinational Corporations for Value Creation. San Francisco: Jossey-Bass Publishers Norburn, D. & Schoenberg, R. (1994). European Cross-Border Acquisitions: How was it for You? Long Range Planning, 27(4), 25-34 Papanastassiou, M. & Pearce, R. (1999). Multinationals Technology and National Competitiveness. Cheltenham: Edward Elgar Pearce, R. (1999). Decentralised R&D and Strategic Competitiveness: Globalised Approaches to Generation and Use of Technology in Multinational Enterprises (MNEs). Research Policy, 28(2-3), 157-178 Persaud, A., Kumar, U. & Kumar, V. (2002). Coordination Structures and Innovative Performance in Global R&D Labs. Canadian Journal of Administrative Science, 19(1), 57-75 Ranft, A. L. (1997) Preserving and Transferring Knowledge-Based Resources During PostAcquisition Implementation. Dissertation submitted for partial fulfilment of requirements for the degree of Doctor of Philosophy in the Kenan-Flagler School of Business. University of North Carolina, Chapel Hill. Ranft, A. L. & Lord, M. (2000). Acquiring new Knowledge: The Role of Retaining Human Capital in Acquisitions of High-tech Firms. The Journal of High Technology Management Research, 11(2), 295-319 Ronstadt, R. C. (1978). International R&D: The Establishment and Evolution of Research and Development Abroad by Seven U.S. Multinationals. Journal of International Business Studies, 9(1), 7-24 Rynes, S. L., Bartunek, J. M. & Daft, R. L. (2001). Across the Great Divide: Knowledge Creation and Transfer between Practitioners and Academics. Academy of Management Journal, 44(2), 340-355. Sanchez, R., Heene, A. & Thomas, H. (1996). Dynamics of Competence-Based Competition Theory and Practice in the New Strategic Management. Oxford: Pergamon Schmid, S. (2000). Foreign Subsidiaries as Centres of Competence: Empirical Evidence from Japanese MNCs. In Recent Studies in Interorganizational and International Business Research, eds. J. Larimo & S. Kock, Sören. Proceedings of the University of Vaasa, Reports 58, 82-204. Shrivastava, P. (1986). Post-merger Integration. Journal of Business Strategy. 7(1), 65-76 Serapio, M., Dalton, D. & Yoshida, P. G. (2000). Research Technology Management. 43(1), 2-4 Suverkrup, C. & Hauschildt, J. (1990). Innovations and US-German Acquisitions. In Management of Technology II – The Key to Global Competitiveness, eds. T. Khalil & B. Bayrakta. Norcross: Industrial Engineering and Management Press, 441-450. Taggart, J. H. (1997). R&D Complexity in UK Subsidiaries of Manufacturing Multinational Corporations. Technovation, 17(2), 73-82 Taggart, J. H. (1998). Determinants of Increasing R&D Complexity in Affiliates of Manufacturing Multinational Corporations in the UK. R&D Management, 28(2), 101-110
30
Jens Gammelgaard
Taggart, J. H. (1999). US MNC Subsidiaries in the UK: Characteristics and Strategic Role. In International Business Organization, eds. F. Burton, M. Chapman & A. Cross. Houndsmills: MacMillan Press Taggart, J. H. & Hood, N. (1999). Determinants of Autonomy in Multinational Corporation Subsidiaries. European Management Journal, 17(2), 226-236 Teece, D. J., Pisano, G. & Shuen, A. (1997). Dynamic Capabilities and Strategic Management. Strategic Management Journal, 18(7), 509-533 Wernerfelt, B. (1984) A Resource-based View of the Firm. Strategic Management Journal, 5(2), 171-180 Weston, J. F., Johnson, B. A. & Siu, J. A. (1999). Mergers and Acquisitions in the Global Chemical Industry. Business Economics, 32(4), 23-31 White, R. E. & Poynter, T. A. (1984). Strategies for Foreign-Owned Subsidiaries in Canada. Business Quarterly, 49(2), 59-69 Zander, I. (1999). How do you mean ‘Global’? An Empirical Investigation of Innovation Networks in the Multinational Corporation. Research Policy, 28(2/3), 195-213
In: Outsourcing, Teamwork and Business Management ISBN: 978-1-60456-956-8 Editor: Karl E. Carettas, pp. 31-45 © 2009 Nova Science Publishers, Inc.
Chapter 3
MANAGEMENT AND CONTROL IN SERVICE FIRMS: BRIDGING THE GAP BETWEEN ORGANIZATION STUDIES AND SERVICE MANAGEMENT Per Skålén* Service Research Center, Karlstad University SE-651 88 Karlstad, Sweden
Abstract The present chapter draws together previous research within the boundaries of organization studies and services management, and then proposes an alternative focus for studying management and control in service firms. The chapter argues that organization studies have contributed to the study of management and control in service firms but without actually focusing on the central issues that services management research has argued should characterize management and control in service firms. By bringing together the two traditions an approach to empirical research that would deepen understanding of management and control in service organizations is outlined.
Keywords: management, control, service firms, service research, organization studies, marketing.
Introduction Contemporary Western society has been characterized as a society dominated by a culture of consumption, with person’s role as a consumer being central to the construction of individual identity (Abercrombie 1991; Bauman 1988). This perspective has been adopted by management scholars, who have argued that the customer essentially ‘manages’ a modern organization because the prevailing customer orientation of contemporary firms implies that the customer is guiding the design, development, and control of the organization (du Gay and *
E-mail address:
[email protected]. Phone +46-54-700 2112. Fax +46 54-83 65 52
32
Per Skålén
Salaman 1992; du Gay 1996). However, this view has been criticized as not being based on sound empirical data (Fournier and Gray 1999; Korczynski et al. 2000; Korczynski 2004). In contrast to the view that customers effectively control firms, Korzcynski et al. (2000) found empirically that, although front-line employees are apparently controlled by customers’ demands to some extent, this is not due to the ‘culture of the customer’; rather it is due to technologies and procedures associated with recruitment, induction, on-the-job training, and performance assessment. Moreover, Korzcynski (2004) found empirically that back-office personnel are almost unaffected by customer demands; rather, their work is characterized by bureaucratic procedures. This empirical work (Korczynski et al. 2000; Korczynski 2004) has certainly contributed to a better understanding of management and control in service firms. In contrast to the argument that a customer-oriented culture is formed spontaneously in service firms as a result of the consumer mentality that dominates contemporary society, this empirical work has persuasively demonstrated that customer orientation is fostered by management in a deliberate manner. However, the influence of the prevailing idea of ‘customer orientation’ in services management research must be taken into account when studying the fostering of customer orientation by management in service firms. The present chapter therefore explores how services management (in general) has elaborated and adapted the idea of customer orientation for service firms, and how this has affected approaches to management and control in service firms. Within the literature on services management, customer-oriented models have seldom been viewed as devices for managing and controlling service firms; however, the present text contends that it is plausible to adopt this position. By drawing on empirical research in services management, it is argued that the customer orientation of service firms has had a more profound influence on management and control of service firms than has been acknowledged by Korczynski (2004) and others (Peccei and Rosenthal 2000; Sturdy 1998; Edwards et al. 1998). In fact, it could be argued that du Gay (1996) and du Gay and Salaman (1992) were not entirely in error in arguing that service firms are more or less ‘colonized’ by a ‘culture of the customer’. However, the present study argues that any such ‘colonization’ has been effected through means other than those posited by du Gay (1996) and du Gay and Salaman (1992). Management and control are thus central concepts in the present chapter. The concept of ‘control’ has traditionally been associated with the theory and practice of management accounting. Indeed, it has been defined as “… a formal system for gathering and communicating data for the ends of aiding and coordinating collective decisions in the light of the overall goals or objectives of an organization” (Horngren and Sundem 1990 in Macintosh 1994 p. 2). However, in accordance with Macintosh (1994), the present chapter adopts a broader understanding of management and control—encompassing all activities that are aimed at directing human behavior in organizations. In a service setting, this broad range of activities is often directed towards implementing a ‘customer orientation’ in the service firm by affecting the values, norms, actions, emotions, and thoughts of the personnel (Korzcynski 2004; Sturdy 1998). The chapter is structured as follows. Following this introduction, the role of ‘services management’ as an approach to managing and controlling service firms is discussed. The study then contrasts the empirical work of Korzcynski et al. (2000) with normative theories of services management. In the third section, empirical work within services management is
Management and Control in Service Firms
33
examined, and the impact of a ‘services-management approach’ to management and control is explored. The fourth section presents a discussion of the findings. Then the final conclusions and suggestions for future research is presented.
Service Management Since the 1950s, marketing discourse has undergone significant change from a narrow focus on market communication to the promotion of a ‘customer orientation’ throughout the organization (Skålén et al. 2006; Vargo and Lusch 2004). Marketing has established itself as a form of managerial theory. In this regard, services management can be understood as the latest and most elaborated contribution. Indeed, services management, as a field of research and a form of consultancy practice, has had a major impact on both marketing theory (Vargo and Lusch 2004) and management practice (Brownlie et al. 1999; Schneide, 2004). Before services management first began to be formulated as a distinctive field in the mid 1970s (Berry and Parasuraman 1993; Shostack 1977), the marketing discourse had generally been articulated on the assumption that the level of interaction between employees and customers should be limited and indirect (Kotler 1976). In contrast to this, services management was based on the reverse assumption. A central tenet of services-management scholarship has thus always been that every member of an organization should behave as a marketer for the organization if it is to survive and prosper (Grönroos 1982; Gummesson 1990; Parasuraman et al. 1985). According to this basic tenet of services-management scholarship, the personnel are essential to the service. From a management-and-control perspective, the major contribution of services management has been the development of theory and practice for managing and controlling the personnel who are deemed essential to the service. Grönroos (1994), for example, identified five distinctive attributes of services management, of which three had a direct connection to management and control—(i) that the services-management perspective must pervade the whole organization (and not be restricted to the marketing department); (ii) that quality should be considered as a strategic management issue (and not a separate issue for the production department); and (iii) that personnel should embrace the quality orientation of the firm as their guiding principle of behavior; that is, they should think and act according to quality prerogatives. This focus on customer orientation in service firms has been especially evident in the conceptualization of ‘service quality’, which has been at the centre of the servicesmanagement research agenda since the early 1980s (Brady and Cronin 2001; Brown et al. 1994; Parasuraman et al. 1985). Grönroos (1982; 1983; 1984) made an early and important contribution in this regard when he conceptualized quality as being not only the output of a service act (so-called ‘technical quality’), but also a result of how the service is delivered (socalled ‘functional quality’). From a management perspective, this conceptualization of quality placed the personnel at the centre of intervention and control—because they were posited as the agents who delivered the service and who thus had the greatest influence on the quality evaluation made by the customer. Grönroos’ (1982; 1983; 1984) conceptualization provided the foundation for several models of customer-perceived service quality that aim to measure, manage, and control the
34
Per Skålén
effect of personnel on the overall quality. The ‘gap model’ (Parasuraman et al. 1985; 1988; 1994) has probably been the most important contribution to the field of customer-perceived service quality, and even though competing models differ in some respects, the fundamental approach of the ‘gap model’ is representative of other contributions. The ‘gap model’ conceptualized the measurement of service quality in terms of a comparison between a customer’s expectations and customer’s perceptions (the so-called ‘disconfirmation paradigm’). This approach to the measurement of quality has remained the predominant model in the literature on quality and customer satisfaction (Oliver 1997). In introducing the ‘gap model’, Parasuraman et al. (1985 p. 42) argued that ‘service quality’ could not be understood in terms of ‘goods quality’—because “… the characteristics of services … have to be acknowledged for a full understanding of service quality”. In creating their model, Parasuraman et al (1985) used an explanatory research design. Four service categories were chosen for their investigation: retail banking, credit card provision, securities brokerage, and product repair and maintenance. A single firm represented each service category. In-depth open-ended personal interviews were conducted with 14 executives (three or four from each firm), and 12 focus-group interviews were conducted with customers of the firms. According to Parasuraman et al. (1985 p. 44, emphasis in original removed), the main conclusion to be drawn from the interviews with the executives was that: … a set of key discrepancies or gaps exists regarding executive perceptions of service quality and the tasks associated with service delivery to consumers. These gaps can be major hurdles in attempting to deliver a service which consumers would perceive as being of high quality.
The gaps identified by Parasuraman et al. (1985) were:
∗ ∗ ∗ ∗
the gap between what customers expect from a service and management’s perception of customers’ expectation; the gap between management’s perception of customers’ expectations and service quality specifications; the gap between service-quality specifications and the actual service delivery; and the gap between service delivery and external communications.
The analysis of the focus-group interviews provided support for the ‘disconfirmation paradigm’—according to which service quality can be considered: (i) ‘satisfactory’ if perceptions equal expectations; (ii) ‘unsatisfactory’ if expectations are below perceptions; and ‘excellent’ if perceptions exceed expectations (Parasuraman et al. 1985; 1988). Although Cronin and Taylor (1994) and Parasuraman et al. (1994) later presented slightly different interpretations of the ‘gap model’, the summary presented above is satisfactory for the purposes of the present study. In later versions of the ‘gap model’, the notion of a ‘zone of tolerance’ was introduced— according to which customer expectations do not have to match customer perceptions exactly for service quality to be considered satisfactory; rather, some variation is allowable, as follows: (i) ‘satisfactory’ service quality is said to range within the zone of tolerance (from
Management and Control in Service Firms
35
‘adequate’ to ‘desired’); (ii) ‘excellent’ service quality exceeds the zone of tolerance; and (iii) ‘unsatisfactory’ service quality fails to reach the zone of tolerance (Strandvik 1994). To return to the four ‘gaps’ identified by Parasuraman et al. (1985), a customer’s perception of a service is dependent on the size and direction of gaps 1–4, whereas the customer’s expectation is dependent on: (i) past experience with the service; (ii) word-ofmouth communication regarding the service; and (iii) personal needs. Taken together, the four ‘gaps’ constitute a perception–expectation construct (designated ‘gap 5’), which thus has a heuristic position in the model—bringing together the customer side and the organizational side of the model (see Figure 1).
Customer Personal needs
Word of mouth communication
Past experience
Expected service Gap 5 Perceived service
Organization
Service delivery Gap 3
Gap 4
External communications to consumers
Translation of perceptions into service quality specifications
Gap 1
Gap 2 Management perceptions of customer expectations
Source Adapted from Parasuraman et al. (1985:44)
Figure 1. The gap-model.
On the basis of the focus-group interviews, Parasuraman et al. (1985 p. 46) also found that: … regardless of the type of service, consumers used basically similar criteria in evaluating service quality. These criteria seem to fall into 10 key categories which are labeled ‘service quality determinants’.
36
Per Skålén
In later versions of the model, these ten service-quality determinants were reduced to five, and eventually to three. However, in the most common presentations of the model, five determinants are usually depicted (Brady and Cronin 2001; Parasurman et al. 1988):
∗ ∗ ∗ ∗ ∗
reliability: the ability to perform the promised service dependably and accurately; responsiveness: a willingness to help the customers and to provide prompt service; empathy: the provision of caring, individualized attention to customers; assurance: the knowledge and courtesy of employees, and their ability to inspire trust and confidence; and tangibles: the physical facilities equipment associated with the service, and the appearance of personnel.
These service-quality determinants provided greater precision to the construct of perceived service quality, and offered a foundation for the creation of a standardized scale for the measurement of service quality. The 22-item instrument that Parasuraman et al. (1988) developed, known as ‘SERVQUAL’, operationalised the five quality determinants and made it possible to measure service quality on a large scale for the first time by using the 22 questions to assess customer’s expectations and perceptions regarding a service delivery. The ‘gap model’ and the SERVQUAL instrument have been accompanied with practical advice to managers on how to act on the information provided (Zeithaml et al. 1990)—that is, advice on how to manage and control service firms on the basis of the model. The fundamental argument in this regard has been that corrective action is required when the quality evaluation falls below the zone of tolerance—that is, when customer perceived quality is classed as unsatisfactory. Zeithaml et al. (1990) provided 17 reasons for the occurrence of various ‘gaps’, followed by suggestions for solving the problems. In discussing ‘Gap 3’—the gap between service-quality specifications and actual service delivery—Zeithaml et al. (1990) offered seven possible reasons. One of these was role conflict, which was defined by Zeithaml et al. (1990 p. 92) as the “… extent to which employees perceive that they cannot satisfy all the demands of all the individuals (internal and external customers) they must serve”. Guidelines were then offered as to how problems of role conflict should be resolved: “If the company defines service roles and standards in terms of customers’ expectations, role conflict is minimized” (Zeithaml et al. 1990, p. 98). Since the development of the ‘gap model’, debate on service quality has focused on two major issues. First, there has been discussion on the measurement of perceived service quality, and whether only perceptions (not expectations) should be measured; however, the disconfirmation approach still appears to dominate research (Brady and Cronin 2001). Secondly, certain elaborations of the ‘gap model’ have been suggested, in which service quality has been conceptualized as consisting of: (i) interaction quality (the attitudes and the behavior of employees); (ii) physical environment (including design issues, social factors, and non-visual aspects); and (iii) outcome quality (waiting time, customer perceptions of outcome, and tangibles) (Brady and Cronin 2001). However, the present text is not concerned to debate the details of the nature of service quality; rather, the focus of this study is on the approach to management and control that is
Management and Control in Service Firms
37
implicit in the discourse on services management and service quality. In this regard, it is possible to identify seven implied principles in the literature:
∗ ∗ ∗ ∗ ∗ ∗ ∗
that the management and control of service firms should be contingent on the customer’s demands; that the customer’s evaluation of a service firm is dependent on the actions of the personnel, and that the personnel are thus the prime focus of management control in services management; that good management and control of the personnel requires customer-perceived service quality to be measured; that the measurement of customer-perceived service quality is contingent upon certain quality factors that account for the quality a firm delivers; that customers differ in their evaluation of these quality factors, and that the importance of these various quality factors to the customer should be evaluated by measuring expectations and perceptions; that any ‘gaps’ that are detected between customer expectations and customer perceptions should be addressed by management according to the suggestions offered in the literature; and that it is pivotal to establish an integrated system of management and control in which all of these principles are incorporated.
Empirical Research on Management and Control of Service Firms in Organization Studies There is a lack of empirical research explicitly focusing management and control in service firms within industrial sociology and organizational studies. Korczynski et al. (2000) have produced the best study of the subject matter in their qualitative study of five call centers, which focused on customer-service representatives (that is, front-line employees). In reviewing the work of Korczynski et al. (2000), the focus of the present chapter is the extent to which this work addressed the attributes of management and control noted above. Korczynski et al. (2000) clearly demonstrated that the paradigm of ‘customer orientation’ informed the management and control of recruitment, induction, training, and performance in the firms they studied. During recruitment, the firms sought to hire personnel with customeroriented attitudes, and this focus continued during induction. As Korczynski et al. (2000 p. 675) observed: “… in all of the sites [studied] staff undertook courses on customer service skills before starting their job proper”. Moreover, their skills in this regard were consolidated during training and measured during actual work performance. However, management and control in the firms studied by Korczynski et al. (2000) was not based on measurement of how customers actually perceived the service encounter; rather, it was based on how the employees themselves preferred to be served—a technique referred to as “self-as-customer orientation”. Employees were asked to put themselves in the position
38
Per Skålén
of the customer, and thus learn appropriate behavior in accordance with the kind of interaction they, themselves, would like to have with personnel in service firms. The training of the personnel was thus based on the idea of taking the customer’s perspective, which appears to be in accordance with a services-management approach to management and control. However, the approach assumes that the norms and attitudes held by the personnel are representative of those of customers, but according to research within service management (as outlined above), it cannot be taken for granted that employees actually have the ability to interpret what kind of service the customers desire in various service settings and personal circumstances. For example, most customers probably expect more empathy from hospital staff than from fast-food restaurant staff, and they probably desire more empathy if they are diagnosed as suffering from cancer than a minor ailment. Although service personnel might have a general understanding of these differing expectations, it is reasonable to suggest that such a general understanding could be developed and ‘fine tuned’ by utilizing objective measurements of customer perceptions and expectations, rather than relying on the subjective views of service personnel. Indeed, it might be argued that the research of Korczynski et al. (2000) revealed management practices that are more in accordance with the views of du Gay (1996) and du Gay and Salaman (1992), who argued that management and control within service firms institutionalize general norms that prevail in contemporary consumer society (as apprehended, in this case, by the personnel). The research of Korczynski et al. (2000) also gave insight into whether the management and control of front-line employees was based on the full range of service-quality factors. In fact, only one of the quality factors was consistently acted upon—that of empathy. According to Korczynski et al. (2000 p. 675): “… in all the sites [studied] management promoted and made use of the norm of customer empathy”. Indeed, a recruitment officer saw this as a reason for the presence of so many homosexual men in the worksites studied: “… recruitment agencies like sending them [gay men] to us because they think they’re empathic” (Korczynski et al. 2000, p. 674). This apparent emphasis on empathy fails to take account of the fact that empathic behavior is not the only important factor in promoting service quality. Indeed, it could be argued that empathic behavior is not as important as other service-quality factors in certain service situations. For example, when ordering a hamburger, responsiveness and reliability are likely to be more important than empathy to many customers. Moreover, empathy might also have been of lesser importance in at least some of the service encounters studied by Korczynski et al. (2000). For example, in financial services and telecommunication services, how much ‘empathy’ does a regular customer really desire when opening an account or when switching a subscription to another mobile-phone vendor? As was suggested in the example of ordering a hamburger, most customers of financial services and telecommunications services would probably wish to have prompt (responsive) and accurate (reliable) services. The service-management approach to management and control also emphasizes the importance of measuring customers’ perceptions and expectations of service quality. It is apparent (from the discussion presented above) that the firms studied by Korczynski et al. (2000) did not use the results of such measurement in the management and control of their employees. The question remains as to whether the firms measured customer-perceived quality at all. The evidence reported by Korczynski et al. (2000) suggests that they did not.
Management and Control in Service Firms
39
However, this does not mean that measurement of behavior did not exist at all; indeed, the opposite seems to have prevailed. According to Korczynski et al. (2000, p. 675), one worker remarked: “…you get measured on how many times you scratch your shoulder”. Measurement was thus associated with bureaucratic control, as was noted by Korczynski et al. (2000, p. 674): “Control in the sites was informed by both pervasive (bureaucratic) measurement and monitoring, and the use of customer related norms”. This bureaucratic type of measurement consisted, at least at one site, in call monitoring. Although one of the professed aims of such monitoring was to understand customers’ perceptions of quality, Korczynski et al. (2000, p. 675) acknowledged that “… monitoring and measurement was … incomplete because a key output related to customer perception of quality … which management could only occasionally proxy through monitoring of calls” [emphasis added]. The lack of measurement of customer-perceived service quality meant that the sites studied by Korczynski et al. (2000) did not have other characteristics that might have been expected in an organization that claimed to have introduced management and control on the basis of services management—such as techniques for establishing what level of quality customers desired, measurement of the quality delivered, approaches for addressing any ‘gaps’ between customer expectations and perceptions, and an integrated servicesmanagement control system. Other studies in the field organization studies suggest that the sites studied by Korczynski et al. (2000) are not unique in this respect. Although previous research of management and control in service organizations is scanty, it would seem that this question (of the systematic use of the services-management approach to management and control) has not been addressed in previous research (Korczynski 2004; Peccei and Rosenthal 2000; Sturdy 1998; Edwards et al. 1998).
Empirical Research on Management and Control in Service Management From the above discussion, the question arises as to whether organizations that have incorporated the services-management perspective on management and control into their daily operations actually exist. The problem in answering this question is the lack of previous research on this specific subject. It might be expected that this would be a well-researched topic in services management; however, this is not the case. Although scholars have been keen to develop managerial practices on the basis of services-management, there has been surprisingly little interest in evaluating these managerial practices through systematic empirical studies. Such studies appear to be completely absent in organization studies; and within the literature on services management and marketing, only a few empirical studies have touched upon this question (Homburg et al. 2002; Johnsson 1996; Lynn et al. 2000; Kantsperger and Kunz 2005). Of these, the study of Kantsperger and Kunz (2005), who investigated the management of overall service quality in customer-care centers, is of particular interest in the context of the present examination of the work of Korsczynski et al. (2000) and Korsczynski (2004). Kantsperger and Kunz (2005) were guided by a managerial rationale—to delineate which service-quality factors affect customer satisfaction and customer loyalty, and thus profitability. Despite this managerial focus, and despite the quantitative design of the study,
40
Per Skålén
the authors did provide some interesting insights into management and control in service firms. The paper reported on the interplay among three levels—(i) management; (ii) employees; and (iii) customers. On the management level, three ‘management styles’ were considered: (i) ‘customer orientation’; (ii) ‘quality orientation’; and (iii) ‘employee orientation’. According to Kantsperger and Kunz (2005 p. 137), these management styles “… are supposed to be important for the effective and efficient management of a customer care center”. Kantsperger and Kunz (2005 p. 137) went on to observe: Quality orientation reflects whether modern quality management techniques are applied in order to enable excellent quality within the customer care center. It encompasses management practices that indicate systematic procedures of performance and quality measurement within the customer care center. Employee orientation indicates whether management shows initiative in order to empower and support their employees. Employee orientation of the management can be described for example by implemented feedback procedures or training skills within the firm. Finally, customer orientation shows the weight put on understanding the customer and his needs and the importance of improving services to safeguard customer retention.
The items that comprised the ‘quality orientation’ in the study included: (i) an integrated internal-performance measurement; (ii) implementation of a quality-management-system; and (iii) quality perceived as a matter for all employees. The items that comprised the ‘employee orientation’ included: (i) employees provided with performance feedback and suggestions for improvement; and (ii) the existence of surveys to measure employee satisfaction. The items that comprised the ‘customer orientation’ included: (i) use of techniques for tracking customer expectations; (ii) the customer being placed at the centre of strategic planning; (iii) service standards being based on consistent and continuous analysis of customers needs; and (iv) explicit knowledge regarding customer benefit being available. On the second level—the employee level—three constructs were considered: (i) employee satisfaction; (ii) quality ‘overstretch’; and (iii) employee loyalty. Based on previous research, ‘employee satisfaction’ was considered to be equivalent to the employees’ attitudes to work; ’employee satisfaction’ was treated as a driver of employee loyalty. The term ‘quality overstretch’ referred to whether work was perceived by employees as being demanding in terms of required abilities and skills; this construct was also correlated with employee loyalty. All three factors were found to be important for the motivation and performance of employees. The constructs at the third level—the customer level—included ‘customer satisfaction’ and ‘customer loyalty’, both of which were posited as drivers of profitability. The conceptualization of the management of service quality used by Kantsperger and Kunz (2005) had many commonalities with the seven principles implied by the servicesmanagement literature (as listed above). The conceptualization advanced by Kantsperger and Kunz (2005) had the following features: (i) the object of management is to control the personnel; (ii) quality is a matter for all organisational members; (iii) service quality is measured from the customer perspective (although the study did not identify the quality factors upon which these measurements were based); (iv) both customers’ expectations and their perceptions are included in the measurement; (v) this information is to be acted upon by
Management and Control in Service Firms
41
the management (that is, service standards and strategic planning are contingent upon customer information); and (vi) an integrated quality system should be implemented. Despite the similarities, there were also minor differences between the conceptualization used by Kantsperger and Kunz (2005) and that identified from the above review of services management from a management and control perspective. In particular, some aspects of the ‘quality orientation’ and the ‘employee orientation’ actually had more in common with the conception of ‘quality’ in the literature on total quality management (TQM) (Hackman and Wageman 1997) than that of the services-management literature. TQM has typically placed more emphasis on internal performance measures, and on the provision of feedback to employees on the basis of such measures—rather than on measures derived from customers’ perceptions, which have been typically associated with services-management research. Kantsperger and Kunz (2005 p. 147) reported that multiple regression analysis of their path model had revealed that ‘employee satisfaction’ was “… the most important predictor for customer satisfaction and employee loyalty … [and that] … employee satisfaction is the main mediator of a customer orientated management style to the customers’ level”. They thus concluded that customer-oriented management had a positive effect on employee satisfaction. However, they also found that ‘quality orientation’ had a negative effect on employee satisfaction. As Kantsperger and Kunz (2005, p 147) put it: … quality orientation seems to discourage the employees in some way … a strict quality orientation could signal a low level of trust toward the employees that in response could decrease intrinsic motivation and employee satisfaction.
However, in interpreting these results, it should be kept in mind that the ‘quality orientation’ posited by these authors was largely influenced by a ‘TQM approach’ to quality. In fact, a low level of trust as an effect of adapting a TQM approach to management and control was also reported by Korczinsky et al. (2000), who pointed out the inherent contradiction between the bureaucratic procedures associated with TQM (such as call monitoring) and the commitment to the empowerment of employees that service-management scholars have found to be pivotal if customers are to be satisfied. From the perspective of the present chapter, it seems that a services-management approach to management and control has various effects on employees. Some of these effects are in accordance with the aims of services management (such as a positive correlation with profitability), whereas others are not in accordance with those aims (such as a negative correlation with employee satisfaction). Furthermore, it seems that service firms really utilize a services-management approach to management and control. However, to understand how management and control influenced by services management really operates in service firms, this process needs closer scrutiny than has yet been the case in the services-management literature.
Discussion The above discussion raises the question of why previous research within organization studies has made little attempt to study management and control of service firms based on the
42
Per Skålén
services-management approach to the subject matter. Several explanations (or partial explanations) can be offered. Part of the answer probably resides in the historical fact that services management was not, at least initially, conceptualized as a managerial device—in contrast to other initiatives, such as TQM, which were designed for practical application in the management of firms. Unlike TQM and similar management programs, services management has been historically associated with academic research, and in this domain it has certainly produced an impressive body of academic theory. Nevertheless, despite its roots in the academia, almost every piece of research that it produces has a managerial and prescriptive rationale. Given that truth and knowledge is said to equate with power (Foucault 1977; 1981), it is likely that academic managerial discourse effects practice—especially when the techniques and models produced by academic research are designed to create local and contextualized truths, which is certainly the case with the customer-perceived measurement models of services management. The ‘gap model’ frames and delivers ‘truths’ with respect to how service quality should be measured in general terms—that is, that the managerial focus should be on the service-quality determinants of reliability, responsiveness, empathy, assurance, and trust. According to this ‘general truth’, organizations can use the ‘gap model’ to ascertain the levels of reliability, responsiveness, and so on that the customers of a given organization are likely to desire and perceive—and this establishes a ‘local truth’. By acting upon these ‘truths’, management can legitimize its customer-oriented approach to the control of personnel—in accordance with customers’ demands. A second part of the explanation of why previous research has made little attempt to study management and control in service firms informed by services-management research is the lack of ‘branding’ within service management. If an organization implements ‘TQM’ this is usually made known to the environment of the organization. In contrast, because services management is not ‘branded’ in the same way as TQM, an organization’s adoption of a services-management approach to control is seldom overtly stated. The lack of ‘branding’ of services management as a managerial device also means that it can be difficult to detect it in real-life organizations. In fact, practitioners can be unaware that they are actually using services management. A third reason for the relative lack of study of service firms that have adopted a servicesmanagement approach to control might be the very conceptualization of ‘control’ in previous research. In this regard, it is instructive to note that Korczinsky (2004 p. 103) quoted Edwards (1979, p. 17) with approval in seeing ‘control’ as “the ability of … managers to obtain desired work behavior from workers”. This prescriptive notion of control can be contrasted with Foucault’s (2000) more subtle notion of ‘government’—which he defined as the “conduct of conduct”. On one level, this implies that government is about leading, directing, and guiding others in a more or less prescriptive way, but it also allows for a more subtle interpretation— that government is not only about leading others, but also about leading oneself in selfregulation. The “conduct of conduct” thus becomes the deliberate direction of people’s articulated sets of behaviors by themselves (Dean 1999). This notion of self-regulation is pivotal in understanding how management and control can be based upon service management. As previously noted, services management focuses on shaping the values, norms, attitudes, and emotions of employees, but the notion of selfregulation posits that such factors have to be shaped and changed by the employees themselves. Managers can try to influence them, but managers cannot determine them. It is
Management and Control in Service Firms
43
thus apparent that the managerial focus cannot be restricted to formal procedures for managing and controlling employees—such as training and induction focused upon by Korczynski et al. (2000). Rather, the managerial focus must include consideration of how the employees perceive and constitute themselves as a workforce. Managers cannot focus only on the ‘what’ of power and government; they must also focus on the ‘how’ of power and government in terms of the “conduct of conduct”.
Conclusions This chapter suggests that inspiration for future research in management and control of service firms should be drawn from organization studies, in addition to inspiration from services management itself. Methodologically, the design of the research should be on a par with the qualitative work of Korczynski (2004) and Korczynski et al. (2000). This is certainly the only way to learn more about how management and control actually work in service firms. On the other hand, future research also has to pay attention to the services-management literature when it comes to the object and focus of empirical study and analysis. In this latter endeavor, future research could find inspiration in the present text’s explication and clarification of how management and control should work according to the servicesmanagement literature. Combining these perspectives in the way proposed here would deepen understanding on how service organizations are managed and controlled. The present chapter has also discussed the epistemological and ontological foundation of previous research into management and control of service firms within industrial sociology and organization studies, and has set this within a subtle conceptualization of the notion of ‘control’. The chapter has argued that the social ontology of the modern consumer society in which service firms operate must be taken into account, as must be the role of knowledge in legitimizing power. It is thus argued that a deeper understanding of management and control in service firms is likely to be achieved if power is conceptualized as embedded in knowledge, rather than as a commodity possessed by a particular agent. Moreover, government needs to be understood as the “conduct of conduct”—that is, as an exercise whereby workers orientate themselves, and thus manage themselves, in terms of the norms produced by the managerial ‘truths’ that have a legitimate status in the particular organization.
References Abercrombie, N. (1991) ‘The Privilege of the Producer’. In Keat R. and Abercrombie N., (eds.) Enterprise Culture. London: Routledge. Bauman, Z. (1988) ‘Is there a Postmodern Sociology?’, Theory, Culture and Society, 5: 217-37. Berry, L.L. and Parasuraman, A. (1993) ‘Building a New Academic Field: The Case of Services Marketing’, Journal of Retailing, 69: 13-61. Brady, M.K. and Cronin J.J. (2001) ‘Some New Thoughts on Conceptualizing Perceived Service Quality: A Hierarchical Approach’, Journal of Marketing, 65: 34-49.
44
Per Skålén
Brown, S.W., Fisk, R.P. and Bitner, M.J. (1994) ‘The Development and Emergence of Services Marketing Thought’, International Journal of Service Industry Management, 5: 21-48. Brownlie, D., Saren, M., Wensley, R. and Whittington, R. (1999) ‘Marketing Disequilibrium: On Redress and Restoration’. In Brownlie, D., Saren, M., Wensley, R. and Whittington, R. (eds.) Rethinking Marketing: Towards Critical Marketing Accountings. London: Sage. Cronin, J.J. and Taylor, S.A. (1994) ‘SERVPERF Versus SERVQUAL: Reconciling Performance-Based and Perceptions-Minus-Expectations Measurement of Service Quality’, Journal of Marketing, 58: 125-31. Dean, M. (1999) Governmentality: Power and Rule in Modern Society. London: Sage. du Gay, P. (1996) Consumption and Identity at Work. London: Sage. du Gay, P. and Salaman, G. (1992) ‘The Cult[ure] of the Customer’, Journal of Management Studies, 29: 615-33. Fournier, V. and Gray, C. (1999) ‘Too Much, Too Little and Too Often: A Critique of du Gay’s Analysis of Enterprise’, Organization, 6: 107-28. Edwards, P., Collinson, M. and Rees, C. (1998) ‘The Determinants of Employee Responses to TQM: Six Case Studies’, Organization Studies, 19: 449-75. Foucault, M. (1977) Discipline and Punish: The Birth of the Prison. London: Penguin. Foucault, M. (1981) The Will to Knowledge: The History of Sexuality, Vol. 1. London: Penguin. Foucault, M. (2000) ‘The Subject and Power’. In Faubion, J.D. (eds.) Power: The Essential Works of Foucault: Volume 3. New York: The Free Press. Grönroos, C. (1982) ‘An Applied Service Marketing Theory’, European Journal of Marketing, 16 (7): 30-41. Grönroos, C. (1983) Strategic management and marketing in the service sector, Lund: Studentlitteratur. Grönroos, C. (1984) ‘A Service Quality Model and its Marketing Implications’, European Journal of Marketing, 18 (4): 36-44. Grönroos, C (1994) ‘From Scientific Management to Service Management: A Management Perspective for the Age of Service Competition’, International Journal of Service Industry Management, 5: 5-20. Gummesson, E. (1990), ‘Marketing-orientation Revisited: The Crucial Role of the Part-time Marketer’, European Journal of Marketing, 25: 60-75. Hackman, J.R. and Wageman R. (1997) ‘Total Quality Management: Empirical, Conceptual and Practical Issues’, Administrative Science Quarterly, 40: 309-43. Homburg, C., Hoyer, W.D. and Fassnacht, M. (2002) ‘Service orientation of a Retailer’s Business Strategy: Dimensions, Antecedents, and Performance Outcomes’, Journal of Marketing, 66: 86-101. Johnsson, J.W. (1996) ‘Linking Employee Perceptions of Service Climate to Customer Satisfaction’, Personnel Psychology, 49: 831-51. Kantsperger R. and Kunz, W.H. (2005) ‘Managing Overall Service Quality in Customer Care Centres: Empirical Findings of a Multi-Perspective Approach’, International Journal of Service Industry Management, 16: 135-51. Korcynski, M., Shire, K., Frenkel, S. and Tam, M. (2000) ‘Service Work in Consumer Capitalism: Customers, Control and Contradictions’, Work, Employment and Society, 14: 669-87.
Management and Control in Service Firms
45
Korczynski, M. (2004) ‘Back-office Service Work: Bureaucracy Challenged?’, Work, Employment and Society, 18: 97-114. Kotler, P. (1976) Marketing Management: Analysis, planning and control, third edition. London: Prentice-Hall. Lynn, M.L., Lytle, R.S. and Bobek, S. (2000) ‘Service Orientation in Transitional Markets: Does it Matter?’, European Journal of Marketing, 34: 279-98. Macintosh, N.B. (1994) Management Accounting and Control Systems: An Organizational and Behavioral Approach. Chiseter: Wiley. Oliver, R.L. (1997) Satisfaction: A Behavioural Perspective on the Consumer. New York: McGraw-Hill. Parasuraman, A., Zeithaml, V.A. and Berry, L.L. (1985) ‘A Conceptual Model of Service Quality and it’s Implications for Future Research’, Journal of Marketing, 49: 253-68. Parasuraman, A., Zeithaml, V.A. and Berry, L.L. (1988) ‘SERVQUAL: A Multiple-Item Scale for Measuring Consumer Perceptions of Service Quality’, Journal of Retailing, 64: 12-37. Parasuraman, A., Zeithaml, V.A. and Berry, L.L. (1994) ‘Reassesment of Expectations as a Comparison Standard in Measuring Service Quality: Implications for Further Research’, Journal of Marketing, 58: 111-24. Peccei R. and Rosenthal, P. (2000) ‘Front-line Responses to Customer Orientation Programmes: A Theorethical and Empirical Analysis’, International Journal of Human Resource Management, 11: 562-90. Schneider, B. and White, S.S. (2004) Service Quality: Research Perspectives. London: Sage. Shostack, L.G. (1977) ‘Breaking Free from Product Marketing’, Journal of Marketing, 42: 73-80. Skålén, P., Fellesson, M. and Fougère, M. (2006) ‘The Governmentality of Marketing Discourse’, Scandinavian Journal of Management, 22 (4): 275-91. Strandvik, T. (1994) Tolerance Zones in Perceived Service Quality. Helsinki: Swedish School of Economics and Business Administration. Sturdy, A. (1998) ‘Customer Care in a Consumer Society: Smiling and Sometimes Meaning it?’, Organization, 5: 27-53. Vargo, S.L. and Lusch, R.F. (2004) ‘Evolving to a New Dominant Logic for Marketing’, Journal of Marketing, 68: 1-17. Zeithaml, V.A., Berry, L.L. and Pararsuraman, A. (1990) Delivering Quality Service: Balancing Customer Perceptions and Expectations. New York: The Free Press.
In: Outsourcing, Teamwork and Business Management ISBN: 978-1-60456-956-8 Editor: Karl E. Carettas, pp. 47-72 © 2009 Nova Science Publishers, Inc.
Chapter 4
ACTIVITY AWARENESS AND COMPLEX TEAMWORK John M. Carroll, Mary Beth Rosson, Craig H. Ganoe, Marcela Borge, Jamika D. Burge, Umer Farooq, Gregorio Convertino, Paula M. Bach, Helena Mentis and Hao Jiang The Pennsylvania State University, PA USA
Abstract Collaborators must attain and maintain reciprocal awareness of shared activity in order to coordinate effectively (Dourish & Bellotti, 1992). They need to be assured that their partners are ‘there’ in some sense, which is not always evident or simple in computer-mediated collaboration. They need to know what tools and resources their counterparts can access, who they know that might know something, or know how to do something that would be critical. They need to know what relevant information their collaborators know, and what they expect, as well as their attitudes and goals. They need to know what criteria their partners will use to evaluate joint outcomes, the moment-to-moment focus of their attention and action during the collaborative work, and how the view of the shared plan and the work actually accomplished evolves over time. Research on collaboration and technology support for collaboration has identified several types of awareness: social awareness, action awareness, workspace awareness, situation awareness (for an excellent review, see Schmidt, 2002). Most investigations of awareness research have focused on synchronous phenomena: awareness of who is participating in an ongoing activity, awareness of what each person is currently doing in that activity context, and awareness of how the team as a whole is performing. Asynchronous awareness phenomena, for example those supported by version control systems, shared calendars, and project management software, have received less attention. Our research has focused on activity awareness, a programmatic concept for the mutual awareness of partners in a shared activity of significant scope and duration. Activity awareness transcends synchronous awareness of where a partner's cursor is pointing, where the partner is looking, etc. It involves monitoring and integrating many different kinds of information at different levels of analysis, such as events, tasks, goals, social interactions and their meanings, group values and norms, and more. It involves monitoring and integrating
48
John Carroll more-or-less continuingly to learn about developing circumstances and the initiatives, reactions, and sense making of other people with respect to on-going and anticipated courses of action. Activity awareness is not merely a matter of coordinating state information. It is continually negotiated and constructed throughout the course of a collaborative interaction. Thus, it is a process that is constitutive of collaboration. In the balance of this paper we will first describe fieldwork characterizing routine social practices to establish and maintain activity awareness during complex teamwork in regional emergency management planning, emergency room operations, collaborative education, open source software development, scientific collaborations, and management of nonprofit community groups. We then describe software systems we have developed to support activity awareness in complex teamwork in some of these contexts. We close with some discussion of the challenges of supporting activity awareness.
Regional Emergency Management One domain we have investigated is regional emergency planning and response. To many people, emergency planning and response is associated with urban areas – cities – and with network hubs for transportation, commerce, and utilities, such as airports, harbors, and power plants. The scale of emergencies can of course be far greater when the setting is a population center or a piece of major infrastructure. But disaster events such as tornados, lightning, house fires, local flooding, and multi-car collisions can obviously occur anywhere. One important characteristic of relatively more rural settings is that there are few or no full-time first responders to plan or to carry out emergency response operations. For locales that are neither urban centers nor infrastructure hubs, emergency planning and response depends critically on volunteer effort. For about a year and a half during 2004-2005, we “shadowed” the Emergency Coordinator for the Centre Region Council of Governments in Pennsylvania (http://www.crcog.net/). The Center Region is a group of six municipalities in Centre County, Pennsylvania, roughly surrounding the main campus of Pennsylvania State University (University Park). The six municipalities plus Penn State cooperate with respect to issues and functions that can be addressed more effectively on a regional basis, than at the lever of individual townships. For example, hiring an Emergency Coordinator would have been a burden for each organization on its own, and would also be less effective than pooling resources both in supporting the Emergency Coordinator and in better coordinating planning and response. The Emergency Coordinator works with the many area volunteer fire companies, nonprofit emergency medical service organizations, and municipal police forces. The Coordinator oversees a continual planning process in which planning documents are developed and reviewed. These documents are used to help participating organizations understand their individual roles, to individually recruit and train their personnel, and to guide walkthroughs and other inter-organizational training exercises. These planning documents comprise the shared common ground that the various organizations and their personnel would take as a starting point for actual first response operations. (For more details on this study see Schafer, Ganoe & Carroll, 2007; Schafer, Carroll, Haynes, & Abrams, 2008). Regional emergency management as we observed it in this study is highly complex teamwork. It is also surprisingly, and even a bit scarily, improvisational. We closely observed a regional planning and training tabletop simulation of an airport emergency scenario. This
Activity Awareness and Complex Teamwork
49
event took most of a day, and involved personnel from about a dozen regional organizations. We were struck by the complexity of such an operation (even for a relatively tiny regional airport facility), and also by the fact that such walkthrough exercises occur at most two times a year. Much of the time and effort of the participants is spent re-establishing the common ground of shared meanings and expectations that was developed at the previous planning and training event but that had eroded through the course of the intervening several months. Most of the first responders we observed are volunteers or part-time staff. Even when they are full time professionals – as in the case of police officers, the work they do together, planning and training for coordinated multi-organization operations is more or less stolen from their more focal and immediate responsibilities to their primary organization and constituency. On reflection it could hardly be otherwise. Each participating organization already has its own functions and responsibilities, and for the most part, each has barely adequate resourcing to address these core responsibilities. Accordingly, the broader-scope, regional exercises and planning is compromised. In the Centre Region, this is mitigated by the existence of the Emergency Coordinator. Unfortunately, very few regions across Pennsylvania have organized themselves in this way. Most have no Emergency Coordinator. We saw considerable and explicit effort directed at establishing and maintaining activity awareness. The tabletop walkthrough exercise served as a biannual synchronization of the various first response organizations. Members worked through the day monitoring and integrating what others were doing and saying. They checked their understandings, they tried to make sense for themselves of the plans. By and large, these were highly experienced people, so it is significant that in such a setting they were still learning about the role, reactions, and sense making of their partners with respect to anticipated courses of action. But we also concluded that these groups are poorly, and perhaps inadequately supported with respect to activity awareness. We subsequently met with the Emergency Coordinator and demonstrated interactive shared map tools that support collaborative annotation. He felt this sort of computer support would be useful for planning interactions when people could not meet face to face, and that such plans, codified as a sequence of annotations on an interactive map, could be used to train new personnel. He suggested that such software could show trainees both the geo-spatial plans that had been developed, and the annotation process from which the plan derived (see follow up studies in Convertino et al. 2008, To Appear).
Developmental Professional Communities Another domain we have been exploring and supporting is communities of individuals who share professional goals and are organized to develop one another in ways that will promote attainment of these goals. These communities are developmental, in that they understand (either tacitly or explicitly) the phases their members must transition through as part of their professional growth and share the motivation to promote such growth. One implication of the developmental phases is that a community holds expectations about what sorts of community-building behavior is appropriate for different developmental levels. For example a more senior member is expected to serve as a role model, to monitor and provide feedback to more junior members, and so on. More junior members are expected to seek out relationships with senior members who have followed paths similar to theirs.
50
John Carroll
We are studying the seeding and evolutionary growth of a developmental community of young women considering or pursuing education and careers in information sciences and technology (IST; Rosson & Carroll, 2006). The wConnect community currently includes individuals at four phases: high school girls who are considering college but disinclined toward information technology as a career; first and second year undergraduate women who are considering IST as a major; third and fourth year undergraduate women who have committed to IST as a major and are developing their specializations within the program; and IST alumni who have joined the professional workforce (Rosson et al., 2008; 2009). Although the community-building activities involve a few face-to-face activities (e.g., workshops held by undergraduates at high schools, and by alumni at Penn State), most of the activity takes place online. As a result, supporting activity awareness within the community is an important requirement for the tools that we are building. Our toolset is diverse, recognizing the multi-faceted goals and activities of community members at different phases: customized “programming” tools for use by high school students; extensions to the Facebook social networking platform for new members; and online community authoring tools for the more advanced members of the community. There are at least four aspects of community building that we are exploring with respect to activity awareness. One of these is the recruiting of new members. As the community forms, we are engaging different members in outreach to different constituencies. For instance, first or second year students reach out to their former high schools, leveraging their social networks from earlier years. More advanced students reach out to other undergraduates, leveraging their own university-oriented social network. Alumni reach out to undergraduates who they knew – or who know other IST individuals that they knew – while at Penn State. Tracking the status of this outreach and its results is difficult, as it is a highly distributed and dynamic process. Thus we are developing a membership database as a core information structure, defining different levels of recruitment “status” and monitoring the related activities carried out by different members. The second and third activities are at the heart of the developmental mission of the community. Individual members or groups of members are designing the activities that can promote development, for instance hands-on workshops that can be delivered at an area high school, or an online meeting with an alumna who will share experiences regarding her first weeks on her new job. The design of these activities is emergent, and depends greatly on critical mass – we have seen that no one member is likely to build and deliver an activity on her own. Thus we have been exploring mechanisms for tracking and visualizing the status of these activities as collaborative efforts. In a related vein, the tools that are being used to develop the activities are constantly evolving, whether as specialized applications within Facebook, or as customized tools for community building or for the high school workshops. These activities tool need the support of awareness tools that track and externalize the status of tool-building efforts. Finally, one of the assumptions that has guided the wConnect project is that the combination of outreach to members at different phases and the more concrete development and delivery of community activities will promote member participation as the women see themselves as active participants in the creation and growth of their own community. This in turn should lead to the accrual of social capital and feelings of connectivity and trust. We are tracking these as outcomes in our research project, but at the same time are exploring
Activity Awareness and Complex Teamwork
51
mechanisms for enhancing the awareness of the increasing breadth and strength of the social ties that are at the foundation of wConnect.
Hospital Emergency Room Operations The Emergency Department (ED) is the main point of entry to hospitals for unplanned patient admittance, and for many in today’s US healthcare crisis it is also the primary point of care. It is a dynamic, unpredictable, and complex environment filled with continuous communication and interruptions (Spencer, Logan, Coiera, 2002). This 24-hour a day operation engages multiple shifts of autonomous healthcare workers who must navigate this complex and high pressure situation in a coordinated fashion with limited resources (Amouth, et al., 2005). Due to the high rate of medical errors in the ED, hospitals are beginning to push for the integration of electronic patient records and other information systems in order to increase efficiency (i.e. lower costs) and increase safety (i.e. lower deaths). There have been a number of problems noted with this move to the paperless hospital (Ash, Berg, & Coiera, 2004; Embi, Yackel, Logan, Bowea, Cooney, & Gorman, 2004) – primarily due to the lack of understanding how personnel coordinate and communicate in the ED. Despite the amount of information these systems are providing, many of them fail to recognize the natural need for activity awareness. ED nurses especially need quick access to information and a general understanding of the environment in order to do their jobs. For instance, in the practice of medical error recovery, nurses cite surveillance of the entire patient care environment and awareness of the big picture as some of the strategies they take in identifying errors in the ED (Henneman, Blank, Gawlinski, & Henneman, 2006). For more than a year we have been observing operations at the Mt. Nittany Medical Center in State College, Pennsylvania. The ED in this 201-bed acute-care facility is a Trauma 2 center encompassing 20 beds and a 7-bed basic care unit. During triage, a nurse determines the severity of the patient complaint and decides on if their case can be handled in basic care (e.g. a splint or stitches) or if it requires more encompassing emergency room care. The basic care unit is normally staffed with a registered nurse, ED technicians, and physician assistant whereas the emergency room is staffed with approximately a charge nurse, 5 shift nurses, 2 ED technicians, and 2 physicians along with volunteers and ED secretaries. In the emergency room the locus of the activity is on patient care and bed management. The former activity can take from 1 to 12+ hours while the later is an ongoing process. Patient care is the understood primary activity of the health care workers in an emergency room. This includes the tasks of diagnosis and resolution of the health complaint – resolution taking the form of providing care such as medication or admittance to the hospital. Both care and resolution directly relate to the second locus of activity, which is less recognized: bed management. Bed management includes both internal bed management as well as moving people out of the ED. Coordination within these two activities occurs between the health care workers within the Emergency Department as well as between the ED and other departments within the medical center. Activity awareness between departments includes coordination of timing, order, and status. For instance, ordering tests such as a CT scan and ultrasound require the charge nurse to coordinate the order of the tests (the attending physician may want the results
52
John Carroll
of the CT before the ultrasound) as well as the timing of when the procedure can be done (these separate departments have different lengths of time for each procedure and also have outpatient procedures). Order can also apply to the ordering of patients for the same procedure due to severity of condition or other timing issues. Thus, the charge nurse needs to maintain a full understanding of all patients that are in the ED and coordinate that with their understanding of the load and typical situation in each of the other departments in order to make these assessments. In addition, there are different types of status to convey between departments. For instance, when a patient has been indicated for a procedure, when the patient has been sent to the proper department for that procedure, when the patient has returned, and when the results are ready for review by the attending physician. The activity of bed management also requires status between departments – a bed is requested, a bed is assigned but not ready, a bed is ready, and patient has been discharged to the hospital. Activity awareness between healthcare workers within the emergency department primarily requires coordination of status. This includes status of requests, status of results, and status of load. Requests can include medication administration or test requests whereas results can include both an indication of the completion of these tests as well as the results themselves. Status of load is most applicable to the triage nurse who must assess the bed and patient load status of both the ED and basic care in order to make decisions as to where and how quickly to place the incoming patients. At a moments notice, the status can change – an ambulance comes in with a cardiac arrest patient – and this must be conveyed quickly and seamlessly to the triage nurse. There is another underlying awareness that pervades the ED. The triage nurse may also be aware of the affective load that can enhance or be detrimental to the progress of activity in the ED. Particularly stressful days or an unruly patient can cause the attending healthcare workers to change their strategy for achieving their tasks. An affective awareness also can have implications on activity awareness in how the triage nurse foresees the impact of their future assignments. All of these observations have implications for the design of information systems used in an ED. Current information systems support status but not quite fully enough. There is no support for timing or order – it is all in the head of the charge nurse. Bringing the appropriate information to the surface from each of the attending nurses can help in this awareness and decision making process.
Open Source Software Projects In the domain of open source software development, we are studying the integration of usability activities. Open source software refers to the “development of software through collaborative, informal networks of professional or amateur programmers, and the networked distribution making it available to developers and end-users free of charge” (Ghosh, 2005). As such, this development involves a complex ecology of developers and users. Open source communities produce software in a complex ecology. The ecology supports a socio-technical infrastructure where developers collaborate over electronic networks. Information flows through networks and informal channels of communication. The communities are project centric. Developers self-select work. The work features many users and developers looking
Activity Awareness and Complex Teamwork
53
for problems, suggesting changes, fixing problems, and using the software they are working on. This work is fast. Software versions are released often. Because the work is distributed, a technological infrastructure supports it. Communication technologies such as email, interrelay chat (IRC), and web forums support coordination and collaboration. Despite challenges, such as weakened social presence, open source communication occurs through electronic media, mainly mailing lists because no other alternatives exist (Yamauchi, Yokozawa, Shinohara, & Ishida, 2000). The rational culture of open source communities is characterized by a propensity to put forth logically plausible solutions that lead to technologically superior options. This rationality is evident in email communications. Coordination is managing dependencies among activities (Malone & Crowston, 1994). Coordination challenges particular to open source software include organizing a distributed team of volunteers in order to align their work for a release (Michlmayr, Hunt, & Probert, 2007) and bringing together modules of code (Narduzzo & Rossi, 2005). Modularity, the breaking up of software architecture into many small components, has been a key element in the success of open source software development. In addition, Lindman (2007) found that coordination is based on technical determination. In this sense of technical determination, open source developers believe that problems have technical solutions which will be found when the best developers focus on the problems. A major challenge for coordination in open source software development is that developers tend to act before declaring their commitment (Yamauchi, Yokozawa, Shinohara, & Ishida, 2000). This bias toward action leads developers to coordinate their work only after they have completed their self-initiated tasks. Although communication and coordination among developers and users involves monitoring and integrating many different kinds of information, it is mostly at similar levels of analysis. The level of analysis at which developers and users articulate problems and solutions is a techno-centric, rationalist perspective. Because of this perspective, developers and users of open source software have difficulty with the user-centric nature of usability. Therefore, activity awareness surrounding usability occurs on at least two different levels of analysis, making the integration of usability activities particularly challenging. The usability of open source software is problematic (Nichols & Twidale, 2003). Usability is the effectiveness, efficiency, satisfaction, and fun experienced by an identified group of users when attempting to achieve a specified set of tasks in a particular context. Open source communities, however, generally lack both the resources and knowledge to integrate usability activities that ensure a usable product. Such activities include bringing on usability experts or adopting usability best practices. A major challenge to integrating usability activities with respect to activity awareness includes sense-making across levels of analysis. This challenge arises because technical developers and users have difficulty negotiating information about usability. Usability information tends to be fluid, focusing on the dynamic interaction of people and technology. Information about usability holds no clear answers. Conversely, technical information about software is always about whether the software works properly, that is, it does not crash. Without usability experts, discussions about usability issues tend to be self-referent (Nichols & Twidale, 2006). This means that the scope of usability is limited to the experience of technical developers and users. However, the user base of open source software is much broader than this. Because discussions are open and easily accessible, studying activity awareness in open source software development is feasible. Tracking issues such as usability bugs or feature changes reveals the difficulty developers and users have with usability. In preliminary studies
54
John Carroll
of discussions on usability issues in bug trackers and email lists of three open source projects, we found that most contributors are not aware of usability activities as a whole. Furthermore, the projects that deliberately integrate usability into their community have infrastructures such as dedicated usability experts who share information and post on usability email lists. One challenge for usability experts is to track usability problems in the bug tracker and ascertain whether they are being fixed or whether they are related to other issues. Finally, one email discussion thread about the usability of a proposed feature revealed that when a usability expert posts research supporting rationale for the feature, the users and developers do not address the research results in their discussions and continue using self-referent rationale or analogy based on logic to support their claim that a feature is usable or not. To address such activity awareness issues related to the difficulty contributors have with transcending the different levels of analysis between user-centric and techno-centric perspectives, open source projects can encourage their usability experts to continue posting rationale with detailed explanations for certain design choices that either enhance or decrease usability. Whether usability experts are available or not, open source projects interested in supporting usability activities can do so by increasing the awareness of such activities with deliberate tools that track usability issues and offer a toolbox of usability design techniques that help developers get beyond analysis of usability issues based on their subjective experience.
Local Non-profit Organizations (NPOs) One of the ways we address activity awareness issues at the local non-profit level is with our work in community informatics. Non-profit organizations provide a unique look at how citizens are involved in their community. We are using participatory design methods with local NPOs as stakeholders to inform the design of mobile applications. Encouraging the use of place-based mobile applications by NPOs and the community can encourage new kinds of civic engagement. These applications provide collaborative opportunities both within and between the NPOs themselves, and, perhaps more interestingly, allow non-profits to interact with the local community. Non-profit organizations tend to have limited resources at their disposal (McPhail, et. al, 1998; Trigg and Bødker, 1994; Te’eni and Speltz, 1993). This is due, in part, to a heavy focus on recruiting volunteers (Corder, 2001). Many of the people who make up the nonprofit workforce have limited experience and skills in managing information systems, which impedes training and adoption of important technologies at the non-profit level. Couple this with high volunteer turnover, and the integration of information technologies becomes quite difficult. At other times, non-profit organizations are working with outdated equipment, and their solutions for technical problems are often developed offhand (Merkel, et. al, 2007). Despite these limitations, we have worked with non-profits to help them manage their collective technical expertise (Merkel, et. al, 2004) and incorporate information systems into their daily activities (Carroll, et al., 2000). In our current work, we are helping these organizations collaborate via technology in innovative ways. The collaborative activities at the non-profit level can be described in two ways: those activities among the non-profits, themselves, and those activities between non-profits and the rest of the community. From a non-profit perspective, collaborating with other non-profits
Activity Awareness and Complex Teamwork
55
may not be mutually beneficial. The mission statement of the potential non-profit, as well as those mechanisms that define inter-organizational operation can be barriers to successful alliances (Guo and Muhittin, 2005). However, when non-profit collaborations do happen, the end result can benefit the community at large. Within non-profit organizations, collaboration happens “when different non-profit organizations work together to address problems through joint effort, resources, and decision making and share ownership of the final product or service” (Guo and Muhittin, 2005). Pooling resources to work toward a shared goal enables non-profits to do more than they could on their own. Even though some organizations may be reluctant to collaborate—resisting in an effort to maintain their independence—they may choose to work with other non-profits in ways that help them increase their share of resources (Galaskiewicz, 1985). Except for the tangible benefits to the community, people are generally unaware of and have no say in the specific activities that govern the operation of non-profit organizations. We are interested in those activities that engage the non-profit community along with other community members (residents and local government, for example). We are interested in studying ways that people collaborate using a community wireless network. Results from participant interviews and surveys bear strongly on our thinking in terms of developing applications that are most meaningful to users. Specifically, we are developing locationsensitive applications that allow non-profit organizations to interact with the community, and vice-versa. We have developed several application designs and scenarios in an effort to understand the kinds of applications that support awareness of their community activities. Place-based blogging is an application that would support communicative interactions about local events. For example, a non-profit specializing in water quality could host a blog at a local stream to discuss the impact of runoff and engage visitors in their observations and ideas for improvement at the site. Volunteering-on-the-fly is another application that both supports non-profit activities and serves as a vehicle for informing the surrounding community. With this application, volunteers can be mobilized in real-time, and, if necessary, quickly. Awareness is a primary consideration for these kinds of activities, since collaboration depends on people receiving and subsequently responding to the notification. Integrating information about tasks and user goals are critical for studying activity awareness on a community wireless network, and it is important that these levels of analyses be brought to bear. For example, non-profit goals are constrained by their overall mission, whereas resident community members are motivated by goals that are more personal. Our design concepts explore what is important for activity awareness. They show: 1) how the activities of non-profit organizations can be integrated into the community and 2) how the community, by and large, might interact with non-profits. By exploring these collaborations, we learn how to better design information systems for wireless communities.
Scientific Communities We are investigating the feasibility and effectiveness of supporting activity awareness to support complex teamwork in collaboratories. The primary focus of collaboratories is on the social and collaborative aspects of distributed scientists working together online. Through collaboratories, scientific communities can share key intellectual resources that allow
56
John Carroll
colleagues located anywhere to access, view, manipulate, and have discussions about these artifacts (Finholt and Olson, 1997; Kouzes et al. 1996). The network of people and resources in collaboratories relevant to one’s current scientific activity is quite dynamic; through the course of a project, one may move through a series of foci. The number and variety of communities, forums, and channels for such interactions are vast, even within one collaboratory. Indeed, the recent growth of interdisciplinary scholarship—particularly in the sciences—has made contemporary scientific communities more complex and open but also less identifiable. It is a challenge for scientists to keep track of research issues interesting to them and understand how they are influencing communities over time. A concrete way to think about this challenge is that contemporary scientists need to be aware of a far wider range of colleagues and research topics to engage in effective teamwork over time. But clearly the solution is not to thoroughly read or even manually browse ever-widening swaths of research. That is not humanly possible. Fortunately, these new needs are concurrent with powerful new infrastructures for research in the form of collaboratories integrated with digital libraries and the Internet homepages of researchers throughout the world. The answer to the challenge of how scientists can become aware of relevant colleagues and resources can be addressed, at least in part, by investigating new sorts of digital tools to apprise them of events and teamwork pertaining to their current scholarly activities in collaboratories. Our study context is CiteSeer (http://citeseerx.ist.psu.edu), a scholarly digital library for computer science. We are investigating the support for activity awareness in the CiteSeer collaboratory. While activity awareness can be supported in many ways, a requirements elicitation survey (Farooq et al. 2007a) of CiteSeer users led us to consider three types of activity awareness mechanisms. (1) Awareness of collaborators’ creative activity in distributed settings. A central aspect of and reason for scientific collaboration is creativity. We are studying creativity in distributed teamwork in the context of complex tasks such as collaborating on writing a research paper. Based on an empirical investigation (Farooq et al. 2007b), we identified four breakdowns where activity awareness could support complex teamwork. First, during cognitive conflict and dissent, one of the creative breakdowns we observed was the underconsideration of minority ideas. This was mainly due to normalization or majority influence in the group, resulting in the dismissal of dissenting ideas that may have been novel. Second, the novel ideas generated and narrowed down by group members in prior interactions did not fully carry over to subsequent interactions, were not readily available for review, and/or could not be easily integrated. As a result, novel ideas were easily lost, either for part of the group interaction or for the entire duration of the task. Third, groups made hasty decision in choosing which ideas to converge on. This resulted in a lack of critical evaluation of perspectives. Fourth, we noted that the groups exercised weak reflexivity during convergent thinking during which the members dissipated in their collective effort to develop a coherent product. The creative breakdowns we identified, in general, highlight the need for groups to have their own work re-presented to them through activity awareness mechanisms. For example, our second breakdown suggested that novel ideas got lost. If an idea has not been commented upon and no one has ranked it based on some specified time threshold, the system can make the group aware that the idea has been dormant for some time and may prompt the group members to comment on it. In this way, the group is made cognizant of a possibly good
Activity Awareness and Complex Teamwork
57
idea that may otherwise get ignored. In general, recommendations to collaborators can serve as reminders to critically evaluate different perspectives in the idea workspace. (2) Awareness of scholarly resources through notification systems. Scholars have always organized themselves into communities of peers to track the intellectual work they produce. In today’s digital era, the number and variety of professional channels for tracking the network of intellectual resources relevant to one’s scholarly activity is highly dynamic: through the course of a career, or even of a project, one may move through a series of research foci. Awareness of scholarly resources presents novel opportunities to the brute force solution of thoroughly reading or even manually browsing ever-widening swaths of research. One concrete way to operationalize this is to proactively subscribe to dynamic feeds in order to stay aware of new publication events (e.g., an alert to notify the publication of a paper by an author who cites one’s prior work). Over time, this collected body of publication information identifies: who is currently active in the field, what they are working on, how it relates to other work in the field and who else is interested in this work. While a manual CiteSeer search on “tactile interfaces” might give you all the instances of that phrase in that site’s database at that point in time and who wrote them, aggregating what occurred over time with who created and/or is seeking those resources will answer questions like: Is this topic active now or when was it active in the past? Who else is publishing or even searching in this area? Are they active now and/or when were they active in the past? (3) Awareness of intellectual trends as defined by the community. The contemporary Web has popularized social bookmarking services. These services allow users to stay aware of web resources that are of interest to them by specifying keywords or tags. We focus on social bookmarking services for scholarly communities in which users collectively organize and tag intellectual resources. Tagging as a long-term activity is a form of activity awareness. Part of this awareness is knowing what the scholarly community values and what it doesn’t value. This can include intellectual trends and state of a field, most popular papers in a research area, “hot” topics and research themes, and so forth. Supporting tag reuse and tag convergence is important so that the social nature of tagging can make such awareness information easily accessible and shared. For CiteSeer, we are considering these heuristics. We have the goal of explicitly encouraging tag reuse and convergence as a means of facilitating a more effective and collaborative scientific community. In the domain of scientific communities, we have focused on complex teamwork and support for activity awareness around scientific literature where the vehicle for sharing much of it is written media such as journals and presentations at symposia or conferences. The need for activity awareness also ties to other aspects of scientific collaboration. Traditionally, study of collaboration in e-science has focused on collaboratories based around remote work with unique and/or costly instruments (radio telescopes, supercomputers, etc) or very large datasets (census records, the human genome, etc). Activity on these research instruments and datasets are also means of documenting trends in science. Trends in the use of a scientific instrument could be monitored through its configuration parameters and subjects of its use, while trends in operations on large data sets could be monitored through the filters and formulas of analysis projected on the data. From a purely research perspective, privacy and intellectual property concerns noted, awareness of activity on shared data and resources can help researchers understand and develop their own activity as positioned in networks of research
58
John Carroll
colleagues and research topics. How such activity data can be aggregated and visualized in a meaningful way for scientists may vary from field to field and warrants further study.
Collaborative Education in Technology and the Classroom The ability to work in a collaborative team is an expected and desired outcome in many educational domains, such as business, law and usability engineering. At the same time, collaborative teamwork is also a means to improve other learning outcomes, such as an increase in domain knowledge or better argumentation and reasoning skills (Kollar, Fischer, & Slotta, 2005). In other words, collaboration is a means to an end for learning and educational goals. The ability to collaborate effectively may not be a natural skill in students, but rather something they have to learn and practice in order to develop competence. However, in order to be able to develop collaborative competence students need to engage in the types of interactions that lead to mutual awareness of group processes as well as content related tasks. Interactions leading to knowledge convergence and integration of collective perspectives and knowledge bases require that teams share and negotiate ideas effectively. In order to accomplish this it is necessary to structure collaborative activities so as to scaffold the use of metacognitive processes such as planning, reflecting, so students can better control their group processes. For modern education, preparing students to participate in collaborative endeavors by developing these critical skills becomes an essential responsibility, one that we are addressing through technological development and classwork instruction.
Collaborative Education and Technology Researchers in management and group research generally see teamwork ability as a fundamental knowledge set necessary to achieve shared goals, along with task-work knowledge. Studies from these domains have paid close attention to aspects of group work, such as team formation, team process and outcomes (for review, see Kozlowski & Bell, 2003). From a human development perspective, scholars, such as Vygotsky (1978) and Tomasello (1999), claimed that engaging with others during learning can play a key role in developing higher mental functions, such as higher order memory, problem-solving and reasoning. In this line of research, social interaction is seen as a major and necessary mechanism that helps shape higher human mental functions. For Vygotsky (1978), higher level mental functions develop at inter-psychological (social) level first, and then by participating in social engagements, individuals internalize those higher mental functions and contents into intra-psychological (individual) plane. The example of child pointing shows that the behavior of pointing is symbolized and signified socially, and later becomes internalized as a mental function of an individual kid to seek help from more competent adults. The concept of zone of proximal development (Vygotsky, 1978) suggests that help from more competent peers or instructors can lead to better human development in students. Inspired by the insights from management, group research and developmental psychology, and driven by the practical need of teamwork in students, we believe that computer-supported learning environments should improve both teamwork skills and task-related knowledge in students, to help achieve educational goals.
Activity Awareness and Complex Teamwork
59
Advanced information technologies enable distributed collaboration both synchronously and asynchronously, but using technology mediated channels also impedes collaboration. For example, distributed systems usually have lower channel capacities to deliver rich information than face-to-face communication does (Trevino, Lengel & Daft, 1987). Although technologies have advantages in many aspects, such as revisibility, traceability, and less production block, they are not good at delivering social-emotional cues with high fidelity. In face-to-face situations, collaborators perceive those cues directly and are often forced into engagement. Recently, researchers have pointed out that a key component not well supported in Computer Supported Collaborative Learning (CSCL) systems is social interaction, which, in fact, is the base for collaborative work (Kreijns, Kirschner & Jochems, 2003; Kreijns, Kirschner, Jochems & van Buuren, 2007). Kreijns et al (2003; 2007) reviewed related research and argued that the current CSCL systems have failed to support social interaction, compared with functional and educational supports with technologies. They further identified two pitfalls that contribute to this failure. First, people assume that social interaction will always happen or be meaningful, as long as communication happens. However, this is not always true. Second, many researchers restrict social interaction to cognitive processes, ignoring other dimensions of social interaction, such as affection, commitment, and affiliation (Nardi, 2005). Schön (1983) argued that a fully functioning practitioner always tacitly engages in a personal self-critique in the context of professional activity, which we believe includes the psychological collaborative environment, such as group status (e.g., group efficacy, group coherence). To improve collaborative work and collaborative learning, CSCL systems must support awareness at team level that keep members informed during their course of action, synchronously and/or asynchronously (Carroll, et al, 2006; Farooq, et al, 2008). Activity awareness (Carroll, et al, 2006) is an overarching and promising framework that addresses both individual and collaborative work. Activity is an encompassing analytical unit, which includes contextual factors, human intentions, means (artifacts) to achieve goals, and influence of community where the practice bounds. The framework of activity awareness aims to support awareness from communication level to human development level. In this framework, the four components—common ground, community of practice, social capital and human development—all relate to collaborative work and human mental development. Our previous studies have shown successful cases in which activity awareness contributes to the collaborative work among students and teachers (Ganoe, Somervell, Neale, Isenhour, Carroll, Rosson, & McCrickard, 2003; Carroll, Neale, Isenhour, Rosson, & McCrickard, 2003). To support various social interactions and improve collaboration skills, CSCL systems must support awareness to highlight activities taking place and their consequences. CSCL systems should allow and encourage activities other than task execution, and those activities could be, for example, planning, discussing, commenting and (re)negotiating. At the same time, team level awareness should be made ostensive to team members and instructors, so that the collaborations among teachers and students groups are possible, in terms of group facilitating or tutoring. Some current group systems (e.g., Google Groups) and CSCL systems (e.g., BRIDGE—Basic Resources for Integrated Distributed Group Environments) have taken steps towards supporting team level activity awareness. For instance, Google Groups (http://groups.google.com/) provides each registered group a profile on which statistical information about group activity displays, such as intensity of group activity, how many documents and discussion have been posted. Many studies (e.g., Bostrom, Anson & Clawson,
60
John Carroll
1993) showed that group facilitators, persons who observe group activities and intervene when necessary, can improve group performance and group processes. Classroom BRIDGE (Ganoe et al, 2003) has shown the activity awareness and effect of this kind. In this case, Classroom BRIDGE uses a timeline to visually show intensity and productivity of teams, and a teacher who was aware of something problematic going on took actions to improve and correct performance of a student group. In CSCL systems, the technical support of task or expertise activity and that of social interaction should go hand in hand. CSCL systems can improve learning outcomes, in terms of collaboration skills and other desired outcomes, by supporting more social interactions with activity awareness, sharing information about team and other members. Many important constructs and concepts from social psychology, sociology and management are critical for collaborative work, such as efficacy, esteem, trust, social capital, social loafing, and conflict. One issue worthy to explore is to support these concepts with activity awareness. For example, group conflict is very important information of which team members should be aware. Conflict resolving is a team process, which helps a team reach consensus and shared vision. It also leads to group cohesion and potential viability. Given a CSCL system, if it can highlight conflicts (e.g. schedules, goals, plans, interests) in legible forms and thus provides awareness of those conflicts, student groups, as well as their facilitators, will be able to perceive those conflicts easily and take further actions. When a conflict is resolved, the system can also visually provide such awareness, so that members and their facilitators can see the progress of the team. If we put this information on a time scale, the development of the team will be much easier to perceive. With this type of awareness, the systems can help teams and their facilitators identify the potential breakdowns and develop solution for students. This is especially critical to a system with educational purpose.
