Valuation in Life Sciences: A Practical Guide, 3rd Edition

Valuation in Life Sciences Third edition Boris Bogdan · Ralph Villiger Valuation in Life Sciences A Practical Guide ...

Author: Boris Bogdan | Ralph Villiger

309 downloads 2041 Views 5MB Size Report

This content was uploaded by our users and we assume good faith they have the permission to share this book. If you own the copyright to this book and it is wrongfully on our website, we offer a simple DMCA procedure to remove your content from our site. Start by pressing the button below!

Report copyright / DMCA form

DOWNLOAD PDF

Valuation in Life Sciences Third edition

Boris Bogdan · Ralph Villiger

Valuation in Life Sciences A Practical Guide Third edition

123

Dr. Boris Bogdan Avance, Basel GmbH B¨aumleingasse 2 4051 Basel Switzerland [email protected]

Ralph Villiger Avance, Basel GmbH B¨aumleingasse 2 4051 Basel Switzerland [email protected]

ISBN 978-3-642-10819-8 e-ISBN 978-3-642-10820-4 DOI 10.1007/978-3-642-10820-4 Springer Heidelberg Dordrecht London New York Library of Congress Control Number: 2010924856 c Springer-Verlag Berlin Heidelberg 2007, 2008, 2010 This work is subject to copyright. All rights are reserved, whether the whole or part of the material is concerned, specifically the rights of translation, reprinting, reuse of illustrations, recitation, broadcasting, reproduction on microfilm or in any other way, and storage in data banks. Duplication of this publication or parts thereof is permitted only under the provisions of the German Copyright Law of September 9, 1965, in its current version, and permission for use must always be obtained from Springer. Violations are liable to prosecution under the German Copyright Law. The use of general descriptive names, registered names, trademarks, etc. in this publication does not imply, even in the absence of a specific statement, that such names are exempt from the relevant protective laws and regulations and therefore free for general use. Cover design: WMXDesign GmbH, Heidelberg Printed on acid-free paper Springer is part of Springer Science+Business Media (www.springer.com)

Foreword to the First Edition

Capital, whether it originates from venture, public equity, private, or government sources, continues to be biotech’s source of sustenance. In fact, access to financing can make or break a company, regardless of whether it has Nobel Prize winning science or a top-flight business school management team. In order to attract capital there has to be a “value proposition” that is sufficiently engaging and well constructed that will not only capture the imagination of investors but also make them want to ultimately invest. This book recognizes that there is no consent on how to apply valuation methodologies in life sciences. One of the complicating factors is that, compared to other industries, valuation of biotech innovation is much more demanding. The long 10-15-year development and clinical trials process still represents the main risks faced by any biotech company. Added to that is the fact that getting a drug across the regulatory goal line and receiving Food and Drug Administration approval (or other regulatory agency approval in the United States or elsewhere in the world) for marketing is no longer good enough. Since the biotechnology industry is not isolated from the major influences affecting the overall healthcare industry, the reality of de facto healthcare cost controls will mean a significant reduction in “peak sales” for individual drugs. Science will no longer be the deciding factor; payor agents will play a far more activist role via reimbursement criteria, co-payment arrangements and similar rationing technologies. Reimbursement planning starts with clinical planning. At the dawn of the biotechnology industry, early investors had very few metrics with which to base and compare their investment decisions and valuations of such fledgling companies as Amgen and Genentech were “negotiated events”. Investors bought into the hopes, dreams and expectations that the scientists and entrepreneurs were selling. Fast forward to today and the industry is still extolling those same hopes and dreams that its innovation can make a difference … meet an unmet medical need with a breakthrough therapeutic; create better crops; improve the environment, or protect the population from the threat of bioterrorism.

VI

Foreword to the First Edition

By and large, the investment community and capital markets have liked what they have heard. For the comparatively short 30-plus year history, this industry has attracted almost $400 billion. Capital continues to flow into biotechnology companies at incredible rates and this year alone we predict that a new milestone will be reached for the industry with $40 billion being invested in the United States alone. We have come a long way in a very short space of time. Those of us who fund innovative, but relatively immature enterprises not only have to recognize the best-in-class companies but also understand how the future will unfold in the light of constant technology change. The maturing industry that is now melding its scientific entrepreneurship with the needs and culture of its larger commercial reality is an ever-increasing challenge. In the early days, the dream of an emerging biotechnology company was to become a fully integrated pharmaceutical company and lever its intellectual property portfolio into a stream of blockbuster drugs. To demonstrate how dramatically this business model and landscape can change – the focus on a one-size fits-all blockbuster drug is a strategy that may have run its course. The best selling drug of 2020 is somewhere between the bench and pre-clinic trials … but who knows if big pharma, as we know it, will develop it. The successful companies of the future will be those that marry both molecular diagnostics with targeted drugs and deliver effective personalized therapies. We are witnessing an unprecedented rate of evolution and transformation in the life sciences and healthcare. We are in an era of personalized medicine where the promise is compelling for the future of health care … personalized medicine offers earlier and more precise diagnoses, treatments tailored to the individual, reduction of side affects and adverse reactions to drugs, breakthroughs in treatment, and ultimately prevention of major diseases such as cancer, diabetes and Alzheimer’s. On the other side of the ledger the move to a more personalized, predictive and preventive medicine (the three “P’s”) world that will revolutionalize the health care system, as we know it today, is challenging pharmaceutical and biotechnology companies alike to adapt. Up until recently, their focus has been on the discovery of blockbuster drugs. Now, with the convergence of IT and genomics, smarter drug delivery and “labs” on a chip are moving us down an inevitable path towards targeted personalized medicines – away from “blockbusterology” – and spotting “early warning” signs of impending health problems. Science is leading into a promising realm of new technologies such as theranostics, responder/non-responder effect of biomarkers, molecular diagnostics and the re-emergence of genomics and proteomics companies in an era of tougher regulations governing the safety of drugs with payors looking to decrease the costs of health care.

Foreword to the First Edition

VII

Creating a successful and profitable biotechnology enterprise is therefore just as complicated as the genetic and cellular innovations that the company is trying to commercialize. Conservative estimates tell us that a biotech company will need close to $2 billion (according to recent estimates) and at least 12-15 years before any patient receives its novel drug therapy and returns on this massive investment can be generated. Along the way, the company will face many challenges: For its product – numerous and stringent regulatory hurdles; for the corporation – fickle financial markets; product competition; rapidly changing technologies; and difficulties in recruiting and retaining skilled staff. Taking all these factors into account small wonder that the chance of bringing a drug to market is about 1 in 1000! Thus, as investors, we are taking a chance, it’s true, but we view this as a very necessary investment in the future of mankind – and that gives us courage. Also, there is an opportunity for moving the science forward into commercialization, and for creating value. This, of course, requires a tremendous amount of due diligence and dogged research, but those of us who have stayed with this industry for decades know that the effort is well worth it. While, as investors we have the courage to be part of a better future our investment decisions have to be tempered with reality and on arriving at appropriately calculated valuations of the company’s worth … after all investment comes at a price to the entrepreneur. As a long-time investor in life sciences and someone who has been involved in the process of “valuation” for many years this is where the rubber meets the road in the deal making process. Tension on both sides of the negotiation table might be ameliorated if there was one agreed on set of rules about what a piece of technology or biotech companies are actually worth. The claim is made that the process of valuation in biotechnology is one part science and one part art. I would argue that there is a third essential ingredient, and perhaps the most important, and that is passion. Without it the biotechnology industry, perhaps would never have come into being. Any assistance that can help us figure out, in a more rational way, the complexity of the “value proposition” will help both communities at the bargaining table reach more realistic valuations is welcomed. A company’s value lies in its potential to generate a stream of profits in the future. Profits can be generated from sales of drugs, services but also from upfront, milestone and royalty payments. All valuation exercises are based on determining a company’s future, which requires many assumptions such as: the state of the market targeted and the potential share obtainable; the

VIII

Foreword to the First Edition

company’s intellectual property and its freedom to operate; third, the ability of management to deliver on the business plan; and fourthly, the absolute size of the financing and what’s needed to get to the next value inflection point. There is no “magic bullet” when it comes to valuation and it has remained a “black box” process for even the most seasoned investment professionals. The authors have, therefore, done an admirable job in setting out a roadmap for valuation in life sciences. They take us from this often times subjective operation to the more rational scientific process. The detailed worked examples are particular helpful in reducing the process into a more practical operation. This text will become a useful addition to the reference shelves of both the entrepreneur and investment professional alike and must reading for anyone contemplating raising investment capital. G. Steven Burrill, CEO, Burrill & Company San Francisco

November 2006

Table of Contents

Introduction................................................................................................ 1 The Life Sciences Industry ..................................................................... 2 Academia............................................................................................ 2 Biotech................................................................................................ 3 Pharma................................................................................................ 4 Valuation ................................................................................................ 5 The Role of Valuation ........................................................................ 5 Current Problems in Valuation ........................................................... 6 Basics of Valuation................................................................................... 11 Introduction .......................................................................................... 11 Fundamentals........................................................................................ 11 Cash Flows ....................................................................................... 11 Discounting....................................................................................... 13 Valuation Methods ............................................................................... 29 Discounted Cash Flows Valuation ................................................... 29 Decision Tree.................................................................................... 32 Real Options ..................................................................................... 36 Difference Between Real Options and Financial Options ................ 59 Valuation of Financial Options......................................................... 60 Valuation in Life Sciences ....................................................................... 67 Fundamentals........................................................................................ 67 Drug Development ........................................................................... 67 Medical Device Development and Approval ................................... 70 Cost................................................................................................... 72 Success Rates.................................................................................... 74 Time.................................................................................................. 76 Peak Sales......................................................................................... 77

X

Table of Contents

Discounting....................................................................................... 96 Volatility......................................................................................... 100 Project Valuation ................................................................................ 103 Project Valuation: DCF .................................................................. 105 Project Valuation: Decision Tree ................................................... 113 Project Valuation: Real Options ..................................................... 121 Simulations of Projects................................................................... 134 Sensitivity Analysis and Sensitivity-Simulations........................... 143 License Contract Valuation ................................................................ 146 Fundamentals of Licensing............................................................. 146 Valuation of License Contracts: DCF ............................................ 152 Valuation of License Contracts: Real Options ............................... 160 Advanced Topics in Project and License Contract Valuation ............ 169 Early-Stage Licensing with Possible Sublicensing......................... 169 Simulation of License Contracts..................................................... 194 Discovery and Preclinical License Deals: What’s Realistic? ......... 195 How to Read a Deal........................................................................ 197 The Virtual Company Model.......................................................... 201 Negotiating License Deals.............................................................. 208 Valuing License Contracts with Multiple Indications .................... 218 Valuing Projects with Multiple Markets ........................................ 232 Technology Valuation and Feed Rate................................................. 233 IP Valuation........................................................................................ 235 Pipeline Valuation .............................................................................. 243 Portfolio Management ........................................................................ 245 Introduction .................................................................................... 245 Qualitative Approaches .................................................................. 247 Quantitative Approaches ................................................................ 248 Optimal Portfolio Structure ............................................................ 250 Company and Stock Valuation ........................................................... 255 Introduction to Company Valuation ............................................... 255 Theory ............................................................................................ 258 Interpretation of Valuation Results................................................. 273 Simulation on Company Level ....................................................... 275 Case Company Valuation ............................................................... 275 Stock Valuation .............................................................................. 278 Advanced Company Valuation....................................................... 284 Mergers and Acquisitions............................................................... 290 Other, Less Serious Approaches..................................................... 293

Table of Contents

XI

Exercises ................................................................................................. 305 Introduction ........................................................................................ 305 Exercises............................................................................................. 305 Solutions ............................................................................................. 311 Case Studies............................................................................................ 331 Introduction ........................................................................................ 331 Case Study 1 ....................................................................................... 331 Case Study 2 ....................................................................................... 338 Case Study 3 ....................................................................................... 346 Case Study 4 ....................................................................................... 354 References............................................................................................... 363 Index........................................................................................................ 365

Abbreviations

BLA CAPM CNS CVS DCF DF EMEA ESOP EUT FDA GIT IND IP IPR IRR MCPM M&A NCE NDA NME NPV PDE PMA rNPV ROV R&D VBA WACC

Biologics license application Capital asset pricing model Central nervous system Cardiovascular system Discounted cash flows Discount factor European medicines control agency Employee stock options plans Expected utility theory Food and drug administration Gastrointestinal tract Investigational new drug Intellectual property Intellectual property rights Internal rate of return Market-derived capital pricing model Mergers and acquisitions New chemical entity New drug application New molecular entity Net present value Partial differential equations Pre-market approval Risk adjusted net present value Real options valuation Research and development Visual basic for applications Weighted average cost of capital

Introduction

Valuation is a hot topic among life sciences professionals. There is no clear understanding on how to use the different valuation approaches and how to determine input parameters. Some do not value at all, arguing that it is not possible to get realistic and objective numbers out of it. Some claim it to be an art. In the following chapters we will provide the user with a concise valuation manual, providing transparency and practical insight for all dealing with valuation in life sciences: project and portfolio managers, licensing executives, business developers, technology transfer managers, entrepreneurs, investors, and analysts. The purpose of the book is to explain how to apply discounted cash flow and real options valuation to life sciences projects, i.e. to license contracts, patents, and firms. We explain the fundamentals and the pitfalls with case studies so that the reader is capable of performing the valuations on his own and repeat the theory in the exercises and case studies. The book is structured in five parts: In the first part, the introduction, we discuss the role of the players in the life sciences industry and their particular interests. We describe why valuation is important to them, where they need it, and the current problems to it. The second part deals with the input parameters required for valuation in life sciences, i.e. success rates, costs, peak sales, and timelines. The third part outlines the principles of valuation. The reader gets an overview of the essential tools for valuation. The theory outlined lays the foundation of discounted cash flows, decision trees, and real options valuation, the two most common methods to quantify life sciences projects. The fourth part is the core of the book and guides the reader through theoretical aspects and practical applications of valuation in life sciences. We discuss the differences and the usability of DCF, decision trees, and ROV in valuation of projects, license contracts, technologies, IP, and firms and calculate the case studies in a way that the user can adapt and reproduce them in his own spreadsheets.

B. Bogdan, R. Villiger, Valuation in Life Sciences, 3rd ed., DOI 10.1007/978-3-642-10820-4_1, © Springer-Verlag Berlin Heidelberg 2010

2

Introduction

The last part of the book contains Excel exercises the reader can use as a template for future valuations. The case studies will particularly be useful for professors and students and for readers who want to make a test-run before applying their skills to real world projects. By the end of the book, the reader should have the knowledge to confidently apply valuation, strengthening his negotiation power and ability to take decisions. We have outlined the most important methods used in valuation of life science projects, rNPV, decision trees, and real options. The fact that real options valuation fills more pages in the book does not mean that we think it is more important than the other methods, but the complexity make detailed explanations necessary. We have seen in our mandates that the real options approach has important advantages, especially as a framework stimulating the strategic thinking of management. Nevertheless, the approach has not yet been accepted by all parties in the industry, probably also due to a lack of consensus in the literature. Nevertheless, most of the important players observe closely the development of real options analysis because they acknowledge that it addresses an important topic, namely uncertainty.

The Life Sciences Industry Academia The majority of groundbreaking advances in science come from academic research. Since the universities are in constant lack of funding, a logical step is that they commercialise their research findings. There are two major ways in doing so: First, the scientist himself builds a start-up around his research, or second, the university licenses or sells the rights to the research findings to a third party. In both cases the university gets remunerated for the cession of IP rights, either with an equity stake in a start-up company, or with license payments over several years or even decades. Income from licensed IP, if successfully managed, can mean significant revenues to universities and research institutions. The key figures on licensing at the US universities, hospitals, and research institutions are listed below (source: Association of University Technology Managers (AUTM), 2005 & 2008): • In 1992 1229 licenses were executed, in 2004 4783 and finally in 2007 5109. • Since 1980, 4543 companies were spun out of US universities, two third of them still active.

The Life Sciences Industry

3

• 555 start-up companies launched in 2007. • The net licensing income increased from $218 million to $1,385 billion in 2004. The figures confirm the trend of generating revenues from out-licensing IP. Next to tuition fees, state funding, and private donations, license income has become an important funding pillar for academia. Biotech After the burst of the biotech bubble in 2000/2001, the sector has slowly been recovering before the new downturn in the financial crisis. The numbers below show the year 2007 before the financial downturn in 2008 (Ernst&Young, 2008): • Revenues increase by 7% to $80 billion • Industry net loss (public) decreased to $2.7 billion • Total potential value of deals announced during the 2007 – including mergers, acquisitions and strategic alliances – was close to US$ 60 billion. Products produced by the biotechnology sector (“biologics” = protein based therapeutics) are expected to show a long-term growth rate of 12 – 13% p.a. compared to the 5 – 8% p.a. growth of drug sales in general. The global biotech financing has increased from under $20 billion in 2002, to $35 billion in 2005, and to fell to $29.9 billion in 2007, with the main growth stemming from license revenues and collaborations. Biotech is an essential source of drugs for the pharmaceutical industry. Apart from the rise in partnering revenues, the global venture capital investments into the biotech sector hit a new record high in 2007 with $7.5 billion. Despite the maturing of the biotech sector, the majority of companies is still years away from launching their first product and generating profits and therefore in a constant need for cash. Licensing agreements offer access to immediate cash for biotech and to innovative products for pharma and are likely to grow as a source of revenues for both.

4

Introduction

Pharma New drug output, measured by new molecular entities approval has been on a steady decline since 1995, from 28 to 18 in 2006 (FDA). 17 years ago, the first drug achieved the US$ 1 billion sales mark (definition of a blockbuster drug); today there are over 60. Companies depend on the sales of an individual branded drug as we have seen with Lilly’s loss of Prozac exclusivity in 2001, Bristol Myers-Squibb’s loss of Taxol and Glucophage in 2001/2002, and Schering-Plough’s loss of Claritin exclusivity in 2002. Analysts forecast that pharmaceutical revenues could fall US$ 40 billion short of investors’ expectations by 2007 due to patent expiries. More conservative estimates of drug sales lost to generics due to patent expiry range from US$ 7 billion in 2004 to US$ 17 billion in 2007 (Lehmann Brothers, 2003) Several factors have caused a decline in R&D output in the pharmaceutical industry (Ma et al. 2004). First, drug discovery has relied much more on new genomic screening strategies in the recent past. A myriad of new targets have been found this way, but these targets are not clinically validated and thus have a much higher failure rate than existing targets. Zemmel and Ma estimate the success rate of validated targets to be 50% higher. Second, regulatory obstacles are increasing. Recent scandals like Vioxx force the FDA to review their approval policy and will most likely lead to tougher regulatory requirements, finally increasing the cost of drug development. Third, the landscape in drug development is changing. Treatments are getting more efficacious and numerous. With more than 5,000 companies in drug development to date, competition is fiercer; it requires better drugs to build a successful pipeline today than 20 years ago. Companies who missed the train struggle to get back on track. The pharmaceutical sector has belonged to the most attractive industries over the last decade with yearly growth rates over 10%, but at present growth is slowing down. As the US, Europe, and Japan fight with rising health care costs and an increasing debt burden, the pharmaceutical industry experiences much more price pressure and will hardly be able to maintain current growth rates. On the other hand, new markets are opening up. India and China offer a population of more than 2 billion persons. Enforcement of patent laws will allow fast growth in this area. In the future, Biotech will play an essential role in fuelling the pipeline of big pharma. According to Ernst&Young, 66 biotech-pharma mergers took place in Europe in 2005 alone (Ernst&Young, 2005). The trend of inlicensing products from biotech and acquiring attractive companies will

Valuation

5

strengthen and might reach a point where, according to Novartis, 50% of the company’s products will derive from licensing. As already described earlier, the total value of deals in 2007 amounted to the record figure of $ 60 billion, with Europe accounting for $34 billion.

Valuation The Role of Valuation Today, most academic institutions have an office for technology transfer, but only few of them possess the abilities to negotiate sustainable license income. The standard way to value a project relies on experience and gut feeling; often the first best offer is accepted. Later, when the project proves to be better than anticipated, the university tries to renegotiate deal terms. With simple and small deals this is negligible, but once a university starts to handle large contracts, small mistakes in the legal and financial structure of the deal mean a difference of millions of dollars to the university. Research institutions with a long-standing licensing experience often have capable people with sound knowledge on valuation and deal structuring in their technology transfer office. These institutions succeed in generating a respectable share of their research funding through licensing; other institutions of similar size, but with inexperienced personnel, do not make hardly any money with their inventions. But even experienced licensing staff struggles to attribute the right value to a complicated license contract, as the contracts often include options and sublicensing clauses. The valuation is demanding and can only be performed using advanced methodologies, as we will show later. Nevertheless, a clear understanding of the basics of valuation helps to value most of the license contracts with a rational framework and generates a sound negotiation position for universities. Biotech is often academia’s counterparty and must therefore be able to value the same contracts. In the following situations, biotech has to deal with valuation: • Fund raising • Licensing • Portfolio management • M&A

6

Introduction

Biotech often faces the question of the company’s value. When seeking to raise funds, as every biotech does, it has to be sure how much stake to give away in return of cash. Venture capitalists are tough partners in negotiations; a sound idea of the own company’s value is the prerequisite for the founders to keep a respectable stake in the company. Licensing is the second situation where valuation plays a central role for biotech. Dealing with big pharma, biotech has to know the value of its compounds. While big pharma has experts for valuation and negotiation, in the biotech company the same person is in charge of fundraising and finding licensing partners. As these deals are vital for biotech, the company’s management should not underestimate the importance of valuation. Pharma will not start negotiating without a solid valuation, based on thorough due diligence, showing the leeway of the negotiations; Biotech should follow that example. The third application of valuation to biotech is portfolio management. For advanced biotech companies, knowing which projects generate value, and which do not, is key to success. In this regard, we should mention that those companies that only have one product in their pipeline do not need to value it. Abandoning the project would mean closing the company. M&A and IPOs are usually a smaller problem for biotech. Normally by that stage, companies have their advising investment bank performing the M&A and the IPO. But even an experienced investment bank does not guarantee the right valuation. The executive management of the biotech should at least understand and agree on how the bank generated the proposed value. Pharma, most often negotiating with biotech, encounters valuation in the same situations as biotech, sitting on the other side of the negotiation table. While biotech, being a young industry, often does not have the capabilities in valuation, pharma has its licensing, business development, or M&A team with all the necessary skills. Current Problems in Valuation At present, there is no consensus on how to apply valuation in life sciences. Due to inconsistent description and application of the different methods and input parameters, people feel insecure when applying valuation. Compared to other industries, valuation in life sciences is more demanding due to the inherent complexity and length of R&D. Main concerns are the choice of the right valuation method, the methodology itself, the input parameters and the interpretation of the results.

Valuation

7

Valuation methods and methodology. There are two major quantitative valuation approaches applied in the sector, DCF and real options. While DCF has been the gold standard for years, real options valuation is gaining grounds and is regarded as a possible alternative in life sciences. Both methods have their advantages and drawbacks. DCF, when applied to early stage projects, generally yields negative values; nevertheless the industry is profitable. Consequently, managers do not trust their valuations and disregard the recommendation retrieved from the valuations. Projects in early development are continued despite their negative DCF values. Often this translates into a general refusal of quantitative methods. As soon as it comes to licensing and M&A, the companies urgently need a valuation method that displays the correct value of early stage projects. Real options valuation on the other hand has been developed to overcome the shortfalls of DCF. It is still regarded as too complex and highly theoretic, compared to the easy to use DCF method. Today, there is no standard on how to apply the method to life sciences valuation. A multitude of different explanations and forms of real options do not facilitate its acceptance. This is partly attributable to the often repeated claim that real options should be treated like financial options. In fact, real options are just an expansion of DCF taking into account real flexibility of the management. The inappropriate use of the Black-Scholes formula, the discount rate, or the volatility are a root for disagreement on real options and hinder a broader use of the method. Too many eye-catching but apparently wrong case studies severely harmed real options valuation’s image. Properly understood and applied, the method represents the reality much better than DCF does, assuming that estimated input parameters for the valuation can change over time and that the management can react to this. We will discuss the advantages and pitfalls of both methods in the life sciences sector and propose when and how to use DCF or real options. Input parameters. Apart from the problem of using the right methodology, there is no consensus on how to define the essential input parameters in life sciences valuation. Mainly four parameters are problematic: • Discount rate • Costs • Success rate • Peak sales

8

Introduction

The discount rate is not only a problem in life sciences; it is also a problem in valuation in general. It is the most important input factor with extreme impact on the valuation results. Unfortunately, the discount rate is to a large part not observable and relies mainly on theoretic models. Past and present methods to assess the discount rate have failed to generate consistent numbers. The CAPM, which has served as a basis for the determination of the discount rate for almost five decades, does not deliver usable results for most life sciences companies. Moreover, it is of no use for early stage companies, not publicly listed. New methods like the market derived capital pricing model (MCPM) have evolved, but again do not yield applicable results in life sciences. We will outline alternative methods to define the discount rate. The second parameter posing difficulties are the costs. People not involved in everyday valuation will find it difficult to estimate the right costs of clinical trials and preclinical research, as well as of the post approval phase. There are useful public sources for industry data that are often not known. Problematic is furthermore the differentiation between costs for small companies like early stage biotech and costs for big pharmaceutical companies. A thorough discussion of the essential numbers for the life sciences sector will equip the user with a set of numbers for life sciences valuation. The success rates, like the discount rate, have a tremendous impact on the valuation. Success rates and discount rates deal with risk and how to account for it in the valuation. People struggle with quantification of risk. The success rates need to be carefully chosen and adapted to the different valuation methods and sectors. A special chapter will explain in detail the success rates and the implications on DCF and real options valuation. Peak sales, like the costs and the success rates, are a major value driver in the valuation, but hard to estimate. The assessment of sales potential for a product ten years from market is most often not more than a good guess. There needs to be a consistent way in how to define the sales figures, that unfortunately does not exist today. The reader will learn how to handle situations where peak sales are not predictable and practically apply the theory to the case studies. Interpretation of results. Because of the above-discussed pitfalls, valuation results are not considered reliable and are not comparable. Ideally, valuation should yield reproducible numbers with the different methods allowing the user to base decisions on the outcome. Today this is not the reality. The case studies will instruct on the significance of the results and the sensitivity of the valuations on manipulation of input parameters.

Valuation

9

The goal of the book is to serve as a roadmap for valuation in life science. The reader will know when to use the different valuation methods, how to choose the right input parameters, and most importantly, the reader will understand how to use the valuation results. The reader can use the book to learn the fundamentals without prior knowledge in finance, valuation, or life science. Valuation professionals can focus on the theoretical and practical part to deepen their existing expertise and to practice. The book includes advanced chapters for professionals seeking advanced modeling concepts. These chapters should fuel new ideas in the industry.

Basics of Valuation

Introduction All companies deal with valuation from time to time. Capital budgeting, company and asset valuation, or value based management rely on valuation. Two approaches are the foundation of valuation, discounted cash flow valuation and relative valuation. The first one is a bottom-up approach where the present value of an asset’s future cash flows is calculated, the second determines the value of an asset by comparing it to similar other assets. While relative valuation is well applicable by common sense, DCF needs considerable understanding of the relevant input parameters. As DCF is a vital approach to valuation in life sciences and the basis of decision tree analysis and real options valuation, it is worthwhile to discuss in detail how the method is properly applied. We discuss in the following chapters the reasoning behind DCF and how to define the input parameters to value an asset. We also discuss the current problems to this valuation approach, such as the problem of risk and uncertainty, and some methods that try to overcome these problems.

Fundamentals Cash Flows The cash flow is, as its name describes, money that flows in or out of a company. Cash flows can be classified into: • Cash flows from operations: Cash flows from day-to-day, incomeproducing activities. • Cash flows from investing activities: Net cash flow from investing activities, defined as divestments minus investments. • Cash flows from financing activities: Net cash flow from financing activities. Issue of new debt and equity minus repayment of debt, minus equity return, minus payment of dividends. B. Bogdan, R. Villiger, Valuation in Life Sciences, 3rd ed., DOI 10.1007/978-3-642-10820-4_2, © Springer-Verlag Berlin Heidelberg 2010

12

Basics of Valuation

• Change in liquidity: Liquidity at the end of a period minus the liquidity at the beginning the same period. The cash flows of a company are presented either in the statement of cash flows or in the accounting statement of cash flows. The latter is useful to see the change in accounting cash and differs from the statement of cash flows mainly in the interest expense. Below is Virtual Corp.’s statement of cash flows for one accounting period: Table 2.1. Virtual corporation statement of cash flows Cash Flow of the Firm

(’000s)

Operating Cash Flow Capital Spending Additions to net working capital Total

$200 (100) (50) $50

Cash Flow to Investors of the Firm Debt $45 Equity $5 $50 Total

Extracting the valuation relevant cash flows from these statements is not always straightforward. For the purpose of the valuation theory outlined in this book, we will therefore not refer to accounting cash flows such as depreciation or to terms like additions to net working capital. We assume that all investments are immediately depreciated. Nevertheless, the final valuation result remains with both approaches the same. The essential input for any valuation are the cash flows. We have to identify and estimate all relevant cash flows in terms of: • Size • Time • Probability Cash flow size. A cash flow is either positive, i.e. a revenue, or negative and therefore an expense, with an absolute size. A negative cash flow is either noted with a negative sign or between brackets, i.e. $ 50 expense is $ –50 or $ (50).

Fundamentals

13

Time of cash flows. We need to define the time when a cash flow occurs as the time influences the value of money. The discount rate accounts for this. Probability of cash flows. Once we know the size and the time of a cash flow, we have to estimate the probability of it. A cash flow that is certain has more value than a cash flow that is uncertain. We will learn in the following chapters, that valuation is the process of defining the cash flows size, time, and probability, the risk adjusting and netting of all relevant cash flows and finally calculating the net present value by discounting. Once we have exactly identified and described all cash flows for the valuation, the major part of the work is done. Discounting In valuation we compare cash flows that: • occur at different points in time, • are not accurately predictable in their size, • occur with different likelihood. The first point alludes to the time value of money, the second to the uncertainty of the estimate, and the third to the risk that the cash flow occurs at all. Discounting takes full account of time value and uncertainty. Probabilities (or success rates in the life sciences context) cover the shortfall risk. Time value of money. The concept of time value bases on the fact that people prefer a dollar today to a dollar tomorrow. A dollar today has more value than dollar in the future. To keep its value, money must accrete. This value increment is called interest. Somebody investing a dollar in a project wants to get more than the dollar he invested after one year, because he could also have put his money on the bank and earn interest on it. Interests are generally compounded annually. An interest rate of 5% means that the investor of $ 1,000 earns after one year $ 50 of interest. After another year we earn 5% of the meanwhile $ 1,050, ending up with $ 1,102.50. This example shows that compounding of interests has an impact on the final amount. In the second year, the investor receives not only 5% of the initial amount, but also 5% on the interests he earned in the first year. We determine the risk free interest rate by taking the yields of treasury bonds of the US or the EURO zone. The risk free interest rates depend on

14

Basics of Valuation

the maturity of bonds and vary from country to country. Usually, interest rates rise with increasing maturity. In Japan, interest rates are at the moment low (0%-2%), while they are higher in Europe (1%-5%) and in the United States (3%-7%). The numbers depend on the state of the economy in these countries and are subject to fluctuations. For valuation purposes, we will take interest rates of bonds with long maturities, i.e. 10 to 30 years. Risk. Next to the time value, discounting reproduces the risk of the investment and the willingness of the investor to take risk. We estimate future costs and revenues to a certain degree, but the real cash flow can be higher or lower than our initial estimate. This uncertainty is typical for predictions. However, people tend to dislike under-performance or loss more than they do like over-performance or gain. They are willing to pay a fee in order to avoid any shortfalls. Insurances base their business on this asymmetric attitude called risk aversion. Discounting must now compensate not only for the loss of value over time, but also for the impending difference between the expected and the actual return. Consequently, uncertain investments should reward the investor at a higher rate than safe investments, as it is more likely that the actual return is closer to the expected return E(R).

Low Risk Investment

E(R)

High Risk Investment

E(R)

Fig. 2.1. Expected return of assets with different risk profile

The valuation can reflect this increased return expectancy by adding a spread on top of the rate that displays the above-described value loss over time. This spread depends on the uncertainty of the cash flow estimates and can range from 0% up to 20%. Typically, companies use a spread between 5% and 8%. Discrete and continuous compounding. Discounting is one of the most technical parts of valuation. It is worthwhile to spend some time to get the technicalities right. We differentiate between discrete compounding and continuous compounding. Discrete compounding pays the interests ac-

Fundamentals

15

crued over a period T–t at the end of the same period. With r being the annualized interest rate an amount of money ST becomes:

ST = S t (1 + rdis )

(T − t )

(2.1)

Continuous compounding assumes that interests are paid out and reinvested continuously, i.e. the investor earns at every moment already interests on the interests earned in the previous glimpse of time. Continuous compounding yields therefore a higher interest. After a time period t with interest rate r the amount S becomes, if continuously compounded:

ST = S t exp(rcon (T − t ))

(2.2)

Unlike interest calculations where we have to calculate an amount we will receive in the future, discounting is used to determine the present value of a future amount. Instead of moving forward in time, we now move backward in time. This leads to the following discrete and continuous discounting methods, r being the discount rate: S t = ST (1 + rdis )

− (T − t )

S t = ST exp(− rcon (T − t ))

(2.3) (2.4)

Usually interest and discount rates refer to the discrete method, using one year as reference period. The continuous rate rcon that corresponds to the discrete rate rdis can be found with the following relationship:

rcon = ln(1 + rdis )

(2.5)

The Cost of Capital The discount rate of a firm corresponds to the average rate at which the stakeholders want their capital to accrete. A company has two major classes of stakeholders, debtholders or bondholders, and shareholders. Both require compensation for the risk of their investment into the company. Bondholders typically receive the principal of the bond. Only in the case of default of the company, they can lose up to the full amount invested. The spread of the bond that comes on top of the risk free rate must offset this risk. The spread depends on the likelihood of the company defaulting. Rating agencies like Standard & Poor’s, Moody’s, or Fitch give their opinions on the trustworthiness of companies. According to these ratings, the return rates of the bonds are determined. These bond returns then correspond to the

16

Basics of Valuation

cost of debt, i.e. the minimum rate a company must achieve to satisfy the demands of their bondholders. The rate required by the shareholders is called cost of equity. Unfortunately, this parameter is not observable. Several theories have evolved to determine the cost of equity. After this section, we will present the capital asset pricing model and the market-derived capital pricing model as two representatives. Once we have determined cost of debt and cost of equity, we can compute the average cost of capital of a company. The debt part of the company must accrete by the cost of debt, the equity part by the cost of equity. An important property of debt is its tax-deductibility. This makes debt even cheaper capital. Taking account of the ratio between bond- and shareholders and the tax advantages of debt the cost of capital, or in this case the weight-adjusted cost of capital (WACC), becomes (with D being the market value of debt, E equity, and T the tax rate): WACC =

E D rE rD (1 − T ) + D+E D+E

(2.6)

The capital invested in a project must appreciate at least at the discount rate, otherwise the company cannot satisfy the expectations of its stakeholders. This is the reason why the discount rate is often called hurdle rate. In practice we rarely use the wacc, because usually we do not want to know the value of the company but rather the value of equity, i.e. the value represented by the shares. In this case we apply the cost of equity in the valuation. And if we estimate future cash flows based on today’s assumptions, which we always do, we must adjust the valuation for inflation. Estimating future cash flows we would have to consider inflation as well, i.e. in reality the cash flows would be higher than estimated. But instead of doing so for every cash flow, we can also perform this adjustment in the discount rate. For this sake we reduce the discount rate (more precisely: the cost of equity) by the inflation rate. So, if the yearly inflation is expected to be of 2%, we simply subtract 2% from the discount rate and calculate with this. The following example explains this method. Assume that an artist is going to sell a painting for US$ 10,000. But you are only going to sell it in three years, and the expected inflation rate amounts to 2%. So the actual selling price won’t be US$ 10,000 but rather US$ 10,612 (=10,000*(1+2%)3). The discount rate you apply is 13%. So the value of this painting to the artist is US$ 7,355 (=10,612*(1+13%)-3). This value was calculated by first adjusting US$ 10,000 for the inflation

Fundamentals

17

over three years and then discounting it over the same period. We have multiplied a cash flow by the fraction (1+2%)/(1+13%) or more general by the fraction of inflation adjustment (1+ρ) over discount (1+r), where ρ is the expected inflation rate and r is the discount rate. If ρ is much smaller than r then this becomes approximately: t

⎛1+ ρ ⎞ t −t −t ⎜ ⎟ = (1 + ρ ) (1 + r ) ≅ (1 + r − ρ ) 1 r + ⎝ ⎠

(2.7)

This means, that instead of adjusting for inflation and then discounting, we could directly discount at a rate reduced by the inflation. Of course, this is not 100% correct, but if ρ is relatively small the error is negligible. In our case the artist must discount the US$ 10,000 at a rate of 13%–2%=11%. This gives US$ 7,312. We see that it is a shortcut method, because we have a difference of US$ 43 to the exact solution. We can see that if applying continuous discounting this method becomes exact: e ρ c t e − rct = e ρ c t − rc t = e − (rc − ρ c )t

(2.8)

Terminal Value Sometimes we want to know the value of series of yearly cash flows from now to infinity. The bank of England for instance issued a bond that pays out each year a certain fixed coupon, let’s say £ 10. If we want to know the value of this bond we have to calculate the sum of all future discounted cash flows. This can be easily calculated with the following identity: ∞

1

∑ CF (1 + r )

i

=

i =1

CF r

(2.9)

Using a discount rate of 4% the value of the perpetual bond of the Bank of England therefore becomes £ 10/4%=£ 250. If we assume that the cash flow will happen each year, but will also become larger with a constant growth rate µ, then the formula becomes: ∞

1

∑ CF (1 + µ ) (1 + r ) i

i

i =1

≅

CF r−µ

(2.10)

Again, the same imprecision applies like for the inflation adjustment. But the formula above is the one commonly used for terminal value calculations.

18

Basics of Valuation

Capital Asset Pricing Model One way to determine the cost of equity is the capital asset pricing model. This model differentiates between diversifiable and non-diversifiable risks. The non-diversifiable risks are generally denoted as market risk, the diversifiable risks as asset specific risk. We assume that there exists a portfolio that is only exposed to market risk, all diversifiable risks cancel each other out. This portfolio is defined by the modern portfolio theory by Harry Markowitz (Markowitz 1952) and optimises the risk-return ratio (Sharpe ratio). The portfolio is called market portfolio. The return of an asset can then be described as follows:

rasset = β rmarket + (rasset − β rmarket )

(2.11)

ε

rmarket is the return of the market portfolio that is solely exposed to the nondiversifiable market risk. The factor β indicates by what extent the asset is exposed to the market risk. ε then represents the diversifiable part of the asset’s risk. Note that β is chosen such that ε is completely independent of rmarket, i.e. there is no correlation between ε and rmarket. Note that rmarket and ε are random variables and are only predictable to a certain degree. While an investor can reduce or even avoid his exposure to diversifiable risks by holding a large and well-diversified portfolio, he cannot reduce the non-diversifiable risks. He only must be rewarded for the nondiversifiable part of the asset’s risk. The average return the investor therefore claims, i.e. the cost of equity if the asset is a share, is the risk free rate plus a spread on top, proportional to the non-diversifiable risk he accepts to bear. The reward for taking the market risk is called market risk premium. Unfortunately, this measure is not observable and might even be perceived differently by each investor. However, people generally accord that the market risk premium corresponds to the excess of the market portfolio over the risk free rate, rˆmarket − r f , where rˆmarket denotes the average return of the market portfolio. rˆmarket − r f is exactly the reward the market pays for bearing the market risk. Consequently, the cost of equity an investor claims for an equity investment is: rE = r f + β (rˆmarket − r f )

(2.12)

market risk premium

It remains to define the parameters rf, β, and the market risk premium. The risk free rate should not pose major problems. For the market risk premium

Fundamentals

19

one can use the historic excess performance of the market portfolio compared to risk free investments. Unfortunately, the theoretic market portfolio is not known, therefore we don’t know rˆmarket either. A generally accepted bypass is to assume an index as the S&P 500 or FTSE100. rˆmarket then becomes the historical performance of the chosen index. Finally, we have to determine β. This measure is defined by:

β = Corr (rasset , rmarket )

σ asset Cov( rasset , rmarket ) = 2 σ market σ market

(2.13)

Hence, β equals the correlation between the returns of asset and market multiplied with the ratio of their volatilities. A correlation of zero means that the asset is completely independent of the market portfolio and has no market risk component, consequently the β is zero as well. A correlation of one on the other hand means that the asset fluctuates in exactly the same way as the market portfolio, a correlation of minus one means that the asset moves exactly opposite to the market. Drawbacks of CAPM. The CAPM method is the most widely used way to determine the cost of equity. However, it has major flaws. First, it relies to a great extent on the knowledge of the market portfolio. The market portfolio is the portfolio that maximises Sharpe’s risk-return ratio in the universe of all possible investment opportunities, including not only equity and debt markets, but also commodities and illiquid and not transparent markets such as real estate, private equity, art, or even wine. Nobody knows the market portfolio; therefore, we cannot use it as a reference for calculation purposes. Second, all measures in the CAPM are prospective. The β and the risk premium relate to the future. The β of an asset is a notoriously unstable parameter. Historic and future β do not have to be identical. Even historical β depend strongly on the chosen observation period. This instability renders the determination of the cost of capital arbitrary. Often it is possible to achieve any cost of capital simply by choosing an adequate observation period for β and the market risk premium. Third, β intentionally captures solely the non-diversifiable part of risk. The argument goes that the investor can diversify and is therefore not rewarded for diversifiable risks. However, very few investors would agree with this, and especially within the private equity universe, investors have large parts of their fortunes invested in one single company, effectively impeding a sufficient diversification. Furthermore, the company or asset specific risks have an impact on the business of the company. It is little

20

Basics of Valuation

consolation for a company and its investors facing financial distress that the risks causing this situation are diversifiable and therefore should not have been included in any valuation anyway. Diversifiable risks can have significant impact and should therefore be included in the valuation. This is particularly true for life sciences companies as we will see further down. Market-Derived Capital Pricing Model The market-derived capital pricing model (McNulty et al. 2002) tries to overcome the major flaws of CAPM. First, it does not rely on an unobservable risk factor like the market portfolio. Second, MCPM uses explicitly prospective and observable parameters, and avoids reliance on arbitrarily measured parameters like β. Third, and most importantly, MCPM accounts for all risks of a company, whether they are diversifiable or not, trying to reflect better the risk perception of managers and investors. The fundamental idea of MCPM is that shareholders should at least earn the cost of debt. We assume that shareholders insure this rate of return with a put option. The cost of equity then equals the cost of debt plus the annualised cost of the put option. As a consequence of this definition, the cost of equity depends on the time horizon. The put option has the forward price of the share at the cost of debt as strike price. rE (T ) = rD +

(

put 1, (1 + rD ) , T 1 1 − rD rD (1 + rD )T T

)

(2.14)

The put option uses 1 as the value of the underlying, the forward price (1+rD)T as the strike, and T as maturity. In many cases, MCPM yields intuitively better results than CAPM. MCPM however is not free of problems. First, the requirement that shareholders should be able to insure a return that is equivalent to the cost of debt lacks a stringent argumentation. Why should shareholders be able to avoid the downside while keeping all the upside? Second, the put option uses a prospective or implied volatility. This is only applicable if options are traded, i.e. when the company has already a respectable market capitalization. For smaller or even unlisted companies the prospective volatility is a very delicate measure. Unfortunately, the cost of equity is relatively sensitive to this parameter. Third, the cost of equity depends on the chosen time horizon. The authors calculate with a time horizon of five years, however, it is not clear how this is determined.

Fundamentals

21

While being an interesting and in many cases plausible alternative to CAPM, MCPM is difficult to apply to small companies. Furthermore, the diversifiable risk is included in the cost of debt and the volatility, leading to some double counting and consequently to a too high cost of equity. As a remedy, one could use the risk free rate instead of the cost of debt. Uncertainty Cash flows are linked to two different natures of uncertainty. The first uncertainty concerns the accuracy of the estimate. We do not know in advance, how large a cash flow will be. Many factors like competition or regulation influence the potential size of the cash flow. This uncertainty can have a negative impact, if for instance a competitor launches a better product. But the uncertainty has also an upside; imagine that the product sells better than anticipated, or costs are lower than expected. The second nature of uncertainty is the technical uncertainty. We do not know in advance, whether a cash flow will occur at all. The reason for uncertain cash flows is usually caused by failure risk, e.g. failure in clinical trials, and quantified by attrition rates. Attrition rates are purely value destructive. If a cash flow has only a probability p to happen, then you cannot count with the full cash flow, but only with its expectation, i.e. p times the cash flow. For valuation purposes, a cash flow must always be multiplied with its probability to happen. This requirement becomes clear when playing heads or tails. Assume you win one dollar with heads and you do not win anything with tails. Obviously, you will win half a dollar on average, which is exactly 50% of the dollar, i.e. the probability times the cash flow. However, if you play just once, you either win or you do not. But if playing already two times, you have 25% chance to win two dollars, 50% chance to win one dollar, and 25% to win nothing at all. You are already more likely to win the average sum. If playing ten times, you have a chance of more than 65% to win 4, 5, or 6 dollars, i.e. close to the average. The more often you play, the more likely you win exactly the average value of the game and the better your are placed to predict the average outcome of the game. Technology projects like in life sciences have the same property. The more projects a company has, the more likely it will end up with the number predicted by the success rates. The figure displays how many projects will succeed (in percents), assuming a one product company (Biotech), a company with a pipeline of ten projects (MidPharma), and a company with a pipeline of 100 projects (Pharma). Each project has a success rate of

22

Basics of Valuation

900 800 700 600 500 400 300 200 100 0 0%

10%

20%

30%

40%

50%

60%

70%

80%

90% 100%

Fig. 2.2. 1,000 simulations of a one-project pipeline (Biotech)

15% to reach market. The outcome for the biotech company is binary, either complete success or complete failure. According to the success rate, the success scenarios make up 15% of all scenarios. In the case of one product the ultimate value must lie somewhere between the value of the good and the bad scenario. However, after learning the result (success or failure) we notice that the project jumps considerably in value. Either it looses all of its value in case of failure, or the value increases significantly, because no uncertainty about the outcome has to be factored in anymore. For a mid-sized pharma company complete failure would mean that all ten projects fail. This probability is rather low ((100%–15%)10=19.6%). With a diversification of ten projects the risk of complete loss could already be lowered by an impressive 65%. Furthermore, the initially expected outcome of either one or two successful projects has a probability of 63%. We see that the diversification not only reduces the downside risk of complete failure, but also makes the outcome more predictable. The value difference between start and end of trials seems more manageable. The bulk part of scenarios will end up with one or two successes, and no scenario contains more than 5 successes. Looking at the simulations of the pharma pipeline of 100 projects we see that in 83% between 10 and 20 projects succeed. On the other hand, it seems almost impossible that more than 30 projects will make it to the market. The actual numbers of the simulations displayed vary between 5 and 29 successful projects. So the risk of complete failure has been totally removed due to a broader diversification, although theoretically it is still

Fundamentals

23

400 350 300 250 200 150 100 50 0 0%

10%

20%

30%

40%

50%

60%

70%

80%

90%

100%

Fig. 2.3. 1,000 simulations of a ten projects pipeline (MidPharma) 900 800 700 600 500 400 300 200 100 0 0%

10%

20%

30%

40%

50%

60%

70%

80%

90%

100%

Fig. 2.4. 1,000 simulations of a 100 projects pipeline (Pharma)

possible that no project reaches the market (=(100%-15%)100=0.00001%). But this would correspond to one simulation out of 10 mn simulations. No company is perfectly diversified. In the contrary, most companies depend on the success of a few projects. In the extreme case we have a one-project company that will either be a success or a complete failure. The value of the company then extremely depends on the success of the project. If the project succeeds, the uncertainty is resolved and its negative effect can be excluded, i.e. we do not have to multiply the cash flows by the success rate linked to that uncertainty. This will lead to a jump in value

24

Basics of Valuation

in the order of (1-p)/p. Assume V to be the value after the uncertainty has been resolved. Consequently the value just before should be around p times V. The value increase is therefore: (2.15) V − pV 1 − p = pV p Consequently, the value jump is higher the smaller the success rate. If the project turns out to be a failure, the value drops to zero. The value of a company or a project corresponds to the average outcome. After negative results related to one important project, people blame valuation not to have predicted the value drop that has occurred meanwhile. Such statements are based on a misconception of valuation. A sound valuation does not predict winning or losing, but it quantifies the odds. Playing or investing many times, one realises the odds, just like in the casino, where in the long run the bank always wins, i.e. the bank realises the odds. Taking another look at the simulations we see that a pharmaceutical company plays the drug development game many times and comes close to winning exactly the odds. Investing in a pharmaceutical company is like placing bets on hundreds of products and we are sure that the one or other project will make it to the market. The risks of such an investment is relatively small because of the inherent diversification. Value jumps due to success and failure of projects offset each other. In a biotech company there are usually no projects that can compensate for the failure of the lead project. The payoff profile of a biotech investment is therefore much more risky. An investment in a biotech company therefore is more like placing a bet on one project. Investors want to be rewarded for this uncertainty when investing in a biotech company. They ask for a higher discount rate compared to investing in a pharma company as we will see. Human beings feel uncomfortable with probabilities. People have a natural aversion against uncertainty and tend to even pay an amount just to avoid risk. The insurance industry makes its living thanks to this human trait. In order to demonstrate this let us play the following simulation. Simulation of Risk Aversion Mr. Human owns US$ 500,000. He has to play one of the three games listed below. In order to avoid playing the game, Mr. Human can pay a certain amount to Mr. Insurance beforehand and can keep the remainder of his fortune.