Collaborative Education in the Classroom It is fairly common for students to think they fully comprehend something when in actuality many holes exist in their conceptual understanding. Collaborative activities have the potential to minimize this likelihood by requiring students to discuss learned concepts, question new ideas, and ultimately challenge each other’s understanding (Matusov & Hayes, 2000; Webb and Palincsar, 1996; Piaget, 1976). These types of social interactions help students realize that mismatches exist between their perceived and their prevailing grasp on new ideas, processes, or activities. Cognitive gains are not the only merits of collaborative activity; group products and outcomes can also benefit from collaborative interactions. In business, for example, successful entrepreneurial ideas are more likely to come from teams than individuals (Gartner, Shaver, Gatewood, and Katz, 1994). Furthermore, entrepreneurial teams are most likely to be successful if they are composed of members possessing diverse knowledge bases, perspectives, and abilities, but only to the extent that these resources can integrated by the group (West, 2007). Unfortunately, herein is where the problem lies and where collaboration often fails. Often teams do not take the time or make the effort to share, analyze, or deliberate ideas and those that do often lack the capability to monitor and regulate interactions so as to effectively utilize these contributions. This type of discourse and knowledge building, where
Activity Awareness and Complex Teamwork
61
individual perspectives merge to form a group’s collective mindset, is what leads to mutual understanding and reciprocal awareness of the activity at hand. Thus due to dysfunctional interactions, those that interfere with rich discourse and knowledge convergence, potential gains from collaborative activities may not come to fruition. These dysfunctional interactions may prevent team members from attaining or maintaining reciprocal awareness of problems, resources, knowledge, needs, expectations, goals, or other crucial fragments of information that the team could utilize. Dysfunctional interactions can result from many different types of problems: social problems, (i.e., disrespectful treatment of others, social loafing, hierarchies, etc.), organization problems, (i.e., time mismanagement, inadequate preparation for meetings, etc.), or a problem with team’s approach to the activity, (i.e., meeting goals, how work is completed, expectations, etc.). While all of these types of problems can lead to dysfunctional interactions, a teams approach to a collaborative activity can be especially problematic. A teams approach can narrow the scope of possible interactions during group meetings. This, in turn, can limit the range of intellectual activities and possible learning/ product outcomes. For example, some teams may decide to forgo discussions, reflective activities, or other processes that they perceive as interfering with their ability to finish a product quickly. For these teams, meetings serve as a means to manage the product workload. They meet to delegate work and check-in on what has been already accomplished rather than to discuss, build on, or explore ideas and understanding. Thus the prevailing group dynamic will be one that emphasizes product completion over the quality of collaborative processes. Even though the instructor’s pedagogical intent may have been for robust collaborative-interactions to take place, in actuality very little intellectual activity will occur during these meetings. Such groups are unlikely to spend time discussing ideas at length or reflecting on their understanding and may therefore fail to reap the full benefits that collaborative activities provide. It is doubtful that students will be highly motivated to spend time managing group processes in classes where collaborative interactions are not modeled or assessed. If students’ grades are primarily determined by end products, quizzes, or finals, then students’ primary objectives will be to do well on these aspects of the course and not waste time elsewhere. By elevating process-work to the same status as content-work we hope to motivate students to monitor and regulate their interactions so they may fully benefit from their collaborative interactions. Now, anyone who has worked with students and used collaborative activities in their classes is probably asking the same question, “how can instructors help students to monitor and regulate group-processes? We believe this can be accomplished by structuring a team’s approach to a collaborative activity through the use of sociocognitive roles and holding them accountable for this work. We call this activity engaging in process-work. This method, originally developed by the ThinkerTools Research group, helps students to manage collaborative interactions through the use of process-oriented roles (Borge, 2007; Borge, White, & Miller, 2005; White & Collins, 2000). These roles break down effective collaborative activity into smaller, manageable parts. Detailed guides help students to operationalize the selection and use of strategies to achieve necessary process-goals. Through role management students learn about, monitor, and practice regulating one piece of collaborative activity. Students can learn about the other aspects of effective collaborativeinteraction by watching team members manage other roles. However, the roles we currently use are slightly different from the originals in that they, along with their guides, were
62
John Carroll
modified for use with adult engineering students. Our four roles focus on: joint planning and decision making, communication, productivity, and mediation of ideas and conflict. Students are provided with instructions and guides to help them perform specific actions related to role management. These guides make explicit the various operations to identify problems, select strategies, and actively participate in improving their group’s interactions. They also distinguish between two different kinds of activities that students must establish and maintain awareness off in order to be a successful group: product work, and processwork. Product work refers to creation of an end product or deliverable. The roles make sure to emphasize careful planning, analyses of requirements and resources, and overall quality of the product. Process-work, on the other hand refers to the creation and maintenance of an effective collaborative environment. To this end the roles focus on the quality and extent of communication, and negotiation of ideas. The roles help students to focus on both aspects of process and product work by planning and reflecting on how to meet specified goals. Below are some example excerpts from the our new communication and productivity guides: Communication Manager: Goals and Strategies Guide Goals: You are trying to create a team in which everyone participates verbally, but are concise and stay on point, members listen and understand each other, they build on each other’s ideas, and they come to a common understanding. Problem: Someone isn’t talking Strategy: Take turns in talking. “Pat and Saul have already contributed. Amir, what do you think...” Strategy: Choose someone who hasn’t spoken in a while. “Sasha hasn’t commented yet, what’ your take on...” Being concise and staying on point: Problem: Someone is rambling or digressing. Strategy: Remind them that a goal of the course/ communication is to give only relevant information that is direct and to the point (concise). “ Let’s get back to discussing trade-offs, I think we are veering off course”
Productivity Manager: Goals and Strategies Guide Goals: You are trying to help the team to monitor its progress and be productive, as well as to evaluate and improve its products. Problem: Deliverables are not clear or concise. Strategy: Have someone outside of you team read your deliverables to ensure that they can be easily understood. Strategy: Schedule time to check each other’s work prior to submitting (peer editing) in order to reduce wordiness and improve readability. Strategy: Read materials on writing style, i.e. Shrunk and White, Elements of Style, or others listed at: http://www.personal.psu.edu/jth/Engl497.html
Activity Awareness and Complex Teamwork
63
The excerpts above illustrate how different roles focus on process-work and productwork, respectively. The communication guide is dealing with discourse issues while the productivity guide is trying to prevent or correct problems with the end-product. Both of these guides contain more information than is presented. Each guide is roughly two-three pages long. Besides revising the tools to support the roles, we are also developing new ways to assess collaborative competence, the ability to effectively monitor and regulate group processes. This requires 1) testing students’ knowledge of the goals, problems, and strategies associated with collaborative activity and then 2) assessing their ability to use this knowledge to prevent or correct problems. We are in the process of creating an interactive video assessment to help us measure the latter ability. This assessment may eventually help instructors to assess the abilities students bring to the table as well as how much they improve during the course of their training. Currently this training method is being implemented in a senior-level usability engineering course. Students are being graded on their group processes, group products, and individual performance in the class. In order to promote use of the roles and proper reflection on goals and strategies, students are required to create archives, or records, of their group meetings. Each student must create an archive written from the lens of their particular role. For example, the communication manager has to write down which goals they will focus on and which strategies they can use to meet these goals. These archives can help to “record” what gets accomplished during a meeting: how the view of a shared plan and project is carried out, and how it evolves over time. Activity awareness is just the beginning. If students are to create and maintain effective collaborative interactions activity awareness has to be utilized by students: as a means to better understand their team members, create and define group expectations, and regulate interactions to meet expectations. Through this type of research we hope to develop better ways of structuring, supporting, and assessing activity-awareness and group-process regulation.
Tool Strategies for Activity Awareness The fieldwork we describe above covers a wide range issues related to activity awareness, which tools might support. Examining these teamwork settings leads to many implications for system design. We have previously described how common ground, communities of practice, social capital and human development are important facets of activity awareness (Carroll et al. 2006). These pertinent properties may not be self evident or easy to capture and display in an information system. The greatest challenge for technology here may not be in some new visualization or interaction technique, but in deciding what details of this awareness data provide the most benefit for the current situation and how to capture aggregate and summarize them. We had suggested three design goals (Ganoe et al. 2003) when designing user interfaces to support activity awareness: integrated presentation of awareness information, incidental creation of awareness information, and public access to all group activity information. To support the four facets we have added aggregation and contrast to this list of design goals and presumed that incidental collection of information is crucial to all the facets. For the
64
John Carroll
purpose of discussion we mainly refer to our distributed emergency management planning prototype in BRIDGE (Convertino et al. 2007 and In Press), but discuss our other fieldwork and software as appropriate. Table 1. Information types and design techniques by sub-processes of activity awareness Process Common Ground Communities of Practice Social Capital Human Development
Examples of Information to be Shared common or similar interests, surroundings common or similar training, experiences, events, level of experience common or similar locales, levels of risk taking knowledge of changing roles and learning
Suggested Design Technique public availability of all information everywhere integration of diverse datasets into metaphors that coincide with the community of practice aggregation of each individual’s parts into the organizational whole contrast changing of individuals over time
The incidental creation of awareness and other types of information is known to be a challenge for CSCW systems. Expanding this over the long-term and making this information accessible across many systems are yet to be resolved challenges. If such context information were available, it would need to be publicly available to accelerate formation of common ground. Constructing common ground could be as basic as knowing that responders are all familiar with the local area when rescuing people from a flash flood. It can be more subtle and detailed like: I know that you grew up near the flood location, while I have been involved in flood emergency planning, just not this specific location. In our emergency management planning prototype, all planners have access to a shared team view where they can see each others’ annotations and telepointers. Expanding awareness of common ground over longer-term activity typically requires capture of some form of team event data. Our Classroom BRIDGE software (Ganoe et al. 2003) tracked significant changes to document objects (text, calendar, chat, etc.) for student group projects and then provided interactive timeline visualizations of that activity to each project team. There was also a public screen that aggregated project activity across all the project teams in the class allowing the teacher and other project teams to view relative progress. Begole et al. (2003) took another approach to this by modeling rhythms in peoples’ daily office activity. Through these typical patterns of a person’s work day, they created a contact list for their Awarenex communication system that predicted where people might be (e.g. lunch) if they weren’t in the office an estimated time of when they might return. They suggest that by overlapping this information for multiple people, it can aid in planning communication and scheduling meetings. Information that is shared detailing awareness for communities of practice can focus in on similar or common education and training, details of work experience (along with the level of experience in those areas), or knowledge of similar methods and/or tools. One goal might be the integrated presentation of this information into a context that fits within the community of practice. LifeLines (Plaisant et al. 1996), for example, provides a graphical view into youth
Activity Awareness and Complex Teamwork
65
records for the Maryland Department of Justice and individual patient records for medical practitioners. The intent here is to provide events in a patient’s personal history as a timeline visualization, which maps to the work context of the system’s user. In our emergency management planning prototype, we provide separate team and role-specific views, where the role view changes for public works, environmental and mass care public officials. Depending on a user’s role, they see different map details and icons (data layers) specialized to his or her role. Supporting social capital, in our emergency planning scenario, includes knowing that we are working together protecting the same community, and that we are first responders taking risks and often volunteering time for the common good. It can also mean relying upon others and knowing where you expect the ones who are backing you up will be, or knowing that others have resources available to share. In non-profit organizations and open source software project communities, this includes maintaining a network of volunteers, keeping track of and managing the development of resources. While some social networking Internet sites have become popular (e.g. Facebook, MySpace) and more synchronous approaches have been explored toward social proxies, such as Babble (Ericsson et al. 2002), support for social aspects of activity awareness are less explored. We see the CiteSeer community as a rich opportunity to provide awareness of how social aspects of activity awareness develop within a scientific community. To support human development, systems must take roles into account but provide flexibility for users to expand upon those roles. Herrmann et al. (2004) examined how role development occurred for four groups over two weeks in the KOLUMBUS collaborative learning environment. They concluded that the development of roles can be detected within on-line groups, and that there is a gap in between the current views of roles in computer science vs. social sciences. Current CSCW systems lean on these static, fixed authorization/permissions based structures of computer science while roles from a social sense are dynamic, always changing and evolving. In our emergency planning prototype, each role can annotate their own role-specific map view, but can also share those annotations to the team map and edit the overall plan with their collaborators. The resulting plans can be stored over time, alternatives tried and changes tracked. In our wConnect professional development community, one might track the long-term progress from being a college recruit with questions, to undergraduates answering questions about their experience within the program, and through alumni describing job experiences. The concept of time itself becomes an important design consideration in activity awareness. Even under the “emergency” response moniker, we can contrast two distinct examples: the highly asynchronous emergency management teams who meet annually (though they are often first responders in their “day jobs”) to plan their response for something we hope never happens versus the hospital emergency room workers constantly seeing new patients as they arrive and moving between then during the process of their stay. Our prototyping in BRIDGE has attempted to bring many of the above design considerations together. With the shift to Web 2.0 and mashup approaches to working online, we anticipate the need to aggregate activity awareness information from many Internet sources into meaningful summaries. To this end, we have begun to explore more standardsbased mechanisms for activity awareness such as RSS for describing changes to different
66
John Carroll
types of content within BRIDGE (Hylton et al. 2007) and also design of RSS feeds as a means of tracking scholarly change within the CiteSeer community Farooq et al. 2007a).
Measuring Awareness in Teamwork More that two decades ago, field researchers had observed that when people work together in the same place for a long time, they tend to naturally align and integrate their activities in a seamless way. That is, they develop mutual awareness (e.g., Harper et al. 1989) or the emerging ability to monitor and integrate work. However, Collaborative Computing research is still missing apposite methods for measuring and theory for explaining the awareness-building process in distributed and long-term collaborations. Two research traditions have investigated awareness in tool-supported collaboration. Researchers in Human Factors have studied situation awareness in synchronous, informationintensive settings. They focused mostly on non-discretional technology users, such as aircraft crews and air traffic controllers. Key contributions from this tradition have been the development of cognitive theory and standard empirical measures of situation awareness (e.g., Endsley and Garland 2000). Differently, researchers in the Computer-Supported Cooperative Work (CSCW) research community has been very prolific in developing and informally evaluating novel tools for supporting awareness. Their studies have focused mostly on discretional users, such as knowledge workers using collaborative editors and media spaces. This second tradition has put less effort into the development of operationalized definitions and standardized methods for studying awareness (see Schmidt 2002 about low consistency in the definitions). This makes it difficult for these researchers to accumulate related findings across independent studies. In the various studies of awareness presented in this paper we are working towards better operationalization and measurement of the activity awareness phenomenon in specific domains where the articulation work among collaborators is distributed and protracted in time. Proposed as a programmatic definition in Carroll et al. (2006), Activity Awareness is construed as a broad process that builds on four basic sub-processes. These are the development of respectively common ground, shared work practices, social capital (reciprocity and trust), and human skills, among the actors in the activity (see Table 1). An important goal for future studies on activity awareness is the development of a comprehensive (i.e., multi-method) and theory-grounded approach to measurement. The above-mentioned four sub-processes of activity awareness suggest four useful theory-based foci for measurement. Note however that since the methods were imported from various domains, different sets of methods have been used for measuring such sub-processes in the past (see review in Carroll et al. 2006). For example, typical studies on common ground are laboratory experiments on communication with small groups; the studies of communities of practices are usually based on direct observations of behaviors and artifacts in the field; studies of social capital are often based on surveys over large populations, and the studies of human development tend to use repeated measurements of behaviors and subjective assessments of individuals or groups.
Table 2. Attributes of Activity Systems and Related Methods of Investigation
Regional Emergency Management Development Professional Communities Hospital Emergency Room Operations Open Source software Projects Local Nonprofit Organizations
Unit of analysis Teams of emergency management workers A community of female students
Setting (spatial, temporal) - Space: Organizations in rural townships; face-to-face. - Time: Infrequent meetings, synchronous (e.g., biannual) - High school and university, distributed. - Year-long, asynchronous Teams of - Hospital ED, collocated. healthcare workers - Hours-long and ongoing in a hospital activities, synchronous and division asynchronous Communities: of - Virtual/online, distributed. volunteer - Months-long to year-long, programmers, mainly asynchronous usability experts Communities: non- - Within a town, augmented profit organizations reality, distributed. in one municipality - Months-long to year-long, mainly asynchronous
Scientific Communities
Communities of scientists, collaboratories
Collaborative education in a classroom
A large group (class) and several small (project) teams
Tools used Shared physical maps, telephone, official reports
Activity goal/type Planning and training, tabletop walkthrough exercises
Networking and community building tools Electronic patient records and coordination tools (e.g. whiteboards) email and mailing list, chat, web forums, bug trackers
Professional development, social capital development Patient care (hours) and Observations, bed management artifact analysis (ongoing), safety- critical
Community wireless network, locationsensitive applications (e.g., place-based blogging) - Virtual/online, distributed. Digital library - Months-long to year-long, (CiteSeer), notification mainly asynchronous & social bookmarking tools - Collocated during, distributed Digital library of case between classes, semester-long. studies, groupware, - Months-long, synchronous and course and role asynchronous. materials
Data collection Observations of teams, shadowing of manager Observations and interviews
Data analysis Mostly qualitative
Mostly qualitative Mostly qualitative
Software development, focus on usability activities
Analysis of logs Mostly and artifacts, qualitative surveys
Volunteering (e.g., educational outreach), resource sharing, NPONPO and NPOcommunity coordination Collaborating on publications, codeveloping creative ideas Coursework on Usability Engineering, semester group project
Participant interviews and surveys
Mostly qualitative
Surveys and system logs (e.g., tagging logs) Surveys, videos, system logs (e.g., usage, products)
Qualitative, quantitative (statistical tests) Qualitative, quantitative (statistical tests)
68
John Carroll
A comprehensive and consistent approach to measuring activity awareness can be achieved through operationalizations of this construct in well-specified work domains. Also, in the design of the study, the construct (or its sub-components) should be related to known variables and measures in Collaborative Computing research, such as the properties of the work environment (e.g., space and time arrangement), the task (e.g., work coupling) and the tools (see Neale et al. 2004 for an example). In table 2, we analyze the attributes that characterize the activity systems investigated in each of the studies presented above (i.e., actors, setting, tools, goals). We draw the concept of activity system from activity theory (e.g., Korpela et al. 2001). Then we point to examples of methods that were used for studying activity awareness in correspondence with the analysis of each activity system. In our future work, we plan to develop a comprehensive ‘toolset of research methods’, relating the strengths and weaknesses of the methods to the attributes of the activity system. Such taxonomy could guide researchers while designing their studies on awareness in CSCW.
Acknowledgments and Notes This work was supported in part by the National Science Foundation (0454052, 0634337, 073544, 0749172) and the Office of Naval Research (N00014-05-1-0549), the Knight Foundation (2007-02561), Microsoft Corporation (6098505), and Intel Corporation (34251). Gregorio Convertino is now at the Palo Alto Research Center (PARC). Umer Farooq is now at Microsoft Corporation.
References Amouth, T., Gemo, M., Macq, B. Vanderdonckt, J., El Gariani, A. W., Reynaert M. S., Stamatakis, L., & Thys, F. (2005). Versatile clinical information system design for emergency departments. IEEE Transactions on Information Technology 9(2), p. 174183. Ash, J. S., Berg, M., & Coiera, E. (2004). Some unintended consequences of information technology in health care: The nature of patient care information system-related errors. Journal of the American Medical Informatics Association, 11(2), p. 104-112. Begole, J. B., Tang, J. C., & Hill R.. (2003) Rhythm modeling, visualizations and applications. In Proceedings of the 16th annual ACM symposium on User Interface Software and Technology (UIST 2003), p. 11–20., ACM Press. Borge, M. (2007). Regulating social interactions: Developing a functional theory of collaboration. Unpublished Dissertation, University of California, Berkeley. Borge, M., White, B., & Miller, T. (2005). Regulating social interactions: developing a functional theory of collaboration. Paper presented at the American Educational Research Association, Montreal, Canada. Bostrom, Anson & Clawson (1993) "Group Facilitation and Group Support Systems," in Group Support Systems: New Perspectives, L. Jessup and J. Valacich, Eds. New York: Macmillan. Carroll, J.M., Chin, G., Rosson, M.B., and Neale, D.C. (2000). The development of cooperation: Five years of participatory design in the virtual school. In D. Boyarski and
Activity Awareness and Complex Teamwork
69
W. Kellogg (Eds.), Designing Interactive Systems (DIS ’00) Brooklyn, NY, August 2000). ACM Press New York, NY, 2000. 239-251. Carroll, J. M., Neale, D. C., Isenhour, P. L., Rosson, M. B., & McCrickard, D. S. (2003).Notification and awareness: synchronizing task-oriented collaborative activity. International Journal of Human-Computer Studies, 58(5), 605-632. Carroll, J.M., Rosson, M.B., Convertino, G., Ganoe, C.H. (2006). Awareness and teamwork in computer- supported collaborations. Interacting with Computers, 18(1), pp. 21-46. Convertino G., Mentis H., Rosson M.B., Carroll J.M., Slavkovic, A. Ganoe, C.H. (2008). Articulating Common Ground in Cooperative Work: Content and Process. In Proceedings of the Conference on Human Factors in Computing Systems (CHI ‘08). 1637-1646. ACM Press. Convertino G., Mentis H., Rosson M.B., Slavkovic, A., Carroll J.M. (To Appear). Supporting Content and Process Common Ground in Computer-Supported Teamwork. In Proceedings of the Conference on Human Factors in Computing Systems (CHI ‘09).. ACM Press. Convertino G., Zhao D., Ganoe C., Carroll J.M., Rosson M.B.: (2007). A Role-based MultiView Approach to support GeoCollaboration, Proc. HCI International 2007. Corder, K. (2001). Acquiring New Technology: Comparing Nonprofit and Public Sector Agencies. Administration & Society, 33, 2, (May 2001), 194-219. Crowston, K., & Howison, J. (2005). The Social Structure of Free and Open Source Software Development. First Monday, 10(2). Dourish & Bellotti, 1992 Embi, P. J., Yackel, T. R., Logan, J. R., Bowea, J. L., Cooney, T. G., & Gorman, P. N. (2004). Impacts of computerized physician documentation in a teaching hospital: Perceptions of faculty and resident physicians. Journal of the American Mecial Informatics Association, 11(4), p. 300-309. Endsley, M.R. & Garland D.J. (Eds). (2000) Situation Awareness Analysis and Measurement. Mahwah: Lawrence Erlbaum Associates. 147-173. Erickson T., Halverson C., Kellogg W. A., Laff M. & Wolf T. (2002) Social translucence: designing social infrastructures that make collective activity visible. Communications of the ACM, 45(4): 40–44, April 2002. Farooq, U., Ganoe, C.H., Carroll, J.M., Councill, I.G., and Giles, C.L. (2008): ‘Design and evaluation of awareness mechanisms in CiteSeer’, Information Processing & Management (IP&M): Special issue on digital libraries in the context of users’ broader activities, Volume 44, Issue 2, pp. 596-612. The Netherlands: Elsevier. Farooq, U., Ganoe, C.H., Carroll, J.M., and Giles, C.L. (2007a): ‘Supporting distributed scientific collaboration: Implications for designing the CiteSeer collaboratory’, Proceedings of the Hawaii International Conference on System Sciences (Waikoloa, Hawaii, January 3-6, 2007), IEEE Computer Society, 2007, p. 26c. Farooq, U., Carroll, J.M., and Ganoe, C.H. (2007b): ‘Supporting creativity with awareness in distributed collaboration’, Proceedings of the International GROUP Conference on Supporting Group Work (Sanibel Island, Florida, November 4-7, 2007), pp. 31-40. New York: ACM Press. Galaskiewicz, J. (1985). Interorganizational relations. Annual Review of Sociology, 11, 281-304.
70
John Carroll
Ganoe, C., Somervell, J., Neale, D., Isenhour, P., Carroll, J. M., Rosson, M. B., and McCrickard, D. S. Classroom BRIDGE: Using Collaborative Public and Desktop Timelines to Support Activity Awareness. In Proceedings of the ACM Conference on User Interface Software andTechnology (UIST '03), Vancouver BC Canada, Nov 2003 Gartner, W. B., Shaver, K. G., E.Gatewood, & Katz, J. A. (1994). Finding the entrepreneur in entrepreneurship. Entrepreneurship Theory and Practice, 18(3), 5-9. Ghosh, R. A. (2005). Understanding Free Software Developers: Findings from the FLOSS study. In J. Feller, B. Fitzgerald, S. A. Hissam & K. R. Lakhani (Eds.), Perspectives on Free and Open Source Software. Cambridge: MIT Press. Guo, G., and Muhittin, A. (2005). Understanding Collaboration Among Nonprofit Organizations: Combining Resource Dependency, Institutional, and Network Perspectives. Nonprofit and Voluntary Sector Quarterly, Vol. 34, No. 3, 340-361 (2005). Harper, H.R., Hughes, J.A., & Shapiro, D. Z. (1989). Working in harmony: An examination of computer technology in air traffic control. ECSCW'89: Proceedings of the First European Conference on Computer Supported Cooperative Work, Gatwick, London, 13– 15 September, 1989, pp. 73–86. Henneman, E. A., Blank F. S. J., Gawlinski, A., & Henneman, P. L. (2006). Strategies used by nurses to recover medical errors in an academic emergency department setting. Applied Nursing Research, 19, p.70-77. Herrmann, T., Jahnke I., & Loser K.-U. (2004) The role concept as a basis for designing community systems. In Cooperative Systems Design (COOP’04), pp. 163–178. ISO Press. Howison, J., Inoue, K., & Crowston, K. (2006). Social dynamics of free and open source team communications. Paper presented at the 2nd International Conference on Open Source Software, Lake Como, Italy. Koch, S. (2007). Exploring the Effects of Coordination and Communication Tools on the Efficiency of Open Source Projects using Data Envelopment Analysis. In Open Source Development, Adoption and Innovation (pp. 97-108). Kollar, I., Fischer, F., & Slotta, J.D. 2005. Internal and external collaboration scripts in webbased science learning at schools. Proceedings of CSCL 2005 (pp. 331-340). International Society of the Learning Sciences. Korpela, M., Mursu, A., Soriyan, H.A., 2001. Two times four integrative levels of analysis: a framework. In: Russo, NL, Fitzgerald, B, DeGross, JI (Eds.), Realigning Research and Practice in Information Systems Development. The Social & Organizational Perspective. IFIP TC8/WG8.2 Working Conference, Kluwer Academic, Boston, pp. 367–377. Kreijns, K., Kirschner, P. A., & Jochems, W. (2003). Identifying the pitfalls for social interaction in computer-supported collaborative learning environments: A review of the research. Computers in Human Behavior, 19(3), 335-353. Kreijns, K., Kirschner, P. A., Jochems, W., & van Buuren, H. (2007). Measuring perceivedsociability of computer-supported collaborative learning environments. Computers &Education, 49(2), 176-192. Lindman, J. (2007). Shared Assumption Concerning Technical Determination in Apache Web Server Developer Community. In Open Source Development, Adoption and Innovation (pp. 283-288). Malone, T. W., & Crowston, K. (1994). The interdisciplinary study of coordination. ACM Comput. Surv, 26(1), 87-119.
Activity Awareness and Complex Teamwork
71
Matusov, E., & Hayes, R. (2000). Sociocultural critique of Piaget and Vygotsky. New ideas in psychology, 18, 215-239. McPhail, B., Costantino, T., Bruckmann, D., Barclay, R., and Clement, A. (1998). CAVEAT Exemplar: Participatory Design in a Non-Profit Volunteer Organisation. Computer Supported Cooperative Work: The Journal of Collaborative Computing, 7, 3 (1998), 223241. Michlmayr, M., Hunt, F., & Probert, D. (2007). Release Management in Free Software Projects: Practices and Problems. In Open Source Development, Adoption and Innovation (pp. 295-300). Merkel, C.B., Farooq, U., Ganoe, C.H., Rosson, M.B. and Carroll, J.M. (2007). Managing technology use and learning in nonprofit community organizations: Methodological challenges and opportunities. Proceedings of CHIMIT 2007: First ACM Symposium on Computer-Human Interaction for Management of Information Technology. March 30-31, 2007: Cambridge, MA. Merkel, C.B., Xiao, L., Farooq, U., Ganoe, C.H., Lee, R., Carroll, J.M. and Rosson, M.B. (2004). Participatory design in community computing contexts: Tales from the field. Proceedings of the Participatory Design Conference (Toronto, Canada, July 27-31). New York: ACM Press, pp. 1-10. Narduzzo, A., & Rossi, A. (2005). The Role of Modularity in Free/Open Source Software Development. In S. Koch (Ed.), Free/Open Source Software Development (pp. 84-102). Hershey: Idea Group Inc. Neale, D.C., Carroll, J.M., & Rosson, M.B. (2004). Evaluating Computer-Supported Cooperative Work: Models and Frameworks, Proceedings of CSCW 2004, New York: ACM Press. Nichols, D. M., & Twidale, M. B. (2003). The Usability of Open Source Software. First Monday, 8(1). Nichols, D. M., & Twidale, M. B. (2006). Usability processes in open source projects. Software Process: Improvement and Practice, 11(2), 149-162. Piaget, J. (1976). The Grasp of Consciousness: Action and concept in the young child. Cambridge, MA: Harvard University Press. Plaisant, C., Milash B., Rose A., Widoff S., Shneiderman B. (1996) Lifelines: Visualizing personal histories. In Proceedings of the Conference on Human Factors in Computing Systems (CHI ’96), p. 221–227. ACM Press. Roberts, J., Hann, I.-H., & Slaughter, S. (2006). Communication Networks in an Open Source Software Project. In E. Damiani, B. Fitzgerald, W. Scacchi & G. Succi (Eds.), IFIP International Federation for Information Processing (Vol. 203, pp. 297-306). Boston: Springer. Rosson, M.B. & Carroll, J.M. (2006). Developmental learning communities. Journal of Community Informatics, 2(2). Available online at: http://www.ci-journal.net/ viewarticle.php? id=208&layout=html. Rosson, M.B., Ioujanina, A., Paone, T., Sheasley, G., Sinha, H., Ganoe, C., Carroll, J.M. & Mahar, J. (2009). A scaffolded introduction to dynamic website development for female high school students. Proceedings SIGCSE 2009. Chattanooga, TN, March 2009. Rosson, M.B., Sinha, H., Zhao, D., Carroll, J.M., Ganoe, G. & Mahar, J. (2008). Cultivating a landscape of online places for a developmental learning community. Proceedings of
72
John Carroll
International Conference on Advanced Learning Technologies – ICALT 2008 (pp. 311315). Washington, DC: IEEE. Schafer, W.A., Ganoe, C.H. & Carroll, J.M. 2007. Supporting Community Emergency Management through a Geocollaboration Software Infrastructure. Journal of ComputerSupported Cooperative Work, 16, 501-537. Schafer, W., Carroll, J.M., Haynes, S. & Abrams, S. 2008 in press. Emergency management planning as collaborative community work. Journal of Homeland Security and Emergency Management, 5(1), Article 10, 1-17. Schmidt, K. 2002. The Problem with ‘Awareness’: Introductory Remarks. Journal of Computer-Supported Cooperative Work, 11, 285-298. Schön, D.A. 1983. The reflective practitioner. New York: Basic Books Spencer, R., Logan, P., & Coiera, E. (2002). Supporting communication in the emergency department. Te’eni, D., and Speltz, N. (1993). Management information systems in cultural institutions. In D. Young, R. Hollister, V. Hodgkinson, & Associates (Eds.), Governing, leading, and managing nonprofit organizations. San Francisco: Jossey-Bass. Tomasello, M. (1999) The cultural origins of human cognition. Harvard University Press. Trevino, L., R. Lengel, R. Daft. 1987. Media symbolism, media rich-ness, and media choice in organizations: A symbolic interac-tionist perspective. Comm. Res. 14(5) 553–574 Trigg, R.H., and Bødker, S. (1994). From Implementation to Design: Tailoring and the Emergence of Systematization in CSCW. In Proceedings of the ACM Conference on Computer Supported Cooperative Work (CSCW ‘94) Chapel Hill, NC, October 22-26, 1994). ACM Press, New York, NY, 1994, 45-54. Webb, N., & Palincsar, A. S. (1996). Group processes in the classroom. In D. C. Berliner & R. C. Calfee (Eds.), Handbook of Educational Psychology (pp. 841-873). New York: Simon & Schuster Macmillan. West, G. P. (2007). Collective cognition: When entrepreneurial teams, not individuals, make decisions. Entrepreneurship Theory and Practice, 31(1), 77-102. White, B., & Collins, A. (2000). Models of core capabilities. Working Paper. ThinkerTools Research Group, UC Berkeley. Vygotsky, L. S. (1978). Mind in society. Cambridge, MA: Harvard University Press Yamauchi, Y., Yokozawa, M., Shinohara, T., & Ishida, T. (2000). Collaboration with Lean Media: How Open-Source Software Succeeds. Paper presented at the CSCW'00, Philadelphia, PA USA.
In: Outsourcing, Teamwork and Business Management ISBN: 978-1-60456-956-8 Editor: Karl E. Carettas, pp. 73-82 © 2009 Nova Science Publishers, Inc.
Chapter 5
TEAMWORK IN TODAY’S WORLD Carol Boswell1 and Sharon Cannon2 Center of Excellence in Evidence-Based Practice, Texas Tech University Health Sciences Center, 800 West 4th, Odessa, Texas 79763 USA
Abstract Teamwork in all venues today requires management and staff to utilize a variety of approaches to ensure safety in the workplace. The Institute of Medicine’s report is just one example of an organization identifying the importance of effective collaboration by all employees to ensure the wellbeing of engaged participants. Communication is crucial in the response to the call for improved partnership within the workplace. Communication tools such as SBAR (situation, background, assessment, and recommendation), LDS (Let’s Do Something) leadership style, and Huddles can be readily utilized to facilitate effective and comprehensive delivery of key information essential for teamwork regardless of settings. This chapter will discuss these communication tools through the use of safety issue examples such as safe medication administration. Aspects of teamwork will be defined, delineated, and applied. The application of evidence-based practice guidelines would serve as the foundation for the discussion related to effective and successful implementation of a sound alliance within any workplace setting.
Introduction Teamwork in today’s work is applicable to any field, regardless of the profession, place, employment, or position. While many would not think that manufacturing and nursing have any connection, they both require teamwork to provide products and services. There is an old cliché’ that “no man is an island”. That statement holds true today with the demands for highly skilled employees in the workforce resulting from increasing technology, the emphasis on safety in the workplace, and safety in products and services. In 2008, Dr. Beth Mancini (personal communication) used a quote from Henry Ford – “Coming together is a beginning. 1 2
E-mail address:
[email protected]. (432) 335-5150 work. (432) 335-5169 fax E-mail address:
[email protected]. (432) 335-5150 work. (432) 335-5169 fax.
74
Carol Boswell and Sharon Cannon
Keeping together is progress, Working together is success”. Effective collaboration as a team is the guiding force for various approaches to ensure quality of products and services in a safe work environment. Lencioni (2002) describes the impact of teamwork as “…the ultimate competitive advantage, both because it is so powerful and so rare” (p. vii). He further indicates that teamwork is elusive because team members are not perfect human beings. Basic to teamwork is the need to collaborate and communicate among team members. This appears to be a rather simplistic approach but a closer look at collaboration, communication, the required aspects of teamwork and guidelines to use the evidence in practice is needed.
Use of Collaboration to Ensure Wellbeing Collaboration reflects the integration of individuals’ and organizations’ shared aspirations as efforts are made to move the mission of the project, activity, or service forward in a positive, unified manner. Collaboration entails the use of intellectual efforts as the group comes together to address a common challenge or project. The work resulting from the group activity exceeds the efforts of any one member or single member endeavor. The team develops synergistic outcomes as the combined strengths of the group outclass any negativity that may restrict the success of the venture. When a diverse group is convened, the different backgrounds and cultures provide a unique opportunity to embrace original and creative solutions to any challenge under discussion. The group, as a whole, embraces the different strengths brought to the table by the members while offsetting the weaknesses that may be encountered. Critchley, Edwards, and Fallon (2007) state that the ability to effect ongoing change within an organization results from the dedication to insuring that the hearts and minds of the membership of the group are engaged with the determined mission. By exciting the members toward a common objective, the relationship and engagement by the members supports and advances the overall work of the team. Collaboration by a multi-disciplinary team fosters the significance of solidarity. Within the study by Kovner, Brewer, Fairchild, Poornima, Kim, & Djukic (2007), newly licensed RNs identified the work-group relationship to be positive and cohesive, but the relationships with other health care providers as more neutral in nature. As the different partnerships within a multi-disciplinary team are developed, each group of individuals should come together with support and encouragement. The tribal/confrontational nature which is sometimes evident between disciplines can result in a blaming culture rather than a supportive one. As a means for getting past the blaming game which can frequently happens, each member of the team needs to accept personal accountability for their own actions. Too frequently, time and energy is restricted to placing the blame and not looking at the individual’s responsibility for the challenge under consideration. Miller (2004) states “personal accountability is about each of us holding ourselves accountable for our own thinking and behaviors and the results they produce” (p. 64). Individuals can not change other people, they can only frustrate and anger them. As the team accepts the challenge to change, the realization that each individual accepting responsibility for key aspects allows the development and synergy of the collaborative effort to be successful. Miller (2006) lists five essential ideas or principles useful when embracing personal accountability. The five concepts are: •
learning the skills/activities needed to address the issue,
Teamwork in Today’s World • • • •
75
accepting ownership for any individual pieces of the puzzle , embracing creativity for identifying innovative paths to solutions, serving others in any and all settings , and building trust in the relationships encountered
By endeavoring to apply the Golden Rule – “Do for others, what you would like done for you” -- a practical and rewarding means of advancing the goals and/or mission for the project or organization can result. Making sure that the majority of the outcomes are a “win-win” endeavor not a “win-lose” situation allows for positive outcomes at all levels and within any enterprise.
Communication According to Critchley, Edwards, and Fallon (2007), “clear communication helps each member of a staff understand why their organization must meet the targets and operational standards that have been set for it, it is important to encourage commitment rather than compliance” (p. 10). The ability to provide translucent communication to each member of a team is paramount in the workplace. From observing workplace conflicts, the primary challenge, which seems to be the foundation of most conflicts, is the lack of unambiguous communication of the facts and expectations. The dedication of time and energy initially within any organization or team to ensure that the targets and operational standards are effective communicated to everyone involved saves frustration and energy as the project, service, or task is planned and implemented. By taking this time and responsibility, the overall teamwork and collaboration for the endeavor is placed into a positive enterprise. A model to improve teamwork requires an examination of a leadership model. In 2006, Cannon and Boswell developed a simple, flexible leadership model for use by leaders in evidence-based practice (Cohn, Cannon, & Boswell, 2006). The “Let’s Do Something” (LDS) model was applied to nursing but is also applicable for any organization. The model isn’t linear, requires risk taking, doesn’t have to have a completely implemented plan, and the leader isn’t named or elected but rather takes responsibility for action. The first component of the model is vision. The team must have a common vision/mission/goal. A need is recognized and agreement leads to change. The second component of the model is leadership. In this instance, the leader is not appointed or elected but rather someone who is willing to take a risk and does so. Leadership in this model allows a member to organize one project but be a member for a different project. This allows for individual expertise in multiple areas and doesn’t weigh down a leader in a “be all, end all” situation. The final component of the LDS model is networking. Networking, as a form of collaboration, allows individual expert skills and abilities to contribute to the success of the project. Working together as a team is exciting and satisfying. Even at times when the project is stymied, unsuccessful members share a common bond. An important aspect of the LDS model is the provision of growth. Strong or different personalities bring unique perspectives. The model accommodates both. Success and even failure can generate other visions and/or possibilities. The team generates energy to sustain required actions. A second communication venue can be seen in a group meeting used on nursing units to facilitate the communication between team members called “Huddles”. The concept behind
76
Carol Boswell and Sharon Cannon
the Huddles process is to provide a time for the work group to come together for a quick, ontarget discussion about the key challenges confronting the group within the immediate work session. As a result, this communication methodology could be utilized in any setting or situation where conflicts, concerns, and challenges present within the work environment. By bringing the work group together for the quick discussion of potential threats to the work environment, solutions and/or management plans for the situation can be considered and agreed upon by the group. Each member of the team can then participate in the successful management of the working experience. Guidelines for the material to be potentially discussed during the Huddles meeting must be agreed upon by the participants prior to the session. What type of information, by who, how often, are all issues which should be addressed within the guidelines. The guidelines would be unique for the different sites where the Huddles session is held. The information provided within a nursing unit would be quite different from the information presented by a manufacturing unit. The information communicated via the Huddles process does not address the overview of the work day but should be used to convey information needed to facilitate the work environment. Within a nursing setting, the information could include which patients are at risk for falls, which families may need additional support due to bad health news, upcoming mandatory educational sessions, potential committee meetings, new opportunities for involvement with the management of the unit, or which staff members are overwhelmed due to critical situations with client assignments. The timing of the meeting is based on the needs of the particular unit while restricting the length of the meeting to approximately 15 to 20 minutes total. Short, sweet, and to the point are the ideas for getting key information to the individuals needing the information to improve the work environment and quality of the work setting. Another form of communication skills revolves around the initials SBAR within health care. The concepts of SBAR relate to the provision of effective communication between individuals. The initials stand for: S = Situation, B = Background, A = Assessment and R = Recommendation. By organizing any communication of a situation around these four aspects, the message delivered is apt to be complete and functional. This organization of communications can be effectively instituted in any communication setting where a challenge or critical situation is perceived. The originator of the communication related to the problem or challenge is expected to orchestrate the information provided around the four aspects. The circumstances of the problem/challenge are supported by the provision of the background information to provide context for the process. After providing these two aspects, the originator of the communication provides an assessment of the current situation then ends with recommendations for the possible management of the challenge/problem. The receiver of the communication can then take the information provided in this organized manner and work with the originator to solve the problem and/or challenge. This communication process requires that the originator of the communication comes to the communication situation with a rational and structured delivery of information. So often the problem or challenge is communicated only to the point of the situation and maybe the background. Within this communication process, the instigator has to round out the process by including an assessment and recommendations based on their interpretation of the situation. Transforming Care at the Bedside (TCAB) is a response to the need for healthcare quality (Robert Wood Johnson Foundation and the Institute for Healthcare Improvement, 2006). The goal is to provide nurses and other healthcare providers’ guidance at the bedside to effect better clinical results. This process in turn, helps to reduce staff turnover as staff are directly
Teamwork in Today’s World
77
involved in the decision-making process. The intent is to improve the quality of care and improve staff satisfaction. This intent is of primary importance for patient safety in a time of healthcare worker shortages and the high cost of healthcare. TCAB puts the point of service responsibility where the care is being delivered. This approach involves the staff in designing the work to reduce waste and increase patient satisfaction. The staff generates the ideas and focuses on working as a team. Staff is encouraged to create change and share ideas to improve patient safety and satisfaction. This team concept is the glue that bonds individuals to accomplish organizational goals. It is a bottom up approach as opposed to the traditional top to bottom directions. There are four steps to TCAB: 1.) assemble a front-line team to generate new ideas, 2.) test ideas and measure results, 3.) implement and spread, and 4.) collaboration to share learning. The idea behind TCAB and its four steps can be transferred to other businesses. Few would argue that frontline staff can make or break a business. A rude or indifferent receptionist sets a negative environment for customers and often leads to customer dissatisfaction. Putting on emphasis on teamwork and allowing employees at the point of service to generate, implement, and evaluate ideas to improve quality of services/products can be exciting, cost effective, and prevent staff turnover. The TCAB approach, regardless of setting, business, or organizations, is an excellent example of how “working smarter, not harder” can be accomplished.
Aspects of Teamwork Change is inevitable and an integral part of teamwork. According to Reineck (2007), “organizational change often occurs in response to natural forces, and ultimately, change is shaped by people” (p. 388). Teamwork is that process of taking the strengths and weaknesses of the participants, intertwining them, and arriving at agreement as to the direction the working group needs to take to meet and address the identified challenge or situation. The natural forces come to bear by distinguishing the situation and/or challenge requiring attention from the group. Once the circumstances are recognized, the team progresses to manipulate and mold the process based on the values and beliefs maintained by the group. MacPhee (2007) stated “effective teamwork depends on leadership clarity, role clarity, shared goals, and frequent communication” (p. 407). Each of these pieces must be considered as a team initiates the work to be done. As each of these aspects is managed, the resulting critical outcome for the team is the development of trust. Trust by each member within the team allows for the open, effective management of key aspects necessary to positively impact the challenges and projects engaged. As an individual considers and struggles with team work, the philosophy presented by Lundin, Christensen, and Paul (2003) concerning the principles of Fish! can be used to improve and champion the effort. The Fish! philosophy of motivation and leadership was initiated and developed at a fish market in Seattle, Washington. Three principles central to the effective use of teamwork are: Find IT, Live IT, and Coach IT. The IT relates to knowing the vision and/or goal for the group or team. Only when the vision (IT) is firmly established can a path for acquiring it be sought and found. Once the vision is securely identified, each member of the team must commit to engaging in the opportunities which drive the vision toward completion. If the opportunities related to the vision are ignored, the overall success of the
78
Carol Boswell and Sharon Cannon
team will be negatively impacted. Once the team has identified the vision and discovered occasions to put it into practice, then the team needs to find venues for engaging other members and peers within the vision. According to Lundin, Christensen, and Paul (2003), change is the easy part -- sustaining the change is the aspect which requires time, energy, and commitment. Maintaining the change brought about by the team necessitates that the members of the team are willing to pledge dedication to the vision. Within the development of a team, each member must take responsibility for their own thoughts, opinions, and values. Values clarification within the group can be a positive formula for moving the team toward a common goal or mission. Each member within the team must identify and support the values central to the organization. When members with the team are each directed toward their own values or outcomes, the success of the team can be impacted. Buy-in by the members within the team to the mutual goals of the group is paramount for success. For most challenges/problems addressed by any team, the outcomes are seen as “win/win”, “lose/win”, or “win/lose”. Each team should strive to make the resulting situation a win/win result. Having acknowledged this, it is apparent in some situations when values clarification is not apparent that the members within the group are focused on ensuring success from their viewpoint. When individual needs and expectations serve as the motivation, the outcome does not always provide a winning solution for all of the membership. By laying a foundation that is reflective of the needs of all the participants within the situation, the outcome/solution results in a successful product for everyone. Within the delineation of the values, a clear definition of the problem is an overriding principle for the successful management of the situation. Without an understandable acknowledgement of the dilemma, the focus for the team can result in a convoluted path toward a solution. Another key aspect within teamwork as a win/win solution is attempted revolves around the inclusion of each member within the affected group. Consideration of all viewpoints as solutions and recommendations are identified is imperative. The resolutions can then be based upon the maximization of the benefits recognized while minimizing the obstacles. Objectivity and openness by all participants within the team culminates in a win/win outcome for the project encountered.
Evidence-Based Practice Guidelines While evidence-based practice tends to be directed primarily toward clinically identified care settings, the concepts embraced by it can be applied to any setting or location. According to Boswell and Cannon (2007), evidence-based practice is “viewed to be a research-based, decision-making process utilized to guide the delivery of holistic patient care by nurses” (p. 10). Evidence-based practice utilizes the research results as the foundation for the care, services, and projects incorporated into the provision of health care or any other employment setting. Evidence-based practice allows individuals the opportunity to based their recommendations on tested information. The recommendations then move from a “my opinion is better than yours” to the experts support this recommendation. Justice and Fey (2004) state that evidence-based practice highlights the methodical, purposeful combination of science, craft, data and theory. They go on to state that evidence-based practice is a “shift from a craft-based and intuition-driven practice to one in which consideration of the preponderance of evidence from scientific investigations is systematically integrated with
Teamwork in Today’s World
79
clinical expertise and contextual factors” (p. 4-5). Each and every consumer driven agency benefits from the inclusion of tested venues of product delivery. The use of validated outcomes within the provision of any service provides support and foundation for the resulting service and/or project. By using tested processes within the delivery of services, the outcome is supportive and positive. One of the key aspects within the use of evidence-based practice is the incorporation of the leveling of the evidence. Several different agencies provide versions of levels of evidence. Overall, the evidence levels reflect moving from strong experimental design studies to unstructured qualitative types of studies (see Table 1). While it is hoped that recommendations are based on the strongest level of evidence, each level of evidence provides support for the recommendation. As any organization/profession strives to base the services, projects, and/or products on best practices, the research unique to that topic has to be carefully considered and analyzed. Only research that is defendable should be incorporated into daily practice. DiCenso, Guyatt, and Ciliska (2005) provide ideas which can be used to gauge the effectiveness of the outcomes being considered (see Table 2). Any solution to a problem/challenge should be weighed again the evidence that is available concerning that situation. By contemplating what is known about a given state of affairs, informed decisions can be made. The decisions then come with support and underpinning to provide a clearer understanding of the potential benefits and weaknesses which may be involved in the situation. Evidence-based practices force an agency, company, team, or individual to carefully consider all viewpoints related to the situation under investigation or consideration. By including all perspectives within the discussion, a comprehensive resolution can be considered. Table 1. Levels of Evidence for Evidence-based Practice Systematic reviews of randomized trials Single randomized trial Systematic review of observational studies addressing patient-important outcomes Single observational study addressing patient-important outcomes Physiologic studies (e.g., studies of blood pressure, cardiac output, exercise capacity, bone density) Unsystematic clinical observations DiCenso, A., Guyatt, G., Ciliska, D. (2005). Evidence-based nursing: A guide to clinical practice. St. Louis, MO: Elsevier Mosby.
Larue (2007) lists five steps used in the performance of evidence-based practice. The five steps are: 1. 2. 3. 4. 5.
Ask an answerable question Find the best evidence Critically appraise the evidence Act on the evidence, and Evaluate the performance (Larue, 2007, p. 14).
80
Carol Boswell and Sharon Cannon Table 2. Ideas to Ponder When Considering Options
Are the conclusions defensible? Was randomization used within the process, if so how? Were clusters comparable or were the examinations adjusted to offset variations? Was the group randomization communicated effectively? Was follow-up achieved?
What are the outcomes? What was the intervention outcome? How accurate was the assessment of the intervention consequence?
How can the outcomes be applied to the work environment? Were the study subjects similar to the work environment setting? Were key conclusions contemplated? Are the identified interventions profitable for the work environment based on the possible results and expenses? Adapted from DiCenso, A., Guyatt, G., Ciliska, D. (2005). Evidence-based nursing: A guide to clinical practice. St. Louis, MO: Elsevier Mosby.
As is evident from these five steps, the process is directed toward ensuring that the correct problem/challenge is addressed. Any agency can incorporate these pieces into the management of analysis and trials confronting the organization. So often the problem with management of problems is the lack of asking the right question. Melnyk and FineoutOverholt (2005) call this process of identifying the question as formulating the burning question. When any challenge is identified within an organization, an answerable question must be identified. If the question is not answerable, time and energy should not be wasted on that situation. Sometimes, the process requires the rewording of the question into a focus that does allow an outcome or answer. The group/team must carefully consider what is really the foundational question or concern. So often the group/team centers the process on peripheral questions or concerns instead of identifying the crucial question which tends to be the driving force within the situation. If the question is not answerable, then any solutions arrived at will not address the real problem under consideration. Once the crucial question is considered, peripheral questions can be prioritized based upon the potential consequences. To guide the development or clarification of an answerable question, the use of the PICOT format helps. Each letter within the format allows for the clarification of the question. The “P” stands for the population. Elucidation of the populace which is affected by the problem is important. The “I” represents the intervention or problem to be considered. Again, this part of the question is key to moving the process toward success and must be clearly delineated. The “C” reflects any comparison to the intervention or problem. Not all problems or interventions will have a comparison aspect. The “O” signifies the outcome. The outcome is another key aspect within the answerable question. The team must determine the outcome expected to be able to
Teamwork in Today’s World
81
state that the problem was solved. The “T” depicts any time limit within the problem. Again, time may not be a integral part of all questions. As can be seen, an answerable question within evidence-based practice must have the population, intervention/problem, and outcome identified each time but can be strengthen when any comparison and/or time limit can be included. The overall benefit for using evidence-based practice within the workplace environment is twofold. Through the use of evidence-based practice, professionals can effectively consider treatments, products, and services endorsed by individuals based upon the preponderance of evidence rather than on just “best guess”. Each treatment, product, and service has to run the gauntlet to prove the effectiveness of the process for utilization within the work environment. Aspects of care and/or products are then validated as to the functionality of it for a given setting or endeavor. Accountability for the effectiveness of the treatment, product, or service becomes an integral aspect within the utilization of it within the work place. According to Justice and Fey (2004), “EBP requires a synergy between the research community, which is charged with accumulating evidence, and the clinical community, which is charged with examining the preponderance of evidence, to make decisions about the best ways to evaluate and treat children” (p. 30-32). This same synergy is evident from any utilization of EBP. The best of the research information is supported and validated by the application of the service, treatment, or product within the real world of industry. By accepting the accountability for corroborating services, treatments, and products with accumulated evidence, a win/win situation results as the best solution can be determined for the challenge confronted.
Conclusion According to Erenstein and McCaffrey (2007), “it is obvious then that there is a need for dialogue between healthcare administrators and nurses concerning the needs of each to create an environment where nurses can act as professional care providers, which will benefit both the healthcare institution and patients” (p. 306). While this is true for healthcare, it can easily be applied to any employment setting. Administrators and staff members must come together to create an environment that is supportive and manageable for all of the employees and clients. As the employees are comfortable with the work environment, the functionality and successfulness of the organization will be positively impacted. Everyone will benefit from the administrators to the employees to the consumers. Whether using leadership, communication, or evidence-based models, the key to success is shared responsibility, accountability, and teamwork.
References Boswell, C. & Cannon, S. (2007). Introduction to nursing research: Incorporating evidencebased practice. Sudbury, MA: Jones & Bartlett Publishers. Cohn, K.H., Cannon, S., & Boswell, C. (2006). Let’s Do Something: A Cutting-edge Collaboration Strategy. In K.H. Cohn (Ed.) Collaborate for Success! Breakthrough strategies for engaging physicians, nurses, and hospital executives. (pp 73-86). Chicago: Health Administration Press.
82
Carol Boswell and Sharon Cannon
Critchley, D. Edwards, C., & Fallon, R. (2007). The importance of good teamwork. Nursing Management, (14)7, 8-12. DiCenso, A., Guyatt, G., Ciliska, D. (2005). Evidence-based nursing: A guide to clinical practice. St. Louis, MO: Elsevier Mosby. Erenstein, C. & McCaffrey, R. (200, November/December). How healthcare work environments influence nurse retention. Holistic Nursing Practice. 303-307. Justice, L.M., & Fey, M.E. (2004, September 21). Evidence-based practice in schools: Integrating craft and theory with science and data. The ASHA leader, pp. 4-5, 30-32. Kovner, C.T., Brewer, C.S., Fairchild, S., Poornima, S. Kim, H., & Djukic, M. (2007). Newly licensed RNs’ characteristics, work attitudes, and intentions to work. AJN, 107(9), 58-70. Larue, E. (2007, Fall). Evidence-based practice requires much more than identifying the best evidence. Pitt Nurse. Pittsburgh, PA; University of Pittsburgh School of Nursing Magazine. Lencioni, P. ( 2002). The five dysfunctions of a team: A leadership fable. San Francisco, CA: Jossey-Bass. Lundin, S.C., Christensen, J., & Paul, H. (2003). Fish Sticks. New York, NY: Hyperion. MacPhee, M. (2007). Strategies and tools for managing change. Journal of Nursing Administration, 37(9), 405-413. Melnyk, B.M. & Fineout-Overholt, E. (2005). Evidence-based practice in nursing and healthcare. Philadelphia, PA: Lippincott Williams & Wilkins. Miller, J.G. (2006). Flipping the switch. New York, NY: G.P. Pubnam’s Sons. Miller, J.G. (2004). QBQ! The question behind the question. New York, NY: G.P. Pubnam’s Sons. Reineck, C. (2007). Models of change. Journal of Nursing Administration, 37(9), 388-391. Robert Wood Johnson Foundation and the Institute for Healthcare Improvement, (2006). Transforming care at the bedside: A new era in nursing. RWJF/IHI.
In: Outsourcing, Teamwork and Business Management ISBN: 978-1-60456-956-8 Editor: Karl E. Carettas, pp. 83-98 © 2009 Nova Science Publishers, Inc.
Chapter 6
TEAMWORK AND PBL-BASED TEACHER EDUCATION: A STUDY ON PROSPECTIVE SCIENCE TEACHERS’ OPINIONS Laurinda Leite and Esmeralda Esteves University of Minho, Braga, Portugal
Abstract It is fully accepted that successful science learning depends at least in part on teachers’ teaching competences. In addition, there is some evidence that teachers tend to teach as they were taught. Therefore, in order to develop teachers’ innovative and student-centered teaching competences, methods courses should acknowledge teaching methodologies similar to those that prospective teachers will be required to use in their future as teachers. Problem-Based-Learning (PBL) and teamwork are student-centered teaching approaches that may foster the development of relevant competences for students as citizens. In fact, PBL can promote learning how to learn competences while teamwork can foster the development of social and interpersonal skills. However, to successfully put these teaching approaches into practice teachers need to fully acknowledge big role changes. Hence, in order to prepare innovative teachers, PBL, together with teamwork, should be used in initial science teachers’ education programs. This chapter describes how 38 prospective physical sciences teachers evaluate teamwork and PBL carried out within a methods course to approach a module on Using the lab for physical sciences teaching. Data were collected by means of a questionnaire, a self-evaluation grid and a videotaped discussion focusing on the diverse parts of a PBL sequence: problems formulation from a scenario, problem solving and synthesis and evaluation. Results indicate that prospective teachers valued PBL and teamwork, as they felt that the latter helped them to cope with the new roles that they were required to undertake throughout the PBL sequence. However, the facilitating effect of teamwork seems to be insufficient to lead students to fully overcome their difficulties with more unusual tasks. Nevertheless, it seems that they may be prone to use these teaching approaches when they become science teachers.
84
Laurinda Leite and Esmeralda Esteves
Introduction Modern societies are highly dependent on science and technology. Whether it is a bicycle, a plane, the television, the computer, a mobile phone, an i-pod, an ATM, a fridge or a rice-cooker, the truth is that everybody uses technology. As technology is science (inter)dependent, it can be stated that everybody uses science in a more or less conscious way. A dramatic consequence of this is that many people may be using science and technology in non-appropriate ways. Nowadays, ozone layer depletion and global warming are good examples of consequences of non-appropriate ways of using science and technology. By appropriate way of using science and technology it is meant the use of science knowledge and technological devices for the well being of the individual as well as for the development of society without putting at risk the future of both the mankind and the planet. The increasing awareness of politicians about the importance of this issue has led to the development of educational policies aiming at developing citizens’ scientific literacy and consciousness of the interrelationships among science, technology, society and environment. To successfully putt into practice those policies, a deep change into the usual target competences of science courses is required (Ratcliffe & Grace, 2003). If the argument for having every child attending schools for a few years has to do with reaching a level of literacy that enables him/her to behave as an active and informed citizen, then school has to acknowledge procedural competences rather than to solely concentrate on the conceptual ones. The reason for this lies in the fact that science develops so fast that unless school equips students with learning how to learn competences, they will be left behind soon after leaving school. In addition, the increasing specialization of science and technology claims for interdisciplinary rather than for individual work. Hence, high level of social skills development is more and more required so that people can not only interact and communicate with each other but also do it in a way as to positively contribute to the global advanced society. Problem-Based-Learning (PBL) together with teamwork may foster the development of relevant competences for students as citizens, including learning how to learn competences and social and interpersonal skills. However, a successful implementation of these teaching approaches requires big role changes that teachers need to fully acknowledge. As there is some evidence that teachers tend to teach the way they were taught (Gil & Pessoa, 1994; Martínez et al., 2001), methods courses for prospective science teachers should also use PBL together with teamwork to approach science education issues in order to improve the chance of leading teachers to be willing and able to use them. Despite the great educational potential of PBL (Barell, 2007), this methodology has not yet concentrated enough the Portuguese teacher educators’ attention. However, there are a few examples of experiences of putting PBL into practice, together with teamwork, at science teacher education contexts (Edwards & Hammer, 2004; Esteves & Leite, 2005; Peterson & Treagust, 1998). The engagement of prospective teachers in such experiences requires them to undergo big changes in the way they were used to behave as students in school and university, as they need to move from individual passive listeners to active social learners. Thus, the point is how they feel when they become protagonists of such an enterprise.
Teamwork and PBL-Based Teacher Education
85
Therefore, the objective of this chapter is to analyze how prospective science teachers view them as part of a PBL environment based on teamwork.
Theoretical Background Science Teacher Education for Students’ Lifelong Learning How teaching is approached affects the way students learn (Yeung et al., 2003). In addition, pre-service teachers actively construct beliefs about their future role as teachers through the teaching experiences they are submitted to as students (Hancock & Gallard, 2004). However, there is some empirical evidence that learning about teaching is best accomplished by direct experience of the teacher-learner and by opportunities to critically analyze that experience (Hancock & Gallard, 2004). Hence, it can be argued that how teacher education is approached affects the way prospective teachers conceptualize teaching. Also, engaging prospective teachers in experiences similar to those they are expected to put into practice in their future, as teachers, may be a good strategy to foster innovation in teaching. There is no single way of teaching any subject. Also, there is no teacher using a single teaching approach irrespective of the students he/she is teaching and the goals they are expected to attain. However, there are teachers that tend to value some teaching approaches instead of others. Research (Akinoğlu & Tandoğan, 2007; Jenkins, 2006) suggests that teachers tend to value content rather than process and to adopt teacher centered teaching approaches rather than the students centered ones. Teacher centered teaching approaches may be good to “fill in” students’ memories or even to teach them a lot of content but they are of little use if students’ differentiation of what they know and what they don’t know is at stake (Lambros, 2007). In a changing world, memorization is of reduced use. It does not account for even the smallest changes. What modern societies need is citizens that can easily identify their lack of knowledge with confidence, that can access new information and integrate it with the existing knowledge and that use new information to solve problems (Barell, 2007; Lambros, 2007). Therefore, instead of teaching for memorizing, teachers need to teach for coping with changes. This means that not only teachers need to adequate their teaching but also that students need to change their learning strategies as well as their conceptualizations of learning and their roles as learners. Besides, society is changing and students’ interests are changing too. What motivates students today may probably become of no interest to them in the near future. A consequence of this is that teachers need to develop lifelong learning skills too, so that they can maintain updated their science knowledge and their teaching methods repertoires. This means that they need to learn how to learn about science education resources and methods in order to maintain their pedagogical content knowledge updated at any time. A way of helping them to attain this goal is to teach them about such issues through PBL. In adopting such an approach, teacher educators are giving their student teachers the opportunity to acquire specific content knowledge but also to develop self-direct learning and lifelong learning skills, to think critically, to analyze and to solve problems, to enhance communication skills and information management skills, to work cooperatively (which includes negotiation, mediation and appreciation) and to develop attitudes and behaviors (Lambros, 2007). However, it is worth mentioning that if PBL is to lead to effective learning, teacher educators (like any other
86
Laurinda Leite and Esmeralda Esteves
teacher) need to be equipped with process skills competences (such as handling group dynamics, questioning skills, facilitating meta-cognition etc.) and to be able to identify, articulate and assess those skills (Tan, 2004). As there is some empirical evidence that PBL can assist students in learning concepts and also in working with others (Peterson & Treagust, 1998; Koh, 2008), it can be argued that prospective teachers should “be taught” through PBL (Sage, 2001) at least in science methods courses, so that they can learn important concepts, get used to work with others and perceive the relevance and the viability of the teaching approach, namely with regard to the development of lifelong learning skills. Peterson & Treagust (1998) used a PBL approach, including small group discussion and a peer teaching activity, to develop pre-service teachers’ knowledge base for teaching and pedagogical reasoning ability. Those authors noticed that randomly selected pre-service teachers developed their knowledge base and pedagogical reasoning. In addition, the authors believe that together these two domains of competences are relevant for problem solving in primary teaching. Dalghren et al. (1998) analyzed the integration of environmental education in teacher training through PBL. Data collected from seven teachers reveal different ways of conceiving the role of the teacher. It was found to range from supportive to directive, being the former more compatible with PBL than the latter. Edwards & Hammer (2004) conducted a research study with 54 Australian pre-service teachers that required them to learn the unit Child Development trough PBL. The results of the study indicate that a few students found it hard to understand what they were asked to do and felt several difficulties when working with other people. However, the majority of the participants found learning realistic and empowering both as a student and a future teacher. According to Dalghren et al. (1998), “the implementation of PBL is more than a change of methods; it requires reflections on the part of the teachers on their conceptions of knowledge and learning” (p. 446). The acknowledgement of such change may be at the basis of the positive results of PBL in physician competence after medical school education (Koh et al., 2008) and can also apply to other professionals, namely to the teachers.
Problem Based Learning into Practice Solving problems is a very common activity in science classes. Nevertheless, problems are usually associated with calculations and used to check whether, or not, students learnt what they were taught. However, according to Barell (2007), “a problem is any doubt, difficulty or uncertainty that invites or needs some kind of resolution” (p. 3). Thus, problems can play three different roles in the teaching and learning processes (Dumas-Carré & Goffard, 1997; Watts, 1991), as they can be used: -
To evaluate students’ learning, being solved after teaching and learning; To deepen students’ learning, being solved during the teaching and learning processes; As a starting point for teaching and learning, being given to the students or formulated by themselves at the very beginning of the teaching and learning processes.
Teamwork and PBL-Based Teacher Education
87
When problems play the latter role it is usual to say that Problem-Based Learning (PBL) is taking place. However, this newer role of problems has been conceptualized in several different ways (Savin-Baden & Major, 2004), ranging from teacher’s centered use of problems to teach (e.g., lecture-based cases and case-based lectures) to students’ centered use of problems to learn. The latter way of using problems is the one that deserves most consensuses among science educators as having to do with PBL (Chin & Chia, 2004; Evensen & Hmelo, 2000; Hmelo-Silver, 2004; Lambros, 2002; Savin-Baden & Major, 2004, Gandra, 2001; Leite & Afonso, 2001; Leite & Esteves, 2005) and it is also the meaning acknowledged in this chapter. Hence, PBL assumes the idea that by engaging in an inquiry process (Barell, 2007) students can “learn by doing”. Therefore, PBL acknowledges the idea that students can play an active role in learning (Barron et al., 1998). Transferring Barrow’s (1986, quoted by Yeung et al., 2003) primary educational objectives in the clinical context to the science education context, it can be argued that PBL can promote the development of effective reasoning, the organization of a knowledge base for use in science education contexts, the development of effective self-directed learning skills and an increased motivation for learning science. Thus, PBL is not only a way of learning but also a way of learning about learning and a way of promoting lifelong learning (Hmelo-Silver, 2004) as well as a way of developing self-directed learning skills (Yeung et al., 2003; Tan, 2004). In addition, if PBL is organized in such a way as to start with everyday problems, which are multidisciplinary by nature, than it can both facilitate school knowledge integration and contribute to better equip citizens to face socio-scientific problems in their daily lives (Nagel, 1996; Barron et al., 1998; Savin-Baden, 2000; Yeung et al., 2003). An analysis of relevant literature (e.g., Albanese & Mitchell, 1993; Barell, 2007; Barrows & Tamblyn, 1980; Gandra, 2001; Lambros, 2002; Leite & Afonso, 2001; Davis & Harden, 1999) leads to the conclusion that a PBL sequence includes four main phases. During the first phase, the teacher selects a scenario or problem context believed to lead students to raise problems that are both motivating to them and relevant from the curriculum point of view. The context should deal with a problem situation as real as possible and it can be introduced to the students through a videotape, an audiotape, a text, etc (Davis & Harden, 1999; Lambros, 2002; Lambros, 2004). Problems to be solved are crucial as “in PBL, the content of a problem serves as a major vehicle that influences each student’s direction and motivation to learning” (Yeung et al., 2003, p. 341). In the second phase, students are faced with the scenario or problem context and, in small groups, they are asked to formulate questions focusing on: “What issues/concepts am I already familiar with?”, “What are the issues/concepts I do not known anything about?/ What are the issues/concepts that I do not understand? What have I never heard about?” and “What would I like to know about this?/What would I like to develop further?”. Afterwards, students are asked to discuss the questions formulated, to analyze their relevance and interdependency and to place the selected problems to be solved in a (provisional) logic sequence. The third phase is devoted to finding out solutions for the problems. In order to devise a strategy, students have to answer questions like “what do I already know about this problem?” What do I need to do to efficiently solve this problem?” and “What sources of information should I use to find out an appropriate solution?” (Duch, Groh & Allen, 2001; Lambros, 2004). Afterwards, they have to put into practice the strategy, using whatever resources they feel that can be useful to carry the task successfully, that means to reach solutions for a problem or solutions for a set of problems that were worked together.
88
Laurinda Leite and Esmeralda Esteves
Finally, in phase four takes place when students presented the solutions to the whole class and the teacher promotes a global analysis of the process (fourth phase). Its aim is synthesizing learning and analyzing the contribution of the leaning process followed to both science education and education for citizenship. Questions like “What did I learn?”, “What was not fully understood?” and “What sorts of strategies and behaviors worked better?”. As the emphasis is on learning rather than on teaching, the main role of the teacher is to create conditions for students to learn and to help them to become aware of the successes and failures of the learning process.
Teamwork and the Development of Interpersonal Skills in a PBL Context In PBL as well as in daily life, problems offer a challenge to students as they appear before people have a solution to them and even before people become aware of what is needed in order to reach it. Hence, to solve a problem both students and citizens would have to make a journey from the known to the unknown, relying on their previous knowledge but developing it in order to successfully overcome the challenge offered by the problem. Thus, during this process they will develop specific knowledge, as they will acquire a deeper comprehension of the scientific concepts and principles that are associated with the problem to be solved. In addition, they will enlarge their repertoire of general competences, as some adjustment of procedural knowledge, reasoning and communication skills is expected to occur during the problem-solving process (Lambros, 2002; Lambros, 2004). Although students may not initially be fully aware of that, the probability of successfully overcome the challenge underlying the problem is larger if students work cooperatively instead of working in an individual basis on the problem-solving task. Cooperative learning is mostly influenced by Vygotsky’s (1986) perspective on the benefits of learning in social settings. It seeks to lead students to create meaning in school experiences by offering them “opportunities to collaborate, converse, and reflect upon important information and often does so by having students become teams of investigators or analysts” (Vermette & Foote, 2001, p.33). However, putting students together does not give any guaranty of cooperation and successful learning (Gillies, 2003). In addition, in the every day world, the word cooperation encompasses meanings that are not appropriate for a small group work context (Holloway, 1992), as they not only stress the limitations rather than the potentialities of cooperation but also are often associated with inequality and coercion. Besides, “the experience of being in groups can be powerfully emotionally charged both positively and negatively” (Cartney & Rouse, 2006). This is the reason why some authors argue for the need of not only teaching students how to work in groups before using small group work as a strategy for promoting content learning (Peeters, 2005) but also using dimension, ability and gender criteria for organizing groups (Gillies, 2003). In fact, communication within small groups may be facilitated if the nature of the group provides conditions for a set of social skills to be put into action. Based on a review of literature, Gillies (2003) listed the following skills: listening to each other during group discussions; acknowledging others’ ideas and considering their perspectives on issues; stating ideas freely; resolving conflicts democratically; sharing tasks equitably; allocating resources fairly among group members. An analysis of the nature of such skills indicates that the teacher has an important role to play in order to promote the social integration of the group members
Teamwork and PBL-Based Teacher Education
89
(Cartney & Rouse, 2006) and to help each one of them to understand that no one will succeed unless the group as a whole succeeds (Gillies, 2003). Some studies indicate that students involved in cooperative learning instruction environments had significantly higher results than those taught by traditional approaches (Acar & Tarhan, 2006; Fernández-Santander, 2008; Gillies, 2003) and showed more motivation to learn than if they worked alone (Gillies, 2003). In addition, a recent research study (Grindstaff & Richmon, 2008) suggests that the social dimension can offer positive support to the emotional, the cognitive and the technical dimensions of learning. The benefits of cooperative learning are the main reason why some authors (Lambros, 2004; Woods, 2000) argue for carrying out PBL in small groups. Besides, if PBL is carried out in small groups, alongside with cooperation within the group, there can be competition among the diverse groups. As Cartney and Rouse (2006) put it, “a group can be seen as a complex interweaving of internal and external worlds, individual and group defensive mechanisms, shaped by intense emotions” (p.86). In fact, as solving the problem becomes the group’s goal, every member of the small group engages into the tasks in an attempt not only to succeed but also to override the success of the other groups. To achieve such aim, the group members need to play and share diverse roles (Hogan, 1999) and to acknowledge the ways of working as well as the learning strategies of theirs counterparts (Barron, 2000). Moreover, they need to see the group as a unit in itself, “with the whole being more than the sum of the parts” (Cartney & Rouse, 2006, p.86). Being a complex task, problem solving requires students to share their thinking in order to create a common knowledge product (Hogan, 1999). Hence, solving problems in small groups provides students with experience of in-depth, higher-order collaborative thinking and offers them opportunities to develop cognitive as well as social and interpersonal skills. In fact, it leads students to discuss prior knowledge, to look for new data and information, to integrate knowledge, to cooperate with each other, to mediate conflicts, to negotiate meanings and approaches, to divide tasks, to develop arguments and critical reasoning, to discuss issues and to make judgments (Lambros, 2007). However, research (Gillies, 2003) has shown that to attain these aims, small group work needs to be structured and students as well as teachers need to be trained to use it. It is worth noticing that success on the small group task depends on a set of factors and different task contexts give rise to different group and individual behaviors (Hogan, 1999). This is the reason why the author argues for the situated nature of collaboratively constructed learning. Therefore, evaluation cannot concentrate only on cognitive learning but rather it should monitor the diverse activities within the scope of all the phases of the PBL cycle and deal with the contribution of each member to the group (Duch & Groh, 2001), both as an individual and as a member of the group. In doing so, evaluation fosters a metacognitive attitude and develops students’ social responsibility.
Methodology Sample A group of 38 Prospective Physical Sciences Teachers (PTs) attending the fourth year of a fifth yearlong undergraduate teacher education program participated in the study. They
90
Laurinda Leite and Esmeralda Esteves
represent 100% of the PTs that were taking a course on Methodologies of Teaching Physics and Chemistry that was taught by the authors. The majority of the PTs were about 21 or 22 years old and they were attending the course for the first time. Despite the fact that a few experiences on active learning already exist at the University of Minho (namely at the School of Health Sciences), PTs enrolled in this study had a single previous short experience on learning through PBL within the same course (Leite & Esteves, 2005).
Teaching Methodology Subjects were invited to watch a video clip on a four science teachers’ meeting (“What a confusion!…”) debating the issue of how to use the laboratory to teach science. Those teachers were having a discussion on the concept, role, varieties and evaluation of laboratory activities but they did not reach a conclusion about the issues raised. After watching the video, PTs were organized in small groups and asked to formulate questions that the video raised to them. Then, each group presented the questions to the whole class, discussed them, selected the relevant ones, and organized them in sets that were ordered based on the dependency of one set of questions on the answers of the other. Afterwards, PTs were asked to work in small groups (four or five members) to find out answers for each group of questions. To do so, they could use a variety of materials (like books, papers, dissertations, textbooks, etc) that were made available to them and look for other information sources to complement them. An oral presentation (followed by class discussion) by each small group was expected after reaching answers to each set of questions. After the presentation of the answers to the last set of questions, a global analysis of the learning that had taken place during the module as well as of the “teaching” approach adopted was carried out.
Data Collection Data were collected by means of a questionnaire, focusing on PTs’ opinions on PBL, and through a videotape of PTs’ final discussion and evaluation of the whole approach. The questionnaire used in this study is an improved version of the one mentioned in Leite & Esteves (2005). In addition, a self-evaluation grid was also filled-in by the PTs after the final discussion. The objective of the grid was to get information from students about the way they felt throughout the three students centered PBL cycle phases (formulating problem from scenarios, solving problems and presenting and discussing the problem solutions) as well as while members of a small group. The items related to the latter part of the grid were used in Leite & Esteves (2006). A minimum of two and a maximum of five items were formulated for each aspect. PTs were asked to rate each item on a four points scale (ranging from 1insufficient to 4-very good).
Data Analysis Answers to open questions as well as the videotape were content analyzed in order to find out the main issues pointed out by the PTs.
Teamwork and PBL-Based Teacher Education
91
Data collected through the self-evaluation grid were analyzed per aspect, so that mean and standard deviation values were obtained.
Results Through the questionnaire, PTs were asked to make an overall evaluation of the classes by pointing out positive and negative aspects. Content analysis of their answers geared ideas that were clustered around several positive and negative aspects. As far as positive aspects are concerned, table 1 shows that students pointed out issues related to students’ role, teacher’s role and the types of activities carried out. Table 1. Positive aspects pointed out by the PTs (N =38) Dimensions Students’ role Teacher’s role
Types of activities
Learning No answer
Positive aspects Students’ active role Teacher’s passive role Problem-solving activities Small group work Discussion between work groups Video watching Question formulation and organization Learning promotion through group work
f 12 6 11 10 3 3 2 1 3
Twelve PTs explicitly stated that they valued the fact that they were asked to play a more active role than they were used to in other classes. In addition, six PTs mentioned that enjoyed the fact that the teacher played a passive role. In the final discussion, subjects pointed out that although the teacher played a passive role, they nevertheless felt that she was always there, “keeping them on the track”, and giving them the necessary guidance to avoid time waste or insecurity feelings. In fact, moving from lecture-based classes to PBL environments requires changes of roles that are taken as necessary conditions for PBL to succeed (Yeung et al., 2003). Nevertheless, students need guidance not only to perform the cognitive task but also to successfully cooperate within the group. This is the reason why teachers continue having an outstanding role in PBL environments. The results of the questionnaire indicate that the activities enjoyed by larger numbers of PTs are problem solving and small group work (table 1). These types of activities are noy only typical from PBL environments (Davis & Harden, 1999; Lambros, 2004; Watts, 1991) but bearing in mind the literature mentioned in the theoretical part of this chapter, they were expected to be welcome by the PTs. The final discussion revealed that PTs enjoyed solving the problems because they focus on a resource – the lab – that is familiar to them (as science students) and that raises curiosity to them, as prospective teachers. As far as the functioning of the small group is concerned, when answering to the questionnaire, seven PTs stated that their groups did not work as such while the remaining 31 PTs mentioned that it was important for them to be allowed to work in mall groups because cooperation within the group
92
Laurinda Leite and Esmeralda Esteves
facilitated their job. It may have happened that in a few small groups there were different levels of engagement of the diverse elements of the group and that these interfered with the group work, as expected (Davis & Harden, 1999; Lambros, 2004; Watts, 1991). However, in the final discussion students recognized that if they had not been allowed to work in small groups, PBL would have turned learning into a too hard and not rewarding task for them. As far as the negative aspects are concerned, PTs mentioned mainly issues related to the tasks performed and to the workload associate with them (table 2). Table 2. Negative aspects pointed out by the PTs (N =38) Dimensions Tasks performed Workload Time available No negative aspect No answer
Negative aspects Presentation of problem answers Information search Work after the classes Workload in the classes Time restriction to the exploration of related issues
f 23 6 3 2 1 1 3
The most negative effect seems to have been due to the way the problem answers were presented to the whole class. In fact, the majority of the subjects (n=23) stated that the strategy used for the groups to present the problem answers (oral presentation of each group results) to the whole class was boring and repetitive. The requirement of having all the groups working on the same problems and making an oral presentation of the answers they got to their counterparts led to PTs hearing basically the same answers several times. In order to avoid such negative effect, other ways of presenting results from PBL work (namely portfolios, posters, reports), that are suggested by authors like Davis & Harden (1999) and Lambros (2002; 2004), should be considered. A few PTs (n=6) did write in the questionnaire that they did not enjoy doing literature search. This result was not surprising, as there is some evidence that Portuguese students do not enjoy reading too much. However, PBL requires reading and interpretation, even though they can be used alongside with other ways of collecting information. A few PTs also pointed out that the teaching approach used in the module increased their workload both in the classroom (n=2) and outside it (n=3). However, in the final discussion the majority of the participants in the study argued that the extra work was useful, and made them fell better prepared than usual for the final exam. One PT mentioned that the time devoted to the study of this module was not enough to explore and to deepen issues related to it. Another PT stated that he/she found no negative aspect on the set of classes focusing on the module. The majority of the PTs stated that contents within the scope of the module were approached at normal or deep levels (table 3). This means that they felt that there is no negative difference between the level at which they learned the contents themselves and the level at which teachers were used to teach contents to them. The final discussion suggests that PTs attribute this result to having the opportunity of doing literature search in diverse types of information sources.