Fundamentals

25

1. Mr. Insurance tosses a coin (and it is supposed to be a fair coin with 50% probability for head and 50% probability for tail). If the result is head then Mr. Insurance completes Mr. Human’s fortune to one million dollars. If it is tail, then Mr. Human gives his fortune to Mr. Insurance. So basically, they play for Mr. Human’s fortune. 2. Mr. Insurance tosses 10 coins. For each head Mr. Human receives US$ 50,000. For the coins with tail Mr. Human gives US$ 50,000 to Mr. Insurance. 3. Mr. Insurance tosses 100 coins. For each head Mr. Human receives US$ 5,000. For the coins with tail Mr. Human gives US$ 5,000 to Mr. Insurance. How much is Mr. Human ready to pay to Mr. Insurance in order to avoid the game? In gamble 1 Mr. Human faces a potential loss of US$ 500,000, but also a potential gain of US$ 500,000. From the initial fortune of US$ 500,000 he could go to US$ 0 or to US$ 1,000,000. Since each outcome has a probability of 50% the game is fair; on average Mr. Human will still have a fortune of US$ 500,000. Mr. Human, however – if he lives up to his name – will be more worried about the loss of half a million dollars than he would be pleased by the gain of the same amount. Doubling the fortune does not satisfy Mr. Human to the same degree as losing it disappoints him. He is therefore ready to pay something in order to avoid putting his fortune at risk. The amount Mr. Human is ready to pay depends on his personal attitude towards risk – his risk preference. In gamble 2 Mr. Human only faces gradual losses or gains. Complete loss is rather unlikely with all coins showing tail (=50%10=0.1%). The outcome is not binary anymore. In 25% of all possible cases Mr. Human can even keep his US$ 500,000, because exactly 5 coins show head. He could even say that a loss of US$ 100,000 is still bearable, so he actually only cares about the 17% of all cases where he looses more than US$ 100,000. Clearly, Mr. Human thinks that gamble 2 is already less dangerous than gamble 1 and consequently he would spend less money to avoid the gamble. In gamble 3 finally, the chances to end up with less than US$ 400,000 are less than 2%, so very unlikely. While in gamble 1 Mr. Human was maybe ready to spend US$ 100,000 in order to keep at least parts of his fortune, in gamble 3 Mr. Human doesn’t even face a noticeable risk to loose as much. If Mr. Human still wants to insure his fortune and avoid playing, then obviously to a much lower price than in gamble 1 and even than in gamble 2.

26

Basics of Valuation 30%

Probability

25% 20% 15% 10% 5% 0% 0

0,1

0,2

0,3

0,4

0,5

0,6

0,7

0,8

0,9

1

US$ Mio

Fig. 2.5. Payoff profile of gamble 2 9% 8%

Probability

7% 6% 5% 4% 3% 2% 1% 0% 0

0,1

0,2

0,3

0,4

0,5

0,6

0,7

0,8

0,9

1

US$ mn

Fig. 2.6. Payoff profile for gamble 3

The three gambles offered by Mr. Insurance show that Mr. Human is sensitive to risk, and that the higher the risk, the less comfortable he feels, although the fair value remains the same (on average Mr. Human keeps his fortune in each gamble). As a consequence, investors require or expect a higher rate of return for an investment that bears a lot of risk compared to a relatively safe invest-

Fundamentals

27

ment. This translates directly into a higher discount rate. Imagine a business plan assumes to realise US$ 100 mn cash in 10 years; and for the sake of the example we do not question this assumption. If the investor will receive these US$ 100 mn with certainty after ten years, then he is ready to invest at nearly the risk free rate, let’s say 5%. This would then mean that the venture is worth 100*(1+5%)-10=US$ 61 mn, because if one lets US$ 61 mn accrete at 5% they become exactly US$ 100 mn after ten years. However, if the investor is not sure whether he will receive maybe only US$ 80 mn, but maybe also US$ 120 mn after ten years, i.e. there is some uncertainty linked to the investment, then the investor feels uncomfortable with the 5% discount rate. Finally he could put his money on a bank account and earn exactly 5% without running any risk. The investor therefore requires a higher rate of return for the money he puts into the venture. He therefore wants to start from a much lower initial value to have a higher return once achieving the US$ 100 mn. If we start at US$ 55 mn and expect them to become US$ 100 mn in ten years, then they have to accrete at a rate of 6.2%. Alternatively, if we discount the US$ 100 mn over ten years at a rate of 6.2% we get a present value of US$ 55 mn. The main conclusions out of this short example are first that the expected return and the discount rate correspond to each other and Second that more risk leads to a higher expected rate of return/higher discount rate and therefore to a lower initial value. We can relate the investment opportunity directly to the preceding simulation about risk aversion. Imagine the investor owns US$ 61 mn and has the choice between putting it on a bank account and earning the risk free rate of 5% (playing it safely), or he invests it in the described venture that could turn out to generate US$ 80 mn or US$ 120 mn. Obviously the investor would opt for the first option, because the value of the two options is the same, but he has no downside risk on the bank account. But what if he could keep some of his US$ 61 mn and participate with remainder in the risky venture? The few millions that the investor can keep would correspond to the premium he would pay in order to avoid the investment. Or in other words, the amount of money that makes him indifferent between the two investments. If this amount equals US$ 6 mn, this means that he thinks investing US$ 55 mn in the risky venture corresponds to putting US$ 61 mn to the bank account. Or: if he can avoid the risky investment, he would only do it up to a premium of US$ 6 mn. But if the insurance is more expensive, then he prefers the risky investment. So, in the investor’s eyes, the additional risk is worth a 1.2% premium on the discount rate.

28

Basics of Valuation

As we have seen in the simulation diversification reduces the need of a risk premium in the rate of return. So, if an investor is diversified, then he basically does not have to claim a high rate of return. This is exactly the idea of CAPM. In CAPM we assume that all investors are well diversified. However, as we will see later, this assumption is not reflected in reality. First, most investors are far from being well diversified. And diversification is also a strange animal, because in periods of financial distress, when diversification is needed most dearly, all values tend to fall together, i.e. a previously assumed well-diversified portfolio suddenly behaves as if it were one risky asset. Second, diversification is a value added by the investor to his portfolio and the investor prefers harvesting the fruits of his work – the diversification – on his own, rather than letting others benefit from it. Because if he would apply a lower discount rate to his investments because of his diversification, the co-investors in the same venture profit from his relatively low discount rate by getting less diluted. As we will see in the discussion about discount rates in the life science industry we opt for a theory that takes account of the risk profile of the company rather than the investor’s portfolio. Why should the investor when investing in biotech give away his premium he earned by building up a diversified portfolio? The same applies to big pharma when in-licensing a project. The hurdle rate, i.e. the discount rate, is usually higher for in-licensing project candidates than for in-house projects. If we agree that mainly the risk profile of the company matters for the discount rate we come to the conclusion, that a start-up company certainly 14% 12% 10% 8% 6% 4% 2% 0% 0

10

20

30

40

50

60

70

80

Fig. 2.7. Distribution of success with 100 projects

90

100

Valuation Methods

29

10% 9% 8% 7% 6% 5% 4% 3% 2% 1% 0% 0

20

40

60

80

100

120

140

160

180

200

Fig. 2.8. Distribution of success with 200 projects

has to expect to be discounted at a much higher discount rate than a pharmaceutical company, similar to the above-displayed risk profiles of biotech, mid pharma and big pharma. Nevertheless, even for a pharmaceutical company it is impossible to be perfectly diversified, as show the two figures for companies with 100 and 200 projects. Although scientific uncertainty, as represented by success rates, is diversifiable, no company or investor will achieve a degree of diversification where the value is certainly realised. This would be perfect diversification. The 200 projects pipeline, the double of the 100 projects pipeline, still produces outcomes between 10 and 30 commercialised projects. It is not difficult to imagine that the value of 10 and 30 products differ substantially. Consequently, every investor wants to be rewarded for this risk despite of its diversifiable nature. CAPM does not account for this portion of risk, exactly because of its apparentlynot-to-be-rewarded diversifiability.

Valuation Methods Discounted Cash Flows Valuation

We have learned in the chapter on discounting that we can calculate the value of future cash flows back to today by discounting them. We then receive the present value of the cash flows. The discounted cash flow valua-

30

Basics of Valuation

tion is finally nothing else than netting the present value of all future cash flows. The result is the net present value (NPV) of the cash flows. The terms DCF and NPV are therefore equivalent. If we furthermore adjust the cash flows with the probability (risk) that they occur, we get the risk adjusted net present value (rNPV). In the following we will use the term DCF for the method to calculate the risk adjusted net present value. The rNPV is calculated with the following formula: rNPV = − I o +

T

t =1

I0 rCFt r T

= = = =

rCFt

∑ (1 + r )

(2.16) t

Investment into the project at time 0 (=CF0) Risk adjusted cash flow at time t Discount rate Endpoint of the project (if today is t = 0, T = duration of the project)

In order to illustrate how to use the DCF valuation, we now calculate the rNVP of the project Supersolution. The project is still in development for one year. The investment for the development is due today and amount to $ 50,000. After one year, we have a 50% chance that the project passes the development phase and will finally be launched. The launch, including marketing and production expenses, amounts to $ 500,000. After the launch, the project generates the following revenues: Year 2 $ 100,000, year 3 $ 200,000, year 4 $ 300,000, year 5 $ 200,000, and in the final year 6 $ 100,000. Once the project is on the market, the expenses are 10% of the revenues. The company developing the project applies a discount rate of 15% to value all its investments: Table 2.2. rNPV calculation of project Supersolution Year Expenses

0

1

Revenues Net CF Probability

2

4

5

6

$ 100,000 $ 200,000 $ 300,000 $ 200,000 $ 100,000 ($ 50,000) ($ 500,000) $ 90,000 $ 180,000 $ 270,000 $ 180,000 $ 90,000 100%

50%

50%

Risk adjusted CF ($ 50,000) ($ 250,000) $ 45,000 Discount

3

($ 50,000) ($ 500,000) ($ 10,000) ($ 20,000) ($ 30,000) ($ 20,000) ($ 10,000)

100%

87%

76%

rpCF

($ 50,000) ($ 217,391) $ 34,026

rNPV

($ 32,801)

50%

50%

50%

$ 90,000 $ 135,000 $ 90,000

50% $ 45,000

66%

57%

50%

43%

$ 59,176

$ 77,187

$ 44,746

$ 19,455

Valuation Methods

31

To calculate the risk adjusted net present value of the project we: 1. List the expenses (negative cash flows) for each year. 2. List the revenues (positive cash flows) for each year. 3. Net the cash flows. 4. Risk adjust the discounted cash flows with the success rate. 5. Discount the cash flows by multiplying the risk adjusted net cash flows with the discount factor. The result is the risk adjusted present cash flows (rpCF). 6. Sum the risk adjusted present values of all cash flows. The result is the rNPV, the risk adjusted net present value of all cash flows. The discount factor is calculated in the following way: DFt =

(2.17)

1 (1 + r ) t

The risk adjusted net present value of the project is ($ 32,801). The company should therefore not invest in the project. In general, projects that yield a positive value should be continued, those with a negative value abandoned. We will discuss in the chapter on portfolio management how to decide which projects to halt and which not. Internal rate of return (IRR). The project above yields a negative value if we value it with a discount rate of 15%. The rate of 15% is the hurdle rate the company applies to decide if projects should be continued or abandoned. Projects that give a positive rNPV when discounted with 15% have a return that is higher than the demanded 15%. These projects pass the hurdle rate and are continued. We now calculate for our example the discount rate that yields exactly a project value of zero. The resulting discount rate is the internal rate of return of the project. rNPV = − I 0 +

T

rCFt

∑ (1 + IRR) t =0

t

=0

(2.18)

The IRR of our project is 10.13%. At this discount rate the project yields a value of zero. The figure below displays the relation between the discount rate and the rNPV:

Value

32

Basics of Valuation $ 120.000 $ 100.000 $ 80.000 $ 60.000 $ 40.000 $ 20.000 $ 0,0% ($ 20.000) ($ 40.000) ($ 60.000) ($ 80.000)

5,0%

10,0%

15,0%

20,0%

Discount Rate

Fig. 2.9. Influence of discount rate on project value

The internal rate of return is therefore the exact calculation of the return of a project. The money invested into the project above therefore accretes with 10.13%. As long as the IRR is below the discount rate, the project has a negative value. Decision Tree

Some projects have not a predefined development path or strategy. At some points in the life cycle of the project the original plan might be altered, depending on an event like own trial results or the entry of a new competitor. Some authors already consider a risk adjusted net present value calculation as a decision tree, because at the end of each trial we decide whether to continue or to abandon the project, as displayed in the figure.

Launch and Commercialisation 50% R&D 50%

Abandon

Fig. 2.10. Plan of project Supersolution as decision tree

Valuation Methods

33

In the rNPV calculation we did not have to bother too much about the abandoned branch of the decision tree, because no cash flows were linked to that branch. We could therefore focus on the successful branch, but adjusting all cash flows for launch and commercialisation with the probability of that branch. In reality, project plans can involve more decision points. In project Supersolution it is for instance possible, that the R&D phase can produce not only two possible results like success or failure, but three: Either the product does not work, or it does work, but not to the extent as originally expected, or it works even better than expected. In the second case the company is required to add another R&D phase of one year at US$ 50,000 to improve the product such that it then meets the original expectations. This corresponds to the following figure.

Better potential than expected 30% 67% R&D

Potential as expected

2nd R&D 30%

33%

Abandon

40% Abandon Fig. 2.11. More detailed decision tree for project Supersolution

We see that the overall success rate of the project remains 50% (30%+30%*67%=50%). When constructing the decision tree it is important that at each decision point the probabilities of all subsequent options add up to 100%. At the first decision point the project is abandoned with a probability of 40%, further developed with 30% and launched with 30%, adding up to 100%. At the second decision point we abandon the project with 33% and launch it with the remaining 67%. When calculating the value of the project we first have to calculate the value for each end-branch. These are the immediately abandoned scenario, the abandoned scenario after the 2nd R&D phase, the immediate launch,

34

Basics of Valuation

and the launch after the 2nd R&D phase. The abandoned scenarios have no cash flows anymore and their values are US$ 0. The immediate launch scenario assumes launch costs of US$ 500,000 and then sales that are 33% higher than originally expected. The launch scenario after the 2nd R&D phase then corresponds to the original launch scenario. Year Expenses

1

2

Revenues Net CF Discount pCF NPV

3

4

5

6

($ 500.000) ($ 13.300) ($ 26.700) ($ 40.000) ($ 26.700)

($ 13.300)

$ 133.000

$ 267.000

$ 400.000

$ 267.000

$ 133.000

($ 500.000)

$ 119.700

$ 240.300

$ 360.000

$ 240.300

$ 119.700

100%

87%

76%

66%

57%

50%

($ 500.000)

$ 104.087

$ 181.701

$ 236.706

$ 137.392

$ 59.512

$ 219.398

Discount rate

15%

Fig. 2.12. NPV of immediate launch scenario Year Expenses

2

3

Revenues Net CF Discount pCF NPV Discount rate

4

5

6

7

($ 500.000) ($ 10.000) ($ 20.000) ($ 30.000) ($ 20.000)

($ 10.000)

$ 100.000

$ 200.000

$ 300.000

$ 200.000

$ 100.000

($ 500.000)

$ 90.000

$ 180.000

$ 270.000

$ 180.000

$ 90.000

100%

87%

76%

66%

57%

50%

($ 500.000)

$ 78.261

$ 136.106

$ 177.529

$ 102.916

$ 44.746

$ 39.558 15%

Fig. 2.13. NPV of launch scenario after

With this we get the provisional values at the end branches of the decision tree as displayed in the figure. From here we must first calculate the middle option, i.e. when a second R&D phase is necessary. The outcome of this second R&D phase is either US$ 39,558 with a probability of 67% or US$ 0 with probability of 33%, on average US$ 26,504. Since this is only the average value after the second R&D phase we have to discount this value for one year and subtract the R&D costs from it to get the value of the middle branch. With this we have all necessary values for the decision point right after the first R&D phase.

Valuation Methods

US$ 219,398 30% 67%

US$ 39,558

33%

US$ 0

nd

2 R&D

R&D

30%

40% US$ 0 Fig. 2.14. Provisional values for project Supersolution Year

1

Expenses

2

($ 50.000)

Value

$ 39.558

Net CF

($ 50.000)

Probability rCF

$ 39.558

100%

67%

($ 50.000)

$ 26.504

Discount

100%

87%

pCF

($ 50.000)

$ 23.047

NPV

($ 26.953)

Fig. 2.15. Calculation of middle branch

US$ 219,398 30%

R&D

(US$ 26,953) 30%

40% US$ 0 Fig. 2.16. Continuation values after first R&D phase

35

36

Basics of Valuation

If the first R&D phase reveals that a second phase is necessary, then the company abandons the project, because it would have to invest US$ 50,000 in a project with negative rNPV at that time. We therefore assume that reaching the middle branch is equivalent to abandoning the project, but this time not because of technical reasons, but because of economic considerations. The final value is then calculated in a similar way like the value of the middle branch. The first R&D phase leads after one year to an average value of US$ 65,819 (=30%*US$ 219,398). This has to be discounted by for one year and reduced by the R&D costs, giving us a final value of US$ 7,234. If we would not include the option to abandon the project in the middle branch, so if we calculated with the actual value of (US$ 26,953), we would receive a final value of US$ 203. The assumption that the company abandons negative projects increased the project value therefore by US$ 7,031. We will see that real options valuation focuses particularly on this value added by management. Decision tree is a very useful method that encourages thinking through a project plan in more detail. It has however the problem that mostly the used probabilities are not known to the degree of detail that is required in the decision tree. It is common to know the success rate of projects, but we find rarely data quantifying the probability of success after the first or only after the second R&D phase. The decision tree method is therefore a prone to subjective assumptions. Real Options

General Aspects DCF convinces by its simplicity. Nevertheless, this simplicity is only achieved at the cost of some strong hypothesis. In the DCF method we assume that the market does not change, i.e. once we have fixed our estimate of the peak sales, we do not question this number anymore. One can argue that the estimated peak sales correspond to the average of what we can expect. Whatever happens, there is about the same probability that the actual outcome lies above or below the estimated number. On average, the sales equal what we have predicted. So why not just calculate with the estimated peak sales?

Valuation Methods

37

This reasoning bears two dangers. First by assuming one scenario, i.e. the average, we run the risk to take this number for granted. This can lead to over-reliance on this scenario that is still only a vague estimate. Second, the reduction of the future to just one scenario avoids considering alternatives in case the estimate turns out to be wrong or must be adjusted due to new information. This would correspond to a static management unable to react to a changed marketplace. Of course, this does not mean that managers who use DCF just twiddle their thumbs. Even with DCF, you can revaluate the project once some value drivers change. This is common practice and does not depend on the valuation method. However, DCF does not take into account this practice. DCF values the project as if the management takes a one-time go or no go decision. It would be reasonable to assume that the company alters its plan if one critical measure deteriorates or maybe improves way above expectations. The company has various options at hand: it can scale down the project, put it on hold, halt it completely, or in the good case accelerate or expand it. The management constantly decides what to do in order to increase value. In some scenarios, although not in the most probable ones, management takes actions that differ from the initial project plan. DCF does not consider this managerial flexibility. The real options approach takes into account that some decisions in a project’s or company’s life can or even must occur at a later stage depending on future market conditions. These can be fundamentally different from current market conditions; therefore, future decisions can only be anticipated to a certain degree. Consequently, management has to maintain flexibility to react to changes. This flexibility is used either to increase profits or to avoid losses. Either way it generates value. If a project includes such options where the management can influence the ongoing of the project, this should flow into the valuation process. Flexibility is always used to increase the value of a project or company, therefore a valuation method that captures this flexibility returns a superior value than one that does not. This value creation is not captured by DCF valuation. The novelty of real options valuation compared to DCF is that some future decisions about the ongoing of the project or company are conditioned on the respective market conditions and not vaguely anticipated. In real options theory the development of a project or company is depending on the changing market conditions, while in DCF the course of the project is predefined, no matter what happens.

38

Basics of Valuation

In order to correctly model the decisions, it is necessary to model the conditions, or parameters, they depend on. For this, real options theory makes use of the extensive research in the field of financial options that deal with the same issues. The market fluctuates and it is not sure from the beginning what payoff a holder of a derivative will incur at expiry. While in financial derivatives the different ways to execute are predefined, management has to recognise the different ways how to adapt the company’s business to the changed environment. Finally, it does not suffice to recognise that the company can still adapt its decisions at a later moment in time. We also need to quantify this flexibility. Some projects might have several probable alternatives and consequently, flexibility is important and represents a considerable part of its value. Other projects are unlikely to be redirected, mostly clearly profitable projects; their inherent flexibility is almost worthless. We must model the different ways to react to a changed market environment. These different reaction types define the different embedded real options in a project or a company. The existing literature provides six categories of real options based upon the types of managerial flexibility: • Option to defer • Option to expand or contract • Option to abandon or license • Option to switch • Option to stage investments • Option to grow Option to defer. An investment cannot be recuperated anymore once it is triggered. Hence it should only be triggered if the subsequent revenues are expected to be higher. The option to defer is finally a trade-off between risk and return. Investing immediately gives the company earlier, less discounted, revenues. Waiting might resolve some uncertainty about the market. For some projects it might be advisable to wait although they have a positive NPV due to their risk profile that might better be cleared prior to investing. An option to defer is specially valuable if the project outcome is highly uncertain and the investment irreversible. But waiting involves also the risk of loosing the first mover advantage. Therefore, in a competitive environment the option to defer might be worth less. Costs

Valuation Methods

39

to keep the option alive, e.g. renting costs, have a negative impact on its value as well. Option to expand or contract. Options to expand or contract incorporate the possibility to change the existing scale of the project according to the market movements. A company may build production facilities in excess of the expected demand. The company may incur additional revenues in times of high demand. On the other hand the company should engineer their plants in a way that they can be partially shut down in periods of low demand, reducing the fix costs. Testing a new market with a scout product and expanding if the test run proves successful is another example of this option type. Option to abandon or license. If a project fails to meet the expectations and is even deficitary, the company’s management should consider abandoning the project, possibly avoiding losses. The technology of the project might be licensed and the infrastructure sold. The company can recuperate a salvation value. The more reversible the investments, the higher the salvation value. Option to switch. A company can reconsider its production location and switch to a cheaper place. It is also imaginable that a production plant will be modified in order to fabricate another product. In these cases the company switches either the production costs or the sales revenues. Option to stage investments. Some projects require a staged investment. The firm resolves in every subsequent stage further uncertainties. Based on the learned facts, the project is revaluated and consequently continued or abandoned after each phase. These staged investments are modeled as compound options. After each period the company has the option to continue or not. Continuing corresponds to the acquisition of a new option. A staged investment process is common to R&D projects and to start-ups. Option to grow. Growth options are the most common and most widely cited real options. A company is supposed to have the possibility to expand its activities to other countries, to other clients, or to other products, if the initial activity proves to be successful. In a first stage the company resolves uncertainty about the market conditions, such as product acceptance, demand, and price politics. Once these doubts are removed, the company decides whether it is worth to expand – to grow – or not. Finally growth

40

Basics of Valuation

options are nothing else than staged investments. But normally growth options are considered to be just a one-stage option, while R&D real options are multi-stage options. As we will see later, real options valuation in life sciences mainly requires the option to abandon a project once it is not profitable anymore. Other options, e.g. options to grow or options to switch, are useful in certain instances but not for everyday valuation. We have already mentioned that it is the duty of the management to recognize the multiple options that the company holds in its hands and react to them. Different strategies offer different options. Real options valuation must then be used to compare the different strategies. As we see later, real options are not only a strategic tool for value aspects, but they also provide important information about risk and liquidity management. Nevertheless, there are also some dangers attached to the real options approach. So-called options and later opportunities should not be misused to justify any kind of investments. Real options still demand a thorough valuation process; they can be worth less than the initial investment to acquire the option. Furthermore, projects tend to create their proper dynamics and might be difficult to abandon. It is not sufficient to change the valuation method from DCF to real options; it is necessary to adapt the entire decision taking process to this new method. Real options assume that the management takes a decision in the future based on the best available information according to rational decision criteria. The valuation loses its justification if the management does not stick to this practice anymore. Eventually real options valuation is more a strategic than a valuation tool; it is a new way of management. We have mentioned above that the options – or the flexibility linked to them – represent a value that does not go into the DCF valuation. Usually real options valuation returns a higher value than DCF. This does not mean that the project is suddenly worth more if valued with real options. The project is still the same. And even when the management uses DCF, it revaluates a project periodically and adjusts its hypothesis. But the DCF valuation does not take this periodical revaluation into account; a revaluation that might trigger new investments or the abandonment of the project. DCF simply omits an aspect of the project life. In some cases this leads to a missed opportunity. We discuss this aspect in further detail in the following chapters. Finally, we can oppose DCF and real options valuation in the following manner:

Valuation Methods

41

Table 2.3. Comparison of DCF and real options Advantages

Disadvantages

DCF

Easy to implement and to understand Misses the value of flexibility and market uncertainty Standard in all sectors of the economy Not suitable for risk management Real Captures market uncertainty and the Relies on more hypothesis and Options management’s ability to react requires more data Suitable for risk management

Technical

Improves strategic thinking if properly understood

Resolution Methods Financial mathematics offers four different methods to value an option: 1. Formula (closed form solution) 2. Trees 3. Simulations 4. Finite differences All methods yield the same result, but the ways to get there are different. We discuss the methods with focus on their applicability to the valuation of real options. Some methods have been developed for special purposes and do not necessarily suit the requirements of real options valuation. First, the assumptions of real options are relatively modest compared to financial options. Second, the users of real options valuation most often do not have the required mathematical knowledge to apply the approach in practice. The complexity of the valuation method should therefore be as simple as possible, but not any simpler. No rocket science is required. Above all, the method should show the particularities that are considered in real options valuation and should not hide them in complicated algorithms. The decision takers need an efficient, explanatory and understandable tool, not a scientific high-end black box. The managers must be able to defend and explain their decisions without a long monologue about assumptions and technicalities. The valuations should support them in the communication of value and risk of a certain investment. We will now investigate the four valuation techniques with these thoughts in mind.

42

Basics of Valuation

Formulae. Formulae are well suited for the valuation of simple call and put options determined by five input parameters (underlying, strike, maturity, interest rate, and volatility). Unfortunately, it is virtually impossible to capture the complexity of real options in project valuation with a closed form formula. The complexity quickly surmounts the mathematical knowledge of most users and ends up being a black box. Finally, it is not helpful to use a formula, as the real option framework should not only serve to generate a number, but also as a managerial instrument to visualise possible scenarios and the impact of revaluations and decisions. We will further discuss the applicability of formulae to real options in the chapter about the differences between real options and financial options. Trees. Trees are a simplistic model of future market movements and are broadly used to value financial options. Binomial trees subdivide the time to maturity in small time steps and assume that in each time step the market, or in what we will see later, the peak sales, can go both up or down, each scenario with a certain probability. Binomial trees are best spanned in a way that after a step up and a step down the market is in the same state again. We call this type of tree a recombinant binomial tree. Usually the tree is spanned until the time of option expiry, the time steps can be chosen to satisfy the accuracy requirements, yearly, quarterly, monthly, or even daily. Normally, the smaller the time steps, the more accurate the final value. The option value is obtained by calculating the tree back from the final leaves to the root. At expiry, the option value is clear if the market state is known, and the tree precisely indicates the market state for every leave. After that, we place ourselves in a node one time step before expiry. From there, we have to deal with a Market size

Time Fig. 2.17. Recombinant binomial tree

Valuation Methods

43

simplified situation: We have just two possibilities, whereof we know the probability. Being able to deal with such simplified situations, we can value the option for each possible situation at every time step before expiry. Working back the tree time step by time step, we finally reach the root node of the tree. By subdividing the initial problem into clearly arranged situations, we can calculate a value for the option. Trees are easy to understand. They visualize what can happen to the market. While a formula yields a value in a magic way, the link to the value drivers is missing. A tree allows comprehending how and when a value driver impacts the value. Everyone resolving a tree has no problems to say how the final value is built up. Simulations. Simulations follow an even simpler concept than trees. We determine the future beforehand, e.g. we assume certain peak sales numbers, and then attribute a value to the option given that future. This is repeated many times with an algorithm that randomly chooses the outcome. The simulated scenarios must on average correspond to the assumed probability distribution. Once all scenarios are simulated, there is no uncertainty anymore. Every manager can tell how much an investment is worth if he already knows the future. The simulations return not only an average value of the option; they also give a good impression about the uncertainty of the value and are therefore an excellent risk measurement tool.

Fig. 2.18. Simulation paths

Simulation techniques can very well deal with more complicated assumptions, but struggle with some special option structures called path-dependency. Nevertheless, simulations enjoy great popularity in the financial world and are well suited for real options valuation. Unfortunately, simulations

44

Basics of Valuation

require programming expertise and are thus often too complex. We will give some guidelines on how to use simulations in a separate chapter on simulations. Finite differences. Option valuation has first been approached using a mathematical concept called partial differential equations (PDEs). PDEs are a very flexible concept that can deal with many additional hypotheses. They however only display characteristics and properties of the option value, e.g. how it behaves by modifying some parameters, but do not provide the value. For simple hypothesis it is possible, to derive the option value out of a PDE; most of the times however this is impossible. Similar to binomial trees, finite differences subdivide the problem into small concise subproblems. While it is impossible to predict the value development on a large scale, finite differences use PDEs to forecast the value development on a small scale. Finite differences are widely used in engineering, biology, or even meteorology; they are extremely powerful to make forecasts, but require a high degree of proficiency with mathematical concepts and programming. Table 2.4. Comparison of resolution methods Method

Advantages

Disadvantages

Formula

Easy to use

Calculation process not visible

Fast

Only for simple option structures

Sensitivities

Simple assumptions

Easy to understand

Rigid

Trees

Visualisation can deal with more complicated options Simulations Easy to understand Visualisation

Time consuming Problems with path dependency

Can deal with more complicated assumptions Finite differences

Can deal with more complicated assumptions

Calculation process not visible, hard to understand

Can deal with more complicated options

Technically demanding

Valuation Methods

45

For real options valuation of biopharmaceutical projects trees are most suitable. A simple formula cannot cope with the project structure. We will also discuss simulations but omit finite differences because of their complexity. Binomial Tree Valuation in Real Options We will show in the following chapters in detail how to apply binomial trees to real options valuation. It is essential that the reader understand how to construct the tree and how to use it in the valuation. It will be the basis in all subsequent chapters dealing with real options. The application of the Black-Scholes formula to real options valuation will be discussed in a separate chapter. Using a tree, we model the market development, i.e. possible changes of the sales revenues. While we know the market state today, we are already uncertain about the market state after one time step, let’s say next month.

Initial market state

Market states after one time step

∆t

Market states after tw o time steps

∆t Fig. 2.19. Binomial tree

∆t

Final market states

46

Basics of Valuation

The market can either improve or deteriorate. The up- or down-steps of the tree reflect this. The further we go into the future, the less certain we are about the forecast. The tree therefore widens the farther we move in time. Each node represents another market state for the corresponding time. The end nodes represent market states at the final time point of the tree. In real options valuation we model the market up to the last decision point of the project we want to value. We then place ourselves in each of the end nodes, i.e. at the time of the last decision and assume a market that corresponds to the node. The market state in the end node gives us all future cash flows, revenues and expenses. This situation is no different than if we just have to decide without any further uncertainty and can be calculated using standard DCF valuation for all future cash flows of the project after the last node of the tree. In some end nodes the market state might be so bad that the DCF value is negative, i.e. the expected revenues do not provide enough compensation for the necessary investments. In these states, we abandon the project. Consequently, we put the value of the project to zero by abandoning it, because there are no future cash flows anymore. Putting the value to zero instead of keeping the negative NPV corresponds to the option to abandon. In these states, the project value is increased from a negative value to zero, while in the other states where the project is profitable, the values are kept the same. The average value of all scenarios is therefore higher than without putting the negative values to zero. This difference in value is directly linked to the option to abandon, which avoids losses and is the source for the different values in DCF and real options valuation. In a next step we place ourselves in the nodes one time step before the nodes we have just solved. The uncertainty in these states is not too complex. There are just two possibilities, either we reach after one time step the upper state or the lower state, in our case the upper or lower end node. For both scenarios we know the probability and the value of the project in these states – we have just calculated them. We can therefore take the expectation of the project values, discount it and account for the cash flows that are due in this very time step. Assessing the value for each node at that time step, we have already calculated the values for the final two layers of the tree. Applying the same procedure to each previous layer we can work back the tree to its root. Finally, the value of the project corresponds to the value calculated for the root node, because this is exactly the time and market state of today. We will now explain each step of the tree valuation outlined above in detail and illustrate it with an example.

Valuation Methods

47

Revenues

Expenses

Value

Project abandoned

Fig. 2.20. Calculation of end nodes

Step 1: Determine the project parameters, i.e. the underlying S, growth rate µ, volatility σ, the time step size ∆t, the expenditures, the success rates, and the duration of the project. Before constructing the tree, we have to fix the input parameters of the project for the valuation. Not all are already relevant for the tree; but it is better to define the parameters before starting to set up the Excel sheet. We first define the underlying of the binomial tree, S. In most cases the underlying of the tree is the peak sales or the market state. Once we have the peak sales, we estimate their growth rate and their uncertainty, i.e. their volatility. If we know all characteristics of a project, i.e. all test results of the development; we can more or less accurately determine the peak sales at the time of launch. Before, i.e. during development, many factors have an impact on the estimated peak sales. New knowledge on application, safety and efficacy, a change in the competitive environment, e.g. another company launches a similar product, or an economic downturn, influence the estimate of the project’s peak sales S. We do not know how all these risk factors will evolve, but we assume that first, the general market for the medication is going to grow at the growth rate µ and second, the volatility of the peak sales estimate S is σ. The volatility is a measure of uncertainty, i.e. the exactness of our peak sales estimate today. The more uncertain the

48

Basics of Valuation

peak sales estimate today is, the higher is the volatility of the estimate. We will further discuss volatility in the chapter on life sciences valuation. Finally, we determine the project related parameters, i.e. the duration of the different phases, the success rates and costs of the phases. In our example, we assume that a company develops a product. The product first has to pass two development phases, where it is tested on feasibility and other concerns like compatibility with other products, compliance with regulation or safety. We estimate the peak sales to be $ 150,000. The volatility of our estimate is 30%. The project is at the start of the first development phase that lasts 2 years. We now decide whether to go ahead or not, depending on the project value. After two years, the project will finish phase 1 testing with a success rate of 75%. Again, depending on the outcome of phase 1 testing and the market conditions at that time, we will perform another valuation; we will only start phase 2 testing if the project has a positive value. Phase 2 testing lasts 1 year and will be successfully completed in 80% of the cases. The costs for phase 1 and phase 2 testing are $ 25,000, and $ 30,000, respectively. If the project passes phase 2 testing successfully, we will value the project again. Depending on the valuation, we will launch the project. The costs to launch the project are $ 150,000 and are due at the beginning of the fourth year. Our company discounts the project at a rate of 15%. We are looking at 5 years of sales. The operating margin is 50%. Table 2.5. Project characteristics Estimated peak sales Volatility Margin Discount rate Time on market

$ 150,000 30% 50% 15% 5 years

Phase 1 testing Duration Success rate Costs

2 years 75% $ 25,000

Phase 2 testing Success rate Costs

1 year 80% $ 30,000

Launch costs

$ 150,000

Valuation Methods

49

Step 2: Spanning the tree: With the assumptions made in step 1 we can proceed to spanning the tree. The underlying, in our case the peak sales estimate, move either up or down during the next time step ∆t. But on average the estimate move to Seµ ∆t (or to S(1+µ)∆t if µ is discretely compounded). As another condition we want the tree to recombine, i.e. the market gets to the same state after a down- and then an up step as after first an up- and then a down-step.

Recombinant tree

Non-recombinant tree

Fig. 2.21. Recombinant versus non-recombinant tree

As we see in the figure, this feature reduces the complexity of our tree considerably. The size of the up and down step is determined by the volatility and by the time interval. The higher the volatility is, the larger is the angle between the branches. The formulae are displayed in the box below. If we design the tree in the way shown in the box, all the conditions are fulfilled. The average value after the time step ∆t is pS u + (1 − p ) S d = Se µ∆t . Due to the condition d=1/u, the tree is recombinant and the uncertainty σ is correctly implemented in the tree. Therefore, we can build a tree with these relationships from now until the last decision point. The choice of the time step influences the accuracy of the valuation. The shorter the time step, the more accurate the valuation, but the more fastidious the calculation. For real option purposes time steps of one, three or six months, in some cases even of one year are sufficient. step up: u = eσ

∆t

step down: d = e −σ

(2.19) ∆t

(2.20)

50

Basics of Valuation

uS p S

1-p dS

∆t Fig. 2.22. Step up and step down

probability up: p =

e µ∆t − d u−d

probability down: 1 − p

(2.21) (2.22)

We now span the tree for our example. For simplicity, we use time steps of one year. The estimated peak sales today are $ 150’000, the growth rate is assumed to be zero, and the volatility of the estimate is 30%. Using the formula of the box below, we now calculate the step size for each year. u = eσ

∆t

= e30%

1

= 1.35

d=

1 1 = = 0.74 u 1.35

p=

e µ∆t − d e0⋅1 − 0.74 = = 43% u−d 1.35 − 0.74

In year two we expect the peak sales estimate to change with the probability p to 1.35 times the initial sales estimate of $ 150,000, resulting in $ 202,479. With a probability of 1–p, the estimate goes down by 0.74 to $ 111,123. The probabilities for the step up, p, and down, 1–p, are again calculated according to the formula in the box. We receive for p 43%, and for 1–p 57%. This means that in every year we expect our sales estimate to increase by the factor 1.35 with a probability of 43%, and to decrease by the factor 0.74 with a probability of 57%. As we have already outlined, the changes in the peak sales estimate derive from the changes of the market, e.g. new information about the project in terms of applications, safety, or customer acceptance. This flexibility of the estimated sales figure is an important difference to

Valuation Methods

51

DCF, where we assume that the estimate once fixed, stays constant during the entire project life. Obviously, it is more realistic to assume that an estimate of future sales figures is subject to change and fluctuation. We later see how this is applied to life sciences projects. The figure below displays the tree spanned to the market entry of the product. The figures in the top of the boxes are the estimated peak sales, and below the probabilities to reach that state.

Year 1 Phase 1

Year 2 Phase 1

Year 3 Phase 2

Year 4 Market 368.940

150.000 100%

202.479 43% 111.123 57%

273.318

8%

18% 150.000 49% 82.322

202.479 31% 111.123 42%

33%

60.985 19%

Fig. 2.23. Binomial tree displaying market states and their probabilities

We see that in the best case our peak sales estimate is $ 368,940, but we expect to reach that number at the time of market entry only in 8% of all cases. The worst case, $ 60,985, is reached with a probability of 19%. The following figure exhibits a tree modelling a market of initial 100 units, an annual growth rate of 0% and an annual volatility of 40%. We chose 6 months time increments. After 5 years the tree allows 11 different states of the market, the range lasts from 6 to 1,692. At first sight this seems to overweigh the high values and therefore a positive market development. If we take a closer look at the tree, the numbers keep their justification. If an ant walks along the tree, always from the left to the right, then there are 1,024 different paths at the ant’s choice (there are 10 points where the ant has to chose to go up or down – between two possibilities – hence there are 210=1,024 different paths). But only one path ends at the very top – if the ant each time chooses to go upwards. On the other hand, there are many more

52

Basics of Valuation

10'000 1'000 100 10 1 0

1

2

3

4

5

Fig. 2.24. Tree for a five years’ period

possibilities to reach one of the middle states. The ant is supposed to go upwards with a probability of 49% and downwards with a probability of 51% – these are the tree parameters that correspond to a growth rate of zero, a volatility of 40%, and a time step of 0.5 years. The different number of paths and their probability induces different weights (probabilities) for each end state. The following graph displays this relationship. The fact that it is slightly more probable that the market goes down at each node (51%) shifts the probability curve on the graph a bit to the left. The possibility that a medication, whose peak sales are estimated today at $ 100 mn, would reach peak sales of $ 1,692 mn in five years seems not realistic. Nevertheless one can argue that it is possible, though not 30%

300

25%

250

20%

200

15%

150

10%

100

5%

50

0%

0 6

10

18

32

57 paths

100 176 310 546 961 1692 probability

Fig. 2.25. Frequency and probability of different end states

Valuation Methods

53

realistic, that competitive products suddenly deal with severe problems; the applications of the drug is consequently expanded to other diseases or the efficacy exceeds anything ever seen and a major demand for this kind of drug arises. This way the medication could become a blockbuster. Roche experienced a similar scenario with the threat of the avian flu pandemic. This sudden change of market state let sales of Tamiflu rise from CHF 330 mn in 2004 to CHF 1,558 mn in 2005. Still this scenario is comparable to winning the lottery, but someone finally wins it. The tree considers that by attributing a minimal probability to this end state. The most probable end states are 57, 100 and 176 with more than 65% of probability all together. On first sight some end states of the tree might seem unrealistic, but the tree captures exactly the probabilistic hypotheses. The more unrealistic a state is, the lower the probability attributed to it by the tree. Step 3: Solving the end states. After having set up the tree we now solve the tree from the end where no uncertainty is left, i.e. where the decisions can be taken according to clear criteria. The tree puts you into hypothetical states at the final decision points, i.e. at the end nodes of the tree. In each state, the tree gives you a best estimate of the potential of the project, if this state were actually achieved. With this input, we are able to calculate the value of the project. We know the future payoff of the project, as the end node tells us the size of the peak sales and there are no more decision points ahead. Using these peak sales along with the expenditures to launch and market the project; we can calculate the net present value of the project at the end node. Next we decide whether to invest or not according to the value of the DCF calculation: Invest if the discounted (and probability-adjusted) future revenues exceed the investments, abandon otherwise. This way the value of the project is necessarily positive, as all projects with a negative rNPV are not continued. The project either is profitable, i.e. the revenues are higher than the expenses, or the project is abandoned, in which case the value is zero. This is the first implementation of the option idea. Nobody would voluntarily make losses; one rather exercises the option to abandon the project and floors the future losses to zero. The value of the DCF calculation is now attributed to all end nodes in the tree. In all cases where we have a negative value we put the end node value to zero. Resetting the value to zero is only done when we are at a decision point, i.e. start of a new phase, here the market phase. In our example we start with solving the most upper node, with expected peak sales of $ 368’940. The sales will be constant for 5 years and then stop. We now have to find the value for this end node by multiplying the yearly sales with the margin of 50%, and then discount them back to the date of launch, i.e. beginning of year 4.

54

Basics of Valuation Table 2.6. Calculation of NPV in the end nodes (in $) Year 4

Year 5

Year 6

Year 7

Year 8

Discount

Market 100%

Market 87%

Market 76%

Market 66%

Market 57%

Sales Operating Profit DCF NPV

368,940 184,470 184,470 711,129

368,940 184,470 160,409

368,940 184,470 139,486

368,940 184,470 121,292

368,940 184,470 105,471

Sales Operating Profit DCF NPV

202,479 101,239 101,239 390,276

202,479 101,239 88,034

202,479 101,239 76,552

202,479 101,239 66,567

202,479 101,239 57,884

Sales Operating Profit DCF NPV

111,123 55,561 55,561 214,188

111,123 55,561 48,314

111,123 55,561 42,012

111,123 55,561 36,533

111,123 55,561 31,767

Sales Operating Profit DCF NPV

60,985 30,493 30,493 117,549

60,985 30,493 26,515

60,985 30,493 23,057

60,985 30,493 20,049

60,985 30,493 17,434

The table above represents the calculation of the net present value of the project’s operating profit at the time of launch for the end nodes. We first calculate the operating profit from the sales figures for each year. In a next step we discount these operating profits with the discount factor corresponding to the respective year. The discount factors are calculated with the formula (2.13). We then sum up the discounted yearly operating profits. The results for the four different states are displayed as the NPV. Now we still have to take the decision if we launch the product or not. We therefore subtract the expenses from the net present value of the operating income and receive the following numbers, displayed in bold for each scenario (see Table 2.6). We see that only in the case of the lowest end node we do not launch the project. The future income would not payoff for the launch expenses and the project would not be profitable. For all further calculations we set the

Valuation Methods

55

Year 1 Year 2 Year 3 Year 4 Phase 1 Phase 1 Phase 2 Market 711,129 561,129 390,276 240,276 214,188 64,188 117,549 -32,451 Fig. 2.26. Project value in the endnodes

value of this box to zero, as we do not proceed in this node, avoiding the expected losses. Step 4: Working back the tree to the previous phase. We now place ourselves in one node right before an end state. This node represents a state ∆t before the last decision point. What happens between then and the decision point, i.e. the end node of the tree where we already have the value? First, the results of phase 2 will be known. This is crucial for the ongoing of the project; a negative result corresponds to the end of the project. Second, the market potential is going to fluctuate, either a step up or a step down in the tree. Third, the time simply advances by ∆t, this forces us to discount future cash flows by this time difference. We consider these three points for the calculation of the project value at our specific node:

our node

' t Fig. 2.27. One time step in the binomial tree

56

Basics of Valuation

1. The results are part of a project’s technical uncertainty. The outcome is not known in advance. We can only make use of statistical success rates. We then adjust the project value by this probability. 2. The tree reflects the market uncertainty, there are two end states following our node. The expected value of the project in our state is therefore the average of the project value at the two following end states. 3. Finally the whole value must be discounted for the time interval ∆t. If VTup and VTdown are the project values of the two consecutive end states, then the value VT − ∆t of the project at our node can be expressed in the following way:

1 VT − ∆t = N P ( pVTup + (1 − p)VTdown )

(1+ r )∆t 1 2

(2.23)

3

P is the statistical success rate of the phase, p the probability of the step up. Now we calculate the project value for all nodes at T-∆t with the formula above. Then we have to go another time step back to the nodes at T-2∆t. Note that if between T-∆t and T there is no end of a phase, i.e. if we are in the middle of a phase, we do not have to account for a failure of a phase. This must only be done at the very last time step of a phase. The value VT − ∆t is therefore: 1 VT − ∆t = ( pVTup + (1 − p)VTdown ) (1 + r )∆t 2

(2.24)

3

Step (1) drops out, only the expectation (2) and the discounting (3) enters into the calculation. Applying this scheme we work back the tree to the root, carefully considering the difference in the calculations at nodes where we take decisions, i.e. after test phases, and where not. We calculate the value for the tree, starting with the most upper node for year three. According to the formula, we get: 1 ⎤ ⎡ V3273 = ⎢80%(43% ⋅ 561,129 + 57% ⋅ 240,276 ) − 30,000 1 + 15% ⎥⎦ ⎣ V3273 = 232,134 We repeat the same for the other two nodes:

Valuation Methods

57

1 ⎤ ⎡ − 30,000 V3150 = ⎢80%(43% ⋅ 240,276 + 57% ⋅ 64,188) 1 + 15% ⎥⎦ ⎣ V3150 = 66,781 1 ⎤ ⎡ − 30,000 V382 = ⎢80%(43% ⋅ 64,188 + 57% ⋅ 0 ) 1 + 15% ⎥⎦ ⎣ V382 = − 10,998 As we are again at a decision point, the negative value in the lowest node is set to zero, because we do not continue the project in this case and therefore avoid the losses. This actually means that the lowest point at launch is never reached, because the project would already be abandoned beforehand. Year 1 Year 2 Year 3 Year 4 Phase 1 Phase 1 Phase 2 Market

561,129 232,134 240,276 66,781 64,188 0 0 Fig. 2.28. Calculating back the tree to year 2

Step 5: Working back the tree to the root. The resolution of the rest of the tree is now a repetition of the previous steps. At each start of a phase you decide at each node, if it makes sense to continue the project. For this you compare the value of the project you have obtained in the above-described way with the costs of the phase. Replace the previous project value either with the positive difference between value and investment or put it to zero in case the difference is negative and the project consequently abandoned. Again, this is the implementation of an option in the valuation process. Then work back one time step and adjust for the technical uncertainty of the previous phase (1) as well as you take the expectation (2) and discount all

58

Basics of Valuation

values for another ∆t (3). After that, you work back to the start of the previous phase by only averaging (2) and discounting (3). Move back to the root of the tree applying these steps for each phase. In our example, for year 2, where we do not have any decisions to take, we calculate the following:

V2202 = 75% ⋅ (43% ⋅ 232,134 + 57% ⋅ 66,781)

1 (1 + 15%)

V2202 = 89,445 V2111 = 75% ⋅ (43% ⋅ 66,781 + 57% ⋅ 0)

1 (1 + 15%)

V2111 = 18,534 The expenses for the phase 1 are triggered at the beginning of year one. We can now calculate the value for the root node the same way as we have calculated the value for year 3. V1150 = ( 43% ⋅ 89,445 + 57% ⋅ 18,534)

1

(1 + 15% )

− 25,000

V = 17,357 The tree then looks the following: Year 1 Phase 1

Year 2 Phase 1

Year 3 Phase 2

Year 4 Market

561,129 232,134 89,445 17,357

240,276 66,781

18,534

64,188 0 0

Fig. 2.29. Tree solved up to the root node

The root node value represents the real option value of the project today. The project has a real option value of $ 17,357.