Teamwork and PBL-Based Teacher Education
93
The results presented so far indicate that PTs enjoyed playing an active role on learning and felt that they could learn as much as they were used to learn in teacher centered classes. Table 3. Level at which contents were approached (N =38) Level Very superficial Superficial Normal Deep Very deep No answer
f 0 0 22 14 0 2
Table 4 shows the results obtained by means of the self-evaluation grid with regard to the way students belonging to a small group evaluate their role as members of the group and the performance of the group they belong to. Table 4. PTs evaluation of their performance and the performance of their groups (N=38) Group I II III IV V VI VII VIII IX X
Student’s self evaluation Mean SD 3,3 0,43 3,0 0,00 3,5 0,50 3,0 0,00 3,3 0,43 3,0 0,00 3,0 0,00 3,8 0,43 3,3 0,83 3,0 0,00
Student’s own group evaluation Mean SD 3,0 0,00 3,3 0,43 3,5 0,50 3,0 0,00 3,3 0,43 3,0 0,00 3,3 0,43 3,8 0,43 3,3 0,43 3,0 0,00
The results indicate that students in each group tend to value individual performances as much as the group performance. In fact, only three groups got mean scores different from the mean of the scores of their members: group 1 got a mean score lower than the means of the scores of its members; groups 2 and 7 got higher scores than the mean of the scores of their members. In these three groups standard deviation is larger for the groups with higher mean scores, what means that there is less agreement among their students than in the other case. The largest value of SD was obtained for group 9 students’ selfevaluation, and it is also larger than the SD of the group mean score. It is worth noticing that group 9 had a difficult student that was used to feel insecure about her performance and therefore valued herself as a member of the group quite low (2 - fair). In addition it is worth noticing that all mean scores are equal or higher than 3 (good) and that two groups (group 3 and 8) got self-evaluation scores as well as group evaluation scores that are equal or higher
94
Laurinda Leite and Esmeralda Esteves
than 3,5. This means that some students rated 4 (very good) both their individual performance and/or the group performance. Therefore, it seems that PTs enjoyed working in small groups and felt satisfied with their own performance and the performance of the group. Table 5 shows the results obtained through the self-evaluation grid with regard to the way students evaluate their performance throughout the three student centered phases of the PBL cycle. Members of four groups (2, 4, 5 and 10) rated their performance on the first PBL student centered phase (problem formulation, selection and hierarquization) bellow 3 (bellow good, that is, fair or insufficient). Phase 2 includes two sub-phases: 2a – use of literature; 2b – selection and treatment of information. Groups 2, 3, 4, 5 and 10 rated bellow 3 their performances in the former sub-phase. Only three groups (3, 6 and 10) did the same for sub-phase 2b. As far as phase 3 is concerned, it was divided into two subphases: 3a preparation of the presentation; 3b – presentation to the class. In the former case three groups (3, 4 and 10) got mean scores bellow 3. In the latter case, the majority of the groups (1 to 5, 8 and 10) got mean scores lower than 3. This means that the part of the cycle in which most groups of students felt that they performed worse is “making a presentation to the whole class was”. In addition, group 10 is a group whose members seem unhappy with their performance in all the phases of the PBL cycle. This result may be related to the fact that students were not used to speak in public, and were not used to present and discuss results with the whole class. However, it was very important that they could develop such competences as the year after they had to use those competences in their teaching practice classes. Table 5. PTs evaluation of their performance in the diverse phases of the PBL cycle (N=38)
Group I II III IV V VI VII VIII IX X
Phase I Mean SD 3,2 0,60 2,9 0,11 3,4 0,11 2,9 0,12 2,4 0,51 3,0 0,00 3,1 0,11 3,0 0,64 3,2 0,31 2,8 0,18
Phase II a Mean SD 3,0 0,50 2,4 0,22 2,8 0,25 2,3 0,24 2,8 0,25 3,8 0,25 3,3 0,43 3,3 0,83 3,0 0,35 2,3 0,25
PBL cycle phases Phase II b Phase III a Mean SD Mean SD 3,4 0,36 3,3 0,18 3,1 0,14 3,0 0,00 2,9 0,28 3,4 0,11 3,0 0,00 2,8 0,24 3,1 0,72 3,4 0,22 2,9 0,16 3,2 0,12 3,1 0,14 2,9 0,28 3,1 0,14 3,5 0,20 3,3 0,43 3,0 0,30 2,6 0,43 2,6 0,28
Phase III b Mean SD 2,8 0,56 2,8 0,43 2,8 0,43 2,7 0,47 2,8 0,43 3,7 0,47 3,1 0,41 2,8 0,47 3,3 0,62 2,9 0,22
Table 6 suggests that group 10 is the one whose members evaluate worse the functioning of the small group. Although its mean score is a bit above three (good), it is very low if it is taken into account that half of the groups got mean scores above 3.5 (out of a maximum of 4). These results suggest that students were quite happy with the way their small groups
Teamwork and PBL-Based Teacher Education
95
functioned and, therefore, they did not feel the group as a factor that could interfere negatively with the learning tasks. Table 6. PTs evaluation of the functioning of the group (N=38) Group I II III IV V VI VII VIII IX X
Mean
SD
3,8 3,5 3,8 3,6 3,3 3,6 3,2 3,6 3,5 3,1
0,17 0,17 0,14 0,27 0,30 0,28 0,58 0,49 0,38 0,22
Conclusion The results of this study suggest that prospective teachers valued PBL, did not have trouble with group work but reacted differently to the diverse phases and sub-phases of the PBL cycle, being the presentation sub-phase the one which PTs were more unhappy with. This reaction may be due to the fact that the participants in the study were unused not only to speak in public but also to argue for or against ideas. In fact, the final presentation required them to explain their answers, to argue for the points of view of their own group and against some points of view of other groups and to acknowledge that not all their answers were perfect. This involved not only communication and interpersonal skills but also content knowledge on the issue of using the lab for teaching science. The results obtained were expected based on the literature reviewed but they add to the knowledge base available in that they call our attention to the fact that although students show an overall positive reaction to small group based PBL, the intensity of this reaction depends on the activities they are asked to do. In addition, they suggest that despite the importance of having all students presenting the product of their work as a way of guaranteeing that they all invest in the job, the potential of alternative ways to oral presentations need to be explored in order to prevent the risk of having students experiencing the feeling of déjà vue which may negatively interfere with their motivation to debate the products obtained by the diverse groups. A poster may be an alternative to an oral presentation that is worthwhile testing. Finally, it can be argued that PTs’ positive reaction to a PBL based teacher education contexts indicates that they may be prone to use these teaching approaches when they become science teachers. The point is what kind of institutional support they will need and get first of all from their teaching practice supervisors and afterwards from their school authorities. Surely, opportunities to carry out PBL experiences during teaching practice will influence their future engagement with PBL.
96
Laurinda Leite and Esmeralda Esteves
References Acar, B. & Tarhan, L. (2006). Effect of cooperative learning strategies on students’ understanding of concepts in electrochemistry. International Journal of Science and Mathematics Education, 5, 349-373. Akinoğlu, O. & Tandoğan, R. (2007). The Effects of Problem-Based Active Learning in Science Education on Students´ Academic Achievement, Attitude and Concept Learning. Eurasia Journal of Mathematics, Science & Technology Education, 3 (1), 71-81. Albanese, M. & Mitchell, S. (1993). Problem based learning: a review of literature on its outcomes and implementation issues. Academic Medicine, 68, 52-81. Barell, J. (2007). Problem-based learning: An inquiry approach. Thousand Oaks: Corwin Press. Barron, B. et al. (1998). Doing with understanding: Lessons from research on Problem- and Project-Based Learning. The Journal of the Learning Sciences, 7 (3, 4), 271-311. Barrows, H. & Tamblyn, R. (1980). Problem-Based Learning: An approach to medical education. New York: Springer. Cartney, P. & Rouse, A. (2006). The emotional impact of learning in small groups: Highlighting the impact on student progression and retention. Teaching in Higher Education, 11(1), 79-91. Chin, C. & Chia, L. (2004). Problem-Based Learning: Using students’ questions to drive knowledge construction. Science Education, 88 (5), 707-727. Dahlgren, M. et al. (1998). PBL from the teachers’ perspective. Higher Education, 36, 437447. Davis, M. & Harden, R. (1999). Problem-Based Learning: A practical guide. Dundee: AMEE. Duch, B. & Groh, S. (2001). Assessment strategies in a Problem-Based Learning course?. In B. Duch et al. (Eds), The Power of Problem-Based Learning (pp. 95-108). Virginia: Stylus. Duch, B., Groh, S. & Allen, D. (2001). Why Problem-Based Learning? A case study of institutional change in undergraduate education. In B. Duch et al. (Eds), The Power of Problem-Based Learning (pp. 3-12). Virginia: Stylus. Dumas-Carré, A. & Goffard, M. (1997). Rénover les activités de résolution de problèmes en physique: concepts et démarches. Paris: Armand Colin. Edwards, S. & Hammer, M. (2004). Teacher education and Problem-Based Learning: Exploring the issues and identifying the benefits. Paper presented at the Conference of the Australian Association for Research in Education. Melbourne, November. Esteves, E. & Leite, L. (2005). Learning how to use the laboratory through problem based learning: A pilot study in an undergarduate physical sciences teacher education programme [Proceedings of the ATEE Conference. Amesterdam]. http://www.atee2005.nl/search/paperworks.php?contrid=121. Evensen, D. & Hmelo, C. (Ed.) (2000). Problem-Based Learning: A research perspective on learning interactions. Mahwah: Publishers. Fernández-Santander, A. (2008). Cooperative learning combined with short periods of lecturing. Biochemistry and Molecular Biology Education, 36(1), 34-38.
Teamwork and PBL-Based Teacher Education
97
Gandra, P. (2001). A Aprendizagem da Física Baseada na Resolução de Problemas. Um estudo com alunos do 9º ano de escolaridade na área temática “Transportes e Segurança”. Unpublished master dissertation, University of Minho. Gil, D. & Pessoa, A. (1994). Enseñanza de las ciencias y la matemática. Madrid: Popular. Gillies, R. (2003). Structuring cooperative group work in classrooms. International Journal of Educational Research, 39, 35-49. Grindstaff, K. & Richmond, G. (2008). Learners’ perceptions of the role of peers in research experience: implications for the apprenticeship process, scientific inquiry, and collaborative work. Journal of Research in Science Teaching, 45(2), 251-271. Hancock, E. & Gallard, A. (2004). Preservice science teachers’ beliefs about teaching and learning: The influence of K-12 field experiences. Journal of Science Teacher Education, 15(4), 281-291. Hintz, M. (2005). Can problem based learning address content and process? Biochemistry and Molecular Biology Education, 33(5), 363-368. Hmelo-Silver, C. (2004). Problem-Based Learning: What and how do students learn? Educational Psychology Review, 16 (3), 235-266. Hogan, K. (1999). Sociocognitive roles in science group discourse. International Journal of Science Education, 21(8), 855-882. Holloway, S. (1992). A potential wolf in sheep’s clothing: the ambiguity of cooperation. Journal of Education, 174(2), 80-99. Jenkins, E. (2006). The student voice and school science education. Studies in Science Education, 42, 49-88. Koh, G. et al. (2008). The effects of problem based learning during school on physician competency: a systematic review. Canadian Medical Association Journal, 178(1), 34-41. Lambros, A. (2002). Problem-Based Learning in K-8 classrooms. Thousand Oaks: Corwin Press. Lambros, A. (2004). Problem-Based Learning in middle and high school classrooms. Thousand Oaks: Corwin Press. Lambros, A. (2007). Transforming education with problem based learning. In B. Almeida et al. (Eds.), Encontro de Educação em Física: Do ensino básico ao superior no século XXI – livro de actas (pp. 65-72). Braga: University of Minho. Leite, L. & Afonso, A. (2001). Aprendizagem baseada na resolução de problemas: Características, organização e supervisão. Boletín das Ciencias, 48, 253-260. Leite, L. & Esteves, E. (2005). Ensino orientado para a Aprendizagem baseada na resolução de problemas na Licenciatura em ensino de Física e Química. In B. Silva & L. Almeida (Eds.), Actas do Congresso Galaico-Português de Psicopedagogia (Cd-Rom, pp. 17511768). Braga: University of Minho. Leite, L. & Esteves, E. (2006). Trabalho em grupo e aprendizagem baseada na resolução de problemas: Um estudo com futuros professores de Física e Química. Proceedings of the PBL2006ABP Congress (CD-Rom). Lima: PUCP. Martínez, M. et al (2001). Qué pensamiento profesional y curricular tienen los futuros profesores de ciencias de secundaria? Enseñanza de las Ciencias, 19 (1), 67-87. Nagel, N. (1996). Learning through Real-World Problem Solving: The Power of Integrative Teaching. Thousand Oaks: Corwin Press. Peeters, L. (2005). Méthodes pour enseigner et apprendre en groupe. Bruxelles: De Boeck.
98
Laurinda Leite and Esmeralda Esteves
Peterson, R. & Treagust, D. (1998). Learning to teach primary science through problem based learning. Science Education, 82, 215-237. Ratcliffe, M. & Grace, M.(2003). Science Education for Citizenship: Teaching SocioScientific Issues. Maidenhaid: Open University Press. Sage, S. (2001). Using Problem-Based Learning to teach Problem-Based Learning. In B. Levin (Ed.), Energizing Teacher Education and professional development with ProblemBased-Learning (pp. 87-107). Alexandria, Virginia: Association for Supervision and Curriculum Development. Savin-Baden, M. & Major, C. (2004). Foundations of Problem-Based Learning. Buckingham: Open University Press. Savin-Baden, M. (2000). Problem-Based Learning in higher education: Untold stories. Maidenhaid: Open University Press. Tan, O. (2004). Students’ experiences in Problem-Based Learning: Three blind mice episode or educational innovation? Innovations in Education and Teaching International, 41 (2), 169-184. Vermette, P. & Foote, C. (2001). Constructivist philosophy and cooperative learning practice: Towards integration and reconciliation in secondary classrooms. American Secondary Education, 30(1), 26-37. Vygotsky, L. (1986). Pensamiento y language: Teoria del desarollo cultural de las funciones psiquicas. Buenos Aires: Editorial la Pleyade. Watts, M. (1991). The science of problem-solving. London: Cassell Education. Woods, D. (2000). Problem-based learning: How to gain the most from PBL (2nd ed.). Hamilton: McMaster University, The Bookstore. Yeung, E. et al. (2003). Problem design in problem-based learning: Evaluating students’ learning and self-directed learning practice. Innovations in Education and Teaching International, 40 (3), 237-244.
In: Outsourcing, Teamwork and Business Management ISBN: 978-1-60456-956-8 Editor: Karl E. Carettas, pp. 99-113 © 2009 Nova Science Publishers, Inc.
Chapter 7
DEVELOPING EFFECTIVE TEAMS AND PROTECTING THE VULNERABLE: AN INTERPROFESSIONAL JOURNEY Susan Morison1 and Moira Stewart2 1
Centre for Excellence in Interprofessional Education, School of Medicine, Dentistry and Biomedical Science, Queen’s University, Belfast, Ireland 2 Department of Child Health, School of Medicine, Dentistry and Biomedical Science, Queen’s University, Belfast, Ireland
Abstract This chapter examines teamwork in medical care and makes particular reference to the pediatric team. It considers the characteristics of an effective team and the perceived benefits of team working in healthcare. It discusses the important role that education, and in particular interprofessional education, might have in helping to prepare a future workforce capable of effective patient focused team working. The contributory effect of different professional cultures is also examined and arguments are presented that reflect on the meliorating role of appropriate interprofessional education. The theoretical arguments are illustrated with reference to recent highly publicized and significant failures by teams responsible for the health and well-being of children. Recent case studies and judicial reviews from the United Kingdom and the United States are discussed.
Introduction The delivery of comprehensive healthcare across the lifespan is dependent on the skills and expertise of professionals from a wide range of disciplines. Although healthcare professionals and related organizations profess commitment to team working, the continued failure of medical teams has been identified as one of the underlying reasons for recent problems in the planning and delivery of healthcare. Moreover, it is in relation to society’s most vulnerable people that the failure of healthcare teams is most apparent.
100
Susan Morison and Moira Stewart
Increasing pressure to develop a healthcare workforce able to work collaboratively and function effectively in a wide variety of team settings generates a concomitant need to define and contextualize the skills required by an effective team worker. A corollary of this is recognition that these skills are not innate but need to be learned and therefore opportunities are required to educate and train healthcare professionals to work effectively in teams. In recent years interprofessional education, where students and practitioners from different healthcare professions learn together to promote collaborative practice (Barr 2005), has been seen as the optimum way to achieve this goal. This chapter begins by examining ideas about developing better teams, teamwork skills and support structures, and the barriers to effective team working created by issues such as professional culture and status. It considers why it is important that team working skills are learned and how this might be achieved. Some of the key difficulties encountered by teams responsible for the healthcare and protection of children are then outlined. Finally, there is a discussion about the role of interprofessional education in helping healthcare professionals to become more effective as team workers.
Developing Effective Teams The debate about the necessity for team working has resulted in general consensus that teamwork in healthcare is beneficial and emphasis has now shifted to identifying the most effective ways of using teams and the competences required by team members. Many of the recent calls for healthcare reform have indicated that teamwork is viewed as key to delivering more effective and efficient patient-centered care (Department of Health 2001, Laming 2003 and Institute of Medicine 2003). Research undertaken in the public sector has indicated that through teamwork, tasks can be performed more thoroughly, efficiently and cost-effectively and it has also been shown that teamwork can result in improved job satisfaction and staff who are more supportive of one another (Hayes 1997, Salas et al 2003). Teamwork has been credited with providing staff with an increased sense of being valued, and with being able to positively transform organizations (Katzenbach and Smith 1993 and Hayes 1997). Furthermore, within healthcare it has been suggested that teamwork can help to bring about a reduction in medical errors and improve the quality of patient care (Calman 1995, Groff 2003, Salas et al 2003 and Oandasan et al 2006). Team working involves more rather than less individual responsibility. Along with the tasks inherent within each professional role, there is a commitment to the well being of other team members and to making best use of available expertise and resources (National Initiative for Children’s Health Care Quality Project Advisory Committee 2001). The cost implications of failures in team working are also substantial. In the United States, during 1985 to 1996 it was estimated that $3.45 of the cost of every Emergency Room patient visit was attributable to litigated cash payouts associated with poor teamwork (Risser et al 1999). The discipline of Management has stimulated much thought on teamwork but the complex issues particular to healthcare cannot be underestimated. At any one time healthcare professionals are often members of a variety of different teams, all involved in patient care. Healthcare team members can therefore easily be described as members of a uni-professional, interprofessional and multi-professional team. Although these professionals may have a common goal - the provision of best quality patient care - their education and training,
Developing Effective Teams and Protecting the Vulnerable
101
management accountability and professional development, may be located within uniprofessional or other organizational structures. The setting of a team, for example whether hospital or community-based, will also determine team membership and function (Oandasen et al 2006). Teams must also be located in areas where a collective contribution is essential and there is clear distinction between common tasks that need to be undertaken by a team and distinct tasks, better done at an individual level. The Pediatric team is particularly useful to illustrate the complexity of team interaction. Its members will come from different professions, (medicine, nursing, social work, psychology, therapies allied to medicine and the police) will work across community and hospital services, and will involve not only the child but his or her carers. What is essential to the development of all effective healthcare teams however, is that the patient is regarded as the focal point (Academic Health Center Task Force on Interdisciplinary Health Team Development 1996). The literature on teamwork has generated a variety of ideas about the skills required by members of an effective team. An effective team should comprise of individuals with a diversity of knowledge, skills and expertise that extend the range of care provided to patients (Davis 1995 and Heath 1998). The team must be skilful in collaborating and hence communicating, decision-making, sharing knowledge and reaching consensus. Team members need to have a strong team identity and sense of belonging, share an integrated set of mutually valued and agreed goals that all members are committed to, and they should be flexible and responsive in their team practice (Brill 1976, West and Wallace 1991, Davis 1995 Cott 1998 and Borrill et al 2000). In addition they must be able to listen and respond constructively, to support other members and recognize when and how to help them (Katzenbach and Smith, 1993, Oandasan 2006). Team members need to be sensitive to group dynamics and have an agreed strategy for dealing with problems should the need arise and this is especially important with regard to the issue of status (Oandsan 2006). Motivation, trust, honesty, openness, consistency, respect clarity of purpose and of team roles and responsibilities are also considered important characteristics of a successful team (Davis 1995, Sundstrum 1998 and Boaden and Leaviss 2000). Finally, effective teams are those able to respond to difficult situations and adjust their practice and strategies appropriately (Salas et al 2003). Team structure is also important and members need to be empowered to function effectively (Hayes 1997). Effective teams have a two-fold level of operation. As well as achieving their patient-centered goal they must succeed in developing a team structure that allows them to do this. Task definition, cohesion, the establishment of rules and clearly defined team roles enable the team to function effectively. The larger the team, the easier it is for members to disengage and absolve from individual responsibility. Small teams are more easily organized and managed, but must include key elements in ensuring optimal patient outcomes. Teams need a clear support structure which is results-driven and which has effective monitoring of performance and feedback (Davis 1995, Sundstrum 1998). Teams require external support and recognition (Davis 1995) and essentially a working environment that enables members to participate and allows freedom within the team for different methods of working to be used (West and Wallace 1991 and Hayes 1997). The culture of the workplace must be one that recognizes that teamwork has an important role to play in optimizing patient care and is willing to make changes to accommodate this (Clements, Dault and Priest 2007). The team should be, and be recognized as, more effective as a whole than if the individuals were working separately and members must respect and
102
Susan Morison and Moira Stewart
trust one another’s skills, abilities, opinion and contribution, even where they include members with different skills, professional backgrounds, responsibilities and status (Hayes 1997). Standards of excellence should be established that can be attained through performance and the team must have fair and impartial leadership to champion and drive forward change (Davis 1995 and Clements, Dault and Priest 2007). Recognition of improvements in health care, following introduction of team working should be shared equally across all members of the team. In contrast, unless team objectives are clearly defined, team members may only pay lip service to teamwork and the result may be an ineffective team defined by its lack of belief in the efficacy of team working; resistance to sharing control and being dependent on others for achievement; feeling judged and scrutinized by others in the team, and lacking selfconfidence (Sundstrom 1998, Davis 1995, Katzenback and Smith 1993 and West and Wallace 1991). Lack of opportunity for learning about teamwork and the continuation of uniprofessional strongholds are also indications that teamwork may not be effective (Oandasan 2006). Encouraging team members to appreciate shared responsibility and trusting others rather than giving preference to individual responsibility and performance above that of the team, is essential. The effectiveness of a team should be judged not only by the team themselves but also by the patient and managers (Sundstrom 1998) and ideally the team should satisfy the expectations of its members and people outside the team with an interest in its action.
Leadership, Culture and Status Within healthcare, one of the most important and widespread working partnership is that between doctors and nurses. The doctor-nurse relationship and its effect on teamwork has been the subject of much discussion and particularly problems associated with the power struggle between an established, powerful, well-educated and traditionally male-dominated profession, and an emergent, female-dominated profession. The medical profession has long been identified as needing to relinquish some power to enable, where appropriate, joint decision making, nurses as team leaders and mutual respect. For a long time, interprofessional education and training has been seen as a means of addressing this (Hoekelman 1975, Keddy et al 1986, Casey and Smith 1997 and Salvage and Smith 2000) but it has taken recent major medical crises to provide the impetus for its implementation. Communication breakdown has consistently been linked to problems within the doctornurse relationship (Kalisch and Kalisch 1977, Benner 1984, Mackay et al 1995). Similarly, conflict has been identified as resulting from role misunderstanding (Clifford 1985, Webster 1985, Walby et al 1994) and particularly in relation to issues of responsibility and accountability where the professional perspectives and cultures are very different. For example, where Medicine stresses individual responsibility and independent judgment, Nursing is more bound by rules and codes of practice. Conflict has been identified in teams where doctors do not seek or listen to the opinion of nurses and do not seem to respect their experience, skills or knowledge (Mackay et al 1995). Doctors have also been found to have difficulty in delegating responsibilities to non-doctors (NHS Confederation 2002). Conversely, although changes in education have given nurses more confidence in their
Developing Effective Teams and Protecting the Vulnerable
103
dealings with doctors, some nurses remain reluctant to volunteer opinions and take a role in decision-making. Team leadership is inextricably bound in these issues of tribal conflict. Often, “leadership” is equated with “responsibility”, and within healthcare, ultimate responsibility for overall management has traditionally been the role of the doctor (Mackay et al 1995). One consequence is that there is both a perception and the reality that healthcare teams will be led by medical staff. Whether this leadership role is on the basis of traditional stereotypes, training and experience during early educational programmes or an innate reluctance on the part of other professionals to take on leadership, and the responsibilities implicit within the task, remains unclear. In the UK the implementation of the European Working Time Directive, which limits hours worked by junior doctors, has led to initiatives to redefine the roles and responsibilities of various healthcare professionals, (National Workforce Project 2007) which may lay more emphasis on the knowledge and skills within disciplines other than medicine. Enhanced status may, in turn, change dynamics within team working. It is unlikely however, that these difficulties will be overcome entirely by directives alone. Rather, these problems reinforce the growing recognition that healthcare professionals need to be educated to work in teams. Education for teamwork must be aware and take account of these tribal difficulties and boundaries and needs to encourage healthcare professionals to respect and value the distinctiveness of each profession (Ross and Southgate 2000) in order to work together effectively to improve patient care.
Learning to Work as a Team Member Developing a team with the requisite skills and structures is particularly difficult within healthcare where different professionals may join the team with different values concerning teamwork, depending upon their professional socialization and personal experiences and beliefs (Cott 1998). There is growing acknowledgement of the need to develop education and training specifically aimed at teaching team working skills. Importantly these skills, once learned, are transferable to the various teams that healthcare professionals are engaged in, and transmittable to colleagues (Salas et al 2003). Ultimately the effect of this should be to improve patient care and reduce adverse events and medico-legal complaints. Although it may be possible to assemble an interprofessional group, further effort is required to encourage cohesion and the development of team characteristics (Hayes 1997). It is often mistakenly accepted that teamwork skills will develop if people work together but this is not necessarily the case and the skills of collaboration need to be learned and teams to be developed (Horder 2000). In particular providing opportunities for team members to learn effective communication and listening skills and to explore some of the barriers, such as status, that prevent the achievement of team goals, can help to encourage the development of a team ethos (Hayes 1997). Further, as the development of team working skills is a long rather than short-term effort, many aspects of work experience can be utilized to help the development of team skills. Feedback on performance is regarded as particularly important in encouraging the development of effective team working and encouraging collaborative problem-solving approaches are appropriate and effective (Hayes 1997). Despite useful scraps of advice it is still the case that only a limited amount of research activity has been directed towards the development of education and training to improve team working
104
Susan Morison and Moira Stewart
(Leggatt 2007). Indeed, educational institutions and providers are regarded as being slow to embrace the need to develop such initiatives (Oandasan 2006). The arguments above do not only apply only to healthcare workers in practice. These skills need to begin to be learned during undergraduate or pre-qualification education and training (Salas et al 2003 and Leggatt 2007). Undergraduate education must ensure that students become effective communicators, understand how their role connects with those of others, and that they are familiar with problem solving approaches to learning. These are core professional skills, common to students from all healthcare professions but especially important in preparation for teamwork. Interprofessional education would appear to be one important approach that can help to provide a range of learning about teamwork. One of the essential aims of interprofessional education is to encourage healthcare students and professionals with different skills, abilities, values and status to appreciate and respect one another's different but complementary roles and appreciate how these different skills can contribute to the whole. Beginning to learn together during undergraduate education and continuing this into practice should encourage healthcare professionals to view collaboration as usual and core to their professional development rather than as something additional or irrelevant. To ensure the development and implementation of appropriate educational opportunities there needs to be a framework that facilitates interprofessional communication and collaboration at all levels – from students through to practitioners and policy-makers. Each profession has its own clearly identified knowledge and skills base which, in the case of UK healthcare professions, is described by the regulatory body responsible for ensuring that professional standards are met and maintained (General Medical Council and Nursing and Midwifery Council for example). In the United States, professional standards are mainly overseen by state government agencies. However, professional organizations such as the American Medical Association, and American Nurses Association have an integral role in setting standards for licensing and determining scope of practice. These professional bodies need to lead by example and work collaboratively and communicate effectively if appropriate, integrated strategies for education for teamwork are to result.
Team Work in Child Care Although an essential feature of most patient care, healthcare teams operate at different levels - from the fundamental level, where administrative and clerical staff facilitate the contact between patient and professional, to the other end of the spectrum where the problem is more complex and severe and where there is greater need for an integrated, team approach (Oandasan 2006). There are patient groups, such as children, for whom, due to their dependence on others for basic levels of care, the emphasis on team working is especially important. The importance and complexity of the Pediatric team provides fertile ground for exploration and research to deconstruct its functions and aid understanding of how better to prevent mistakes and improve outcomes for patients and families. Health is more than the absence of disease, and includes psychosocial and physical wellbeing. In the case of children, health allows each child the opportunity to achieve their potential for growth, development, education and physical well-being. Technological advances mean that more children with previously incurable conditions are surviving, at least
Developing Effective Teams and Protecting the Vulnerable
105
beyond the short-term, and it is now estimated that 15-20% of all children have chronic disease/disability that requires ongoing intervention from professionals (Newacheck et al 1998). Their management extends beyond the child and immediate family, into schools, community with societal and economic implications and demands a high level of interprofessional involvement. In the UK, as a result of the Court Report of 1976, team and team working are well established within child health services. This report was a major government commissioned document which reviewed services for children in the UK and made a series of recommendations which have been the foundation for present day pediatric and child health services and highlighted the need for better integration of services across all levels of healthcare and involving professionals from a wide variety of backgrounds (Committee on Child Health Services 1976). More recently professional guidelines on commissioning tertiary and specialized services for children, stress the need for multiprofessional teams and clinical networks to facilitate the delivery of integrated care pathways (RCPCH 2004). The individual child with chronic disease and their family will almost certainly encounter team working at some level. However, even in many established teams, deficiencies in team working and communication have been identified as major factors in failures in delivery of care to vulnerable patient groups such as children with disability and those in need of child protection. It is important that current and future developments retain the former and eliminate the latter. Ensuring that Government policy, which dictates professional practice, takes cognizance of this, is one way forward. Over the past 10 years the UK has witnessed two high profile judicial reviews which have had had a major impact on the processes which must be in place to achieve improvements in standards of care. Both reviews highlighted deficiencies in communication and team working which directly impacted on the death of children. The death of a child, following months of “truly unimaginable abuse”, led to the Victoria Climbie Inquiry (Laming 2003). The report made many recommendations, several to do with professionals working together. In 1999, the Government responded with a White Paper, Working Together to Safeguard Children, which was a joint document from the UK Department of Health, Education and Employment and the Home Office (Department of Health 2006). It took account of principles contained within United Nations Convention on the Rights of the Child (UNCRC), the European Convention on Human Rights (ECHR) and the requirements under the Children Act 1989. While not having the full force of statute, professionals in the UK are required to comply with its recommendations on practice to promote children’s welfare and protect them from abuse and neglect. As the title suggests, the emphasis is on interprofessional working, sharing of information and responsibility to achieve the best outcomes for the child and his/her family. It therefore goes beyond drawing up new legislation to an expectation that it will be implemented with evidence of joined up working to protect children (Keenan 2006). The public inquiry set up to examine the management of care of children receiving complex heart surgery at the Bristol Royal Infirmary was presented to Parliament by the Secretary of State in 2001 (Department of Health 2001). Over 30 children under the age of one year died following open-heart surgery between 1991 and 1995. The Inquiry concluded that there were a number of contributing factors, including system failures, clinical competences, uni-professional power imbalances and lack of critical evaluation of outcomes. The Inquiry was particularly critical of the problem of poor teamwork and its implications for
106
Susan Morison and Moira Stewart
performance and outcome, despite the general recognition of the crucial importance of effective teamwork in this complex and relatively uncommon area of surgery. One key recommendation was that all “prospective healthcare professionals, therefore, must receive education and training in the meaning of teamwork, how to work effectively in multidisciplinary teams, how to deal with the issues of accountability which arise in teams, and the role of teams in providing healthcare.” (Department of Health 2001 p338). As discussed earlier, the existence of a team does not automatically confirm team working at a level that improves patient care, avoids adverse outcomes and makes best use of professional expertise and this is particularly important in the Pediatric team. Because of the team’s complexity it is vital to have effective, timely communication, mutual respect and a willingness to defer to others. There must be clearly defined areas of responsibility, measures to identify failures in provision of care and proper attention to issues of clinical governance. The child who is brought for professional attention, no matter whether prior failures in care giving have caused the problem, is less at risk than the child discharged from care because of non-attendance. A recent UK pilot project, investigating child deaths (CEMACH 2008a and b) found evidence in a number of child deaths of lack of clearly defined care pathways for vulnerable children who were never given the opportunity to have their needs addressed. Uniprofessional concerns without sharing of information can mean that the true extent of problems are ignored. Furthermore, CEMACH concluded that within acute health care, decision-making still lies within the remit of medical staff and when there are differences in interpretation of findings, there is a reluctance to challenge the perceived authority of doctors. Even when there is evidence of interprofessional team working, such as within Emergency departments, the continued existence of hierarchical structures and uni-professional roles mean that more weight is given to decisions made by medical staff. The CEMACH enquiry concluded that, for the majority of child deaths that were felt to be avoidable, there were multiple factors contributing to “system error” and that the outcome might have been different had parents and nursing staff felt empowered to follow up their concerns (CEMACH 2008a). Shared responsibility can, in practice, mean that no one healthcare professional or professional group feels the need to intervene to address problems. This is especially relevant within Child Protection. The Inquiry into the death of Victoria Climbie was particularly critical of the lack of joined up working, so that while uni-professional concerns were recorded, the failure to address these at interprofessional level, meant that appropriate action was not taken and the child died following months of abuse and many opportunities for avoiding actions (Laming 2003). The need to share information and responsibility in difficult areas of work such as child protection exceed other issues such as confidentiality and fear of criticism. Improvements in ability to audit activity and increased emphasis on clinical governance should result in better patient care. The American Academy of Pediatrics aims to provide the best and safest healthcare for infants, children, adolescents and young people. In a policy statement on Principles of Patient Safety in Pediatrics, it recognized that teamwork and communication are the basis for achieving changes in willingness to recognize that the potential for errors does exist within medicine. Collaboration and pooling of professional expertise may then allow improvements in patient safety above those possible if engaging
Developing Effective Teams and Protecting the Vulnerable
107
with uni-professional groups (American Academy of Pediatrics National Initiative for Children’s Health Care Quality Project Advisory Committee 2001). As expertise at uni-professional level increases so does the size and complexity of the team. The expectations of the public generally, and parents in particular, is that all possible options for diagnosis, management and care are explored. The demands on the skills needed to combine all of these elements within a coherent team working structure mean that learning to work together must be an integral part of pre and post qualification curricula. It is only with appropriately educated, responsive team members that teams of the future can expand in a flexible manner to meet these new and increasing demands.
Role of Interprofessional Education All healthcare teams have a duty to care for and protect patients and yet, paradoxically, ineffective team functioning has resulted in society’s most vulnerable being exposed to harm by those who should be empowered to prevent it. In recognition of the need to rectify these failings and ensure that future healthcare workers are prepared for complex, collaborative practice, policy makers and professional regulators in the UK and USA have responded by including learning about teamwork and communication as core aspects of healthcare curricula at both student and practitioner levels (DoH 2005, GMC 2000, 2001 and 2002, NMC 2001 and Institute of Medicine 2003). Interprofessional education, or learning together, is regarded as a way to provide the educational basis for a more flexible workforce able to adapt to the challenges of new approaches to clinical practice and governance, and respond to patients’ changing expectations (Department of Health 2000). The premise underpinning developments in interprofessional education is that through learning together students and practitioners from different healthcare professions will learn to understand and respect one another’s roles and responsibilities and to communicate and collaborate effectively to improve patient care. These are precisely the knowledge and skills that have been identified as lacking in ineffective or failing healthcare teams. There is growing evidence to support the contention that interprofessional education does help to improve collaborative practice (Barr 2005) although further empirical and theoretical research continues to be necessary to convince skeptics of its effectiveness. If interprofessional education is to make a significant contribution as a vehicle for education for teamwork then it would seem appropriate that it should begin during prequalification education and continue into practice. Although a variety of studies have been conducted to evaluate the effectiveness of interprofessional education designed for students and practitioners, it has now been recognized that each of these levels has a different focus (Barr 2005, Morison 2008). Interprofessional education designed for pre-qualification students should essentially be focused on preparation for future professional practice and should promote better understanding of teamwork, collaboration and professional roles and responsibilities. At practitioner level it is more about encouraging those already working in professional practice to develop their team working skills. Indeed examples of successful interprofessional initiatives designed to improve teams working with children and families have already been reported (Sicher et al 2000 and Larivaara and Taanila 2004). Although a greater number of reported studies evaluating the effectiveness of interprofessional education are concerned with practitioners (Barr 2005), where direct effects
108
Susan Morison and Moira Stewart
on practice are comparatively easy to measure, an increasing body of evidence is emerging to support the contention that appropriately focused pre-qualification interprofessional education is beneficial to students (Salas et al 2003, Barr et al 2005 and Morison and Jenkins 2007). Not only can it make a valuable contribution to improving team working skills it can also encourage a more inclusive approach to professional practice (Harden 1998, Miller 1999, Morison et al 2003 and Salas et al 2003) thus helping to dispel professional silos. The dominance of uni-professional approaches contributes to the preservation of status quo (Salas et al 2003) and traditional attitudes and it is therefore essential to break this stronghold if teamwork and education for teamwork is to succeed. Central to the argument for the early introduction of interprofessional education is the premise that effective team working involves more than having knowledge of different roles and responsibilities - indeed, ineffective teams are characterized by their lack of knowledge, understanding and appreciation of others’ roles and responsibilities. Interprofessional education can enable students to acquire deeper understanding of the roles of others, develop appropriate attitudes and have a genuine commitment to the team ethos (Morison and Jenkins 2007). Advocates of the introduction of early interprofessional learning also recognize that status, tribalism and professional stereotyping and identity, barriers to successful teamwork, are equally obstacles to the development of interprofessional education (Headrick et al 1998, Barr et al 2005, Morison and O’Boyle 2008 and Morison et al 2008) and these issues need to be addressed early in the education continuum (Mandy et al 2004, Adams et al 2006, Morison and O’Boyle 2008 and Morison et al 2008) if collaborative practices and open-minded attitudes are to result. Interprofessional education can help healthcare students and professionals to develop a common understanding of teamwork and appreciation of the value of different roles and responsibilities in patient care and management. Essentially, interprofessional education should be designed as an interactive learning opportunity and implemented where common skills and knowledge can be found, and ideally linked to an area where patient-centered teamwork is a feature of practice (Parsell and Bligh 1999, Hall and Weaver 2001, Horsburgh et al 2001 and Morison et al 2003). Healthcare of children and child protection are areas with these overlapping requirements. As discussed earlier, the problems that have been identified in failing pediatric teams include poor communication, inadequate sharing of information, lack of clear guidelines on responsibility and accountability and over-reliance on medical decision-making. These are all areas where interprofessional education can make a difference in providing early opportunities for students to learn together about these issues and there is evidence to support the efficacy of this approach (Barr et al 2005). Through shared, interactive learning opportunities students are enabled to gain an interprofessional perspective on individual practices. Those with responsibility for healthcare education, at all levels, also need to be convinced that interprofessional education can improve team practice, which in turn will improve patient care and be cost-effective. It is vital to have learning and practice environments that will support interprofessional education and foster the development of a teamwork culture. Salas (2003) argues that key to engendering this culture is to be able to identify when a team, rather than a single professional is essential and to provide adequate training and support. There is already sufficient evidence about the failures of child protection teams to suggest that this area of clinical practice will benefit hugely from the introduction of interprofessional learning opportunities early in the pre-qualification curriculum.
Developing Effective Teams and Protecting the Vulnerable
109
The logistical problems of introducing interprofessional education, including already currently over-stretched curricula and restricted timetables cannot be underestimated. However, an interprofessional approach to learning needs to be promoted as a different, rather than an additional, way of working. A further requirement is that team working skills must be perceived as of equal or greater importance to the acquisition of factual knowledge and practical skills. Failure to do so on the part of teachers and students immediately negates their pivotal role in providing high quality patient care. Similarly, students need to be aware that they will be expected to demonstrate continuation and progression as they advance through their pre-qualification education and into practice and that these skills are recognized by their profession’s leaders. There remains one further stumbling block to the acceptance of interprofessional education and whole-hearted commitment to team working and this concerns the blurring and merging of professional roles. The dominant view is that interprofessional education should promote teamwork and exemplify how different skills and expertise can be combined in teams to provide effective patient care that retains respect for the integrity and contribution of each profession (Heath 1998, Calman 1999, Pirrie 1999, McPherson 2001). Healthcare professionals may still need some reassurance from policy-makers and interprofessional educators that there is no hidden agenda to abolish professional boundaries and identities behind the promotion of interprofessional learning. Instead it should be seen as vital for the protection of patients and particularly those who have been failed by professionals in the past.
Conclusion Advances in healthcare have increased the need for delivery of services to be shared by increasing numbers of professionals. The effectiveness and acceptability of such services depend on team working and the skills associated with it. It is now generally accepted that teamwork is beneficial to patient care and the characteristics of an effective team have been defined. Conversely, failures in team working have been identified as major sources of adverse outcomes. There is a clear need for healthcare workers to learn team working skills and have a better understanding of the complexities of team working but little has been done to address this. Issues such as status and professional culture that can adversely affect team effectiveness also need to be acknowledged and included in educational initiatives. Effective teamwork can improve clinical practice and contribute to reduction of errors. This is especially important in teams responsible for children and vulnerable groups. Common goals for the team should involve not only patient-centered benefits but also effective communication, mutual respect for the skills and expertise of others and the willingness to share responsibility. Education, and in particular, interprofessional education has a vital role to play and will be most beneficial if appropriate opportunities to learn together are identified and introduced early in the curricula of healthcare students and continued into practice. Essentially interprofessional education has the potential to enable students and healthcare professionals from different backgrounds to learn about and experience teamwork and its role in protecting patients by improving patient care, management and safety.
110
Susan Morison and Moira Stewart
References American Academy of Pediatrics National Initiative for Children's Health Care Quality Project Advisory Committee (2001). Principles of Patient Safety in Pediatrics. Pediatrics, 107, 6, 1473-1475. Academic Health Center Task Force on Interdisciplinary Health Team Development (1996). Developing Health Care Teams: A Report by the Academic Health Center Task Force on Interdisciplinary Health Team Development. Minnesota: University of Minnesota. Adams, K. Hean, S. Sturgis, P. and Macleod Clark, J. (2006). Investigating the factors influencing professional identity of first-year health and social care students. Learning in Health and Social Care, 5, 2, 55-68. Barr, H. Freeth, D. Hammick, M. Koppel, I. and Reeves, S. (2005). Effective Interprofessional education: Argument, Assumption and Evidence. London: Blackwell Publishing. Benner, P. (1984). From Novice to Expert: Excellence and Power in clinical practice. California, Addison. Boaden, N. and Leaviss, J. (2000). Putting teamwork in context. Medical Education, 34, 921927. Borrill, C. West, M.A. Shapiro, D. Rees, A. (2000). Team working and effectiveness in health care. British Journal of Health Care, 6, 364-71. Brill, N. (1976). Teamwork: Working together in the Human Services. Toronto: Lipincott. Calman, K. and Hine, D. (1995). A Policy Framework for Commissioning Cancer Services. A report by the Expert Advisory Group on Cancer to the Chief Medical Officers of England and Wales. London: HMSO. Casey, N. and Smith, R. (1997). Bringing nurses and doctors closer together. British Medical Journal, 314, 617-618. Confidential Enquiry into Maternal and Child Health (2008a). Why Children Die: A pilot Study 2006. London: CEMACH Confidential Enquiry into Maternal and Child Health (2008b). Why Children Die: A pilot Study 2006, Children and Young People’s Report. London: CEMACH Clements, D. Dault, M. and Priest, A. (2007). Effective Teamwork in Healthcare: research and reality. HealthcarePapers, 7(Sp), 26-34. Clifford, C. (1985). Nurse-Doctor relationships ; is there cause for concern? Nursing practice, 2, 102-108. Committee on Child Health Services (Chairman Court D) (1976). Fit for the Future. London: HMSO Cott, C. (1998). Structure and meaning in multidisciplinary teamwork. Sociology of Health and Illness 20,6, 848-873. Davis, JR. (1995). Interdisciplinary Courses and Team Teaching. Arizona: American Council on Education and the Oryx Press. Department of Health (2000). The NHS Plan: a plan for investment, a plan for reform. London: TSO. Department of Health (2001). Learning from Bristol: The Report of the Public Inquiry into Children’s heart Surgery at the Bristol Royal Infirmary 1984-95. London: TSO.
Developing Effective Teams and Protecting the Vulnerable
111
Department of Health (2005). Curriculum for the foundation years in postgraduate education and training. London: Department of Health. Department of Health (2006). Working together to protect children. London: TSO. General Medical Council (2000). Teamworking in Medicine. London: GMC. General Medical Council (2001). Good Medical Practice. London: GMC. General Medical Council (2002). Tomorrow’s Doctors. London: GMC Groff, H. (2003). Teamwork Failures Noted in Malpractice Claims. Risk Management Foundation Harvard Medical Institutions, 23, 3-4. Hall, P. and Weaver, L. (2001). Interdisciplinary education and teamwork: a long and winding road. Medical Education, 35, 867-875. Harden, R.M. (1998). Effective multiprofessional education: a three-dimensional perspective. Medical Teacher, 20, 402-408. Hayes, N. (1997). Successful Team Management. London: Thomson International Business Press. Heath, I. (1998). A seamless service. British Medical Journal, 317, 1723-1724. Headrick, L. Wilcock, P. and Batalden, P. (1998). Interprofessional working and continuing medical education. British Medical Journal, 316, 771-774. Hoekelman, R. (1975.) Nurse-Physician relationships. American Journal of Nursing, 75, 7, 1150-1152. Horder, J. (2000). Leadership in a multiprofessional context. Medical Education, 34, 203-205. Horsburgh, M. Lamdin, R. and Williamson, E. (2001). Multiprofessional learning: the attitudes of medical, nursing and pharmacy students to shared learning. Medical Education, 35, 876-883. Institute of Medicine of the National Academies (2003). Health Professions Education: A Bridge to Quality. Washington: The National Academies Press. Kalisch, B. and Kalisch, A. (1977). An analysis of the sources of physician-nurse conflict in Muff, J (ed) (1982) Socialization, Sexism and Stereotyping: Women's Issues in Nursing. Cleveland: Prospect Heights. Katzenbach, J.R. and Smith, D.K. (1993). The Wisdom of Teams: creating the high performance organisation. Boston: Harvard Business School Press. Keddy, B. Jones Gills, M. Jacobs, P. Burton, H. and Rogers, M. (1986). The doctor-nurse relationship: an historical perspective. Journal of Advanced Nursing, 11, 745-753. Keenan, C. (2006). Lessons from America? Learning from child protection policy in the USA. Child and Family Law Quarterly, 18, 1, 43-66. Laming, L.J. (2003). Inquiry into the death of Victoria Climbie. London: TSO. Leggat, S.G. (2007). Effective healthcare teams require effective team members: defining teamwork competencies. BMC Health Services Research, 7, 17. Available from url: www.biomedcentral.com/1472-6963/7/17 Mackay, L. Soothill, K. and Webb, C. (1995). Troubled times: The context for interprofessional collaboration? In: Soothill, K. Mackay, L. and Webb, C. (eds) Interprofessional Relations in Health Care. London: Arnold, pp5-10. Mandy, A. Milton, C. Mandy, P. (2004). Professional stereotyping and interprofessional education. Learning in Health and Social Care,3,3,154-170. McPherson, K. Headrick, L. and Moss, F. (2001). Working and learning together: good quality care depends on it, but how can we achieve it? Quality in Health Care 10 (Suppl II): ii46-ii53
112
Susan Morison and Moira Stewart
Miller, C. Ross, N. and Freeman, M. (1999). The role of collaborative shared learning in pre and post registration education in nursing, midwifery and health visiting. London: ENB. Morison, S. Boohan, M. Jenkins, J. and Moutray, M. (2003). Facilitating undergraduate interprofessional learning in healthcare: comparing classroom and clinical learning for nursing and medical students. Learning in Health and Social Care, 2, 92-104. Morison, S. and Jenkins, J. (2007). Sustained effects of interprofessional education on student attitudes to communication and team working depend on learning together in the ward as well as the classroom. Medical Teacher, 29, 5, 450 - 456. Morison, S. Marley, J. Stevenson, M. and Milner, S. (2008). Preparing for the dental team: investigating the views of dental and dental care professional students. European Journal of Dental Education, 12, 23-28. Morison, S. and O’Boyle, A. (2008) Developing professional identity: a study of the perceptions of first year nursing, medical, dental and pharmacy students. In Callarra L (ed) Nursing Education Challenges in the 21st Century. New York: Nova Science Publishers Inc. Pirrie, A. Wilson, V. Harden, R. and Elsegood, J. (1999). Promoting cohesive practice in healthcare. In: AMEE Education Guide No 12: Multiprofessional Education. Dundee: AMEE. Parsell, G. and Bligh J. (1999). Educational Principles Underpinning Successful Shared Learning. In: AMEE Education Guide No 12: Multiprofessional Education. Dundee: AMEE. NHS Confederation (2002). The problem of unhappy doctors: what are the causes and what can we do? London: NHS Confederation. NHS (2007). Introducing NHS National Workforce Projects. Manchester: NHS Newacheck, P. Strickland, B. Shonkoff, J. Perrin, J. McPherson, M. McManus, M. Lauver, C. Fox, H. and Arango, P. (1998) An Epidemiologic Profile of Children with Special Health Care Needs. Pediatrics, 102, 1, 117-123. Nursing and Midwifery Council (2001). Fitness for Practice and Purpose. London: NMC. Oandasan, I. et al (2006). Teamwork in Healthcare: Promoting Effective Teamwork in Healthcare in Canada. Ottowa: Canadian Health Services Research Foundation. Risser, D.T. Rice, M.M. Salisbury, M.L. Simon, R. Jay, G.D. and Berns, S.D. (1999). The Potential for Improved Teamwork to Reduce Medical Errors in the Emergency Department. Annals of Emergency Medicine, 34,3,373–83. Ross, F. and Southgate, L. (2000). Learning together in medical and nursing training: aspirations and activity. Medical Education, 34,739-743. Royal College of Paediatrics and Child Health (2004). Commissioning Tertiary and Specialist Services for Children and Young People. London: RCP. Salas, E. Sims, D.E. Klein, C. and Burke, C.S. ( 2003). Can Teamwork Enhance Patient Safety? Risk Management Foundation Harvard Medical Institutions, 23, 5-9 Salvage, J. and Smith, R. (2000). Doctors and nurses: doing it differently. British Medical Journal, 320, 1019-20. Sicher, P. et al (2000). Developing child abuse prevention, identification and treatment systems in Eastern Europe. Journal of the American Academy of Child and Adolescent Psychiatry, 39, 660-67. Sundstom, E. (1998). Supporting work team effectiveness: Best management practices for fostering high performance. San Francisco: Jossey-Bass.
Developing Effective Teams and Protecting the Vulnerable
113
Walby, S. Greenwell, J. Mackay, L. and Soothill, K. (1994). Medicine and Nursing: professions in a changing health service. London: Sage. Webster, D. (1985). Medical Students' Views Of the Role of the Nurse. Nursing Research, 34, 5, 313-317. West, M.A. and Wallace, M. (1991). Innovation in healthcare teams. European Journal of Social Psychology, 21, 303-15.
In: Outsourcing, Teamwork and Business Management ISBN: 978-1-60456-956-8 Editor: Karl E. Carettas, pp. 115-134 © 2009 Nova Science Publishers, Inc.
Chapter 8
THE IMPACT OF ENGINEERING DESIGN ON OUTSOURCING DECISIONS Mahmood Al-Kindi* and Ali A. Yassine Product Development Research Laboratory, Department of Industrial and Enterprise Systems Engineering University of Illinois at Urbana-Champaign Urbana, IL 61801 USA
Abstract Many models in the literature examine outsourcing based on product modularity; however, modularity is assumed to be known and exogenous. In reality, modularity is a decision variable defined (i.e. built into the product) during the engineering design phase of the product development (PD) process. This paper bridges the gap between the outsourcing literature and the engineering design literature by incorporating into the outsourcing decision model detailed engineering design information regarding the time spent on various engineering design activities within the PD process (e.g., system design, detailed design, and testing and integration). In this paper, a mathematical model is developed to study the impact of outsourcing and time spent in the various engineering design activities on firm’s revenue (represented by a marketing window) for different PD scenarios. These scenarios differ in four major factors: technological capability of the firm and its suppliers, design task size and complexity, nature of detailed design work (i.e., fraction of rework), and amount of outsourcing. It is shown that this model is a convex optimization problem that admits a global optimum; however, no explicit closed-form solution could be obtained and the problem was solved using optimization software. The optimal solution reveals several interesting managerial insights regarding the impact of the various engineering design decisions on the outsourcing decision. First, spending more time in system design leads to higher outsourcing fraction and vice versa; that is, well defined product architectures lead to higher outsourcing. Second, higher firm capability makes outsourcing less attractive. Third, outsourcing is found to be more attractive at the medium task sizes compared to larger or smaller tasks. Fourth, a product with a complex architecture will lead the firm to spend more time in system design and thus outsource more. Lastly, as the *
E-mail address:
[email protected]. Tel. (217)333-7621. (Corresponding author)
116
Mahmood Al-Kindi and Ali A. Yassine rework fraction of detailed design increases, it is better to spend more time in system design and outsource more.
Keywords: outsourcing decision, product design & development, engineering design, modularity, product complexity, task size, firm capability.
Nomencalture Ttotal : o d
T
:
Product life cycle. Total time for the whole process [unit time] Nominal time spent in detailed design [unit time]
TdoOEM : Detailed design duration for OEM, including all rework TdoSupplier : Detailed design duration for Supplier, including all rework TS :
Time spent in system design [ unit time]
Tm :
Marketing window
POEM : Success probability forthe OEM at testing and Integration PSupplier : Success probability forthe OEM at testing and Integration
α : γ: λ: μ: −
μ: ν: ε: G:
Fration of rework [%] Shape parameter for product complexity in system design [γ ≥ 0] Task size at detailed design phase [unit time] OEM (Firm) capability [μ ≥ 1] Modified OEM capability Information gap between OEM and Supplier Outsourcing fraction[ε ≥ 0] Product generations
1. Introduction The aggressive nature of competition in today’s markets makes product development (PD) a central point of contest. The benefit goes to the companies that are able to efficiently introduce new products into the market. These firms guide their development effort toward three main objectives: low price, high quality, and long marketing window. Although these objectives are often clashing or conflicting, they must be compromised using an optimal PD process. The product development (PD) process is a sequence of all the essential tasks that a firm must perform to develop, manufacture and sell a product (Ulrich and Eppinger, 2004). These tasks include marketing research (where customer needs are identified and product attributes are specified), system design (which includes product architecture definition, subsystem identification and interfaces), engineering design (also referred to as detailed design and includes fully specifying the product dimensions, tolerances, and material), prototyping (which includes product validation and testing), manufacturing planning (which includes
The Impact of Engineering Design on Outsourcing Decisions
117
process design for ramp-up and full production), and supply chain design (which may involve a large number of suppliers). Outsourcing plays a significant role within the above product development process. Firms outsource for various reason. The main reason is simply cost considerations. Generally, buying components from a supplier costs less than making them in-house. Another driving force is capacity. Sometimes an Original Equipment Manufacturer (OEM) does not have enough production capacity to make a component or product. Lack of knowledge, expertise, or even quality, could be another motive for outsourcing. Additionally, firms tend to outsource noncore (or noncritical) components or subsystems that either have low influence or impact on product performance (or quality), which may be critical to the marketability or image of the product, or that do not constitute a core competence for OEM’s long-term survivability and competitiveness. Outsourcing has various other dark sides, which also make it less attractive. A major one is the loss of knowledge or basically eroding knowledge know-how for producing the outsourced components or subsystems by reducing cumulating learning-by-doing. As a result, an OEM could be trapped into outsourcing (long-term dependence on supplier for knowledge and know-how), which makes it more expensive (than what it used to be) to build the components in-house in the future. Another down side is the potential for loss of local jobs within the community, which ultimately could result in loss of market share in local markets. The careful allocation of time and resources to various development activities, in addition to determining the right amount of outsourcing, are critical for successful and profitable product development (PD). Essentially, PD is a sequential process where various decisions in each stage are made in conjunction with both former stages and later ones. Therefore, spending more time in system design activity during PD leads to more modular products, which may in turn be more amenable to outsourcing. This makes outsourcing decisions directly related to design decisions and particularly the effort spent on each phase (or stage) of the development process (i.e., system design, detailed design, and integration). Clearly, the probability of integration failure for a modular architecture is lower than that for an integral architecture due to well defined product modules (e.g., subsystems) and the various interactions between them. However, spending less time in system design may lead to higher probability of integration failure for the outsourced modules due to the nature of product architecture and the lack of proper or ideal communication with the OEM (compared to communications between two departments within the same organization) and lack of the systems perspective that usually resides within the OEM and hard to communicate to the supplier of an outsourced module. Few outsourcing models take into account details of the engineering design process; mainly discussing the impact of product modularity as laminated exogenous variable. In this paper, we consider detailed design process information (e.g. time spent on each design activity) in order to arrive at an optimal outsourcing strategy. Thus, our model bridges the PD process literature and the outsourcing literature into a unified model. A good PD outsourcing model should not only include time and cost considerations, but also does not ignore the loss of knowledge which affects negatively the OEM’s future capability. Although, it might take years to erode design knowledge and know-how within a development organization, we consider the average of the long term capability as the OEM’s current capability in order to account for this long-term loss. These various assumptions are described in details in later sections.
118
Mahmood Al-Kindi and Ali A. Yassine
In this paper, a mathematical model is formulated to maximize OEM’s revenues through a larger marketing window. The decision variables are the time spent in system design and outsourcing fraction. Different scenarios within the product development process will be investigated. These scenarios differ in product complexity (i.e. newness), task size (e.g., small and large), firm capability (i.e. resource levels and expertise), and nature of detailed design work. It is shown that this model is a convex optimization problem that admits a global optimum; however, no explicit closed-form solution could be obtained and the problem was solved using optimization software. The rest of the paper is organized as follows. The relevant literature is described and discussed in the next section. In Section 3, we present the formulation for the proposed model; particularly, time spent in system design and detailed design, failure probability during systems testing and integration for both OEM and supplier, and finally marketing window and revenues. Model analysis and simulation procedures are in Section 4. Section 5 contains sensitivity analyses results and their discussion. Finally, we present our conclusions in Section 6.
2. Literature Review 2.1. Outsourcing Literature Outsourcing decisions are critical due to their impact on the PD process (Ulrich and Ellison, 2005; Staudenmayer et al., 2005). For instance, Anderson and Parker (2002) presented a model that accounted for learning effect in make/buy decisions. However, their model captured the effect of learning for an infinite time horizon, which is generally not the case in most outsourcing scenarios. Firms switch their policy whenever they experience that outsourcing is not beneficial. Moreover, they assumed that outsourcing decisions are made once during this infinite time. It would be more realistic to study outsourcing for one generation of a product by including the effect of this decision on future capability if the OEM decides to quit outsourcing and make the component in-house. Ülkü et al. (2005) examined how the process adoption is impacted by the make/buy decision. They argued that outsourcing does not grantee faster time-to market, which was explained by insufficient client base at the supplier. This result needs more explanation and study involving the various factors that could delay time-to-market like quality assurance. Ülkü and Schmidt (2005) found that firm capabilities play an important role in product architecture. They studied the interrelation between the degree of modularity and outsourcing the product design. Moreover, the model mainly consisted of three factors: component performance, total product performance, and market characteristic. However, the model was deterministic and did not include time allocations at the various development phases. Another study done by Mikkola (2003) linked modularity of the product to outsourcing and firm’s learning. She argued that outsourcing creates certain degree of inter-dependence. However, most of these insights regarding outsourcing were made through case observations and did not include other factors that link the interrelation between the OEM and the supplier. Along the same stream, Argyers (1996) found that OEM and supplier capabilities have a role in the outsourcing decision. This study was conducted based on observing various product development processes. Novak and Eppinger (2001) developed a statistical model to study the
The Impact of Engineering Design on Outsourcing Decisions
119
relation between product complexity and sourcing. Their model showed evidence of complementarity between product complexity and vertical integration. Becker and Zirpoli (2004) presented insights that linked various product development aspects (outsourcing, performance integration and product architecture). Their insights were based on an extensive ten-year data from European automotive industries. Moreover, they argued that outsourcing influences product performance.
2.2. Product Development Literature Several studies encourage speed to market in order to achieve considerable amount of market share (Lilien and Yoon, 1990; Meyer and Utterback, 1995; Smith, 1999; Langerak and Hultink, 2006). However, this may lead to an immature new product (i.e. with lower than optimal performance or quality) which reduces product’s demand and overall firm profits (Cohen et al., 1996; Bayus, 1997; Karlsson and Ahlstrom, 1999). Bayus et al. (1997) addressed the question of when should a firm introduce a new product and what should its performance level be? They assumed that every firm in the market has the same capability and discussed the time to introduce the new product to market relative to market’s leader introduction time. Bayus (1997) continued the flow by addressing normative research by relating various market, demand, and cost conditions to maximize the profit. Nevertheless, the model did not include characteristics of the product regarding its complexity. Cohen et al. (1996) generated a mathematical model of time-to-market decision given a fixed performance level. Yet, the model was limited to constant market demand and price. Along similar lines, Calantone and Di Benedetto (2000) addressed the same tradeoff (i.e. between product performance and time-to-market) when product development phases are overlapped and jointly working towards performance improvement. This model did not include other important factors that can affect performance, such as firm capability, product complexity, and it was deterministic. Other researchers studied PD in a stochastic manner as one major issue is the failure at testing and integration phase during PD. Morali and Soyer (2002) studied the optimal stopping in software testing. They developed a sequential decision model to find the optimal release time. Similarly, Eppinger et al. (1997) used a signal flow approach to analyze PD. The model computed PD lead-time for a probabilistic changing design environment.
3. Model Description and Formulation We consider a typical product development process which consists of four phases or stages: system design, detailed design, testing and integration, and marketing (Ulrich and Eppinger, 2004). The process is mainly performed sequentially as shown in Figure 1. The development team first spends time to define the product architecture (i.e. specifying subsystems and their interfaces). We assume that the more time spent on system design, the less chance the product fails during testing and integration. That is, more time spent on system design results in more modular product architecture (with clear and well defined subsystem boundaries), which in turn leads to less chance of failing testing and integration.
120
Mahmood Al-Kindi and Ali A. Yassine Detailed design
System design
in-house Market
Testing & Integration Fail
Pass
Fail
Supplie
Product life cycle
Figure 1. Product development process major phases.
Once the product architecture and the various subsystems are identified, the OEM (also referred to as the ‘firm’) is ready to make its outsourcing decision. For any outsourcing amount, both the firm and the supplier will work concurrently on their assigned tasks consuming a nominal time which is predicated by their associated expertise. At this stage (and after performing detailed design for the first time, which consumes the nominal allocated duration) the detailed design teams (i.e., OEM and supplier) believe that they have achieved 100% quality (based on the pre-specified design requirements or specifications). However, during testing and integration, a fraction of this quality will be confirmed and hence some of the previously performed work needs to be reworked (this also called ‘design iteration’). We assume that the supplier is more likely to fail at this stage due to the lack of intense communication with OEM (crossing organizational boundaries) and lack of complete information about the product or perfect specifications. In addition, we assume that design iteration takes place during detailed design phase only and do not require system redesign. The number of design iterations is related to the failure probability, which is, in turn, a function of the time spent on system design. The OEM and supplier must rework whenever they fail during testing and integration in order to achieve the required specifications and product performance before the product is launched to the market. This is a work policy assumption, which we made in our model. However, other work policies could be implemented in an extension to this model, such as working only up to a predefined maximum number of design iterations. Finally, for simplicity we normalize the product life cycle (i.e., the sum of system design time, detailed design time, and marketing window) to 1. We also assume that there exists a fixed and known marketing window beyond which there is no significant demand for the product under development. Finally, it is worth noting that excluding the production or manufacturing stage from our hypothetical model, as shown in Figure 1, does not detract from the value of the model for two reasons. First, one can assume that this model reflects a software development process,
The Impact of Engineering Design on Outsourcing Decisions
121
where production (i.e. making copies of the developed software) is a relatively minor issue compared to development. Second, even for other types of products, we assume that the production stage is not impacted by any of the upstream development decisions and does not impact any downstream marketing decisions (assuming that any quality issues during production are not a result of design decisions, but rather are an intrinsic property of the existing production system). Based on the above model description and assumptions, we, next, formulate the various mathematical constructs required for capturing the trade-off between product quality and the narrowing marketing window.
3.1. System Design Stage - Product complexity The product development process starts with system design where the team spends Ts time units to generate the product architecture, define major subsystems and interfaces, and set target costs and specifications. We assume that the success of the product during testing and integration phase is solely determined by the amount of time spent on system design. Moreover, the supplier will generally have lower probability of integration success; however, longer times spent on system design will reduce the gap between the probabilities of success of the supplier and the OEM. Therefore, the success probability during testing and integration (P) is expressed as:
POEM (Ts ) = Tsγ PSupplier (Ts ) = Tsγ +ν
(1)
The success probability is assumed monotonically increasing with respect to Ts. The shape parameters γ and ν determine how fast the success probability reaches 100% and the gap between the success probabilities, respectively, as shown in Figure 2. The probability of success gap, ν, can be seen as lack of information, communication and translation (e.g. language barrier or measurements system), or even perfect specification. Moreover, the supplier does not completely know the product architecture and its various subsystems (i.e., does not have the systems perspective than an OEM may have). Figure 2 illustrates that the supplier needs to spend more time to get to the same success probability as the OEM. Equation (1) indicates that as Ts increases, the probability of success (P) raises rapidly with low values of γ and slowly with higher values of γ. The value of the shape parameter γ is determined by the complexity (or degree of newness or innovativeness) of the product from the firm’s perspective (Johannessen et al., 2001). When the product is relatively complex, the firm needs to spend more time to develop the product architecture (i.e. decompose the system into modules / subsystems and define various interfaces) (Bashir and Thomson, 1999). This can be seen when the product is totally new to the firm (compared to a derivative product), firms struggle more at early stages and leads to frequent failures during testing and integration (Sethi, 2000).
122
Mahmood Al-Kindi and Ali A. Yassine
1.1 OEM Supplier
1 0.9
Pass probability
0.8 0.7 0.6 0.5 0.4 0.3 0.2 0.1
0
0.2
0.4
0.6
0.8
1
1.2
1.4
time
Figure 2. Success probabilities for OEM and supplier versus time in system design Ts.
3.2. Detailed Design Stage - Task Size and Firm Capability During this phase, a complete specification of the geometry, tolerance, and material for all parts and subassemblies of the product is performed. As shown in Figure 1, the OEM decides prior to this stage what to outsource to the supplier. The OEM should outsource the right fraction ε to maximize its marketing window. The time spent on this stage is generally effected by two factors: task size and firm capability. Intuitively task size increases the duration of detailed design. However, firm capability (e.g., knowledge or expertise) reduces 1 the duration for any task size exponentially. Thus, the nominal duration of detailed design (i.e. the estimated time to be spent on detailed design, without any rework or design iteration considered) is:
Tdo = λ μ
(2)
Figure 3 illustrates that given any task size, then different firms with different could have different completion times depending on their capabilities. Note that the task size ranges between 0 and 1. Therefore, any task size which is closer to 0 is assumed to be very small and can be finished instantaneously. On the other hand, a task with a size close to 1 is considered to be very large and consumes the whole product life cycle in order to finish. For example, three different firms with capabilities µ=1, µ=2, and µ=3 performing a task of size λ=0.4, will complete this task in 0.4, 0.16, and 0.064 time units, respectively (see Figure 3). Also, note that a firm’s capability ranges between 0 (no capability) to ∞ (extreme capability). Thus, varying task size λ and firm capability µ allows us to examine different detailed design conditions. 1
This relation is supported by standard learning curve theory (Wright 1936).
The Impact of Engineering Design on Outsourcing Decisions
123
1 µ=1 µ=2 µ=3
0.9
Actual Completion Time
0.8 0.7 0.6 0.5 0.4 0.3 0.2 0.1 0
0
0.1
0.2
0.3
0.4 0.5 0.6 Task Size (unit-time)
0.7
0.8
0.9
1
Figure 3. Nominal detailed design duration with firm capability and task size.
3.3. OEM and Supplier Detailed Design Duration The OEM and supplier work concurrently to finish their assigned detailed design work. The outsourcing fraction reduces the task size for the OEM to (1-ε) λ, where the supplier will have a task size of ελ. But the outsourcing fraction will decrease the OEM long-term development capability due to loss of knowledge and development know-how regarding the outsourced modules. Thus to model this long-term effect (of outsourcing) on current OEM capability, we perform a net present value calculation on the current capability for Gth generations of the product. It is reasonable to assume that current OEM capability will be obsolete after the Gth generation of the product. Therefore, the modified OEM capability is estimated by:
μ=
μ ⎞ 1⎛ G ⎜∑ ⎟ G ⎝ n =1 (1 + ε ) n ⎠
(3)
In Figure 4, suppose that the OEM current capability µ=1 and the OEM outsources 50%, then the modified OEM capability is 0.2, 0.35, and 0.5 for G=10, G=5, and G=3, respectively. It is obvious that the OEM will be affected worst when G is higher because the need for knowledge is more crucial when the OEM has protracted product generations ahead; hence, the long term effect of outsourcing will be more noticeable. Although firm capability may also increase from one generation to the next due to learning-by-doing, we present here the worst case scenario where knowledge loss due to outsourcing is more prominent. Furthermore, we assume that the supplier is initially as
124
Mahmood Al-Kindi and Ali A. Yassine
capable as the OEM. Finally, the nominal detailed design duration for both OEM and supplier becomes as follows:
TdoOEM = ( (1 − ε )λ ) o d Supplier
T
= ( ε .λ )
μ
(4)
μ
G=10
Capability Level
1 0.9
G=5
0.8
G=3
0.7 0.6 0.5 0.4 0.3 0.2 0.1 0 0
0.2
0.4
0.6
0.8
1
Outsourcing Fraction
Figure 4. Outsourcing fraction effect on OEM’s capability for different product generations.
3.4. Testing and Integration Stage – Design Iteration or Rework Once the OEM and supplier complete their assigned detailed design work, the outcome is checked during the testing and integration phase, where there is a chance that the work of the OEM, supplier, or both does not pass satisfactorily and requires modifications. The probability of success (or failure) is directly related to the time spent in system design as previously discussed. However, when failure occurs and some rework is necessary, the fraction of rework (α) depends on the nature of the detailed design (e.g. a CAD model is easier to rework than a clay model). The OEM work policy is that the product must achieve the required performance or quality prior to marketing. That is, the OEM and supplier must pass this phase in order for the product to be introduced into the market. The duration of this process is simply the maximum time spent on detailed design (including design iteration) by either the OEM or supplier. Each time the design fails, only a fraction of the prior work is approved and therefore both the OEM and supplier need to work on the remaining design issues. Although the probability of failure is constant, the amount of rework is reduced in each subsequent design iteration. Note that when the OEM or supplier fails the first time, they will spend a rework time of α Tod . However, the second time they fail, they need to spend
The Impact of Engineering Design on Outsourcing Decisions
125
only α2 Tod and so on. Thus, the expected time that detailed design stage consumes, considering all rework to be performed (i.e. design iterations), is expressed as:2
E (Tdo OEM ) =
Tdo OEM 1 − α (1 − POEM )
E (Tdo Supplier ) =
Tdo Supplier
(5)
1 − α (1 − PSupplier )
Since both the OEM and the Supplier are working concurrently, the total time consumed in detailed design stage is simply max{E (TdoOEM ), E (TdoSupplier )} .
3.5. Marketing Window Tm Time-to-market plays a major role in OEM’s revenues. Having a large marketing window will increase the total sales volume of the product simply because the OEM will have sufficient amount of time to sell the product. Our objective is to maximize OEM’s revenues 3 by maximizing the marketing window, Tm. Thus the overall model can be summarized as a well bounded non-linear convex optimization problem as follows:
f (Tm , Ts , ε ) = max Tm * * Ts ,ε
s.t
(6)
μ ⎡ ( (1 − ε )λ ) μ ⎤ ε .λ ) ( ⎥ = TTotal Tm + Ts + max ⎢ , γ γ +ν ⎢1 − α (1 − Ts ) 1 − α (1 − Ts ) ⎥ ⎣ ⎦
4. Analysis It is not easy to find an optimal solution analytically since it requires trial and error iterations. However, it is easier to get an analytical solution to the optimization problem in (6) for the binary outsourcing decision (i.e. ε =0 or ε =1) as described in Section 4.1; otherwise, using optimization software becomes necessary to solve Equation (6), as described in Section 4.2.
E (Tdo OEM ) = Tdo OEM .POEM + Tdo OEM (1 + α ) POEM (1 − POEM ) 2
+Tdo OEM (1 + α + α 2 ) POEM (1 − POEM ) 2 + .. E (Tdo Supplier ) = Tdo Supplier .PSupplier + Tdo Supplier (1 + α ) PSupplier (1 − PSupplier ) +Tdo Supplier (1 + α + α 2 ) PSupplier (1 − PSupplier ) 2 + ..
3
Generally, revenues have many key inputs such as product demand, price, and marketing window. Assuming constant demand and price, then the marketing window becomes a good proxy for revenues.
126
Mahmood Al-Kindi and Ali A. Yassine
4.1. Binary Outsourcing Decision (Complete in-House or Complete Outsourcing) In some cases, the OEM makes the decision to fully outsource ε=1 or make every thing in-house ε=0. The only constraint that the problem has is time limitation. Assuming that each time the product fails at testing and integration then the detailed design team will rework the whole task (i.e. α=1). By substituting the equality constraint into the objective function then the objective function can be rewritten as:
f (Ts ) = Ttotal − Ts −
λμ Ts γ
(7)
From which the optimal time to be spent in system design given that OEM will perform the complete in-house:
(
Ts* = γ .λ μ
)
1 1+γ
(8)
When relaxing the assumption of α=1(worst case scenario), then the optimal time to be spent in system design is expressed as:
(
T ≤ γ .λ * s
)
1
μ 1+γ
(9)
Note that the optimal time spent in system design is an increasing function in product architecture complexity and task size. However, it is a decreasing function in firm capapbility. However, if the OEM decided to perform complete outsourcing then the optimal time spent in system design is shown in Equation (10), which indicates that when both OEM and the supplier has the same capability then it is optimal to spend more time in system design when the OEM has to outsource 100%.
(
Ts* ≤ (γ + ν ).λ μ
)
1 1+γ +ν
(10)
4.2. Partial Outsourcing ( 0< ε <1) In this kind of outsourcing the OEM must outsource the right amount in order to maximize the marketing window. Our objective is study to the influence of the various model parameters on the optimal time in system design and outsourcing amount. It is not easy to get an analytical closed-form solution to the optimization problem in Equation (6); therefore, we revert to using optimization software (e.g., MatLab) instead to get the optimal time spent on system design and outsourcing fraction given various PD scenarios. Our goal is to get
The Impact of Engineering Design on Outsourcing Decisions
127
managerial insights regarding the optimal strategy which maximizes the marketing window given different PD scenarios. Performing an extensive study on various model inputs within their reasonable ranges gives a clear picture for their impact on these two decision variables (Ts and ε). To save computational time, we construct reasonable ranges for each input parameter as shown in Table 1. For instance, the information gap at which the supplier has a higher probability of failure will be around 0-0.3. Furthermore, the task size will not probably be more than 5060% of total project time Ttotal . Therefore, λ will be within the range of (0.05-0.8). Table 1. Ranges for input parameters Input parameter Firm capability (µ) Task size (λ) information gap (ν) Product complexity (γ) Nature of detailed design work (α)
Reasonable Range 1-6 0.05-0.8 0-0.30 0-0.5 0-1.0
Following these various inputs (shown in Table 1), optimal solutions for time in system design and outsourcing fraction can be obtained using MatLab. Since there are five inputs (firm capability, task size, information gap, product architecture complexity, and nature of detailed design work) with six levels each (Low1-2-3-4-5-6High), then the total optimal solutions obtained are for 56=15625 PD scenarios.
0.16
Time In System Design
0.14 0.12 0.10 0.08 0.06 0.04 0.02 0.0
0.2
0.4 0.6 %Outsourcing
0.8
Figure 5. Time spent in system design versus outsourcing fraction.
1.0
128
Mahmood Al-Kindi and Ali A. Yassine
To see how the outsourcing fraction relates to the time spent in system design and vice versa, we solved all 56 data points for Ts* and ε*. Then, taking the average of Ts* at various ranges for ε* (e.g. ε* [0-0.1] the average Ts * = 0.03). The result is shown in Figure 10. The figure shows an increasing relationship between the time spent in system design and the outsourcing fraction. Therefore, spending more time in system design will lead to higher amount of outsourcing or if the OEM outsource more than it is preferred to spend more time in system design.
5. Sensitivity Analysis To perform the sensitivity analyses, we combined all possible cases within the reasonable range shown in Table 1. Initially, we started with a current capability µ=1, a task size λ = 0.05, a product complexity γ=0, an information gap ν = 0, and a rework fraction α=0. Once all five input parameters are declared, then an optimal time in system design, Ts*, and outsourcing fraction, ε*, are computed and stored. Then, the rework fraction, α, is incremented (by a small step size) while keeping all other four inputs at their starting values. Once the rework fraction reaches its maximum value (i.e. α=1), then a new value for the information gap (ν=0+step-size) is introduced and the rework fraction is back again to its low level (α=0). Similar update is made for product complexity, γ, when the information gap reaches its maximum value (ν = 1). Once product complexity reaches its maximum level, then a new update is made for the task size. Firm capability, µ, is incremented, when the task size reaches its maximum value of 1, until it reaches its upper limit µmax. Finally, we performed extensive explorations of the stored optimal solutions in order to get the full picture behind the optimal strategy in terms of design time allocations and outsourcing for the various PD scenarios (i.e. input parameters).