Valuation Methods

59

Difference Between Real Options and Financial Options

In the following chapter we discuss the difference between real options and financial options and the implications on using the Black-Scholes formula for real options. The chapter is intended for readers interested in advanced discussions of real options and is not fundamental for the following chapters of the book. Real options owe their name to their similarity to financial options. Almost every introduction to real options compares an investment opportunity with a financial call option. This suits to explain the general concept of real options. Unfortunately, virtually all literature about real options only concentrates on the second word of the name, “options”. The word “real” is not only a descriptive of the different area of application – real economy vs. financial economy – it is also an important statement about the very nature of these options and has significant impact on quantification. We start with a short explanation of the fundamentals of financial option pricing, i.e. about hedging. Then we elucidate the difference between real and financial options and explain its influence on valuation. Financial option valuation. The concept of financial option valuation has been introduced by Fisher Black and Myron Scholes with their seminal paper in 1973 (Black 1973). Since then, the basics haven’t changed significantly. They have shown that a financial option is replicated with a portfolio of the underlying and a bond. This allows mitigating all risk. And since the portfolio is without risk, it must earn the risk free rate, otherwise there would be an arbitrage opportunity. The simplest way to visualise this concept is by means of a binomial tree. First, we model the fluctuations of the underlying of the option. We assume that the value of the underlying can go up or down during the next time step. Assuming today’s value of the underlying being S, after one time step it can be either uS or dS, with u denoting the up step, and d the down step. Each scenario, the up or the down step, has a certain probability p or 1-p. The parameters u, d, and p should be set in a way such that they model correctly the underlying’s fluctuation. In the Black-Scholes concept it is usually assumed that the underlying follows a geometric Brownian motion with growth rate µ (also called drift) and volatility σ. We give two examples of parameter sets u, d, and p that model correctly a geometric Brownian motion with parameters µ and σ. The table below summarises the parameters to model the fluctuation of the underlying:

60

Basics of Valuation Table 2.7. Parameter sets for binomial trees Set 1

Set 2

u

eσ

e µ∆t +σ

∆t

d

e −σ

e µ∆t −σ

∆t

p

∆t ∆t

e µ∆t − e −σ eσ

∆t

− e −σ

1 − e −σ

∆t ∆t

eσ

∆t

∆t

− e −σ

∆t

Second, we define the replicating portfolio. The portfolio consists of the option, a portion ∆ of the underlying (We use the notation ∆ for consistency with the literature. Don’t confuse ∆ with the time increment ∆t). The remaining part is allocated to a risk-free bond, which we can neglect in what follows. ∆ is chosen such that, no matter what happens to the underlying, the return of the portfolio is the same. This is achieved with the following reasoning (V, Vup, and Vdown representing the value of the option at the different nodes of the binomial tree): Vup + ∆uS = Vdown + ∆dS

(2.25)

Solving for ∆ we get: ∆=−

Vup − Vdown

(2.26)

uS − dS

Valuation of Financial Options

1. Model the underlying 2. Build a replicating portfolio 3. Calculate the expected risk-free discounted value using risk-neutral probabilities Third, we have to find a way, how to value the option. From the fact that the portfolio has by construction always the same value after one time step in all nodes of that time, we conclude that it is risk free. A risk free asset or portfolio must earn the risk-free rate rf, otherwise we would have an arbitrage opportunity.

Valuation Methods

61

Vup V Vdown

't Fig. 2.30. Time step in option valuation

It would now be easy to deduce the value of the option from the values of the option at the leafs of the tree. The value of the option however is not equal to the expected and discounted future value of the option, because we cannot consider the option alone but only in conjunction with the replicating portfolio (otherwise we would not have a risk-free asset). Let us therefore define a fictive number q such that:

V =e

− r f ∆t

(qV

up

+ (1 − q )Vdown )

(2.27)

The purpose of the introduction of q is a valuation algorithm that can focus on the option alone instead of always having to consider the replication portfolio. We have to stress that the introduction of q is a trick. Using the postulated property of q, the fact that V + ∆S = e

− r f ∆t

(V

up

+ ∆uS )

(2.28)

and the definition of ∆ we can deduce the value of q. Vup + ∆uS − ∆Se

q=

r f ∆t

= qVup + (1 − q )Vdown

(

Vup − Vdown + ∆S u − e

r f ∆t

)

(2.29) (2.30)

Vup − Vdown

r ∆t r ∆t (2.31) u−e f e f −d = u−d u−d Using q according to its definition we can now calculate the value of the option V in a simple way:

q =1−

62

Basics of Valuation

V =e

− r f ∆t

(qV

up

+ (1 − q )Vdown )

(2.32)

q is sometimes called risk-neutral probability, because its use corresponds to a discounted expectation, using q as probability within the tree and discounting at the risk-free rate. Note that usually discounting something uncertain at the risk-free rate is not accepted, because investors want to be rewarded with a risk premium. Nonetheless, in the case of financial option valuation this is allowed because it is only a trick. In fact, we discount the whole portfolio at the risk free rate, and the whole portfolio does not fluctuate. The risk-neutral probabilities for the above mentioned parameter sets are: Table 2.8. Risk free formulae for binomial tree Set 1

Set 2

σ ∆t

u

e

d

e −σ

q

e f − e −σ eσ ∆t − e −σ

∆t

r ∆t

∆t ∆t

e µ∆t +σ

∆t

e µ∆t −σ

∆t

(r

e f eσ

)

− µ ∆t ∆t

− e −σ ∆t − e −σ ∆t

We recognise especially from parameter set 1 that using risk-neutral probabilities corresponds to a change of the underlying’s drift parameter µ to rf. In mathematical terms we have changed the underlying’s real-world probability measure to its risk-neutral probability measure. But remember, this is only a trick for the purpose of the valuation, the underlying still follows a geometric Brownian motion with parameters µ and σ, whether there are some options or not. Real options valuation. We now follow the same steps and quantify real options. First, we model the uncertainty of the underlying. This can be either estimated revenues of a product, the gold or oil price, or another economic metric. For simplicity let us assume that the underlying follows equally a geometric Brownian motion with parameters µ and σ. Second, we build the replicating portfolio. This now leads us to the very difference between real options and financial options. While financial options are a contract that defines rights on a traded asset in a liquid market, real options are business opportunities. The decision to seize such opportunities depends on an economic metric, the underlying of the real option. This underlying

Valuation Methods

63

however can be a virtual number, an estimate, e.g. the estimated revenues of a not yet developed product, or a share of a specific market the company considers expanding in. Most often the underlying is not tradable, which makes it impossible to build a replicating portfolio. In the case of commodity companies like mines or petrol companies, the underlying is traded in a liquid market. Nonetheless, the question we have to ask is not “Can we build a replicating portfolio?” but rather “Do we build a replicating portfolio?” If we do build a replicating portfolio, then we can use the valuation of financial options as described above. If we do not, then we cannot use the valuation of financial options, because the fundamental concept of replication is not respected. We have to keep in mind that financial options are valued as part of a risk-free portfolio, and only this justifies the use of the riskfree rate in the risk-neutral probabilities. If you valued real options with riskneutral probabilities you would value something you do not have, namely a business opportunity that is hedged. Even oil companies and mines may only apply risk-neutral valuation if they hedge their real options. Hedging the business opportunities would correspond to pocket in the value of the business opportunity and then hedging it until the end of the life cycle. This would be real arbitrage, a very smart way to run a business. Unfortunately, we have not heard of any company doing so. The reason might be simple: Real options model the uncertainty of some value drivers. But there are still a lot of other factors not built in the model and not hedgeable, and managers cannot trust their models such that this would justify large financial transactions to hedge their business opportunities. For the moment, it seems that the model risk outweighs the business risk. But, let us return to the standard real options that cannot be replicated anyway. We have to acknowledge that we must not use the elegant valuation method that applies to financial options. So, the second point of financial option valuation cannot be executed. Consequently, the third point that is based on the second cannot be respected either. So what can we do? Let us go back to the actual meaning of “real options”. Real opportunities are business opportunities that require decisions in the future. Until then, some value relevant parameters can fluctuate. The decision will eventually be based on the parameter set that is encountered at the time of decision. Hence, in contrast to other valuation methods, using real options we just model some additional decisions that influence the course of the project. Typically, these decisions are abandonment or expansion. The decisions and the subsequent scenarios occur with some probability. These probabilities depend on the probability distribution of the decision driving parameters. It is therefore straightforward, first, to model the fluctuation of

64

Basics of Valuation

these decision relevant parameters, second, to anticipate for each possible scenario what management would decide, and third, to calculate the expected value of all these scenarios. In the binomial tree, this means that the value of the real option can be described by:

V = e − ra ∆t ( pVup + (1 − p )Vdown )

(2.33)

We notice two differences with respect to financial option valuation. First, we use the real world probability p that models correctly the likelihood that either the one or the other scenario occurs. Second, we use the risk-adjusted discount rate ra instead of the risk-free rate r. Since the value of the option fluctuates with the underlying, the option holder faces some risk, which he wants to be rewarded for. Again, the financial option was part of a risk-free portfolio, so there was no risk that would have to be compensated. In this context, real options valuation is nothing else than a risk-adjusted expectation. Risk-neutrality or change of probability measures like for financial options lack any justification. The difference between real options and other valuation methods lies solely in the modelling of future decisions and their underlying drivers. Some authors try to bypass this apparent inconvenient that avoids the direct application of Black-Scholes by assuming that the real option can be hedged. They suppose that there is a traded asset that could be used to replicate the real option. This hypothesis is called market asset disclaimer (MAD). MAD is about as useful to plan your agenda assuming the week has eight days because you are short in time. The planning might work fine, but at the latest on the eighth day reality will wake you up. In corporate finance the investors’ wake-up call uses to be little tender. Consequences. All formulae from quantitative finance are considered to be applied to financial option valuation. Therefore, we have to adapt them prior to using them for real options valuation. In the case of the binomial tree we have to use the real-world probabilities and use the risk-adjusted discount rate when working back the tree. Black and Scholes’ partial differential equation (Black et al. 1973) has to be modified in the following sense: 1 Vt + µSVS + σ 2 S 2VSS − raV = 0 2

(2.34)

For the financial option version of this PDE we have to replace µ and ra with r. This formula can be derived by using the Hamilton-Jacobi-Bellmann equation.

Valuation Methods

65

The well-known and often used option formula, equally derived by Black and Scholes, must read: V = Se ( µ − ra )t N (d1 ) − Ke ra t N (d 2 )

(2.35)

1 2⎞ ⎛ ⎜ µ + σ ⎟t 2 ⎠ d1 = ⎝ , d 2 = d1 − σ t σ t

(2.36)

Conclusion. Black and Scholes derived an elegant method to value financial options as part of a risk free portfolio. Consequently, they could discount at the risk-free rate and use risk-neutral probabilities. In contrast to that, real options cannot be, or, equally important, are not hedged. This implies that real options valuation simply becomes an expectation calculus discounted at the risk-adjusted rate. If you use risk-neutral valuation for real options and you do not hedge, you value something you do not have – a business opportunity as part of a risk-free portfolio – instead of something you have – a business opportunity with real risks.

Valuation in Life Sciences

In the following part we discuss how the theory outlined in the previous chapter is applied to valuation in life science. First we look at the input parameters related to life science valuation and then at some special aspects, i.e. discounting, volatility and peak sales prediction, before moving on to the actual valuation. We start with the valuation of standard projects and then gradually move to the more complex problems. Every section includes a case study to illustrate practically how to use the theory. The exercises in the last part of the book complement the cases.

Fundamentals In the following chapter the reader learns more about the data required for valuation in life sciences companies. We first discuss drug development and then medical device development. The chapter shall serve as a reference for those readers not familiar with life science. For reasons of clarity, we have not included all publicly available data, but have generated a combination the user can apply to valuation. Drug Development

Drug Discovery Drug discovery following basic research can in most cases be broken down into four phases: Target identification, target validation, lead identification, and lead optimisation. In the following we briefly describe each phase for those readers not familiar with the topic. Target identification: Biological research studies the basic cellular processes in the healthy and pathologic state. By comparing these states, disease responsible actors are identified as possible drug targets. These targets are mainly proteins, such as cell surface receptors. Once a target has been

B. Bogdan, R. Villiger, Valuation in Life Sciences, 3rd ed., DOI 10.1007/978-3-642-10820-4_3, © Springer-Verlag Berlin Heidelberg 2010

68

Valuation in Life Sciences

selected, its interacting partners, biologic, and biochemical function is examined. A disease model is then set up and studied. Cultured human cells or animals such as mice serve as model. Target validation: In search for the suitable target, each target for a given disease is compared to others and its potential in regulating biologic processes is assessed. The most promising targets are selected for further drug development. Hit identification: In the next step, compounds acting on the selected target and leading to the looked-for change are discovered and characterized. Usually large libraries consisting of thousands to millions of molecules are used. Compounds interacting with the target, leading to the looked-for change, are selected, if present in the library. The identified compounds are called leads. Lead optimisation: Lead optimisation is the stage during which medicinal chemists attempt to improve primary leads to achieve the best compromise between improved activity, bioavailability and safety. Often during this same stage of development, lead prioritisation studies are conducted in living organisms (in vivo) and in cells in the test tube (in vitro) to compare various lead compounds and how they are metabolised and affect the body. Once the optimised lead is chosen, it is taken to the last stage prior to human testing, the preclinical studies. Clinical Trials Preclinical testing: Once a lead compound is selected, its safety has to be tested in animals before human testing takes place. Preclinical tests in animals are designed to study the effect of the drug and its metabolites in the living organism. Toxicity, metabolic pathways, and excretion are studied in detail. Preclinical studies are needed to see if the drug can safely be tested in humans, or if it causes dangerous toxicity. Due to the poor predictive value of current animal model, previously elaborated hypothesis on the mechanism of action often fail in this phase. It might be necessary to reformulate the drug, costing a lot of time. When preclinical testing is done, several drugs that seemed promising are abandoned; the further development is stopped due to problems of poor absorption, toxicity or simply because the drug doesn’t work.

Fundamentals

69

Clinical phase 1 trials: Phase 1 is the first phase of drug testing in humans. Phase 1 is conducted in a small group of healthy volunteers (20-80) and is designed to evaluate safety, dose range and side effects. The same questions as in preclinical testing are now assessed in humans. Absorption of the drug (bioavailability) into the blood stream is studied, as well as the mechanism of action in men. This involves the assessment of biochemical and physiological effects of the drug, including the correlation of action and effect of the drug with its chemical structure. Furthermore, the safety profile of the drug is monitored and side effects are carefully studied and documented. A license, which is granted upon study of preclinical data, is required to start human drug trials. In the USA the Federal Drugs Agency (FDA), in Europe the European Medicine Controls Agency (EMEA) is responsible for approval of human testing. In the US prior to starting human clinical trials, an investigational new drug application (IND) must be filed to the FDA. Only if the IND is approved, clinical trial can start. In most countries, an independent committee, periodically reviewing results to ensure proper conduct and respect of ethical issues, monitors human trials. If the drug shows no serious side effects and no toxicity, it can be taken to phase 1I trials. Phase 1 trials are often the endpoint for new drugs, as many fail to meet the expectations in men. Clinical phase 2 trials: In phase 2 trials, the drug is studied in a group of people (100-300) having the disease to be treated. Clinical phase 2a studies are small studies to define the dose; phase 2b studies are designed to prove effectiveness. The goal of phase 2 trials is to show the proof-of-concept of the drug, i.e. that it is effective in the treatment of the target disease. Nevertheless, pharmacodynamics and pharmacokinetics need to be assessed again, as the drug effect might differ in the patient group from healthy volunteers. Special attention is needed in patients with impaired renal or liver function concerning the accumulation of the drug or its metabolites. Another factor to be considered are the interactions between the drug under investigation and concomitant medications. In this stage, the drug might not show the wished effects in the patient group, which may mean the drug has to undergo modification or reformulation. However, in addition to all the absorption and excretion questions the drug also has to demonstrate a clear efficacy in the target patient group. New drugs often do not show a clear benefit over existing treatments in terms of efficacy, safety or delivery and will therefore have less chance to be approved by regulatory authorities and fail when marketed. Thus, most companies decide not to continue to develop a drug with poor efficacy at this point in testing.

70

Valuation in Life Sciences

Clinical phase 3 trials: If the drug looks promising at this stage, it is taken forward to large-scale phase 3 trials. The drug is tested in a large group of patients (500-20,000) with the condition to be treated. The main aim of this phase is to confirm the effectiveness of the treatment, to disclose any side effects, and to establish the right dosage for the future patient treatment. Questions to be answered are how effective the drug really is, whether it is more effective than other competing treatments on the market, and finally how should the patient labelling be. Approval phase: A company files a new drug application after successful clinical trials. The FDA or EMEA reviews the data and decides if the company is granted a marketing approval. Possibly, the regulatory authority asks for further clinical trials or even rejects marketing approval. The time for reviewing the drug ranges from 1 to 2 years. Phase 4 trials: Phase 4 trials are conducted after the drug has been launched. These trials give more information about long-term safety and efficacy of the drug. Alternative formulations, different dosage or new patient populations are tested. Medical Device Development and Approval

Development and Approval in Europe In the European Union notified bodies (NB) play a similar role for medical device development as the FDA in the US. The NBs are controlled organizations that can review medical devices in development and issue a CE-mark, necessary for certain devices for marketing approval. Once a medical device has obtained the CE-mark, it can be marketed in the entire European Union. In the US, the manufacturer of a class III device (see below) must prove that the device is safe and effective, usually requiring a large randomised prospective clinical trial. In Europe, the manufacturer most show that the device is safe and its use is as intended by the manufacturer. This usually can be demonstrated in a small clinical trial involving less than 50 patients (Kaplan et al. 2004). This major difference between the US and Europe makes it much more costly to receive marketing approval in the US. Development and Approval in the US The FDA regulates the medical device development in the US. Devices have different pathways to approval based on their classification:

Fundamentals

71

• Class I: General controls: crutches, band aids • Class II: Special controls: wheelchairs, tampons • Class III: Pre-Market Approval (PMA): heart valves (known to present hazards requiring clinical demonstration of safety and effectiveness) – OR – not enough known about safety or effectiveness to assign to Class I or II Class I devices are the simplest devices, posing the smallest risk and require only general controls. Class II devices are of moderate risk. They require a 510(k) Pre-Market Notification prior to marketing. Class III devices are of high risk. They usually require a PMA prior to marketing. There are only a small number of Class III preamendements 510(k) devices in the US. 40 to 50 PMAs are approved each year in the US. If the FDA defines a device as Class III device, that type will always be a class III device, except if the FDA approves a reclassification petition and downclassifies the device. All competitors that develop similar products have to follow the PMA process in order to market their device in the US. The 510(k) Pre-Market Notification. A 510(k) application demonstrates that a new device is substantially equivalent to another device that is legally on the market without a PMA. If FDA agrees that the new device is substantially equivalent, it can be marketed. Clinical data are not required in most 510(k) applications; however if clinical data are necessary to demonstrate substantial equivalence, the clinical studies need to be conducted in compliance with the requirements of the IDE regulations, IRB review and informed consent (21 CFR parts 812, 56 and 50, respectively). The 510(k) process is a relatively rapid and flexible approval process. The goal of the 510(k) process is: “Demonstration of Substantial Equivalence to a device that was on the U.S. market prior to May 28, 1976, or to a device that has already gone though the 510(k) clearance process.” A PMA device cannot serve as a 510(k) predictive device. Clinical data might be required to demonstrate substantial equivalence to a predictive device. The trials are in most cases smaller and simpler than most PMA clinical trials. The Pre-Market Approval Application. PMA is necessary to market a device that is not substantially equivalent to a device marketed in the US approved under 510(k). The PMA process is more complex than the 501(k) process and the review time with about one year much longer. Unlike most 510(k), a

72

Valuation in Life Sciences

detailed manufacturing section is required. Prior to the final approval of the PMA, the device manufacturing facility must be inspected and approved. The PMA process includes clinical trial that must meet the following requirements: • Well controlled investigations • Partially controlled investigations • Objective trials without matched controls • Well documented case histories conducted by qualified experts • Report of significant human experience with a marketed device from which it can be concluded that there is reasonable assurance of safety and effectiveness of a device under its condition of use. The vast majority of PMA studies are designed as well controlled studies, with randomised patient groups. International data can be used for the PMA approval if it follows the 21 CFR 814.15 guideline. The study must: • Have been conducted in accordance with the declaration of Helsinki, or other ethical procedures, whichever is stricter • Utilise a patient selection similar to a U.S. patient selection. • Utilise a standard of care and medical practice similar to that in the U.S. • Must be performed by competent investigators • Generate data, including source documentation, that are available for audit by FDA. Data on medical device development is not readily accessible and there are no published success rates, durations or costs for the different types of medical devices. Valuations for medical devices require more work in terms of defining the input parameters. Cost

In a recent study on the cost of drug development, DiMasi et al. estimate that it costs US$ 802 mn (year 2000 dollars, capitalised costs) to bring a new drug on the market (DiMasi et al. 2003). In a 2007 publication this number even increased to US$ 1,241 mn. This figure includes all costs, including drug failures, and is a capitalised instead of discounted perspective (see special paragraph at the end of this section). Therefore, it is not ap-

Fundamentals

73

propriate as a starting point to estimate future expenditures in drug development. We cannot assume that a small biotech company will spend a couple of hundred million dollars to finally bring a drug to market. In the following, we will discuss the costs of drug development in more detail based on real life figures. Lead optimisation: The costs for lead optimisation are very variable and depend on the size of the company. A smaller company spends US$ 0.56 mn (average US$ 2.5 mn), while large pharmaceutical companies spend US$ 3-8 mn (average US$ 6 mn). Preclinical costs: The preclinical costs are again highly variable and depend as well on the company size. Small drug development firms spend US$ 1-9 mn (average US$ 3 mn), while the big pharmaceuticals spend 511 mn (average US$ 7 mn). The costs for the clinical phases encompass all necessary expenses to conduct the trials, including project management, drug supplies, toxicology, investigator fees, or study design. The costs outlined below are an average and vary according to the different disease groups, as certain studies are inherently more complex than others. Nevertheless, the costs are a good approximation to be used for the purpose of the valuation. Adjusting the patient number per phase for the different disease groups allows getting a more precise estimate of the costs. The costs furthermore represent a full clinical trial investigation. It is possible to conduct all those trials at less cost. Also, the costs for big pharma are higher than outlined below. DiMasi can be used as a source for big pharma drug development costs. His figures are not representative for young companies. A rule of thumb proposes that the drug development costs for biotech are a factor five smaller than what big pharma spends for the same phase. Clinical phase 1 costs: The costs for an average phase 1 trial based on 100 patients excluding drug supply and non-clinical costs amounts to US$ 2 mn. If we include all costs, the expenditures for the phase 1 are US$ 4-5 mn. Clinical phase 2 costs: Phase 2 costs for clinical expenditures are US$ 78 mn, and US$ 10-11 including non-clinical costs. This number is based on a trial with 300 patients. Smaller phase trial may cost less than $ 5 mn. Clinical phase 3 costs: The clinical phase 3 is the most expensive phase as it involves the largest number of subjects. Average costs range from US$ 10-50 mn for the clinical trial costs and US$ 30-60 mn including nonclinical costs.

74

Valuation in Life Sciences

Approval costs: The approval cost including submission in the US and Europe are in the range of US$ 3 mn. The table below summarises the complete drug development costs for a small to medium-size biotech/pharma company. Of course, these costs can vary considerably. The data below is therefore only a rough estimation what a complete clinical trial programme could cost for an average drug (source: Avance). Table 3.1. Drug development costs Phase

Cost

Lead Optimisation Preclinical Phase Clinical Phase 1 Clinical Phase 2 Clinical Phase 3 Approval

US$ 2-3 mn US$ 2-3 mn US$ 1-5 mn US$ 3-11 mn US$ 10-60 mn US$ 2-4 mn

Success Rates

Even though the pharmaceutical industry has a long-standing history of drug development, the numbers describing the success of the projects are well kept and not accessible. There are a number of publications on the topic, most of which derive from the TUFTS Centre for the Study of Drug Development and which base mainly on data provided by big pharma. The success rate is the chance that a project entering a development phase reaches the next phase. Success rates are assessed per phase and, once the drug is in clinical development, depend on the disease group. There are also major differences between the success rates of chemical and biological drugs. The success rates for lead optimisation are 70%, and 65% for preclinical development for the average drug, independent of the disease group. Combining the different publicly available publications on success rates, we get the following numbers (DiMasi 2001, Kola 2004, Avance). In the table CP means clinical phase, and cumul. stands for the cumulative success rate from beginning of clinical phase 1 to reach the market. The cumulative success rates in the table correspond to the figures published by DiMasi where available (DiMasi 2001).