5.1. Firm Capability µ In this section we consider the average for all optimal solutions (time in system design Ts and ε* optimal outsourcing fraction) at each level of firm capability µ. The result is shown in Figure 6. The figure shows that as OEM capability decreases, the amount of outsourcing increases. Obviously, to maximize the marketing window, lower OEM capability forces the OEM to spend more time in system design in order to minimize the integration failure probability. Therefore, it is better to split the work and get the benefit of the supplier’s superior capability. On the other hand, having a high OEM capability makes outsourcing less attractive since the OEM can do the job relatively fast and outsourcing will only force the OEM to spend more time in system design (in order to minimize the integration failure probability). Consequently, it is optimal to spend less time in system design as firm capability increases. *
The Impact of Engineering Design on Outsourcing Decisions
129
6 4 Firm Capability
2
0.50
0.75
0.6 0.4
Time in System Design 1.00
0.2
%Outsourcing
Figure 6. The effect of OEM capability on outsourcing fraction and time in system design.
5.2. Task Size λ Using similar analysis (i.e. taking the average of Ts* and ε* at each level of task size λ), the result is summarized in Figure 7. It can be seen that as the task size increases the OEM spends more time in system design which means that it is optimal not to gamble with large
0.5 %Outsourcing 0.4 0.3 1.05 0.90
0.4 0.75
0.6 Task Size
0.8
0.60
Time in System Design
Figure 7. The effect of OEM capability on outsourcing fraction and time in system design.
130
Mahmood Al-Kindi and Ali A. Yassine
task sizes and a well defined product architecture is recommended. However, the amount of outscoring is higher at medium task sizes and lower at the edges. For large task sizes, it is optimal to outsource less (in this case <30% ), because, it is risky to outsource a high fraction due to the higher chance of failure that the supplier has and the possibility of taking more time to achieve the required performance. On the other hand, when the task size is small, outsourcing forces the OEM to spend more time in system design since detailed design does not consume much time for small task sizes.
5.3. Product Complexity γ Another important parameter that has an impact on the outsourcing and time allocation decisions is product architecture complexity. In Figure 8, as the complexity of the product increases, then it is essential to spend more time in system design. Moreover, the outsourcing fraction increase as the product becomes more complex. That is, for more complex products, the OEM will outsource more and spend more time in system design. It seems that splitting the task size will minimize the maximum expected time of OEM and the supplier. Therefore, achieving higher success probability for larger task sizes takes more time than having two small tasks with higher chance of integration failure.
0.40 %Outsourcing
0.35 1.0 0.30
0.8 0.6 0.1
0.2
0.4 0.3
Time in System Design
0.4
Product Complexity
Figure 8. The effect of OEM capability on outsourcing fraction and time in system design.
5.4. Fraction of Rework (Detailed Design Nature) α This parameter appears during the testing and integration phase as it merely dictates how much the OEM and Supplier should rework each time they fail integration and testing. Figure 9 indicates that when the fraction of rework is high, then the outsourcing fraction is low and
The Impact of Engineering Design on Outsourcing Decisions
131
the time spent in system design is high. However, at low rework fractions the outsourcing fraction is increased and the time in system design is shorter.
1.00 Dtld Design Nature 0.75
0.50
0.2 1.0
0.4
0.8
%Outscourcing
0.6 0.4
Time in System Design
0.6
Figure 9. Effect of rework fraction on outsourcing fraction and time in system design.
5.5. Supplier Total Information Gap υ This parameter is responsible for the higher failure probability (during testing and integration) of the supplier compared to the OEM. It can be visualized as misscommunications or miss-translation between OEM and the supplier. As shown in Figure 10, as this gap increases, then it is better to spend more time in system design and outsource more. The reason is that since the OEM spends considerable time in system design, then it will end up with a well defined product architecture, where the OEM and supplier may have similar probability of success.
0.348 0.342 %Outsourcing
0.336 0.066
0.330
0.064 0.062 0.1
0.060 0.2
System Design
0.3
Supplier Info. Gap
Figure 10. Effect of information gap on outsourcing fraction and time in system design.
132
Mahmood Al-Kindi and Ali A. Yassine
6. Summary and Conclusion Literature suggests that outsourcing decisions are affected by product architecture and modularity, task size, firm capability, and marketing window. Therefore, in this paper, a mathematical model is developed to link all these factors and their effect on the outsourcing decision. We consider product modularity as a feature that can be obtained by spending more time in system design. This mathematical model is described as a well-bounded optimization problem where its objective is to maximize the marketing window subjected to time limitation (project total time). Different product development scenarios are examined by varying various input parameters in order to obtain the optimal strategy regarding time in system design and outsourcing fraction. A closed form solution is obtained for the binary outsourcing decision. However, for the partial outsourcing case, we use optimization software to get the optimal solution. Sensitively analyses were performed to get an overview of the impact of these inputs on the optimal solution. The optimal solution reveals several interesting managerial insights regarding the impact of the various engineering design decisions on the outsourcing decision. First, spending more time in system design leads to higher outsourcing fraction and vice versa; that is, well defined product architectures lead to higher outsourcing. Second, higher firm capability makes outsourcing less attractive. Third, outsourcing is found to be more attractive at the medium task sizes compared to larger or smaller tasks. Fourth, a product with a complex architecture will lead the firm to spend more time in system design and thus outsource more. Lastly, the nature of the detailed design work, which determines the rework fraction, has an impact of both system time and outsourcing decisions; namely, as the rework fraction of detailed design increases, it is better to spend more time in system design and outsource more.
References Anderson, E., and Parker, G., “The Effect of Learning on the Make/Buy decision,” Production and Operation Management, 11(3), 2002, pp. 313-339. Argyres, N. “ Evidence on the Role of Firm Capabilities in Vertical Integration Decisions,” Strategic Management Journal, Vol 17, pp. 129-150. Bashir, H.A. & Thomson, V. (1999) Metrics for design projects: a review, Design Studies, 20, pp. 263- 277. Bayus, B. “Speed to Market and New Product Performance Trade-offs,” Journal of Product Innovation Management, No.14, 1997, pp. 485-497. Bayus, B., Jain, S., and Rao, A., “Too Little, Too Early: Introduction and New Product Performance in the Personal Digital Assistant Industry,” Journal of Marketing Research, 34(1), 1997, pp. 50-63. Becker, M., and Zirpoli, F., “Organizing New Product Development: Knowledge HollowingOut and Knowledge Integration- The FIAT Auto Case,” International Journal of Operations & Production Management, 23(9), 2003, pp.1033-1061. Becker, M., and Zirpoli, F., “Organizing Complex Product Development: Outsourcing, performance Integration and the Role of Product Architecture,”
The Impact of Engineering Design on Outsourcing Decisions
133
Calantone, R., and Anthony, D. “Performance and Time-To-Market: Accelerating Cycle Time with Overlapping Stages,” IEEE Transactions on Engineering Management, 47(2), 2000, pp. 232-244. Calantone, R., Di Benedetto, C.A. “Performance and Time to market: Accelerating Cycle Time with Overlapping Stages,” IEEE Transactions on Engineering Management, 47(2), 2000, pp. 232-244. Cohen, M., Eliashberg, J., Ho, T., “New Product Development: The Performance and Timeto-Market Tradeoff,” Management Science, 42(2), 1996, pp. 173-186. Eppinger, S, Nukala, M., Whitney, D., “Generalized Models of Design Iteration Using Signal Flow Graphs,” Research in Engineering Design, 9(2), 1997, pp. 1122-123. Henserson, R., Clark, K., “Architectural Innovation: The Reconfiguration of Existing Product technologies and the failure of Established Firms,” Administrative Science Quarterly, 35(1), mar. 1990, pp. 9-30. Hsuan, J., “Modularization in Black Box Design: Implication for Supplier-Buyer Partnership,” Paper Prepared for DRUIC Winter Conference, Holte, Denmark, 1997. von Hipple, E. , “Task Partitioning: An innovation process available” Research Policy 19, 1990, pp.407-418. Joglekar, N., Yassine, A., Eppinger, S., Whitney, D., “Performance of Coupled Product Development Activities with a Deadline,” Management Science, 47(12), 2001, pp. 16051620. Johannessen, J-A., Olsen, B., Lumpkin, G.T., “Innovation as Newness: What is new, how new, and new to whom?” European Journal of Innovation Management, 4(1) 2001, pp. 20-31. Karlsson, C., Ahlstrom, P., “Technological level and Product Development Cycle Time,” Journal of Product Innovation Management 16, 1999, pp.352–362. Langerak, F., Hultink, E.J., “The Impact of Product Innovativeness on the Link between Development Speed and New Product Profitability,” Journal of Lilien, Gary L. and Yoon, E. (1990). The Timing of Competitive Entry: An Exploratory Study of New Industrial Products. Management Science 36, May 1990, pp. 568–85. McFadden, D., "Econometric Models for Probabilistic Choice Among Products," Journal of Business, Vol. 53, 1980, pp 513-530. Meyer, M., Utterback, J., “Product Development Cycle Time and Commercial Success,” IEEE Transactions on Engineering Management, 42(4), 1995, pp. 297-304. Mikkola, J. “ Modularity, Component Outsourcing, and Inter-Firm Learning,” R&D Management Vol.33 (4), 2003. Morali, N. and Soyer, R. “Optimal Stopping in Software Testing,” Naval Research Logistics, Vol. 50, 2003. Novak, S. and Eppinger, S. “ Sourcing by Design: Product Complexity and the Supply Chain,” Management Science, Vol.47 (1), 2001. Sethi, R., “New product Quality and Product Development Teams,” Journal of Marketing, Vol. 64, Apr. 2000, pp. 1-14. Smith, P.G., “From Experience: Reaping Benefits from Speed to Market,” Journal of Product Innovation Management 16(3), 1999, pp.222–30. Staudenmayer, N., Tripsas, M., Tucci, C., “Interfirm Modularity and its Implications for Product Development,” Journal of Product Innovation Management 22, 2005, pp.303– 321.
134
Mahmood Al-Kindi and Ali A. Yassine
Tatikonda, M.V., Rosenthal, S.R, “Technology novelty, project complexity, and product development project execution success: a deeper look at task uncertainty in product innovation,” IEEE Transactions on Engineering Management, 47(1), 2000, pp. 74-87. Ülkü, S. and Schmidt, G., “Matching Product Architecture and Supply Chain Design,” MSOM Conference 2005, Northwestern University, Kellogg School of Management, Evanston, Illinois. June 27-28, 2005. Ülkü, S., Tokaty, L.B., Yucesan, E., “The Impact of Outsourcing and Manufacturing on Timing of Entry in Uncertain Markets,” Production and Operation Management, 14(3), 2005, pp. 301-314. Ulrich, K., Eppinger, S., Product Design and Development, 3rd Edition, McGraw-Hill, Inc., New York, 2004. Ulrich, K., Ellison, D., “Beyond Make-Buy: Internalization and Integration of Design and Production,” Production and Operation Management, 14(3), 2005, pp. 315-330. Yassine, A., Joglekar, N., Braha, D., Eppinger, S., Whitney, D., “Information Hiding in Product Development: The Design Churn Effect,” Research in Engineering Design, Volume 14, No. 3, 2003. Yassine, A., Sreenivas, R., Zhu, J., “Managing the Exchange of Information in Product Development,” European Journal of Operational Research 184 (2008) pp. 311-326.
In: Outsourcing, Teamwork and Business Management ISBN: 978-1-60456-956-8 Editor: Karl E. Carettas, pp. 135-166 © 2009 Nova Science Publishers, Inc.
Chapter 9
TIME ALLOCATION AND OUTSOURCING WITHIN HOUSEHOLDS: DIFFERENCES IN LIFESTYLE BETWEEN NATIVE DUTCH AND IMMIGRANTS IN THE NETHERLANDS J.R. Cornelisse-Vermaat1,*, H. Maassen van den Brink2, J.A.C. van Ophem3 and G. Antonides4 1
Department of Marketing and Consumer Behaviour, Wageningen University Hollandseweg 1, 6706 KN Wageningen, The Netherlands 2 Department of General Economics, University of Amsterdam; Roetersstraat 11, 1018 WB Amsterdam, The Netherlands. 3 Economics of Consumers and Households. Wageningen University Hollandseweg 1, 6706 KN Wageningen, The Netherlands. 4 Wageningen University Hollandseweg 1, 6706 KN Wageningen, The Netherlands
Abstract Due to the increased female labour participation in the past decades, households lack time to perform all households and care activities. At present in the Netherlands, households can outsource home cleaning to a cleaning lady/man, cooking to restaurants (or people can eat ready meals or takeaway food), and childcare can be outsourced to day care centres. The increased household income, attributed to higher female labour participation, gives more possibilities to outsource domestic work. Outsourcing could not only be determined by socioeconomic and demographic variables, culture (or ethnicity) can also be of importance in explaining outsourcing within households. This chapter aims to determine the time households spend on domestic tasks and care activities and whether differences in lifestyle and ethnicity are related to outsourcing behaviour. Time spent on household and care activities is estimated with a model including socioeconomic and demographic variables and including some lifestyle determinants. Household expenditures on different types of outsourcing possibilities within households are measured and differences are drawn between native Dutch and non-western immigrants. For *
E-mail address:
[email protected]. Tel: +31-317-482437
136
J.R. Cornelisse-Vermaat, H. Maassen van den Brink, J.A.C. van Ophem et al. the analyses a sample (2001) is used that consists of Dutch, Turks, Surinamese/Antilleans, and Moroccans (N=2551). The analyses show that immigrants spend less time on household and care activities compared to native Dutch. Both Household income and level of education are determinants of the expenditures on outsourcing of domestic tasks and care activities. Native Dutch and Surinamese/Antilleans have comparable expenditures on home cleaning and childcare, whereas Moroccans and Turks spend more on takeaway food and delivery food. The results reveal differences as well as similarities in lifestyle between native Dutch and non-western immigrants.
Keywords: time allocation, outsourcing, food habits, lifestyle, ethnicity, households
1. Introduction Due to the increased female labour participation of the past decades, households have lack of time to perform all household and care activities. There are three strategies to solve this problem: 1) outsourcing household and care activities; 2) substitution of household and care tasks by domestic appliances; and 3) time arrangement (adjusting working hours or shop opening hours) (SCP, 2000a; Van Ophem and De Hoog, 1995; and Van Dam et al., 1994). In this chapter, we will focus on the outsourcing of household and care activities and its determinants. Outsourcing is defined as an arrangement for a particular service outside the household (either private or subsidised) to take care of household activities. Since the end of the 1990s, household and care activities have been outsourced more frequently (RIVM, 2004; SCP, 2000a; and Tijdens et al., 2000). Households can decide to outsource household and care activities more and more because they have more financial budget, and because of the increasing female labour participation (implying that more women combine labour and care). Over the years, the number of outsourcing possibilities has grown, which gives households more options to outsource their household and care tasks. Still, until now, not much research has been performed on the determinants of outsourcing and the relation between the determinants of household and care time. One could assume that the time-saving aspect is an important reason for outsourcing household and care activities. At present in the Netherlands, households can outsource home cleaning to a cleaning lady/man, cooking to restaurants (or people can eat ready-to-eat-meals, takeaway food or delivery food), and childcare to day-care centres. The increased household income, attributed to higher female labour participation, gives more possibilities to outsource domestic work to others. 82 percent of the Dutch double-income households eat takeaway food more than once per month, against 62 percent of the single-income households. About half of the Dutch households with children living at home make use of childcare. Between 1995 and 2000, the expenses on childcare have more than doubled (SCP, 2000a). Until now, not much is known about the outsourcing behaviour of immigrants in the Netherlands and the same holds for the differences in outsourcing between immigrants and the native Dutch. Earlier research shows that immigrants in the Netherlands make less use of formal childcare than the native Dutch (NIBUD, 2004). Culture involves many health-related notions amongst which are nutrition and lifestyle (Uniken Venema et al., 1995). Outsourcing behaviour could not only be determined by socio-economic and demographic variables,
Time Allocation and Outsourcing within Households
137
culture (or ethnicity) can also be of importance in explaining outsourcing within households. Immigrants could choose different (cheaper) outsourcing possibilities than the native Dutch. People make different choices (less or more healthy) with respect to outsourcing food preparation. These choices may be culture-driven (Cornelisse and Maassen van den Brink, 2007). In this chapter we will study whether native Dutch make different choices in outsourcing not only their meal preparation, but also their childcare and home cleaning compared to immigrants in the Netherlands and how that is related to the time allocation within their households. This chapter aims: 1) to study the determinants of the demand for household and care time and the demand for outsourcing household and care activities; and 2) to investigate whether immigrants differ from the native Dutch in their time allocation and outsourcing behaviour. The chapter is structured as follows. Section 2 provides an overview of the theory and empirical results on time allocation and household production. Section 3 describes the data on time allocation and outsourcing behaviour among the native Dutch and the studied immigrant groups. Section 4 gives the empirical model for the demand for household and care time and the demand for outsourcing. In section 5, the estimations on time allocated to household and care tasks and money spent on outsourcing household and care tasks are given. In section 6, we conclude and discuss some issues of this chapter.
2. Time Allocation, Outsourcing, and Household Production. A Review of the Literature 2.1. Time Allocation Research and Time Allocation in the Netherlands Becker (1965) gives an important theoretical argument for outsourcing. Households can be considered as small production companies that try to maximise their output, restricted by their time and budget. If time spent on the labour market is more valuable than time spent on home production, it could be profitable to outsource home production (of course, this also depends on the price of outsourcing). It is therefore more likely for working wives to outsource household activities, since that will place a higher marginal value on their household production time than non-working wives (Kim, 1989). Surveys on time allocation show that unpaid household labour is divided on the basis of gender, which was also the case at the end of the 19th century (Mokyr, 2000). Males spend the most time on market labour, while females do the most household labour. On average, women spend twice as much time on household labour than men do (Bittman et al., 1998). More than 80 percent of women’s time is spent on non-market activities. In general, time men spend on household activities is about the same in all countries, except Norway, Sweden, and Japan. Japanese men do not spend any time at all on childcare and household activities, while Norwegian and Swedish men spend relatively more time on these activities within their households (Aronsson et al., 2001; Strober and Miling Kaneko Chan, 1998; Yamada et al., 1999; Maassen van den Brink and Groot, 1997). In most developed countries, working wives perform about two-third of the total household activities. In Spain, even in households where the spouses’ earnings and level of education are very
138
J.R. Cornelisse-Vermaat, H. Maassen van den Brink, J.A.C. van Ophem et al.
similar, their time spent on household activities differ significantly (Álvarez and Miles, 2003). During past decades, the time pattern in households is one of the ‘masculanisation’ of unpaid work: women’s time spent on unpaid work is increasingly coming down to meet men’s, rather than men’s unpaid work-time rising to meet women’s (Bittman et al., 1998). Although men’s household and care time has increased over the decades, women still spend more hours on these activities than men and women still spend less hours on labour market than men (Sousa-Poza, 2001). Women’s time allocated to household and care activities depends on several social, economic, and demographic factors, while men’s household and care time is largely invariant to these factors (Ciscel et al., 2000). Other research shows that a change in mother’s working hours has less influence on the parent’s time with their children than a change in the father’s working hours (Hersch and Stratton, 1994). Every five years, the Social and Cultural Planning Bureau investigates time allocation of the Dutch population. Over the last decades, the time allocation of the Dutch population has changed. Table 1 shows some results of the Time Budget Research (TBO) from 1980 to 2000. Table 1. Time allocation of women and men in The Netherlands between 1980 and 2000 (people aged ≥ 18) Hours per week
Labour Education (formal and informal) Household and care activities Sleeping, eating, personal care Leisure TOTAL
1980 6.8 1.8 35.8 78.6 46.0 168.0
women 1990 9.7 2.0 30.6 77.2 47.5 168.0
2000 12.9 2.3 28.2 79.8 45.0 168.0
1980 28.8 3.3 10.5 78.9 49.8 168.0
men 1990 25.4 3.6 12.3 77.7 49.0 168.0
2000 26.0 2.7 13.6 77.4 48.3 168.0
Source: SCP, 2004a
For women, time spent on market labour has doubled in two decades, while for men time spent on market labour has decreased by almost 3 hours per week. Because of the increase of labour time for women, one could expect a decrease in time spent on household and care activities for women, which is indeed the case. Over the years, time spent on household and care activities by men has increased by 3 hours per week, indicating that the division of time between males and females in the households has become more balanced. Still, in 2000, women spend more than twice as much time on domestic and care activities than men do. The total time spent on labour, household, and care activities are quite close to each other for both sexes (41.1 hours for females and 42.6 hours for males). Over the years, the total time households spend on household and care activities has decreased with 5 hours, which could have several causes: 1) the technologies used for household and care tasks are have improved; 2) people do the household and care tasks more efficiently; 3) more tasks are left undone; or 4) more activities are being outsourced, which relieves the ‘household burden’. We will come back to this issue later. Table 2 gives an overview of time allocation in the last two decades, according to households with or without children.
Time Allocation and Outsourcing within Households
139
Table 2. Time allocation of households with or without children living at home in the Netherlands between 1980 and 2000 (people aged ≥ 18) Hours per week
Labour Education (formal and informal) Household and care activities Sleeping, eating, personal care Leisure TOTAL
couples with children 1980 1990 2000 13.8 17.0 19.7 1.6 1.3 1.4 31.3 30.3 29.8 77.7 76.8 76.9 43.6 42.6 40.2 168.0 168.0 168.0
couples without children 1980 1990 2000 14.6 14.6 16.3 1.9 2.3 1.4 18.9 19.1 19.2 81.2 81.0 80.5 51.4 51.0 50.6 168.0 168.0 168.0
Source: SCP, 2004a
Over twenty years, for both households with and without children, the number of hours spent on market labour has increased and leisure has decreased. Households with children have more working hours than households without children, whereas households without children have more time for education (although, not in 2000). The time spent on household and care time in households with children is about a third more compared to the households without children. This increased household and care time is compensated with less time for sleeping etc. and leisure. Although little is known about the time allocation of immigrants in the Netherlands, there are data on labour participation of immigrants in the Netherlands that give an indication of the time division in these households. Surinamese/Antillean women are more often employed than Turkish and Moroccan women. Table 3 gives the labour participation and working hours of the different groups in the Netherlands. Table 3. Labour participation and working hours among the native Dutch and immigrants in the Netherlands in 2001 of people aged between 15 - 64 (in percentages) native Dutch women men Labour participation Working hours
Surinamese women men
Antilleans women men
Moroccans women men
Turks women men
55
80
59
66
48
62
26
56
33
62
19 46 36 100
2 10 88 100
10 39 51 100
3 12 85 100
16 46 38 100
3 20 77 100
12 37 51 100
2 14 83 100
18 43 40 100
3 10 87 100
Source: SCP, 2002
Table 3 shows that the Dutch men have the highest labour participation, whereas the Moroccans have the lowest labour participation for both males and females. This could imply that Turks and Moroccans have more time left for household and care activities than the native Dutch have. From the native Dutch women with paid work about two-third works parttime (defined as less than 34 hours per week), which is more than the women in immigrant groups where 40 to 50 percent works part-time. Among women, the labour participation of
140
J.R. Cornelisse-Vermaat, H. Maassen van den Brink, J.A.C. van Ophem et al.
the Surinamese is the highest, followed by the native Dutch and the Antilleans. The workinghours pattern of Moroccan and Turkish men and women is quite comparable. In comparison with other EU countries, the Netherlands has a relative high prevalence of part-time workers: 70 percent for females and 19 percent for males. In other EU countries these figures are about 35 percent for females and 5 percent for males (SCP, 2002). In the past decade in the Netherlands, female working hours have doubled and men also work more hours. The time spent on household activities has not changed much, while free time has decreased. This indicates that households are under some pressure. However, the increased household income gives opportunities to outsource household tasks.
2.2. Outsourcing of Household Tasks and Care Activities As discussed in the previous section, in the Netherlands the number of households with two working partners is increasing. This increase is mainly caused by an increase in labour supply by women with young children during the past decade (Maassen van den Brink and Groot, 1997). Between 1980 and 2005, the fraction of people combining paid labour with care increased from 16 percent to 38 percent for people between the ages of 20 and 64 (SCP, 2006a). The increased female labour participation is barely accompanied by a decrease in male labour participation. Earlier research shows that households with higher educated wives are more likely to make use of outsourcing opportunities than households with lower educated wives (Bellante and Forrester, 1984; Soberon-Ferres and Dardis, 1991). With respect to outsourcing household activities, both the household’s income and the wife’s income are important. Soberon-Ferrer and Dardis (1991) find that unearned income (non-labour income), wives’ wages, wives’ education, and being white are significant factors in outsourcing home cleaning. Spitze (1999) and Oropesa (1993) found that in the United States, higher-income households receive more paid household help than lower-income households. This is also true in Dutch households (Lambriex and Siegers, 1993). In the United States, the top-income households spend more than twelve times as much on housekeeping services and four times as much on food away from home than the lowest income group (Cohen, 1998). In the Netherlands, mainly two-earner households (largely one-and-a-half income households) outsource the care for children to paid childcare centres (47 percent), but also one-earner households make use of this service (20 percent). Except labour participation, there are lots of other reasons to send children to play groups. Parents could do that for pedagogical reasons, to let their children play with other children, or to have more leisure time for themselves. In the Netherlands, outsourcing the care for children is not only positively related to income, but to level of education (SCP, 2000a). Especially two-earner households use formal childcare like childcare centres (at the parent’s work), host family, or childcare outside of school. However, for all kinds of households, informal childcare (grandparents, brothers/sisters, or friends) is still the most popular way of outsourcing childcare. 54 percent of the two-earner households and 28 percent of the one-earner households use informal care. About 40 percent of the Dutch households with children living at home use childcare (SCP, 2006).
Time Allocation and Outsourcing within Households
141
Table 4 shows that immigrants make less use of formal childcare than the native Dutch do. The table shows the use of formal childcare for both one-earner and two-earner households with children living at home. The amounts are excluding subsidies. Table 4. Use of formal childcare by native Dutch and immigrant households in the Netherlands Frequency (%) the native Dutch Surinamese/Antilleans Turks Moroccans
36.2 29.6 18.1 16.2
Expenditures of users per month (€) 218 151 61 100
Source: NIBUD, 2004
Table 4 shows that the Surinamese/Antilleans are comparable to the native Dutch with respect to the use of formal childcare. It seems strange that the Turks make a little more use of formal childcare than Moroccans, but pay about 40 percent less than the Moroccans do. This can be attributed to the use of subsidised pre-school education used by the Turks (SCP, 2003). The time women spend on taking care of their children is quite stable, despite their increasing working hours. The care of children is still mainly a task for women, even when childcare services are used (Maassen van den Brink and Groot, 1997). Sousa-Poza et al. (2001) find that for wives, the presence of children, marital status, and hourly wage rate are important factors influencing time spent on household activities and childcare. Out-of-home childcare is not a substitute taking care of one’s own children. Parents use out-of-home childcare to combine work with the time for their children and cut down on private leisure and household activities (Hallberg and Klevmarken, 2003). Home cleaning is mainly outsourced by two-earner households, especially when they have children. Outsourcing home cleaning is also positively related to household income, and income and level of education of the female in the household (Tijdens et al., 2000; SCP, 2000a; Lambriex and Siegers, 1993). Between 1975 and 1995, the number of two-earner households outsourcing home cleaning more than doubled from 10 to 25 percent. The number of one-earner households outsourcing home cleaning has been quite stable in that time period (around 5 percent). Other help for domestic work (not specified) is also primarily used by two-earner households (29 percent) and is positively related with having children living at home (Tijdens et al., 2000). Tijdens et al. (2000) show that using a household help (mainly for cleaning, doing the laundry, and ironing) is negatively correlated to the time spent on household activities. The number of children and having children aged 0 – 5 are positively correlated to time spent on household activities. Table 5 shows that between 1992 and 2000, the expenditures on household services in the Netherlands have more than doubled.
142
J.R. Cornelisse-Vermaat, H. Maassen van den Brink, J.A.C. van Ophem et al. Table 5. Mean expenditures on paid household services 1992 – 2000 for all Dutch households (in Euros per year)
Childcare Expenditures general household help Window cleaner Launderette/ Dry cleaner’s TOTAL
1992 60 87 22 11 180
1995 87 102 25 11 225
2000 184 160 24 17 380
Source: CBS Statline, 2003, the figures are not corrected for inflation
In the Netherlands, between 1995 and 2000, the expenditures on childcare have doubled. The expenditures on wage for service personnel (like household help) have also increased between 1995 and 2000, but less than the expenditure on childcare. Compared to the other categories, the expenditures on window cleaning and clothing care have been rather constant. Data from the Social and Cultural Planning Bureau (SCP) show that, in the Netherlands, meal preparation is mainly outsourced by two-earner households: 60 percent eats out more than once per month, and 82 percent eat takeaway food more than once per month (35 percent even more than once a week). For one-earner households these figures are 26 percent for restaurant visits more than once per month and 62 percent for takeaway food more than once per month (SCP, 2000a). Visiting restaurants is highly positively correlated to income and negatively correlated to having children. Takeaway food is cheaper and less time-consuming, which explains why in particular middle-income households with children eat takeaway food (SCP, 2000a). Between 1980 and 1999, the expenditures on outsourcing meal preparation in the Netherlands increased from 3.1 to 4.2 percent of the total household budget (CBS, 2001). Time spent on cooking and dishwashing significantly decreases with takeaway food and increases with the number of children living at home, cohabiting without children, and having children aged above 15 years and older (Tijdens et al., 2000). Time spent on cooking also decreases with the working hours of both males and females (Labriex and Siegers, 1993). Research in other western countries shows that outsourcing meal preparation is positively related with income, employment status, urban location, and the number of people in the household aged above 14 (Heiman et al., 2001; Mihalopoulos and Demoussis, 2001; and Manrique and Jensen, 1998). The expenditures on food away from home are positively associated with income and education (Mihalopoulos and Demoussis, 2001). Higher educated people are more likely to participate in preparing meals, but the amount of time allocated on meal preparation is less than the time lower educated people allocate to meal preparation (Florkowski et al., 2000). Research on outsourcing behaviour of immigrants is scarce, some studies have been found in the United States and in Switzerland. Both US studies show that blacks (although the question arises whether these people should be considered as immigrants or not) spend less on food away from home and domestic services, but more on clothing care than whites (Soberon-Ferrer and Dardis, 1991; Bellante and Foster, 1984). In Switzerland, when corrected for wage rate, immigrants spend more time on household activities, which could imply that they outsource less (Sousa-Poza et al., 2001). Also in the Netherlands, immigrants may outsource less household tasks, or other tasks than the native Dutch. Recent research shows a
Time Allocation and Outsourcing within Households
143
large unfulfilled demand for childcare amongst immigrants, the cause of which is unknown1 (Tijdens et al., 2000). A difference in culture can be an explanation: immigrant mothers might want to outsource the care for their children, but their husbands do not give their permission because of their values and beliefs or religion (cultural background).
2.3. Household Production and Time Allocation Models In 1965, Becker introduced a model that would dramatically change the way households were studied in traditional microeconomic theory. Becker’s idea was that households obtain utility not only indirectly from market goods and services purchased on the market, but also directly from household commodities. Household commodities are products that the members of the household produce by combining their time with market goods. Market goods and services are not the only inputs in this process; the other input is the household’s time. Becker’s theory is called the ‘New Home Economics’ (NHE) theory. In the model of household production, maximising household utilities directly depends on household commodities. According to this approach, a household both consumes and produces, and maximises welfare subject to time and budget constraints. Welfare is a function of household commodities, which are produced using market goods and time (for example, a clean house and prepared meals). So, the production process requires input: time of the household members and market goods; and generates one or more outputs: commodities. This idea is formalised by introducing household production functions, where the commodities resulting from these production processes generate utility in turn. Becker’s theory has transformed the household from a passive maximizer of utility from market goods into an active maximizer also engaged in extensive production and investment activities (Becker, 1965). According to Gronau (1977, 1980), the analysis can be substantially simplified by the assumption that output of the home production process is a perfect substitute for goods that can be bought on the market2. Where neo-classical theory only distinguishes labour and leisure, the NHE theory makes a distinction between work, leisure and household labour. In later work, Gronau (1986) reduces trade-offs between spending one more hour on home production and selling this hour on the labour market to a comparison of the wage rate and the marginal product value of household production time3. In this way the role of the utility function is limited to the decision to allocate the time that is not used for home production activities to leisure and work. In the extreme case, work at home and work in the market are perfect substitutes4. A person is indifferent to the composition of the goods and services (s)he consumes, whether they are produced at home or purchased in the market. Browning and Chiappori (1998) have added important new insight in using the household production theory. For individuals, the household production theory implies efficient outcomes for each
1
People were asked whether they found formal childcare important in the welfare policy making of the municipality they lived in (Tijdens et al., 2000). 2 This is not always true. For example, when outsourcing meal preparation, a ready-to-eat meal is not always a perfect substitute for a home made meal. Nevertheless, this assumption is necessary to simplify the model (and in many cases, market goods and commodities are close substitutes). 3 Which is then equalized in equilibrium (Gronau, 1986). 4 Under the assumption that the person(s) in the household is/are able to work and has/have the opportunity to participate on the labour market.
144
J.R. Cornelisse-Vermaat, H. Maassen van den Brink, J.A.C. van Ophem et al.
decision. Browning and Chiappori prove this is also true in multi-person households, and call it ‘collective setting’. For several decades, household production has received much attention in economic literature. Between 1965 and 1990, the economic theory was mainly focused on extending theoretical notions, not on empirical research because of econometric difficulties and lack of data. The empirical research during the mentioned period was among others done to solve econometric issues, like corner solutions (Kooreman and Kapteyn, 1987; Gronau, 1986; Graham and Green, 1984; and Gronau, 1977). Homan et al. (1991) used Becker’s model to estimate the monetary value of household production in Dutch households. During the past years, more attention has been paid to empirical research on household production (see for example Ermisch, 2003; Maassen van den Brink and Groot, 1997). Empirical household production research is now also performed in countries where in the past such research was not common, like Spain (Manrique and Jensen, 1998), and Bulgaria (Florkowski et al., 2000). For an overview of applications of the NHE theory to paid labour and household work, see Ermisch (2003), Vogel et al. (2003), Jenkins et al. (1998), and Maassen van den Brink and Groot (1997). Although Becker’s ideas have been criticised over the years5, the NHE theory has been leading in economic theory in time allocation and has led to many results in opening the “black box” of household production. In the (recent) past, Becker’s model has been used for constructing household production models including outsourcing. These models were mainly constructed to explain household’s expenditures on outsourcing meal preparation and home cleaning. Hallberg and Klevmarken (2003) have developed a model with time spent with children as a utility function depending among others on out-of-home childcare. Mihalopoulos and Demoussis (2001) and Manrique and Jensen (1998) use Becker’s model to estimate expenditures on food away from home (and at home). Florkowski et al. (2000) construct a household production model to predict the time allocated to meal preparation. Kim (1989) uses the model to investigate the determinants of time-saving tendencies in Canadian households.
3. Description of the Data on Time Allocation and Outsourcing The data was collected during a telephone survey in the Netherlands between September and November 2001 by a Dutch market research organization (DESAN). The Dutch subsample was drawn randomly from the total pool of phone numbers (about 6.8 million) administered by the Dutch Telephone Company in 2001. The immigrant sub-samples were drawn from a sample of 80,000 names owned by DESAN and used for bi-yearly investigation. The immigrants were selected on the basis of their name(s) (indicating their ethnicity). DESAN looked up these names in the total pool of phone numbers mentioned above. 5
For example, Chiappori (1997, 1988), Lundberg and Pollak (1994, 1993), Sawhill (1980) criticise the fact that Becker assumes altruism within a multi-person household. Others argue the fact that Becker attaches too much weight on purely economic motives, leaving not much room for social and psychological motives (see among others Kirzner, 1999 and Haagsma, 1995).
Time Allocation and Outsourcing within Households
145
The purpose of the data gathering was to obtain samples of four different ethnic groups, each containing 50 percent women and at least 25 percent households with children living at home. The average response rate in all ethnic groups was 23.4%. The total sample (N=2551) consists of 4 different groups: the native Dutch (N=701), the Surinamese/Antilleans (N=701), Moroccans (N=449), and Turks (N=700). During the telephone interview it was explained to the respondents that the survey was held to obtain more insight in the life situation of the Dutch population. The respondents were asked about their outsourcing behaviour, their time allocation, as well as socio-economic and demographic circumstances (see Appendix I for variables). In our analyses we use total time spent on household and care activities as a dependent variable. Total time includes time spent on household tasks like home cleaning and doing laundry, time spent on cooking6, and time spent on childcare, and is the sum of the time both partners in the household spend on these time categories7. Table 6 gives an overview of time allocation of the native Dutch and immigrants in our sample. Table 6. Time allocation for the native Dutch and immigrants according to gender (N=2551) Hours per day (24 hours) native Dutch Working time female male HH activities* female male Leisure** female male Sleeping, eating, pers. care female male Education female male TOTAL
3.9 3.1 5.2 2.9 3.6 1.5 5.9 6.1 5.7 11.1 11.0 11.3 0.2 0.2 0.3 24.0
Surinamese/ Antilleans 5.4 4.7 6.3 2.6 3.4 1.7 5.5 5.3 5.7 10.1 10.3 9.9 0.3 0.3 0.4 24.0
Moroccans
Turks
4.6 2.8 6.1 3.2 4.8 1.7 5.5 5.4 5.5 10.2 10.4 10.1 0.5 0.6 0.2 24.0
3.1 1.9 4.7 3.3 4.3 2.2 6.9 6.9 7.0 10.1 10.4 9.7 0.5 0.5 0.4 24.0
* household activities include meal preparation, domestic tasks, and care for children ** incl. time spent on newspapers, TV, partner, family, computer, books, and remaining free time
Table 6 shows that the Turks in our sample have the highest mean household and care time, followed by the Moroccans. The Surinamese/Antilleans spend less time on household 6
Our data had one combined time category for eating and cooking. The most recent data from Time Budget Research in the Netherlands (TBO 2000) shows that from the total time spent on eating and cooking (for all meals), one third is spent on cooking and preparing meals. We multiplied our time category for eating and cooking with one third in order to get an estimation of the time spent on cooking only.
146
J.R. Cornelisse-Vermaat, H. Maassen van den Brink, J.A.C. van Ophem et al.
and care time than the native Dutch, which can be ascribed to more working hours for the Surinamese/Antilleans. Compared to the other groups, Surinamese/Antilleans have relatively more working hours. As mentioned before, this could be explained by the fact that for the native Dutch mainly (not-working or part-time working) women participated in the research. Time spent on sleeping and personal care is about the same for all groups. Table 7 shows the time allocation of the four groups divided over households with children and households without children. Analysis of variances was conducted in order to investigate whether the differences between the time categories differ significantly over the four groups. Table 7. Time Allocation for the native Dutch and immigrants for different households
Working time b, c hh with children hh without children HH activities a, b, c hh with children hh without children Leisure a, b, c hh with children hh without children Sleeping, eating, pers. care a, b, c hh with children hh without children Education a, b, c hh with children hh without children TOTAL
native Dutch (n=701)
Surinamese/ Antilleans (n=701)
Moroccans (n=449)
Turks (n=700)
4.3 3.6
5.8 4.9
4.2 5.1
3.1 3.4
4.2 1.8
3.3 2.0
4.3 1.5
3.9 1.6
4.8 6.8
4.7 6.2
5.2 5.9
6.9 7.0
10.5 11.6
9.9 10.3
10.1 10.5
9.8 10.9
0.2 0.2 24.0
0.2 0.5 24.0
0.2 1.1 24.0
0.2 1.0 24.0
a: significant main effects of having children across the four groups (ANOVA) b: significant main effects of ethnicity across households with and without children (ANOVA) c: significant interaction effects (ANOVA)
The native Dutch and Surinamese/Antillean respondents with children spend more time on paid labour, while the Turkish and Moroccan respondents with children spend less time on paid labour. The ethnic groups differ significantly in number of working hours (F = 25.563, p <0.01). For all time categories an interaction effect between ethnicity and children living at home exists (working time F = 4.455, p <0.01; household and care activities F = 9.466, p <0.01; leisure F = 8.133, p <0.01; sleeping F = 2.523 p <0.10, study F = 14.195, p <0.01). This means that for all time categories the different ethnic groups respond differently in their time allocation when children are present. Time allocated to household and care activities is
7
In our data, we had a time division over 24 hours of the respondent and we had information on the working hours of the partner. We assume partners spend leisure, sleeping, and remaining time together (Hallberg, 2003; and based on Table 4), leaving the rest of the partner’s time to be spent on household activities.
Time Allocation and Outsourcing within Households
147
significantly higher in households with children in all groups, which is caused by the time needed to take care of children. Households with children at home differ significantly from households without children at home in their time spent on household activities (F = 413.391, p <0.01). Surinamese/Antillean households show the lowest number of hours spent on household activities, which is compensated by the higher number of working hours. The four ethnic groups and the different household types (with or without children at home) differ significantly in leisure time (F = 21.672, p <0.01 for ethnic groups and F = 52.224, p <0.01 children living at home). The native Dutch and Surinamese/Antilleans ‘lose’ more leisure time than the Turks or Moroccans when they have children. As also seen in Table 5.6, the Turkish respondents have the most leisure time, which is associated with fewer working hours. For all groups, it is shown that households with children spend less time on sleeping and personal care than households without children (significant effect F = 34.052, p <0.01), which can be attributed to the time needed for childcare. Except for the native Dutch, all groups spend less time on education when they have children. In this time category, ethnic groups differ significantly (F = 19.691, p <0.01), households with children living at home differ significantly from households without children living at home (F = 92.734, p <0.01), and there is a significant interaction effect (F = 14.195, p <0.01). When having children, all groups spend more time on household and care activities, less time on sleeping, eating, and personal care, and less time on leisure. The native Dutch and Surinamese/Antilleans make more working hours when they have children, while Moroccans and Turks make fewer working hours when they have children. Table 8 gives an overview of the use of outsourcing methods by the native Dutch, Surinamese/Antillean, Moroccan, and Turkish households. Again, analysis of variances is done to investigate whether the differences in outsourcing between the groups are significant for each outsourcing category. Table 8. Use of outsourcing methods per group (in percentages)
Household help*** Window cleaner*** Launderettea* Carwash*** Childcareb* Takeaway food*** Delivery food*** Eating out***
native Dutch (N=701) 18.7 14.8 14.4 23.0 41.3 61.5 14.8 54.6
Surinamese/ Antill. (N=701) 13.1 21.8 15.5 31.7 32.5 64.1 22.3 45.1
Moroccans (N=449) 4.2 10.2 14.9 36.5 33.5 50.6 19.4 37.3
Turks (N=700) 3.4 12.0 11.1 35.3 27.8 61.3 20.6 38.1
a: and dry cleaner’s b: when children living at home * p < 0.10, ** p < 0.05, *** p < 0.01 (ANOVA)
The analyses of variance show that the four groups differ significantly in all outsourcing categories. Almost a fifth of the Dutch households uses household help. Surinamese/Antilleans make less use of this service, while Moroccans and Turks scarcely use
148
J.R. Cornelisse-Vermaat, H. Maassen van den Brink, J.A.C. van Ophem et al.
a household help. The use of a window cleaner is about the same for all groups, except for Surinamese/Antilleans who make use of this service about 10 percent more than the other three groups. The native Dutch make less use of the carwash than immigrants of whom about a third makes use of the carwash. About 30 percent of the Moroccans and Turks in our sample make use of childcare, which is higher than the NIBUD figures (2004) in Table 4. However, Table 9 shows that the Turks pay less for childcare than the other groups. This difference can be attributed to the fact that NIBUD (2004) only measured the use of formal childcare, while we measured the use of both formal and informal childcare8. Turks probably outsource childcare more to (unpaid) family and friends. Turks outsource meal preparation relatively frequently. This is also seen in Table 8, although a larger part of the native Dutch eat out. Table 9 gives insight into the expenditures on outsourcing by households that make use of it, separately for households with and without children. Table 9. Average monthly amount in Euros* spent on outsourcing for all households according to different household types (standard error in parentheses)
90 (6.7)
Surinamese/ Antilleans 96 (9.7)
111 (13.1) 75 (6.1) 173 (22.2)
110 (16.4) 84 (11.4) 163 (22.7)
45 (22.0) 64 (11.4) 101 (17.3)
78 (25.2) 50 (16.7) 95 (18.7)
Window cleaner
6 (0.5)
8 (0.7)
6 (0.7)
6 (0.5)
hh with children hh without children Launderette & dry cleaner’s
7 (0.9) 6 (0.7) 14 (1.1)
8 (0.7) 8 (1.4) 28 (2.8)
6 (0.7) 6 (1.5) 26 (3.6)
6 (0.6) 4 (1.1) 28 (3.4)
hh with children hh without children Carwash
15 (1.8) 13 (1.3) 9 (0.5)
24 (2.4) 32 (4.7) 12 (0.7)
30 (5.4) 20 (3.7) 11 (0.9)
26 (3.6) 35 (8.1) 10 (0.7)
hh with children hh without children Takeaway food
9 (0.8) 9 (0.8) 39 (1.8)
14 (1.0) 11 (1.1) 48 (2.3)
11 (1.2) 11 (1.3) 52 (4.7)
10 (0.7) 13 (1.7) 49 (3.9)
hh with children hh without children Delivery food
44 (2.7) 33 (2.2) 34 (2.8)
48 (3.0) 48 (3.5) 39 (5.5)
47 (5.2) 58 (8.3) 55 (11.0)
48 (5.0) 51 (5.3) 49 (6.4)
hh with children hh without children Eating out
33 (3.1) 35 (4.5) 88 (6.5)
38 (5.7) 39 (9.5) 88 (10.2)
73 (17.5) 25 (2.7) 69 (4.4)
42 (6.9) 62 (13.0) 72 (6.6)
104 (15.4) 78 (5.1)
88 (11.8) 88 (15.2)
61 (5.9) 76 (6.5)
64 (4.3) 86 (16.6)
Dutch Household help hh with children hh without children Childcare (hh with children)
hh with children hh without children
Moroccans
Turks
54 (11.7)
68 (17.4)
* rounded off to the nearest Euro 8
We measure both formal and informal childcare, because the number of hours childcare is outsourced affects the total time parents spend on care for their children.
Time Allocation and Outsourcing within Households
149
Moroccans spend the most on takeaway food and delivery food. Table 8 showed that the Turks eat takeaway food the most frequently. Apparently, they choose cheaper takeaway food than Moroccans. The relatively large standard errors for almost all categories indicate that expenses on outsourcing facilities differ among households. The native Dutch and Surinamese/Antilleans spend the most on eating out. Although the groups differ in their outsourcing behaviour with respect to meal preparation, their monthly expenditures are about the same for all groups. The native Dutch outsource home cleaning the most frequently, while they pay relative little. This could be explained by the use of black labour for household cleaning (these households pay a lower price for household cleaning, because no tax is paid), which might be higher among the native Dutch. The frequency of outsourcing childcare is about the same among all groups. The native Dutch have the highest costs for outsourcing childcare per month, which can be caused by the fact that the native Dutch outsource their childcare more hours per month than the other groups or by the fact that immigrant groups outsource childcare more to family and friends (informal childcare), or they make more use of subsidized pre-school education. Another cause on the difference in expenditures on childcare is the fact that the costs of formal childcare depends on income (households with higher incomes have to pay more than households with lower incomes). There are only small differences in the outsourcing behaviour between households with and without children living at home. In most categories, the differences are small, except for childcare and household help. There appear to be some differences for household help, delivery food, and eating out. In all groups, households9 with children pay more for household help than households without children. Moroccans spend more on delivery food when they have children, while Turks spend less on delivery food when they have children. Since eating out is in general more expensive than takeaway food or delivery food, one could expect that with the arrival of children a household will spend less on eating out. This is true for Moroccans and Turks, but the native Dutch spend significantly more on eating out when they have children, indicating they will still eat out and even take the children with them. The Dutch also spend more on takeaway food when they have children. These habits could indicate that the native Dutch outsource meal preparation more when they have children, perhaps because of lack of time.
4. Empirical Model of the Demand for Household and Care Time and Outsourcing We use the household production theory of Becker (1965) and Gronau (1986, 1972) as a framework to construct linear specifications of the demand functions for household and care time and the demand for outsourcing household and care activities. We are aware of the fact that the number of working hours ( N i ) is endogenous. However, in real-life, it will be difficult to adjust the time spent on market labour (especially in the short run) when compared with household and care time or leisure. Therefore, we 9
Except for the Moroccan households, but those amounts are based on only a few respondents.
150
J.R. Cornelisse-Vermaat, H. Maassen van den Brink, J.A.C. van Ophem et al.
consider N i as given for each partner (see also Van Ophem and De Hoog, 1995a; Lambriex and Siegers, 1993; and Homan, 1998), and equal to N i . The following demand functions can be constructed for demand for household and care activities time10 ( H , measured in hours per day) and the demand for outsourcing goods and services ( O , measured in expenditures per month11) for a two-person household (couples):
with
H = α 0 + α 1Y + α 2 N f + α 3 N m + α 4 D + ε 1
(1)
O = β 0 + β 1Y + β 2 N f + β 3 N m + β 4 D + ε 2
(2)
α 0 and β 0 are the constant terms, α 1 to α 4 , and β 1 to β 4 coefficients to be
estimated. Y is the household income per month (net income of each partner, and/or other income like social security or children’s allowance). N f and N m are the market working hours for female and male per week12. D is a vector of socio-economic and demographic variables and includes dummies for ethnicity (with the native Dutch as reference for Surinamese/Antilleans, Moroccans, and Turks13), age of each partner, level of education of each partner, children living at home, living in an urban area, being a homeowner, religious affiliation, and health. ε 1 and ε 2 are the stochastic disturbance terms with a normal distribution and zero mean. A high number of working hours correlates with a high household income. In households with a high number of working hours, household and care activities may be outsourced to “buy” time for activities with their children (Hallberg and Klevmarken, 2003). Therefore, working hours are expected to have a negative relationship with household and care time, but to have a positive relationship with outsourcing. Because of their lower income levels, immigrants are expected to outsource less than the native Dutch. Education is also be an indication of high working hours and a high income and therefore lower household and care time, but higher expenditures on outsourcing. The effect of age on household and care time can be either positive or negative. Older people have more time for household and care tasks, but a poorer health situation can cause a decrease in household and care time. Younger people could outsource more, because they work more hours on the labour market and have less time for household and care tasks. On
10
We study the following household and care tasks: cooking, childcare, and other household tasks, like cleaning, doing the laundry etc.
11
These are measured as
Po O , because our cross-section data do not give enough information about the prices. For simplicity reasons we write O in equation (2). The same holds for the prices of goods X , they are not
included in the empirical model. Working hours are measured per week and household and care time is measured per day, while income and expenditures are measured per month. This difference in measurements does not affect the estimation results. We use these figures this way, because they are easier to interpret. 13 Where we assume that both partners in each household have the same ethnicity, which is true for about 80 percent of the households in our sample. 12
Time Allocation and Outsourcing within Households
151
the other hand, older people can have more money to outsource household and care tasks and may have a larger need for it. As discussed in section 2, earlier research shows that household and care time is higher when having (young) children in the household. Households with (young) children are expected to outsource more, for example childcare. People who own their home are assumed to have a higher income; therefore, a positive effect on time spent on household and care activities is hypothesised. A positive relation between living in an urban area and outsourcing is expected, since living in an urban area will give more opportunities to outsource childcare and household tasks. People with a religious background may have a more traditional time division between men and women, and therefore spend more time on household and care activities and make less use of outsourcing facilities. The effect of health can be either positive or negative, since people with a poor health need more time to do household and care activities or need to outsource many of these activities.
5. Estimation Results In order to be able to estimate the demand for outsourcing, we have only included the respondents in the sample who reported making use of the several outsourcing options and have expenditures on outsourcing (this means that we leave out people who have no expenditures on outsourcing) (n=2170). In order to investigate the effect of outsourcing on the total time households spend on household and care both demand equations are estimated. For the estimations of (1) and (2), the Ordinary Least Squares regression method is used14 (see section 4). Table 10 gives the estimation results for couples living in a household (with or without children) for both equations. Besides the estimations, the standardized coefficients are presented as well to give a one-sight overview of the importance of the variables. The working hours of both males and females have a significantly negative effect on the household and care time. The expenditures on outsourcing do change significantly due to working hours of each partner. Having (very) young children in the household increases the household and care time and the expenditures on outsourcing significantly. As children are older (from the age of 12), these effects become smaller. And as the standardized coefficients show, the effect of having (very) young children is larger on household and care time than on outsourcing. Very young children need more care than older children. And as children become older they will be able to help in the household, which decreases the need for outsourcing. Lower educated women spend more hours on household and care time. There could be several reasons to explain this effect: 1) in general, lower educated women have less working hours on the market; 2) lower educated women may find it important to take care of their husband or may spend more time on preparing meals that will be shared with families and friends, while higher educated women may not consider a clean house as very important; and 3) there might also be a slight change that lower educated women are less efficient in their household and care tasks, although this cannot be proved.
14
Appendix I comprises a list of the variables used in the estimations.
152
J.R. Cornelisse-Vermaat, H. Maassen van den Brink, J.A.C. van Ophem et al.
Table 10. Estimation results of OLS regression on the demand for household and care time and the demand for outsourcing for couples with or without children (t-values in parentheses)
Constant Native Dutch Surinamese/Antilleans Moroccans Turks Working hours female Constant for wrk hrs female not obs. Working hours male Constant for wrk hrs male not obs. Household income Children at home 0-3 Children at home 4-11 Children at home 12-15 Children at home 16-25 Children at home > 25 Low level of education female Medium level of education female High level of education female Low level of education male Medium level of education male High level of education male Living in an urban area Religious affiliation Health Homeowner Age female 18 – 34 Age female 35 – 44 Age female 45 – 64 Age female ≥ 65 Age male 18 – 34 Age male 35 – 44 Age male 45 – 64 Age male ≥ 65 # Observations Adjusted R2 F statistic
Total household and care time (5.1) stand. coeff 5.175 (6.231)*** ref. group 0.758 (2.241)** 0.067 1.299 (3.126)*** 0.102 0.503 (1.392) 0.050 -0.055 (-3.842)*** -0.190 -0.005 (-0.011) -0.001 -0.028 (-2.385)** -0.111 0.513 (0.896) 0.044 0.005 (0.793) 0.024 2.111 (7.204)*** 0.206 1.122 (4.131)*** 0.118 0.612 (1.890)* 0.053 0.078 (0.232) 0.006 ref. group ref. group 0.485 (1.656)* 0.051 0.079 (0.201) 0.007 ref. group 0.197 (0.690) 0.021 -0.502 (-1.414) -0.046 0.292 (1.099) 0.028 0.329 (1.312) 0.033 0.500 (4.156)*** 0.106 -0.095 (-0.365) -0.010 ref. group 0.666 (1.757)* 0.065 -0.480 (-0.840) -0.040 -2.378 (-2.343)** -0.094 ref. group 0.121 (0.348) 0.012 0.231 (0.443) 0.021 -0.031 (-0.036) -0.002 1452 0.169 12.351
Log total outsourcing expenditures (5.2) stand. coeff 4.185 (15.503)*** ref. group -0.282 (-2.569)*** -0.078 -0.370 (-2.742)*** -0.089 -0.412 (-3.516)*** -0.126 -0.001 (-0.158) -0.008 -0.214 (-1.448) -0.071 0.005 (1.349) 0.063 -0.085 (-0.545) -0.022 0.000 (4.897)*** 0.152 0.207 (2.174)** 0.062 -0.063 (-0.711) -0.020 0.143 (1.356) 0.038 0.068 (0.621) 0.017 ref. group ref. group 0.265 (2.780)*** 0.087 0.498 (3.876)*** 0.127 ref. group 0.330 (3.556)*** 0.106 0.283 (2.457)** 0.080 0.140 (1.623) 0.041 0.009 (0.108) 0.003 -0.001 (-0.036) -0.001 0.063 (0.749) 0.021 ref. group 0.103 (0.833) 0.031 -0.051 (0.277) -0.013 -0.318 (-0.965) -0.039 ref. group -0.206 (-1.813)* -0.064 -0.391 (-2.305)** -0.110 -0.275 (-0.985) -0.042 1452 0.163 11.901
* p< 0.10 ** p<0 .05 *** p<0.01
The level of education of both females and males positively affects the total amount spent on outsourcing. A social-class effect could explain this result; in higher social classes it is more common (and more accepted) that females work and more household and care
Time Allocation and Outsourcing within Households
153
activities need to be outsourced. This social-class effect explains the higher household and care time hours for Moroccans. Compared to the native Dutch, Moroccan have a lower socio-economic status and less female labour participation. A culture where women are expected to stay at home and take care of the children can also be an explanation for the higher household and care time of Moroccans. Surinamese/Antilleans also have more household and care time than the native Dutch, but their socio-economic situation is quite comparable to that of the Dutch. This result could suggest that compared to the native Dutch, the Surinamese/Antilleans consider other things to be important (like a clean house, or preparing meals). All immigrant groups spend significantly less on outsourcing than the native Dutch. Healthy people spend more time on household and care activities than people with a poorer health (who are not able to do all these activities themselves). Nevertheless, no significant effect is found for health and the expenditures on outsourcing. This indicates that people with a poorer health do not spend more money on outsourcing than healthy people. Although expected, people living in urban areas do not spend more on outsourcing. Also, no effects were found for religious affiliation and being a homeowner. Younger females have significant more household and care time, which is explained by having young children. Women aged above 65 have significant less household and care time, which can be ascribed to the fact that as people become older, it becomes physically harder (and usually health becomes poorer) to do household and care tasks. Women aged above 65 have also less household and care tasks, since no children live at their home anymore. The household and care time of males is not affected by their age. The expenditures on outsourcing decrease significantly with male’s age (in many cases the main provider for household income). Of course, partly this could be explained by the fact that older people do not have (young) children living at home and have fewer working hours (or do not work at all), which decreases the need for outsourcing. Also, older men may have a traditional household where the wife does not do paid labour and does (almost) all household and care activities.For couples (with or without children), income does not affect the total time spent on household and care activities. Income has a significantly positive effect on the expenditures on outsourcing, which was also found in earlier research (Lambriex and Siegers, 1993). For a one-person household (or, better is to speak of ‘singles’ since children could also be present in the household), these results are different, because in such households there is only one person to do all household and care activities and to generate income. Therefore, we have repeated the estimations for singles (including one-person households and one-parent families). Because there is no partner in households of singles, only the working hours, level of education, and age of the respondent are included in the demand functions (1) and (2). The results are shown in Table 11. The results of the estimations in Table 11 give some interesting additional information. It shows that single females spend significantly more hours on household and care tasks. This effect could have overruled the effect for working hours in this analysis. These women could also be single mothers who have to take care of children by themselves.
154
J.R. Cornelisse-Vermaat, H. Maassen van den Brink, J.A.C. van Ophem et al.
Table 11. Estimation results of OLS regression on the demand for household and care time and the demand for outsourcing for singles with or without children (t-values in parentheses) Total household and care time (1) stand. coeff Constant 0.811 (2.107)** Female 0.552 (3.608)*** 0.119 Native Dutch ref. group Surinamese/Antilleans 0.654 (3.483)*** 0.139 Moroccans 0.757 (3.084)*** 0.120 Turks 0.443 (1.837)* 0.075 Working hours 0.001 (0.102) 0.006 Constant for wrk hrs not obs. 1.551 (5.666) 0.321 Household income -0.008 (-1.048) -0.042 Children at home 0-3 2.420 (6.575)*** 0.204 Children at home 4-11 2.533 (8.651) 0.289 Children at home 12-15 0.405 (1.167) 0.038 Children at home 16-25 -0.156 (-0.499) -0.016 Children at home > 25 ref. group Medium level of education -0.186 (-0.960) -0.040 High level of education -0.223 (-0.974) -0.043 Living in an urban area -0.099 (-0.658) -0.021 Religious affiliation 0.203 (1.255) 0.038 Health -0.113 (1.505) -0.047 Homeowner 0.232 (1.197) 0.040 Age 18 – 34 ref. group Age 35 – 44 0.759 (3.363)*** 0.116 Age 45 – 64 0.830 (3.565)*** 0.140 -0.077 (-0.245) -0.010 Age ≥ 65 # Observations 715 Adjusted R2 0.375 F statistic 22.482
Log total outsourcing expenditures (2) stand. coeff 4.200 (14.505)*** -0.055 (-0.479) -0.018 ref. group 0.248 (1.756)* 0.081 -0.120 (-0.649) -0.029 0.219 (1.211) 0.057 -0.004 (-0.641) -0.044 -0.403 (-1.957)* -0.127 0.000 (4.771)*** 0.222 -0.261 (-0.942) -0.034 -0.306 (-1.389) -0.053 0.249 (0.955) 0.036 -0.298 (-1.266) -0.046 ref. group 0.308 (2.109)** 0.102 0.279 (1.615) 0.082 0.160 (1.419) 0.051 -0.041 (-0.327) -0.012 -0.049 (-0.873) -0.032 0.436 (2.995)*** 0.114 ref. group -0.483 (-2.842)*** -0.113 -0.622 (-3.533)*** -0.161 -1.110 (-4.674)*** -0.215 715 0.174 8.554
* p< 0.10 ** p< 0.05 *** p< 0.01
Also in households of singles (with or without children), the household and care time increases when having (very) young children. But no significant effect of children is found for expenditures on outsourcing. The single-parent households can have less money to spend on outsourcing. Still, they will also need childcare if they want to do paid labour. It is therefore interesting to repeat the estimations of outsourcing separated by type as mentioned in section 3 (these results will be presented in Table 13). In households of singles, the level of education does not affect household and care time. While in households of couples lower educated women spend more hours on household and care time. This effect could have been diminished, because there was controlled for gender in the analyses of singles. Also, in households of singles, level of education may not be important, because the household and care tasks should be done anyway since there is no partner to help with the tasks. The expenditures on outsourcing increase with the level of education.
Table 12. Estimation results of OLS regression on the demand for household and care time and the demand for outsourcing for couples households with and without children (t-values in parentheses) Total household and care time (1) without children Constant
Log total outsourcing expenditures (2)
with children
without children
with children
4.566
(3.406)***
6.362
(6.103)***
4.552
(9.441)***
3.946
(11.856)***
ref.
group
ref.
group
ref.
group
ref.
group
Surinamese/Antilleans
1.754
(3.470)***
0.088
(0.201)
-0.345
(-1.895)*
-0.235
(-1.686)*
Moroccans
0.446
(0.582)
1.230
(2.443)**
-0.371
(-1.346)
-0.408
(-2.536)**
Turks
1.048
(1.493)
0.020
(0.046)
-0.568
(-2.251)**
-0.408
(-2.909)***
Working hours female
-0.044
(-1.679)*
-0.048
(-2.686)***
0.007
(0.689)
-0.001
(-0.222)
Constant for wrk hrs female not observed
0.100
(0.108)
0.203
(0.378)
0.305
(0.914)
-0.279
(-1.626)
Working hours male
-0.040
(-1.841)*
-0.026
(-1.874)*
-0.009
(-1.099)
0.010
(2.242)**
Constant for wrk hrs male not observed
-0.636
(-0.635)
1.042
(1.495)
-0.870
(-2.418)**
0.194
(0.872)
Household income
0.001
(0.048)
0.006
(0.743)
0.007
(1.868)*
0.000
(4.470)***
Native Dutch
Children at home 0-3
-
1.345
(3.740)***
-
0.244
(2.127)**
Children at home 4-11
-
0.437
(1.348)
-
-0.040
(-0.390)
Children at home 12-15
-
0.347
(1.005)
-
0.165
(1.499)
Children at home 16-25
-
-0.788
(-1.981)*
-
0.064
(0.505)
Children at home > 25
ref.
group
ref.
group
ref.
group
ref.
group
Low level of education female
ref.
group
ref.
group
ref.
group
ref.
group
Medium level education female
0.372
(0.669)
0.614
(1.779)*
0.218
(1.089)
0.252
(2.288)**
High level education female
0.463
(0.695)
0.046
(0.093)
0.368
(1.538)
0.513
(3.275)***
Table 12. Continued Total household and care time (1) without children
Log total outsourcing expenditures (2)
with children
without children
with children
Low level of education male
ref.
group
ref.
group
ref.
group
ref.
group
Medium level education male
-0.186
(-0.338)
0.309
(0.922)
0.308
(1.558)
0.334
(3.117)***
High level education male
-0.754
(-1.231)
-0.397
(-0.917)
0.378
(1.716)*
0.256
(1.847)*
Living in an urban area
0.135
(0.272)
0.352
(1.127)
0.020
(0.112)
0.154
(1.546)
Religious affiliation
0.254
(0.560)
0.398
(1.329)
-0.086
(-0.525)
0.075
(0.780)
Health
0.677
(3.272)***
0.441
(3.024)***
0.080
(1.079)
-0.016
(-0.345)
Homeowner
-0.004
(-0.010)
-0.182
(-0.569)
0.057
(0.376)
0.035
(0.346)
Age female 18 – 34
ref.
group
ref.
group
ref.
group
ref.
group
Age female 35 – 44
0.576
(0.624)
0.677
(1.579)
0.655
(1.973)**
0.020
(0.147)
Age female 45 – 64
-0.117
(-0.108)
-0.455
(-0.646)
0.648
(1.659)*
-0.199
(-0.844)
-1.330
(-0.931)
-3.296
(-1.816)
0.318
(0.618)
-0.519
(-0.895)
Age male 18 – 34
ref.
group
ref.
group
ref.
group
ref.
group
Age male 35 – 44
-0.193
(-.0246)
0.086
(0.215)
-0.280
(-0.990)
-0.181
(-1.42)
Age male 45 – 64
-0.210
(-0.190)
0.223
(0.369)
-1.218
(-3.067)
-0.221
(-1.142)
Age male ≥ 65
0.162
(0.118)
0.138
(0.105)
-1.047
(-2.188)
0.020
(0.047)
# Observations
366
1085
366
1085
0.078
0.107
0.221
0.145
2.409
6.019
5.729
8.050
Age female
≥
Adjusted R
65
2
F statistic * p< 0.10 ** p< 0.05 *** p< 0.01
Time Allocation and Outsourcing within Households
157
While Table 10 shows that all immigrant groups spend less on outsourcing, in the households of singles (with or without children) this effect was not found. And in households of singles (with or without children), Surinamese/Antilleans spend even significantly more on outsourcing than the native Dutch. This result shows that outsourcing behaviour between couples and singles (with or without children) is very different. Health does not affect households and care time of singles (with or without children), while it affects households and care time for couples (with or without children). If there is only one adult in a household, that person will have to do all household and care activities, also when (s)he is ill.Household income increases the expenditures on outsourcing significantly. Although not significantly, contrary to the results in Table 10, household income decreases the household and care time. Also in the estimations for singles it is found that the expenditures on outsourcing decrease with age, which could be explained by the fact that in small households, less household and care activities need to be done. It is useful to split up the expenditures into different categories, which we will do later. First, we re-estimate the demand functions for households with children and households without children. Table 12 shows the results for couples, with and without children. Having (very) young children significantly increases household and care times as well as the expenditures on outsourcing. As earlier mentioned, very young children need a lot of care, which will be the reason on the increase in household and care time when there are children. A large part of the increase in outsourcing expenditures is caused by the expenditures on childcare. Only in household with children, lower female’s level of education increases household and care time, while both female’s and male’s level of education increase the outsourcing expenditures. A higher level of education is associated with higher income and higher working hours, and therefore a larger demand for outsourcing. If we compare immigrants to the native Dutch, we see that in almost all cases, in both types of households, they spend less on outsourcing (only not significant for Moroccan households without children). As also seen in Table 10 and 11, Surinamese/Antilleans have more household and care time. Table 12 shows that compared to the native Dutch, mainly Surinamese/Antillean households without children have more household and care time. They also spend less on outsourcing, which could indicate that these are households with lower incomes. Household income is significantly positively related to outsourcing expenditures. When looking at the t-values, we see that income is more important in households with children than in households without children. Compared to households without children and about the same income level, there is relatively less income to spend in households with children (because it needs to be shared with more persons). It is expensive to have children, thus households without children have fewer costs and more money to spend. This explains the relative higher importance of household income with respect to outsourcing expenditures in households with children. The determinants of outsourcing home cleaning and childcare can be different than for outsourcing meal preparation. Therefore, separate analyses for different outsourcing categories are performed. This could be done by further subdividing the demand equation for outsourcing (3) into separate equations for different outsourcing categories:
158
J.R. Cornelisse-Vermaat, H. Maassen van den Brink, J.A.C. van Ophem et al.
O1 = γ 10 + γ 11Y + γ 12 N f + γ 13 N m + γ 14 D + ε 12
(3)
O2 = γ 20 + γ 21Y + γ 22 N f + γ 23 N m + γ 24 D + ε 22
(4)
O3 = γ 30 + γ 31Y + γ 32 N f + γ 33 N m + γ 34 D + ε 32
(5)
where O1 represents the expenditures per month on childcare, the household help, and the window cleaner; O2 the expenditures per month on takeaway food, delivery food, and eating out; and O3 the rest of outsourcing expenditures per month, including the carwash and the laundrette/dry cleaner’s. The
γ ’s are coefficients to be estimated and ε 12 – ε 32 are
stochastic disturbance terms and normally distributed. We have focused on the main outsourcing categories (home cleaning, childcare, and outsourcing meal preparation), so equation (5) was not used. Again, the estimations are done both for couples and for singles, with or without children. The results are shown in Table 13. Table 12 already showed that for households that make use of outsourcing, household income is significantly positively related to outsourcing expenditures for both types of households. In Table 13 we also see, that for couples (with or without children), where the woman does not do paid labour (no observed working hours), the household spends less money on outsourcing home cleaning and childcare. These households will have less money to spend, and there is less necessity to outsource childcare, because the mother is at home. For singles (with or without children), the expenditures on childcare and outsourcing home cleaning decrease with working hours. These households might make more use of informal (or subsidised) childcare in stead of formal (more expensive) childcare. As could be expected, the expenditures on childcare and outsourcing home cleaning and childcare increase significantly when they have (very) young children in the household. Very young children need more care than other children, and older children can help the parent(s) with household and care tasks. If households have children between aged between 4 and 11 and/or aged between 12 and 15, the expenditures on outsourcing meal preparation increase significantly. For having children aged between 4 and 11 this can be ascribed to the fact that these children cannot help as much in the household as children aged between 12 and 15. There could also be a generation-effect for having children aged between 12 and 15, since children in this agecategory in general like takeaway food (like MacDonald’s) a lot. When couples have very young children the expenditures on outsourcing meal preparation increase, while for singles with very young children these expenditures decrease, which could be explained by a lower income in households of singles. Moroccans as well as Turks spend less on outsourcing home cleaning and childcare than the native Dutch. As discussed earlier, this could be explained by a lower socio-economic status and a culture-effect. Moroccans, Turks, and Surinamese/Antilleans spend less on outsourcing meal preparation than the native Dutch. For the Turks and Moroccans this could be explained by less income (or lower socio-economic) status as well. However, the analyses have controlled for socio-economic status by income, level of education, and working hours. This indicates that the culture-effect could be stronger than the effect of socio-economic status.
Time Allocation and Outsourcing within Households
159
Table 13. Estimation results of OLS regression on two categories of outsourcing for couples and singles with or without children (t-values in parentheses) Log outsourcing home cleaning and childcare (3) couples singles Constant 0.694 (1.812)* -0.215 (-0.624) Female 0.144 (1.052) Native Dutch ref. group ref. group Surinamese/Antilleans -0.096 (-0.618) 0.256 (1.525) Moroccans -0.344 (-1.794)* -0.110 (-0.502) Turks -0.432 (-2.594)*** -0.028 (-0.130) Working hours female -0.005 (-0.796) Wrk hrs fem. not obs. -0.694 (-3.301)*** Working hours male 0.000 (-0.031) Wrk hrs male not obs. 0.085 (0.323) Working hours -0.013 (-1.992)** Wrk hrs not observed -0.343 (-1.402) Household income 0.000 (4.558)*** 0.000 (5.094)*** Children at home 0-3 1.339 (9.902)*** 1.071 (3.253)*** Children at home 4-11 0.227 (1.812)* 0.905 (3.453)*** Children at home 12-15 -0.198 (-1.326) -0.010 (-0.034) Children at home 16-25 -0.122 (-0.788) -0.596 (-2.131)** Children at home > 25 ref. group ref. group Low level educ. female ref. group Medium level educ. fem. 0.072 (0.531) High level educ. fem. 0.549 (4.010)*** Low level educ. male ref. group Medium level educ. male 0.061 (0.461) High level educ. male 0.460 (2.812)*** Low level of education ref. group Medium level of educ. 0.487 (2.807)*** High level of educ. 0.453 (2.204)** Living in an urban area 0.039 (0.321) -0.097 (-0.722) Religious affiliation -0.007 (-0.062) 0.050 (0.338) Health -0.090 (-1.6230 -0.110 (-1.628) Homeowner 0.195 (1.634) 0.369 (2.219)** Age female 18 – 34 ref. group Age female 35 – 44 0.250 (1.427) Age female 45 – 64 0.083 (0.316) 0.334 (0.721) Age female ≥ 65 Age male 18 – 34 ref. group Age male 35 – 44 0.178 (1.103) Age male 45 – 64 0.087 (0.362) 0.163 (0.411) Age male ≥ 65 Age 18 – 34 ref. group Age 35 – 44 0.248 (1.230) Age 45 – 64 0.702 (3.368)*** 1.527 (5.405)*** Age ≥ 65 # Observations 1452 715 Adjusted R2 0.184 0.171 F statistic 13.563 8.356 * p< 0.10 ** p< 0.05 *** p< 0.01
Log outsourcing food preparation (4) couples singles 4.022 (10.958)*** 4.576 (12.489)*** -0.092 (0.633) ref. group ref. group -0.570 (-3.818)*** -0.022 (-0.121) -0.749 (-4.078)*** -0.441 (-1.889)* -0.569 (-3.558)*** 0.210 (0.877) -0.004 (0.576) -0.140 (0.693) 0.002 (0.475) -0.385 (-1.519) -0.009 (-1.263) -0.521 (-2.001)** 0.000 (4.478)*** 0.000 (3.040)*** -0.320 (-2.467)** -0.839 (-2.394)** -0.070 (-0.587) -0.809 (-2.901)*** 0.319 (2.229)** 0.651 (1.972)** 0.261 (1.754)* -0.091 (-0.304) ref. group ref. group ref. group 0.316 (2.444)** 0.421 (2.412)** ref. group 0.323 (2.557)** 0.295 (1.881)* ref. group 0.150 (0.812) 0.243 (1.111) 0.125 (1.065) 0.156 (1.094) -0.138 (-1.247) -0.290 (-1.842)* 0.064 (1.201) -0.042 (-0.590) -0.013 (-0.115) 0.365 (1.982)** ref. group 0.011 (0.066) -0.164 (-0.649) -0.674 (-1.503) -0.568 (-3.678)*** -0.873 (-3.784)*** -0.750 (-1.973)** ref. group -0.678 (-3.157)*** -1.157 (-5.222)*** -2.341 (-7.792)*** 1452 715 0.136 0.202 9.822 10.031
160
J.R. Cornelisse-Vermaat, H. Maassen van den Brink, J.A.C. van Ophem et al.
In all households the expenditures on outsourcing meal preparation decrease with age. This indicates that outsourcing (probably, mostly in the case of delivery food and takeaway food) meal preparation is mainly done by younger people. In order to calculate the elasticity of outsourcing, the regression on outsourcing expenditures is repeated for the four groups including the log of household income. The two main outsourcing categories are taken: 1) expenditures in childcare, household help, and window cleaner; 2) the expenditures on outsourcing meal preparation (both in log). The analyses as presented in Table 13 are repeated excluding working hours, level of education, and own home (which are related to income). The results are presented in Table 14. The elasticities shown in Table 14 indicate that outsourcing childcare and home cleaning is a luxury good for Surinamese/Antilleans (elasticity ≥ 1). For the native Dutch, Moroccans and Turks outsourcing childcare and home cleaning is a normal good (elasticity between 0 and 1). This means that contrary to the Surinamese/Antilleans, for the native Dutch, Moroccans, and Turks outsourcing childcare and home cleaning does not depend very much on household income. Table 14. Elasticity of expenditures on outsourcing for the native Dutch and immigrant couple (standard errors in parentheses)
0.665 (0.186)
Surinamese/ Moroccans Turks Antilleans 1.090 (0.240) 1.000 (0.303) 0.350 (0.170)
0.441 (0.144)
0.887 (0.217)
Native Dutch Expenditures on childcare, household help, and window cleaner Expenditures on outsourcing meal preparation
0.738 (0.314)
0.871 (0.194)
The elasticity for outsourcing childcare and home cleaning is the smallest for the Turks. Table 8 and 9 explain this elasticity. These tables show that Turks outsource childcare and home cleaning not very frequently compared to the native Dutch and the Surinamese/Antilleans. And if they outsource these activities, they pay a relatively low amount (due to subsidies for childcare and other/cheaper ways of outsourcing home cleaning). The elasticities for outsourcing meal preparation indicate that it is a normal good for both the native Dutch and the studied immigrant groups. Their expenditures on outsourcing meal preparation are not very much affected by a decrease or an increase in income.
6. Conclusions and Discussion In this chapter, we have studied: 1) the determinants of the demand for household and care time and the demand for outsourcing household and care activities; and 2) differences between immigrants and the native Dutch in time allocation and outsourcing behaviour. Scheme 1 shows the expected and confirmed effects of the estimations.
Time Allocation and Outsourcing within Households
161
Scheme 1. Expected and confirmed (significant) effects of the estimations
Variable Surinamese/Antilleans Moroccans Turks Gender (female) Working hours female Working hours male Household income Children living at home Level of education female Level of education male Living in an urban area Religious affiliation Health Homeowner
Household and care time (1) expected confirmed +/– Yes (+) +/– Yes (+) +/– No + Yes – Yes – Yes – No + Yes – No – No – No + No – Yes + No
Outsourcing expenditures (2) expected confirmed – Yes – Yes – Yes +/– No + No + No + Yes + Yes + Yes + Yes + No – No + No + No
+ positive effect, – negative effect, +/– either positive or negative effect expected
For all groups, households have more household and care time when they have (very) young children at home. The native Dutch and the Surinamese/Antilleans work longer when they have children, while for Moroccans and Turks the reverse is true. This can be explained by the fact that the Surinamese/Antilleans and the native Dutch have to compensate for the income-loss due to less working hours of the mothers, or want to earn more money to raise their children. The Turks and Moroccans have fewer possibilities to increase their income, since they are more often unemployed and have less good chances on the labour market. Females spend significantly more hours on household and care tasks than males. For couples (with or without children), Moroccans have more household and care time than the native Dutch, while all immigrant singles (with or without children) have more household and care time. This could be explained by less working hours (for Moroccans and Turks), or by the fact that these groups put more value on other things (taking a lot of time to prepare the meal, or for wives to take care of their husbands and children). In correspondence with the literature, we have found that having (very) young children, level of education, and household income are important determinants for the demand for outsourcing (Sousa-Poza et al., 2001; Spitze, 1999; Cohen, 1998; Manrique and Jensen, 1998; Lambriex and Siegers, 1993; Oropesa, 1993; Soberon-Ferrer and Dardis, 1991; Kim, 1989; and Bellante and Forrester, 1984). In our estimations, income is a more important factor in households with children, compared to households without children. In households with children, relatively less money can be spent, which makes a higher income more important to be able to outsource household and care activities. As expected, all immigrant groups spend less on outsourcing than the native Dutch, which could be explained by their lower socio-economic status (measured by household income and level of education). Generally, not many differences in outsourcing behaviour were found in the four studied groups. Immigrants make less use of childcare, but outsource
162
J.R. Cornelisse-Vermaat, H. Maassen van den Brink, J.A.C. van Ophem et al.
meal preparation more frequently to takeaway food or delivery food (and spend more on these outsourcing categories) than the native Dutch. The native Dutch spend the most on childcare and eating out, but they spend less on the carwash than immigrants. The outsourcing behaviour of immigrants might affect their health. A lower socio-economic status relates to an unbalanced diet, which may be a cause of overweight and health damage in the long term (Cornelisse and Maassen van den Brink, 2007; Hulshof et al., 2003; Roux et al., 2000). When people become older, their expenditures on outsourcing decrease. Mainly their expenditures on outsourcing meal preparation decrease. Particularly people younger than 35 outsource meal preparation. For singles (with or without children) the demand for home cleaning and childcare increases from the age of 45. The calculated elasticities show outsourcing home cleaning and childcare is a normal good for the native Dutch, Moroccans, and Turks (especially for the immigrants the latter can be attributed to subsidies for childcare) and a luxury good for the Surinamese/Antilleans. For immigrants as well as for the native Dutch outsourcing meal preparation is a normal good. This chapter shows that not only socio-economic determinants are of importance regarding the outsourcing behaviour and time allocation within households in the Netherlands. The decision to outsource household and care tasks may is not only an economic choice but may also be culture-driven.
Appendix I. Distribution of the Variables N=2170 (Standard Deviations in Parentheses) Total time spent on household and care activities by both partners in 24 hours (ln) Total outsourcing expenditures per month (ln) Expenditures per month on home cleaning and childcare (ln) Expenditures per month on outsourcing meal preparation (ln) Expenditures per month on rest categories of outsourcing (carwash and launderette/ dry cleaner’s) Working hours female Working hours male Children at home 0-3 y/n Children at home 4-11 y/n Children at home 12-15 y/n Children at home 16-25 y/n Children at home >25 y/n Level of education female* Low Medium High Level of education male Low Medium High
7.86 (4.95) 4.74 (1.51) 1.25 (2.07) 4.10 (2.00) 1.32 (1.68) 12.92 (16.11) 24.75 (20.89) 21.2% 31.3% 15.5% 14.0% 3.3% (reference group) 24.5% (reference group) 39.6% 17.5% 21.6% (reference group) 33.9% 20.3%
Time Allocation and Outsourcing within Households
163
Appendix I. Continued Age female 18 – 34 35 – 44 45 – 64 ≥ 65 Age male 18 – 34 35 – 44 45 – 64 ≥ 65 Surinamese/Antilleans Moroccans Turks Living in an urban area Religious affiliation** Health*** Homeowner Net monthly household income (in Euro)
47.2% (reference group) 26.9% 19.0% 5.4% 39.3% (reference group) 30.7% 22.5% 5.8% 28.0% 15.9% 27.6% 30.3% 30.4% 2.69 (0.98) 40.1% 1613.50 (930.25)
* A low level of education: primary school and vocational education. Medium level of education: lower and higher secondary education, pre-university education, and intermediate vocational education. High level of education: people holding a bachelor’s or a master’s degree. ** going to a church, temple, mosque, or synagogue more than once per month *** scale, 4 = excellent – 0 = poor.
References Álvarez, B. and Miles, D. (2003). Gender effect on housework allocation: evidence from Spanish two-earner couples. Journal of Population Economics 16, 227 – 242. Aronsson, T., Daunfeldt, S. and Wikström. M. (2001). Estimating intra household allocation in a collective model with household production. Journal of Population Economics 14, 569 – 584. Becker, G. S. (1965). A theory of the allocation of time. Economic Journal 75, 493 – 517. Bellante, D. and Foster, A.C. (1984). Working wives and expenditure on services. Journal of Consumer Research 11, 700 – 707. Bittman, M., Meagher, G., Matheson, G. (1998). The changing boundary between home and market. Australian trends in outsourcing domestic labour. Social Policy Research Centre Discussion Paper 86, 1 – 29. Browning, M. and Chiappori, P.A. (1998). Efficient intra-household allocations: a general chracterization and empirical tests. Econometrica 66, 6, 1241 – 1278. Centraal Bureau voor de Statistiek (CBS) (2001). Allochtonen in Nederland 2001 [Immigrants in the Netherlands 2001]. Centraal Bureau voor de Statistiek. Voorburg/ Heerlen. Chiappori, P.A. (1988). Nash bargained household decisions: a comment. International Economic Review 29, 791 – 796.
164
J.R. Cornelisse-Vermaat, H. Maassen van den Brink, J.A.C. van Ophem et al.
Chiappori, P.A. (1997). Introducing household production in collective models of labor supply. Journal of Political Economy 105, 1, 191 –209. Ciscel, D.H., Sharp, D.C., and Heath, J.A. (2000). Family work trend and practices: 1971 to 1991. Journal of Family and Economic Issues 21, 1, 23 – 36. Cohen, P.N. (1998). Replacing housework in the service economy: gender, class, and raceethnicity in service spending. Gender & Society 12, 219 – 231. Cornelisse-Vermaat, J.R. and Maassen van den Brink, H. (2007). Ethnic differences in lifestyle and overweight in the Netherlands. Obesity 15, 2, 483 – 493. Ermisch, J.F. (2003). An economic analysis of the family. Princeton University Press. United States of America. Florkowski, W.J., Moon, W., Resurrection, A.V.A., Jordanov, J., Paraskova, P., Beuchat, L.R., Murgov, K., Chinnan, M.S. (2000). Allocation of time for meal preparation in a transition economy. Agricultural Economics 22, 173 – 183. Graham, F. and Green, C. (1984). Estimating the parameters of a household production function with joint products. Review of Economics and Statistics 66, 277 – 282. Gronau, R. (1977). Leisure, home production, and work – theory of the allocation of time revisited. Journal of Political Economy 85, 1099 – 1123. Gronau, R. (1986). Home production – a survey. In: Ashenfelter, O. and Layard, R. (eds.). Handbook of Labour Economics Volume I. Elsevier Science Publishers BV. Haagsma, R. (1995). Microeconomic models for gender discrimination in labour markets with assymmetric information. Academisch Proefschrift. Universiteit van Amsterdam. Hallberg, D. (2003). Synchronous leisure, jointness and household labour supply. Labour Economics 10, 185 – 203. Hallberg, D. and Klevmarken, A. (2003). Time for children: a study of parent’s time allocation. Journal of Population Economics 16, 205 – 226. Hersch, J. and Stratton, L.S. (1994). Housework, wages, and the division of household time for employed spouses. Economic Issues for Work and Family. AEA papers and proceedings 84, 2, 120 – 125. Heiman, A., Just, D.R., McWilliams, B., Zilberman, D. (2001). Incorporating family interactions and socio-economic variables into family production functions: the case of demand for meats. Agribusiness 17, 4, 455 – 468. Homan, M.E. (1988). The allocation of time and money in one-earner and two-earner families: an economic analysis. Dissertatie. Erasmus Universiteit. Rotterdam Hulshof, K.F.A.M., Brussaard, J.H., Kruizinga, A.G., Telman, J., Löwik, M.R.H. (2003). Socio-economic status, dietary intake and 10 y trends: the Dutch National Food Consumption Survey. European Journal of Clinical Nutrition 57, 128 – 137. Kim, C. (1989). Working wives’ time-saving tendencies: durable ownership, convenience food consumption, and meal purchases. Journal of Economic Psychology 10, 391 – 409. Kirzner, I.M. (1999). Rationality, entrepreneurship and economic ‘imperialism’. In: Dow, S.C.H. and Earl, P.E. (eds.). Economic Organization and Economic Knowledge. Cheltenham: Edward Elgar. Kooreman, P. and Kapteyn, A. (1987). A disaggregated analysis of the allocation of time within the household. Journal of Political Economy 95, 21. Lundberg, S. and Pollak, R. (1993). Separate spheres bargaining and the marriage market. Journal of Political Economy 84, 2, 132 – 137.
Time Allocation and Outsourcing within Households
165
Lundberg, S. and Pollak, R. (1994). Non-cooperative bargaining models of marriage. American Economic Review 101, 6, 988 – 1010. Lambriex, G.E.E.M. and Siegers, J.J. (1993). Een geintegreerde analyse van tijds- en inkomenbesteding [Integrated analysis of time allocation and income allocation]. SWOKA. Instituut voor consumentenonderzoek. Onderzoeksrapport 136. Den Haag. Maassen van den Brink, H. and Groot, W. (1997). A household production model of paid labour, household work and childcare. De Economist 145, 3, 325-343. Manrique, J., and Jensen, H.H. (1998) Working women and expenditures on Food-AwayFrom-Home and At-Home in Spain. Journal of Agricultural Economics 49, 3, 321-333. Mihalopoulos, V.G., Demoussis, M.P. (2001). Greek household consumption of food away from home: a microeconometric approach. European Review of Agricultural Economics 28, 4, 421 – 432. Mokyr, J. Why more work for the mother? Knowledge and household behaviour, 1870-1945. The Journal of Economic History 60, 1, 1 – 41. Nationaal Instituut voor Budgetvoorlichting (NIBUD) (2004). De inkomsten, uitgaven en het financieel beheer van alochtone huishoudens [Income, expenditures, and financial management in immigrant households]. NIBUD. Utrecht. Oropesa, R.S. (1993). Using the service economy to relieve the double burden. Female labor force participation and service purchases. Journal of Family Issues 14, 3, 438 – 473. Rijksinstituut voor Volksgezondheid en Milieu (RIVM) (2004). Ons eten gemeten. Gezonde voeding en veilig voedsel in Nederland [Measuring our food. Healthy and safe food in the Netherlands]. Van Kreijl, C.F. and Knaap, A.G.A.C. (eds.). Bohn Stafleu Van Loghum. Houten. Sawhill, J.V. (1980) Economic perspectives on the family. In: Amsden, A.H. (ed.). The economics of women and work. New York. Soberon-Ferrer, H. and Dardis, R. (1991). Determinants of household expenditures for services. Journal of Consumer Research 17, 385 – 397. Sociaal en Cultureel Planbureau (2000a). De kunst van het combineren. Taakverdeling onder partners [The skill to combine. Division of tasks among partners]. Keuzekamp, S. and Hooghiemstra, E. (eds.). SCP. Den Haag. Sociaal en Cultureel Planbureau (SCP) and Centraal Bureau voor de Statistiek (CBS) (2002). Emancipatiemonitor 2002 [Emancipation monitor 2002]. Portegijs, W., Boelen, A. and Keuzenkamp, S. (eds.) Sociaal en Cultureel Planbureau and Centraal Bureau voor de Statistiek. Den Haag. Sociaal en Cultureel Planbureau (SCP) (2003). Rapportage minderheden 2003. Onderwijs, arbeid en sociaal-culturele integratie [Report on ethnic minorities. Education, market labour, and socio-cultural integration]. Dagevos, J. Gijsberts, M. and Van Praag, C. (eds.). Sociaal en Cultureel Planbureau. Den Haag. Sociaal en Cultureel Planbureau (SCP) (2004a). www.tbo.nl . Table 5.1 and 5.2 were made on our request by the SCP. Sociaal en Cultureel Planbureau (2006). Hoe werkt het met kinderen. Moeders over kinderopvang en werk [How does it work with children. Mothers about childcare and work]. Portegijs, W., Cloïn, M., Ooms, I., Eggingk, E. (eds.) Sociaal en Cultureel Planbureau. Den Haag.
166
J.R. Cornelisse-Vermaat, H. Maassen van den Brink, J.A.C. van Ophem et al.
Sociaal en Cultureel Planbureau (2006a). Emancipatiemonitor 2006. Veranderingen in de leefsituatie en levensloop [Emancipation monitor 2006. Changes in life situation and lifecycle]. Portegijs, W., Hermands, B., Lalta, V. (eds.) Sociaal en Cultureel Planbureau. Den Haag. Sousa-Poza, A., Schmid, H., and Widmer, R. (2001). The allocation and value of time assigned to housework and childcare: analysis for Switzerland. Journal of Population Economics 14, 599 – 618. Spitze, G. (1999). Getting help with housework. Journal of Family Issues 20, 6, 724 – 745. Strober, M. and Miling Kaneko Chan, A. (1998). Husbands, wives, and housework: graduates of Stanford and Tokyo universities. Feminist Economics 4, 3, 97 – 127. Tijdens, K., Lippe, T. Van der, Ruijter, E. (2000). Huishoudelijke arbeid en de zorg voor kinderen: herverdelen of uitbesteden? [Household labour and childcare: redistribution or outsourcing?]. Serie wetenschappelijke Publicaties. Elsevier Bedrijfsinformatie b.v. ‘s Gravenhage. Uniken Venema, H.P. Garretsen, H.F.L. and Maas, P.J. van der (1995). Health of migrants and migrant health policy, the Netherlands as an example. Social Science Medicine 41, 6, 809 – 818. Van Dam, Y.K., Hoog, C. de, Ophem, J.A.C. van (1994). Ten geleide: reflecties op gemak bij voeding [Reflections on convenience and food]. In: Van Dam, Y.K., De Hoog, C., and Van Ophem J.A.C. (eds.). Eten in de jaren negetig. Reflecties op gemaksvoeding [Eating in the 1990’s. Reflections on convenience food]. Landbouw Universiteit Wageningen Van der Lippe, T., Tijdens, K., and de Ruijter, E. (2004). Outsourcing of domestic tasks and time-saving effects. Journal of Family Issues 25, 2, 216 – 240. Van Ophem, J.A.C, and De Hoog, K. (1995). Monetarisering in de huishouding: een historische beschouwing [Monetarisation in households: a historic consideration]. Huishoudstudies 5, 2, 1 – 8. Van Ophem, J.A.C, and De Hoog, K. (1995a). Monetarisering in de huishouding: een analyse van de relatief welgestelden [Monetarisation in households: an analysis of the relatively well-to-do]. Huishoudstudies 5, 2, 8 – 19. Vogel, J. (ed.). European Welfare Production. Institutional Configuration and Distributional Outcome. Reprinted from Social Indicators Research, 63, 3. Kluwer Academic Publishers. The Netherlands. Yamada, Tadashi; Yamada, Tetsuji; Moonwon Kang, J. (1999). A study of time allocation of Japanese households. Japan and the World Economy 11, 41 – 55.
In: Outsourcing, Teamwork and Business Management ISBN: 978-1-60456-956-8 Editor: Karl E. Carettas, pp. 167-189 © 2009 Nova Science Publishers, Inc.
Chapter 10
OUTSOURCING AND PUBLIC SECTOR EFFICIENCY: HOW EFFECTIVE IS OUTSOURCING IN DEALING WITH IMPURE PUBLIC GOODS? Argentino Pessoa* Faculdade de Economia do Porto, Rua Dr. Roberto Frias, 4200-464 Porto, Portugal
Abstract The debate on new public management, together with the shortage of public funds, has had a considerable impact on public administration. Accordingly, many governments have searched positive impacts on the efficiency, equity and quality provision of public services through increasing competition and active participation of the private sector, considering outsourcing as the appropriate instrument to attain such endeavor. However, private involvement in public services provision is controversial. While, on the one hand it is touted as a way to increase efficiency and accountability by turning over choices to individuals in the market place, on the other hand, some argue that it has the potential to produce considerable fraud and corruption if managerial control by the public sector is weak. So, given this context, we aim to assess the private involvement in public services in efficiency terms, putting aside ideological considerations. So, after the introduction, we present a definition of public goods and we characterize their different types, with particular emphasis on “impure” public goods. Section 3, focuses on market failures together with equity considerations as the main reasons that configure the role of the public sector in providing impure public goods, as well as on the possibility of government failures. Section 4 deals with the benefits and costs of outsourcing in the public sector. Section 5 describes the most frequent forms of private sector involvement in the provision of impure public goods, as well as the advantages and disadvantages of the different options. Section 6 carries out some comments on the need for regulation. Finally, section 7 concludes.
Keywords: Contracting out, impure public goods, market/government failures, private sector involvement, public sector. JEL codes: H42, H50, I18, I28, L33. *
E-mail address:
[email protected]
168
Argentino Pessoa
1. Introduction Since the end of the 1970s the disenchantment of the public administration has become apparent: the dominating view is that the government has changed from the provider of public goods to a tax burden for the citizens. The public administration is regarded as “bureaucratic” in the sense of too big, inefficient and unable to improve (e.g. Heclo, 1981), and composed of structures that often develop an independent logic maximizing their own survival and growth. Furthermore, government systems and government workers are often seen as too slow, too inflexible, too focused on process, and excessively indifferent to results (Gurwitt 2000). Partly owing to this disillusionment and partly due to fiscal pressures there was a wave of public sector reforms throughout the world, since the 1980s. Many reforms in this wave share some characteristics that later have been known as New Public Management (NPM). NPM may be characterized as a move away from the standardized bureaucratic system towards greater flexibility, performance measurement, cost cutting (e.g. Hood, 1991; Boston, 1996), and more focus on results than on procedures (Minouge et al., 1998). This involves both a new philosophy of administration and a new pack of tools, which seek to enhance the efficiency of the public sector and the control that government has over it. The key hypothesis in the NPM-reform is that more market orientation in the public sector will lead to greater cost-efficiency for governments, without having negative side effects on other objectives and considerations. So, NPM have continuously supported the use of private sector management principles of planning, measurement and evaluation, the empowerment of midlevel management and the orientation of organizations to the needs of customers1. As a management philosophy, NPM look for to achieve efficiency gains by applying competition, as it is known in the private sector, to organizations of public sector, emphasizing economic and leadership principles. It addresses beneficiaries of public services much like customers (another similarity with the private sector) and, likewise, citizens as shareholders. Accordingly, NPM seeks to alter the way in which public servants are held responsible to the public, assuming that if citizens are aware of the performance of public services they will increase the political pressure placed on elected and appointed public servants, thereby enhancing both managerial and allocative efficiency in the public sector. A great deal of tools advocated by the NPM, and present in this wave of reforms, are forms of private involvement in the provision of public services. Among such forms we find different involvement degrees from the complete divestiture of former public services to the outsourcing of specific public services by private firms and a particular policy instrument known as Public-Private sector Partnership (PPP)2. Accordingly, an important policy-question resulting from the NPM debate is in what extent some of the so-called public goods can be better supplied by private providers, and if so, under what conditions. Given the above considerations, we will analyze the involvement of the private sector in providing public goods and services. The remainder of this chapter is as follows. In section 2 1
This also implies the implementation of specific performance indicators used in private organizations to create a performance-based culture. These indicators when applied in the public sector can function as targets leading to better efficiency and effectiveness. As in the private sector, increasing external-related outcomes can have a deep impact on internal control mechanisms, as managers and public servants become more responsive to their duties and more conscious and dedicated to serve their public customers. 2 For a characterization of PPPs, and an application to the infrastructures sector in developing countries, see Pessoa (2008). For a different point of view see Savas (2005).
Outsourcing and Public Sector Efficiency
169
we’ll make a short outline of what “public goods” are and the problems involved in their public production and provision, distinguishing their different types and emphasizing some examples of “impure” public goods. Section 3, focuses on market failures together with equity considerations, as the main reasons that configure the role of the public sector in providing impure public goods, as well as on the possibility of government failures. Section 4 discusses the ‘make’ versus ‘contracting out’ decisions, highlighting benefits and costs of outsourcing in the public sector. Section 5 describes the most frequent forms of private sector involvement in the provision of public goods, as well as the advantages and disadvantages of the different options. Section 6 carries out some comments on the need of regulation and the requirements needed in order to assure an effective regulatory framework. Finally, the paper closes with some conclusions for discussion and future research.
2. Public Goods vs. Private Goods According to public economics literature, we can distinguish between public goods and private goods3. In efficiency terms, the distinction between public goods and private goods is based on two characteristics — rivalry in consumption and excludability — rather than in the nature of the agent that provides for them. However, those characteristics are not of an absolute kind. Depending upon the degree of each characteristic, goods and services can thus be classified from the pure public good on one extreme to a pure private good on the other. Table 1. Private goods vs. public goods Characteristics Excludable
Rival Pure Private Goods
Not excludable
“Tragedy of the commons”
Non rival Exclusion would cause inefficiency Ex: highway “Pure” Public Goods
As is apparent in table 1, a “pure” public good is a scientific term used to describe a hypothetical good that is non-rival in consumption and, simultaneously, has a zero degree of exclusion4. In the real economy, pure public goods don’t exist. The goods that are nearer this concept are Defense and Administration of Justice. By contrast, a pure private good is a supposed good whose benefits are completely rival in consumption and which has simultaneously a perfect degree of exclusion. Economics teaches that the pricing mechanism of the market secures an optimal allocation of resources, if certain conditions are met. For private goods, these conditions are satisfied reasonably well over wide areas in the market economy. In these areas, the government normally does not have to get itself involved with matters of resource allocation. On the other hand, there is a wide consensus that the main economic role of government is providing “pure” public goods. 3
Although the notion of public goods has a long tradition in economics, going back to Marshall, Pigou and Wicksell, it was Paul Samuelson (1954) that first characterized and systematized the concept of public goods and the externalities related to them. 4 Usually, the degree to which a good is excludable is the extent to which the owner of the good can charge a fee for its use.
170
Argentino Pessoa
But, in practice, the government does not limit its action to “pure” public goods. The bundle of goods and services provided by the public sector is more varied. In fact, goods and services are supplied by the public sector because different reasons. On the one hand, government offers “pure” public goods, because if it left this function to the private sector there would be an under-production or no production at all – despite the fact that a significant demand exists. Here the reason is the fact that private investors will not be forthcoming because there is no way, or there is only an insufficient way in which they will be able to appropriate the returns on the investment in the provision of such goods and services. But there are also goods that are provided by political decision, and not because of its non-rivalry and no excludability. These goods could in theory be produced by the private sector although in practice they often are not. In fact, as is apparent in table 1 there is a wide variety of intermediate areas between two extreme cases in relation to various degrees of non- rivalry and exclusion. A good may be rival and not excludable as is the case known by the “Tragedy of the commons”. The classic example of such a type of goods is the common land shared by peasants during the precapitalistic era. The cost of one peasant choosing to graze an additional cow on the commons is shared by all of the peasants, but solely one peasant captures the benefit, with an inefficiently high level of grazing as the main result and a potential devastation of the common land. On the other hand, a good may by non-rival and excludable. For instance, if a road is not congested one car may utilize it with no loss of benefit for other cars. However a tollbooth may exclude traffic from such road unless payment is made. Likewise, access to a swimming pool has the potential of exclusion, but below capacity limits each person admitted may consume services without subtracting from the benefits of others. Here the market could be applied, but the existence of at least limited non-rivalry indicates that exclusion would cause inefficiency in the sense that one individual could be made better off by the consumption of the good without fully denying consumption to another. So, although theoretically an unambiguous line can be drawn between the two types of goods (private goods provided for adequately by the market and public goods satisfied through the government action), in practice we need to consider situations where government corrective action is required to secure an allocation of resources that is in line with consumer preferences. Certain goods are satisfied by the market, subject to the exclusion principle, within the limits of effective demands. But, if they are considered so meritorious that their satisfaction ought to be provided for through the government over and above what is supplied by the market and paid for by private buyers, they become a sort of “public” goods. This second type of goods provided by public entities is usually referred to as merit goods, whose typical examples include such services as free education, free health services, subsidized low-cost housing, etc5. Obviously, the satisfaction of merit goods cannot be explained in the same terms as that of “pure” public goods. The difference is essentially one of degree, but this distinction remains of fundamental importance. Although both are two public goods in the sense that the government provides them, different principles are applied. Pure public goods in general constitute a special problem caused by market failures, but the provision of merit goods, because it involves interference with consumer preferences, falls within the scope of consumer autonomy, as private goods 5
See Musgrave (1959, pp. 13-14).
Outsourcing and Public Sector Efficiency
171
are satisfied. Public provision for free educational or for free health services are typical cases in point. Such services are of direct benefit to the particular pupils or residents, but apart from this, everybody enjoys from living in a more educated or healthier community. Thus, goods that come into view of society as merit goods may include substantial positive externalities. But there is another type of goods and services that we must refer to. This is the case of goods and services that are more or less private (in the sense of excludable, appropriable) in nature but their provision and fair distribution is viewed as essential to public interest. They are associated to capital-intensive projects and they have significant ongoing maintenance requirements. As the words indicate “public utilities” such as water supply, gas, and electricity fit this last category of goods probably best, but there are plenty of other examples, like telephone network services, certain modes of transport such as rail, etc. In spite of not being generally named as merit goods, usually the government has had a role to play in the provision and implementation of these goods and services. As a consequence, as shown in Figure 1, both the private sector and the government have an overlapping zone from which some goods are provided to the general population. We may call relevant goods provided in this area as ‘impure public goods’. Most public sector reforms have occurred in the set of activities that deal with this type of goods. We argue that outsourcing must be handled with care if it deals with goods and services included in that overlapping zone.
Private goods: • Rival • Exclusive
Private sector
Impure public goods
Public goods: • Non-rival • Non-exclusive
Governme nt
• •
Merit goods Public utilities
Figure 1. Impure public goods.
The above considerations were motivated by the search of efficiency. But there is another reason why the distinction between private and public goods using two characteristics is not of an absolute kind: the need of equity. Looking at equity, a society might be interested in correcting the final allocation of goods and services as it closely depends on the initial distribution of wealth. Therefore the government might want to correct these inequities by a policy which directly benefits the poorer part of the population, e.g. providing services at a low cost or for free to the poorest part of society. Figure 2 clarifies the basic nature of goods in whose provision government is involved, by using two criteria: efficiency and equity. Efficiency is taken on the horizontal axis while equity corresponds to the vertical axis. It is apparent that goods of type B, usually classified
172
Argentino Pessoa
as merit goods (v. g., education and health), would be located in the intermediate position, between pure public goods (type A), which are provided by the government mainly because of market failures, and more rival goods, like telecommunications (type C), whereas the provision by the government cannot be exclusively justified in efficiency terms6. 100% TYPE C (INFRASTRUCTURES AND
Equity
UTILITIES)
Electricity; Water; … TYPE B (MERIT GOODS) Health; Education;
…
0%
TYPE A (PUBLIC GOODS) Defense; Judicial System; …
Efficiency
100%
Figure 2. Different types of goods provided by government.
It is, however, quite arbitrary to draw the frontier between “pure” public goods and merit goods, as well as to trace the boundary between merit goods and type C goods, because that depends on value judgments. Consequently, the amount of goods provided for by the government is unclear in a political process influenced by tradition, history and other influences, which entail specific social values. For instance, while the noninterventionist tradition that prevails in the US usually claims that merit goods and public utilities must be provided for, to a significant amount, by the private sector, the prevalence of social values in the North European Countries tends to extend the desirable field of merit goods. Goods and services of type C deserve some additional considerations. All type C goods include natural monopoly characteristics arising from persistent economies of scale and scope. These characteristics mean that competition is unlikely to develop, or if it develops, it will be uneconomic because of the duplication of assets. Although technological advances, notably in telecommunications, have reduced some of the natural monopoly characteristics in utilities, permitting economic competition in certain areas of service delivery, nevertheless each one of the utilities retains some natural monopoly features. As a consequence, privatization of these industries, in whole or in part, threatens the introduction of private-sector monopolies that will exploit their economic power in the market place, leading to supernormal profits and consequently reduced consumer welfare. Accordingly, after privatization consumers may suffer from both no or limited choice of goods and services and face monopoly prices. It is well established in the economics profession that the (Pareto) efficient amount of output in an industry occurs where price equals marginal cost. However, a monopolist 6
Obviously, this classification of goods and services (types A, B and C) has practical purposes in the context of this chapter.
Outsourcing and Public Sector Efficiency
173
produces where marginal revenue equals marginal cost and thus produces too little output. It would seem that regulating a monopoly to eliminate the inefficiency is pretty easy — all the regulator has to do is to set price equal to marginal cost, and profit maximization will do the rest. But in practice, this analysis leaves out one important aspect of the problem: at such a price, the monopolist would make negative profits. Price
Dem
MC
AC
PAC A PMC
YAC YMC
Output
Figure 3. A natural monopoly.
An example of this is shown in Figure 3. Here the minimum point of the average cost (AC) curve is to the right of the demand curve, and the intersection of demand and marginal cost (MC) lies below the average cost curve. If a natural monopolist operates where price equals marginal cost, then it will produce an efficient level of output, YMC . But in spite of being efficient the level of output is not profitable. If a regulator sets this level of output, the monopolist would prefer to go out of business because it will be unable to cover its costs. If it is required to produce an output where price equals average cost, YAC, then it will cover its costs, but it will produce too little output relative to the efficient amount. This kind of situation often arises with public utilities. For example, in an electricity company the technology involves very large fixed costs — creating and maintaining the electricity delivery wires — and a very small marginal cost to providing extra units of electricity — once the wires are laid, it costs very little to drive more electricity down the wire. Similarly, a local telephone company involves very large fixed costs for providing the wires and switching network, while the marginal costs of an extra unit of telephone service is very low. When there are large fixed costs and small marginal costs, one can easily get the kind of situation described in Figure 3. Such a situation is referred to as a natural monopoly. If allowing a natural monopolist to set the monopoly price is undesirable due to the Pareto inefficiency, and since forcing the natural monopoly to produce at the competitive price is infeasible due to negative profits, what is the right way? Different countries have adopted different approaches. In some countries the government provides the telephone
174
Argentino Pessoa
service while in others private firms, which are regulated by the government, provide it. Both approaches have their advantages and disadvantages. For example, let us consider the case of government regulation of a natural monopoly. If the regulated firm is to require no subsidy, it must make nonnegative profits, which means it must operate on or above the average cost curve. If it is to provide service to all who are willing to pay for it, it must also operate on the demand curve. Thus the natural operating position for a regulated firm is a point like (PAC, YAC) in Figure 3. Here the firm is selling its product at the average cost of production, so it covers its costs, but it is producing too little output relative to the efficient level of output. Government regulators set the prices that the public utility is allowed to charge. Ideally these prices are supposed to be prices that just allow the firm to break even — that is, to produce at a point where price equals average costs. The problem facing the regulators is to determine just what the true costs of the firm are. Usually there is a public utility commission that investigates the costs of the monopoly in an attempt to determine the true average cost and then to set a price that will cover costs. To end this section some conclusions are mandatory: first, it is obvious that in practice the provision of goods and services faces a diversity of situations, far from the simplistic dichotomy between private goods and public goods; second, given the diversity of situations, the solutions to inefficiencies are not simple; third, as the case of natural monopolies shows, even after implementing a reform that relies on higher involvement with the private sector, government must retain a role to play.
3. Market Failures and Equity Considerations vs. Government Failures a) The role of the market It is almost consensually accepted that the dynamic function of markets in improving efficiency and innovation is the main factor behind the superiority of decentralized economies as compared to the centrally directed economies. The main reason is that the market system translates consumer preferences into market demand in a discernible and efficient way. The type and amount of goods and services being produced depends on the utility they offer to consumers, as compared to the utility that consumers obtain from other goods and services that they could purchase for the same cost. On the supply side, there is the assumption that when a product is produced inefficiently and therefore too costly, competitors that are more efficient can and will (depending on the entry conditions) supply the product at lower prices, and the inefficient firm will either run to produce more efficiently, or in the end it will be driven out of business. In a similar approach, the quality of goods and services is likely to be protected by the market-mechanism. If a business fails to maintain and increase the quality of its products and services, competitors with a better price-quality ratio will force the business to keep up and improve the quality of its products; otherwise it will lose customers. b) Market failures, equity and regulation But in spite of the above-mentioned benefits of the market, there are cases when market forces cannot secure optimal results, and so we are faced with the problem of how the
Outsourcing and Public Sector Efficiency
175
government can interfere to obtain a more efficient resource allocation. The role of the government can be described as consisting of the following: •
•
•
7
Overcoming market failures. Where needs are likely to go unmet because of market failure, there is a role for the government to step in. For goods of type A and type B, market failure means essentially an under provision, or no provision at all. This can occur when the social benefits of services exceed the private benefits with a resulting sub-optimal provision, which normally calls for government provision. As one example, people typically contract sexually-transmitted diseases (STDs) accidentally. By bearing some of the cost of detecting and treating STDs, governments confer benefits not just on the individuals treated, but also on those who may otherwise be at risk of infection. The same can be told about vaccination programs and other forms of diseases control. Another example of market failure in developing countries is the education of girls. Many families fail to see any benefit from sending girls to school or are averse to give up the household labor, or income, they make available. However, as a social investment, girls’ education is crucial because it is associated with improved opportunities for them to live longer, richer, and more rewarding lives — and with better health and social outcomes for their children. Thus, by encouraging the education of girls, through educational scholarships or consciousness-raising campaigns, governments can benefit both girls themselves as well as their families and communities. This example may be extended to the health sector, as the welfare of infants depends heavily on the health status of the mother. For goods of type C market failures mainly relates to the existence of co-ordination malfunctions induced by scale economies. There is the case of external economies that arise when a new highway is built or as the size of a telecommunication service increases. The market failure is that at a given point in time, current prices may not convey the information about prospective expansion that is relevant to attain a lower cost of production (Scitovsky, 1954; Chenery, 1960)7. Equity. To provide goods B (health care or education) and C in rural areas tends to be particularly difficult, and generally unprofitable from a private viewpoint. Not only rural populations are often small or dispersed but also private providers are often scarce or nonexistent8. The public sector is best placed to provide a safety net for citizens who cannot pay market prices for health or education. However, this can be achieved by providing services directly or by creating incentives for the private sector to carry out the task. Regulation. Implementing appropriate regulations to ensure quality and controlling costs. Consumers will usually act as a force for quality, but only if they have sufficient information. Governments can do something as important providers of this information. The existence of asymmetrical information is overwhelming in the
Additionally, the government must deal with other examples of market failures, such as the problem of adverse selection and moral hazard, associated to the privaty run insurance schemes, which leads to an unequal coverage of health care services. Private insurers will only include good risks in their schemes. This behavior makes risk pooling among a society difficult and leaves the bad risks to the public sector. 8 Government clearly has a role providing services here, but it can also act in other ways. It can place obligations on private providers to provide broader access when they occupy a monopoly position or consider subsidizing access to private systems for disadvantaged groups.
176
Argentino Pessoa health sector. When drugs are sold in the open market, the manufacturer is usually better informed on the efficiency and safety of the drug than the purchaser. To address the described market failures the government usually reacts doing something to minimize the effects of asymmetric information, e.g. the official registration of health professionals and official recognition of drug quality. Quite frequently, governments act to put a ceiling on the fees private sector providers can charge. This is controversial, as it causes a market distortion, and should be done with care. However, restrictions may be necessary where there is little competition, no parallel public provision, or where consumers are relatively poorly informed about their needs and the quality of the provision. Pharmaceutical cost is an area where the potential for excess profits is high, and control may be necessary, but in goods of type C the problem is also real.
c) Government failures The above reflections on the role of government have been mainly derived from theoretical considerations. In practice, however, some of the aforementioned points have to be equated with the possibility of government failures. Some government failures result from the absence of the corrective function of prices. For private goods, prices not only reflect relative supply and demand, but also signal interesting profit opportunities, best practice cost, quality and delivery performance, etc. A careful analysis of information from prices is therefore important. However, there are two basic conditions for the ‘disciplining’ action of prices to function well: transparent and condensed prices and a high degree of competition. The public sector has no one of these conditions. For public organizations, prices usually do not contain the correct and needed information. Since these prices are a result of political regulation, they only partially reflect or do not convey at all information on scarcity, utility, quality, and efficiency. Obviously there is a cost price for producing public goods and services. However, the resources for the production of public goods and services come from taxes that are collected largely independently from what they are used for. The allocation of these resources is a matter of policy-decisions and the price that consumers of public goods and services pay is often only slightly related to the actual costs. Furthermore competition is nonexistent, not only owing to technical reasons, such as the scale of production or the need of universal access, but mainly because of an explicit choice rendering in some cases a “governmental monopoly” structure as the most efficient delivery provision. Government failures take place both at the supply-side and at the demand-side of public goods and services. On the demand side it will be difficult or onerous to determine what the real demand for these goods and services is. When calculated, it is generally determined by estimations of the needs for public goods and services, in combination with certain political values and equity considerations. But, unfortunately it is neither easy nor quick to translate these estimations into actual policy implementation. Since there is neither competition nor a price mechanism as a disciplining characteristic, on the supply side the production will be
Outsourcing and Public Sector Efficiency
177
determined by the allocation of budgets, and so there will be little external incentive to lower costs, improve quality, or to satisfy an increasing demand9. There are several ways to deal with these aspects of public failure, but recent public sector reforms rely almost exclusively on two. The first mode is by privatizing certain public services forcing them to operate in a market-like way, in reaction to the perception that government agencies failed constantly to provide high quality services. For example, Kansas privatized its entire child welfare system, in part in response to a widespread sense that under the publicly managed system, children were remaining in foster care too long after removal from their families (Gurwitt 2000). But this first way has been prominent in the utilities (electricity, telecommunications, water, etc.) sector. However, to privatize completely these services requires the setting up of a sophisticated regulatory framework with wide-ranging functions: avoiding private monopolies10, enabling “fair” competition between the incumbent and new entrants, regulating prices and access provision, securing certain national strategic guarantees, etc. The difficulties in liberalizing and privatizing public utilities in most European countries – e.g., electricity, telecommunications – illustrate well the fact that a simple transfer from public to private provision is generally speaking for an insufficient guarantee for increasing efficiency and quality on a long term sustainable basis. The other way consists of segmenting functions of the public services, distinguishing between the ones that by their own characteristics must be dependent on the public sector from the others that can be contracted to the private sector. In other words, public bodies need to assess their functions according to their relevance to their core values providing for them, and contract out all the others11. The remainder of this chapter deals with this second way of solving public failures.
4. Outsourcing in Public Services The outsourcing of activities formerly done by the public sector was popularized by the discussion around the NPM. For the proponents of NPM, outsourcing of public services is typically viewed as a way to increase accountability by turning over choices to individuals in the market place and, consequently, as a means of maximizing economic efficiency — reducing government costs while increasing the scope and quality of service delivery by transferring (or “returning”) government functions to the private sector (e.g., Butler, 1985; Donahue, 1989). But on the other hand, some scholars argue that it has the potential to produce considerable fraud and corruption, if managerial control by the public sector is weak12.
9
Of course, there is some intra-public-sector competition. But this competition is about a bigger share of the budget, and so it is generally more related to internal politics than to a perceived increase in demand for goods and services. 10 As was theoretically argued above, in section 2. 11 See Prahalad and Hamel (1990) for a managerial perspective on this subject. 12 Additionally, there is a fear that outsourcing results in diminishing citizens’ legal rights, as “government sovereignty is extended to private contractors” (Wiesniewski, 1991, p. 378), and in serious problems with accountability. For example, access can be denied and complaints can be ignored. As we’ll see in next sections, regulation can provide some protections against this situation.
178
Argentino Pessoa
However, the risks that outsourcing pose to clients and governments need to be equated with the advantages of promoting innovation through outsourcing. Given that outsourcing of services is more complicated than the purchase of goods, the proper balance between outsourcing and maintaining direct control over program operations (Blank 1999) depends on the nature of the service being provided, and on a net benefit to consumers. So, with these considerations in mind, governmental agencies, as well as private companies, need to consider the costs and benefits of contracting out versus in-house provision. Let’s begin with the potential benefits of outsourcing.
Benefits The strategic management literature suggests that outsourcing can contribute to competitive (or collaborative) advantages in three different ways (Bovaird, 2004): first, providing economies of scale in the provision of certain services; second, providing economies of scope or the ability to exploit more fully the complementary capabilities and competences which exist in the partner organization(s); third, providing opportunities for mutual learning between partners which may be intended to lead to a long-term dynamic process or interchange. If competition in the private sector exists, we can expect the following benefits: •
•
•
•
Improving quality and customer service. Public services are recognized the world over for low-standards of customer care. In recent years, many business sectors have been revolutionized by a new customer-focus. Private providers must develop their businesses and, in most situations, this involves retaining existing customers as well as attracting new ones. For this purpose they need to be highly innovative and also to learn from their competitors, thus aiding the transmission of best practice. One expects that outsourcing in public services can benefit from similar gains. Investing in research and development. Organizations have great difficulty in learning, and they seldom question the underlying basis of their own problems. Especially in the public sector, organizations are often depicted as lacking in innovation and intrinsically resistant to change, stressing conformity instead of creativity, defending the status quo instead of striving for change and improvement. The involvement of the private sector can be a stimulus to carry out research and to develop new techniques13. Improving management standards. Some observers argue that because in the private sector the staff is usually better paid and motivated, the management standards are generally higher and so business can transfer important skills for a great lot of sectors including the ones of health and education (Van Slyke, 2003). Developing new services and market-based systems of rationing. The private sector has an essential role where demand is expanding or the patterns of demand are changing. When these changes happen, it is an increasingly important provider of higher education, for example. Skills development and professional development, for
Outsourcing and Public Sector Efficiency
•
179
instance, can be funded privately, either directly at the level of the firm or through reimbursement mechanisms. It is inevitable that some costly procedures, perhaps of limited efficacy, cannot be funded universally. Here the association between the public sector and the private sector can be very useful: The government is able to set the context of what is considered an essential service available for all. The private sector can control access to other services using the price mechanism. Filling the ‘capability gap’. Another reason for outsourcing is the need to fill a ‘capability gap’, that is, contracting out functions where ‘in-house’ capacity is limited. This may be particularly required for capabilities that are highly specialized as the formation of teachers or the management of schools. With this regard, it is usually argued that contracting-out allows savings on the long-term costs of hiring specialized experts, who may be required only in very specific periods of time being under-occupied for the rest of the time (Sanger, 2001).
b) Reducing costs Alongside the above attributed benefits, the cost side of contracting out should not be overlooked. Proponents of outsourcing in public services argue that contracting out is synonymous with reducing the size and effects of government. This theory suggests that contracting out saves money as the positive pressures of competition force organizations to find ways to work more efficiently. It is basically this idea that has motivated State and local governments to turn to private providers for a wide range of services, from routine matters such as road maintenance and garbage removal to sensitive undertakings such as fire protection and the operation of correction facilities (Sclar 2000). The positive effects of competition are thought to hold true for competition broadly, not only for competition by the private sector or by for-profit corporations. In fact, what matters most is the extent of competition rather than simply whether the public or private sector is the provider (Kettl 2000; Donahue 1989; Osborne and Gaebler 1992). So, the above argument is not valid if the provider is a monopolist, and so the decision about outsourcing must be based on a more pragmatic approach: comparing the costs of in house production with the costs of outsourcing the service provision. Even if we assume the existence of effective competition and well-functioning markets, efficiency calls for that the government must be a smart buyer, a skilled purchasing agent, and a sophisticated examiner of the goods and services it purchases from the private sector. All of this is not for free. Although the policymakers tend to consider only the production costs, there are many other costs that need to be accounted for. First, contracting out will increase transaction costs, including both contracting and monitoring costs14. As Williamson (1979) argues, given the governance structure or institutional context within which governments transactions are negotiated and executed, the contracts with program providers are likely to be complex rather than simple. In such environments, the transaction costs of designing, monitoring and enforcing complex contracts
13
14
Nevertheless, the government must retain an important role in financing basic research that can produce important building blocks for subsequent applications that may improve well-being and for which short-run commercial gains are not apparent. See Coase (1960) for the economic framework in the “make vs. buy” decisions, and Donahue (1989) for its practical applications.
180
Argentino Pessoa
are very likely to be high15. In the light of this, governments must be wary of getting caught up with outsourcing processes that compromise the government’s ability to secure and protect the public interest in the long term. Second, the costs related to the loss of monopsony purchasing power and the social costs arising from equity problems (Robinson, 1990; von Otter and Saltman, 1992) could also be significant16. But, these direct costs are not the only ones that must be controlled in the public services provision. In this specific sector, contracting-out requires maintaining minimum levels of qualified staff in-house in order to specify employment terms clearly and in a way that fits the specific purposes of the activity, or to correct the service provided externally in the event of provider failure. Hence, in the costs point of view contracting out is justified only when one can expect to lower the sum of production costs and the costs of managing the relationship between government and the provider of goods and services (Globerman and Vining, 1996). Contracting has a potential for lowering the first set of costs, but these savings could be more than offset by increases in governance and transaction costs. Where the complexity of the task is high, contestability or market competition is low, and asset specificity — and thus investment risk — is high, governance cost could prove to be tragically high for governments. Because purchase of services is more complicated than acquisition of goods, the former is more frequent than the latter. Of course, there are regional variations: in the USA outsourcing is more generalized in public services than in Europe. In the USA it is used for all the types of services: direct services, support services, and services delivered to third-party clients. Local governments outsource direct services such as solid-waste collection, street repair, street cleaning, snow removal, and gardens maintenance. The average American city contracts out almost a quarter of its common municipal services to the private sector. The average American state contracts out 14 percent of its activities, including the operation of some prisons17. c) Final assessment Research on the quality of outsourcing in public services is very limited, but, like that on cost savings, it appears to give mixed results. A number of experts argue that the different sectors will have different relative strengths, depending on the primary goals of services (Osborne and Gaebler, 1992). The empirical evidence, limited though it is, suggests that the quality of services contracted out might generally be the same or somewhat higher than when these services are provided by the public sector18. With a constant quality, if outsourcing is done in the right fashion, it enables governmental agencies to benefit from the combined force of specialization and competition, and therefore to reduce their costs substantially. The savings provided by adopting outsourcing, seem in some cases significant. Overall, it has been estimated that the benefits of 15
As argued by Van Slyke (2003), outsourcing supporters seldom acknowledge that contracting out leads to additional public management costs such as developing program performance measures and evaluation tools, developing and maintaining management capacity to monitor and oversee contractors, and so on. 16 In addition, some other impacts should be taken into account, too. As Mills (1995) argues, the introduction of contracts may both lead to a fragmentation or lack of co-ordination within the broader public service system, and could have an impact on staff resources with a drain of key personnel to the for-profit providers. 17 See Savas (2005) and the references therein. 18 However, some experts note that these analyses may be somewhat biased in favor of the private sector because public reform often occurs only when public services are particularly ineffective, providing a point of comparison that might not be typical of public-sector provision. The results of several research efforts reflect this complicated picture of service quality.
Outsourcing and Public Sector Efficiency
181
competitive contracting out may allow reductions in costs by as much as 10-20 percent, at the same time as constant quality is maintained19. There has been a lot of experience with government contracting out social service provision to private firms, and there is an extensive literature that examines the serious problems with doing so (Miller, 2001; Wisniewski, 1991). Looking at the private sector in general, and based on the above and on other references (Berman, 1997; Blank, 1999; de Bettignies and Ross, 2004; USGAO 1997; Eggers and Ng 1993; Osborne and Gaebler 1992), we can summarize the strengths and weaknesses of contracting out public services (table 2). Table 2. Pros and cons of contracting out Pros
Cons
Reducing costs for the same level Private providers respond to the population’s willingness to pay for public services. As a result, they will serve those groups in of quality the population who are most willing to pay, such as affluent urban residents. The result will be increased inequity in access Filling the “capabilities gap” and use of public services. Because of lower willingness to pay, private providers will The replacement of direct, undersupply socially desirable services, such as immunizations hierarchical management and personal preventive care. This will worsen allocative structure with contractual relationships between purchasers efficiency in the corresponding sector. Driven by the profit motive, and because they have significant and providers, which will control over demand, private providers will take advantage of increase: clients by supplying more than is required. This is particularly Not only the transparency of significant in health care services. This is inefficient and may prices result in health-weakening actions. But also competition, Private providers can also take advantage of clients by providing Which in turn will lead to a gain low-quality services, which may result in welfare losses. The actual effect of these four major worries is as greater as there in efficiency. is lack of competition.
As is apparent from the analysis of table 2, several factors come into play in reaching efficient decisions20. Factors like the need to fill a “capability gap” or to reduce costs would advise the contracting out of some functions. If this is the case, public bodies face the need of, at least, maintaining quality constant. Such decisions should be based on the identification of the agency’s core functions and consideration of the costs and benefits of contracting out versus in-house provision21. This means that outsourcing in public services 19 20
21
See Domberger (1998, p. 163). Many other concerns have been highlighted along time. McKean and Browning (1975) discuss how and why “overlooking any relevant objectives could lead to poor choices”; Grizzle (1985) examines the serious attention that needs to be given to selecting output measures in terms of multiple criteria: their relevance, validity, reliability, accuracy, comparability, and cost. The above considerations about the ‘make vs. contracting-out decision’ in the provision of public goods were stylized without considering the level of development of the countries where decisions are taken. However, in a developing country context there are other additional problems arising from asymmetry of information. Many private providers that deal with the Governments of developing countries come from more developed countries with more experience of consumer preferences (Pessoa, 2008).
182
Argentino Pessoa
may not have the result of creating what has been termed the “hollow state” (Bovaird, 2004). A ‘hollow state’ is one having the double sin of low capacity and weak legitimacy — the latter resulting from, or worsened by, the shrinkage of the governmental core functions. The main reasons in opposition to outsourcing, and more generally to the private sector involvement, can be minimized if there is a competent regulation. However the corrective effect of regulation also depends on the specific form of private sector involvement. So in the next two sections we’ll deal with these issues, beginning with a review of the typical forms of private sector involvement in the public sector and after that, we’ll make some considerations about the need of regulation.
5. Forms of Private Sector Involvement in Public Services As highlighted in the previous section, the provision of public services has undergone major changes in the last two decades with many developed and developing countries choosing to move away from the traditional public sector model of service provision and to introduce private sector participation. The involvement of the private sector in public services has followed, in general, six basic forms ranging from short-term service contracts to divestiture. •
•
• •
Short term service contracts. In this option, specific tasks, usually everyday maintenance jobs, are contracted to the private sector, but overall services management remains within the public sector. This type of contracts has been implemented in many countries with good records of success and is often seen as a first step towards a more definitive collaboration. In order to define the compensation to the private sector partner, two types of contract are frequent. In a quantity-based maintenance contract, the remuneration of the contractor is based on unit prices defined in the contract and the quantities are measured on site. The other type — performance-based maintenance contract — is derived from the previous type of arrangement, by shifting the focus from administration (maintenance activities and resources) to certain performance conditions valued by the users. In this case, the payment is based on a fee directly stated in the agreement and linked to performance indicators. Management contract. A management contract is an arrangement by which a private company is entrusted with various types of tasks relating to the organization and maintenance operations, usually performed by the public authority. This type of contract involves the payment of a fee to the private company. Usually, the function of the private firm is to respond to day-to-day routine maintenance needs by contracting private companies, on behalf of the public entity. Lease. In this form, a private company rents the assets of a utility, and maintains and operates them, in return for the right to revenues. Greenfield projects. In this option, very usual in public works, the private sector develops, finances and operates bulk facilities. Under a BOT (Build-OperateTransfer) the responsibility of the concessionaire is not limited to the operation and maintenance of the infrastructure, but it also includes a component of initial
Outsourcing and Public Sector Efficiency
•
•
183
construction, upgrading or major road rehabilitation. Massive investment and consequent mobilization of private funding sources are therefore required from this company, which is to be repaid from the revenue collected from service users (usually tolls). The BOT arrangement stresses the responsibility of the private entity during construction and operation of the infrastructure and the transfer of the assets to the public entity at the end of the operation period. The high initial investment required from the private sector and the consequent long concession period turn the distribution of risk between the parties into a key element of success in such schemes22. Concession. In a concession a public entity owns the assets, but it contracts with the private sector for operations, maintenance and investment. For instance, a road concession is an arrangement under which, the owner of the road, delegates to a private entity (concessionaire) the responsibility for providing and maintaining a specified level of service to road users in exchange for the right to collect revenue from those users. Besides the issues inherent in a concession agreement, an operation and maintenance concession is similar in scope and approach to what is required and negotiated in a typical operation and maintenance agreement between private parties under a BOT-type arrangement. A concession is more typical for goods of Type C, but there are other cases where such an option is applied to goods and services of type A and B: for instance in delivering educational services. This was the case of the city of Bogotá, Colombia, in 1999, which launched an educational program without precedent in the history of the country. The program, called Concession Schools, consisted of public education in 25 schools provided by the private sector for a period of 15 years. The public sector provided the infrastructure, selected the students (from income strata 1 and 2), and paid a pre-agreed sum per full-time student per year (approximately $1,200,000 Colombian pesos, according to Villa and Duarte (2004))23. Divestiture: an asset or public enterprise is either partially / totally sold, or shut down. Where state-owned enterprises are abundant, the word “denationalization” is frequently used instead of divestiture.
Table 3 shows some illustrative examples and the advantages and drawbacks of the different options. As is apparent from the analysis of the table, if the principal reason for private sector participation is the large potential for gains in efficiency in the public sector, it may be expected that projects with higher level of private sector involvement deliver more efficiency gains. However, the consequent risk of failure grows correspondingly. One the other hand, options that yield higher social benefits also tend to demand a higher level of government commitment, and also require a better prepared institutional framework.
22
Many variations on this type of contract have been implemented with a consequently growing number of acronyms used to label them (BOOT, BOO, BTO, DBO — Design-build-operate). 23 For details and an assessment of the Colombian Concession program, see Rodriguez (2005).
184
Argentino Pessoa Table 3. Pros and cons of the different private involvement options Option
Service contract
Management contract
Lease
Greenfield projects
Concession
Divestiture
Typical duration
6 months to 2 years
3-5 years
10 to 15 years
Examples Several contracts in the water sector of Mexico City: i) consumer census, mapping the network, metering; ii) regularization of billing; iii) loss detection and reduction. Waste collection in:Caracas, Seoul, Bangkok, Jakarta, Lagos
Water supply in Guinea (Conakry and 16 other towns, in 1989)
Design-build-operate Solid Waste in Hong 15 to 30 Kong: for refuse transfer years stations and a chemical waste plant. Water and sewerage concession began in 25 to 30 Manila in 1997; years Concession schools in Colombia, Bogotá, 1999 Indefinite, Privatization of utilities but may be like electricity, limited by Telecommunications, a license etc.
Pros
Can inject good technical expertise
Gains in managerial efficiency Commercial risk borne by the private sector, giving strong performance incentives Good way of getting efficient delivery of bulk services, with private investment
Cons
Unlikely to greatly improve performance where overall management is weak
Gains can be difficult to enforce; public entity remains responsible for investment Administratively demanding; public entity remains responsible for investments Not a good solution if supporting distribution systems are in bad shape, or traffic levels are uncertain
Potential for high efficiency in operations and investment
Requires considerable commitment and regulatory capacity
Potential for high efficiency gains
Requires credible regulatory framework
6. Conditions for Private Sector Involvement and the Need of Regulation Both macroeconomic and microeconomic conditions affect the involvement of the private sector in the provision of public goods in a specific country. Concerning the macro level, political factors are important: without an overall political environment supporting both private for-profit and not-for-profit activities no significant participation of the private sector in public goods provision can be established. In countries where civil society and/or the private sector are discriminated, the government will remain the dominant supplier of public goods and services. Concerning the micro-level, several conditions are also important. First of all, there must be an interest and a commitment of some individuals and firms to make the
Outsourcing and Public Sector Efficiency
185
involvement happen. If there is an interest and an acceptance of the different partners to be involved, then one has to look at the capacities of the different actors. In this respect, we have to consider not only the skills of the staff to provide specific services, but also the financial availability for an engagement in service provision and the overall organizational and management structure. Ultimately, the sustainability of the reforms and the ability of the public sector to use money more effectively in leveraging private money will depend significantly on the political commitment to design and carry out effective regulatory policies. Although regulation is above all fundamental in divestiture of utilities, it is also important in other forms of private involvement on provision of basic public services. Within the framework of NPM reforms, in order that the new, privatized market be efficient and equitable, it must be well regulated so that it operates in ways that maximize social returns. Justifications for expanded outsourcing of public services clearly recognize this: “Capacity in the government to contract out and to regulate is required” (World Bank, 2001, p. 17); “strengthening the capability of the state to develop and supervise health and education systems is thus critical” [and so] “major capacity and institution-building of public sector agencies is required to fulfill this role” (World Bank, 2002, p. 18). Accordingly, the need of efficiency calls for the existence of independent regulatory bodies. So, the main changes in the last two decades in the provision of public services, both in developed and developing countries, call for strong and competent economic regulation, in order to ensure that the interests of all parties are protected. Such protection is necessary first and foremost, to defend the customers’ interests but also those of the public and private 24 parties to a contract . The role of institutions in charge of carrying out regulatory functions is even more important in developing countries than in developed ones. In the former, owing to several reasons that affect differently the two groups of countries, a much more intrusive and demanding form of regulation is required. Besides the reduced educational level of the population and the scarcity of infrastructures, which may restrict the availability and circulation of information, many other reasons affect the effectiveness of the regulation in developing countries. However, in developing countries the need for regulation is even more vital, because they are usually characterized by non-competitive industry structures and/or lack of capital market discipline. In such environments, too little market information is revealed and information asymmetries are overwhelming. In addition, regulators in developing countries face other specific challenges, when large portions of the customer base for infrastructure services are poor and unconnected, tariffs are being kept artificially low, baseline information for decisions tends to be limited or unreliable and the regulators have difficulties in establishing their credibility and in implementing sound governance arrangements. As already argued (Pessoa, 2008), to be effective, regulators are required to fill three qualities: competence, this quality being measured by access to technical expertise in a wide variety of areas; independence, both from government interference and from capture by service providers and interest groups; and legitimacy, i.e., both long-lasting by existing legal principles and practices and being transparent and accountable. Many, if not all, regulators 24
Since the beginning of utility reforms in the late 1980s – early 1990s, it is estimated that about 200 regulators in some 130 countries have been granted the functions of regulating public services such as telecommunications, water, and electricity (World Bank, 2004).
186
Argentino Pessoa
lack one or all of the qualities required for effective regulation. These deficiencies can result from different reasons, including limited resources, repeated political interference in regulatory decisions, difficulty in attracting and retaining competent staff, and short or no history of performing regulatory functions. All these deficiencies are particularly apparent in the case of countries emerging from social conflict or where the political environment makes it difficult to set up any kind of independent institution. Where there is lack of independence we can’t prospect either great legitimacy or competence. This lack in turn limits the capacity of agencies in charge of regulation to act as effective regulators, i.e. to promote adequate levels of investment in the regulated sector through the setting of tariffs that recover costs without depriving part of the society from using the services, to attract private investment and/or to monitor the public sector for superior performance. Of course, regulatory functions can also be contracted out, but there must be a core of functions that governments cannot give up.
7. Conclusion As it is well known, the role of governments in formerly developed countries started from the very limited scope of Adam Smith’s “small government” that provides only defense and the administration of justice. However, it is widely acknowledged that the relative share of government fiscal activities (in short, the public sector) tended to increase steadily in the national economy towards a big government. However, if one looks at the role of the government’s performance in practice, one has to recognize that, due to allocative inefficiency, operational inefficiency and equity problems, sometimes it poses more problems than solutions. Additionally, if public services are provided for free and are accessible, then the quality is often so bad that people prefer to go to a private provider and to pay fees with a certain guarantee of a quality treatment. But if people prefer a private provider even if they have to pay fees, a question arises: Why not “contracting out”? The answer to the above question must take into account that outsourcing services is not so easy as contracting out goods. This explains why for instance, in education, there is considerable contracting out to the private sector, for things like building schools or running a cafeteria, but these experiences with well-defined school inputs have little to do with the core functions of educational public agencies, where outsourcing is much scarcer. Given the possibility of outsourcing, public bodies are confronted with the decision of whether they should produce a service internally or contract it out. The choice between the two options must be founded in an analysis that equates benefits with costs. As argued in section 4, it is not easy to compute the total costs associated to outsourcing. Particularly, the costs with contracting and regulation are generally overlooked, as well as the need of an augmented regulation when outsourcing is extended to another service is usually ignored. The design of rules and regulation and their enforcement are crucial in efficiency and equity grounds, where government decides contracting out services or involves itself in a partnership with the private sector. If the public interest is to be secured, outsourcing requires that the public sector be equipped with staff with the relevant contract-management experience, policy expertise, negotiation, bargaining, and mediation skills, oversight and inspection capabilities, and the necessary communication and political skills to manage programs with third parties in a complex political and economic environment. If this
Outsourcing and Public Sector Efficiency
187
capability exists, outsourcing in public services may not have the result of creating what has been termed the ‘hollow state’ (Bovaird, 2004); nevertheless, outsourcing must not be considered as a panacea to the problems posed by the public provision.
References Berman P. (1997). Supply-Side Approaches to Optimizing Private Health Sector Growth, in W. Newbrander, (ed.). Private Health Sector in Asia. Issues and Implications, Chichester. Blank, R. M. (1999). When Can Public Policy Makers Rely on Private Markets? The Effective Provision of Social Services. Working Paper Series, Paper 7099, April 1999. Cambridge, MA: National Bureau of Economic Research. Boston, J. (1996). Public Management: The New Zealand Model. Melbourne: Oxford University Press. Bovaird, T. (2004). Public-Private Partnerships: from Contested Concepts to Prevalent Practice. International Review of Administrative Sciences, Vol.70 (2), pp. 199-215. Butler, S. M. (1985). Privatizing Federal Spending: A Strategy to Eliminate the Budget Deficit. New York: Universe Books. Chenery, H. B. (1960). Patterns of Industrial Growth, American Economic Review, vol. 50, pp. 624-654. Coase, R. H. (1960). The Problem of Social Cost. Journal of Law and Economics, vol. 3, pp. 1-44. de Bettignies, J-E. & Ross, T. (2004). The Economics of Public Private Partnerships. Canadian Public Policy, Vol. 30 (2), pp. 135-154. Domberger, S. (1998). The Contracting Organization, Oxford University Press. Donahue J. (1989). The Privatization Decision: Public Ends, Private Means. New York: Basic Books. Eggers, W. D. & Ng, R. (1993). Social and Health Service Privatization: A Survey of County and State Governments. Policy Study No. 168, Privatization Center. Los Angeles: Reason Foundation. Globerman, S. & Vining, A. (1996). A framework for Evaluating the Government Contracting-Out Decision. Public Administration Review, Vol. 56, pp. 577-580. Grizzle, G. A. (1985). Performance measures for budget justifications: Developing a selection strategy. Public Productivity and Management Review, 9, pp. 328-341. Gurwitt, R. (2000). Focus: Child Welfare, The Lonely Leap. Governing Magazine, July. Available at www.governing.come/archive/2000/jul/child.txt. Heclo, H. (1981). Towards a New Welfare State, in P. Flora & A. J. Heidenheimer (eds.), The Development of Welfare States in Europe and America, pp. 137-180. New Brunswick: Transaction. Hood, C. (1991). A Public Management for All Seasons? Public Administration, Vol. 69(1), pp. 3-19. Kettl, D. F. (2000). The global public management revolution: a report on the transformation of governance. Washington, DC: Brookings Institution Press. McKean, R. N. & Browning, J. (1975). Externalities from government and non-profit sectors. Canadian Journal of Economics, 8(4), 574-590.
188
Argentino Pessoa
Miller, G. (2001). Introduction. In G. Miller, W. Hildreth, & J. Rabin (eds.) Performancebased budgeting, pp. 1-16. Boulder, CO: Westview Press. Mills, A. (1995). Improving the Efficiency of Public Sector Health Services in Developing Countries: Bureaucratic versus market approaches. Departmental Publication No. 17, London School of Hygiene and Tropical Medicine. Minouge, M., Polidano, C. & Hulme, D. (1998). Beyond the New Public Management: Changing Ideas and Practice in Governance. Cheltenham, Mass.: Edward Elgar. Musgrave, R. A. (1959). The Theory of Public Finance. New York: McGraw-Hill. Osborne, D. & Gaebler, T. (1992). Reinventing Government: How the Entrepreneurial Spirit Is Transforming the Public Sector. Reading, MA: Addison-Wesley Publishing Company. Pessoa, A. (2008). Public-Private Partnerships In Developing Countries: Are Infrastructures Responding To The New ODA Strategy? Journal of International Development, 20, 311325. Prahalad, C. K. & Hamel, G. (1990). The Core Competence of the Corporation, Harvard Business Review, May-June, 79-91. Robinson R. (1990). Competition and Health Care. A Comparative Analysis of UK Plans and US Experience. Research Report No. 6. London: Kings Fund Institute. Rodríguez, A. (2005). Case Study: Public School Concession model of Bogotá, Colombia, non-published document, the World Bank. Samuelson, P. (1954). The Pure theory of Public Expenditure, Review of Economics and Statistics, 36(4), pp. 387-389. Sanger, M. B. (2001). When the Private Sector Competes, Washington, DC: Brookings Institution Press. Savas, E. S. (2005). Privatization in the City: Successes, Failures, Lessons. Washington, DC: CQ Press, chapter 1. Scitovsky, T. (1954). Two Concepts of External Economies. Journal of Political Economy, 62, 143-51, reedited in N. Agarwala & S. Singh (eds.) (1958), The Economics of Underdevelopment, London: Oxford University Press. Sclar, E. D. (2000). You Don’t Always Get What You Pay For: The Economics of Privatization. A Century Foundation Book. Ithaca, NY: Cornell University Press. Sedjari, A. (2004). Public Private Partnerships as a Tool for Modernizing Public Administration. International Review of Administrative Sciences, Vol. 70(2), 291-306. USGAO (1997). Social Service Privatization: Expansion Poses Challenges in Ensuring Accountability for Program Results. Publication no. GAO/HEHS-98-6, October. Washington, DC: U.S. General Accounting Office. Van Slyke, D. (2003). The Mythology of Privatization in Contracting for Social Services. Public Administration Review, Vol. 63(3), 277-296. Villa, L. & Duarte, J. (2004). Colegios en Concesión de Bogotá: una experiencia innovadora de gestion escolar en Colombia. In J.C. Navarro, J.Vargas, J. Duarte, & G. Arévalo (Eds.), Alianzas publico-privadas en educación: innovaciones en América latina. Washington, DC: BID. von Otter C. & Saltman, R. B. (1992). Planned Markets and Public Competition. Strategic Reform in Northern European Health Systems. Buckingham: Open University Press. Williamson, O. (1979). Transaction-cost economics: The governance of contractual relations. Journal of Law and Economics, 22(2), 233-261.
Outsourcing and Public Sector Efficiency
189
Wisniewski, S. (1991). Analyzing the contracting-out of government services: Relevant costbenefit considerations. Public Budgeting and Finance, 11, pp. 95-107. World Bank (2002). Private sector development strategy: Directions for the World Bank Group. Washington DC: World Bank. World Bank (2004). Contracting Out Utility Regulatory Functions, Final Report. Washington DC: World Bank.
In: Outsourcing, Teamwork and Business Management ISBN: 978-1-60456-956-8 Editor: Karl E. Carettas, pp. 191-212 © 2009 Nova Science Publishers, Inc.
Chapter 11
EXAMINATION OF DEDICATED RELATIONSHIPS BETWEEN AUTOMOTIVE SUPPLIERS AND CARMAKERS: EVIDENCE ON THE FLAGSHIP / 5 PARTNERS MODEL Bart Kamp* Rotterdam School of Management, Erasmus University Rotterdam, PO Box 1738 3000 DR Rotterdam, The Netherlands, Room T11-53
Abstract The Flagship / 5 Partners model argues that key suppliers are dedicated exclusively to flagship firms and that flagship firms work on an exclusive basis with key suppliers. As the F/5P model is partly rooted in empirical analyses of the North-American automotive industry, it is interesting to test its external validity on the European car industry. Similarly, it seems relevant to test its claim of exclusive buyer-supplier relationships as there are also scholars that report on non-exclusive b2b practices. This paper analyzes the component supply relationships for 32 car models to test the exclusivity presumptions of the F/5P model. Results show that industry-wide client bases on behalf of suppliers and multiple sourcing practices by carmakers are a stronger empirical reality than exclusive flagship firm-key supplier relationships. Outcomes also indicate that previously in-house parts of carmakers successfully succeed in establishing client relationships with third party OEMs.
Key words: buyer-supplier relationships, single sourcing, automotive industry, flagship firms.
Biographical Notes Dr. Bart Kamp is lecturer in International Entrepreneurship and New Business Venturing at the Erasmus University Rotterdam (the Netherlands). He also holds the position of Senior Policy Adviser at Technum-Resource Analysis, one of the largest Belgian consulting firms in *
E-mail address:
[email protected]
192
Bart Kamp
strategic policy and industry matters. As such, he is project manager of a major European research assignment called ‘Benchmarking car manufacturing performance and supportive public policy in European regions dependent on car industry’. He is a member of the International GERPISA Network of Researchers on the Automotive Industry. Previous academic positions held include: guest lecturer at the University of Antwerp Management School (Belgium), visiting researcher at the Department of Applied Economy of the Universitat Autònoma de Barcelona (Spain) and visiting researcher at the Department of Organization Sciences at the University of Tilburg (the Netherlands).
Acknowledgements The author would like to thank Prof. Dr. Yannick Lung (Montesquieu University of Bordeaux IV, France) for his valuable comments on this work.
Introduction With the studying of and attempts to learn from the Japanese (car) manufacturers (Cusumano, 1985; Dore, 1987; Womack, Jones and Roos, 1990), it not only became fashionable for Western companies to follow the logistics and lean enterprise principals of the Japanese companies (Monden, 1981a, 1981b, 1992), but also to take the supplier relationship management practices of the Japanese as a benchmark for own functioning (Asanuma, 1988; Cusumano and Takeishi, 1991). In that respect, the conclusion was drawn that -among otherscarmakers from North America and Europe would have to reduce the enormous amounts of direct supply relationships they maintained with parts and component manufacturers (Boston Consulting Group, 1993; Shimokawa, 1999; Veloso, 2000). In operational terms, what was believed necessary to update the antiquate Western car manufacturer-supplier relationships was to come to a tiering of all supply relationships (with only the first tier suppliers maintaining a direct relationship with the car manufacturer) and to adopt single sourcing policies with regard to the delivery of automotive components for the assembly of car models, instead of having multiple suppliers for the same component. It is the concept of single sourcing that forms the main issue of inquiry in the present article, and the question to what extent it is implemented. It is not difficult to imagine cases where, for a specific component, a single sourcing policy is practiced with regard to a single car model or even an individual assembly plant. For instance: a specific manufacturer of seats being the sole provider of seats for a particular model or a particular assembly plant. However, one could also hypothesize going beyond that level, more in the spirit of textbook lessons on Japanese inter-firm relationships. This is what the so-called “Flagship / 5 Partner model” (Rugman, 1999; Rugman and D’Cruz, 2000) does. It presumes that multinational corporations, such as carmakers, entertain single and exclusive relationships with key suppliers for their entire global operations. Moreover, the Flagship / 5 Partners model (Rugman, 1999; Rugman and D’Cruz, 2000) claims that key suppliers are dedicated exclusively to flagship firms (see e.g. 2000, p. 11, 36-37, 65, 84-86, 96). Conceptual support for single sourcing practices can be derived also from other frameworks, like transaction cost economics (e.g. Williamson, 1981). Also in the Network
Examination of Dedicated Relationships…
193
and Interaction Approach to b2b relationships there is a strong emphasis on mutual orientation between firms in buyer-supplier interactions/relationships (Cunningham, 1980; Ford, 1980, 1997), which lead to bonding, relationship-specific investments, adaptations from both sides and lasting relationships (Johanson and Mattsson, 1987; Forsgren, Hägg, Håkansson, Johanson and Mattsson, 1995; Laage-Hellman, 1997). However, these two approaches do neither exclude the simultaneous existence of multiple buyer relationships on behalf of a specific supplier, nor of multiple supplier relationships on behalf of a specific buyer. The Flagship / 5 Partners model, though, seems to be provide a stronger argument for single sourcing. Therefore, attention is focused on this model’s premises when analyzing the concept and practice of single sourcing. As the conceiving of the Flagship / 5 Partners model (hereafter: F/5P model) is partly rooted in empirical analyses of the North-American automotive industry, it is an interesting challenge to see to which extent such exclusive relationships can be observed in automotive business reality. Especially since there are also scholars like Lieberman and Montgomery (1988) and Wells and Rawlinson (1994), who refer to the existence of industry-wide suppliers. Moreover, a fair share of the historical anecdotal evidence1 that exists, as well as particular theoretical reasoning (see e.g. works in the Transaction Cost Economics tradition from Williamson, 1975, 1981, 1985 and works in the tradition of the Resource-Based View from Wernerfelt, 1984 and Barney, 1991), would question such exclusivity behaviour for being too risky for both sides. Among others as it would lead to high asset specificity of resources or pronounced relationship-specific investments (Dyer, 1997) with the risk of low value of such resources in other circumstances than under incumbent relationships (Klein, Crawford and Alchian, 1978). Also, engaging in exclusive relationships presupposes a level of trust between business partners that may not be realistic in the Western world (Sako, 1992; Dyer, 1997). Instead, there may be too much fear for opportunism on behalf of the other side to prevent the emergence of exclusive relationships in many occasions (Williamson, 1985). Given these antitheses, this article tests the exclusivity presumptions of the F/5P model via a survey on buyer-supplier relationships in the automotive sector. It investigates to which extent single sourcing policies and exclusive buyer-supplier relationships are in play in the car industry beyond the level of individual car models or assembly plants. Based on a sample of 32 European car models, the buyer-supplier relationships for a specific set of automotive components are analyzed. The survey allows to screen the validity of the exclusivity claim regarding buyer-supplier relationships and to assess the existence of industry-wide suppliers. It also enables to comment upon the client trajectories that several component manufacturers have “travelled” since gaining independence from their former owners i.e. Delphi from GM, Visteon from Ford and Faurecia from PSA. Results show that industry-wide client bases on behalf of suppliers and multiple sourcing relationships on behalf of carmakers are a stronger empirical reality than exclusive flagship firm-key supplier relationships. Outcomes also indicate that previously in-house parts of carmakers that became independent have succeeded in establishing client relationships with multiple third party carmakers.
1
For instance, Lopez de Arriortua’s approach to saving GM several billions of dollars in the early 1990s made sure that no supplier would be interested in an exclusive relationship with GM. And the other US carmakers had only slightly better reputations, since they also started asking across-the-board price reductions from their suppliers.
194
Bart Kamp
The Flagship / 5 Partners Model The F/5P model (Rugman and D’Cruz, 2000) can be seen as a conceptual framework for understanding the formation and evolution of business networks. Their creators claim that the model is inspired by successful keiretsus (p. 8, 33, 57, 86, 180). The term "flagship firm” refers to a company operating at the core of an extensive business network. Usually the flagship has long-term relational contracts with a set of five partners. The five partners consist of the flagship firm itself, key suppliers, key customers, competitors, and non-business infrastructure. The present article focuses on the relationships between the flagship firm and the key suppliers. From Rugman and D’Cruz (2000, pp. 36-37, 86, p. 94 and notably pp. 166-170) it can be understood that the difference between key and non-key suppliers in an automotive industry setting, is the difference between first tier and lower tier suppliers. They argue that the first tier suppliers provide a key input to the network whereas the other suppliers do not. The F/5P model is based on the assumption that sustainable competitiveness can best be achieved through co-operative relationships in a business network structure. It argues that inter-organizational cooperation makes firms depend less on internalization as the sole means to leverage their firm-specific advantages and gain competitive advantage. Instead of depending on internalization, the F/5P model posits that a flagship firm focuses on sharing knowledge among partners in order to facilitate inter-organizational learning. Rugman and D’Cruz (2000) compare their F/5P model with Lorenzoni and BadenFuller’s strategic network framework (1993, 1995). Lorenzoni and Baden-Fuller speak of a central firm instead of a flagship firm. Like Rugman and D’Cruz, they also stress asymmetrical dependence between the central firm and the partners. The F/5P model argues that the behaviour of key suppliers can be explained through their asymmetrical dependence on the flagship firm’s strategy and their exclusive dedication to a certain flagship from a single industry perspective (Rugman and D’Cruz, 2000; Rugman and Verbeke, 2003). This implies that the flagship firm is able to provide a strategic perspective for all five partners pertaining to their networks and that each true partner does not need a fully separate strategy except to be a "key" partner of a specific flagship firm. In the F/5P model, the actors involved are supposed to share a common global strategy and purpose, with the flagship firm having the resources and perspective to lead the network and strategically manage its activities. The partners yield the strategic leadership to the flagship because it is the flagship’s product and global strategy that made the partners join the network in the first place. Key suppliers are expected to give near or total exclusivity to the flagship firm (Rugman and D’Cruz, 2000, p. 2, 9, 31, 41, 65, 84-86, 96). The supposed benefit to suppliers of such a deal is that they should receive increased order volumes and multi-year supply contracts, strategic and technological guidance and learning, and capture a higher proportion of value-added activities that are outsourced by the flagship firm (p. 58). The reason for flagship firms to work on an exclusive basis with key suppliers is to harness their own global strategy. Global competition in industries is supposed to take place increasingly between F/5P-like business networks and there is a trend for multinationals as flagship firms to depend less on internalization (Rugman and D’Cruz, 2000, p. 57; Rugman and Verbeke, 2003). Consequently, sharing knowledge among network partners in a collaborative atmosphere facilitates interorganizational learning and fosters competitive advantage of both a hub multinational and the partners with which it works.