Fundamentals

75

Table 3.2. Drug development success rates Disease Group

CP 1

CP 2

CP 3

Arthritis/Pain CNS CV GIT Immunology Infections Metabolism Oncology Ophthalmology Respiratory Urology Women’s Health

76.9% 66.2% 62.7% 66.8% 64.8% 70.8% 47.8% 64.4% 66.0% 63.4% 50.0% 39.0%

38.1% 45.6% 43.3% 49.1% 44.6% 51.2% 52.0% 41.8% 39.0% 41.1% 38.0% 42.0%

78.1% 61.8% 76.3% 71.0% 65.2% 79.9% 78.9% 65.4% 64.0% 59.9% 67.0% 48.0%

Approv Cumul. 89.1% 77.9% 84.4% 85.9% 81.6% 96.9% 92.8% 89.7% 92.0% 76.9% 79.0% 59.0%

20.4% 14.5% 17.5% 20.0% 15.4% 28.1% 18.2% 15.8% 15.2% 12.0% 10.1% 4.6%

The attrition rates for biologicals are at present much lower than for new chemical entities but recent trends indicate that they approach the ones of new chemical entities. We therefore suggest using the figures in the table for the purpose of valuation. The reasons for failure in drug development have been described by DiMasi (DiMasi 2001). The abandonment for drugs with their first IND filing from 1987 to 1992 happened due to the following reasons: Table 3.3. Reasons for abandonment of drug development Economics Efficacy Safety Other

33.8% 37.6% 19.6% 9.0%

The primary reasons to terminate a drug development project are lack of efficacy and safety, both technical reasons, and lack of profitability of the project. The latter reason, accounting for a third of all drug development terminations, has important implications to valuation. A prominent example of economic abandonment is Novartis halting the development of NKS104 (pitavastatin) during Phase II trials. We will discuss this in more detail in the chapter on basics in life sciences valuation.

76

Valuation in Life Sciences

Novartis Halts Cholesterol Drug Development

Saturday, December 17, 2005 13:00 IST Basel Novartis has decided to stop the development of NKS104 (pitavastatin), a lipid-lowering agent in phase II for the treatment of elevated total cholesterol, after data from recent investigational trials showed the compound was no longer competitive enough for Novartis to invest further resources. The company intends to seek licensing partners for this compound. As a result, Novartis intends to record an impairment of US$ 266 million in the fourth quarter of 2005 to fully write off the remaining value of this asset. The European rights to this compound were acquired under a licensing agreement from Kowa. Novartis already recorded an impairment of US$ 66 million in the third quarter related to the acquired and capitalised marketing rights for NKS104, informs a company release. Novartis AG is a world leader in pharmaceuticals and consumer health. In 2004, the group’s businesses achieved sales of US$ 28.2 billion and pro forma net income of US$ 5.6 billion. Source: pharmabiz.com Time

Drug development time depends like the success rates on the disease group and the chemical/biological properties for the drug. Below is a table summarising the different average phase lengths for the average drug (source: Avance). Table 3.4. Duration of drug development Phase

Length

Lead Optimisation Preclinical Clinical phase 1 Clinical phase 2 Clinical phase 3 Approval

20-40 months 10-12 months 18-22 months 24-28 months 28-32 months 16-20 months

Fundamentals

77

Peak Sales

There are different definitions of peak sales. We refer to the maximum sales the drug will reach in its life cycle as peak sales. The peak sales number is the net sales the company makes with the drug. The peak sales from existing drugs can be clustered into disease group sales. Table 3.5 gives an overview of the most recent average sales for the major disease groups. The numbers might seem high, but the reader has to keep in mind that first, only the sales of successful drugs are published, second, drugs that do not seem profitable are abandoned during development, third, the average sales are distorted by a few blockbuster drugs lifting the average figure, and fourth, that the sales figures of small population drugs are often not published. If we want to have a sales figure that does not include the unlikely scenario of reaching blockbuster drug status we take the median sales of the drugs currently on market (cf. Table 3.6). For early stage compounds, we recommend to use the median sales as an estimate for potential peak sales. This serves as the basis for conservative valuations. If there is a clear sign that the drug might reach the extremes of the sales distributions, i.e. become a blockbuster, the average number can be used as the estimate for future peak sales. At later stages in development, the sales estimate can be calculated in more detail with a bottom-up approach. Estimation of Peak Sales The estimation of peak sales raises difficulties when preparing the input parameters for valuation. With early stage projects it is virtually impossible to estimate the peak sales. This uncertainty gets resolved with the progression of the project trough development. In the following, we discuss how to predict the sales for projects corresponding to their development stage. Obviously, very early in project development we do not have a clear picture of the future sales potential. We often do not even know what exact disease we can treat with the compound. This is especially true for cancer drug development. Even though there are signs when testing the drug in animals, the indication for clinical testing will only get evident after clinical phase 1. The sales estimate has always to be as precise as possible and justifiable.

78

Valuation in Life Sciences Table 3.5. Disease category average sales Category

Sales

Anti-infectives Blood Bones & Joints Cancer CNS CVS GIT Inflammatory Metabolic Ophthalmologics Pain Respiratory Skin Urology Women

US$ 385 mn US$ 981 mn US$ 127 mn US$ 488 mn US$ 746 mn US$ 466 mn US$ 792 mn US$ 571 mn US$ 803 mn US$ 272 mn US$ 693 mn US$ 646 mn US$ 98 mn US$ 690 mn US$ 514 mn

Table 3.6. Disease category median sales Category

Sales

Anti-infectives Blood Bones & Joints Cancer CNS CVS GIT Inflammatory Metabolic Ophthalmologics Pain Respiratory Skin Urology Women

US$ 265 mn US$ 302 mn US$ 127 mn US$ 344 mn US$ 422 mn US$ 145 mn US$ 299 mn US$ 349 mn US$ 371 mn US$ 157 mn US$ 274 mn US$ 213 mn US$ 69 mn US$ 685 mn US$ 386 mn

Fundamentals

79

It is not believable to predict the peak sales of a cancer drug in preclinical testing based on the number of subjects that will receive the drug once it is on the market. In preclinical testing one often does not even know in which cancer the drug might be efficacious. But we can say that the drug might reach the average peak sales of the most similar comparable basket of drugs that are currently on the market. I.e. if we are looking at a new cytotoxic drug, we can put together the subset of all cytotoxic drugs on the market today and use this figure as a comparable. Of course, the sales figure needs to be projected into the future. We have to consider competition and the behaviour of the disease. For a drug that is even further back in development we could simply use the average sales of all drugs that are currently on the market. In this situation we argue that we do not know enough about the drug, but the most closely comparable reference is the average drug today on the market. We always take the best comparable for early stage projects. If we want to make a conservative estimate for the future peak sales we can take the median of drug sales. This way we diminish the influence of a few blockbusters on the peak sales figure. We can also estimate the peak sales with a top down approach. We define the size of the market, e.g. market for anti-hypertension drugs. Then the future market share of the drug is estimated, e.g. first year 2%, second year 3% etc. Important is to predict the market size, if the market is growing or shrinking and by how much each year. Already modest growth rates at the time of drug development mean enormous changes. For a drug that is in clinical phase I the estimated time to launch might be 8 years. The current market size is $ 800 mn, growing by 4% each year. The predicted market share of the drug in the first year is 5%. At time of launch, the market size will be $ 1.094 bn, and the predicted market share $ 55 mn, instead of $ 40 mn (5% of $ 800 mn), a 37% increase. Once the project advances, we learn more about the details on the drug, its safety and efficacy profile, and its anticipated indication. We know what is the best dosing schedule and we can define a price for the drug based on the current drugs approved on the market today. This approach is called bottom-up and needs considerable due diligence, but allows to predict more precisely the sales potential. The estimate gets much more imprecise if we are looking at a first in class device or drug where we cannot foresee how the regulatory authorities will handle the drug labelling and how the insurers will reimburse the drug.

80

Valuation in Life Sciences

Bottom-Up Calculation Calculating the Patient Number Cook has very well described in his book “Forecasting in the Pharmaceutical Industry” how to calculate peak sales (Cook, “Forecasting in the Pharmaceutical Industry”, 2006). Population. Using the bottom-up approach you define the population of interest, this might be a single country, the USA, Asia, or the world. Doing this you need to consider that the population is best chosen as one with a uniform incidence or prevalence rate. Epidemiology. In the next step we have to define the epidemiological data. This is either the prevalence, i.e. the number of existing cases, or the incidence, i.e. the number of new cases per year. The latter is important in diseases like cancer, stroke, or myocardial infarction, because these are event driven. The incidence or prevalence can be broadly defined or we can go in more detail, i.e. breast cancer in women, or breast cancer in women above 50, or tubular breast cancer in women above 50. We should choose the input as close as possible to the patient population to be treated by the drug we calculate the sales for. Once we know the rate of the disease in the population as defined by incidence or prevalence, we calculate the number of existing or new cases per year, i.e. how many patients have the disease in the population we have chosen. Knowing how many patients have the disease, we need to know in how many patients the disease actually causes symptoms. Symptomatic population. We therefore have to multiply the patient population with the symptomatic rate. This is only necessary if the figures we use for the prevalence/incidence actually encompass all patients, also the ones that do not have any symptoms. Very often it is not known how many patients have a disease but no symptoms, so care is needed in this step of the calculation. We now have the symptomatic population. Diagnosed patients. Of these patients, not all are diagnosed as having the disease. Some might never be diagnosed as having the disease, others are wrongly diagnosed as having another disease. Multiplying the symptomatic population with the diagnosis rate, we get the diagnosed population. Access to health care. Of the diagnosed population, not everybody has access to medicine or wants to have access to medicine. Some might live too far from the doctor, others simply do not have the means to afford health

Fundamentals

81

insurance or there is no functioning health care system in place. After applying the access rate we have the number of patients that have the disease, show symptoms, are diagnosed, and have access to health care. Treatment rate. Now, as not all diagnosed patients are actually treated, we need to apply the treatment rate to the potential population. The final figure we now receive is the number of patients, which are finally treated. When predicting sales, it is crucial to determine beforehand for which time point in the future the sales are calculated. If the launch of the drug is in 10 years, the parameters can change. Especially the diagnosis rate can quickly increase due to new techniques and approaches in medicine. This is of course hard to predict, but there might already be a trend visible today. All of the parameters above can be multiplied with a growth factor. This factor can also be less than one, e.g. if the population decreases or the symptomatic population decreases due to improvements in prevention or the health care system in general. We also need to take care of future treatments that have an impact on the patient population we are looking at. A new treatment might reduce the number of patients with refractory glioblastoma that we are targeting. Calculating the Market Share Up to now we have calculated the population that might be treated with the drug we are looking at. The difficult part is now to determine how many of these patients we will reach with our drug and how fast we will reach the peak share. The share we reach in the patient population depends on multiple factors. Depending on the indication, the following factors have an influence: • Safety • Efficacy • Origin (biological vs. synthetic) • Duration of action • Time to onset of action • Application modus/patient convenience • Price/reimbursement • Dosing schedule • …

82

Valuation in Life Sciences

These factors represent the properties of the drug. The importance of each single factor depends on the indication and on the unmet medical needs. If for instance we have a new drug that the patients have to take at home which is now a tablet instead of a subcutaneous. Injection, this is an important medical need that is met. The importance of the single properties is best evaluated by interviewing patients, prescribers, and reimbursement authorities. For the price, we will need to create a price – volume curve (elasticity), i.e. how many treatments are we able to sell depending on the price of the treatment. But it is not only the drug that decides how much of it is sold, but largely also the capabilities of the company. Influencing factors here are: • Marketing power of the company • Presence of the company in the target market • Strength of the company in the disease area • Influence of the company on key opinion leaders • Strength of the company brand • Importance of the product in the company pipeline • … All of the factors above influence the share of patients we might reach and how fast. Now, to determine the share, we need to determine the performance of the drug and the company in the respective areas, compared to the treatments that will already be on market when we reach the market and against the treatments, which might follow later on. This can get quite complex as is illustrated in the short example that follows. Let us assume a drug that will reach market in 2 years, drug B. Drug A is already on the market. 5 years later drug C might follow, 10 years maybe drug D. Thereafter our patent expires and we assume that we will have no significant share anymore. The possible scenarios we need to consider are the following: Table 3.7. Different competition scenarios Scenario Drug A Drug C Drug D 1 2 3 4

X X X X

X X

X X

Fundamentals

83

Drug B has to compete with drug A in any case, since it is already on the market. We will then also have to consider the scenarios when either drug C, drug D, or both enter the market. For each of these scenarios we need to determine how large the share of each drug can be. We are only interested in the events where our drug, drug B, will be on the market. And each of the scenarios will have a certain probability defined by the probability that each of the drugs reaches commercialisation or not. We could even assume the cases when drug A would be withdrawn from the market. For every single combination we will perform a scoring based share analysis. The peak sales we then finally expect to reach will be the probability-weighted sum of the different scenarios. But this is only party what we are looking for, because we need to we might have a sequential change of scenarios. We also need to define how fast we reach the penetration depending on the scenario. In the following we will do an example to illustrate how the scoring method is applied to calculate the market share.

Fast uptake

Immediate

Delayed

100%

100%

90%

90%

80%

80%

70%

70%

60%

60%

50%

50%

40%

40%

30%

30%

20%

20%

10%

10%

0%

0%

Slow uptake

0

1

2

3

4

5

6

7

8

9

10

100%

100%

90%

90%

80%

80%

70%

70%

60%

60%

50%

50%

40%

40%

30%

30%

20%

20%

0

1

2

3

4

5

6

7

8

9

10

0

1

2

3

4

5

6

7

8

9

10

10%

10%

0%

0% 0

1

2

3

4

5

6

7

8

9

10

Immediate Fig. 3.1. Sales uptake curves

Delayed

84

Valuation in Life Sciences

Sales Curve The peak sales correspond to the sales once the company has reached its maximum market share. It is unrealistic to assume that a company immediately reaches its full market penetration. We therefore include an uptake curve in the modelling of our sales. This uptake curve describes how fast a company reaches its full market share. This uptake curve typically has two properties. First, we must specify after what period we expect the full share to be reached. In underserved smaller markets/diseases, with a highly unmet medical need, that are well known and organised and where the patients and practitioners are clearly defined this can be as fast as 2 to 3 years, for larger markets with a less clear-cut profile this might take as long as 10 years or more. Second, the uptake can start immediately or it can have some delay, usually after approval the sales only start several months later, and only grow slowly. The figure below displays these two properties graphically. Multiplying the uptake value with the peak sales estimate we receive the sales estimates for each year. Price Elasticity The number of treatments that can be sold depends on the price of the treatment. Doctors hesitate less to prescribe a drug that costs the patient or the insurance only a few dollars. But a drug that costs US$ 30,000 requires a good reason; otherwise the patient or the health insurance will not pay it. This dependence between price and treatment is called price elasticity. Sales

90% 80% 70% 60% 50% 40% 30% 20% 10% 0%

60 50 40 30 20 10 0

2,000

4,000

6,000

8,000

10,000

Price US$

Fig. 3.2. Price elasticity and corresponding sales

0 12,000

Sales US$ mn

Treatments

Elasticity

Fundamentals

85

During market research one tries to find out how this price elasticity looks like by interviewing prescribers, users, and pyors. It is typically a decaying curve. The black line in the figure displays this. Depending on the price practitioners do prescribe the treatment, or prescribe it to at least some patients. Out of the elasticity curve we can calculate the sales figures. In the figure we have assumed that in total 25,000 patients could be treated. We see that with a price set between US$ 4,000 and US$ 4,500 the company can expect the most sales. Case Study: Peak Sales A multi-national pharmaceutical company develops drug Beta against an allergic disease. Drug Beta is an emergency treatment for severe allergic reactions. The patients have a hereditary disease, which causes periodically allergic attacks. Some patients have only one attack per year, some up to ten. At the onset of symptoms, the patients take the drug to avoid an allergic rash of the skin, which is very painful. Our drug is a tablet, which is taken at the onset of the first symptoms. We are now in the process of starting the pre-launch activities and have to calculate the peak sales in a country with a population of 60 mn people. Going trough the medical literature and conducting interviews we come up with the following data. We have seen that due to improvements in diagnosis and understanding of the disease, the treatment rate and diagnosis rate are improving. We estimate that the upper boundary of the diagnosis rate is 80% and 75% for Table 3.8. Data Parameter Population Prevalence (severe attacks) Symptomatic rate Average number of attacks per year Diagnosis rate Access to health care Treatment rate Number of doses per attack Price per dose

Value

Growth rate

60,000,000 1% 15/100,000 1% 100% 0% 7.3 0% 60% 5% (max 80%) 90% 0% 60% 2% (max 75%) 2 0% US$ 4,000 2%

86

Valuation in Life Sciences

the treatment rate. For the sake of simplicity we assume linear increase in diagnosis and treatment rates of 5% and 2% each year, respectively. After one year already 65% will be diagnosed, after two years 70%, and so forth. In reality these increments become smaller the more the rate approaches its maximum. Using the data above we now calculate the number of attacks that are treated per year today and in future years (see Table 3.9). There is currently only one treatment on the market, which is only available in the subcutaneous form, drug Alpha. Patients need the drug to avoid severe allergic reactions and to avoid going to the hospital. All of the treated attacks at present are treated with drug Alpha. The subcutaneous form on the market is a biological with a risk of viral contamination. Our drug on the other hand is synthetically produced and does not have this risk of viral contamination. The onset of action is a bit slower for drug Beta, but we have a slightly higher efficacy. Interviewing the prescribing specialists we find out that the unmet medical needs in this indication are the following: • Oral dosing of the drug (leading need) • No risk of viral contamination (second most important need) Table 3.9. Number of treated attacks Year 2007 2008 2009 2010 2011 2012 2013 2014 2015 2016 2017 2018 2019 2020

Number of treated attacks 21,287 24,549 28,114 32,002 36,232 38,424 40,715 43,110 45,012 46,372 47,772 49,215 50,701 52,232

Fundamentals

87

Efficacy of the present treatment is satisfactory in the eyes of the specialists. Another important point in the view of specialists and payers is the price of the drug. We do expect to see another treatment reaching the market in 5 years, drug Gamma. This will be an oral follow-on treatment of the present subcutaneous drug on the market. We expect that it will have a very similar safety and efficacy profile like our drug. There are currently no other known competitors. We now decide to calculate the possible share of the market using a scoring method. We only analyse the cases where our drug reaches market and compare it to drug Alpha and Gamma. This means we have to assume the following scenarios:

Alpha

Scenario 1

Beta Gamma

Alpha

Scenario 2

Beta

Beta

Scenario 3 Gamma

Scenario 4

Beta

Fig. 3.3. Market scenarios

Each scenario has a certain probability. We are, of course, only interested in the scenarios where Beta reaches the market, which is not certain yet. To understand the potential of Beta, we have to look at the following scenarios: • Scenario 1: Alpha, Beta and Gamma on market • Scenario 2: Alpha and Beta on market • Scenario 3: Beta and Gamma on market • Scenario 4: Only Beta on market

88

Valuation in Life Sciences

We include the possibility where Alpha is not on the market due to the high saftety risk of the drug. This corresponds to a relatively high withdrawal risk. There would also be the scenario that Alpha remains the sole treatment, or that there is Alpha and Gamma, or only Gamma, without Beta. But out of the scenarios with Beta, the four scenarios in the figure are possible. We assume a success rate of 67% for Gamma as a follow-on drug. The withdrawal risk of Alpha is estimated at 20% if Beta reaches the market, and an additional 10% if Gamma reaches the market. This gives rise to the following probabilities. Note that the probabilities have to sum up to 100% – one of these scenarios must happen if Beta reaches the market. Table 3.10. Probabilities of scenarios Scenario

Probability

Scenario 1 Scenario 2 Scenario 3 Scenario 4

57% 13% 10% 20%

We now need to evaluate the performance of our drug in all the above scenarios and want to calculate what share of all treated attacks will be treated by drug Beta. For this we conduct a market study, establish the price elasticity, and determine the most likely price for Beta Second, we must determine the potential peak market share of each drug in each scenario. For this we set up a scoring sheet and come up with the following: Table 3.11. Scoring model for scenario 1

Application form Efficacy Safety Marketing power Importance of drug Total

Weight

Alpha

25% 25% 25% 25% 0% 100%

2 7 2 3 3 3.5

Beta Gamma 10 6 8 7 8 7.75

10 7 8 3 6 7

Fundamentals

89

Table 3.12. Scoring model for scenario 2

Application form Efficacy Safety Marketing power Importance of drug Total

Weight

Alpha

Beta

30% 10% 30% 25% 5% 100%

2 7 2 3 3 2.8

10 6 8 7 8 8.15

Table 3.13. Scoring model for scenario 3 Weight Application form Efficacy Safety Marketing power Importance of drug Total

10% 30% 10% 45% 5% 100%

Beta Gamma 10 6 8 7 8 7.15

10 7 8 3 6 5.55

The weights depend on the scenario. In scenario one we assume that the medical need of a non-infectious and oral treatment has been met. Therefore no property has particular weight. In scenario 2 this medical need has not been solved yet. That’s why safety and application form have a higher weight. In scenario 3 finally we estimate that mainly the marketing power will be decisive between two almost equipotent drugs. For scenario 4 we obviously do not have to calculate a score, because Beta will serve the whole market. The total score should be an indicator of the relative strength of a drug compared to its competitors. For the two scenarios this allows for the following peak market shares: Table 3.14. Market shares

Scenario 1 Scenario 2 Scenario 3 Scenario 4

Alpha

Beta

Gamma

19% 26%

43% 74% 56% 100%

38% 44%

90

Valuation in Life Sciences

This scoring model can be seen on the one hand as alternative to the elasticity model, which also gives an idea how much treatments can be sold. On the other hand it should be in line with the elasticity model, because price and market share are closely related to each other.We now have to model the penetration of the market. This is a dynamic process, especially when several drugs are involved. We assume the following dynamics: Beta quickly gains market share from Alpha due to its superior safety profile and better application form. Gamma, once Beta is on the market only slowly reaches peak market share as the medical need has already been met by Beta and the marketing power of the pharma launching Gamma is weaker than the one of Beta. Alpha

Beta

Gamma

100% 90% 80% 70% 60% 50% 40% 30% 20% 10% 0% 0

1

2

3

4

5

6

7

8

9

10

11

12

13

11

12

13

Fig. 3.4. Market shares for scenario 1 Alpha

Beta

Gamma

1 0,9 0,8 0,7 0,6 0,5 0,4 0,3 0,2 0,1 0 0

1

2

3

4

5

6

7

8

9

10

Fig. 3.5. Market shares for scenario 2

Fundamentals Alpha

Beta

91

Gamma

1 0,9 0,8 0,7 0,6 0,5 0,4 0,3 0,2 0,1 0 0

1

2

3

4

5

6

7

8

9

10

11

12

13

11

12

13

Fig. 3.6. Market shares for scenario 3 Alpha

Beta

Gamma

100% 90% 80% 70% 60% 50% 40% 30% 20% 10% 0% 0

1

2

3

4

5

6

7

8

9

10

Fig. 3.7. Market shares for scenario 4

Once we have established the market dynamics we can then calculate the sales curves for Beta by multiplying its market share with the number of treated attacks, the number of doses per attack, and the price per dose. This gives rise to Fig. 3.7. We see that on average Beta will generate sales up to US$ 300 mn until the dynamics between the market players settles. From then on Beta profits again from the general growth of the market. Discussion The presented case study illustrates that we cannot come up with one peak sales figure without taking account of the different scenarios. Usually the sales depend strongly on the scenario, i.e. on the available alternative

92

Valuation in Life Sciences Scenario 1

Scenario 2

Scenario 3

Scenario 4

Average

600

500

Sales (US$ mn)

400

300

200

100

0 2006

2008

2010

2012

2014

2016

2018

2020

2022

Fig. 3.8. Sales curves

treatments. Especially for late stage projects this involves a lot of interviews with prescribers, payers, and future users and a thorough analysis of the industry pipeline for the indication that is considered. The estimates can then be benchmarked to sales figures of already marketed drugs. For early stage treatments this benchmarking becomes more important, because many of the critical properties such as safety, efficacy profile, and pricing are simply not known yet. Special Topic: The US$ 1 bn Drug Fairytale It takes US$ 802 mn to take one drug successfully to market. This was Joseph DiMasi’s central message in his article in 2003 about R&D costs. This number has been cited many times and has been abused in many ways. US$ 802 mn should discourage any rational investor from investing into drug development. In 2007 DiMasi rerun his study with new data, especially with a new set of success rates. Overall success rates increased over the last years and moved from 21% to 30% from IND to approval, i.e. a drug entering clinical development has a 30% chance to reach market. But to our big surprise the new number on drug development costs even increased to an astonishing US$ 1,241 mn per drug. We expected that an