Examination of Dedicated Relationships…
195
To conclude with, Rugman and D’Cruz (2000, p. 56) state that “… partners may compete in business systems not related to that of the business network, [and therefore] it should be emphasized that the flagship firm’s asymmetric strategic control extends only to those aspects of its partners’ business systems committed to the network.” This would mean that a versatile component company active in e.g. the automotive industry and the defence industry, could at the same time be a key supplier to one flagship firm from the automotive business and to another from the defence industry. However, it also implies that a specific supplier of key inputs to a certain flagship firm of a business network does not operate as key supplier to another flagship firm operating in the same business system, like the automobile industry.
Empirical Observations The rationales behind the F/5P exclusivity propositions are in line with general trends among carmakers to reduce the amount of suppliers from which they source (Lamming, 1993; Womack, Jones and Roos, 1990; Sako and Warburton, 1999; Veloso, 2000) and to form long term relationships with key suppliers (Pyke and Johnson, 2003; Maurer, Dietz and Lang, 2004). In this respect, the following figures are very illustrative. Between 1984 and 2001 Renault reduced the amount of suppliers of automotive parts from which it sourced from 1800 to 507 (Gorgeu and Mathieu, 1995; Renault, 1986-2002RA; Renault, 1991-2002AE). From 1983 to 2001, VW reduced its worldwide supplier sample from 30,000 to 4,532 (Volkswagen, 1984, p. 37; Grohn, 2002). Identical trends can be observed for other carmakers, like PSA: from 900 direct suppliers in 1986 to approximately 500 in the year 2000, BMW: from 1400 to 600 in the same period, Ford: from 2400 to 1200 in the same period, and Chrysler: from 3000 to 600 in the same period (the Economist Intelligence Unit, various years; Ward’s, various years). Furthermore, Chrysler’s average contract length nearly doubled over a ten-year period (Helper, 1991; Dyer, 1997). Finally, after playing with the thought of following a “single firm per module supply across the world” strategy (Veloso, 2000), Ford now plans to cut by more than half the amount of suppliers from whom it buys the 20 parts that make up about half of its global purchasing bill. Meanwhile, it will offer larger, long term contracts on models built from 2008 onwards to the ones that will remain as preferred suppliers and build more collaborative relationships via a so-called “Aligned Business Framework” (Automotive News, 2006). It is also in line with tendencies towards global instead of local-for-local sourcing (see for example the annual reports on behalf of automotive OEMs, as well as: Boston Consulting Group, 1993; Shimokawa, 1999) and single and module-based sourcing in detriment of multiple and fragmented sourcing (see again the annual reports of automotive OEMs, as well as: Lamming, 1993; Dyer, Sung Cho and Chu, 1998). It can also be underpinned by empirical observations that indicate that many overseas supplier establishments follow from corporate negotiations with carmakers for whom they already work and to which they want to keep committed (Volpato, 1997; Carillo and Gonzalez Lopez, 1999; Shimokawa, 1999; Gonzalez Lopez, 2000; Frigant and Lung, 2002). At the same time, however, one can find indices of opposite practices. Empirical evidence demonstrates that carmakers encourage suppliers to work for multiple clients (Florence, 1996), and suppliers may effectively follow a multiple client strategy or an industry-wide car constructor servicing policy (O hUallachain and Wasserman, 1999; Rutherford, 2000, p. 743; Rhys, 2000; Valeo, 2000). Similarly, conceptual thinking from the
196
Bart Kamp
network and interaction approach to b2b relationships emphasizes multiple inter-firm ties (Håkansson and Shehota, 1989; Laage-Hellman, 1997, p. 16). Finally, recent times also witnessed a selective return of carmakers to multiple sourcing for one and the same component in order to spread risks and not to place all eggs in one basket. See e.g. front-end production and assembly activities for the Renault Clio (Renault, Marzo 2001G, p. 42) as an exponent of a more general practice of working with multiple suppliers per component (Torrico, 2002). Also Veloso (2000) argues that VW and Renault tend to relate with more than one partner for the development and supply of a specific component. All these observations are at odds with the exclusivity principles, preached by the F/5P model. Table 1. Synthetic overview of the benefits and costs of single sourcing Advantages of single sourcing
Disadvantages of single sourcing
Possibility to reduce costs through exploitation of scale economies (Marshall, 1890; Penrose, 1959; Chandler, 1962) Simplification of parts procurement and quality control activities and dealing with a more surveyable supplier base with, consequently, lower transaction, agency and quality inspection costs (Williamson, 1975, 1981, 1985; Jensen and Meckling, 1976; Fama and Jensen, 1983) Better possibilities to build trust relationships with a reduced number of suppliers (Sako, 1992) High (resource) dependence upon a sole supplier (Barney, 1991; Wernerfelt, 1984), in terms of product flow (e.g. in case of strikes, fire or calamities) and quality of products sourced High switching costs, especially in case of high asset specificity (Williamson, 1975, 1981, 1985) Information asymmetry: supplier tends to understand better the technical processes behind product and production, and has a better view on cost savings possibilities. This results in weaker bargaining power on behalf of the buyer and can lead to opportunism on behalf of suppliers (Williamson, 1975; Kapoor and Gupta, 1997; Ford et al, 1998) Single sourcing provides lesser opportunities for learning effects, to increase absorptive capacity and to identify best practices (Cyert and March, 1963; Levitt and March, 1988; Cohen and Levinthal, 1990) Single sourcing may provide insufficient incentives to suppliers to offer the best possible price, quality, service and overall performance, due to a lack of competition with peers (Ettorre, 1995; Carter and Narasimhan, 1996) If demand is high, it may supersede the available production capacity of a single supplier and multiple sourcing provides a better way to deal with demand uncertainty (Williamson, 1975; Ford et al, 1998)
Source: own elaboration.
On the part of suppliers, it is neither self-evident that they would follow a single client model as there is ample proof of the existence of industry-wide suppliers (Lieberman and Montgomery, 1988; Wells and Rawlinson, 1994). Besides spreading risks and enlarging turnover and learning possibilities, lack of trust is another source that feeds multiple client strategies. Conssequently, there is a trade-off between the benefits of exclusive and non-dedicated buyer-supplier relationships. In the table below, we summarize and conceptualize the main (dis)advantages of single sourcing, notably from the buyer’s perspective. Consequently, the F/5P principles of buyer-supplier exclusivity are far from undisputed and the degree with which it can be encountered in practice deserves a closer look.
Examination of Dedicated Relationships…
197
Research Design The chosen research setting to explore the tenability of the F/5P assumption that key suppliers are exclusively dedicated to a flagship firm and that flagship firms engage in exclusive relationships with key suppliers, is the European automotive industry. This choice is invoked by the following motives. First of all as the automotive industry is a prominent example of a sector where one encounters buyer-supplier networks on a large scale (Castells, 1996; Dyer, 1996; Fine and Whitney, 1996). Secondly as the conceivers of the F/5P model also apply their concepts on the North-American car industry and source from practices in the car industry of that part of the world, it also follows that the car industry is a good test case for assessing the validity of the ground rationales underneath the F/5P theory. Thirdly, with regard to the automotive industry, the phenomenon of single sourcing is related to the exclusivity point raised by the F/5P model. In fact, whereas the practicing of single sourcing is widely reported on single plant and or model level, the F/5P model goes one step furhter and predicts single sourcing and dedicated supply on a corporate level (that is: a buyer company sourcing a specific component from one supplier for its whole product range or a supplier selling a specific component exclusively to a single buyer company). Thus, also from this perspective the automotive industry makes sense as a relevant and interesting research setting. Finally, the fact that the present research focuses on flagship firm-key supplier practices in the European automotive industry means it may also allow drawing up comparisons between North-American and European flagship firm-key supplier practices. Within the automotive industry setting, the carmakers can be seen as the flagship firms, whereas key suppliers are notably the first tier suppliers. The carmakers are the companies that assemble and integrate the various components and systems obtained from suppliers into a final product for the end consumers market. They are the hub or central buyer firm of the automotive buyer-supplier networks. By key or first tier suppliers the F/5P model refers to those suppliers whose components and systems are a vital input to the final product and which already form an integrated product or module in itself (“system integration”), see e.g. Rugman and D’Cruz, 2000, pp. 3637, 86, p. 94 and notably pp. 166-170). As a consequence, the article focuses on examples like complete cockpit systems instead of identifying who are possibly the suppliers of clockwork instruments, rubber panels and panel support items behind the system integrator of the cockpit as a whole. The data used to deal with our research question were gathered via a medium-sized survey on the first tier suppliers that OEMs work with for several car models. In total, with regard to 32 car models assembled in Great Britain and on the European continent, coming from 5 car groups, the first tier supplier relationships for a specific set of automotive components were analyzed. All car models are from the smaller and middle class segment. Moreover, most car models stemming from the same car groups are platform interrelated. This contributes to basing the analyses upon an internally coherent set of research units. Data came from own field work with regard to the suppliers for the VW Polo and the Renault Clio, supplemented by breakdowns of the supplier networks around car models obtained via specialized automotive journals. Finally, data were obtained from annual reports from several first tier suppliers to final assembly plants throughout Europe.
198
Bart Kamp
The car models that were screened are presented in the following tables. To highlight the interrelatedness of models and brands, they are grouped platform-wise as well as based on shared group ownership. With regard to this research sample, the following analyses were carried out. One, with regard to the following components it was investigated whether a single car concern used one or more suppliers for the car models screened, both across platforms and on the level of a single platform:2 • • • • • • • • • •
Exhaust system Suspension Steering direction Coil springs Front-end carrier Wing mirrors Fuel system Cockpit Seats Head lamps
Two, for each of the identified first tier suppliers active in the delivery of the aforementioned components, it was established whether it maintained supply relationships with one or more carmakers. Finally, the research looks into the penetration of former in-house suppliers or internal “departments” of certain carmakers i.e. Visteon (ex-Ford), Delphi (ex-GM) and Faurecia (PSA-linked) into the supply chains of previously unrelated carmakers for these suppliers. Volkswagen “Supermini” Platform PQ 25: VW Polo SEAT Cordoba
Audi SEAT Volkswagen Aktien Gesellschaft Platform PQ 35: Platform PL 45: VW Golf Audi A6 VW Touran Audi A3 Skoda Octavia SEAT Altea
Peugeot PSA Platform PF1: Citroën C2 Citroën C3 Peugeot 1007
Skoda “Super Passat” Platform PL 62: Audi A8 VW Phaeton
Citroën
Platform PF2: Citroën C4 Peugeot 307
Platform PF3: Peugeot 407
Platform PF4: Peugeot 807
Figure 1. Continued on next page.
2
The components whose suppliers’ we have examined form a representative sample of automotive parts that are commonly outsourced by car manufacturers today. Moreover, only those components were withheld for analysis for which the subsequent suppliers of most of the screened 32 car models could be traced back. In other words, we focused on those components for which most supplier identity indicators were available.
Examination of Dedicated Relationships… Renault
Dacia Renault-Nissan Alliance Platform C : Renault Scenic
Platform B : Dacia Logan Renault Modus Renault Clio Nissan Micra
Nissan Platform M2: Renault Espace
Opel / Vauxhall Platform Gamma: Opel / Vauxhall Meriva
Ford
General Motors Platform GM 2700 / 3000: Opel/Vauxhall Astra
Volvo Ford Motor Company Platform BE 91: Ford Streetka
Platform B 2: Ford Fusion
199
Saab Platform EPSILON: Opel/Vauxhall Signum Saab 9-3 Jaguar Platform C 170: Ford Focus Ford Focus C-Max Volvo S40
Source: own elaboration.
Figure 1. Overview and classification of research sample.
Results In the next figures and tables the following issues are illustrated. Figure 2 presents the number of component-specific suppliers per car constructor as a whole. This serves as an indicator for car constructor practices of single sourcing per component and of maintaining exclusive supply relationships with regard to individual components. Figure 3 reports the number of suppliers an OEM engages for specific components on the level of platformrelated car models. This serves as an additional indicator for the existence of single sourcing practices. Figure 4 shows, per component type, the number of clients to which component-specific suppliers service. This serves as an oligo-/monopolization indicator per component “segment” and as an indicator for the existence of industry-wide suppliers c.q. of dedicated supply practices per component “segment”. Table 1 presents the amount of components that formerly in-house departments of specific car constructor sell to the carmakers that formed our sample. This serves as an indicator for the penetration of such formerly internalized supply activities into third party client relations and the degree with which formerly in-house suppliers of OEMs have succeeded in establishing business relationships with competitor OEMs.
Nr. suppliers 12
10
8
VAG PSA 6
Renault GM Ford
4
2
0 Sample size per OEM (N car models)
Exhaust
Suspension
Steering direction
Coil springs
Front-end carrier
Wing mirrors
Fuel system
Cockpit
Seats
Head lamps
Source: own elaboration.
Figure 2. Number of component-specific suppliers per car constructor (results as measured on a corporate level).
Nr. suppliers
6
5
4
VAG PQ 35 PSA PF 1 3
Renault B Ford C 170
2
1
0 Sample size per
Exhaust
Suspension
St eer ing dir ect ion
Coil spr ings
Fr ont - end car rier
Wing mir ror s
Fuel syst em
Cockpit
Seat s
plat f or m ( N car models)
Source: own elaboration.
Figure 3. Number of component-specific suppliers per car constructor (results as measured on platform level).
Head lamps
Average N of OEM-client relationships per category of component supply 12
10
8
Sample size per component category (N of suppliers)
6
Average N of OEM-client relationships per supplier
4
2
la m ps H ea d
S ea ts
C oc kp it
Fu el sy st em
m irr or s W in g
ca rr ie r Fr on t-e nd
C oi ls pr in gs
di re ct io n
on
S te er in g
S us pe ns i
E xh au st
0
Source: own elaboration.
Figure 4. Number of carmakers to whom suppliers deliver specific components.
Examination of Dedicated Relationships…
203
Table 2. Number of supplied components to sample of OEMs by (formerly) in-house departments of specific carmakers Clients
Visteon (ex-Ford): interior and exterior trim parts, lighting, climate systems, powertrain systems, steering systems, chassis components, cooling systems VW 1 Ford 5 GM 1 PSA 5 Renault 2 Source: own elaboration.
Delphi (ex-GM): engine management systems, climate control, cockpit and interior products, cabling, cooling systems, steering systems 4 1 4 2 1
Faurecia (PSAlinked): exhaust systems, seats, acoustic package, cockpit, door panels, front-end module 6 4 3 6 6
From the previous graphics the following can be learned. According to the F/5P thesis that flagship firms practice single sourcing for the supply of key components, values for “number of component-specific suppliers per car constructor” (see Figures 2 and 3) should generally be “1”. However, as one can see the empirical results provide a rather different picture. On a corporate level, only Ford demonstrates signs of single sourcing per component for a whole range of car models. I.e. for the five car models screened of Ford, it appears that regarding suspensions, wing mirrors and fuel systems it makes use of the same supplier. On the level of platform-related car models, one encounters somewhat more indications of exclusive supply relationships for specific components. All reviewed platforms appear to have one component-specific single supplier case. For VAG’s PQ 35 platform this is the case with the steering direction: for all five car models pertaining to the indicated platform that were screened, the steering direction was delivered by the same supplier. PSA’s PF 1 platform, 3 car models screened, appears to work on a single supply basis with a particular supplier for wing mirrors, fuel systems and seats. Renault’s B platform, 4 car models screened, appears to work on an exclusive basis with a single supplier for the supply of seats and head lamps. Finally, Ford’s C 170 platform, 3 car models screened, revealed single sourcing signs for the supply of suspensions, front-end carriers, wing mirrors and cockpits. On a whole, however, there is a strong case to assume that OEMs engage with multiple suppliers for all kinds of key inputs to their final products. On a corporate level, for the conjoint of car constructor-component combinations only in 2,5% of the possible cases indications of single sourcing are encountered. On the level of platform-related car models, this is 25%. The latter percentage indicates that single sourcing is not even the general rule for component supply on platform-level. The fact that our research sample has a reduced size, makes one even think that if it were larger, this would become even more manifest. The same is probably true for the findings on a corporate level. Therefore, notwithstanding the fact that single sourcing practices are certainly a reality for individual car models, on more aggregated levels of analysis, i.e. on platform-level and on corporate level, the phenomenon exists to a lot lesser degree. In fact, multiple sourcing appears to be the standard. This becomes most manifest in the case of VAG, although it has to be stated this was the company for which most car models were screened (which may inflate the impression that VAG exceeds the other OEMs in multiple sourcing). All in all, the overall picture is one of multiplicity of
204
Bart Kamp
OEM-supplier relationships per component. Not even on platform-level, except for a minority of components, OEMs seem to engage in single supplier relationships. As regards insights on suppliers working exclusively for specific OEMs, we turn to Figure 4. there the average number of clients serviced by suppliers of specific components are portrayed. If a supplier of a specific component were to service only one OEM, he would get a value “1”. As our sample contains but 4 OEMs, the maximum value for OEM relationships a supplier can entertain is “4”. The closer to 1 the average amount of OEM relationships entertained by a set of component-specific suppliers is, the more this indicates exclusive supply relationships. If the average score is closer to 4, the more this indicates the existence of industry-wide suppliers. Overall, the average scores lie well above the value 1. Only for the components with a rather fragmented supplier landscape, i.e. with large numbers of suppliers, like suspensions, front-end carriers and cockpits, the average number of OEMs serviced per individual supplier is rather low (below 2). On the other hand, for components where one encounters a more concentrated supplier landscape, as for coil springs, fuel systems and seats, one encounters significantly higher values. To a certain extent, the latter is logical as the more oligopolized or monopolized a component segments is, the more clients the few (remaining) suppliers from this segment will service. For one reason, as their clients will not have a lot of choice. The opposite is not automatically true, if a component segment counts with a large number of suppliers, this does not have to mean that each of them will only service a small number of the final clients sample. For this depends on the sourcing strategies of the buyer OEMs. If they insist on multiple supplier relationships for specific components, then also in the event of a fragmented supplier landscape each individual supplier can still entertain multiple supplier relationships. This is illustrated by the fact that also in the cockpit, suspension and front-end carrier segments one encounters many examples of suppliers that service various OEMs and exclusive relationships are also here far from being the rule. Thus, also in these segments this leads to average values well above 1. And although we do not measure extremely high values (close to 4) for the components segments with large numbers of suppliers, still the results in Figure 3 point at a general situation of non-dedicatedness on behalf of suppliers vis-à-vis OEMs and ample evidence of industry-wide suppliers. It also fortifies the impression of OEMs practicing multiple sourcing strategies. A final finding that follows from our sample material, is that formerly dedicated suppliers have succeeded in establishing multiple OEM client bases. From Table 1 it clearly appears that both Visteon, Delphi and Faurecia –all three of them offering a highly diversified package of components and functions- have effectively penetrated into the purchasing channels of several OEMs that are competitors of their (former) owning hierarchies. This is most strongly the case for Faurecia. With regard to this company it was possible to establish that it supplies multiple functions and components to all five OEMs from our sample. This is what our own primary data indicates. Moreover, secondary analysis on the annual reports of these three companies reveals that they also supply several other OEMs, often with the same kind of (key) components. Therefore, they also can be classified as industry-wide suppliers.
Discussion and Implications Results show that industry-wide suppliers are a stronger empirical reality than exclusive buyer-supplier relationships. More than single sourcing practices per component on behalf of
Examination of Dedicated Relationships…
205
carmakers, one observes sourcing from multiple supply relationships. In fact, several of the analyzed carmakers publicly express that this is a policy they follow, also on the level of individual car models. Noteworthy examples here are VW and Renault. For example with the introduction of the latest Golf, Volkswagen continued its policy of splitting parts contracts among multiple suppliers (Automotive News Europe, September 22, 2003, p. 22T) using two or more suppliers on several key components. Whereas having multiple suppliers per component for a single car model may not be that typical, based on our sample it can be concluded that -when looking at the level of platform-related models or on a corporate levelit is clearly mainstream that multiple component-specific supplier relationships are maintained. And this sourcing behaviour appears to be representative for all carmakers that were screened. The practice of sourcing specific components from multiple suppliers is understandable. It forms a way to regulate lock-in effects and dependence upon external relationships (Grabher, 1993; Uzzi, 1997; Kamp, 2005). Likewise, it fosters learning possibilities for carmakers. By engaging in multiple supply relationships, carmakers learn from different parties (Lane and Lubatkin, 1998; Nooteboom, 2002) and as such they are better able to judge the functioning of suppliers. Moreover, this can limit the know-how differential and, therefore, dependence on behalf of carmakers upon suppliers that can arise from outsourcing. The fact that several carmakers maintain certain component development and manufacturing activities in-house can also be interpreted as a way to dispose of countervailing power to dependence on external suppliers. For instance, Renault maintains certain production and development activities in-house and VW has created joint ventures with third parties for the conceiving and production of several key components. While multiple sourcing points at carmakers’ attempt to not depend entirely and exclusively on a single supply relation, the maintenance of in-house component capacity may indicate that carmakers intent to not rely completely on outside companies at all. As such, this is also a sign that carmakers sense that they (can) loose too much control over the forth bringing of vital car components. Another observation that results from our survey is that not all carmakers engage to the same extent in sourcing of “mega-modules” or of contracting suppliers for far-reaching “system integration”. This may also be based on the desire to avoid a too strong reliance on suppliers. In concreto by keeping the integration of automotive conjoints below a certain (system or modular) level, carmakers avoid losing too much grip on their conceptual and technological development and on their composition in parts and technologies. Through less aggregated conjoints, the purchasing management is able to maintain a stronger hold on the overall design and cost of a conjoint and exercise more control over supplier relationships. It also avoids becoming too dependent on suppliers through co-location patterns related to modular production. In fact, whenever OEMs have adopted a modular strategy at assembly plants (e.g. VW Resende, MCC Hambach, Skoda Mlada Boleslav, SEAT Martorell, Volkswagen Pamplona, Ford Valencia), it has strengthened spatial clustering and agglomeration of suppliers around such plants (Frigant and Lung, 2002), adding to mutual dependence between buyers and suppliers through higher site-related asset specificity 5williamson, 1981). The former has led several scholars to argue that this increases suppliers’ bargaining power (Millington, Millington and Cowburn, 1998; Van Hoek and Weken, 1998). Furthermore, also because it is posited that “… few suppliers really know all the technologies involved in more complex modules such as cockpits or front-ends. As a result
206
Bart Kamp
[several] suppliers are unable to manage all the sub-systems and having them integrate them does not always offer benefits” (SupplierBusiness, 2003, p. 20), carmakers may be more interested in less integrated modules. It also offers an additional explanation for a purchasing policy that does not rely on single supplier relationships for complex and vital components. On the other hand, there are multi-faceted supplier companies like Visteon, Delphi, Siemens and Bosch that dispose of knowledge and capabilities that many carmakers do not possess anymore (Grohn, 2002) and with regard to whom the rise of a know-how differential is a real possibility. This adds to the point that certainly not all suppliers depend asymmetrically on their car constructor clients. At least, the ongoing specialization of (certain) suppliers in their respective fields means that asymmetry in buyer-supplier relationships can be contained and that buyers can even come to depend on their suppliers. Moreover, as several of those suppliers service multiple buyers (even in multiple industries), their dependence on single carmakers should not be overestimated. Consequently, the non-distinction of the F/5P model into different “leagues” of suppliers as regards dependence and possibilities to influence car constructor firm strategies becomes problematic. As a matter of fact, it seems more appropriate to argue that some suppliers can be termed “influential suppliers” (Lamming, 1993; Grohn, 2002), whereas others –in spite of being first tier suppliers- have a rather weak impact on the strategies of car manufacturing firms. In brief, one can indeed expect that high-technology component suppliers like Bosch, Siemens and even commodity suppliers like Alusuisse exert significant influence on the strategies of car manufacturing firms and are less depending on (single) carmakers. However, one should expect suppliers responsible for the delivery of, for instance, mirrors, to do so to a far lesser degree. Precisely from the viewpoint of reducing dependence, also the establishment that suppliers do not devote themselves to a single client relationship is comprehensible. Not only from a direct commercial perspective, but also from a market and business intelligence point of view: relating to multiple clients is rational. While attending one single client, suppliers can lose valuable feedback about their market environment. Working for different clients enhances suppliers’ learning curve, which influences their competitiveness in a positive way. This is something from which also carmakers can take profit. Hence, carmakers are also interested in their suppliers servicing multiple carmakers, as is testified by Pries (1999). Therefore, a managerial implication of our findings for purchasers is to be critical towards single sourcing practices. Instead, engaging with multiple suppliers will help to spread risks and increase learning opportunities. Concerted competition between and development of supplier capabilities helps improve performance and rentability for all parties involved. For sellers, the main implication is to be critical towards preferred vendor statuses. They ought to engage also with multiple buyers in order to spread their risks and steepen their learning curve. The previous considerations question the rationality behind the exclusivity claims of the F/5P model as such. Similarly, our findings lead to the conclusion that the assumption of the F/5P model regarding exclusive buyer-supplier relationships in a flagship-key suppliers setting are not valid in the European automotive context. Consequently, how then should the present results be judged in view of the observation that the conceptual and empirical fundaments underlying the F/5P model are partly rooted in the automotive industry (see e.g. Rugman and D’Cruz, 2000, p. 19, 37), and Rugman and D’Cruz’ assessment that they are applicable to this sector (2000, p. 160)?
Examination of Dedicated Relationships…
207
One can try to reconcile the different findings by pointing at the fact that OEMs may behave differently from one Triad-leg (e.g. North America or Europe) to another. In fact, Rugman and D’Cruz (2000, pp. 179-180) insinuate this themselves by claiming that automotive flagship firms operate regional business networks and not global ones. In other words, that modus operandi between OEMs and key suppliers can be different from one place to another. This is confirmed by studies on Japanese transplants outside Japan (Rutherford, 2001). However, there is more reason to believe that the F/5P model is not representative for buyer-supplier relationships in today’s automotive industry. Ultimately, it can be hypothesized that the way the F/5P model presents OEM-supplier relationships has never been the standard all together, not even in the North-American automotive industry. In support of this thought one can point at the lack of trust that has characterized buyer-supplier relationships in the U.S. automotive industry even before the arrival of purchasing managers like José Ignacio Lopez de Arriortua (Dyer and Chu, 1997). Only the trend towards trimming supplier bases, also among North-American OEMs, provides reason to believe that the remaining suppliers will entertain relationships of longer term, but by no means it equals exclusivity and mutuality. In this regard, Liker and Choi (2004) register how today, on the one hand, American OEMs want myriad suppliers for everything (p. 109), and on the other hand, set vendors against each other and then do business with the last supplier standing (p. 110). The fact that our sample also analyzed relationships of American and Canadian carmakers (GM, Ford) and suppliers (e.g. Magna, TRW, Visteon, Delphi), and witnessed that all of them relate with different suppliers or buyers respectively, contributes to the idea that also in North America sourcing from multiple suppliers and supplying to multiple clients is a normal practice. This is also voiced by Dyer, Sung Cho and Chu (1998) who reported that American OEMs adopt a rather laissez faire posture vis-à-vis suppliers that serve multiple carmakers. A final point is that the F/5P model’s conception of OEM-supplier or flagship-supply partner relationships –in terms of mutual exclusivity- may be due to the erroneous parallel that tends to be drawn with keiretsus as the basis and legitimization for assuming exclusive buyer-supplier relationships. As Nishiguchi (1993, p. 7) states: “A general misconception or confusion … equates the so-called harmonious Japanese supplier relations with an exclusionary single sourcing policy of automakers (further confused with keiretsu)”. Other scholars likewise established how single sourcing is in fact less common in Japan than usually thought (Richardson, 1993; Hines, 1995). Moreover, Lincoln, Harland and Brennan (1998) and Rutherford (2001) reported that significant changes were in fact taking place in the supplier relationships in Japan, and the keiretsu system is being adapted with primary suppliers being less dedicated solely to one major car assembler. Instead, they are increasingly supplying multiple automotive OEMs. Furthermore, Rutherford (2001) reports on erosion of long term loyalty of assemblers towards suppliers and a turn to price-based competition among suppliers and ditto relationships between automotive buyer and supplier firms. This has especially been the case with Nissan, who warned many of its traditional parts manufacturers that they are on their own now. The shift away from “old school” keiretsu principles at Nissan may of course have to do with its financial situation and its marriage with Renault, which may have infused more European principles on supplier relationships. For Toyota does not seem to move away from the traditional keiretsu patterns at the same pace. At the same time, many keiretsu suppliers intend –after setting up transplants abroad-
208
Bart Kamp
following its mother hen car assembler, to diversify the client base of their overseas subsidiaries, either by gaining access to other Japanese OEMs abroad or to non-Japanese OEMs (Parker, Rutherford and Koshiba, 2000). Also Liker and Choi (2004) point at the fact that Western executives incorrectly assume that in the keiretsu model companies are locked into buying components from specific suppliers. Instead, they report that neither Toyota nor Honda depends on a single source for anything and both develop two to three suppliers for every component or raw material they buy (p. 109). As a general point of conclusion then, it can be argued that exclusive buyer-supplier relationships certainly exist, especially on the level of single supply plants or final assembly plants, but it is illogical to assume that it is a general practice for (buyer or supplier) companies as a whole. On the part of suppliers, it is hard to conceive how they would render themselves completely at the mercy of one client. Several reasons to not expect this have been forwarded throughout this article (for example lock-in dangers and dependence on one client). On the part of buyer firms, it would mean that they miss out on suppliers working for multiple carmakers that bring in experiences from elsewhere. This holds true in Japan, Europe and North America. Outcomes from our sample also indicate that previously in-house parts of carmakers that came to stand on their own feet (Delphi, Visteon and Faurecia) have successfully established client relationships with third party carmakers. Therefore, even organizations that formerly pertained to direct competitors appear able to establish business relationships with additional clients. This also adds up to the impression that exclusive buyer-supplier relationships are not representative for the way carmakers and first tier suppliers relate to each other in automotive business networks.
References Asanuma, B. (1998), Japanese manufacturer-supplier relationships in international perspective: the automobile case, Working paper no. 8, Faculty of Economics, Kyoto University Automotive News Europe (2002-2004), Car cutaways Automotive News (2006), 4 join Ford top supplier list, 19 June 2006 Barney, J.B (1991), Firm resources and sustained competitive advantage, In: Journal of Management, 17 (1), pp. 99-120 Boston Consulting Group (1993), The evolving challenge for the European automotive components industry, Brussels Carillo, J. and S. Gonzalez Lopez (Février 1999), Relaciones cliente-proveedor de empresas automotrices alemanas en Mexico, Actes du GERPISA, Division internationale du travail et relations constructeurs-fournisseurs, pp. 65-76 Carter, J. R. and R. Narasimhan (1996), Is purchasing really strategic?, International Journal of Purchasing and Materials Management. 30 (13), pp. 3-13 Castells, M. (1996), The rise of the network society, Oxford Chandler, A.D. (1962), Strategy and structure: chapters in the history of the industrial enterprise, MIT Press, Cambridge MA
Examination of Dedicated Relationships…
209
Cohen, W. and D. Levinthal (1990) Absorptive capacity: A new perspective on learning and innovation, Administrative Science Quarterly, 35, pp. 128-52 Cunningham, M. T. (1980), International Marketing and Purchasing of Industrial Goods:Features of a European Research Project, European Journal of Marketing, 14 (5/6), pp. 322-338 Cusumano, M.A. (1985), The Japanese automobile industry: Technology management at Nissan and Toyota, Harvard University Press, Cambridge, Mass. Cusumano, M.A. and A. Takeishi (1991), Supplier relations and management: A survey of Japanese, Japanese-transplant, and US auto plants, Strategic Management Journal, 12, pp. 563-588 Dore, R. (1987), Taking Japan seriously, The Athlone Press, London Dyer, J.H. (1996), Specialized supplier networks as a source of competitive advantage: evidence from the auto industry, Strategic Management Journal, 17, pp. 271-291 Dyer, J.H. (1997), Effective interfirm collaboration: How firms minimize transaction costs and maximize transaction value, Strategic Management Journal, 18, pp. 535-556 Dyer, Jeffrey H., Sung Cho, D. and W. Chu (1998), Strategic Supplier Segmentation: The Next ‘Best Practice’ in Supply Chain Management, California Management Review, 40 (2), pp. 57-77 the Economist Intelligence Unit (several years), Motor business international and components business international, various editions/issues Ettorre, B. (1995), A strategy session with Prahalad, C.K., Purchasing performance measurement’s influence on concurrent engineering activities customer-supplier relationships: a model of performance in relational exchanges buyer seller relations’ transaction cost, International Journal of Business and Marketing. Fama, E.F. and M.C. Jensen (1983), Agency problems and residual claims, In: Journal of Law and Economics, 26 (June), pp. 327-349 FASA-Renault (1966-2002), Informes Anuales 1965-2001, Madrid Fine, Ch. H. and D.E. Whitney (1996), Is the make-buy decision process a core competence?, MIT Florence, A. (25-26 January 1996), Supply chain interactions and supplier relationships, Paper presented at «Colloquium on Regional strategies for innovation and competitiveness in the automotive industry», Cardiff Ford, D. (1980), Developments of buyer-supplier relationships in industrial markets, European journal of marketing, 14, pp. 339-353 Ford (Ed.), D. (1997), Understanding business markets, London: Dryden Press Ford, D.; Gadde, L-E; Håkansson, H.; Lundgren, A.; Snehota, I.; Turnbull, P. and D. Wilson (1998), Managing Business Relationships, John Wiley & Sons Ltd., Chichester Forsgren, M., Hägg, I., Håkansson, H., Johanson, J. and L.G. Mattsson (1995), Firms in Networks: A New Perspective on Competitive Power. Studia Oeconomiae Negotiorum #38. Uppsala: Acta Universitatis Upsaliensis Frigant, V. and Y. Lung (December 2002), Geographical proximity and supplying relationships in modular production, International Journal of Urban and Regional Research, 26 (4), pp. 742-755 Gilodi (2001): Interview with Mr. A. Gilodi, Member of Renault’s Corporate Purchasing Management, 7th of June 2001, Technocentre Renault Guyancourt
210
Bart Kamp
Gonzalez Lopez, S. (Avril 2000), Estrategias corporativas y espacios locales: las empresas automotrices en la zona de Toluca, Mexico, Actes du GERPISA, 29, pp. 67-90 Gorgeu, A. and R. Mathieu (1995), Les liens de Renault avec ses fournisseurs: équipementiers et sous-traîtants, Actes du GERPISA, Nr. 14, pp. 41-58 Grabher, G. (1993), The embedded firm: on the socio-economics of industrial networks, London Grohn (2002): Interview with Mr. F. Grohn, Member of Volkswagen’s corporate management for global and forward sourcing, 21st of March 2002, Wolfsburg Håkansson, H. and Shehota, I. (1989), No business is an island: The network concept of business strategy, Scandinavian Journal of Management, 5 (3), pp. 187-200 Hines, P. (1995), Purchasing for lean production: the new strategic agenda, International Journal of Purchasing & Materials Management, 32, pp. 2-10 Hoek, R. Van and Weken, H. (1998), How Modular Production can Contribute to Integration in Inbound and Outbound Logistics, International Journal of Logistics: Research and Applications, 1 (1), pp. 39-56 Jensen, M. and W. Meckling (1976), Theory of the firm: managerial behavior, agency cost, and capital structure, In: Journal of financial economics, 3, October, pp. 305-360 Johanson, J. and L-G. Mattsson (1987), Inter-organizational relations in industrial systems: a network approach compared with the transaction cost approach, International studies of management and organization, XVII (1), pp. 34-38 Kamp, B. (October 2005), Formation and evolution of buyer-supplier relationships: Conceiving dynamism in actor composition of business networks, Industrial Marketing Management, 34 (7), pp. 658-668 Kapoor, V. and Gupta, A. (1997), Aggressive Sourcing: A Free-Market Approach, Sloan Management Review, Fall, pp. 21-31 Klein, B., Crawford, R.G. and A.A. Alchian (1978), Vertical integration, appropriable rents, and the competitive contracting process, Journal of Law and Economics, 21, pp. 297-326 Laage-Hellman, J. (1997), Business networks in Japan: supplier-customer interaction in product development, London: Routledge Lamming, R. (1993), Beyond partnership: strategies for innovation and lean supply, London Lane, P. J., and M. Lubatkin (1998), Relative absorptive capacity and interorganizational learning, Strategic Management Journal, 19, pp. 461–477 Levitt, B. and J.G. March (1988), Organizational learning, In: Annual review of sociology, 14, pp. 319-340 Lieberman, M.B. and D.B. Montgomery (1988), First-mover advantages, Strategic Management Journal, 9 (1), pp. 41-58 Liker, J.K. and Th.Y. Choi (December 2004), Building deep supplier relationships, Harvard Business Review, 82 (12), pp. 104-112 Lincoln, M, Harland, C. and R. Brennan (1998), Developing supplier relationships. British Journal of Management, 8, pp. 57-71 Lorenzoni, G. and C. Baden-Fuller (1993), Creating a Strategic Center to Manage a Web of partners, Working paper, June, University of Bath, School of Management, Bath BA2 7AY Lorenzoni, G. and C. Baden-Fuller (1995), Creating a strategic center to manage a web of partners, California Management Review, 37 (3) Marshall, A. (1890), Principles of economics: an introductory volume, London: MacMillan
Examination of Dedicated Relationships…
211
Maurer, A., Dietz, F. and N. Lang (2004), Beyond cost reduction: Reinventing the automotive OEM-supplier interface, the Boston Consulting Group Millington, A.I., Millington,C.E.S. and Cowburn,M. (1998) “Local assembly units in the motor components industry”, International Journal of Operations & Production Management, 18 (2), pp. 180-194 Monden, Y. (January 1981a), What makes the Toyota production system really tick?, Industrial Engineering, pp. 36-46 Monden, Y. (September 1981b), How Toyota shortened supply lot production time, waiting time and conveyance time, Industrial Engineering, pp. 22-30 Monden, Y. (1992), Cost management of the new manufacturing age: innovations in the Japanese automotive industry, Productivity Press, Cambridge, Mass. Nishiguchi, T. (1993), Governing competitive supplier relations: new auto-industry evidence, MIT/IMVP Annual sponsors’ briefing meeting, Cape Cod (Mass.) Nooteboom, B. (March 2002), A Balanced Theory of Sourcing, Collaboration and Networks, ERIM Report Series Reference No. ERS-2002-24-ORG O hUallachain, B. and D. Wasserman (1999), Vertical integration in a lean supply chain: Brazilian automobile components parts, Economic Geography, 75 (1), pp. 21-43 Parker, P., Rutherford, T. and T. Koshiba (2000), New directions in Canada’s Japanese owned automobile plants, In: Bowles, P. and L.T. Woods (eds.), Japan after the economic miracle: in search of new directions, Kluwer Academic Publishers, Lancaster, UK, pp. 85-103 Penrose, E.T. (1959), The Theory of the Growth of the Firm, New York: John Wiley Pries, L. (Février1999), The dialectics of automobile assemblers and suppliers restructuring and globalization of the German «big three», Actes du GERPISA, Division internationale du travail et relations constructeurs-fournisseurs, pp. 77-92 Pyke, D.F. and M.E. Johnson (2003), The Practice of Supply Chain Management, Kluwer Academic Publishers, pp. 77–90 Régie Nationale des Usines Renault (1986F-2002F), Financial Report 1985-2001, Paris Renault (1991AE-2002AE), Atlas économique 1990-2001, Paris Renault (2001G-2002G), Global Magazine, Nrs. 1-11 (Febrero 2001 – Marzo 2002), Madrid Renault (1986RA-2002RA), Rapport Annuel 1985-2001, Paris Rhys, G. (2000), Supplier parks: Economies of scale play a vital role, Financial Times Survey – Industry briefs, FT Auto 0200/Outsourcing Richardson, J. (1993), Parallel sourcing and supplier performance in the Japanese automobile industry, Strategic Management Journal, 14, pp. 339-350 Rugman, A. (1999), Multinational enterprises and the end of global strategy, Oxford University working paper, Oxford Rugman, A. and J. D’Cruz (2000), Multinationals as flagship firms, New York Rugman, A. and A. Verbeke (2003), Multinational enterprises and clusters, Management International Review, 43 (3), pp. 151-169 Rutherford, T.D. (2000), Japanese investment and buyer-supplier relations in the Canadian automotive industry, Regional Studies, 34 (8), pp. 739-751 Rutherford, T. (2001), Mutual adaptation: Japanese automobile transplants in North America and the restructuring of buyer-supplier relationships, Environments, 29 (3), pp. 73-89 Sako, M. (1992), Prices, quality and trust: Inter-firm relations in Britain and Japan, Cambridge
212
Bart Kamp
Sako, M. and M. Warburton (1999), Preliminary report of European research team Modularization and outsourcing project, MIT/IMVP Annual forum, Boston Shimokawa, K. (1999), Reorganization of automobile industry and structural change of the automobile component industry, MIT/IMVP Publications, Boston SupplierBusiness (August 2003), Volkswagen changes purchasing strategy, Issue 1 Uzzi, B. (1997), Social structure and competition in interfirm networks: the paradox of embeddedness, Administrative Science Quarterly, 42 (1), pp. 35-67 Valeo (2000), Supplier parks: Economies of scale play a vital role, Financial Times Survey – Industry briefs, FT Auto 0200/Outsourcing Veloso, F. (2000), The automotive supply chain organization: Global trends and perspectives, Working Paper, MIT, Cambridge, September 8 Volkswagen (1981-2002), Geschäftsbericht 1980-2001, Wolfsburg Volpato, G. (Février 1997), Fiat Auto and Magnetti Marelli: toward globalization, Actes du GERPISA, Nr. 22, pp. 69-98 Ward’s (several years), Automotive yearbook Wells, P. and M. Rawlinson (1994), The new European automobile industry, New York Wernerfelt, B. (1984), A resource-based view of the firm, Strategic Management Journal, 5 (2), pp. 171-180 Williamson, O.E. (1975), Markets and hierarchies: Analysis and antitrust implications, New York Williamson, O.E. (November 1981), The economics of organization: The transaction cost approach, American Journal of Sociology, 87 (3), pp. 548-575 Williamson, O.E. (1985), The economic institutions of capitalism, New York Womack, J.P., Jones, D.T. and D. Roos (1990), The machine that changed the world, New York
In: Outsourcing, Team Work and Business Management ISBN 978-1-60456-956-8 c 2009 Nova Science Publishers, Inc. Editor: Karl E. Carettas, pp. 213-234
Chapter 12
A LLOCATING O UTSOURCED WARRANTY S ERVICE C ONTRACTS Michelle Opp1∗, Ivo Adan4†, Vidyadhar G. Kulkarni2‡ and Jayashankar M. Swaminathan3§ 1 SAS Institute Inc., Cary, NC 27513 USA 2 Department of Statistics and Operations Research 3 Kenan-Flagler Business School University of North Carolina, Chapel Hill, NC 27599 USA 4 Department of Mathematics and Computer Science Technisch Universiteit Eindhoven, The Netherlands
Abstract Motivated by our interactions with a leading manufacturer of computers, in this paper we consider static allocation as applied to the problem of minimizing the costs of outsourcing warranty services to repair vendors. Under static allocation, a manufacturer assigns each item to one of several contracted repair vendors; every time a particular item fails, it is sent to its preassigned vendor for repair. In our model, the manufacturer incurs a repair cost each time an item needs repair and also incurs a goodwill cost while items are undergoing repair. We model each service vendor as a finite population multi-server queueing system and formulate the resulting outsourcing problem as an integer-variable resource allocation problem. After establishing convexity results regarding the queue lengths at the repair vendors, we show that marginal allocation is optimal. Through a detailed computational study we compare the optimal algorithm with five static allocation heuristics in terms of time and optimality gap. Our study indicates that the optimal algorithm takes less than a minute to solve industry size problems on average. Further, the commonly used heuristics are far away from the optimal on average, thus emphasizing the benefits of the optimal allocation algorithm. We also compare the optimal static allocation to two simple dynamic allocation heuristics. The results of this study further validate the use of static allocation as a justifiable and easy-to-implement policy. Among other computational insights we show ∗
E-mail address: E-mail address: ‡ E-mail address: § E-mail address: †
[email protected] [email protected] [email protected] [email protected]
214
Michelle Opp, Ivo Adan, Vidyadhar G. Kulkarni et al. that when the number of items to be allocated is large, a single-server approximation leads to optimal allocations in most of the cases.
1. Introduction The last decade has seen an explosion in the degree of outsourcing of various business operations that were traditionally performed in-house. Specifically, warranty services, which is a major component of the manufacturing and retail industry, has experienced a rising trend in terms of outsourcing. Several large companies in the past few years have begun to outsource their repair activities. For example, electronic equipment manufacturers routinely augment in-house servicing of warranties by contracting outside vendors to repair the items (machines like personal computers, or components like hard drives) that are sold under warranty. According to a recent report by Merrill Lynch, the warranty services represents a 100 billion dollar opportunity for electronics manufacturers and subcontractors (Serant [19]). Several vendors in the recent past have made large investments to enhance their capabilities in this dimension. For example, Solectron has acquired seven repair companies for mobile phones, desktop computers and notebooks since 1999 (Serant [18]). Outsourcing repair and warranty services provides an opportunity for the original equipment manufacturer (OEM) to improve turnaround times and have better asset utilization by using the core competencies of the vendors specializing in repairs. However, it also poses downward risks in terms of customer satisfaction, particularly in cases where poor experience in repair services may translate into future lost sales. Thus, it is very important for the OEM to not only pay attention to cost reductions but also customer experience while managing the outsourced warranty and repair services. Our motivation for this research comes from extensive discussions with the warranty and repair services division of a prominent local computer manufacturer, although these issues are encountered by other manufacturers in the electronic industry as well. A typical large manufacturer of personal computers (PC) sells them with a warranty that covers parts and labor for the maintenance of the PC for a stipulated period of time. The charges for repairs are spelled out in the contract. Generally, the customer pays no money if the computer is under warranty, and the manufacturer absorbs the entire cost. While there are many important issues in the outsourcing of warranty repair services to subcontracted vendors, we focus mainly on the problem of allocation. That is, when a manufacturer has contracts with several repair vendors, it must decide which vendor will be used to repair each failed item. In particular, we are interested in the problem of minimizing the costs of outsourcing warranty services to alternative repair vendors using static allocation. Under static allocation, each item is assigned to one of the repair vendors; every time a particular item fails, it is sent to its preassigned vendor for repair. In our model, the cost structure is such that the manufacturer incurs a repair cost each time an item needs repair and also incurs a goodwill cost for the delay experienced by a customer while items are awaiting or undergoing repair. Using this cost structure, our model is one of partitioning the customer base into several groups, where each group has its own multi-server queue for service. Therefore, although we have chosen to focus on static allocation of warranty repair services, the model could easily be adapted to many situations in which a finite population of customers is serviced
Allocating Outsourced Warranty Service Contracts
215
by one of several possible service centers. The purpose of this study is to gain insights into the performance of the actual system; the algorithms developed were not designed with the sole intent of being a decision-support model. One may expect that a dynamic model—that is, an item under warranty is not assigned to a particular vendor until the time of failure, and the current workload of each vendor is taken into account when deciding the assignment—would perform better than a static model. However, for problems of practical size, computing the optimal dynamic assigments is intractable. In addition, the static model has significant practical advantages in terms of its simplicity and ease of implementation. For example, the manufacturer does not need to keep track of the current state at each vendor, and does not even need to maintain a centralized call center for handling warranty complaints. Rather, it can delegate this function to the individual repair vendors. We will show through a numerical study that the optimal static allocation of items to outsourced vendors is indeed a justifiable policy when compared to other static allocation heuristics, and also when compared to rather simple (non-optimal) dynamic allocation heuristics, such as join-the-shortest-queue or myopic dynamic allocation. See Opp et al [15] for more details on dynamic policies. Furthermore, there may be situations which necessitate the use of static allocation over dynamic allocation, such as when the manufacturer does not have complete information about the current state of each vendor. For example, the manufacturer would certainly know how many items it has routed to each vendor, but if a vendor does not immediately report the repair completions to the manufacturer, the state information that the vendor possesses would be incorrect. Although the application of static allocation models to warranty outsourcing is new, static allocation models are relatively common in areas such as load balancing (Comb´ e and Boxma [5], Hordijk, Loeve, and Tiggelman [12], Cheng and Muntz [4]), server allocation (Rolfe [17], Dyer and Proll [7], Shanthikumar and Yao [20], [21]), and portfolio selection (Zipkin [23]). Using performance measures such as the mean waiting time of a customer or the total number of customers in the system, Comb´ e and Boxma [5] consider two different types of static allocation — probabilistic allocation and pattern allocation — for an open network with single-server queues. Under probabilistic allocation, a customer is routed to queue i with probability pi , independent of the number of customers in each of the queues. Pattern allocation uses an infinite string of integers (with possibly repeating subsequences) {a1 , . . . , an , . . . }, where an denotes the number of the queue to which the nth customer in the arrival process is routed. Several authors have considered the problem of allocating servers, rather than customers, in a queueing network. Rolfe [17] considers the case of Poisson arrivals (hence, an open queueing network), where the objective is to minimize the expected queueing time of customers in the system. He shows that marginal allocation is optimal in the case of constant service times, and suggests that it is optimal for exponential service times; Dyer and Proll [7] then prove this conjecture. Shanthikumar and Yao [20], [21] consider optimal allocation of servers in a closed network in order to maximize the throughput of the system. Heinhold [10] develops a model to explain the allocation of customers to automobile inspection stations, where the customers are free to choose any inspection station. Heinhold suggests that this approach can be used for similar problems: for example, the allocation of customers to supermarkets. The key difference in this model is that the customer chooses
216
Michelle Opp, Ivo Adan, Vidyadhar G. Kulkarni et al.
to which station he is routed. Our model is most similar to Hordijk, Loeve, and Tiggelman [13], who analyze a closed queueing network in which customers must be assigned to one of two parallel single-server queues, and the routing decision may not depend on the numbers of customers in the queues. However, our model allows an arbitrary number of queues, each of which may have multiple servers. In addition, the particular cost structure of the warranty outsourcing problem includes a fixed cost of routing a customer to a given queue, as well as a holding cost while the customer is awaiting or undergoing repair. We have encountered no other static allocation models incorporating this cost structure. To compute the optimal routing policy, Hordijk, Loeve, and Tiggelman [13] use an algorithm based on successive approximation, which, because of the complex state space, can only be used for a system with a small number of customers (see also Hordijk and Loeve [11]). In Hordijk and Loeve [12], closed queueing networks of quasi-reversible queues are considered, and techniques of linear programming are used to find an optimal deterministic routing of the customers, based on any cost function that depends on the state of the queueing network. In this paper, we show that marginal allocation (similar to Rolfe [17] and Dyer and Proll [7] for server allocation) can be used to compute the optimal static allocation of customers to the queues. The rest of the paper is organized as follows. In Section 2. we present our model and formulate it as a resource allocation problem. In Section 3. we describe our solution method when (i) the objective function is convex; (ii) the objective function is concave; and (iii) the objective function may be neither convex nor concave. In Section 4. we present results from a detailed computational study that compares the optimal algorithm with five static allocation heuristics as well as two simple dynamic allocation heuristics. Section 5. contains concluding remarks and general managerial implications.
2. Model Formulation We model a manufacturer that needs to allocate K identical items (computers in our case) under warranty among V vendors; one of the vendors could be the in-house repair facility of the manufacturer. This closed population assumption is realistic in cases where the manufacturer makes batch sales; for example, the sale of 10,000 computers to a large university. More importantly, however, the closed population model can be used to approximate a system that has, in steady state, a fixed number of identical items covered under warranty. The manufacturer knows the number of servers (si ) and their average rate of repair (µi ) at each of the vendors. We assume that all servers at a single vendor are identical and that the service times are exponential. We also assume that the time between failures for a single item is exponentially distributed with rate λ. The exponential assumption for time between failure and for repair is common in the reliability literature (Blischke and Murthy [2]). Note, however, that the exponential assumption for failure times is not required. For the machine interference model, Bunday and Scraton [3] show that the expected number of machines not running in steady state depends on the failure time distribution only through its mean. Another way to see this result is to think of this particular instance of the machine interference model as a closed network of two stations, one with an infinite number of
Allocating Outsourced Warranty Service Contracts
217
servers and general service times, and the other with s servers and exponential service times. The service time in the infinite-server station models the time until failure for each customer (machine). It is a well-known result in the theory of queueing networks that the limiting distribution of the state of the network depends on the service time distribution in the infinite-server queue only through its mean (Baskett et al. [1]). Therefore, since all we will need in the cost calculation for our model is the expected number of machines not running (that is, the expected number of machines at thes-server station), the results depend only on the mean time to failure and not on the distribution of the time to failure. The contract with vendor i (i = 1, . . . , V ) specifies that the manufacturer will pay the vendor a fixed amount ci for each repair performed by the vendor. Such “fixed fee per repair” contracts, independent of the type of repair and the cost to the vendor, are common when all the materials and parts are charged to the manufacturer and the repair vendor only charges for the labor costs, or for items where the materials and parts costs are minimal compared to the labor costs of repair. In addition, we assume that the manufacturer incurs a goodwill cost at a rate of hi per unit time that an item spends in service at vendor i. This is usually the same for all vendors (i.e., hi = h for i = 1, . . . , V ) since it reflects the loss of goodwill for the manufacturer directly from the customer, independent of the repair vendor used. This goodwill cost prevents the manufacturer from overloading an inexpensive vendor with poor service, which would result in unacceptably large turnaround times for warranty service. Note that by assigning the same goodwill delay cost to all customers, we are assuming that all customers are identical. Under the above assumptions, the manufacturer must decidexi , the number of items to allocate to vendor i, to minimize the expected total warranty cost. There are two types of allocations that are commonly adopted by manufacturers. In a static allocation, the manufacturer projects total volume of sales of a particular product and then allocates all the items under warranty to the different vendors at the beginning; this allocation then remains in effect for the entire contract period. All repairs for a customer are handled by the assigned repair vendor. In a dynamic allocation, as each item experiences a failure, it is dynamically assigned to a vendor based on the current congestions. We focus our attention on the static allocation in this paper, for reasons described in Section 1.. However, in Section 4., we include a computational comparison of the optimal static allocation with simple dynamic allocation heuristics. Before we continue with the formulation of the model, we shall comment briefly on our model assumptions and the applicability of static allocation for outsourcing warranty repair contracts. We have stated that the manufacturer knows the service rate (µi ) and number of servers (si ) at each repair vendor. In reality, capacity for a repair vendor is often shared among several manufacturers. Therefore, we would also have to model a delay at the repair vendor to account for external customers. However, modeling such a delay could become very complicated and would require even more unrealistic assumptions—if it is unlikely that a manufacturer knows the service parameters of a repair vendor, then it is even more unlikely that the manufacturer knows the external rate of arrivals (failures) from other manufacturers that utilize that repair vendor. On the other hand, for very large manufacturers, it is not uncommon for repair vendors to dedicate a portion of their repair facility exclusively to the large manufacturer. We will assume this is the case; in other words, the manufacturer effectively enforces preemptive
218
Michelle Opp, Ivo Adan, Vidyadhar G. Kulkarni et al.
priority in its contracts with outsourced repair vendors. Furthermore, we also assume that the manufacturer does know the service parameters µi and si for each vendor. One can imagine that if a vendor does not perform according to its disclosed service capacity, the manufacturer will not continue to do business with that vendor. Therefore, in steady state, the relationship between the manufacturer and vendors is one that encourages full disclosure of information. Under these assumptions, and given that xi items have been allocated to vendor i, the repair process can be modelled by an M/M/si /∞/xi queue with arrival rate λ, service rate µi , and finite population xi . Such a modeling technique is commonly used for queueing problems with a fixed number of customers; see, for example, Bunday and Scraton [3] and the references therein. Figure 1 shows the population dynamics in this system. Station 1 includes all items that are properly functioning (that is, not undergoing repair); in this figure, there aren such items. These items each fail with rate λ, so items move from station 1 to station 2 with rate nλ. Station 2 is the repair queue, and consists of si servers each having repair rate µi . Therefore, items move from station 2 to station 1 with rate µi min(xi − n, si ), where xi − n is the number of items at station 2 (because there aren items at station 1 and a finite population of xi items). Station 1
Station 2 nλ
-
j1 j2
n
.. .
µi min(xi − n, si )
jsi
Figure 1. Finite-population queueing system.
Let Li (xi ) be the expected number of items at vendor i when xi items are allocated to vendor i (that is, Li (xi ) is the expected number of items at station 2 in Figure 1). Directly computing Li (xi ) from the probability distribution is time-consuming and inefficient. However, the values of Li (xi ) can be computed recursively using mean value analysis, as described in Section 2.1.. Among the xi items that are allocated to vendor i, the expected number of properly functioning items is xi − Li (xi ), and each such item has failure rate λ. Therefore, the expected number of arrivals to the ith vendor per unit time is Λi (xi ) = λ(xi − Li (xi )). The manufacturer’s expected cost per unit time for repairs at the ith vendor includes a per unit cost ci every time there is an arrival to the ith vendor, plus a goodwill cost hi for each item that is waiting at vendor i, and is given as follows: fi (xi ) = ci Λi (xi ) + hi Li (xi ) = λci xi + (hi − λci )Li (xi )
(1)
Allocating Outsourced Warranty Service Contracts
219
2.1. Recursive Calculation of L(x) In this section, we describe how to efficiently compute Li (xi ), which is needed in the cost function fi (xi ) given in (1). Mean value analysis is commonly used to compute statistics such as mean queue sizes, mean waiting times, and throughput in a closed queueing network; see, for example, Reiser and Lavenberg [16]. The principal advantage of mean value analysis is that the product terms and normalization constants do not need to be computed, thereby avoiding problems with numerical stability. For the sake of completeness, in this section we describe the approach of mean value analysis as applied to the static allocation problem. To simplify notation, we drop the subscript i. First consider the M/M/s/s loss model with arrival rate λ and service rate µ. The blocking probability, B(r, s), is given by r s /s! B(r, s) = Ps , i i=0 r /i!
s > 0,
(2)
where r = λ/µ. The blocking probability can be computed using the following recursion, starting with B(r, 0) = 1: B(r, s) =
B(r, s − 1)r/s , 1 + B(r, s − 1)r/s
s > 0.
(3)
Now consider the M/M/s model with arrival rate λ and service rate µ. The probability of waiting, PW (r, s), is given by PW (r, s) =
r s /s! , Ps−1 i (1 − r/s) i=0 r /i! + r s /s!
s > 0.
(4)
It follows from (2) and (4) that PW (r, s) =
B(r, s − 1)r/s , 1 − r/s + B(r, s − 1)r/s
s > 0.
(5)
Next consider a closed two-station network similar to that shown in Figure 1; station 1 is an ample-server station with service rate λ and station 2 is a multi-server station with s servers and service rate µ. Customers move from station 1 to station 2 and from station 2 to station 1 indefinitely. The number of circulating customers isx. The probability that all s servers in station 2 are busy is given by x < s, 0, Px−s i PB (ρ, x) = (6) i=0 ρ /i! , x ≥ s, Px−s i x−s /(x − s)! i=0 ρ /i! + A(ρ, x)ρ where ρ = sµ/λ and s X ρi s(s − 1) · · · (s − i + 1) , A(ρ, x) = (x − s + 1) · · · (x − s + i)si i=1
x ≥ s.
220
Michelle Opp, Ivo Adan, Vidyadhar G. Kulkarni et al.
Note that the number of terms in A(ρ, x) is s, independent of x. From (6) we obtain, for x ≥ s, 1 , PB (ρ, x) = 1 + A(ρ, x)B(sµ/λ, x − s) where B(sµ/λ, x−s) denotes the blocking probability in the loss model withx−s servers. To compute the mean number of customers in station 2, we use the mean value approach. Let Lq (ρ, x) be the mean number of customers waiting in the queue at station 2 when there are x circulating customers in the network, let Wq (ρ, x) be the mean waiting time in the queue at station 2, and let Λ(ρ, x) be the throughput of station 2. Then the following relations hold (which are based on the arrival theorem and Little’s law):
Wq (ρ, x) = Λ(ρ, x) =
PB (ρ, x − 1) + Lq (ρ, x − 1) sµ x 1/λ + Wq (ρ, x) + 1/µ
Lq (ρ, x) = Λ(ρ, x)Wq (ρ, x)
These relations are recursive in the population sizex, starting from Lq (ρ, 0) = PB (ρ, 0) = 0. Then L(ρ, x) can be computed as L(ρ, x) = Lq (ρ, x) + Λ(ρ, x)/µ. When s = 1, L(µ/λ, x) can be simplified as follows: L(µ/λ, x) = x −
µ (1 − B(µ/λ, x)) , λ
where B(µ/λ, x) can be computed recursively using (3).
2.2. Optimization Problem To find the optimal allocation of the K items among the V vendors, the manufacturer solves a resource allocation problem with integer variables (see Gross [9], Fox [8], Ibaraki and Katoh [14]). min
V P
fi(xi )
i=1
s.t.
V P
xi = K
i=1
xi ≥ 0 and integer,
i = 1, . . . , V
To use existing methods for solving this problem, we require convexity properties of the cost functions fi(xi ) = λci xi + (hi − λci )Li (xi ). The convexity is established using the following Corollary; we drop the subscript i for ease of notation. Corollary 1. L(x) is convex with respect to x.
Allocating Outsourced Warranty Service Contracts
221
We first give the intuition behind the proof before stating the formal proof below. We can think of the system at a single vendor as a closed two-station network withx customers, similar to that shown in Figure 1. Station 1 is an ample-server station with service rateλ. Station 2 consists of s servers, each with service rate µ. There are x jobs circulating between the two stations. Let LA (x) denote the mean number of jobs in station 1. To show thatL(x) = x−LA (x) is convex in x, we can show that LA (x) is concave in x, i.e., LA (x) − LA (x − 1) ≥ LA (x + 1) − LA (x).
(7)
Suppose that x − 1 jobs are red and one is blue. In station 2, the red jobs are serviced with preemptive priority over the blue job. Let ρ(x) denote the fraction of time that the blue job spends in station 1. Because the blue job does not exist for the red jobs, it follows that the mean number of red jobs in station 1 is LA (x − 1), and hence ρ(x) = LA (x) − LA (x − 1). Thus to establish (7) it suffices to show that the utilization rate ρ(x) is nonincreasing in x. To do so, we compare the systems with x and x + 1 circulating jobs, each system having exactly one low priority blue job. When an additional red job is added to the system, it increases the blue job’s mean waiting time in station 2, due to the preemptive priority of the red jobs. The blue job spends a larger fraction of time in station 2 and a smaller fraction of time in station 1; therefore, ρ(x) is decreasing in x. The formal proof of Corollary 1 follows directly from the proof of throughput concavity by Dowdy et al. [6]; see also Shanthikumar and Yao [22]. Proof. Consider the closed two-station network with x customers described above. Let T2 be the long-run average throughput at station 2. The long-run average throughput at station 1 is given by T1 = λ(x − L(x)). In a closed network, we must have T1 = T2 ; i.e., T2 = λ(x − L(x)), or L(x) = x − T2 /λ. Dowdy et al. [6] prove that T2 is a concave function of x; hence, it follows that L(x) is a convex function of x. Therefore, from Corollary 1, fi (xi ) is convex if hi ≥ λci , and concave if hi ≤ λci .
3. Solution Method In practice, it is reasonable to assume that hi ≥ λci , because if hi < λci , there is no incentive to repair the item; the cost of holding it at the vendor is smaller than the expected cost to repair it as it fails over time. Therefore, whenhi < λci , the manufacturer’s motivation is not to repair items quickly, but rather to repair items slowly (or not at all) so that they remain in a failed state as long as possible. In doing so, the manufacturer would delay the time until the item fails again, thereby reducing the frequency with which the fixed cost for failure must be paid. However, even though the case hi < λci is extremely rare in practice and can only arise under unusual circumstances, we are interested in a complete solution to the problem under any possible relationship between hi and λci . Therefore, in this section, we present the optimal algorithm for the case when (i) all fi (xi ) are convex; (ii) all fi (xi ) are concave; and (iii) some fi(xi ) are convex and other fj (xj ) are concave.
222
Michelle Opp, Ivo Adan, Vidyadhar G. Kulkarni et al.
3.1. Convex Case When hi ≥ λci for all i = 1 . . . , V , then all fi (xi ) are convex. This problem is the separable convex resource allocation problem. The optimal allocation in this case can be found using marginal allocation, first proposed by Gross [9]: • Step 0: Set xi = 0, i = 1, . . . , V • Step 1: Choose a j ∈ argmin{fi (xi + 1) − fi(xi )} i=1,...,V
• Step 2: Set xj = xj + 1 • Step 3: If
V P
xi < K go to Step 1; else, stop.
i=1
An important implication of the optimality of marginal allocation is that a new customer who enters the system can be allocated without recalculating the optimal allocation for the K original customers. That is, when a new customer is added to the system, we continue to assume a closed queueing network, but with K + 1 customers instead of K customers. If we have already found the optimal allocation ofK customers, we need to perform only one more iteration of the marginal allocation algorithm to determine the allocation of the new customer. Note, however, that this does not imply that marginal allocation works for all dynamically changing population sizes. For example, if a customer leaves the system, the optimal allocation might be to reassign customers to different vendors, but reassignment is not allowed in our static allocation model.
3.2. Concave Case When hi ≤ ci λ for all i = 1, . . . , V , then all fi(xi ) are concave. The objective function is concave, and the optimum occurs at an extreme point. Therefore, the solution method is trivial: choose a vendor j for which fj (K) is minimum, and allocate all K items to that vendor.
3.3. Mixed Case When at least one vendor i has hi ≥ ci λ, and at least one vendor j has hj < cj λ, the objective function is neither convex nor concave. Let A+ = {i : hi ≥ ci λ} and let A− = {ii : h < ci λ}; that is, A+ is the set of vendors for which the corresponding objective term is convex, and A− is the set of vendors for which the corresponding objective term is concave. For a fixed value k ∈ {0, . . . , K}, consider the following problem. P P fi (xi ) + fi (xi ) z(k) = min i∈A+
s.t.
P
i∈A−
xi = k
i∈A+
P
(8)
xi = K − k
i∈A−
xi ≥ 0 and integer,
i = 1, . . . , V
Allocating Outsourced Warranty Service Contracts
223
Problem (8) gives us the optimal solution for a fixed k; we must then optimize over k = 0, . . . , K to obtain the optimal solution to the original problem, z∗ =
min z(k).
(9)
k=0,...,K
Note, however, that solving (8) is equivalent to separately solving problems (10) and (11), below: P fi (xi ) z + (k) = min i∈A+
s.t.
P
xi = k
(10)
i∈A+
xi ≥ 0 and integer, z − (k) = min
P
fi (xi )
P
xi = K − k
i ∈ A+
i∈A−
s.t.
(11)
i∈A−
xi ≥ 0 and integer,
i ∈ A−
Therefore, for a fixed value k, problem (10) can be solved using marginal allocation as in Section 3.1.. Furthermore, the computation of the optimal allocation fork = K requires the computation of the optimal allocation for allk < K, so finding the optimal solutions for all k = 0, . . . , K is no more difficult than finding an optimal solution fork = K (although it does require more storage). Similarly, for a fixed value K − k, problem (11) can be easily solved: choose a vendor j ∈ A− for which fj (K − k) is minimum, and allocate all K − k items to that vendor.
4. Computational Study In this section, we present results from our computational study. Our main aim in this numerical study was (i) to test the ability of the optimal algorithm to handle industry size problems; (ii) to compare the performance of the optimal algorithm with commonly used heuristics; (iii) to develop insights on the effect of service rates of vendors on the optimal allocation; and (iv) to test the effect of utilizing a single-server approximation of vendors on the optimal allocation. Based on our interaction with our industrial contact we constructed a data set that adequately represented the complexity in the real environment. In particular, our parameter choices are as follows. We consider examples with four vendors (V = 4), a fixed failure rate of λ = 1.2 failures per year, and a fixed holding cost of h = $1000 per year. For each value of K ∈ {100, 1000, 10000, 100000} and for each vendor i = 1, . . . , 4, we choose a low value of µi , denoted µi , and a high value of µi , denoted µi . Similarly, we choose low and high values of ci and si , denoted ci , ci , si , and si , respectively. We then create a single trial for each combination of µi , si, and ci such that µi ∈ {µi , µi }, ci ∈ {ci , ci }, and
224
Michelle Opp, Ivo Adan, Vidyadhar G. Kulkarni et al.
si ∈ {si , si } (i = 1, . . . , 4). This gives us 212 = 4096 trials for each value of K, for a total of 16,384 trials. We then repeated this process five times with different low and high values for each parameter, for a total of 81,920 trials. The data is given in Table 8 through Table 11 in the Appendix.
4.1. Running Time of Optimal Algorithm Table 1 gives the average running time, in seconds, for computing the optimal allocation for each value of K. Note that the running time is approximately linear in K. Our results indicate that the optimal algorithm is very fast and can handle industry size problems within a minute in most cases. Table 1. Mean time to compute the optimal allocation K 100 1,000 10,000 100,000
Time (in seconds) 0.0233 0.2071 2.1349 24.5890
4.2. Heuristics In this section we compare the optimal static allocation to several common allocation heuristics. Section 4.2.1. first focuses on optimal static allocation as compared to other (non-optimal) static allocation heuristics. The running times for these heuristics is negligible; therefore, we investigate the optimality gap when using these very fast heuristics. Section 4.2.2. then compares optimal static allocation to two common dynamic allocation heuristics. The cost of using either of the dynamic allocation heuristics is computed using simulation. Therefore, rather than using the entire set of data described in Section 4., we make comparisons based on 50 examples with K = 100 and 50 examples with K = 1,000. 4.2.1. Static Heuristics We chose five common static allocation heuristics for comparison with the optimal static allocation algorithm. • H1: Equal allocation to all vendors. Note that if K/V is fractional, the allocation is bK/V c to each of the first V − K + V bK/V c vendors, and bK/V c + 1 to each of the remaining K − V bK/V c vendors. This allocation policy might be used when the manufacturer has little or no information about each vendor. • H2: Allocation is proportional to 1/ci , which favors low-cost vendors. This allocation might be used when the manufacturer has cost information about the vendors,
Allocating Outsourced Warranty Service Contracts
225
but does not have good estimates of vendors’ service rates. Note that if the allocation is fractional, the allocations to vendors 1, 2, and 3 are rounded to the nearest integer, and the allocation to vendor 4 is adjusted accordingly so the total allocation equals K. • H3: Allocation is proportional to si µi /ci , which favors vendors with low cost ci and high maximum expected service rate si µi . This allocation policy might be preferred when the vendor has accurate estimates both of vendors’ costs and service rates. Note that fractional allocations are adjusted as in H2. • H4: All items are allocated to the vendor with the smallest cost ci . This heuristic is an “all-or-nothing” version of H2. • H5: All items are allocated to the vendor with the largest value of si µi /ci . This heuristic is an “all-or-nothing” version of H3. The time required to compute an allocation based on any of these heuristics is negligible; however, there is a significant loss of optimality, as summarized in Tables 2 through 4. Table 2 gives the number of times that each heuristic (H1,. . .,H5) performed best out of the five heuristics. The column sums are greater than 20,480, the number of trials for each value of K, since some trials had two or more heuristics giving the same allocation. Table 2 shows that heuristics H4 and H5 generally performed best among the five heuristics. For example, for the case K = 100,000, heuristic H4 (send all items to the lowest-cost vendor) performed at least as well as the other heuristics in approximately 72.5% of the trials. Table 2. Number of times each heuristic performed best (among all 5 heuristics) Heuristic
K = 100
K = 1,000
K = 10,000
K = 100,000
H1
235
966
419
519
H2
925
1213
1045
1275
H3
1590
558
1556
1338
H4
12471
14998
14876
14858
H5
12331
10348
9376
9891
Table 3 gives the number of times (out of 20,480) that the cost produced by each heuristic’s allocation was exactly equal to the optimal cost. Again, heuristic H4 performs best. For K = 100, heuristic H4 gives the optimal cost in nearly 49% of the trials, and for K = 100,000, it gives the optimal cost in 72.5% of the trials. The values in Table 3 are zero for heuristics H1, H2, and H3, because these heuristics produce proportional allocations; it is unlikely that the optimal allocation will be exactly proportional to the problem data. From the results in Table 2 and Table 3, it may seem that heuristic H4 is a suitable heuristic for solving this allocation problem. However, Table 4 shows the mean relative
226
Michelle Opp, Ivo Adan, Vidyadhar G. Kulkarni et al. Table 3. Number of times each heuristic’s cost was equal to the optimal cost Heuristic
K = 100
K = 1,000
K = 10,000
K = 100,000
H1
0
0
0
0
H2
0
0
0
0
H3
0
0
0
0
H4
9976
14432
14528
14848
H5
9512
8174
6763
7401
difference between the costs of the heuristic allocation and the optimal allocation. For K = 100,000, the cost of heuristic H4’s allocation is, on average, 166% away from optimal. In other words, heuristic H4 performs well quite often, but can also perform extremely poorly. This is a result of H4’s myopic “all-or-nothing” structure. It chooses a single vendor based on cost parameters alone, and if the cheapest vendor also happens to be very slow, the results can be disastrous. Table 4 shows that heuristic H5 appears to be the best heuristic overall, but even this allocation method is anywhere from 23% to 45% away from optimal. Table 4. Mean relative difference between costs of heuristic and optimal allocations Heuristic
K = 100
K = 1,000
K = 10,000
K = 100,000
H1
0.8354
0.8313
1.1189
1.0345
H2
0.6790
0.6409
0.8819
0.8056
H3
0.3337
0.4768
0.4775
0.5206
H4
1.1574
1.5569
1.4259
1.6621
H5
0.2296
0.3998
0.4214
0.4495
4.2.2. Dynamic Heuristics Under dynamic allocation, an item under warranty is not assigned to a particular vendor until the time of failure, and the current workload of each vendor is taken into account when deciding the assignment. One can explicitly formulate the dynamic allocation problem as a continuous time Markov decision process. Due to the size and nature of the state space, however, the problem of finding an optimal dynamic policy is intractable. In this section, therefore, we compare the performance of the optimal static allocation to two simple dynamic allocation heuristics. In Opp, Glazebrook, and Kulkarni [15], we use techniques of policy improvement and restless bandit models to develop more sophisticated dynamic allocation policies than the simple heuristics used in this section.
Allocating Outsourced Warranty Service Contracts
227
We chose the following simple dynamic allocation heuristics for comparison with the optimal static allocation algorithm: • Join the Shortest Queue (JSQ): At the time of failure, an incoming item is sent to a vendor with the shortest queue length. If more than one vendor has minimal queue length, the item is sent to the vendor among them with a smallest value of the fixed cost, ci . • Individually Optimal (IO): At the time of failure, an incoming item is sent to a vendor for which the total cost associated with that particular item alone is minimal; in other words, this heuristic myopically routes the incoming items. Specifically, the item is sent to a vendor with the smallest value of IOi (xi ), where hi xi < si , ci + , µi IOi (xi ) = (x + 1)hi ci + i , xi ≥ si . si µi The relative difference between the costs of optimal static allocation and the two dynamic allocation heuristics are summarized in Table 5. For example, for the 50 examples with K = 100, the cost of the optimal static algorithm ranged from 27.55% lower to 17.11% higher than the cost of using join-the-shortest-queue, with an average of 0.21% higher. For the trials with K = 100, the optimal static allocation performed, on average, about as well as the join-the-shortest-queue heuristic, and about 4% worse than the individually optimal dynamic heuristic. However, for the 50 trials with K = 1,000, the optimal static allocation performed about 1.7% worse than the join-the-shortest-queue heuristic, but more than 5% better than the individually optimal dynamic heuristic. Table 5. Relative difference between costs of optimal static allocation and dynamic allocation heuristics, using 50 examples with K = 100 and 50 examples with K = 1,000 K = 100
K = 1,000
JSQ
IO
JSQ
IO
Min
-0.2755
-0.0479
-0.2297
-0.2782
Max
0.1711
0.1822
0.1325
0.0433
Mean
0.0021
0.0396
0.0169
-0.0555
Std. Dev.
0.1026
0.0481
0.0552
0.0938
These results show that one should not blindly use dynamic allocation, with the assumption that any dynamic allocation policy will perform uniformly better than static allocation. In fact, for many instances, one does not lose much by using static allocation rather than
228
Michelle Opp, Ivo Adan, Vidyadhar G. Kulkarni et al.
simple dynamic allocation, provided one uses the optimal static allocation. Given the simplicity and ease of implementation of static allocation policies as compared to dynamic allocation, it seems that optimal static allocation is, indeed, an attractive policy for minimizing the costs of outsourcing warranty repair services.
4.3. Single-Server Approximation In finding the optimal static allocation, most of the computation time involves the calculation of L(x) required in the cost function. For the multi-server case, the calculation for L(x) in Section 2.1. is quite complicated; however, this is greatly simplified when there is only a single server. Therefore, in this section we explore the tradeoff between loss of optimality and increased speed of calculation when approximating a vendor withsi servers and rate µi by a vendor with a single server and service rate si µi . First, note that for all values of K in the computational study reported in Section 4., the values of si for each of the four vendors were restricted to the interval from 1 to 20. This is to model the realistic setting where a vendor may have several repairpeople—it is quite common for a vendor to have up to a few dozen repair employees, but it is difficult to imagine any situation where a repair vendor has thousands, or even hundreds, of employees assigned to a single manufacturer. In this section, we will extend this a bit further, tos = 50, in order to show what will happen with the single-server approximation of the cost of an individual vendor as the number of servers increases within a realistic range. For example, suppose the failure rate is 1.2 failures per item per year and the holding cost is $1,000 per year, and consider a repair vendor that has a fixed cost of $100 per repair and an individual service rate of 12,500 repairs per year. Suppose the number of servers ranges from 1 to 50; when there is only a single server, the maximum number of items that the vendor can handle in order to satisfy (infinite-population) stability conditions is 10,416. Figure 3 shows the relative difference between the true cost (using the exact calculation for L(x)) and the approximate cost (using a single-server approximation in the calculation for L(x)) as a function of s when there are 10,000 items assigned to the vendor. The gap clearly increases with s; however, in even the worst case (corresponding to s = 50), the difference between the true cost and the approximation is less than 0.07%. Furthermore, plots of relative difference between the two costs even for thex = 100 and x = 1,000 are nearly identical to Figure 3; therefore, the plots are not included here. However, both of these cases also have a maximum relative difference of approximately 0.07%. In other words, using a single-server approximation to compute L(x) provides a very close approximation of the cost function, provided the number of servers is within a realistic range. As seen in Figure 3, the cost function for a single vendor does not change much when using a single-server approximation. Therefore, one would expect that the allocation to the vendors also will show little change if we use a single-server approximation for each of the repair vendors. For the 81,920 trials described in Section 4., we determine the approximate policy using a single server for each vendor and service rate si µi , i = 1, . . . , V , and we then compute the cost to the manufacturer of following this approximate policy. In 72.9% of the experiments, the allocation under the single-server approximation was exactly the
Allocating Outsourced Warranty Service Contracts
229
Figure 3. Relative difference between true cost and single-server approximate cost.
same as the optimal allocation. Furthermore, this number increases withK: for K = 100, it is 57.1%, and for K = 100,000, it is 79.7%. In only 13.2% of the experiments was the difference between the optimal and approximate allocations for any vendor greater than 10, and this number decreases with K: for K = 100, it is 28.2%, and for K = 100,000, it is only 5.1%. Table 6 shows the mean relative difference between the cost of the optimal allocation and the cost of using the allocation prescribed by the single-server approximation. For K = 100, the cost of the single-server approximate allocation is about 2% away from optimal. However, for the larger values of K, this difference is negligible (e.g., 9.06E-9 for K = 100,000). Table 7 gives the average running time for determining the single-server approximate allocation. Note that the running time is again nearly linear in K, and is approximately 62% to 68% faster (depending on K) than finding the optimal allocation. However, this includes only the time required to determine the allocation, and not the time required to compute the true expected cost of using this allocation. Computing the true expected cost of the allocation would bring the total running time near, or even above, that required to compute the optimal allocation (which automatically computes the expected cost). Therefore, this approximation may be advantageous when an approximation to the expected cost of the policy is sufficient. In addition, the results in this section show that the single-server approximation is also adequate in cases where the manufacturer may not have detailed in-
230
Michelle Opp, Ivo Adan, Vidyadhar G. Kulkarni et al.
Table 6. Mean relative difference: cost of optimal allocation vs. cost of single-server approximate allocation K
Rel. Diff
100
0.02137
1,000
0.00036
10,000
1.58E-6
100,000
9.06E-9
formation about si and µi for each vendor, but rather knows only the overall service rate at each vendor. Table 7. Mean running time to determine single-server approximate allocation K
Time 100
0.0090
1,000
0.0795
10,000
0.7943
100,000
7.9280
4.4. Effect of Service Rates Finally, we compare the optimal allocations when we change all vendors from relatively slow service rates to faster service rates. That is, we compare µ = (µ1 , µ2 , µ3 , µ4 ) with µ = (µ1 , µ2 , µ3 , µ4 ), keeping all other data unchanged. This gives us 5(28 ) = 1280 experiments for each value of K, for a total of 5120 trials. In approximately 43% of the 5120 trials, the allocation was exactly the same using µ or µ. In particular, the optimal solution using µ was 100% allocation to a single vendor. Therefore, one vendor was already fast enough to handle all items at relatively low cost, so increasing the service rates of the vendors did not change the allocation. In the remaining 57% of the trials, the allocation did change when usingµ as compared to µ. The pattern we see from these results is that increasing the service rate for all vendors serves to shrink the vendor base. That is, if allocation is positive for exactlyn of the V vendors when the service rates are given by µ, then the allocation will be positive for m ≤ n vendors when the service rates are given by µ. For example, for K = 100 the allocation might switch from x∗ = (11, 35, 31, 23) when the service rates are given by µ to x∗ = (0, 0, 42, 58) when the service rates are given by µ. This is intuitive, since increasing the service rates means the manufacturer can allocate more items to any given vendor without significantly increasing the waiting times at the vendor; hence, fewer vendors will
Allocating Outsourced Warranty Service Contracts
231
be required to handle all K items. Therefore, the manufacturer has an added advantage when the vendors have fast service rates: not only are the holding costs lower, but the manufacturer will require fewer vendors, which will likely reduce administration costs of maintaining the contracts.
5.
Conclusions
In this paper, we consider the static allocation of items under warranty to alternative repair vendors. Modeling the costs to include both a fixed cost for repair as well as a goodwill holding cost to discourage extraordinarily long turnaround times, we develop an efficient optimal algorithm, where each vendor is modeled as a multi-server finite population queue. This algorithm is not specific to the application of outsourcing warranty repairs, but can be used for any static allocation problem that follows this cost and decision structure. Through a detailed computational study we demonstrate that the optimal algorithm can handle real problem size and also performs much better than common static allocation heuristics. Furthermore, we show that the performance of the optimal static allocation is comparable to that of two simple dynamic allocation heuristics, join-the-shortest-queue and myopic routing. When factoring in the simplicity of finding the optimal static allocation, as well as the relative ease of implementing a static allocation policy, we see that optimal static allocation is indeed a viable choice for the manufacturer seeking to minimize outsourcing costs associated with warranty repairs. Furthermore, there are often situations in which static allocation is is the only option, when the manufacturer does not have reliable information about the current state at each repair vendor. Among other computational insights we show that when the number of items to be allocated is large, a single-server approximation leads to optimal allocations in most of the cases. This approximation may be necessary when the manufacturer does not have detailed information about a vendor (such as the number of servers and individual service rate), but rather knows only the overall service rate, or effective capacity, of the vendor. Note, however, that this approximation is valid if the number of servers is comparatively small; when the number of servers is very large compared to the number of items allocated to the vendor, an infinite-server approximation may be more appropriate than a single-server approximation. There are many possible future directions to this research, which we plan to explore in separate papers. First, we assume all customers are of the same type. In many environments, firms have different types of customers who are promised different levels of service. One could extend the static allocation model to include multiple priority classes of customers, where the holding cost for a customer depends on his priority class. Second, we assume that a fixed fee contract is given. One could alternatively think of this as a contracting problem where vendors and the manufacturer negotiate the service levels and the fees. Presumably, there is a tradeoff between service and cost, so a faster service rate at a vendor would command a higher fee from the manufacturer. In such a situation, the manufacturer’s allocation decision should be determined jointly with the optimal service rate and cost parameters. Another key assumption of this paper is that the manufacturer has a fixed number of items to allocate to the repair vendors, which results in a closed population model. In
232
Michelle Opp, Ivo Adan, Vidyadhar G. Kulkarni et al.
reality, the number of items covered under warranty might not be fixed. For example, items enter the population as they are sold with a warranty, while other items exit the population as the warranty coverage expires. We plan to explore static allocation as it applies to a changing population size. That is, each item is assigned to a vendor at the point of sale, and every time an item experiences a failure, it is sent to its assigned vendor for repair. When allocating a new item, the manufacturer would know how many items are already allocated to each vendor; however, the number of items actually being repaired at each vendor is not observed.
Appendix The data for the computational study in Section 4. are shown in the following four tables. Table 8. Data for K = 100 Group 1 Low High
Group 2 Low High
Group 3 Low High
Group 4 Low High
Group 5 Low High
c1 c2 c3 c4
61.96 67.78 55.60 46.74
157.95 139.79 162.36 173.32
53.74 67.26 60.16 52.48
142.18 131.73 159.83 184.44
55.21 57.36 56.67 47.14
156.39 138.43 153.66 171.87
68.74 62.44 59.99 52.52
136.34 142.75 156.30 184.59
67.92 59.81 58.92 50.79
148.64 139.28 147.37 176.17
µ1 µ2 µ3 µ4
7.86 36.31 24.37 37.53
114.84 121.02 125.54 160.63
17.31 23.67 31.14 30.84
106.76 125.29 125.85 153.60
17.24 30.14 23.89 28.18
103.86 112.45 128.15 162.53
28.10 29.09 22.14 32.65
122.55 112.73 126.12 162.77
19.62 33.22 25.88 28.06
122.05 120.57 123.27 155.17
s1 s2 s3 s4
3 2 2 1
11 15 13 5
1 1 4 2
11 17 14 6
3 1 5 1
10 12 16 8
1 1 3 2
11 14 13 9
1 1 5 1
11 16 13 7
Table 9. Data for K = 1,000 Low
Group 1 High
Low
Group 2 High
Low
Group 3 High
Low
Group 4 High
Low
Group 5 High
c1 c2 c3 c4
57.22 61.60 57.22 51.80
151.98 132.25 150.48 174.18
57.08 56.37 51.51 49.09
148.10 135.24 158.45 172.90
56.02 59.84 56.18 49.43
144.55 134.34 162.12 164.35
65.72 59.99 51.55 47.84
156.69 145.78 156.70 170.03
61.70 60.98 56.96 47.94
147.00 146.47 160.59 180.51
µ1 µ2 µ3 µ4
192.56 354.42 260.29 293.02
1505.98 1802.65 1872.17 2789.58
164.23 351.78 266.40 300.54
1494.62 1831.40 1976.67 2818.41
172.24 363.88 281.11 316.07
1484.92 1886.53 1920.60 2800.25
161.62 352.22 256.34 296.71
1474.44 1839.73 1881.34 2828.71
183.77 327.16 246.10 292.63
1504.23 1867.52 1933.35 2786.68
s1 s2 s3 s4
1 1 2 1
8 11 20 8
3 2 4 1
11 16 17 9
3 2 4 1
12 12 7 7
2 2 3 1
10 15 10 8
3 2 2 2
9 14 14 9
Allocating Outsourced Warranty Service Contracts
233
Table 10. Data for K = 10,000 Group 1 High
Low
Low
Group 2 High
Low
Group 3 High
Low
Group 4 High
Low
Group 5 High
c1 c2 c3 c4
61.47 62.91 69.78 53.30
159.59 142.89 151.38 188.90
60.00 62.03 52.97 59.50
157.77 142.18 160.27 161.24
63.28 63.68 57.38 55.38
145.57 135.74 154.11 170.13
58.65 66.60 53.48 55.30
156.70 129.53 147.12 172.95
58.07 61.72 53.33 45.85
151.52 138.93 157.18 164.70
µ1 µ2 µ3 µ4
2122.30 1880.42 2206.53 2161.15
17961.00 21047.19 16014.71 19811.02
2064.06 1835.40 2216.44 2132.91
18055.41 20907.20 15964.07 19899.36
2107.26 1849.31 2207.74 2204.27
18005.26 20976.57 15924.84 19866.39
2083.28 1822.75 2221.18 2140.53
18081.09 20989.56 16046.04 19937.76
2067.61 1848.79 2204.84 2196.55
18066.65 20940.04 15987.73 19923.36
s1 s2 s3 s4
1 1 5 2
10 15 16 8
1 2 4 1
12 16 7 5
1 1 3 1
10 14 11 7
1 2 5 2
12 17 9 5
2 1 5 2
8 12 12 6
Table 11. Data for K = 100,000 Group 1 Low
Group 2 High
Low
Group 3 High
Low
Group 4 High
Low
Group 5 High
Low
High
c1 c2 c3 c4
68.48 68.00 46.53 47.23
144.85 139.92 167.47 178.57
57.12 58.34 64.81 48.11
135.99 153.69 154.85 180.42
61.31 65.25 54.10 48.61
160.78 140.77 162.40 172.51
60.54 68.89 57.21 55.84
145.40 140.03 166.45 177.11
53.64 61.60 57.39 53.83
136.95 138.17 149.61 174.74
µ1 µ2 µ3 µ4
21036.39 23030.79 20095.03 20943.71
222033.28 191618.05 207623.40 289047.94
21322.31 25869.71 19963.75 21419.29
218480.41 188166.68 189419.52 296212.16
21274.77 24899.77 15707.37 19386.49
222887.26 192447.63 212650.20 290130.16
21455.08 22520.36 18505.10 21519.96
228451.93 197853.31 198175.58 302640.41
21105.24 23503.24 19055.07 21386.20
232381.33 201670.44 221346.82 298925.98
s1 s2 s3 s4
3 2 4 1
11 15 14 6
2 2 2 2
11 13 20 9
2 1 4 1
9 12 14 7
1 2 5 1
8 12 20 5
3 1 4 1
8 14 19 6
References [1] F. Baskett, M. Chandy, R. Muntz, and F. Palacious. Open, closed, and mixed networks of queues with different classes of customers. Journal of the ACM, 22:248–260, 1975. [2] W. R. Blischke and D. N. P. Murthy. Warranty Cost Analysis. Marcel Dekker, New York, 1994. [3] B. D. Bunday and R. E. Scraton. The G/M/r machine interference model. European Journal of Operational Research, 4:399–402, 1980. [4] W. C. Cheng and R. R. Muntz. Optimal routing for closed queueing networks. Performance Evaluation, 13:3–15, 1991. [5] M. B. Comb´e and O. J. Boxma. Optimization of static traffic allocation policies. Theoretical Computer Science, 125:17–43, 1994.
234
Michelle Opp, Ivo Adan, Vidyadhar G. Kulkarni et al.
[6] L. W. Dowdy, D. L. Eager, K. D. Gordon, and L. V. Saxton. Throughput concavity and response time convexity. Information Processing Letters, 19:209–212, 1984. [7] M. E. Dyer and L. G. Proll. On the validity of marginal analysis for allocating servers in m/m/c queues. Management Science, 23:1019–1022, 1977. [8] B. L. Fox. Discrete optimization via marginal analysis. Management Science, 13(3):210–216, 1966. [9] O. Gross. A class of discrete type minimization problems. Technical Report RM1644, RAND Corp., 1956. [10] M. Heinhold. An operational research approach to allocation of clients to a certain class of service institutions. J. Opl. Res. Soc., 29:273–276, 1978. [11] A. Hordijk and J. A. Loeve. Undiscounted Markov decision chains with partial information: an algorithm for computing a locally optimal periodic policy. ZOR – Mathematical Methods of Operations Research, 40:163–181, 1994. [12] A. Hordijk and J. A. Loeve. Optimal static customer routing in a closed queueing network. Statistica Neerlandica, 54:148–159, 2000. [13] A. Hordijk, J. A. Loeve, and J. Tiggelman. Analysis of a finite-source customer assignment model with no state information. Mathematical Methods of Operations Research, 47:317–336, 1998. [14] T. Ibaraki and N. Katoh. Resource Allocation Problems: Algorithmic Approaches. MIT Press, Cambridge, Mass., 1988. [15] M. Opp, K. D. Glazebrook, and V. G. Kulkarni. Outsourcing warranty repairs: Dynamic allocation. Naval Research Logistics, 52(5):381–398, 2005. [16] M. Reiser and S. S. Lavenberg. Mean-value analysis of closed multichain queuing networks. Journal of the Association for Computing Machinery, 27(2):313–322, 1980. [17] A. J. Rolfe. A note on marginal allocation in multiple-server service systems. Management Science, 17:656–658, 1971. [18] C. Serant. EMS Companies Going in for Repairs. EBN, Dec. 17, 2001. [19] C. Serant. Solectron to Provide Xbox Support. EBN, Oct. 22, 2001. [20] J. G. Shanthikumar and D. D. Yao. Optimal server allocation in a system of multiserver stations. Management Science, 33:1173–1180, 1987. [21] J. G. Shanthikumar and D. D. Yao. On server allocation in multiple center manufacturing systems. Operations Research, 36:333–342, 1988. [22] J. G. Shanthikumar and D. D. Yao. Second-order properties of the throughput of a closed queueing network. Mathematics of Operations Research, 13:524–534, 1988. [23] P. H. Zipkin. Simple ranking methods for allocation of one resource. Management Science, 26:34–43, 1980.
In: Outsourcing, Teamwork and Business Management ISBN: 978-1-60456-956-8 Editor: Karl E. Carettas, pp. 235-265 © 2009 Nova Science Publishers, Inc.
Chapter 13
THE IMPORTANCE OF CONTEXT IN DETERMINING CONSUMER RESPONSE TO FOOD SAFETY EVENTS: THE CASE OF MAD COW DISEASE DISCOVERY IN CANADA, JAPAN AND THE UNITED STATES1 Sayed Saghaian, Leigh Maynard and Michael Reed University of Kentucky USA
Introduction No food safety threat inspires more dread than variant Creutzfeldt - Jakob disease (vCJD), an irreversible brain wasting disease contracted from eating beef infected with Bovine Spongiform Encephalopathy (BSE). Consumers often refer to vCJD and BSE interchangeably as “mad cow disease,” which can induce fear through uncertain identification, long incubation periods, and devastating symptoms. Even though most countries have experienced very few BSE cases, such as Japan, the United States, and Canada, and the risk may be exceedingly low, previous studies found conflicting consumer responses to BSE in these low-incidence countries. This chapter helps reconcile disparities in previous findings by demonstrating the importance of context in determining consumer reactions to a BSE scare. The results help focus attention on strategic risk management options that agribusiness managers can use to guard against shocks to retail beef demand if and when future BSE events occur. This chapter consists of three parts that present two complementary statistical analyses. In the first part, we show consumer reactions to BSE in Japan using Directed Acyclic Graphs and historical price and quantity decompositions. The Japanese beef markets faced two subsequent cases of BSE discoveries in 2001, eroding consumer confidence in beef supply channels with huge economic losses to the Japanese beef industry. In the second part, we look at BSE’s impact along the U.S. supply chain using similar contemporary time-series methods. The U.S. beef industry faced BSE in 2003, which led to differential impacts on farm, 1
This chapter is Journal Paper Number 08-04-028 of the Kentucky Agricultural Experiment Station.
236
Sayed Saghaian, Leigh Maynard and Michael Reed
wholesale, and retail markets. Relative to the U.S., Japanese consumers have a strong preference for domestically produced beef, encouraged by retail country-of-origin labeling and BSE media coverage critical of imported beef. Consistent with these differences in preferences, marketing, and information, we observe more negative and more nuanced reactions to BSE in Japan versus the U.S. The third part highlights contextual differences in Canada. A double-hurdle model of Canadian fast food beef purchases shows no significant BSE impacts on the likelihood or quantity of fast food beef item purchases. When applied to Canadian supermarket beef purchases, however, a striking pattern emerges. After the initial BSE event in 2003, when media coverage focused mainly on the plight of ranchers, beef demand increased significantly. Moreover, demand increased the most in Alberta, the center of Canada’s beef industry. Following two later BSE events, beef demand fell significantly. The results illustrate the importance of context along at least five dimensions: the food purchase venue, the geographic proximity of consumers to BSE events, the ordering of BSE events, the role of supplier behavior, and the nature of media coverage.
Part I: BSE Discoveries in Japan The first case of BSE in Japan was reported in September 2001. BSE is a fatal neurological disease which typically occurs in adult animals. BSE is primary transmitted by feeding of diseased animal products. Consumption of contaminated beef by humans is suspected to cause vCJD. BSE discovery in Japan caused considerable economic damage to the Japanese beef and food service industries, in part due to the actions of Japanese officials that eroded consumer confidence (McCluskey, et al. 2004). The Japanese government’s response to the crisis was an aggressive marketing campaign promoting the safety of Japanese beef, which had salient impacts on Japanese consumer reactions (Fox and Peterson, 2002). The impact of food safety scares has been extensively investigated in the literature. These studies generally show that food safety scares affect prices and demand adversely, and that consumers are willing to pay higher premiums for safety and quality assurance (e.g., Marsh, Schroeder and Mintert, 2004; Piggott and Marsh, 2004; McCluskey, et al., 2004; Peterson and Chen, 2005; Livanis and Moss, 2005; and Chopra and Bessler, 2005). Increased awareness of food safety problems appears to create opportunities for branding, labeling, and product differentiation through traceability and beef quality. Labeling of credence attributes for beef reduces information costs to consumers and can result in increased demand for qualityassured products. Beef producers and retailers can differentiate their beef products to earn higher premiums, by using beef safety and quality as a strategic response to consumers’ beef safety concerns. We investigated the impact of BSE events on Japanese retail meat demand with beef differentiated by quality and country of origin. We used a cointegrated vector error correction (VEC) model, directed acyclic graphs, and historical decomposition to find Japanese consumer responses to the sudden, unexpected beef safety scares. Directed graphs allow the errors among the endogenous variables to be incorporated into the forecasted effects of these meat market shocks over time. We traced the dynamic effects of these shocks on retail-level meat prices and quantities over time to see if these changes were consistent with wellinformed consumer behavior.
The Importance of Context in Determining Consumer Response…
237
The Data Fish, poultry, and four beef types differentiated by type and country of origin, namely, U.S., Australian, Japanese wagyu, and Japanese dairy beef, were evaluated. The sample data for fish and poultry were not distinguished by source of origin. The sample contained 105 observations from April 1994 to December 2002. Retail beef prices and quantities were obtained from Agriculture and Livestock Industries Corporations (ALIC) data. Beef prices were the weighted prices of four cuts (chuck, loin, round and flank) reported by ALIC based on Nikkei Point-of-Sales data. Retail prices and quantities for fish and poultry were obtained from the Retail Price Survey by the Statistical Bureau Ministry of Public Management, Home Affairs, Post and Telecommunications. Fish prices were the weighted average of tuna, horse mackerel, flounder, yellow tail and cuttlefish. The fish types selected are composed of high, medium and low-end fish types and reflect the most representative fish series for which data were available and complete.
The Empirical Model Empirical models applied to food safety events cover a wide range, from the commonly used Almost Ideal Demand System, the Rotterdam demand system, and related variations, to models of contingent valuation, experimental auctions and conjoint analysis. The methodological approach used in this study includes a VEC model, directed acyclic graphs, and historical decomposition to investigate the dynamics of price and quantity changes. The VEC model not only allows estimates of short-run relationships for the price and quantity series, but it also preserves the long-run relationships among the variables. Historical decompositions aid in providing a visual explanation of the shock’s impact on price and quantity series in the neighborhood of each incident. Orthogonal innovations are constructed using graph theory to determine causal patterns behind the correlation in contemporaneous innovations of the VEC model. The first step is stationarity testing of each series using the Augmented Dickey-Fuller (ADF) test. The test involves running a regression of the first difference of the series against the series lagged one period, lagged difference terms, and a constant. The stationarity test checks the series for unit roots. The non-stationary series are integrated of order one or I(1) with the first differences being stationary or I(0). Johansen’s co-integration test is performed to determine whether the series are co-integrated. Having a co-integrating equation captures the long-run relationship among the variables. In this process, a matrix is found to capture the longrun relationship among the variables. That matrix is decomposed into two matrices, α and β, where the matrix β contains the co-integrating vectors representing the underlying long-run relationship, and the α matrix describes the speed of adjustment at which each variable moves back to its long-run equilibrium (Johansen and Juselius, 1992; Schmidt, 2000). The VEC model’s covariance matrix helps when investigating the causal relationship among the variables with directed acyclic graphs (Bessler and Akleman, 1998; Saghaian, Hasan, and Reed, 2002). With this method, an algorithm determines the causal structure behind the correlation in the errors of the variables (Swanson and Granger, 1997). Finally, historical decompositions break down the price/quantity series into historical shocks in each series to determine their responses in a time interval neighborhood of the BSE events.
238
Sayed Saghaian, Leigh Maynard and Michael Reed
The Results Table 1 presents the OLS estimation results of the unit-root tests. The second column of the table shows the tests failed to reject the null hypothesis of zero first-order autocorrelation at the 5% level of significance except for fish price, poultry price, US quantity, and dairy quantity. The right-most column of Table 1 gives the results of the ADF test for the first difference transformation of the series. The null hypothesis is rejected for all variables after first differencing. Table 1. Augmented Dickey-Fuller (ADF)a Test Results. Variable US Beef Price AUS Beef Price Wagyu Beef Price Dairy Beef Price Fish Price Poultry Price US Beef Quantity AUS Beef Quantity Wagyu Beef Quantity Dairy Beef Quantity Fish Quantity Poultry Quantity
Test Results for Variables in Levels
Test Results for Variables after First-Differencing
2.64 2.39 2.13 1.64 3.02* 4.01** 3.40* 0.90 1.40 2.76* 2.30 1.98
9.47** 11.60** 11.11** 13.05** 7.45** 9.40** 10.09** 12.37** 11.48** 12.86** 4.44** 4.14**
Note: ** 1% significance level, * 5% significance level. a Test statistics are in absolute value and compared to MacKinnon (1996) one-sided p-value. Source: Saghaian and Reed (2007).
Table 2. Johansen Cointegration Test Results for Prices 5% Critical Null Hypothesisa Trace Statistics Value 185.46 95.75 r = 0*
r ≤ 1* r ≤ 2* r ≤3
a
109.64 56.34 25.44
69.82 47.86 29.80
Johansen Cointegration Test Results for Quantities 5% Critical Null Hypothesisa Trace Statistics Value 170.37 95.75 r = 0* 75.10 69.82 r ≤ 1* 43.86 47.86 r≤2
r is the cointegrating rank, MacKinnon-Haug-Michelis (1999) p-value. * 5% significance level. Source: Saghaian and Reed (2007).
Eigenvalue 0.53 0.41 0.26 0.15
Eigenvalue 0.61 0.27 0.19
The Importance of Context in Determining Consumer Response…
239
Table 2 presents the results of co-integration tests for the price and quantity series. The null hypothesis that r = 0 , r ≤ 1 , and r ≤ 2 was rejected at the 5% level for the price series, but the null hypothesis of r ≤ 3 could not be rejected at the 5% level. For quantity, the null hypothesis that r = 0 and r ≤ 1 was rejected at the 5% level, but the null of r ≤ 2 could not be rejected. Thus there are long-term relationships among the variables and the VEC model is appropriate in order to determine the directed graphs and causal patterns for prices and quantities. The residual correlation matrix of the VEC models provided the contemporaneous innovations (errors) that show how errors among the endogenous variables are related. The results show that the strongest correlation exists between the Japanese wagyu and dairy prices (0.68). This makes sense as pricing policies for Japan’s beef industry are consistently applied to wagyu and dairy beef. The results show the residuals associated with the two import origins are slightly correlated with dairy, but U.S. residuals are more strongly correlated to residuals from Japanese wagyu than Australian beef. Finally, there is little correlation in residuals for U.S. and Australian beef prices or among fish, chicken, and beef prices. However, there is much more correlation among the residuals from the quantity model. Correlations between US and wagyu, US and dairy, and wagyu and dairy are all high among the beef quantities. Fish and poultry consumption residuals are also highly correlated. Most correlation coefficients among the quantity errors are 0.40 or above, much higher than for prices. A formal test of contemporaneous causal structures is performed by orthogonalizing innovations to obtain the historical decomposition functions. The TETRAD IV software is applied to the correlation matrix to generate the causal patterns among the price and quantity series on innovations from the endogenous variables in each system (Spirtes et al., 1999). The historical decompositions include a 12-month horizon for each endogenous variable.
The BSE Impact in Japan Imported beef prices in Japan fell immediately in response to the BSE discovery, but domestic beef prices actually increased. However, ultimately all beef prices were adversely impacted by the BSE discovery. U.S. beef import prices fell the most dramatically immediately after the BSE discovery and saw the widest difference between the actual and forecasted prices. U.S. beef prices rebounded after the first two months, but they took another quick dive after December, reaching their lowest point in May, approximately seven months after the outbreak (Figure 1). Australian beef prices had a similar pattern to U.S. prices, but during the first month there was no dramatic drop. Japanese wagyu and dairy beef prices rose after the BSE outbreak; certainly not what one would expect, but by December, those prices began to fall absolutely relative to what they would have been without an outbreak. The pattern for these beef prices is explained by the way Japanese authorities handled the news of the BSE discovery. The immediate negative responses observed for U.S beef prices is attributed to the published remarks of a Japanese meat company who blamed imported beef as the most likely source of BSE in Japan, as explained by McCluskey, et al. (2004). Two weeks after publicly announcing the first confirmed case, a second case contradicted the government’s assurances
240
Sayed Saghaian, Leigh Maynard and Michael Reed
of healthy domestic animals, prompting anxiety among consumers and leading to further decline in all beef prices. USA Beef Price (BSE01)
0.928
JW Price (BSE01)
1.74
0.912
1.72
0.896 1.70 0.880
(Yen/Gram)
(Yen/gram)
1.68 0.864
0.848
1.66
1.64 0.832
1.62
0.816
0.800
1.60 Jul 2001
Aug
Sep
Oct
Nov
Dec
Jan
Feb
Mar
Apr
May
Jun 2002
Jul 2001
Aug
Sep
July 2001 - June 2002
Nov
Dec
Jan
Feb
Mar
Apr
May
Jun 2002
Mar
Apr
May
Jun 2002
July 2001 - June 2002
Aus Beef Price (BSE01)
0.70
Oct
JD Price (BSE01)
1.28
1.26
0.65 1.24
(Yen/Gram)
(Yen/Gram)
1.22 0.60
1.20
1.18 0.55
1.16
0.50
Jul 2001
Aug
Sep
Oct
Nov
Dec
Jan
Feb
Mar
Apr
May
Jun 2002
1.14
July 2001 - June 2002
Jul 2001
Aug
Sep
Oct
Nov
Dec
Jan
Feb
July 2001 - June 2002
Actual Price: ______________ Forecasted price before the event: ---------------------Vertical line is the Event of interest: | Source: Saghaian, Maynard, and Reed (2007)
Figure 1. Impact of BSE 2001 on Beef Prices in Japan.
The reliability of suppliers and perceived differences among suppliers may explain the impact of a food safety incident on consumers and their loss of confidence or trust (Bocker and Hanf, 2000). When it comes to food safety and reliability, consumers differentiate among product brands and origins, and trust in suppliers and retailers plays a major role in their purchasing decisions. Because consumers are unaware of unsafe food, a priori, and rely on supplier credibility and reputation; any news of a food safety scare involving a particular supplier impacts their perceptions and judgments regarding the reliability of that supplier. Becker et al (1996) showed that ‘trust/safety’ was one of the main factors influencing German consumer choice of a particular meat-product retailer. In an experimental study, Bocker
The Importance of Context in Determining Consumer Response…
241
(2002) tested the hypothesis that consumer reaction to a food safety scare could be explained by differences among perceived reliability of suppliers. BSE impact on cosumption of US 4.2 3.9
Log Per Capita
3.6 3.3 3.0 2.7 2.4 2.1 1.8
Apr
May
Jun Jul 2001
Aug
Sep
Oct
Nov
Dec
Jan
Feb
Mar
Apr
May
Apr
May
Jun 2002
BSE impact on cosumption of WAGYU 3.75 3.50
Log Per Capita
3.25 3.00 2.75 2.50 2.25 2.00
Apr
May
Jun Jul 2001
Aug
Sep
Oct
Nov
Dec
Jan
Feb
Mar
Jun 2002
BSE impact on cosumption of DAIRY 4.75 4.50
Log Per Capita
4.25 4.00 3.75 3.50 3.25
Apr
May
Jun Jul 2001
Aug
Sep
Oct
Nov
Dec
Jan
Feb
Mar
Apr
May
Jun 2002
Source: Saghaian and Reed (2007).
Figure 2. The BSE impact on per capita consumption of U.S., wagyu, and dairy beef in log-form.
242
Sayed Saghaian, Leigh Maynard and Michael Reed
The results indicate Japanese consumers reacted to the differentiation between suppliers in terms of their riskiness, increasing their confidence in the domestically produced beef; Japanese consumers’ beef purchase decision was impacted by the perception of reliability of beef suppliers. Customers did not abandon all beef, but differentiated beef by perceived riskiness of the source. Bocker and Hanf (2000) showed that with BSE in Germany, consumers switched to local butcheries that they trusted more for safer products. With the second announcement of a BSE discovery and the perceived discrepancy in the news and among suppliers, all beef prices were adversely impacted, indicating erosion of consumer trust and confidence in the entire beef industry. Our results show that the negative effects of the first BSE shock on beef prices dissipated after a few months, but the second wave of the scare had a stronger impact on beef prices. Mazzocchi (2005) found similar results regarding two instances of BSE crises in Italy. Our results show that while there was concern about all beef in Japan, Japanese domestic beef prices fell less than imported prices, which suggests that despite the BSE outbreaks, Japanese consumers still had more confidence in domestic beef production. Yet after twelve months, consumption of all types of beef was markedly lower than the predictions without a BSE incident. Consumption changes were drastic for all meats in the short-run. Purchases of US beef, wagyu beef, and dairy beef fell sharply in October and November, immediately after the BSE outbreak (Figure 2). In contrast, consumption of grass fed Australian beef, fish, and poultry increased sharply during the same period (Figure 3). The impact clearly indicates that consumers were mostly suspicious of US beef compared to other beef and had much greater trust in fish and poultry, where consumption drastically increased. This shows consumers switched to meats considered to be free of a BSE threat. These results indicate that Japanese consumer’s purchasing decisions were consistent with the information they were given. While in the short-run consumption of beef decreased, in the longer run, all meat consumption levels were close to the predictions without a BSE incident; but prices, and consequently profit margins, were lower. These results showed that Japanese consumers first reacted negatively to the beef safety scares and changed their buying and consumption habits accordingly, and over time, as the concerns dissipated and beef safety worries diminished, they reverted back to their previous consumption pattern. These results are consistent with discussions made by Mazzocchi (2005). These insights into the habits of consumers and the changing purchasing patterns for meat consumers faced with food safety concerns have strategic implications for supply chain managers and practitioners. It is very important for food firms to be active in providing information to consumers because such information is used in purchasing decisions. Yet the information conveyed must be reliable if trust is to be retained between producers and consumers. Baker (1998) shows consumers are demanding more food safety information; he lists several private, as well as government policy, options (such as labeling, increased production standards and regulatory monitoring) that can help address those concerns.
The Importance of Context in Determining Consumer Response…
243
BSE impact on cosumption of AUS 4.41 4.34
Log Per Capita
4.27 4.20 4.13 4.06 3.99 3.92 3.85
Apr
May
Jun Jul 2001
Aug
Sep
Oct
Nov
Dec
Jan
Feb
Mar
Apr
May
Jun 2002
Apr
May
Jun 2002
May
Jun 2002
BSE impact on cosumption of FISH 6.00 5.95
Log Per Capita
5.90 5.85 5.80 5.75 5.70 5.65
Apr
May
Jun Jul 2001
Aug
Sep
Oct
Nov
Dec
Jan
Feb
Mar
BSE impact on cosumption of POULTRY 6.1 6.0
Log Per Capita
5.9 5.8 5.7 5.6 5.5 5.4
Apr
May
Jun Jul 2001
Aug
Sep
Oct
Nov
Dec
Jan
Feb
Mar
Apr
Source: Saghaian and Reed (2007).
Figure 3. The BSE impact on per capita consumption of fish, poultry, and Australian beef in log-form.
244
Sayed Saghaian, Leigh Maynard and Michael Reed
Remarks The effects of food safety scares are part of a dynamic process, where consumers change consumption during the scare and often return to their past behavior afterward. This study finds that consumers over-react initially to the shock with decreased consumption of the suspect food item, but concern gradually dissipates, leading to the establishment of a new equilibrium. Japanese consumers understood the differences in the two beef safety scares and reacted differently to them. Prices of all beef products were lower twelve months after the BSE discovery, a clear indication that the news of the BSE discovery adversely affected consumers’ perceptions of beef quality and lowered profit margins. Yet, the price decrease for the two imported beef types were more than the price decrease for the two domestically produced beef categories. This indicates that Japanese consumers have a more positive view of their own beef products and this keeps the price of their domestic beef products from falling as much as imported products. Japanese consumers moved away from beef that uses high levels of concentrated feeds (US beef, wagyu beef, and dairy beef) and flocked toward grass fed Australian beef and fish and poultry. These results provide incentives for beef producers and retailers to proactively inform consumers about ongoing beef safety measures, and can potentially provide policy makers a basis for countermeasures and compensations. Beef safety crises have increased the need for robust information technologies in the food marketing system. Time and experience provide a metric for consumers to recalibrate their risk perception and require beef producers and marketers to pay greater attention to beef safety issues and employ quality assurance measures and traceability schemes to address consumer concerns. The BSE situation has certainly created opportunities for producers that have traceable production systems and have quality assurance programs that involve branding and labeling. Proactive information provision in the food marketing systems reduces the impacts of the food scare. Safe food seems to be largely a public good, so industries have an interest to develop protocols together to provide greater safety assurances. A BSE case or salmonella outbreak impacts everyone; one incidence of ‘bad strawberries’ hurts the whole strawberry industry and even related fruits. The U.S. government and the food industry must continue to invest heavily into procedures that will reduce food safety scares in these areas and into information systems that minimize the impacts of food safety shocks.
Part II: BSE Discovery in the U.S. While discovering BSE in other countries such as the United Kingdom and Japan resulted in considerable economic damage to beef producers as well as food service industries, the notable impact of the BSE discovery in the U.S was mostly on the export sector; Japan, a major export market for U.S. beef and beef products stopped all imports. Quickly after the U.S. BSE discovery in December 2003, USDA announced additional beef safety procedures that banned specified risk material (SRM), such as brain, spinal tissue, etc., for cattle over 30 months of age from food supply. This part focuses on the short-run dynamics of price adjustment and transmission to see how the beef safety scare affected the price margins along the U.S. beef supply channel. Research on vertical and spatial price transmission is vast, expanding in different directions
The Importance of Context in Determining Consumer Response…
245
from asymmetric price transmission, price stickiness and incomplete pass through, to market efficiency and integration, to concentration and market power. Given recent structural changes in the U.S. beef industry and its high market concentration, an important research question is the extent that the BSE shock was transmitted through the supply chain and impacted prices at the feedlot, packer, and retail levels. Factors such as product heterogeneity, long-term contracts, and market concentration can potentially influence the degree and dynamics of price transmission, leading to differential price effects along the marketing chain. Differing transmissions for the shock have welfare implications with respect to the efficiency and equity of the marketing system. There are many research articles on market integration and price transmission along the agricultural marketing channels. Early models were primarily static linear equilibrium models, assuming perfectly competitive markets. An important example of this literature is the work of Gardner (1975). Market integration and price transmission theory has evolved extensively since Gardner’s seminal work, expanding the literature in different directions. Heien (1980) added dynamic analysis to address short-run disequilibrium price adjustments. Tiffin and Dawson (2000) found that UK lamb prices were determined in the retail market, and then passed upward along the supply chain. Goodwin and Holt (1999) and Goodwin and Harper (2000) found that retail market shocks were confined to retail markets, but farm markets adjusted to shocks in wholesale markets. Yet, Ben-Kaabia, et al. (2002) found both supply and demand shocks were fully passed through the marketing channel; i.e., they found complete price transmission. The literature on price stickiness and incomplete price transmission provides a detailed explanation for asymmetric price behavior. Price transmission can be asymmetric when there exits different speeds of price adjustment across vertically linked markets. Price asymmetry can exist with respect to magnitude, speed, or a combination of the two. It is important to note that there are different definitions of price asymmetry. In this research the focus is on the different speeds of price adjustment along the beef marketing channel in the feedlot, wholesale and retail markets affecting price margins. The traditional definition of price asymmetry refers to a situation where producer price increases are passed forward quicker and more completely to consumers than price reductions (Pelzman, 2000; Bakucs and Ferto, 2005). In an efficient market prices are transmitted fully and completely. The fact that price dynamics may differ under competitive and noncompetitive market conditions can lead to market inefficiency. McCorrison et al. (1998) demonstrated the role of oligopoly power in determining the price transmission elasticity following a supply shock. Other studies have supported the hypothesis that market concentration and imperfect competition can be the cause of asymmetric price transmission (Miller and Hayenga, 2001; Lloyd, et al., 2003). Luoma, et al. (2004) has argued that market power is the most likely explanation for asymmetric price transmission in the long run. Retailers may keep price levels relatively fixed for long periods when markets are imperfectly competitive, or oligopolies may react quicker to declining margins by utilizing their market power. They do this to maintain market shares, keeping long-run rather than short-run profits in mind. Hence, market power can reduce price transmission along the marketing chain.
246
Sayed Saghaian, Leigh Maynard and Michael Reed
The Data, the Model, and the Empirical Results Weekly data for feedlot, wholesale, and retail prices were assembled from the Livestock Marketing Information Center (LMIC) for 1/5/1991 to 7/2/2005. The vertical structure of the data set begins with feeder cattle followed by live cattle, wholesale, and retail levels. All prices are in dollar per hundred weights ($/cwt). The feedlot price used in this analysis is the Kansas live cattle price. A 1000 lb steer is assumed, and multiplied by this Kansas price, to derive a live steer value. The total wholesale value is the sum of the boxed beef value and the byproduct value -- both are USDA prices from the LMIC. From this assumed 1000 lb steer, a dressing percentage of 63% and a retail yield of 42.7% from the live animal weight are used to calculate the retail value, which is the monthly USDA reported retail price multiplied by the estimated retail yield. The beef prices are the average for all grades. The assumption is that the BSE discovery reported by the news outlets affects quality perception of all beef, consistent with other research in this area (e.g., Piggott and Marsh, 2004). While the beginning date of the BSE scare is well known, there is no way to know exactly how long the beef safety scare will impact on consumer perceptions of safety. In this research we concentrate on the short-run dynamics of price adjustment and price transmission at different market levels in a neighborhood around the BSE shock specified by the historical decomposition graphs, though price transmission patterns could be different before and after the BSE scare. In this part, we closely followed the contemporary non-stationary time-series modeling paradigm of the previous section. First, the temporal properties of the three price series were analyzed using ADF tests. We failed to reject the null hypothesis of a unit root for these variables with two terms, a constant and a trend. Each series was then first differenced and the ADF regressions were re-estimated with a constant but no trend. In each case, we rejected the null hypothesis of a unit root at the 1% level of significance. Second, Johansen’s cointegration tests were employed to determine if a long-run relationship existed among the three variables in the system. These results suggested there are two long-run equilibrium relationships between the three price series (Table 3). Table 3. Johansen Cointegration Test Results Null Hypothesisa
a
r = 0* r ≤ 1* r≤2
Trace Statistics 80.56 17.63 2.03E-05
5% Critical Value 29.80 15.49 3.81
Eigenvalue 0.08 0.02 0.051
r is the cointegrating rank. *Denotes rejection of the hypothesis at the 5% level. Source: Saghaian (2007).
Next, we estimated a VEC model and conducted hypothesis testing within this framework. The VEC model analysis of dynamic adjustments provided a precise measure of price transmission speeds. The empirical estimates of the adjustment speeds are summarized in the top portion of Table 4. The adjustment speeds for wholesale and retail prices were statistically significant at the 1% level. The adjustment speed for feedlot prices was not
The Importance of Context in Determining Consumer Response…
247
statistically significant. The dynamic adjustment speed for wholesale prices was much higher, 0.13, (in absolute value) than retail prices, 0.02. This is an interesting result suggesting that with the beef safety shock, wholesale prices must adjust much more and do it faster than retail prices to restore the long run equilibrium. These results indicate that the speeds of price adjustment vary by market and prices in the wholesale market adjust more than six times faster to the BSE shock than prices in the retail market. Table 4. The Empirical Estimates of Speeds of Adjustment and Diagnostics
ΔPft
ΔPwt
ΔPrt
-0.01
-0.13*
-0.02*
R2
0.14
0.39
0.13
AIC Schwarz Criterion
-4.79 -4.70
-5.38 -5.29
-7.17 -7.08
Variable Speeds of Adjustment Model Diagnostics
* 1% significance level. Source: Saghaian (2007).
Some explanations given in the literature for the causes of price asymmetry are product heterogeneity, long term contracts, and adjustment or menu costs (e.g., Goodwin and Holt, 1999; Zachariasse and Bunte, 2003), which may explain the differential speeds of price adjustment along the U.S. beef marketing channel. Originally, Hicks (1974) and Okun’s (1975) works showed that prices in some sectors of the economy were sticky while prices in other sectors were flexible. According to their arguments, prices of most goods and services are not free to respond to changes in demand in the short run. Bordo (1980) showed some prices respond slowly to policy shocks due to long-term contracts. We employed Granger causality tests and directed acyclic graphs to investigate the causal patterns among the variables. The covariance matrix of the VEC model was used to investigate the causal relationship among the variables by directed acyclic graphs. The results show that innovations in feedlot and wholesale, and in wholesale and retail price variables affect residuals in each other, but there are no arrows to indicate direct causality. Also, there exists no residual relationship between feedlot and retail beef prices. The relationship between feedlot and retail beef price residuals is through wholesale prices. Since directed graph results from the residuals did not provide a clear causality direction, we used pairwise Granger causality tests (with four lags) to investigate causal directions. The results are summarized in Table 5. F-test results indicate that the hypothesis, retail prices do not Granger-cause feedlot prices, is the only one that cannot be rejected. The results show the direction of causality runs especially strong from feedlot to wholesale to retail prices. However, this relationship is not unique (i.e., uni-directional); there are causality relationships going upstream, from retail to wholesale to feedlot prices, as well. The results rejected the hypothesis that price transmission in the U.S. beef sector flows only downward along the supply chain with the direction of causality running from producer to retail prices. These results suggest that prices in the U.S. beef sector are not determined at one end and then passed down or up along the supply channel; pricing patterns in the U.S. beef sector are not just cost or demand driven. Prices are determined
248
Sayed Saghaian, Leigh Maynard and Michael Reed
simultaneously through bidirectional interaction between the different stages (likely through contracts). Table 5. The Results of Pairwise Granger Causality Tests Null Hypothesis Feedlot Price does not Granger Cause Wholesale Price Wholesale Price does not Granger Cause Feedlot Price Wholesale Price does not Granger Cause Retail Price Retail Price does not Granger Cause Wholesale Price Feedlot Price does not Granger Cause Retail Price Retail Price does not Granger Cause Feedlot Price
F-Statistic 41.94** 7.01** 17.07** 4.88* 12.49** 1.73
** 1% significance level, * 5% significance level. Source: Saghaian (2007)
Finally, historical decomposition of feedlot, wholesale, and retail-level prices aided in explaining the behavior of beef prices due to the BSE shock. Figure 4 shows the historical decomposition graphs of the three price series for a four month horizon from the RATS software. The BSE was discovered on December, 23, 2003. Before this date, the actual feedlot, wholesale and retail prices (solid lines) and their forecast prices (dashed lines) followed each other closely with minor differences that are commonly expected between actual price and its forecast. However, they began to depart significantly by the end of December 2003. Historical decomposition of the feedlot prices showed wide departure of actual feedlot prices during the last week of December, reaching its maximum by January 10, 2003. The feedlot prices dropped by 21% during this period. Meanwhile the wholesale prices decreased by 16%. In contrast, the largest negative impact of the BSE shock on retail prices was only about 6%. These results, consistent with the results for adjustment speeds, showed that the beefsafety scare impacts on producers and retailers were quite different. The impact of the BSE shock on feedlot prices (21%), within almost identical time-periods, was more than three times that of retail prices (6%). Also, the effect of the beef safety scare on wholesale prices (16%) was more than twice the effect on retail prices, a clear indication of asymmetric price effects. Overall, the historical decompositions showed, as expected, that the BSE discovery impacted beef prices negatively, but the magnitudes of price effects were substantially different for the three price series, resulting in widened producer-retail price margins. Also, the effect of the shock on the three price series immediately after the event had a one week lag between the stages. Since the BSE discovery was covered by the media and electronic news outlets rather quickly, the estimated one week lag of the beef safety impact along the supply channel may reflect the role of contracts and the fact that, typically, cattle are bought one week before they are slaughtered, rather than reflecting problems with the flow of information through the chain.
The Importance of Context in Determining Consumer Response…
249
Historical Decomposition of FEEDLOT 6.95 6.90
Log of Price
6.85 6.80 6.75 6.70 6.65 6.60
8
15 22 Nov ember
29
6
13 20 Dec ember
27
3
10
17 24 J anuary
31
7
14 21 February
28
7
14 21 February
28
Historical Decomposition of W HOLESALE 7.15 7.10
Log of Price
7.05 7.00 6.95 6.90 6.85 6.80 6.75
8
15 22 Nov ember
29
6
13 20 Dec ember
27
3
10
17 24 J anuary
31
Historical Decomposition of RETAIL 7.62 7.60
Log of Price
7.58 7.56 7.54 7.52 7.50 7.48
8
15 22 Nov ember
29
Actual Price including the BSE shock: Forecasted price before the event: Source: Saghaian (2007)
6
13 20 Dec ember
27
3
10
17 24 J anuary
31
7
14 21 February
28
______________ ---------------------
Figure 4. The BSE impact on U.S. feedlot, wholesale and retail prices in log-form.
250
Sayed Saghaian, Leigh Maynard and Michael Reed
Concluding Remarks In this section we investigated how the U.S. BSE discovery affected feedlot, wholesale, and retail beef prices along the U.S. beef supply channel. First, our results indicated that beef price causality at different stages of the supply channel were bi-directional, influencing and being influenced by each other. Second, the results of the cointegrated VEC model showed that wholesale prices were actually more flexible than retail prices, and the short-run speed of adjustment at the wholesale level was much faster than at the retail level. Third, the historical decomposition results corroborated the results of the dynamic speeds of adjustment showing the BSE shock was distributed unevenly, with the feedlot and wholesale levels taking most of the impact of the negative shock, falling by more than three times and double, respectively, compared to the fall in retail prices. The differential effects of the BSE discovery on the supply channel widened the gross margins between farm and wholesale and wholesale and retail, changing income distribution along the U.S. beef marketing channel.
Part III: BSE Discoveries in Canada BSE was first identified in a Canadian-born cow on May 20, 2003, abruptly ending most beef exports to Canada’s main beef trading partners. On December 23, 2003, United States authorities discovered BSE in a Canadian-born cow in Washington State, and two additional BSE diagnoses occurred in Canada on December 30, 2004 and January 11, 2005. All four discoveries involved animals born in Alberta, Canada’s leading beef-producing province, and Statistics Canada (2006a) estimated that BSE cost Canadian beef producers about $5.3 billion. The present analysis focuses on consumer-level impacts of BSE. Unlike farm-level impacts, less consensus exists on the severity of domestic reductions in demand, but concern remains high among industry members and government agriculture agencies. Unlike the European experience, where 163 vCJD deaths occurred in the United Kingdom alone (NCJDSU, 2008), no deaths have been linked to the Canadian-born BSE events. Maynard, Goddard, and Conley (2008) reviewed prior studies demonstrating substantial retail-level BSE impacts in Europe and Japan (Burton and Young, 1996; Peterson and Chen, 2005), but modest BSE impacts on beef demand in North America (Piggott and Marsh, 2004; Vickner, Bailey, and Dustin, 2006; Peng, McCann-Hiltz, and Goddard, 2005). In addition to the limited human health impact to date, the Canadian government’s response was viewed by many as proactive and transparent, and much media coverage after the May, 2003 BSE discovery focused on the ranchers’ plight (Boyd and Jardine, 2007). Concurrently, however, Canadian consumers expressed serious concern about meat safety, with BSE leading the list of meat-related threats (de Jonge et al., 2006). Maynard, Goddard, and Conley (2008) evaluated the impact of BSE newspaper coverage on fast food beef purchases in Alberta and Ontario. Consumers in Ontario were significantly more likely to stop purchasing fast food beef entrées in months following a surge of BSE media coverage, but those who purchased beef entrées did not reduce their consumption level, on average. Alberta fast food consumers did not appear to respond significantly to BSE media coverage.
The Importance of Context in Determining Consumer Response…
251
The analysis presented here builds on Maynard, Goddard, and Conley (2008) by addressing BSE impacts on beef purchased for at-home consumption in Alberta and Ontario. The results imply conflicting event-specific impacts, suggesting that an undifferentiated media index might understate consumer BSE responses by averaging out conflicting reactions. Accordingly, the analysis of fast food beef purchases is updated here to distinguish among BSE events. The fast food analysis is also expanded to the national level. Taken together, the results present a broad view of Canadian consumer response to BSE in both food-at-home and food-away-from-home markets.
Canadian BSE Impacts on Food-at-Home Beef Purchases Regression models were developed to test whether Alberta and Ontario consumers reacted to BSE either by boycotting or reducing beef purchases. Alberta was selected because Canada’s BSE cases during the study period were all discovered in Alberta-born animals. Alberta is Canada’s leading beef-producing province, and consumers might be expected to support the interests of producers to greater extent than in other provinces. Ontario was selected because it is Canada’s most populous province, and is geographically distant from the source of the BSE discoveries. The data used in the food-at-home analysis were AC Nielsen Homescan data, purchased by the Consumer and Market Demand Agricultural Policy Research Network, hosted at the University of Alberta’s Department of Rural Economy. The data represented household-level meat purchases during calendar years 2002 – 2005. In each year, 9,000 – 10,000 Alberta and Ontario households participated in the panel, often for multiple years. Table 6. Selected Variable Means from Food-at-Home Scanner Data, 2002-2005 # beef purchases / month # pork purchases / month # poultry purchases / month Beef expenditure / month Beef expenditure share Household size Age: 18-34 Age: 35-44 Age: 45-54 Age: 55-64 Age: 65+ Education: < High school Education: High school Education: Some college Education: College Education: Some university Education: University
Alberta 2.13 1.18 1.11 $39.95 36% 2.6 9% 25% 28% 20% 18% 12% 17% 15% 27% 8% 22%
Ontario 2.11 1.25 1.59 $30.27 31% 2.6 7% 25% 25% 21% 23% 13% 16% 15% 23% 8% 24%
252
Sayed Saghaian, Leigh Maynard and Michael Reed
Each observation provided data on a household’s individual meat purchase, including a household ID number, province, primary language, household size, age and presence of children, age of the household head, income, household head education level, purchase date, which of 45 meat types was purchased, quantity purchased, price paid, and codes allowing distinctions among supermarkets, mass merchandise stores, warehouse stores, and other store types. Selected variable means appear in Table 6, illustrating considerable similarity between Alberta and Ontario, with the exception that average beef consumption was relatively higher among Alberta consumers. While the data were rich in number of observations, shortcomings include general product designations that prevented distinctions among beef cuts, and a lack of weight data allowing standardization of quantity units, which in turn prevented calculation of meaningful unit prices. To compensate for these ambiguities, the analysis was performed on multiple measures of beef purchases. Analyses of beef units purchased and beef expenditure share are reported here, but we obtained qualitatively similar results from analyses of beef purchase frequency, beef expenditure, beef unit share, and beef frequency share. The 45 meat type codes were first aggregated into the broader categories of beef, pork, poultry, frozen poultry products, and frozen seafood products. The few remaining meats were game products with exceedingly low purchase frequencies. To provide a temporal basis for comparison across households, purchases were aggregated by household ID and by month, producing over 45,000 Alberta observations, and almost 96,000 Ontario observations. Households may have reacted to BSE discoveries either by ceasing beef purchases entirely, or by altering their level of beef consumption. In many applications, the data generating process for zero observations differs from that of positive observations, typified by distributions with relatively greater probability mass at zero. For example, consumers who never buy beef would produce zero observations, but so might beef consumers who happened to choose a zero quantity during a given period (Burton, Dorsett, and Young, 1996). Doublehurdle models are often used to test for systematic differences between determinants of “participation” (whether or not to buy beef) and “consumption” (how much beef to buy). The number of beef units purchased each month are count data left censored at zero, while beef expenditure share is a continuous variable bounded by the unit interval. Cragg (1971) proposed modeling the participation decision as a binary choice model, and the consumption decision as a truncated tobit model. For the current application, a logit model was used to describe the participation decision, a truncated Poisson model was used for quantity (count data) consumption decisions, and a truncated tobit specification was used for expenditure share (continuous data) consumption decisions. Mullahy (1986), Yen (1999), and Maynard et al. (2004) provide examples of count data double-hurdle models. The general likelihood function for the double-hurdle model is:
L = ∏ Pr(qi = 0) ∏ [Pr(qi > 0) Pr( qi | qi > 0)] , qi =0
qi > 0
where qi denotes quantity of beef entrees purchased by the ith household. The specific likelihood function for the count data double-hurdle model is:
The Importance of Context in Determining Consumer Response…
⎞ ⎛ 1 ⎟⎟ L = ∏ ⎜⎜ qi =0 ⎝ 1 + exp( xiα ) ⎠
⎛ exp( xiα ) ⎞
253
⎛ exp(qi xi β ) ⎞ ⎟⎟ , qi ! ⎠ ⎝
∏ ⎜⎜ 1 + exp( x α ) ⎟⎟ exp[1 − exp( x β )] ⎜⎜ qi >0
⎝
i
i
⎠
where α and β are conformable parameter vectors describing participation and consumption behavior, respectively. In the case of the continuous double-hurdle model used to explain expenditure share wi, the likelihood function is:
⎞ ⎛ 1 ⎟⎟ L = ∏ ⎜⎜ qi = 0 ⎝ 1 + exp( x i α ) ⎠
(
⎛ exp( xiα ) ⎞⎛ f i wi − xi β , σ 2 ⎟⎟⎜⎜ ⎜⎜ ∏ Fi qi > 0 ⎝ 1 + exp( x iα ) ⎠⎝
) ⎞⎟
⎟ , ⎠
where fi and Fi are respectively the pdf and cdf of the standard normal distribution evaluated at xiβ / σ2 (Maddala, 1983, p. 152). Explanatory power of participation models was evaluated using the likelihood ratio index (LRI) measured as one minus the ratio of the unconstrained and intercept-only log-likelihood function values. Explanatory power in the consumption models was evaluated by the R2p statistic for the truncated Poisson model (Greene, 2000, p. 882), while the standard R2 is used for the truncated tobit model.
Figure 5. Retail food-at-home per-unit beef expenditures did not fall dramatically after BSE discoveries in May 2003, December 2003, and January 2005.
While the lack of unit weight data precludes calculation of standardized unit prices, it is possible to measure expenditure per unstandardized unit for each meat category and each month. If one accepts the assumption that average weight per unit within broad meat categories was likely to be stable across time, variation in average per-unit expenditures
254
Sayed Saghaian, Leigh Maynard and Michael Reed
should correlate highly with average price per kg. We used average expenditures per unit, shown in Figure 5, to test anecdotal reports of deep discounting of retail beef products following the first BSE event in May, 2003. Specifically, we regressed weekly national quantity-weighted average beef expenditures on linear and quadratic time trends, a cosinebased seasonality variable, and a BSE dummy variable representing various durations beginning with the week of May 23, 2003. The results showed that retail beef prices were not systematically discounted after the first BSE event. Anecdotally, observers have suggested that retail meat managers were reluctant to distort relative meat prices. Therefore, it seems unlikely that consumer response to BSE was confounded by simultaneously low beef prices. Table 7. Determinants of monthly household beef quantity purchases for at-home consumption a Regressor Intercept
Alberta
Ontario
-0.4319 ** -0.7495
Regressor (cont.) Alberta **
Total meat quantity (t-1) 0.1576
** 0.1361
**
January
0.2696
Beef share (t-1)
1.4791
** 1.7739
**
February
0.0676
Pork share (t-1)
0.3603
** 0.6899
**
March
0.0990
Poultry share (t-1)
0.1399
*
**
April
0.0058
Frz. poultry share (t-1)
-0.3449 ** -0.6075
**
May
0.0855
0.2181
Ontario ** 0.3896 * *
**
0.2268
**
0.1949
**
0.0689
*
0.3469
** **
Frz. seafood share (t-1)
-0.4275 ** -0.7673
**
June
-0.0194
0.1205
Household size
0.2189
**
July
-0.0009
0.0207
Child under 6
-0.3247 ** -0.2602
**
September
-0.2498 ** 0.3171
Child age 6-12
-0.0531
-0.2773
**
October
-0.0211 ** 0.0353
Child age 13-17
-0.0397
-0.0213
November
-0.6500
Age 35-44
-0.1207 ** -0.0526
December
-0.6500 ** -0.3050
**
Age 45-54
0.0237
0.0036
BSE event 1, t+0
0.1542
** 0.1243
**
Age 55-64
0.1037
** 0.0875
**
BSE event 1, t+1
0.2698
** 0.0573
Age 65+
0.0968
** -0.1226
**
BSE event 1, t+2
0.6984
** 0.3927
**
Income < $20K
-0.1141 ** -0.2315
**
BSE event 1, t+3
0.4976
** 0.3849
**
Income $20-$30K
0.0113
-0.1488
**
BSE event 1, t+4
0.0316
-0.0178
Income $30-$40K
-0.0331
-0.1084
**
BSE event 2, t+0
0.2046
** -0.0213
Income $40-$50K
-0.2157 ** -0.0165
BSE event 2, t+1
-0.0235
Income $50-$70K
-0.0422 *
-0.0281
*
BSE event 2, t+2
-0.3609 ** -0.3634
**
< High school
0.0638
*
0.3291
**
BSE event 2, t+3
-0.0284
0.3120
**
High school
0.1335
** 0.2167
**
BSE event 2, t+4
-0.3381 ** 0.1792
**
** 0.1957
*
0.2632
0.2042
** **
**
Some college
0.2500
** 0.1769
**
BSE event 3, t+0
-0.4887 ** -0.0249
College
0.0548
** 0.0763
**
BSE event 3, t+1
-0.1261 *
-0.2101
**
Some university
0.0638
*
0.1082
**
BSE event 3, t+2
-0.3439 ** -0.1194
**
Mass merchandise store -0.1574
0.1152
*
BSE event 3, t+3
0.1992
**
Warehouse store
0.2196
**
BSE event 3, t+4
-0.4614 ** -0.2080
a
0.0736
** 0.3997
values represent Poisson marginal effects, N = 45,146 (Alberta) and 95,906 (Ontario) * and ** denote statistical significance of the underlying parameter at the .05 and .01 levels, respectively
**
The Importance of Context in Determining Consumer Response…
255
Three regressions were estimated on beef purchased for at-home consumption, the dependent variables being beef participation (logit model), beef unit quantity consumption (truncated Poisson model), and beef expenditure share (truncated tobit model). In each regression, the independent variables consisted of a lagged endogenous variable to control for unobserved household heterogeneity, lagged expenditure shares of the five meat categories (beef, pork, poultry, frozen poultry, and frozen seafood), household size, dummy variables indicating the presence of children in three age groups (under 6, 6-12, 13-17), four age group dummy variables with the under-35 age group excluded as the base, five income categories with the $70,000+ category excluded as the base, five education categories with university graduates excluded as the base, dummy variables for purchases at mass merchandise stores and warehouse stores, monthly dummy variables excluding August as the base, and 15 BSE event dummy variables. The BSE discoveries occurring during the study period were treated as three distinct events: one in late May 2003, one in late December 2003, and a pair of BSE diagnoses for which monthly impacts could first be observed in January 2005. For each event, dummy variables were created that separately designated the month of occurrence and four subsequent months.Full regression results are shown in Table 7 for the truncated Poisson model of beef unit quantity consumption. The control variables unrelated to BSE contain useful information about beef purchase patterns, such as the importance of household habits, but the discussion here will focus on the BSE-related parameters. Qualitatively similar results were observed in the beef participation and beef expenditure share consumption models, and the main results concerning BSE impacts are illustrated in Figure 6. One of the main results, and an unexpected one, is that consumers reacted to the first BSE event by significantly increasing beef purchases. Increases were observed in all measures: likelihood of buying beef, quantity of beef units purchased, and beef expenditure share relative to other meats. Comparing the magnitude of quantity marginal effects in Table 7 (e.g., 0.15 to 0.70 units in Alberta) to average monthly purchase quantities (e.g., 2.13 units in Alberta), the results suggest economically significant increases in beef purchases attributable to the first BSE event. Positive impacts were stronger and more immediate in Alberta than in Ontario, perhaps reflecting Albertans’ proximity to the struggling ranching sector. Combined with Boyd and Jardine’s (2007) findings that media coverage following the first BSE event emphasized trade impacts over food safety concerns, our results suggest that many Canadians initially rallied to support ranchers. Consumer perception appeared to change following the second and third BSE events, with negative impacts dominating. As with the first event, the impacts were statistically and economically significant and substantially stronger in Alberta than in Ontario. The results suggest that food safety fears began to mount as multiple cases of BSE were discovered. A detailed analysis of media coverage following the second and third events has not yet been performed. Given the active involvement of government agencies’ in supplying public information about BSE, sometimes in concert with beef industry organizations, it would be policy-relevant to test if the nature of media coverage correlated highly with consumer responses. Overall, the results demonstrate a need to evaluate BSE events individually, rather than measuring an average or net consumer response to BSE.
256
Sayed Saghaian, Leigh Maynard and Michael Reed
Figure 6. BSE impacts varied across province and BSE occurrence.
The Importance of Context in Determining Consumer Response…
257
Canadian BSE Impacts on Food-away-from-Home Beef Purchases The possibility exists that BSE concerns were expressed more strongly in food awayfrom-home purchases, where consumers delegate food sourcing and preparation to restaurants. Consumers receive little information about the safety of beef purchased in restaurants. Fast food beef purchases are of special interest due to the prominence of ground beef products, which are the most likely vehicle for BSE-infected nerve tissue. Building on the context-dependent results of the food-at-home analysis described above, this section reports an extension of an analysis by Maynard, Goddard, and Conley (2008). The original analysis evaluated average responses to BSE media coverage in Alberta and Ontario, while the present analysis uses nationwide data and evaluates responses specific to each BSE discovery. The analysis was performed on the NPD Group, Inc. Consumer Report on Eating Share Trends (CREST) dataset describing Canadian food away-from-home purchases for the period from May, 2000 to May, 2005. The dataset was purchased by the Consumer and Market Demand Agricultural Policy Research Network, and contains individual meal records from approximately 4,000 households per quarter. We restricted our attention to non-breakfast, fast food purchases, and aggregated each household’s purchases by month, producing 44,246 observations. Table 8. Descriptive Statistics of Canadian Fast Food Purchases, May 2000 – May 2005 Beef entree (0/1) Chicken entree (0/1) Pizza entree (0/1) Other entree (0/1) Number of beef entrees Number of chicken entrees Number of pizza entrees Number of other entrees Promotional deal (0/1) Price of beef entrée ($) Price of chicken entrée ($) Price of pizza entrée ($) Price of other entrée ($) Age of household head Married (0/1) Children (0/1) College degree (0/1) Female head works full time (0/1) English second language (0/1) Restaurant within 5 minutes (0/1) Came from work (0/1) Came from home (0/1)
Mean 0.37 0.34 0.28 0.44 1.10 0.75 0.74 0.98 0.37 2.83 3.42 5.49 4.20 48.23 0.69 0.36 0.28 0.36 0.26 0.47 0.20 0.33
Std. Dev. 0.48 0.47 0.45 0.50 1.79 1.47 1.54 1.62 0.43 0.42 0.48 0.52 0.31 14.31 0.46 0.48 0.45 0.48 0.44 0.44 0.35 0.41
N = 44,246, where each observation represents a household's purchases in a given month
258
Sayed Saghaian, Leigh Maynard and Michael Reed
Variables include a household ID number, codes for up to eight items purchased by each individual, the price paid by the entire party, a wide range of socio-economic and demographic variables describing households and individual diners, and detailed data on the time, location, and context of each meal. Table 8 shows descriptive statistics of variables used in the analysis. Specific food codes were categorized into 23 subcategories, of which 14 referred to entrées. Following estimation of entrée unit prices (described below), the entrée subcategories were further aggregated into beef, chicken, pizza, and “other” designations for use in the primary analysis. Beef entrees such as hamburgers were the most commonly reported and most numerous entrée, followed closely by chicken. As in the preceding analysis of beef purchased for at-home consumption, a double-hurdle model was used to distinguish between participation and consumption decisions. The detailed data set offered dozens of potential regressors. Consumer demand theory suggested the inclusion of several variables in the double-hurdle model, and other variables were included when they correlated highly with beef entrée purchases and did not induce severe multicollinearity. Variables relating to the household head refer to the female household head when both genders were present. Following Maynard, Goddard, and Conley (2008), the dependent variable in the participation model was a dummy variable for beef entrée purchases, and in the consumption model the dependent variable was the number of beef entrées purchased by a given household in a given month. Regarding independent variables, a dummy variable called “promotional deal” indicated meals purchased under a special price, coupon, or other offer. Demographic variables included the age of the household head, whether the household heads were married, whether children under age 18 lived in the household, whether the household head had a university degree, whether the female household head worked full time outside the home, and whether English was a second language for the household head. Travel time to the restaurant and the diner’s location prior to the meal were included as regressors. As in the Canadian food-at-home analysis, household habits were controlled using the most recent lagged dependent variables. Individual item prices are not recorded in the dataset, requiring a method for imputing expected average prices for the four entrée categories. The hedonic price estimation method used here follows Maynard, Goddard, and Conley (2008) and is similar to the approach used by Richards and Padilla (2007). Specific food items were assigned to 23 categories, with 5 categories devoted to beef entrées, 4 categories representing chicken entrées, 2 pizza entrée categories, 3 other entrée categories, and 9 categories representing side dishes, desserts, beverages, and other items. The dependent variable of the hedonic regression was monthly household meal expenditures. Independent variables were the household-specific monthly quantity of each food category purchased, and interaction variables of each category quantity with linear, quadratic, and cubic time trends, and with each of the three BSE events, denoted by the month of occurrence and three subsequent months. Given that each interaction variable added 23 regressors to the model, the final hedonic regression contained 161 regressors. The resulting adjusted R-squared value was 0.88. Average expected prices for each entrée type (beef, chicken, pizza, other) were calculated as quantity share-weighted sums of the relevant monthly subcategory prices. The resulting entrée price series appear in Figure 7. Each of the BSE events produced discounted beef prices. Following the first two BSE events, non-beef entrée prices rose, while all entrée prices fell after the third BSE event in January, 2005. The magnitude of BSE-induced price changes was likely tempered by modest farm value as a share of the retail fast food dollar.
The Importance of Context in Determining Consumer Response…
259
Figure 7. Price Variation in Canadian Fast Food Entrée Categories, May 2000 – May 2005
Table 9. BSE had Little Impact on Participation or Consumption of Fast Food Beef Entrees, Canada, May 2000 – May 2005
Lag (Beef entrees (0/1)) Lag (Number of beef entrees) Promotional deal (0/1) Price of beef entrée ($) Price of chicken entrée ($) Price of pizza entrée ($) Price of other entrée ($) Age of household head Married (0/1) Children (0/1) College degree (0/1) Female head works full time (0/1) English second language (0/1) Restaurant within 5 minutes (0/1) Came from work (0/1) Came from home (0/1) BSE event 1 (May 2003) BSE event 2 (Dec 2003) BSE event 3 (Dec 2004 / Jan 2005) Likelihood Ratio Index / Poisson R-squared
Participation Odds Ratio 2.118** 1.097** 1.850** 1.062 1.021 0.884 1.034 0.988** 1.350** 1.200** 0.935* 0.999 0.895** 0.765** 0.838** 0.743** 0.997 1.081 1.088 0.083
Consumption Marginal Effect -0.036 0.166** 0.342** -0.058 0.040 -0.082 0.084 -0.006** 0.520** 0.456** -0.099** -0.103** -0.126** -0.062* -0.193** 0.115** -0.023 -0.059 0.063 0.230
* and ** denote statistical significance of the underlying parameter at the .05 and .01 levels, respectively
260
Sayed Saghaian, Leigh Maynard and Michael Reed
Table 9 shows results from the participation and consumption components of the foodaway-from-home double-hurdle model. For ease of interpretation, odds ratios are shown for the logit participation regression, and marginal effects are shown for the Poisson consumption regression. In both models, almost all of the parameters were statistically significant except the price parameters and, importantly, the BSE event parameters. Explanatory power in both models was modest, and similar to the province-specific results of Maynard, Goddard, and Conley (2008). As in the Canadian food-at-home analysis, prior household behavior was an important predictor of current behavior. If a household recorded at least one beef entrée in its preceding report, it was over twice as likely, on average, to buy a beef entrée than households who listed no beef entrees in the preceding report. Each additional beef entrée recorded in the preceding report was associated with about a 10% higher average probability of a current beef entrée purchase, and about 0.17 more beef entrée purchases. Promotional deals almost doubled the odds of purchasing a beef entrée, and increased the average monthly by 0.34 entrées. Households containing married couples and children were more likely to buy beef entrees, and more of them. Older consumers, and those for whom English was a second language, were significantly less likely to buy fast food beef entrees. The finding that prices were not significant determinants of beef entrée purchases is not entirely surprising when considering the context of fast food purchases. As shown in Figure 7, the relative ordering of fast food entrée prices is stable over time, i.e., beef products are generally less expensive than chicken products, which in turn are less expensive than pizza. Awareness of typical fast food prices is likely to be high among most consumers, and few lower-cost, low-preparation alternatives are available. Convenience and product attribute preferences are likely to be among the most important demand determinants. The failure of BSE events to influence the likelihood or quantity of beef entrée purchases is more surprising, as it conflicts with the statistically and economically significant impacts from the food-at-home analysis. The results correspond roughly with those of Maynard, Goddard, and Conley (2008), who found limited evidence that BSE media coverage affected the likelihood of purchasing fast food beef entrees, but not the quantity, among Ontario consumers. Based on the food-at-home results, we hypothesized that distinguishing among BSE events would also be important in the fast food context, but the results do not support this hypothesis, or challenge the validity of the undifferentiated media index used by Maynard, Goddard, and Conley (2008). The results also fail to support the hypothesis that consumers would perceive greater food safety threats from fast food beef products than from beef products purchased for at-home consumption. The two Canadian analyses presented here were each based on extremely large samples that appeared to be demographically representative of Canada’s general population, based on 2001 Census figures (Statistics Canada, 2006b). It is therefore interesting that responses to BSE would be so different across the food-at-home and food-away-from-home venues. Given the prominence of food safety issues in recent regulatory and trade policy debates, exploring the reasons for divergent behavior across market venues is a useful topic for further study.
The Importance of Context in Determining Consumer Response…
261
Summary In the first section, we examined the impact of two beef safety scares on retail-level per capita meat consumption and prices in Japan. The objective was to investigate Japanese consumer reactions to the news of BSE discoveries at the retail level to better understand consumer responses to beef safety scares, and help the industry restore consumer confidence after food safety crises. This in turn, provides opportunities for national-level product differentiation based on beef quality and traceability. Consumers account for the nature of the contamination and respond to information regarding the origin and type of contaminated beef products with their purchasing decisions. Beef exporters and producer groups can use the study’s findings as a rationale for effective and transparent communication with consumers, and provide credible quality assurances. The findings emphasize the need for beef industry representatives to be immediately involved in providing accurate information when a safety crisis arises. Consumers consider food safety an entitlement, and most are unlikely to pay large premiums for safety assurance unless crises are frequent or locally prominent. Credible efforts that raise consumer confidence in an entire nation’s beef supply may be justified in order to reduce erosion of demand and market share when safety crises do occur. In the second part of this study, we addressed the dynamic impact of the 2003 BSE discovery on the U.S. beef sector. Time-series analysis and historical decomposition of weekly feedlot, wholesale, and retail beef prices were used to address the dynamics of price adjustment and causality along the U.S. beef marketing channel. The results showed that price transmission was bi-directional, determined through interaction between the different stages, and price adjustment was asymmetric with respect to speed and magnitude. The results revealed a differential impact of the exogenous shock on producers and retailers, leading to widening of price margins and pointing to imperfect price transmission, specifically at the retail level, which could have consequences for the efficiency and equity of the U.S. beef marketing channel. In the third part, two large, household-level datasets were used to evaluate the impact of BSE on retail beef purchases in grocery stores and fast food restaurants. Each dataset encompassed the first four Canadian-born cases of BSE, occurring between May, 2003 and January, 2005. Double-hurdle models were estimated to identify BSE impacts on households’ monthly probability of purchasing beef, and monthly frequency, quantity, and/or expenditure share of beef purchases. Consumers in Alberta and Ontario purchased more beef following the initial BSE discovery, even after controlling for other factors, but responded negatively to subsequent BSE events. Consumer response was stronger in Alberta than in Ontario. BSE did not appear to systematically affect Canadian fast food beef purchases.
References Agriculture and Livestock Industries Corporations (ALIC). ALIC Monthly Statistics. Tokyo, ALIC, Various Issues, 1994-2002. Babula, R., Bessler, D., & Payne, W. (2004). Dynamic relationships among U.S. wheatrelated markets: Applying directed acylic graphs to a time series model. Journal of Agricultural and Applied Economics, 36, 1, 1-22.