Fundamentals

93

increase in the success rate translates into lower costs to bring at least one drug to market. To better understand what is behind DiMasi’s figures we will deconstruct his misleading number in detail. DiMasi’s number does not refer to how much a company has possibly to spend to take its drug to the market. Neither does he say how much has been spent on the research and development for drugs that are on the market. No, DiMasi’s numbers are the average expenses it takes to get one drug onto the market, taking into account all possible failures too. So, with a 21% success rate it takes already about 5 clinical phase I projects in order to reach the market with one compound. And in order to inflate the number a little more, DiMasi doesn’t use a forward-looking perspective discounting future expenses, but capitalises all past expenses at the cost of capital of the companies. For each clinical phase we have a mean estimate of costs, success rates, and duration. Furthermore, DiMasi claims that 65% (in 2007, in 2003 it was 69%) of all R&D expenses are spent for clinical trials. This means that per investigational new drug (IND) one has to account for US$ 60 mn (in 2003 only US$ 26 mn) for preclinical expenses. This means that a company has to invest on average US$ 60 mn for R&D to generate one IND, e.g. to bring one drug into clinical development. The table displays the numbers for clinical development DiMasi has used. While the early development costs were high for his 2003 calculation, they become exorbitant for his 2007 calculation. This way the preferential increase of success rates cannot offset the enormous raise of cost assumptions. According to DiMasi the cost estimates are taken from a sample of the ten largest pharmaceutical companies. Of course, this is in no way representative. Biotech companies manage to run clinical trials at much lower expenses, throughout all phases. Table 3.15. DiMasi’s parameters for clinical trials 2003

Phase I Phase II Phase III NDA TOTAL

Costs (US$ mn) 15.2 Success rates 71% Duration (months) 12.3

23.5 44% 26.0

86.3 69% 33.8

0.0 125.0 100% 21% 18.2 90.3

37.7 56% 29.3

96.1 64% 32.9

0.0 166.1 100% 30% 16.0 97.7

2007 Costs (US$ mn) 32.3 Success rates 84% Duration (months) 19.5

94

Valuation in Life Sciences Table 3.16. Attrition weights 2003

Phase I Phase II Phase III

Probability to reach market # it takes that one reaches market

21% 4.7

30% 3.3

69% 1.5

30% 3.3

36% 2.8

64% 1.6

2007 Probability to reach market # it takes that one reaches market

The second point that inflates DiMasi’s number is the fact that costs are not simply added up, but weighted by the success rates. Table 3.16 indicates how many projects it takes in each phase to reach market with one project on average. The last important driver of DiMasi’s number is that he capitalizes all costs at the cost of capital (11% in 2003, 11.5% in 2007). Preclinical costs that lie far in the past are therefore compounded over 10 years at 11%, which means that we do not only account for them 4.6 times because of the attrition weights, but we increase the value by another factor 2.8 because of the capitalisation of costs. An originally relatively reasonable number of US$ 26 mn then contributes US$ 335 mn to DiMasi’s number. The final calculation by DiMasi is shown in the figure. While the 2003 number requires an average drug reaching US$ 310 mn peak sales to break even (which is already far better than the vast majority of all drugs), the 2007 number requires peak sales of US$ 508 mn. It means that one should abandon any project that is estimated to reach sales at its peak of less than US$ 508 mn (using a standard sales curve). We can of course immediately see that most biotechs develop drugs with much smaller peak sales. DiMasi’s figures are as already said in no way usable for other than big pharma companies. DiMasi’s numbers are correct in the way they are defined, but they have two serious flaws. First they are based on a sample of companies that follow the expensive blockbuster model, i.e. only big pharma is included. Therefore the cost estimates are far from being representative. Second, one wants to know how much he must spend, not how much he has spent. We therefore should rather discount the costs back to today than capitalize it from the past to today. Also, when talking about value, we should adjust all costs by their probability. In reality, no biotech will ever get funded by investors with a sum comparable to DiMasi’s number. Biotechs can and do

Fundamentals Preclinical

95

Clinical

1.400 1.200 US$ mn

1.000

629

800 600

466

400

613

200

335

0 2007

2003

Fig. 3.9. DiMasi’s figures

bring drugs on the market for less than US$ 100 mn. DiMasi’s number is only interesting for multinational pharmaceutical companies hunting for blockbusters. With this strategy currently failing and companies moving to targeted and personalized medicine, niche and orphan indications, the costs for a successful drug launch need to be much lower resulting from much higher success rates. The table displays some more realistic assumptions about drug development. While we stick to DiMasi’s 2003 numbers for success rates and duration, we have adapted the clinical costs according to our experience. The last table finally shows the various values we get out of these three data sets. We focus in this analysis only on the clinical costs. As we can see, the eye-catching US$ 629 mn clinical spendings of the 2007 data set can also be interpreted as US$ 79 mn. We however think that this number can even be lowered to US$ 25 mn, the risk adjusted present value of all investment to come for one specific drug. The pharmaceutical companies might be interested in advertising DiMasi’s high number, entrepreneurs Table 3.17. More realistic assumptions Phase I Phase II Phase III

NDA

TOTAL 88

Costs (US$ mn)

5

12

68

3

Success rates

71%

44%

69%

100% 21%

Duration

12.3

26

33.8

18.2

90.3

96

Valuation in Life Sciences Table 3.18. Indicative numbers out of the three datasets 2003

2007

Avance

1

Capitalized and attrition weighted

466

629

261

Attrition weighted (out of pocket)2

274

364

162

Summed up3

125

166

88

59

109

36

45

79

25

311

509

231

Risk Adjusted (out of pocket)

4

5

Discounted and risk adjusted 6

Required peak sales

and investors alike might be more interested in the realistic value of the costs for one project (the US$ 25 mn). The required peak sales allow judging which data set makes most sense. Discounting

Theory and Reality Discounting contains two components, time value and risk. While the time value of money in life sciences is not different to any other industry, the risk of drug and medical device development is special. The CAPM relies on market data. The CAPM uses the famous β to deduce the cost of capital. In theory, the cost of capital should decrease with advanced stage of the company, as the volatility of the company decreases. Nevertheless the statistical βs show sometimes quite a different picture. For some companies the β is even negative. These non-sense βs are a consequence of the low liquidity of most biotech stocks and additional statistical noise. While

1 2 3 4

5

6

DiMasi’s approach DiMasi’s approach but without capitalisation of costs All costs summed up as they are, without adjustment for risk nor time All costs adjusted by the probability that they occur (i.e. multiplied by previous success rates) All costs adjusted for their risk and discounted at the cost of capital (11%) back to IND Minimum peak sales to reach an IRR higher than the cost of capital (preclinical costs included)

Fundamentals

30%

Market Reality

20% 10%

97

MCPM

CAPM

Early stage

Clinical stage

Market

Pharma

Fig. 3.10. Cost of capital as indicated by models and the market

CAPM is a theoretically sound model, it is of little use in practice, as market liquidity and sample size make the statistical measure completely unreliable. MCPM takes care of the fact that a company faces specific risks. The discount rates calculated with MCPM are similar to the CAPM rates for large pharmaceutical companies. This is in line with what we have discussed above; large pharmaceutical are well-diversified and they do not depend excessively on the success of few projects. MCPM then indicates discount rates between 30% and 40% for smaller biotech companies that do not have any product on the market. The discount rates observed in the market, either in valuations for private capital rounds or in analysts’ reports, show a picture in between the two presented theories. While the discount rate of diversified companies comes close to the theoretical cost of capital calculated with CAPM, less diversified and earlier stage companies have to pay their investors a spread to the CAPM rate. Nevertheless, this spread is not as high as MCPM suggests. This confirms our caveats regarding CAPM and MCPM. Table 3.19 displays published discount rates used by companies. Proposed Framework The most critical risk of drug development companies is doubtlessly high attrition. MCPM accounts for this issue, even if in a rather intuitive way. We do not agree with the MCPM idea that shareholders must be able to insure against possible losses. We therefore approach the discount problem in a way that considers the very basic challenge of discounting: the inclusion of risk aversion.

98

Valuation in Life Sciences

We decompose a company’s risk map into risks every company has to deal with and risks that are specific for a company, like a fingerprint of the company. The first category includes financial market risks like interest rates or exchange rates, employee turnover, human errors, war, and other operational risks. Apart from financial risks, it is almost impossible to model these risks. Most operational risks are soft and remain not quantifiable. Nevertheless, these risks are well reflected in an index and CAPM can be used to calculate the discount rate that compensates for this kind of risks. Company specific risks include attrition, cluster risks due to over-reliance on one technology, and market risks linked to single projects or markets Table 3.19. Discount rate comparables Company

Discount Rate

Source

Lilly

18.75%

Annual Report 2004

AstraZeneca AstraZeneca

Rate to discount acquired IP R&D 15% Integrated drug development company with one major marketed product 18% One anti-infective in phase III Risk free rate Risk neutral valuation 8% Projects partnered with AZ

CAT

12.5%

Projects owned by CAT

Genentech Genentech Jerini MedImmune

16%-19% 20%-28% 15% 11.3%

1999 1990 One anti-allergic in review Fully integrated drug development company Vaccine company, CAPM + small company spread

Actelion

Arpida

Berna Biotech 9.9%

Merck KGaA 7%–7.6% Schering Schering

13.5% 14.25%

Chemical company with pharma business Region Europe Region USA

Annual Report 2004

UBS Analyst Report Annual Report 2004 Cash offer by AZ for CAT Cash offer by AZ for CAT Webpage Webpage CS Analyst Report Annual Report 2004 Merger memorandum Crucell-Berna Biotech by PWC Annual Report 2005 Annual Report 2005 Annual Report 2005

Fundamentals

99

Table 3.20. Risks of a drug development company and how to consider them Risks

Shared by the industry

Description

Daily business risk, Characteristic exposures of a company mainly operational like attrition or sales numbers/potential of single products Difficult, because Attrition is modeled with success rates, mostly soft factors market risk is modeled with binomial trees or simulations Base business risk Company specific risks that depend on the that must be pipeline of the company rewarded Base rate Spread

Modeling

Effect on discount rate Part of discount rate Quantification

CAPM applied to industry index

Company specific

EUT applied to pipeline

the company operates in. These risks are well suited for quantitative modeling, even if this is a challenging undertaking. The spread we have to add for this company specific risks must be calculated on the basis of a quantitative model. The less diversified a company, the riskier it is and the more an investor wants to be compensated for taking this risk. The theory that deals with taking risks and how to compensate for risks is called Expected Utility Theory. Analysing the company specific risks using EUT allows calculating a corresponding discount rate that accounts only for the modeled risks. The model can then be calibrated on the market. This discount rate has to be added to the base rate we have calculated using CAPM. The final discount rate then becomes: discount = risk freerate + spread (CAPM ) + spread ( EUT )

(3.1)

Correctly implemented, such a model allows the quantification of a company’s risk profile. Intuitively, a company that has further advanced projects has a more favourable risk profile than an early stage company. Similarly, a company with licensed projects is a less dangerous investment than a company that conducts all projects on its own, putting positive cash flows much more at risk. A quantitative model like the one described above delivers the following discount rates (see Table 3.21). The table describes the projects in the pipeline (column 1-4) and the corresponding discount rate (column 5). The more advanced a pipeline is, the

100

Valuation in Life Sciences Table 3.21. Discount rates (source: Avance)

Project 1

Project 2

Project 3

Project 4

Discount

NDA ($ 500) NDA ($ 300) Ph. 3 ($ 500) Ph. 3 ($ 300) Ph. 3 ($ 500) Ph. 3 ($ 500, lic) Ph. 3 ($ 300) Ph. 3 ($ 500) Ph. 3 ($ 500, lic) Ph. 3 ($ 300) Ph. 2 ($ 500) Ph. 2 ($ 500) Ph. 1 ($ 500) Ph. 3 ($ 500) Ph. 3 ($ 500) Ph. 3 ($ 500)

Ph. 3 ($ 500) Ph. 3 ($ 300) Ph. 3 ($ 500) Ph 3 ($ 300) Ph. 2 ($ 500) Ph. 2 ($ 500, lic) Ph. 2 ($ 300) Ph. 1 ($ 500) Ph. 1 ($ 500, lic) Ph. 1 ($ 300) Ph. 1 ($ 500) Prec. ($ 500) Prec. ($ 500) Ph. 1 ($ 300) Ph. 1 ($ 300) Ph. 1 ($ 300)

Ph. 2 ($ 500) Ph. 2 ($ 300) Ph. 1 ($ 500) Ph. 1 ($ 300) Ph. 1 ($ 500) Ph. 1 ($ 500, lic) Ph. 1 ($ 300) Prec. ($ 500) Prec. ($ 500, lic) Prec. ($ 300) Prec. ($ 500) Prec. ($ 500) Prec. ($ 500)

Ph. 1 ($ 500) Ph. 1 ($ 300) Ph. 1 ($ 500) Ph. 1 ($ 300) Ph. 1 ($ 500) Ph. 1 ($ 500, lic) Ph. 1 ($ 300) Prec. ($ 500) Prec. ($ 500, lic) Prec. ($ 300) Prec. ($ 500) Prec. ($ 500) Prec. ($ 500)

Prec. ($ 300) Prec. ($ 300)

Prec. ($ 300)

12.0% 12.0% 13.4% 13.8% 15.1% 14.2% 16.7% 17.3% 15.3% 18.5% 19.3% 19.5% 19.9% 19.5% 19.2% 18.4%

lower is its discount rate. We can also see that licensing reduces the discount rate. The risk profile is furthermore affected by the disease category; success rates vary across disease categories. A company specialised in cardiovascular diseases has a lower discount rate than a company with a comparable pipeline in women’s health, because success rates for CV drugs are higher. The discount rate also depends on the number of projects in the pipeline. Volatility

The volatility is an input parameter for real options valuation. It is generally understood as an indicator of the underlying’s uncertainty. Since real options valuation is a descendant from financial options valuation we first look at the definition and measurement in finance. Then draw up the particularities of volatility in drug development projects. In quantitative finance the volatility is usually measured as the standard deviation of the daily returns (or better: log-returns) of an asset. The standard deviation indicates by how much the asset price fluctuates in a certain

Fundamentals

101

140 120 100 80 60 40 20 0 0

50

100 20%

150

200

250

50%

Fig. 3.11. Two simulated stock price curves using different volatilities

time. A volatility of 0 corresponds to an asset with no uncertainty, i.e. the asset is completely predictable. The higher the volatility, the less predictable the asset. Indices like the Dow Jones or FTSE 100 have a relatively low volatility (10%-20%) because of their diversification. Single shares on the other hand are affected by company specific events and can have considerably higher volatilities. Volatilities of 50% to 80% are not rare for biotech companies. The volatility measures the fluctuations or the uncertainty of an asset and its price. If the volatility refers to the past changes of the share price we call it historical volatility. In the continuation we are more interested in the future volatility, a measurement of the expected changes of the share price. In general we can say that the volatility is driven by the change of appreciation of the asset’s price by the investors. Various factors influence this appreciation: 1. Uncertainty about results and performance of the company 2. Risk profile of the company 3. Appreciation of the industry of the company and the market in general An exporting company depends naturally on foreign exchange rates. Oil companies depend on the oil price. And all companies depend on the way

102

Valuation in Life Sciences

they seize business opportunities, business scandals and their corporate governance. Some components of the volatility can be attributed to the market and the industry, others are particular to the company. The volatility of a biotech share is mainly driven by the uncertainty of future clinical trials. A positive result leads may let the share price sky rocket, a negative result can mean the end of the company. In real options valuation we require the volatility of the underlying, i.e. the expected peak sales. An analysis of this volatility is jeopardised by the natural life cycle of a product that clearly motivates a large part of changes in sales numbers. An analysis of the volatility must therefore distinguish between changes due to the sales curve and changes due to a changed market potential. Factors different from natural growth and the life cycle that influence the sales numbers comprise: 1. General state of the economy 2. Competition 3. Regulation The volatility of a drug is influenced by the drug characteristics and its position in the market. Fashion drugs (nice to have drugs) depend more on the purchasing power of patients than life saving drugs (need to have drugs). The rise of new competition can seriously impact the sales numbers of a drug. Alternatively, competitors might lose sales due to newly discovered side effects. National health care policies can boost or slow down the sales numbers. Furthermore, we have to consider that a first-inclass drug is likelier to receive competition with improved efficacy, less side effects or increased application convenience. Nevertheless, the uncertainty of sales is rather moderate once the drug is on the market. We estimate the post-development volatility of a drug between 10% and 20%. The above discussed volatility measures the change of the sales potential assuming that the drug and its qualities is known. The volatility that matters for real options valuation describes the uncertainty of the peak sales estimate before commercialization, i.e. when some properties of the drug are still not fully known. The peak sales estimate obviously also depends on the same factors like the post-development volatility, but there are some additional risk drivers. The clinical development reveals important information concerning safety, efficacy, dosage, further applications and patient convenience and allows a benchmarking to other therapies. All these factors can have a positive or a negative impact on the sales potential of the drug. Once clinical development is finished, these issues are well explored. The pre-commercialisation volatility is therefore higher.

Project Valuation

103

We know that in drug development about 30% of all abandonment is caused by economic reasons, i.e. lack of profitability (Dimasi 2001). This is exactly what real options model: the ability to halt a project because it does not make any economic sense. We can therefore adjust the volatility in a way that for a standard drug development project 30% of abandonment is economically motivated. This leads to a volatility of 25% to 35%. Accounting for special factors that impact the estimation of peak sales like a novel mechanism of action, an undeveloped market for a new indication, or a clear competitive environment, we finally arrive at a range of 20% to 50% volatility for the peak sales estimate. The upper end representing a novel approach with many uncertainties concerning indication and target population.

Project Valuation In this section, we explain how to value life sciences projects. The theory on project valuation will lay the grounds for the subsequent chapters; license contract, IP, and firm valuation follow the same principles. In the following, we designate drug or medical device research and development prior to market launch as projects, once on the market as products. The time to market can range from months to years. Typically, there are still future cash flows occurring related to the project, and the projects can be subject to attrition, i.e. being unsuccessful or simply being stopped. Project valuation matters for: 1. Project managers negotiating the budget 2. Portfolio managers reviewing the pipeline 3. Business developers and licensing managers negotiating a license or collaboration deal 4. Investors selling or buying companies 5. Analysts reviewing companies Projects managers have to apply for their budget on a regular basis. Budget debates tend to be highly political and the project managers compete with each other to get as much money as possible. Naturally, they want to present their projects as priority to the company, i.e. as more valuable than the competitor projects. A project manager must therefore understand what parameters drive the projects value and find good justifications why those should be so favourable in his project.

104

Valuation in Life Sciences

Business developers and licensing managers need to value a potential project for licensing from two sides. First, they have to value the project as if it would be developed as a proprietary drug in-house, in this case as a project. The licensor’s value is the minimum price of the license contract; the licensee’s value the maximum price. Second, they have to value the license contract with the proposed license terms to make sure they get a fair deal, i.e. the price of the license contract must lie between the lower and upper limit. We will discuss the implications of valuation to licensing and negotiation in the next sections. Portfolio managers on the other hand do not have any interest in an optimistic valuation of the pipeline. They need to compare the projects in the pipeline objectively in order to optimally allocate resources, to put projects on hold, or even to abandon some. Because project managers have personal incentives in overvaluing their projects, portfolio managers must therefore critically question all parameters when reviewing the pipeline. Investors have various motivations to value a project. Either they want to invest directly into a project or in a basket of projects, i.e. in a company. When investors invest in a company, they try to negotiate a low valuation of the company so they do not have to pay the full price for it. Venture capitalists for example are in the comfortable position to distribute money to clients looking for money. As they are in the dominant negotiation position, they can value the company below its real value and offer the money on a take-it-or-leave-it basis. This advantage disappears as soon as several venture capitalists compete for the offered stake. Once the investors want to sell their investment, they want to get the highest possible price. The valuation then needs to be at the higher end. Analysts will value single projects to get the value of the entire company, just as investors do. They will not expose themselves when performing this task. Consequently, many analysts’ recommendations are alike and provide no new information. An analyst glorifying a young company that goes bankrupt the next month will most likely have to look for a new position. On the other hand, an analyst that is wrong with his prediction but has followed the herd can claim to be at least not the only one. The reader might be surprised about the view of optimistic, realistic, or low valuations. However, the fact that people have different goals when valuing a project leads to different values. A project will not represent the same value to the buyer or the seller. The fact that different interests influence the valuation result needs to be kept in mind when handling the numbers.

Project Valuation

105

We will first discuss how projects are valued with DCF in theory and in the first part of case study 1. We will then look at how to value projects with ROV and discuss the second part of case study 1. Project Valuation: DCF

To value a project we must first define all cash flows related to the project and then determine the applicable discount rate. Once these data are assessed, we can proceed to the actual valuation. Cash Flows Cash flows are characterised by time, size, and probability to occur. Time. During research and development, cash flows are the costs of the individual phases. Mostly, these costs are spread over the whole length of a phase. Once the product approaches commercialisation, the company has to invest in preparing product launch, sales force, and production facilities. Launch costs can incur over several years, sometimes starting even earlier than launch. After launch, the cash flows are linked to sales revenues and operating expenses. Size. All cash flows are negative to the exception of sales revenues and licensing revenues. The sales development is modeled by the peak sales estimate, the growth rate, and the sales curve. The peak sales estimate corresponds to the maximum sales revenue of the product if it would already be on the market today. A product first needs some time to achieve its maximum market share. Therefore we model the sales dynamics with a sales curve, starting at 0% and reaching 100% at the time when the product reaches its maximum market penetration. We account for the difference between today’s market and the market at time of maximum penetration by applying a growth rate. The growth rate displays the annual average increase of the peak sales estimate due to economic, industry specific, and demographic growth. revenuet = SalesCurvet (1 + µ ) 0 PeakSales

(t − t )

penetration at time t

(3.2)

growth adjustment

The sales revenues are reduced by the costs of goods, marketing, distribution, and administrative costs. These reductions are summarised in the operative margin.

106

Valuation in Life Sciences

Probability to occur. After each phase, it is possible that the project has to be abandoned. The success rate of each phase indicates with what probability the project continues to the next phase. This means that only the cash flows of the current development phase are certain, all others occur with the probability that the previous phases are passed successfully. The cash flows of the second phase only happen with the probability that the first phase is successful, i.e. with the success rate of the first phase. The cash flows of the third phase then only happen with the probability that the first and second phase are successful, i.e. the success rate of the first phase times the success rate of the second phase. The cash flows related to commercialisation of the product happen with the probability that the project passes all prior R&D phases successfully. We have now identified all cash flows related to the project to value and defined their time, size, and probability. The next input parameter that we have to determine is the discount rate. Discount Rate With the discount rate we account for the time value of money and the risk. Time value. Typically, the time value of money is the risk free interest rate, which is the rate offered in state issued securities (e.g. T-Bills). The discount rate is composed of the current risk-free intertest rate and a risk premium on top of that. Risk. Generally, risk summarises all sort of uncertainty regarding the cash flows of the project. These encompass technical uncertainty, i.e. the danger that the project must be halted at some stage, the fluctuation of sales revenues and foreign exchange rates, and other deviations from the valuation assumptions. There is no generally accepted theory how to determine the discount rate. Some say that one should apply a different discount rate for every project, others argue for a uniform discount rate on company or portfolio level. It depends what the purpose of the valuation is. An in-licensed product will most often not be discounted with the in-house discount rate but with a higher one to take in consideration the information asymmetry. A portfolio manager will discount all projects with the same discount rate of the company to get a uniform valuation. We propose a framework to define the discount rate as outlined in the section on discounting in life sciences. The number depends on the risk profile of the company, its pipeline, and its projects and allows a very good quantification of these parameters.

Project Valuation

107

Valuation Once we have assessed all input parameters, we can proceed to the actual valuation. Using the discounted cash flows method we must: 1. Adjust each cash flow by its probability to occur. This means that we multiply each cash flow with its probability (P). 2. Discount each cash flow back to the present or to the date we want to know the value for. 3. Sum up all discounted and probability adjusted cash flows to the net present value. In mathematical terms this reads:

NPV =

∑ (1 + r )

− ti

PCFt CFti i

1

2

(3.3)

ti

3

If the net present value is positive, then the valuation suggests continuing the project. A positive value means that the internal rate of return of the project is higher than the hurdle rate, i.e. the discount rate. On the other hand, a negative value should lead to abandonment of the project, since the invested capital yields not the required return. Case Study Project Valuation: DCF Dr. Project Manager loves his cardiovascular drug research project he has just successfully carried trough preclinical testing. The results are encouraging and allow planning for the phase 1 clinical trial. Dr. Project Manager now has to apply for his budget. In order to convince the management that his project is worthwhile, he wants to value it. He knows that the mid-size pharma company he is working in has a very strong clinical research pipeline with good projects, but the company is lacking the cash to fund all of them. Dr. Project Manager wants to show that his project has a high value and an attractive return on investment (ROI). For this he follows the above- described schedule and first defines the cash flows with respect to size, probability, and time. Second, he determines the discount rate, and third, he proceeds to the actual calculation.