262
Sayed Saghaian, Leigh Maynard and Michael Reed
Baker, G.A. (1998). “Strategic implications of consumer food safety preferences.” International Food and Agribusiness Management Review, 1, 4, 451-463. Bakucs, L.Z. & Ferto, I. (2005). Marketing margins and price transmission on the Hungarian pork meat market. Agribusiness, 21, 2, 273-286. Becker, T., Benner, E. and Glitsch, K. (1996). Wandel des Verbraucherverhatens bei Fleisch. Agrarwirtschaft, 45, 267-277. Ben-Kaabia, M, Gill, J.M., & Boshnjaku, L. (2002). Price transmission asymmetries in the Spanish lamb sector. Paper presented at the X. Congress of European Association of Agricultural Economists, Zaragoza, Spain, 28-31, August. Bessler, D.A. & Akleman, D.G. (1998). “Farm prices, retail prices and directed graphs: Results for pork and beef.” American Journal of Agricultural Economics, 80, 1144-1149. Bocker, A. (2002). Consumer response to a food safety incident: exploring the role of supplier differentiation in an experimental study. European Review of Agricultural Economics, 29, 1, 29-50. Bocker, A. & Hanf, C.-H. (2000). Confidence lost—partially—regained: consumers’ response to food scares. Journal of Economic Behavior and Organization, 43, 471-485. Bordo, M. D. (1980). The effects of monetary change on relative commodity prices and the role of long-term contracts.” Journal of Political Economy, 88, 6, 1088-1109. Boyd, A.D. and C.G. Jardine. (2007). Public risk perceptions of BSE and vCJD: An Albertan perspective. Presented at the meetings of the Agricultural Institute of Canada, Edmonton, Alberta. Brown, D.J. and L.F. Schrader. (1990). Cholesterol information and shell egg consumption. American Journal of Agriculture Economics 72: 548-555. Burton, M., R. Dorsett, and T. Young. (1996). Changing preferences for meat: evidence from UK household data, 1973-93. European Review of Agricultural Economics 23: 357-370. Burton, M. and T. Young. (1996). The impact of BSE on the demand for beef and other meats in Great Britain. Applied Economics 28: 687-693. Cameron, A.C. and P.K. Trivedi. (1986). Econometric models based on count data: comparisons and applications of some estimators and tests. Journal of Applied Econometrics 1:29-53. Chopra, A. & Bessler, D.A. (2005). “Impact of BSE and FMD on beef industry in U.K.” Paper presented at the NCR-134 Conference on Applied Commodity Price Analysis, Forecasting, and Market Risk Management, St. Louis, Missouri, April 18-19. Cragg, J.G. (1971). Some statistical models for limited dependent variables with application to the demand for durable goods. Econometrica 39:829-844. de Jonge, J., H. van Trijp, E. Goddard, and L. Frewer. (2006). The reciprocal relationship between public policy making and consumer confidence in food safety: a cross-national perspective. working paper, Wageningen University, Marketing and Consumer Behaviour Group. Fearne, A., Hornibrook, S. & Dedman S. (2001). “The management of perceived risk in the food supply chain: A comparative study of retailer-led beef quality assurance schemes in Germany and Italy.” Imperial College at Wye Research Paper. University of London, Wye, Ashford, Kent. TN255AH. Fox, J.A. & Peterson, H.H. (2002). “Bovine Spongiform Encephalopathy (BSE): Risk and implications for the United States.” Paper presented at the NCR-134 Conference on
The Importance of Context in Determining Consumer Response…
263
Applied Commodity Price Analysis, Forecasting, and Market Risk Management, St. Louis, MO, April, 22-23. Gardner, B.L. (1975). The farm-retail spread in a competitive food industry. American Journal of Agricultural Economics, 57, 399-409. Goodwin, B.K. & Harper, D.C. (2000). Price transmission, threshold behavior, and asymmetric adjustment in the U.S. pork sector. Journal of Agricultural and Applied Economics, 32, 3, 543-553. Goodwin, B.K. & Holt, M.T. (1999). Price transmission and asymmetric adjustment in The U.S. beef sector. American Journal of Agricultural Economics, 81, 3, 630-637. Greene, W.H. 2000. Econometric Analysis, 4th Edition. Upper Saddle River, N.J.: Prentice Hall. Heien, D.M. (1980). Markup pricing in a dynamic model of the food industry. American Journal of Agricultural Economics, 62, 10-18. Hicks, J. (1974). The crisis in Keynesian economics. Basic Books, New York. Johansen, S., & Juselius, K. (1992). “Testing structural hypothesis in a multivariate cointegration analysis of the PPP and UIP for the UK.” Journal of Econometrics, 53, 211244. Livanis, G. & Moss, C.B. (2005). “Price transmission and food scares in the U.S. beef sector.” Paper presented at the American Agricultural Economics Association Meetings, Providence, RI. Lloyd, T., McCorriston, S., Morgan, W., & Rayner, T. (2003) Food scares, market power and relative price adjustment in the UK. Discussion Papers in Economics, University of Nottingham, Discussion Paper No. 03/18. Luoma, A., Luoto, J. & Taipale, M. (2004). Threshold cointegration and asymmetric price transmission in Finish beef and pork markets. Pellervo Economic Research Institute working paper No. 70, November. Maddala, G.S. 1983. Limited-Dependent and Qualitative Variables in Econometrics. Cambridge, U.K.: Cambridge University Press. Marsh, T.L., Schroeder, T.C. & Mintert, J. (2004). “Impacts of meat product recalls on consumer demand in the USA.” Applied Economics, 36, 897-909. Maynard, L.J., E. Goddard, and J. Conley. (2008). Impact of BSE on Beef Purchases in Alberta and Ontario Quick-Serve Restaurants. Canadian Journal of Agricultural Economics 55,3: forthcoming. Maynard, L.J., J.G. Hartell, A.L. Meyer, and J. Hao. (2004). An experimental approach to valuing new differentiated products. Agricultural Economics, 31: 317-325. Mazzocchi, M. (2005). “Modeling consumer reaction to multiple food scares.” Department of Agricultural and Food Economics Research Paper. The University of Reading. McCluskey, J.J., Grimsrud, K.M., Ouchi, H. & Wahl, T.I. (2004). “BSE in Japan: Consumers’ perceptions and willingness to pay for tested beef.” Washington State University Research Paper, TWP-111.Miller, J.D. & Hayenga, M.L. (2001). Price cycles and asymmetric price transmission in the U.S. pork market. American Journal of Agricultural Economics, 83,551-561. Mullahy, J. (1986). Specification and testing of some modified count data models. J. Econometrics 33: 341-365. NCJDSU (National Creutzfeldt-Jakob Disease Surveillance Unit). (2008). CJD Statistics. Western General Hospital, Edinburgh, Scotland.
264
Sayed Saghaian, Leigh Maynard and Michael Reed
Okun, A. (1975). Inflation: Its mechanics and welfare costs. Brooking Papers on Economic Activity, 2, 351-390. Paarlberg, P. L., Lee, J. G. & Seitzinger, A. H. (2003). “Measuring welfare effects of an FMD outbreak in the United States.” Journal of Agricultural and Applied Economics, 35, 1, 53-65. Pelzman, S. (2000). Prices rise faster than they fall. Journal of Political Economy, 108, 3, 446-502. Peng, Y., D. McCann-Hiltz, and E. Goddard. (2005). Consumer demand for meat in Alberta, Canada: Impact of BSE. Presented at the 2005 meetings of the American Agricultural Economics Association, Denver, Co.. Peterson, H.H. & Chen, Y.J. (2005). “The impact of BSE on Japanese retail meat demand.” Agribusiness: an international journal, 21, 3, 313-327. Piggott, N.E. & Marsh, T.L. (2004). “Does food safety information impact U.S. meat demand?” American Journal of Agricultural Economics, 86, 154-174. Pritchett, J. and D. Thilmany. (2005). The cow that stole Christmas? Exploring the role of media coverage in recent BSE outbreaks. Western Economics Forum 4(2): 24-28. RATS (2004) User’s Guide, version 6, Estima. Richard, T.J. and L. Padilla. (2007). Promotion and fast food demand: Where’s the beef? Consumer and Market Demand Agricultural Policy Research Network Project Report CMD-07-03, Edmonton, Alberta. Saghaian, S. (2007). “Beef Safety Shocks and Dynamics of Vertical Price Adjustment: The Case of BSE Discovery in the U.S. Beef Sector.” Agribusiness: An International Journal, Vol. 23, Issue No. 3, 333-348. Saghaian, S., Hasan, M.F. and Reed, M. (2002). “Overshooting of agricultural prices in four Asian economies. Journal of Agricultural and Applied Economics, 34,1, 95-109. Saghaian, S., Maynard, L., & Reed, M. (2007). “The effect of E. Coli 0157:H7, FMD and BSE on Japanese retail beef prices: A historical decomposition.” Agribusiness: An International Journal, 23, 1, 131-147. Saghaian, S, and Reed, M. (2007). “Consumer Reaction to Beef Safety Scares.” International Food and Agribusiness Management Review, Vol. 10, Issue No. 1, 18-35. Schmidt, M. B. (2000). “The dynamic behavior of wages and prices: Co-integration tests within a large macroeconomic system.” Southern Economic Journal, 67, 123-138. Spirtes, P., Scheines, R., Meek, C., Richardson, T., Glymour, C., Boomsam, A., & Hoijtink, H. (1999). TETRAD II-Version 3.1 Tools for Causal Modeling. Pittsburgh PA. Carnegie Mellon University. Spirtes, P., Glymour, C. & Scheines, R. (2000). Causation, Prediction, and Search. New York: Springer-Verlag. Statistics Canada. (2006a). Canada’s beef industry and BSE. accessed October 5, 2006, Last modified June 28, 2006, http://www41.statcan.ca/0920/ceb0920_001_e.htm . Statistics Canada. (2006b). 2001 Census of Canada. accessed September 21, 2006, http://www12.statcan.ca/english/census01/home/Index.cfm . Swanson, N.R., & Granger, C.W.J. (1997). “Impulse response functions based on a causal approach to residual orthogonalization in vector autoregressions.” Journal of American Statistical Association. 92, 357-367. Tiffin, R. & Dawson, P.J. (2000). Structural breaks, cointegration and the farm retail price spread for lamb. Applied economics, 32,1281-1286.
The Importance of Context in Determining Consumer Response…
265
Vickner, S.S., D. Bailey, and A. Dustin. (2006). Estimating the market demand for valueadded beef: Testing for BSE announcement effects using a nested PIGLOG model approach. Presented at the 2006 meetings of the American Agricultural Economics Association, Long Beach, Ca. Yen, S.T. (1999). Gaussian versus count-data hurdle models: Cigarette consumption by women in the U.S. Applied Economics Letters 6: 73-76. Zachariasse, L.C. & Bunte, F.H.J. (2003). How are farmers faring in the changing balance dimensions of the food economy: Exploring the policy issues. Den Haag, 6-7, February.
INDEX A Aβ, 219, 220 academic, 42, 70, 192 access, viii, 2, 4, 5, 8, 11, 13, 16, 17, 21, 26, 47, 51, 56, 63, 64, 85, 170, 175, 176, 177, 179, 181, 185, 208 accountability, x, 74, 81, 101, 102, 106, 108, 167, 177 accounting, 32 accuracy, 181 achievement, 102, 103 ACM, 68, 69, 70, 71, 72, 233 acoustic, 203 acquisitions, 16, 20, 21, 22 activity level, 23 activity theory, 68 acute, 106 Adams, 108, 110 adaptation, 22, 24, 211 adjustment, 88, 237, 244, 245, 246, 247, 248, 250, 261, 263 administration, viii, x, 52, 73, 167, 168, 182, 186, 231 administrative, 104 administrators, 81 adolescents, 106 adult, 62, 157, 236 adults, 58 adverse event, 103 AEA, 164 age, 23, 105, 150, 151, 153, 157, 158, 160, 162, 211, 244, 252, 254, 255, 258 agent, 43, 169, 179 agents, 33 aggregation, 63, 64 agricultural, 245, 264 agricultural market, 245 agriculture, 250 aid, 64, 104, 237 aiding, 32, 178 air, 66, 70
air traffic, 66, 70 airports, 48 Alberta, xii, 236, 250, 251, 252, 254, 255, 257, 261, 262, 263, 264 algorithm, xi, 213, 216, 221, 222, 223, 224, 227, 231, 234, 237 allocative efficiency, 168 alternative, vii, 25, 26, 31, 95, 214, 231 alternatives, 9, 21, 53, 65, 260 altruism, 144 ambiguity, 97 ambulance, 52 American Academy of Pediatrics, 106, 107, 110 Amsterdam, 135, 164 analysis of variance, 147 analysts, 88 anger, 74, 179 animals, 236, 240, 250, 251 annotation, 49 ANOVA, 146, 147 anti-bacterial, 7 antibiotic, 13 antitrust, 212 anxiety, 240 application, viii, 5, 42, 55, 73, 81, 168, 215, 231, 252, 262 argument, 3, 17, 32, 36, 58, 84, 108, 137, 179, 193 Arizona, 110 articulation, 66 Asia, 7, 8, 10, 13, 187 Asian, 8, 10, 264 Asian firms, 8 assessment, viii, 32, 63, 73, 76, 80, 180, 183, 206 assets, 20, 172, 182, 183 assignment, 192, 215, 226, 234 assumptions, 50, 117, 121, 217, 218 asymmetric information, 176 asymmetry, 181, 196, 206, 245, 247 asynchronous, 65, 67 Atlas, 211 ATM, 84 atmosphere, 194 attitudes, viii, 36, 37, 38, 40, 42, 47, 82, 85, 108, 111, 112
268
Index
attractiveness, vii authority, 106, 182 autocorrelation, 238 automakers, 207 automotive sector, 193 autonomy, 15, 19, 20, 24, 25, 26, 170 availability, vii, 2, 21, 64, 185 average costs, 174 averaging, 22, 251 avoidance, 17 awareness, vii, viii, 15, 25, 47, 48, 49, 50, 51, 52, 53, 54, 55, 56, 57, 58, 59, 60, 61, 62, 63, 64, 65, 66, 68, 69, 84, 236
B bacterial, 9 bacterial strains, 9 banking, 34 bargaining, 164, 165, 186, 196, 205 barrier, 13, 121 barriers, 5, 9, 10, 55, 100, 103, 108 basic research, 16, 179 beef, xi, xii, 235, 236, 237, 239, 240, 241, 242, 243, 244, 245, 246, 247, 248, 250, 251, 252, 253, 254, 255, 257, 258, 259, 260, 261, 262, 263, 264, 265 beef products, 236, 244, 254, 257, 260, 261 beef sector, 247, 261, 263 behavior, xii, 32, 33, 36, 38, 39, 42, 49, 58, 175, 210, 236, 244, 245, 248, 253, 260, 263, 264 Belgium, 192 beliefs, 77, 85, 97, 103, 143 benchmark, 192 benefits, ix, xi, 3, 7, 55, 61, 78, 79, 88, 89, 96, 99, 109, 167, 169, 170, 171, 174, 175, 178, 179, 180, 181, 183, 186, 196, 206, 213 Best Practice, 209 beverages, 258 bias, 21, 53 blame, 74 blaming, 74 blog, 55 blood, 79 blood pressure, 79 blurring, 109 Boeing, 20 bonding, 193 bonds, 77 Boston, 14, 70, 71, 111, 168, 187, 192, 195, 208, 211, 212 bounds, 59 brain, 235, 244 Brazilian, 3, 14, 210 breakdown, 56, 102 Britain, 212 brokerage, 34 Brooklyn, 69 brothers, 140
browsing, 57 Brussels, 208 Buenos Aires, 98 buffer, 6 building blocks, 179 Bulgaria, 144 burning, 80 business management, vii buyer, 179, 193, 196, 197, 204, 207, 208, 209
C CAD, 124 campaigns, 175 Canada, xii, 27, 30, 68, 70, 71, 112, 211, 235, 236, 250, 251, 259, 260, 262, 264 Canadian beef, 250 cancer, 38 capacity, 20, 79, 117, 170, 179, 180, 182, 184, 185, 186, 196, 205, 209, 210, 217, 218, 231 capital cost, 4 capitalism, 14, 44, 212 cardiac arrest, 52 cardiac output, 79 carmakers, xi, 191, 192, 193, 195, 196, 197, 198, 199, 202, 203, 205, 206, 207, 208 carrier, 198, 200, 201, 204 case study, 12, 20, 96 category a, 253 cattle, 244, 246, 248 causal relationship, 237, 247 causality, 247, 250, 261 CBS, 142, 163, 165 Census, 260, 264 centralized, 215 CEO, 5, 6 changing population, 222, 232 channels, xi, 52, 56, 57, 59, 204, 235, 245 chemicals, 12 chicken, 239, 257, 258, 259, 260 child abuse, 112 child protection, 106, 108, 111 child welfare, 177 childcare, x, 135, 136, 137, 140, 141, 142, 143, 144, 145, 147, 148, 149, 150, 151, 154, 157, 158, 159, 160, 161, 162, 165, 166 children, ix, 81, 99, 100, 104, 105, 106, 107, 108, 109, 111, 136, 138, 139, 140, 141, 142, 143, 144, 145, 146, 147, 148, 149, 150, 151, 152, 153, 154, 155, 156, 157, 158, 159, 161, 162, 164, 165, 175, 177, 252, 255, 258, 260 China, 1, 2, 5, 6, 7, 8, 9, 11, 12, 13, 14 Christmas, 264 chronic disease, 105 Chrysler, 195 circulation, 185 citizens, ix, 54, 83, 84, 85, 87, 88, 168, 175, 177 citizenship, 88
Index civil society, 184 classes, 61, 67, 86, 91, 92, 93, 94, 152, 231, 233 classification, 172, 199 classroom, 67, 72, 92, 112 classrooms, 97, 98 clay, 124 cleaning, x, 135, 136, 137, 140, 141, 142, 144, 145, 149, 150, 157, 158, 159, 160, 162, 180 clients, 12, 81, 178, 180, 181, 195, 199, 204, 206, 207, 208, 234 clinical trial, 4 clinical trials, 4 cluster analysis, 21 clustering, 21, 22, 205 clusters, 22, 25, 80, 211 Co, 133, 183, 187, 264 codes, 102, 167, 252, 258 coercion, 88 cognition, 72 cognitive, 56, 59, 66, 89, 91 cognitive process, 59 coherence, 59 cohesion, 60, 101, 103 cohort, 23 coil, 204 collaboration, viii, 12, 20, 25, 47, 48, 53, 55, 56, 57, 58, 59, 60, 66, 68, 69, 70, 72, 73, 74, 75, 77, 81, 103, 104, 106, 107, 111, 182, 209, 211 collisions, 48 Colombia, 183, 184, 188 colonization, 32 commerce, 48 commercialization, 5 commodity, 43, 206, 262 commons, 169, 170 communication, viii, 1, 8, 9, 13, 16, 33, 35, 51, 52, 53, 59, 62, 63, 64, 66, 72, 73, 74, 75, 76, 77, 81, 85, 88, 95, 103, 104, 105, 106, 107, 108, 109, 112, 117, 120, 121, 186, 261 communication skills, 76, 85, 88 communication technologies, 1 communities, 49, 52, 53, 55, 56, 57, 63, 64, 65, 66, 71, 175 community, viii, 48, 49, 50, 54, 55, 57, 59, 64, 65, 66, 67, 70, 71, 72, 81, 101, 105, 117, 171 compensation, 182 competence, 8, 19, 58, 63, 86, 117, 185, 186, 209 competency, 97 competition, x, 89, 116, 167, 168, 172, 176, 177, 178, 179, 180, 181, 194, 196, 206, 207, 212, 245 competitive advantage, 2, 3, 9, 74, 194, 208, 209 competitive markets, 245 competitiveness, 117, 194, 206, 209 competitor, 10, 199 complement, 3, 90 complementarity, 119 complexity, ix, 7, 8, 14, 20, 49, 101, 104, 106, 107, 115, 116, 118, 119, 121, 126, 127, 128, 130, 134, 180, 223
269
compliance, 75 complications, 25 components, 20, 53, 59, 117, 192, 193, 197, 198, 199, 202, 203, 204, 205, 206, 208, 209, 210, 211, 214, 260 composition, 143, 205, 210 compounds, 5, 6, 11, 12 comprehension, 88 computation, 223, 228 computer science, 56, 65 computer technology, 70 computing, 71, 215, 218, 224, 234 concentration, 245 conception, 41, 207 conceptualization, 33, 40, 41, 42, 43 conceptualizations, 85 concrete, 50, 56, 57 concurrent engineering, 209 confidence, xi, 36, 85, 102, 235, 236, 240, 242, 261, 262 confidentiality, 106 configuration, 57 conflict, 36, 56, 60, 62, 102, 103, 111, 186 conformity, 178 confusion, 90, 207 Congress, 97, 262 conjecture, 215 connectivity, 50 consciousness, 84 consensus, 60, 100, 101, 169, 250 constraints, 143 construction, 31, 96, 183 consulting, 191 consumer choice, 240 consumers, xi, xii, 34, 35, 81, 172, 174, 176, 178, 197, 236, 240, 242, 244, 245, 250, 251, 252, 255, 257, 260, 261, 262 consumption, 31, 164, 165, 169, 170, 239, 241, 242, 243, 244, 250, 251, 252, 253, 254, 255, 258, 260, 261, 262, 265 consumption habits, 242 contamination, 261 context-dependent, 257 continuous data, 252 contractors, 6, 8, 10, 177, 180 contracts, 10, 179, 180, 182, 183, 184, 194, 195, 205, 214, 217, 218, 231, 245, 247, 248, 262 control, vii, viii, x, 4, 9, 12, 13, 18, 19, 21, 31, 32, 33, 36, 37, 38, 39, 40, 41, 42, 43, 45, 47, 58, 70, 102, 167, 168, 175, 176, 177, 178, 179, 181, 195, 203, 205, 255 convergence, 57, 58, 61 convex, x, 115, 118, 125, 216, 220, 221, 222 cooking, x, 135, 136, 142, 145, 150 cooling, 203 cooperative learning, 89, 96, 98 coordination, 4, 8, 10, 14, 51, 52, 53, 67, 70 Copenhagen, 2, 15 corporations, 179, 192
270
Index
correlation, 25, 41, 237, 239 correlation coefficient, 25, 239 corruption, x, 167, 177 cosine, 254 cost curve, 173, 174 cost effectiveness, 11 cost saving, 2, 8, 180, 196 cost-effective, 108 costs, vii, xi, 4, 6, 8, 11, 13, 23, 25, 51, 117, 121, 149, 157, 167, 169, 173, 174, 175, 176, 177, 178, 179, 180, 181, 186, 196, 209, 213, 214, 217, 225, 226, 227, 228, 231, 236, 247, 264 countermeasures, 244 country-of-origin, xii, 236, 237 couples, 139, 150, 151, 152, 153, 154, 155, 157, 158, 159, 161, 163 coupling, 68 coverage, xii, 175, 232, 236, 250, 255, 257, 260, 264 covering, 21 creativity, 3, 56, 69, 75, 178 credibility, 185, 240 credit, 34 credit card, 34 criticism, 106 cross-cultural, 7, 8 cross-cultural differences, 7 CRS, 27 CT scan, 51 cues, 8, 59 cultural differences, 10 culture, x, 31, 32, 53, 74, 100, 101, 108, 109, 135, 137, 143, 153, 168 curiosity, 91 current prices, 175 curriculum, 87, 108 customer orientation, 31, 32, 33, 37, 40 customers, 16, 18, 20, 32, 33, 34, 36, 37, 38, 39, 40, 41, 42, 77, 168, 174, 178, 185, 194, 214, 215, 216, 217, 218, 219, 220, 221, 222, 231, 233 cycles, 263
D dairy, 237, 238, 239, 241, 242, 244 data gathering, 145 data set, 57, 223, 246, 258 database, 50, 57 death, 105, 106, 111 deaths, 51, 106, 250 decentralisation, 21 decision making, 52, 55, 62, 102 decision-making process, 77, 78 decisions, x, 11, 12, 19, 22, 32, 52, 72, 79, 81, 106, 115, 117, 118, 121, 130, 132, 163, 169, 179, 181, 185, 186, 240, 242, 252, 258, 261 decomposition, 236, 237, 239, 246, 248, 250, 261, 264 defense, 186
definition, xi, 1, 66, 78, 101, 116, 167, 245 delivery, viii, x, 1, 34, 35, 36, 50, 73, 76, 78, 79, 99, 105, 109, 136, 149, 158, 160, 162, 172, 173, 176, 177, 184, 192, 198, 206 Delphi, 193, 198, 203, 204, 206, 207, 208 demand, xii, 14, 119, 120, 125, 137, 143, 149, 150, 151, 152, 153, 154, 155, 157, 160, 161, 162, 164, 170, 173, 174, 176, 177, 178, 181, 183, 196, 235, 236, 237, 245, 247, 250, 258, 260, 261, 262, 263, 264, 265 demand curve, 173, 174 demographic factors, 138 Denmark, 15, 22, 133 density, 79 Department of Justice, 65 dependent variable, 255, 258, 262 desire, 2, 13, 17, 25, 38, 42, 205 destruction, 13, 20 detection, 184 developed countries, 137, 186 developing countries, 168, 175, 181, 182, 185 developing nations, 8 developmental psychology, 58 deviation, 24 dichotomy, 174 diet, 162 dietary, 164 dietary intake, 164 differentiated products, 263 differentiation, 85, 236, 242, 261, 262 direct cost, 180 direct costs, 180 direct investment, 17 direct observation, 66 directives, 103 disability, 105 disaster, 48 discipline, 100, 185 disclosure, 218 discounting, 254 discourse, 33, 37, 42, 60, 61, 63, 97 Discovery, 244, 264 discrimination, 164 diseases, 13, 175 disenchantment, 168 disequilibrium, 245 dissatisfaction, 77 distribution, 22, 23, 52, 150, 171, 183, 184, 216, 217, 218, 250, 253 diversification, 16 diversity, 101, 174 divestiture, 168, 182, 183, 185 division, 19, 67, 138, 139, 146, 151, 164, 214 doctors, 102, 103, 106, 110, 112 domestic tasks, x, 135, 136, 145, 166 dominance, 108 drug discovery, 2, 4, 6 drug targets, 6, 12 drugs, 2, 4, 5, 13, 176
Index drying, 3 duplication, 172 durable goods, 262 duration, viii, 47, 56, 116, 120, 122, 123, 124, 184 duties, 6, 168
E earnings, 137 Eastern Europe, 112 eating, 138, 139, 145, 146, 147, 149, 158, 162, 173, 235 ecology, 52 economic efficiency, 177 economic institutions, 212 economic losses, xi, 235 economic theory, 144 economics, 7, 165, 169, 172, 188, 192, 210, 211, 212, 263, 264 economies of scale, 4, 17, 172, 178 educated women, 151, 154 Education, 58, 60, 70, 85, 87, 89, 91, 93, 95, 96, 97, 98, 99, 103, 105, 107, 109, 110, 111, 112, 138, 139, 145, 146, 150, 165, 172, 251 educational institutions, 104 educational objective, 87 educational policies, 84 educational services, 183 educators, 84, 85, 109 efficient resource allocation, 175 egg, 262 elaboration, 196, 199, 200, 201, 202, 203 elasticity, 160, 245 electricity, 171, 173, 177, 184, 185 electrochemistry, 96 email, 53, 54, 67, 213 Emergency Department, 51 emergency departments, 68 emergency management, viii, 48, 64, 65 emergency planning, 48, 64, 65 emergency response, 48 emerging economies, 1 emotional, 89, 96 emotions, 32, 42, 89 empathy, 36, 38, 42 employee orientation, 40, 41 employees, viii, 5, 6, 10, 20, 21, 22, 23, 32, 33, 36, 37, 38, 40, 41, 42, 43, 73, 77, 81, 228 employment, 7, 73, 78, 81, 142, 180 employment status, 142 empowered, 101, 106, 107 empowerment, 41, 168 encouragement, 74 end-users, 52 energy, 74, 75, 78, 80 engagement, 54, 59, 74, 84, 92, 95, 185 England, 110 enterprise, 14, 75, 84, 183, 192, 208
271
entrepreneurship, 70, 164 environment, 42, 51, 59, 62, 76, 77, 81, 84, 85, 101, 119, 184, 186, 206, 223 epistemological, 43 equality, 126 equilibrium, 143, 237, 244, 245, 246, 247 equipment, 16, 17, 36, 54, 214 equity, x, xi, 167, 169, 171, 174, 176, 180, 186, 245, 261 erosion, 207, 242, 261 estimators, 262 ethnic groups, 145, 146, 147 ethnicity, x, 135, 136, 137, 144, 146, 150, 164 Eurasia, 96 Euro, 148, 163 Europe, 1, 2, 180, 187, 192, 197, 205, 207, 208, 250 evolution, 194, 210 examinations, 80 exclusion, 169, 170 execution, 59, 134 exercise, 43, 49, 79, 205 expenditures, x, 135, 136, 141, 142, 144, 148, 149, 150, 151, 152, 153, 154, 155, 156, 157, 158, 160, 161, 162, 165, 253, 254, 258 experimental design, 79 expert, 54, 75 expertise, 4, 11, 17, 54, 60, 75, 79, 99, 100, 101, 106, 107, 109, 117, 118, 120, 122, 184, 185, 186 explicit knowledge, 40 exploitation, 16, 196 exports, 250 external relations, 205 external validity, xi, 191 externalities, 169
F face-to-face interaction, 8 facilitators, 60 factual knowledge, 109 failure, 25, 59, 75, 99, 106, 117, 118, 119, 120, 124, 127, 128, 130, 131, 133, 175, 177, 180, 183, 215, 216, 217, 218, 221, 223, 226, 227, 228, 232, 260 family, 105, 140, 145, 148, 149, 164, 165 farmers, 265 farming, 6 fast food, xii, 236, 250, 251, 258, 260, 261, 264 fax, 73 fear, 4, 106, 177, 193, 235 fears, 255 February, 249, 254, 265 fee, 169, 182, 217, 231 feedback, 40, 41, 49, 101, 206 feeding, 236 feelings, 50, 91 fees, 176, 186, 231 feet, 208 females, 137, 138, 139, 140, 142, 151, 152, 153
272
Index
fidelity, 59 filters, 57 financial resources, 5 financing, 179 fire, 48, 179, 196 fires, 48 firms, vii, xi, 4, 6, 7, 8, 9, 10, 11, 12, 13, 16, 17, 20, 21, 22, 23, 25, 31, 32, 33, 34, 36, 37, 38, 40, 41, 42, 43, 116, 117, 121, 122, 168, 174, 181, 184, 191, 192, 193, 194, 197, 203, 206, 207, 208, 209, 211, 231, 242 first responders, 48, 49, 65 fish, 77, 237, 238, 239, 242, 243, 244 fixed costs, 173 flank, 237 flexibility, 2, 6, 65, 168 flood, 64 flooding, 48 flow, 3, 14, 119, 196, 248 fluid, 53 focusing, vii, ix, 15, 19, 31, 37, 53, 83, 87, 90, 92 food, x, xii, 135, 136, 137, 140, 142, 144, 147, 148, 149, 158, 159, 160, 162, 164, 165, 166, 235, 236, 237, 240, 241, 242, 244, 250, 251, 255, 257, 258, 260, 261, 262, 263, 264, 265 food industry, 244, 263 food safety, 235, 236, 237, 240, 241, 242, 244, 255, 260, 261, 262, 264 Ford, 73, 193, 195, 196, 199, 200, 201, 203, 205, 207, 208, 209 foreign firms, 22 foreign investment, 5 formal education, 1 Foucault, 42, 44 Fox, 112, 220, 234, 236, 262 fragmentation, 180 France, 192 fraud, x, 167, 177 freedom, 19, 101 fruits, 244 frustration, 75 fuel, 203, 204 function values, 253 funding, 6, 7, 183 funds, x, 167
G Gamma, 199 garbage, 179 gas, 171 gauge, 79 Gaussian, 265 gay men, 38 gender, 88, 137, 145, 154, 164 General Accounting Office, 188 General Motors, 199 generation, 2, 20, 118, 123
genetic disease, 5 genome, 57 Germany, 14, 23, 242, 262 girls, 50, 175 global competition, 3 global demand, 18 global warming, 84 Globalization, 27, 28, 211, 212 goals, viii, 6, 15, 19, 32, 47, 49, 50, 55, 58, 59, 60, 61, 62, 63, 68, 75, 77, 78, 85, 101, 103, 109, 180 goods and services, 16, 143, 150, 169, 170, 171, 172, 174, 176, 179, 180, 183, 184, 247 google, 59 governance, 106, 107, 179, 180, 185, 187, 188 government, x, xi, 2, 4, 5, 7, 42, 43, 55, 104, 105, 167, 168, 169, 170, 171, 172, 173, 174, 175, 176, 177, 178, 179, 180, 181, 183, 184, 185, 186, 187, 189, 236, 239, 242, 244, 250, 255 government failure, xi, 167, 169, 176 government policy, 242 grades, 61, 246 grandparents, 140 grants, 5, 7 graph, 237, 247 grass, 242, 244 gravity, 18 grazing, 170 Great Britain, 197, 262 group activities, 60 group processes, 58, 60, 63 group work, 58, 88, 89, 91, 92, 95, 97 groups, viii, 23, 24, 25, 48, 49, 50, 56, 59, 60, 61, 65, 66, 87, 88, 89, 90, 91, 92, 93, 94, 95, 96, 104, 105, 107, 109, 137, 139, 140, 145, 146, 147, 148, 149, 153, 157, 160, 161, 175, 181, 185, 197, 214, 255, 261 growth, 3, 5, 16, 21, 49, 50, 56, 75, 104, 168 guidance, 76, 91, 194 guidelines, viii, 12, 73, 74, 76, 105, 108 Guinea, 184
H H1, 224, 225, 226 H2, 224, 225, 226 handling, 86, 215 harm, 70, 107 harmony, 70 Harvard, 27, 29, 71, 72, 111, 112, 188, 209, 210 Hawaii, 69 health, ix, 51, 68, 74, 76, 78, 99, 102, 104, 105, 106, 110, 112, 113, 150, 151, 153, 162, 166, 170, 171, 172, 175, 176, 178, 181, 185, 250 health care workers, 51 health services, 105, 170, 171 health status, 175
Index healthcare, ix, 51, 52, 67, 68, 74, 76, 77, 78, 81, 82, 99, 100, 101, 102, 103, 104, 105, 106, 107, 108, 109, 111, 112, 113, 175, 181 hearing, 92 heart, 50, 105, 110 hedonic, 258 Henry Ford, 73 heterogeneity, 21, 245, 247, 255 heuristic, 35, 225, 226, 227 high pressure, 51 high school, 50, 71, 97 higher education, 98, 178 higher-income, 140 high-tech, 17 hip, 101 hips, 205 hiring, 6, 8, 48, 179 holistic, 78 Homeland Security, 72 homogeneity, 21 Honda, 208 honesty, 101 horizon, 118, 239, 248 horse, 237 hospital, 38, 51, 52, 65, 67, 69, 81, 101 hospitals, 4, 51 host, 55, 140 household, x, 135, 136, 137, 138, 139, 140, 141, 142, 143, 144, 145, 146, 147, 148, 149, 150, 151, 152, 153, 154, 155, 156, 157, 158, 160, 161, 162, 163, 164, 165, 175, 252, 254, 255, 257, 258, 259, 260, 262 household commodities, 143 household income, x, 135, 136, 140, 141, 150, 153, 157, 160, 161, 163 household tasks, 140, 142, 145, 150, 151 households, x, 135, 136, 137, 138, 139, 140, 141, 142, 143, 144, 145, 146, 147, 148, 149, 150, 151, 153, 154, 155, 157, 158, 160, 161, 162, 165, 166, 251, 252, 257, 258, 260, 261 housing, 170 hub, 19, 194, 197 human, 28, 32, 57, 58, 59, 63, 65, 66, 72, 74, 250 human behavior, 32 human cognition, 72 human development, 58, 59, 63, 65, 66 human genome, 57 human mental development, 59 humans, 236 Hungarian, 262 husband, 151 hypothesis, 23, 151, 168, 238, 239, 241, 245, 246, 247, 260, 263 hypothesis test, 246
I ice, 245, 246, 247, 264
273
id, 71, 247, 250 identification, 112, 116, 181, 235 identity, 5, 31, 101, 108, 110, 112, 198 Illinois, 115, 134 imbalances, 105 immigrant mothers, 143 immigrants, x, 135, 136, 137, 139, 141, 142, 143, 144, 145, 146, 148, 150, 157, 160, 162 immunological, 5 imperialism, 164 implementation, viii, 19, 40, 73, 84, 86, 96, 102, 103, 104, 168, 171, 176, 215, 228 import prices, 239 imports, 244 incentive, 10, 25, 177, 221 incentives, 4, 175, 184, 196, 244 incidence, 244 inclusion, 78, 79, 258 income, x, 135, 136, 140, 141, 142, 149, 150, 151, 152, 153, 154, 155, 157, 158, 159, 160, 161, 163, 165, 175, 183, 250, 252, 255 income distribution, 250 incomes, 149, 157 incubation, 235 incubation period, 235 incurable, 104 independence, 55, 185, 186, 193 independent variable, 255, 258 India, 1, 2, 5, 6, 7, 8, 9, 11, 12, 13, 14 Indian, 6, 9, 13, 14 indication, 52, 139, 150, 244, 248 indicators, 168, 182, 198 indices, 195 induction, 32, 37, 43 industrial, 2, 21, 37, 43, 208, 209, 210, 223 industry, xi, xii, 1, 2, 3, 5, 9, 10, 12, 13, 14, 17, 19, 23, 81, 172, 185, 191, 192, 193, 194, 195, 197, 206, 207, 208, 209, 211, 212, 213, 214, 223, 224, 235, 236, 239, 242, 244, 245, 250, 255, 261, 262, 263, 264 inefficiency, 169, 170, 173, 186, 245 inequality, 88 inequity, 181 infants, 106, 175 infection, 175 infinite, 118, 215, 216 inflation, 142 information and communication technologies, 1 Information System, 70, 167 information systems, 51, 52, 54, 55, 72, 244 information technology, 50, 68, 71 infrastructure, 48, 52, 53, 182, 183, 185, 194 Innovation, 3, 9, 19, 27, 28, 29, 30, 70, 71, 85, 98, 113, 132, 133, 134, 174, 178, 209, 210 insecurity, 91 insight, 12, 38, 143, 145, 148 inspection, 186, 196, 215 inspiration, 43 institutional change, 96
274
Index
institutions, 14, 72, 104, 185, 234 instruction, 58, 89 instructors, 58, 59, 61, 63 instruments, 57, 197 insurance, 175 intangible, 16, 17 integration, ix, 15, 16, 17, 19, 20, 22, 24, 25, 26, 51, 52, 53, 54, 58, 64, 74, 86, 87, 88, 98, 105, 115, 117, 118, 119, 120, 121, 124, 126, 128, 130, 131, 165, 197, 205, 210, 245, 263 integrity, 109 Intel, 16, 68 intellectual property, 2, 9, 12 intelligence, 206 intensity, 59, 60, 95 intentions, 59, 82 interaction, viii, 1, 7, 12, 16, 33, 36, 38, 48, 53, 56, 58, 59, 60, 61, 63, 70, 101, 146, 147, 196, 210, 223, 248, 258, 261 interaction effect, 146, 147 interaction effects, 146 interactions, viii, xi, 8, 47, 49, 55, 56, 58, 59, 60, 61, 62, 63, 68, 96, 117, 164, 193, 209, 213 interdisciplinary, 56, 70, 84 interest groups, 185 interface, 211 interference, 170, 185, 186, 216, 233 internalization, 194 internationalization, 1 internet, 2, 8, 16, 56, 65 interpersonal skills, ix, 83, 84, 89, 95 interpretation, 25, 42, 76, 92, 106, 260 interrelatedness, 22, 198 interrelationships, 84 interval, 228, 237, 252 intervention, 33, 80, 81, 105 interview, 145 interviews, 5, 34, 35, 55, 67 intrinsic, 41, 121 intrinsic motivation, 41 intuition, 221 investment, 2, 4, 5, 23, 110, 143, 170, 175, 180, 183, 184, 186, 211 investment incentive, 4 investors, 7, 170 ions, 34, 36, 40, 151, 234, 245 IRC, 53 Ireland, 4 island, 73, 210 ISO, 70 Italy, 70, 242, 262 iteration, 120, 122, 124, 222
J January, 69, 209, 211, 248, 249, 250, 253, 254, 255, 258, 261
Japan, xi, xii, 137, 166, 207, 208, 209, 210, 211, 212, 235, 236, 239, 240, 242, 244, 250, 261, 263 Japanese, xi, 7, 13, 17, 20, 26, 29, 137, 166, 192, 207, 208, 209, 211, 235, 236, 237, 239, 242, 244, 261, 264 job satisfaction, 100 jobs, 1, 2, 4, 7, 13, 51, 65, 117, 182, 221 joint ventures, 205 judge, 205 judgment, 102 Jun, 240, 241, 243 jurisdiction, 4 justice, 186
K K-12, 97 keiretsu, 207, 208 Kentucky, 235 Keynesian, 263 knowledge capital, 17 knowledge construction, 96 knowledge transfer, 11, 22, 24, 25
L labeling, xii, 236, 242, 244 labor markets, 4 labour, x, 4, 6, 7, 8, 11, 12, 13, 135, 136, 137, 138, 139, 140, 143, 144, 146, 149, 150, 153, 154, 158, 161, 163, 164, 165, 166, 175, 214, 217 labour market, 137, 138, 143, 150, 161, 164 laminated, 117 land, 170 language, 9, 10, 98, 121, 252, 257, 258, 259, 260 language barrier, 9, 10, 121 large-scale, 4 laundry, 141, 145, 150 law, 58, 220 layoffs, 6, 7 lead, ix, x, 8, 20, 50, 53, 58, 61, 83, 85, 87, 88, 104, 115, 117, 119, 128, 132, 168, 178, 180, 181, 193, 194, 196, 206, 245 leadership, viii, 13, 73, 75, 77, 81, 82, 102, 103, 168, 194 leadership style, viii, 73 leakage, 4, 12 lean production, 210 learners, 84, 85 learning, ix, 16, 49, 58, 59, 60, 61, 64, 65, 70, 71, 74, 77, 83, 84, 85, 86, 87, 88, 89, 90, 92, 93, 95, 96, 97, 98, 102, 104, 107, 108, 109, 111, 112, 118, 122, 178, 194, 196, 205, 206, 209, 210 learning environment, 65, 70 learning outcomes, 58, 60 learning process, 16, 86, 88 learning skills, 87
Index learning task, 95 legislation, 105 leisure, 139, 140, 141, 143, 146, 147, 149, 164 leisure time, 147 lens, 63 liberalization, 1 licensing, 104 liens, 210 life cycle, 116, 120, 166 lifelong learning, 85, 86, 87 lifespan, 99 lifestyle, x, 135, 136, 164 likelihood, xii, 60, 236, 252, 253, 255, 260 Likert scale, 21, 22 limitation, 126, 132 limitations, 54, 88 Lincoln, 207, 210 linear, 75, 149, 216, 224, 229, 245, 254, 258 linear programming, 216 listening, 88, 103 literacy, 84 Livestock, 237, 246, 261 local community, 54 local government, 55 location, 2, 4, 5, 20, 55, 64, 78, 142, 258 locus, 51 logistics, 192 London, 26, 27, 43, 44, 45, 70, 98, 110, 111, 112, 113, 188, 209, 210, 211, 262 long period, 245 long-term, 16, 19, 57, 64, 65, 66, 117, 123, 178, 179, 194, 239, 245, 247, 262 Los Angeles, 187 losses, 181 lower prices, 174 lower-income, 140 loyalty, 39, 40, 41, 207 lysis, 251
M machines, 12, 214, 216, 217 mackerel, 237 macroeconomic, 184, 264 mad cow disease, 235 mainstream, 205 maintenance, 14, 34, 62, 171, 179, 180, 182, 183, 205, 214 males, 138, 139, 140, 142, 151, 152, 153, 161 management, vii, viii, x, 31, 32, 33, 34, 36, 37, 38, 39, 40, 41, 42, 43, 44, 47, 48, 51, 52, 58, 60, 61, 62, 64, 65, 67, 72, 73, 76, 77, 78, 80, 85, 101, 103, 105, 107, 108, 109, 112, 165, 167, 168, 178, 179, 180, 181, 182, 184, 185, 187, 192, 203, 205, 209, 210, 211, 235, 262 management practices, 38, 112, 192 management styles, 40 mandates, 18, 20, 26
275
manufacturer, xi, 20, 176, 192, 213, 214, 215, 216, 217, 218, 220, 221, 224, 228, 229, 230, 231, 232 manufacturing, 1, 3, 4, 17, 73, 76, 116, 120, 192, 205, 206, 211, 214, 234 mapping, 184 marginal costs, 173 marginal product, 143 marginal revenue, 173 marital status, 141 market, x, xi, 2, 6, 7, 11, 13, 15, 16, 18, 22, 24, 33, 77, 116, 117, 118, 119, 120, 124, 133, 137, 138, 139, 143, 144, 149, 150, 151, 161, 163, 164, 165, 167, 168, 169, 170, 172, 174, 175, 176, 177, 180, 185, 188, 197, 206, 236, 244, 245, 246, 247, 260, 261, 262, 263, 265 market access, 15 market concentration, 245 market discipline, 185 market economy, 169 market failure, xi, 167, 169, 170, 172, 175, 176 market prices, 175, 245 market share, 16, 117, 119, 245, 261 marketability, 117 marketing, ix, xii, 31, 33, 39, 44, 115, 116, 118, 119, 120, 121, 122, 124, 125, 126, 127, 128, 132, 209, 236, 244, 245, 247, 250, 261 markets, xi, 16, 17, 18, 23, 116, 117, 174, 179, 209, 235, 236, 245, 251, 261, 263 Markov, 226, 234 marriage, 164, 165, 207 married couples, 260 Maryland, 65 Mathematical Methods, 234 matrix, 237, 239, 247 MCC, 205 meals, x, 135, 142, 143, 145, 151, 153, 258 meanings, viii, 47, 49, 88, 89 measurement, 34, 36, 37, 38, 39, 40, 42, 66, 168, 209 measures, 12, 41, 66, 68, 106, 180, 181, 187, 215, 244, 252, 255 meat, 236, 239, 242, 250, 251, 252, 253, 254, 255, 261, 262, 263, 264 media, xii, 53, 57, 66, 72, 236, 248, 250, 251, 255, 257, 260, 264 mediation, 62, 85, 186 medical care, ix, 99 medical school, 86 medical student, 112 medication, viii, 51, 52, 73 medicine, 101, 103, 106 membership, 50, 74, 78 memorizing, 85 memory, 58 men, 38, 137, 138, 139, 140, 151, 153 merchandise, 252, 254, 255 mergers, 13, 16 merit goods, 170, 171, 172 metacognitive, 58, 89 metaphors, 64
276
Index
metric, 244 Mexico, 184, 208, 210 mice, 98 microeconomic theory, 143 Microsoft, 68 middle class, 197 migrant, 166 migrants, 166 Minnesota, 110 minorities, 165 minority, 56, 204 misconception, 207 Missouri, 262 misunderstanding, 102 MIT, 70, 208, 209, 211, 212, 234 mobile phone, 84, 214 mobile telephony, 20 mobility, 20 modeling, 64, 68, 217, 218, 246, 252 models, ix, xi, 32, 33, 34, 42, 81, 115, 117, 144, 164, 165, 191, 192, 193, 195, 197, 198, 199, 200, 201, 203, 205, 215, 216, 217, 226, 237, 239, 245, 251, 252, 253, 255, 260, 261, 262, 263, 265 modules, 53, 117, 121, 123, 205, 206 modus operandi, 207 mold, 77 molecular structure, 10 money, 137, 151, 153, 154, 157, 158, 161, 164, 179, 185, 214 monopoly, 172, 173, 174, 175, 176 monopsony, 180 Moon, 164 moral hazard, 175 mothers, 143, 153, 161 motivation, 26, 40, 41, 49, 77, 78, 87, 89, 95, 214, 221 motives, 16, 17, 144, 197 mouth, 35 movement, 11, 13 multidisciplinary, 87, 106, 110 multinational companies, 17 multinational corporations, vii, 15 multiple factors, 106 multiple regression, 41 multiple regression analysis, 41 multiplicity, 203 multivariate, 263 mutual respect, 102, 106, 109 mutuality, 207 myopic, 215, 226, 231
N Nash, 163 nation, 261 national, 2, 9, 10, 25, 177, 186, 251, 254 national culture, 25 National Science Foundation, 68
natural, 51, 58, 77, 172, 173, 174 negative relation, 150 negativity, 74 neglect, 105 negotiating, 53, 59 negotiation, 62, 85, 186 nerve, 257 net income, 150 net present value, 123 Netherlands, x, 23, 69, 135, 136, 137, 138, 139, 140, 141, 142, 144, 145, 162, 163, 164, 165, 166, 191, 192, 213 network, vii, 3, 15, 18, 19, 20, 21, 22, 25, 26, 48, 50, 55, 56, 57, 65, 67, 171, 173, 184, 194, 195, 196, 208, 210, 215, 216, 217, 219, 220, 221, 222, 234 networking, 17, 20, 21, 50, 75 neurological disease, 236 New Jersey, 7, 28 New York, 27, 44, 45, 68, 69, 71, 72, 82, 96, 112, 134, 165, 187, 188, 211, 212, 263 New Zealand, 187 newspaper coverage, 250 newspapers, 145 NHS, 102, 110, 112 Ni, 149, 150 Nielsen, 251 non-linear, 125 non-profit, 54, 55, 65, 187 normal, 92, 150, 160, 162, 207, 253 normal distribution, 253 normalization, 56, 219 normalization constant, 219 norms, viii, 9, 32, 38, 39, 42, 43, 47 North America, 192, 207, 208, 211, 250 North Carolina, 29, 213 Norway, 137 not-for-profit, 184 novelty, 134 null hypothesis, 238, 239, 246 nurse, 51, 52, 82, 102 nurses, 51, 52, 70, 76, 78, 81, 102, 103, 110, 112 nursing, 73, 75, 76, 79, 80, 81, 82, 101, 106, 111, 112 nutrition, 136
O obligations, 175 observations, 21, 52, 55, 79, 118, 195, 196, 237, 252, 257 obsolete, 123 odds ratio, 259, 260 offshore, vii, 2, 5, 6, 8, 11, 12 offshoring, vii, 1, 2, 4, 5, 6, 7, 9, 11, 13 older people, 151, 153 oligopolies, 245 oligopoly, 245 online, 50, 55, 65, 67, 71
Index on-the-job training, 32 openness, 78, 101 opportunism, 193, 196 opposition, 182 optimal performance, 119 optimization, x, 115, 118, 125, 126, 132, 234 oral, 90, 92, 95 oral presentations, 95 organization, vii, viii, 3, 4, 20, 31, 32, 33, 39, 41, 42, 43, 48, 49, 61, 73, 74, 75, 76, 78, 79, 80, 81, 87, 91, 117, 144, 178, 182, 210, 212 organizations, vii, 5, 13, 31, 32, 39, 42, 43, 48, 49, 54, 55, 65, 67, 71, 72, 74, 77, 99, 100, 104, 168, 176, 178, 179, 208, 255 orientation, 31, 32, 33, 37, 40, 41, 44, 168, 193 outpatient, 52 outsourcing, vii, ix, x, xi, 2, 3, 4, 5, 9, 10, 11, 13, 14, 115, 116, 117, 118, 119, 120, 123, 125, 126, 127, 128, 129, 130, 131, 132, 135, 136, 137, 140, 141, 142, 143, 144, 145, 147, 148, 149, 150, 151, 152, 153, 154, 155, 156, 157, 158, 159, 160, 161, 162, 163, 166, 167, 168, 169, 171, 177, 178, 179, 180, 181, 182, 185, 186, 187, 205, 212, 213, 214, 215, 216, 217, 228, 231 oversight, 186 overweight, 162, 164 ownership, 25, 55, 75, 164, 198 ozone, 84
P paper, vii, ix, 13, 14, 15, 16, 18, 21, 40, 48, 56, 57, 66, 68, 70, 72, 96, 105, 115, 117, 118, 132, 133, 163, 169, 187, 191, 208, 209, 210, 211, 212, 213, 216, 217, 231, 235, 262, 263 paradox, 212 parameter, 116, 121, 127, 130, 131, 223, 224, 253, 254, 259 parameter vectors, 253 parents, 106, 107, 148 Pareto, 172, 173 Paris, 96, 211 Parliament, 105 partnership, viii, 4, 5, 8, 13, 73, 102, 186, 210 partnerships, 3, 13, 74 passive, 84, 91, 143 path model, 41 pathways, 105, 106 patient care, 51, 68, 78, 100, 101, 103, 104, 106, 107, 108, 109 patient-centered, 100, 101, 108, 109 patients, 52, 65, 76, 81, 101, 104, 107, 109 PBL, ix, 83, 84, 85, 86, 87, 88, 89, 90, 91, 92, 94, 95, 96, 98 pedagogical, 61, 85, 86, 140 pediatric, ix, 99, 105, 108 peer, 62, 86 peers, 57, 58, 78, 97, 196
277
Pennsylvania, 47, 48, 49, 51 per capita, 241, 243 perception, 7, 34, 35, 39, 103, 177, 242, 244, 246, 255 perceptions, 34, 35, 36, 37, 38, 39, 40, 41, 97, 112, 240, 244, 246, 263 perfectly competitive markets, 245 performance, 8, 12, 17, 32, 37, 40, 41, 60, 63, 79, 93, 94, 101, 102, 103, 106, 111, 112, 117, 118, 119, 120, 124, 130, 132, 168, 176, 180, 182, 184, 186, 192, 196, 206, 209, 211, 215, 223, 226, 231 performance indicator, 168 periodic, 234 personal, 1, 5, 34, 35, 38, 59, 65, 71, 73, 74, 103, 138, 139, 146, 147, 181, 214 personal communication, 73 personal computers, 214 personal histories, 71 personal history, 65 personal relations, 5 personal relationship, 5 pharmaceutical, 2, 5, 7, 12, 13 pharmaceutical companies, 2, 5, 7 pharmaceutical industry, 5 pharmacokinetics, 5 Philadelphia, 72, 82 philosophy, 77, 98, 168 phone, 8, 21, 84, 144 physical environment, 36 physical sciences, ix, 83, 96 physical well-being, 104 physicians, 51, 69, 81 pilot study, 96 pipelines, 3 planning, viii, 40, 41, 45, 48, 49, 58, 59, 62, 64, 65, 72, 99, 116, 168 plants, 48, 193, 197, 205, 208, 209, 211 platforms, 17, 19, 198, 203 play, vii, 15, 16, 18, 58, 86, 87, 88, 89, 91, 101, 109, 118, 140, 171, 174, 181, 193, 211, 212 point like, 174 Poisson, 215, 252, 253, 254, 255, 259, 260 police, 48, 49, 101 policy makers, 107, 179, 244 policy making, 143 politicians, 84 politics, 177 poor, 100, 105, 108, 151, 163, 181, 185, 214, 217 poor health, 151 population, xi, 48, 80, 81, 138, 145, 171, 181, 185, 213, 214, 216, 218, 220, 222, 231, 232, 260 population size, 220, 222, 232 pork, 251, 252, 255, 262, 263 portfolio, 215 portfolios, 92 Portugal, 83, 167 positive correlation, 41 positive externalities, 171 positive relation, 150, 151
278
Index
positive relationship, 150 posture, 207 poultry, 237, 238, 239, 242, 243, 244, 251, 252, 254, 255 power, 42, 43, 48, 102, 105, 172, 180, 196, 205, 245, 253, 260, 263 PPP, 168, 263 PPPs, 168 pragmatic, 179 preference, xii, 102, 236 premiums, 236, 261 present value, 123 pressure, 13, 51, 79, 100, 140, 168 prevention, 112 preventive, 181 price changes, 258 price effect, 245, 248 price mechanism, 176, 179 prices, 150, 172, 174, 175, 176, 177, 181, 182, 236, 237, 239, 240, 242, 245, 246, 247, 248, 249, 250, 252, 253, 254, 258, 260, 261, 262, 264 pricing policies, 239 primary data, 204 primary school, 163 prior knowledge, 89 privacy, 57 private, x, xi, 5, 136, 141, 167, 168, 169, 170, 171, 172, 174, 175, 176, 177, 178, 179, 180, 181, 182, 183, 184, 185, 186, 242 private benefits, 175 private good, 169, 170, 174, 176 private investment, 184, 186 private sector, x, xi, 167, 168, 169, 170, 171, 172, 174, 176, 177, 178, 179, 180, 181, 182, 183, 184, 186 private-sector, 172 privatization, 172 proactive, 10, 250 probability, 88, 116, 117, 118, 120, 121, 122, 124, 127, 128, 130, 131, 215, 218, 219, 220, 252, 260, 261 probability distribution, 218 problem solving, ix, 83, 86, 89, 91, 104 Problem-based learning, 96, 98 problem-solving, 58, 88, 98, 103 problem-solving task, 88 procedural knowledge, 88 producers, 236, 242, 244, 248, 250, 251, 261 product attributes, 116 product design, 116, 118, 252 product life cycle, 122 product performance, 117, 119, 120 production, 4, 33, 59, 117, 120, 121, 137, 143, 144, 149, 163, 164, 165, 169, 170, 174, 175, 176, 179, 180, 196, 205, 209, 211, 242, 244 production costs, 179, 180 production function, 143, 164 productivity, 6, 8, 60, 62, 63 professional development, 65, 98, 101, 104, 178
professional growth, 49 professions, 100, 101, 104, 107, 113 profit, 10, 67, 119, 173, 176, 181, 206, 242, 244 profit margin, 242, 244 profitability, 39, 40, 41 profits, 119, 172, 173, 174, 176, 245 program, 7, 50, 65, 89, 178, 179, 180, 183 programming, 50 promote, ix, 13, 49, 50, 63, 83, 87, 88, 100, 105, 107, 109, 186 property, 2, 9, 12, 57, 121 protection, 4, 10, 100, 105, 106, 108, 109, 111, 179, 185 protein, 5 protocols, 244 prototype, 64, 65 prototyping, 65, 116 proximal, 58 proxy, 39, 125 PSA, 193, 195, 198, 200, 201, 203 psychology, 71, 101 public, x, xi, 2, 63, 64, 65, 94, 95, 100, 105, 107, 167, 168, 169, 170, 171, 172, 173, 174, 175, 176, 177, 178, 179, 180, 181, 182, 183, 184, 185, 186, 187, 188, 192, 244, 255, 262 public administration, x, 167, 168 public education, 183 public funds, x, 167 public goods, xi, 167, 168, 169, 170, 171, 172, 174, 176, 181, 184 public interest, 171, 180, 186 public investment, 2 public policy, 192, 262 public sector, x, xi, 100, 167, 168, 169, 170, 171, 175, 177, 178, 179, 180, 182, 183, 185, 186 public service, x, 167, 168, 177, 178, 179, 180, 181, 182, 185, 186, 187 Puerto Rico, 4 pupils, 171 purchasing power, 180
Q qualifications, 1, 13 quality assurance, 118, 236, 244, 261, 262 quality control, 196 quality of service, 77, 177, 180 quality orientation, 33, 40, 41 quality-management-system, 40 questioning, 86 questionnaire, ix, 21, 83, 90, 91, 92 questionnaires, 21 quizzes, 61
R race, 164
Index radio, 57 rail, 171 random, 21, 22 range, 5, 18, 23, 32, 34, 38, 56, 61, 63, 86, 99, 101, 104, 127, 128, 179, 197, 203, 228, 237, 258 rationalist, 53 rationality, 53, 206 raw material, 208 reading, 57, 92 reagents, 9 reality, ix, xi, 67, 103, 110, 115, 191, 193, 203, 204, 217, 232 reasoning, 58, 86, 87, 88, 89, 193 reasoning skills, 58 reciprocity, 66 recognition, 100, 101, 103, 106, 107, 176 reconcile, 207, 235 reconciliation, 98 recovery, 51 recruiting, 50, 54 recursion, 219 redistribution, 166 reduction, 2, 6, 100, 109, 184, 211 reflection, 49, 63 reflexivity, 56 reforms, 168, 171, 177, 185 regional, viii, 48, 49, 180, 207 regression, 151, 152, 154, 155, 159, 160, 237, 255, 258, 260 regression method, 151 regressions, 246, 255 regular, 38 regulation, xi, 42, 63, 167, 169, 174, 176, 177, 182, 185, 186 regulations, 175 regulators, 107, 174, 185, 186 regulatory bodies, 185 regulatory framework, 169, 177, 184 rehabilitation, 183 reimbursement, 179 rejection, 246 relationship, 5, 6, 8, 10, 12, 13, 23, 25, 26, 74, 102, 111, 128, 150, 180, 192, 193, 206, 218, 221, 237, 246, 247, 262 relationships, xi, 4, 7, 9, 10, 12, 13, 19, 49, 74, 75, 110, 111, 181, 191, 192, 193, 194, 195, 196, 197, 198, 199, 202, 203, 204, 205, 206, 207, 208, 209, 210, 211, 237, 239, 246, 247, 261 relevance, 86, 87, 177, 181 reliability, 36, 38, 42, 181, 216, 240, 241, 242 religion, 143 repair, xi, 34, 180, 213, 214, 215, 216, 217, 218, 221, 228, 231, 232 reputation, 240 research, vii, viii, 3, 4, 5, 6, 7, 10, 11, 13, 14, 16, 19, 20, 23, 26, 31, 32, 33, 34, 36, 37, 38, 39, 40, 41, 42, 43, 47, 50, 54, 56, 57, 58, 59, 63, 66, 68, 70, 78, 79, 81, 86, 89, 96, 97, 103, 104, 107, 110, 116, 119, 136, 138, 140, 142, 144, 146, 151, 153, 169,
279
178, 179, 180, 192, 197, 198, 199, 203, 212, 214, 231, 234, 245, 246 Research and Development (R&D), vii, 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15, 16, 17, 18, 19, 20, 21, 22, 23, 24, 25, 26, 27, 28, 29, 58, 133, 178 research design, 34 researchers, 7, 11, 19, 56, 57, 59, 66, 68, 119 residuals, 239, 247 resistance, 102 resolution, 51, 79, 86 resource allocation, xi, 169, 213, 216, 220, 222 resources, vii, viii, 5, 7, 8, 11, 15, 16, 17, 18, 19, 20, 21, 22, 23, 25, 26, 47, 48, 51, 53, 54, 55, 56, 57, 60, 61, 62, 65, 85, 87, 88, 100, 117, 169, 170, 176, 180, 182, 186, 193, 194, 208 response time, 234 responsibilities, 12, 49, 101, 102, 103, 107, 108 responsiveness, 36, 38, 42 restaurant, 38, 142, 258 restaurants, x, 135, 136, 142, 257, 261 restructuring, 211 retail, xi, xii, 34, 214, 235, 236, 245, 246, 247, 248, 249, 250, 254, 258, 261, 262, 264 retention, 40, 82, 96 returns, 170, 185 revenue, ix, 115, 173, 183 rhythms, 64 risk, 6, 9, 10, 14, 16, 19, 20, 64, 75, 76, 84, 95, 106, 175, 180, 183, 184, 193, 235, 244, 262 risk management, 235 risk perception, 244, 262 risks, 2, 65, 175, 178, 196, 206, 214 role conflict, 36 routines, 3, 8 routing, 216, 231, 233, 234 R-squared, 258, 259 rubber, 197 runoff, 55 rural, 48, 67, 175 rural areas, 175 rural population, 175 Rutherford, 195, 207, 208, 211
S safeguard, 40 safety, viii, 51, 67, 73, 77, 106, 109, 175, 176, 236, 240, 242, 244, 246, 247, 248, 250, 257, 261 sales, 5, 125, 214, 216, 217 salmonella, 244 sample, x, 9, 10, 23, 136, 144, 145, 148, 150, 151, 193, 195, 198, 199, 203, 204, 205, 207, 208, 237 SAS, 27 satisfaction, 34, 39, 40, 41, 53, 77, 100, 170, 214 savings, 11, 179, 180 scaffold, 58 scale economies, 175, 196 scarcity, 176, 185
280
Index
scheduling, 64 Schmid, 18, 29, 166 scholarship, 33, 56 scholarships, 175 school, 50, 67, 68, 70, 71, 82, 84, 86, 87, 88, 95, 97, 105, 140, 163, 175, 179, 183, 184, 186, 207, 251, 254 science education, 84, 85, 87, 88, 97 science educators, 87 scientific community, 57, 65 scientists, 6, 8, 13, 17, 55, 56, 58, 67 scores, 93, 94, 204 SCP, 136, 138, 139, 140, 141, 142, 165 scripts, 70 seafood, 252, 254, 255 search, 26, 57, 92, 96, 171, 211 searching, 57 seasonality, 254 Seattle, 77 second language, 257, 258, 259, 260 secondary education, 163 Secretary of State, 105 securities, 34 security, 12, 150 seed, 21 seeding, 50 seeds, 22 selecting, 21, 181 self-regulation, 42 semi-structured interviews, 5 sensitivity, 12, 118, 128 Sensitivity Analysis, 128 series, 12, 56, 57, 105, 237, 238, 239, 246, 248, 258, 261 service provider, 185 service quality, 33, 34, 35, 36, 37, 38, 39, 40, 42, 180 services, vii, x, xi, 1, 3, 14, 16, 31, 32, 33, 34, 37, 38, 39, 40, 41, 42, 43, 57, 73, 74, 77, 78, 79, 81, 101, 105, 109, 140, 141, 142, 143, 150, 163, 165, 167, 168, 169, 170, 171, 172, 174, 175, 176, 177, 178, 179, 180, 181, 182, 183, 184, 185, 186, 187, 189, 213, 214, 228, 247 severity, 51, 52, 250 sexually-transmitted diseases, 175 shape, 58, 121, 184 shaping, 42 shareholders, 168 shares, 255 sharing, 3, 8, 57, 60, 67, 88, 101, 102, 105, 106, 108, 194 sheep, 97 Shell, 3 shock, 237, 242, 244, 245, 246, 247, 248, 249, 250, 261 shocks, 235, 236, 237, 244, 245, 247 short period, 96 short run, 149, 247 shortage, x, 167 short-term, 103, 105, 182
shoulder, 39 side effects, 168 Siemens, 206 sign, 205 significance level, 238, 247, 248 signs, 203 similarity, 168, 252 simulation, 48, 118, 224 Singapore, 4 single-server queue, 215 sites, 19, 20, 37, 38, 39, 65, 76 situation awareness, 66 skeptics, 107 skills, ix, 7, 10, 16, 37, 40, 54, 58, 59, 60, 66, 74, 75, 76, 83, 84, 85, 86, 87, 88, 89, 95, 99, 100, 101, 102, 103, 104, 107, 108, 109, 178, 185, 186 skills base, 104 social awareness, viii, 47 social benefits, 175, 183 social capital, 50, 59, 60, 63, 65, 66 social care, 110 social class, 152 social costs, 180 social factors, 36 social infrastructure, 69 social integration, 88 social network, 10, 50, 65 social presence, 53 social psychology, 60 social responsibility, 89 social sciences, 65 social security, 150 Social Services, 187, 188 social skills, 84, 88 social work, 101 socialization, 103 socioeconomic, x, 135 sociology, 37, 43, 60, 210 software, viii, x, 47, 48, 49, 52, 53, 64, 65, 67, 115, 118, 119, 120, 125, 126, 132, 239, 248 solidarity, 74 solutions, 53, 54, 74, 75, 76, 78, 80, 87, 88, 90, 127, 128, 144, 174, 186, 223 sovereignty, 177 Spain, 137, 144, 165, 192, 262 spatial, 67, 205, 244 specialization, 84, 180, 206 specific knowledge, 21, 88 specificity, 180, 193, 196, 205 spectrum, 104 speed, 3, 119, 228, 237, 245, 246, 247, 250, 261 spheres, 164 splint, 51 springs, 198, 200, 201, 204 St. Louis, 79, 80, 82, 262 stability, 219, 228 stages, 4, 7, 8, 117, 119, 121, 248, 250, 261 stakeholders, 54 standard deviation, 91, 93
Index standard error, 148, 149, 160 standardization, 252 standards, 19, 36, 40, 41, 65, 75, 102, 104, 105, 178, 242 state-owned, 183 state-owned enterprises, 183 statistical analysis, 21 statistics, 11, 219, 238, 258 steady state, 216, 218 stereotypes, 103 stereotyping, 108, 111 stimulus, 178 stochastic, 119, 150, 158 stock, 3 storage, 223 strain, 9 strains, 9 strategic, 15, 16, 17, 19, 22, 25, 26, 33, 40, 41, 177, 178, 192, 194, 195, 208, 210, 211, 235, 236, 242 strategic management, 33, 178 strategic planning, 40, 41 strategies, 10, 15, 16, 17, 19, 21, 22, 26, 51, 61, 62, 63, 81, 85, 88, 89, 96, 101, 104, 136, 196, 204, 206, 209, 210 strategy use, 92 strawberries, 244 strength, 51 stress, 88, 105, 194 strikes, 196 structuring, 61, 63 student group, 60 student teacher, 85 students, ix, 50, 58, 59, 60, 61, 62, 63, 67, 71, 83, 84, 85, 86, 87, 88, 89, 90, 91, 92, 93, 94, 95, 96, 97, 98, 100, 104, 107, 108, 109, 110, 111, 112, 183 subjective, 38, 54, 66 subsidies, 141, 160, 162 subsidy, 174 substitutes, 143 substitution, 136 success rate, 20 suffering, 38 summaries, 65 supercomputers, 57 superiority, 21, 174 supervision, 9 supervisors, 95 suppliers, ix, xi, 3, 9, 12, 115, 117, 191, 192, 193, 194, 195, 196, 197, 198, 199, 200, 201, 202, 203, 204, 205, 206, 207, 208, 211, 240, 241, 242 supply, xi, 3, 4, 5, 7, 11, 14, 117, 140, 164, 171, 174, 176, 184, 191, 192, 194, 195, 196, 197, 198, 199, 202, 203, 204, 205, 208, 210, 211, 212, 235, 242, 244, 245, 247, 248, 250, 261, 262 supply chain, xi, 117, 198, 210, 212, 235, 245, 247, 262 supply shock, 245 support services, 180 surgery, 105, 106, 110
281
surveillance, 51 survivability, 117 survival, 168 surviving, 104 suspensions, 203, 204 sustainability, 185 Sweden, 23, 31, 137 switching, 38, 173, 196 Switzerland, 142, 166 symbiotic, 20, 26 symbolic, 72 symptoms, 235 synchronization, 49 synchronous, viii, 47, 65, 66, 67 synergistic, 26, 74 synergistic effect, 26 synthesis, ix, 83 systems, viii, 25, 47, 48, 51, 52, 54, 55, 57, 59, 60, 64, 65, 68, 70, 72, 112, 117, 118, 121, 168, 175, 178, 184, 185, 195, 197, 203, 204, 210, 221, 234, 244
T takeover, 17, 22, 23, 26 talent, vii, 2, 4, 7, 11 tangible, 16, 55 tangible benefits, 55 targets, 6, 7, 12, 75, 168 tariffs, 185, 186 taxes, 176 taxonomy, 68 TBO, 138, 145 teacher training, 86 teachers, ix, 59, 83, 84, 85, 86, 89, 90, 91, 92, 95, 96, 97, 109, 179 teaching, ix, 4, 69, 83, 84, 85, 86, 88, 90, 92, 94, 95, 97, 103 teaching experience, 85 team leaders, 102 team members, 59, 60, 61, 63, 74, 75, 100, 101, 102, 103, 107, 111 technicians, 51 technology, viii, 17, 19, 20, 21, 23, 47, 50, 53, 54, 59, 63, 66, 68, 71, 73, 84, 173 telecommunication, 38, 175 telecommunications, 38, 172, 177, 185 telecommunications services, 38 telephone, 67, 144, 145, 171, 173 telephony, 20 television, 84 temporal, 67, 246, 252 Tennessee, 20 Texas, 73 textbooks, 90 theory, 32, 33, 42, 66, 68, 78, 82, 122, 137, 143, 144, 149, 163, 164, 170, 179, 188, 197, 217, 237, 245, 258
282
Index
therapeutics, 7 thinking, 55, 56, 74, 89, 195 third party, xi, 191, 193, 199, 208 Thomson, 111, 121, 132 threat, 13, 235, 242 threats, 76, 250, 260 three-dimensional, 111 threshold, 56, 263 time, viii, ix, x, xi, 6, 8, 10, 11, 12, 19, 20, 21, 22, 23, 26, 36, 47, 49, 50, 52, 56, 57, 58, 59, 60, 61, 62, 63, 64, 65, 66, 68, 74, 75, 76, 77, 78, 80, 81, 85, 90, 91, 92, 100, 102, 115, 116, 117, 118, 119, 120, 121, 122, 124, 125, 126, 127, 128, 129, 130, 131, 132, 135, 136, 137, 138, 139, 140, 141, 142, 143, 144, 145, 146, 147, 148, 149, 150, 151, 152, 153, 154, 155, 156, 157, 160, 161, 162, 163, 164, 165, 166, 175, 179, 181, 195, 207, 211, 213, 214, 215, 216, 217, 218, 220, 221, 224, 225, 226, 227, 228, 229, 230, 232, 236, 237, 242, 253, 254, 257, 258, 259, 260, 261 time allocation, 118, 128, 130, 136, 137, 138, 139, 144, 146, 160, 162, 165, 166 time periods, 23 time series, 261 timing, 51, 52, 76 tissue, 244, 257 title, 105 Tokyo, 166, 261 tolerance, 34, 35, 36, 122 tolls, 183 total costs, 186 total product, 118 Toyota, 3, 14, 207, 208, 209, 211 tracking, 40, 50, 57, 66 trade, 1, 255, 260 trade liberalization, 1 trade policy, 260 trade-off, 121, 196 trading, 250 trading partners, 250 tradition, 66, 169, 172, 193 traffic, 66, 70, 170, 184, 233 trainees, 49 training, 6, 8, 10, 13, 32, 37, 38, 40, 43, 48, 49, 54, 63, 64, 67, 86, 100, 102, 103, 104, 106, 108, 111, 112 trans, 174 transaction costs, 179, 180, 209 transaction value, 209 transactions, 179 transcription, 5 transfer, 25, 177, 178, 183, 184 transformation, 187, 238 transition, 49, 164 translation, 121 transmission, 178, 244, 245, 246, 247, 261, 262, 263 transnational, 17 transparency, 181 transparent, 176, 185, 250, 261
transport, 171 transportation, 48 travel, 6, 10 trees, 260 trend, 3, 13, 19, 164, 194, 207, 214, 246 triage, 51, 52 trial, 79, 125, 223 trial and error, 125 tribal, 74, 103 trust, 4, 8, 10, 12, 36, 41, 42, 50, 60, 66, 75, 77, 101, 102, 193, 196, 207, 212, 240, 242 T-test, 24 turnover, 21, 22, 23, 54, 76, 77, 196 tutoring, 59
U ultrasound, 51, 52 uncertainty, 86, 134, 196 undergraduate, 50, 89, 96, 104, 112 undergraduate education, 96, 104 undergraduates, 50, 65 unit cost, 218 United Kingdom, 1, 10, 23, 29, 30, 103, 104, 105, 106, 107, 188, 211, 244, 245, 250, 262, 263 United Nations, 105 United States, ix, 99, 100, 104, 140, 142, 235, 250, 262, 264 universities, 5, 7, 9, 20, 23, 166 university education, 163 urban areas, 48, 153 urban centers, 48 USDA, 244, 246
V vaccination, 175 Valencia, 205 validation, 116 validity, xi, 181, 191, 193, 197, 234, 260 values, viii, 21, 24, 32, 42, 47, 57, 77, 78, 91, 103, 104, 121, 128, 143, 172, 176, 177, 203, 204, 218, 223, 224, 225, 228, 229, 253, 254 variable, ix, 115, 117, 145, 237, 239, 252, 254, 255, 258 variables, x, 21, 22, 68, 118, 127, 135, 136, 145, 150, 151, 164, 220, 236, 237, 238, 239, 246, 247, 255, 258 variance, 25, 147 variation, 23, 34, 253 vector, 150, 236, 264 vein, 50 venue, xii, 75, 236 vertical integration, 119 Victoria, 105, 106, 111 videotape, 87, 90 visible, 69
Index vision, 60, 75, 77, 78, 184 visualization, 63, 65 vocational, 163 vocational education, 163 voice, 97 voids, 106 Volkswagen, 195, 198, 205, 210, 212 vulnerable people, 99 Vygotsky, 58, 71, 72, 88, 98
W wage rate, 141, 142, 143 wages, 140, 164, 264 wagyu beef, 242, 244 waiting times, 219, 230 Wales, 110 walking, 21 warrants, 58 water, 55, 171, 177, 184, 185 water quality, 55 wealth, 171 web, 53, 57, 67, 70, 211 Web 2.0, 65 welfare, 105, 143, 172, 175, 177, 181, 245, 264 welfare loss, 181 welfare system, 177 well-being, viii, ix, 73, 99, 104, 179 western countries, 142 wheat, 261 wholesale, xi, 236, 245, 246, 247, 248, 249, 250, 261
283
winning, 78 wireless, 55, 67 wires, 173 wives, 137, 140, 141, 161, 163, 164, 166 women, 50, 136, 137, 138, 139, 140, 141, 145, 146, 151, 153, 165, 265 work environment, 68, 74, 76, 80, 81 workers, 1, 6, 7, 8, 14, 42, 43, 51, 52, 65, 66, 67, 100, 104, 107, 109, 140, 168 workforce, ix, 4, 5, 6, 7, 8, 9, 10, 13, 43, 50, 54, 73, 99, 100, 107 working hours, 136, 138, 139, 140, 141, 142, 146, 147, 149, 150, 151, 153, 157, 158, 160, 161 workload, 61, 92, 215, 226 workplace, viii, 73, 75, 81, 101 workspace, viii, 47, 57 World Bank, 185, 188, 189 worry, 10 writing, 56, 62
Y YAC, 173, 174 yield, 11, 183, 194, 246 young women, 50
Z zero sum game, 7