108

Valuation in Life Sciences

Cash flow size. Dr. Project Manager draws up the budget for the next phases. As a starting point, he wants to figure out all the to bring the project on the market. After consulting his colleagues in the clinical trial and marketing department he comes up with the following cost assumptions for his project: Table 3.22. Project development costs Clinical Phase 1 Clinical Phase 2 Clinical Phase 3 Approval Launch

$ 9 mn $ 22 mn $ 80 mn $ 3 mn $ 120 mn

The launch costs include the construction of a new production plant and the establishment of a new marketing department. Still left is the estimation of the revenues the product will achieve. Dr. Project Manager thinks that based on the limited knowledge about the efficacy profile of the drug, it is not possible to predict the peak sales yet. He decides to look at all the comparable drugs that are now on the market and to calculate the average sales his project might achieve compared to the present market. The average sales for comparable marketed cardiovascular drugs in 2004 were $ 486 mn. It is still early to determine the sales with a bottom-up approach, therefore Dr. Project Manager opts for a comparative approach. However, he feels that the average number is too optimistic for his project. Therefore he prefers to use an estimate that is not distorted by the unrealistic scenarios to achieve a blockbuster. For this he can use either the median of comparable product sales or the most likely sales. He decides for the most likely sales to exclude too optimistic scenarios and the distortion by niche and under-performing products. The most likely peak sales amount to $ 241 mn. Dr. Project Manager is not sure whether the general market growth between 5% and 10% can be applied to his project. The market growth could very well be linked to the increasing number of therapies available, and for a single product, this market growth could have even negative substitution effects. He therefore decides to use a neutral 0% growth. After defining the peak sales, he still has to elaborate when the peak sales will be reached and when competition sets in. Again, due to limited knowledge on the project, he chooses to use the following standard sales curve for a drug with patent expiry after 12 years:

Project Valuation

109

100% 90% 80% 70% 60% 50% 40% 30% 20% 10% 0% 1

2

3

4

5

6

7

8

9

10 11 12 13 14 15 16 17 18 19 20 21 22

Fig. 3.12. Standard sales curve

Finally, Dr. Project Manager must reduce the sales revenues by the operating costs, which involve costs of goods, costs of marketing and sales, and some general and administrative costs. Whet will be left of the sales revenues is the operating margin. Colleagues from the financial department suggest him to use 65%. Cash flows probability. With this Dr. Project Manager has the size of the future cash flows. He now wants to know how likely they will occur. Therefore, he needs to find out the probabilities of passing the next development phases and obtaining regulatory approval. Studying the literature, he comes up with the following success rates (Kola et al. 2004): Table 3.23. Project success rates Clinical Phase 1 Clinical Phase 2 Clinical Phase 3 Approval

63% 42% 75% 95%

Cash flows timing. Having found the costs and the success rates, Dr. Project Manager still needs to define the timelines for the project. Again, he consults his colleagues and studies the literature. The timeline he assumes for his valuation are:

110

Valuation in Life Sciences Table 3.24. Development time Clinical Phase 1 Clinical Phase 2 Clinical Phase 3 Approval

1 year 2 years 3 years 2 years

He assumes that the clinical costs are equally spread over the length of the phase; the costs for approval are due at the beginning of the approval phase. Launch costs are distributed over three years, one third before launch, one third in the first year, and one third in the second year. Discounting. Dr. Project Manager decides now that his valuation will be based on a standard rNPV calculation. Therefore, he needs to define the discount rate he will apply to the calculations. Not sure about this, he considers if he should take a discount rate specific for his project, or the success rate the company applies generally. Still not smarter after having studied a textbook on valuation he calls up his friend Prof. Corporate Finance. The professor tells him that there is no sound theory on how to adjust the discount rate for project risks. However, he suggests using the company discount rate, as this is certainly lower and yields therefore a higher value. Dr. Project Manager therefore goes ahead with the company internally used discount rate of 14%. Solution Case Study Project Valuation: DCF Determination of cash flows. We deduce the cash flows from the assumptions and attribute them to the right period. The R&D expenses are spread uniformly over the length of the corresponding phase, except the expenses for filing an NDA. The launch costs are distributed over three years, onethird before launch, the other two thirds in the two first years of commercialisation. We then need to compute the sales revenues. In a first step we must know the peak sales in each corresponding year. The peak sales the product achieves in the market in i years from now are: PeakSales(i ) = PeakSales * (1 + µ )

i

(3.4)

In our particular case the growth rate is 0%. We therefore have the same peak sales estimate until launch. By way of example we assume the growth rate to be 5%, in that case the peak sales in year 9 would be: PeakSales(9) = $241mn * (1 + 0.05) = $373.87 mn 9

Project Valuation

111

To deduce the sales revenues from the peak sales we must multiply them with the sales curve: Salesrevenues(i ) = PeakSales(i ) ∗ Salescurve(i )

(3.5)

We are now only missing the operational expenses, defined as follows: OpExpenses(i ) = Salesrevenues(i ) ∗ (100% − MARGIN )

(3.6)

This gives us the final table with all cash flows: Year 1

Year 2

Year 3

Year 4

Year 5

Year 6

Year 7

Year 8

Year 9

Year 10

Phase 1

Phase 2

Phase 2

Phase 3

Phase 3

Phase 3

Approval

Approval

Market

Market

Sales curve Peak Sales

$ 241 mn

$ 241 mn

$ 241 mn

$ 241 mn

$ 241 mn

$ 241 mn

$ 241 mn

$ 241 mn

5%

19%

$ 241 mn

$ 241 mn

Sales revenues

$ 12 mn

$ 46 mn

Operating expenses

($ 4 mn)

($ 16 mn)

R&D Expenses

($ 9 mn)

($ 11 mn) ($ 11 mn) ($ 27 mn) ($ 27 mn) ($ 27 mn)

($ 3 mn)

($ 40 mn) ($ 40 mn) ($ 40 mn)

Year 11

Year 12

Year 13

Year 14

Year 15

Year 16

Year 17

Year 18

Year 19

Year 20

Market

Market

Market

Market

Market

Market

Market

Market

Market

Market

Sales curve

36%

51%

64%

75%

84%

91%

96%

99%

100%

100%

Peak Sales

$ 241 mn

$ 241 mn

$ 241 mn

$ 241 mn

$ 241 mn

$ 241 mn

$ 241 mn

$ 241 mn

$ 241 mn

$ 241 mn

$ 87 mn

$ 123 mn

$ 154 mn

$ 181 mn

$ 202 mn

$ 219 mn

$ 231 mn

$ 239 mn

$ 241 mn

$ 241 mn

Sales revenues Operating expenses

($ 30 mn) ($ 43 mn) ($ 54 mn) ($ 63 mn) ($ 71 mn) ($ 77 mn) ($ 81 mn) ($ 84 mn) ($ 84 mn) ($ 84 mn)

R&D Expenses Year 21

Year 22

Year 23

Year 24

Year 25

Year 26

Year 27

Year 28

Year 29

Year 30

Market

Market

Market

Market

Market

Market

Market

Market

Market

Market

Sales curve

99%

97%

95%

72%

53%

37%

24%

13%

6%

1%

Peak Sales

$ 241 mn

$ 241 mn

$ 241 mn

$ 241 mn

$ 241 mn

$ 241 mn

$ 241 mn

$ 241 mn

$ 241 mn

$ 241 mn

$ 239 mn

$ 234 mn

$ 229 mn

$ 174 mn

$ 128 mn

$ 89 mn

$ 58 mn

$ 31 mn

Sales revenues Operating expenses

($ 84 mn) ($ 82 mn) ($ 80 mn) ($ 61 mn) ($ 45 mn) ($ 31 mn) ($ 20 mn) ($ 11 mn)

$ 14 mn

$ 2 mn

($ 5 mn)

($ 1 mn)

R&D Expenses

Fig. 3.13. Calculation of project value

Probability of cash flows. For each year we compute the probability that the project is still alive. During the first phase the project is certainly alive. The second phase is only achieved if the first phase is successfully passed, i.e only achieved with the success rate. Prob(Phase(i))=Prob(Phase(i-1))*SuccessRate(Phase(i-1))

(3.7)

In particular we have for phase 1 a probability of 100%, for phase 2 a probability of 100%*63%=63%, then for phase 3 a probability of 63%*42*=26% and so on. We then calculate the risk adjusted net cash flows (rnCF); we sum up all cash flows of the same period and multiply with the probability that the project is still alive in that period. rnCF(i)=Prob(i)*(SalesRevenues(i)-OpExpenses(i)-R&DExp(i))

(3.8)

112

Valuation in Life Sciences

For year 9 this is: rnCF(9)=Prob(9)*(SalesRevenues(9)-OpExpenses(9)-R&DExp(9)) (3.9) 19%*($12 mn-$4 mn-$40 mn)=-$6 mn Discounting of cash flows. We discount all risk adjusted net cash flows back to the valuation date; we now have the risk adjusted net present cash flows. rnpCF(i)=rnCF(i)*(1+discount)-i

(3.10)

Again for year 9 this would be: -$6 mn*(1+0.14)-9=-$2 mn The risk adjusted net present value (rNPV) finally is the sum of all risk adjusted net present cash flows. rNPV =

∑ rnpCF (i )

(3.11)

i

Length (y) Costs

Suc. Rate

Phase 1

1

$ 9 mn

63%

Phase 2

2

$ 22 mn

42%

Phase 3

3

$ 80 mn

75%

Approval

2

$ 3 mn

95%

Launch costs

$ 120 mn

Peak Sales

$ 241 mn

Growth Rate

0%

Operating Margin

65%

Discount Rate

14% Year 1

Year 2

Year 3

Year 4

Year 5

Phase 1

Phase 2

Phase 2

Phase 3

Phase 3

Year 6

Year 7

Year 8

Phase 3 Approval Approval

Sales curve Peak Sales

Year 9

Year 10

Year 11

Market

Market

Market

5%

19%

36%

$ 241 mn $ 241 mn $ 241 mn $ 241 mn $ 241 mn $ 241 mn $ 241 mn $ 241 mn $ 241 mn $ 241 mn $ 241 mn

Sales revenues

$ 12 mn

Operating expenses

($ 4 mn) ($ 16 mn) ($ 30 mn)

R&D Expenses Success Rate Probability Risk adjusted net CF Discount

($ 9 mn) ($ 11 mn) ($ 11 mn) ($ 27 mn) ($ 27 mn) ($ 27 mn)

$ 46 mn

$ 87 mn

($ 3 mn) ($ 40 mn) ($ 40 mn) ($ 40 mn)

63%

100%

42%

100%

100%

75%

100%

95%

100%

100%

100%

63%

63%

26%

26%

26%

20%

20%

19%

19%

100% 19%

($ 9 mn)

($ 7 mn)

($ 7 mn)

($ 7 mn)

($ 7 mn)

($ 7 mn)

($ 1 mn)

($ 8 mn)

($ 6 mn)

($ 2 mn)

$ 11 mn

100%

88%

77%

67%

59%

52%

46%

40%

35%

31%

27%

Net present CF

($ 9 mn)

($ 6 mn)

($ 5 mn)

($ 5 mn)

($ 4 mn)

($ 4 mn)

($ 0 mn)

($ 3 mn)

($ 2 mn)

($ 1 mn)

$ 3 mn

rNPV

$ 2.9 mn

IRR

14.6%

Fig. 3.14. rNPV project valuation

The spreadsheet summarises all calculations we have performed so far to calculate the rNPV of the project.

Project Valuation

113

Project Valuation: Decision Tree

The calculation of the project value with DCF is straightforward and fast once we have defined all cash flows. We must first multiply the cash flows with their probability, second discount them back to the valuation date, and third sum them all up. The clarity of the method is, however, to the expense of a relative rigidity. In many project plans it is not known beforehand which development path will be taken. Much of the later development depends on results of previous trials and studies. In DCF we only consider success or failure as possible outcomes of a trial. In reality, trial results are not always as clear-cut as this suggests. Sometimes it becomes clear out of the results that the trial design was deficient, or researchers suspect a type two error (the hypothesis that should be proven with the trial is not supported by the data, although it is actually true). Depending on the data gathered in a trial the design, cost, and success rate of the next clinical study can vary. Also commercialisation is not just one possible scenario as we have already seen in the section about peak sales. Usually the sales estimates depend on other available treatments. Still in development we often do not know which competitors we will face once we are on the market, and which competitors will join us later on the market. If we have already a good idea of what potential scenarios might await us, then decision tree analysis is a good way to calculate the value of a project considering all these possible scenarios. In a first step we must define the scenarios tree, i.e. where along the project plan the development can take different directions. Usually these decision points are placed at the end of a phase. Risk adjusted DCF is the simplest form of a decision tree, which has at the end of each phase the scenarios of continuation and abandonment.

pND p3 p2

1-p2

Review 1-pNDA

Phase 3 1-p3

Phase 2

Market

Abandon

Abandon

Abandon

Fig. 3.15. Decision tree as supposed in a risk adjusted DCF valuation

114

Valuation in Life Sciences

Fig. 3.16. Decision tree for project Cytotoxabine with durations, costs, and success rates

Project Valuation

115

A decision tree can have various forms and there is no standard way to construct it. Usually each project has its particularities, offering special scenarios. The only constraint is that the probabilities of the different scenarios at each decision point must add to 100%. We will explain a decision tree analysis with the example below. Case Study Project Valuation: Decision Tree Analysis Mr. Investor has just come back from a board meeting of one of his portfolio companies, where they discussed the development plan of the project Cytotoxabine. To get more clarity about the economic rationale of the project he wants to perform a valuation himself. First he sketches the development plan. Right now the project enters phase 2 trials for 2nd line treatment and combination with the gold standard Onmarketabine of chronic lymphatic leukemia (CLL). Mr. Investor assumes that in 50% the drug has to be abandoned and in 35% the drug can be taken to phase 3. Since the trial is set-up with relatively few patients he expects that in the remaining 15% the data will be encouraging enough, but not statistically significant, calling for a second phase 2 trial. This second phase 2 trial is expected to have a higher success rate than the first one, since it is already based on some evidence. After a successful phase 3 study the drug is moved to review phase and in parallel a second phase 3 trial for first line treatment will be initiated. Once on market, Cytotoxabine is expected to reach sales of US$ 150 mn as a second line treatment and additional US$ 100 mn with the approval as first line treatment. Launch costs amount to US$ 30 mn for each indication. In case a second phase 2 trial is necessary, patent protection on market will be two years shorter, i.e. only 9 instead of the original 11 years. Mr. Investor now gradually starts to resolve the tree from its endpoints. The value of the abandoned scenarios is US$ 0. For the rest he has to calculate the scenarios of second line indication and first line indication, for each 9 and 11 years of patent protection (first line treatment has again 3 years less patent protection). These scenarios correspond to a DCF valuation of the project if it were at launch. Mr. Investor calculates these four scenarios using a 70% operating margin and a discount rate of 17%. He then comes up with the following tree (Fig. 3.17). Mr. Investor can now work back the decision tree from the end. The value of the project at the decision points is the average value of the value of the branches.

116

Valuation in Life Sciences US$ 324 mn

US$ 0 US$ 180 mn

US$ 0 US$ 0

US$ 0

US$ 300 mn

US$ 0 US$ 153 mn

US$ 0 US$ 0

US$ 0

US$ 0

US$ 0

Fig. 3.17. Values for leafs of decision tree for Cytotoxabine US$ 324 mn

US$ 162 mn

US$ 0

US$ 0

US$ 270 mn

US$ 138 mn

US$ 0

US$ 0

US$ 0

US$ 0

Fig. 3.18. Values for last decision points of decision tree for Cytotoxabine

As a next step, Mr. Investor must know how much the project is worth at the start of review phase for first line treatment, the next earlier decision point. This value is the average value after the review phase, discounted for the elapsed time, i.e. one year, and adjusted by the costs to get there, i.e. the costs of the review phase. This then leads to the following values:

Project Valuation

117

US$ 324 mn

US$ 137 mn

US$ 0

US$ 0

US$ 270 mn

US$ 117 mn

US$ 0

US$ 0

US$ 0

US$ 0

Fig. 3.19. Values at the start of review phase for 1st line treatment

Mr. Investor then works back along the same lines and receives the following intermediary results: US$ 324 mn

US$ 92 mn

US$ 0

US$ 270 mn

US$ 78 mn

US$ 0

US$ 0

US$ 0

Fig. 3.20. Average values at the end of 1st line phase 3 trials

118

Valuation in Life Sciences US$ 302 mn

US$ 0

US$ 247 mn

US$ 0

US$ 0

US$ 0

Fig. 3.21. Values at the end of 2nd line phase 3 trials. The values are the sum of 1st line treatment in phase 3 and 2nd line treatment in review phase. US$ 203 mn

US$ 166 mn

US$ 0

US$ 0

Fig. 3.22. Average values after phase 3 for 2nd line treatment

And the final value is the US$ 37 mn that represent the average value at the end of the phase 2 trial, discounted for two years and reduced by the US$ 10 mn it takes for the trial. This leads us to a project value of US$ 21.7 mn.

Project Valuation

119

US$ 96 mn

US$ 74 mn

US$ 0

US$ 0

Fig. 3.23. Values at the start of phase 3 trial for 2nd line treatment US$ 96 mn

US$ 44 mn

US$ 0

Fig. 3.24. Average value after second phase 2 trial

Decision tree analysis is also a very useful method when assuming different market scenarios. We could easily assume that there is not only one scenario once an indication reaches the market, but maybe two. Once only Cytotoxabine competes with Onmarketabine, once there is also another drug that gets approval. The strength of decision tree analysis, the consideration of a wide range of possibilities, can also turn to a disadvantage. It becomes quickly difficult

120

Valuation in Life Sciences US$ 96 mn

US$ 22 mn

US$ 0

Fig. 3.25. Values after phase 2 trial

US$ 37 mn

Fig. 3.26. Average value after phase 2 trial

to navigate through the calculation. And as shown in the example, often scenarios depend on events, whose probability is difficult to estimate. Whether the probability to directly move on to phase 3 is 35% or maybe only 30% is difficult to determine. Overall the probability of failure in phase 2 should correspond to the attrition rate. For Cytotoxabine this would be 50% for immediate failure and 15%*40% for failure even after the second phase 2 trial. This corresponds to an attrition rate of 56% or to a success rate of 44%. Due to its ability to cover all possible development scenarios and also due to the necessity to think a development plan fully though, decision tree analysis is very popular for in-house valuations in pharmaceutical companies. In the following chapters we will not go into more details on decision tree analysis as the calculations are the same like for DCF. The method can easily be combined with license contract valuation or company valuation using the same principles outlined above.

Project Valuation

121

Project Valuation: Real Options

When calculating the net present value of a project using the discounted cash flow method; we determine the estimate of the future cash flows as if it were certain, with no room to fluctuate. We know that the estimate is not accurate; therefore we have to use a risk premium to remunerate for the uncertainty of the estimate. However, in the calculations we then assume that the size of the cash flows is fixed. Nevertheless, decisions along the project path depend exactly on parameters that may change in the course of time. Imagine that a competing project shows outstanding clinical results. This might impact our sales estimations; and maybe the sales estimations are even that low that they do not justify further investments in the project. Revaluing the project, management then decides to abandon it because of non-profitability. Such scenarios are not considered in the DCF framework. To get a realistic value of the project it is therefore reasonable to model the uncertainty of the parameters that influence project relevant decisions. Life sciences projects have critical decision points at the beginning of each phase. The previous phase gave insight into the characteristics of the product, how it behaves with respect to efficacy, safety, and compared to competitive products. This allows drawing some conclusions on the market potential. If the market potential does not compensate for the coming investments, i.e. if the value of the project is negative, then the project is stopped. Concretely, we should model the economic potential of the project, i.e. the estimated peak sales; the sales potential is most prone to uncertainty and has a crucial impact on the profitability of a project. We therefore have to assess, in addition to the parameters used in the DCF calculation, the volatility of the peak sales estimate. We have dedicated a special section to the determination of the volatility. Real Options Valuation We can now value the project. In a first step, we model the uncertain parameters, i.e. the peak sales estimate; we span a binomial tree from the date of valuation to the time of the last decision. In life sciences, this is mostly the date of filing a new drug application or launch. The time steps of the tree should be chosen in a way that they reflect the project plan. Often a project is scheduled in years, half years, or sometimes in quarters. In the second step, we value the project at each end node of the tree using the discounted cash flows method. Remember that if there are no decisions to

122

Valuation in Life Sciences

Real Options Valuation 1. Span a binomial tree for the peak sales estimate 2. Value the project in each leaf of the tree 3. Work back the tree to the root. For each time step do 1. 2. 3. 4. 5.

Take expectation of up- and down-step Apply success rate if necessary Discount for the elapsed time step Add cash flows Decide if it is a decision point

⎧⎪ ⎫⎪ − dt t + dt ) Vt = max ⎨(1 + r ) Pt ( pVupt + dt + (1 − p)Vdown + CFt ,0⎬

N

⎪⎩ 3 ⎪ 2 4 1 ⎭

(3.12)

5

t+dt

V up p V

t

1-p t+dt

Vdown Fig. 3.27. One time step

make, DCF and real options yield the same value. The third step is working back the tree to the root node. The value of the project is then equal to the value we have calculated for the root node. Working back the tree means that we place ourselves in a node one time step back of the layer we have already valued. Consequently, we know the value of the project if it reaches either the upper or the lower state. To get the value for the state we have placed ourselves in we must complete five actions. First, we must take the expectation of reaching after the time step the upper or the lower state. Second, we need to account for the fact that the project

Project Valuation

123

might not reach the next time step because of attrition. Note that for most of the time steps the transition probability is 100%. Only at the end of a phase, attrition becomes effective, i.e. completing a development phase. Third, We have to account for the time that elapses in this time step; hence, we have to discount the value. Fourth, we must add the cash flows that occur in this time step. The cash flows must not be probability adjusted. The attrition rates are already considered in action 2. Finally, fifth, we must decide whether it is worthwhile to continue the project in that node. Of course, this happens only at decision points, i.e. at the beginning of each phase. A project is usually continued if its value is positive. In case of a negative value, the project is abandoned and consequently its value in this node is set to zero. This fifth action is exactly where the option thinking comes in. We lift a negative value to zero, which is a value increase (by avoiding losses). Management makes use of its possibility to revaluate the project and abandons it if it is not profitable. Using the real option framework we assume that at each decision point in the future the management values the project according to the market state and abandons all projects with a negative value. This thinking is not part of the DCF approach. For all nodes of the tree we have calculated the value of the project in that specific state, i.e. at that moment, with that peak sales estimate. The goal of the valuation is to find out the value of the project at valuation date with the given peak sales estimate. This is exactly the value of the root node. Note that this value could still be negative, or zero, suggesting abandonment of the project. To illustrate the steps of the real options valuation we continue with the case study. Case Study Project Valuation: ROV When Dr. Project Manager calls his friend Prof. Corporate Finance to ask about the discount rate he should best use, the latter also suggests him to use the real options valuation method. Prof. Corporate Finance argues that real options valuation is closer to the reality of a drug development project, since at the beginning of each phase management revalues the project and decides whether to continue it or not, just like in the present situation. This argumentation sounds fair to Dr. Project Manager and he decides to find out how much value management flexibility adds to his project. Volatility. The only parameter he has to determine in addition to the previous DCF calculation is the volatility. Dr. Project Manager remembers

124

Valuation in Life Sciences

that the volatility is a measure of uncertainty, but is confused about the parallel use of volatility and success rates. Furthermore, he is not sure about how to determine the volatility. So, he calls up his friend once again. Prof. Corporate Finance explains him, that the volatility captures only the uncertainty of the peak sales and not of the whole project. A straightforward method to measure the volatility would therefore be calculating the volatility of adjusted drug sales. These might range from 10% to 20%. But Prof. Corporate Finance argues, that the volatility does not actually refer to the peak sales, but to the peak sales estimate. This is impacted during R&D by the clinical results that allow conclusions on the commercial performance of the drug. New information during R&D is not included in the volatility of marketed drug sales. He suggests to add an R&D spread on top of the volatility of marketed products of about 15%. Dr. Project Manager decides to use a volatility of 30%, representing an average risk profile (for a detailed description on how to calculate the volatility consult the section in Basics and Special Aspects). Solution Case Study Project Valuation: ROV Real options valuation with a binomial tree is done in three steps. First, span the tree, second, attribute a value to each end node, and third work back the tree to the root node. Construction of the binomial tree. We model the underlying peak sales estimate with a binomial tree following the parameters µ=0% (growth rate) and σ=30% (volatility). Still using yearly time steps this gives us:

(

)

(

)

u = exp σ dt = exp 0.3 ⋅ 1 = 1.35 d=

p=

1 1 = = 0.74 u 1.35

(1 + µ )dt − d = (1 + 0)1 − 0.74 = 43% u−d

1.35 − 0.74

Starting at valuation date we mount the tree to the final decision point, in our case beginning of NDA. NDA because we assume that this is the last decision point. We therefore model 6 time steps (1 year for phase 1, 2 years for phase2, and 3 years for phase 3). This gives us the following tree:

Project Valuation

125

Year 1

Year 2

Year 3

Year 4

Year 5

Year 6

Year 7

Phase 1

Phase 2

Phase 2

Phase 3

Phase 3

Phase 3

Approval $ 1'458 mn

$ 1'080 mn

1%

$ 800 mn

1%

$ 800 mn

$ 593 mn

3%

$ 593 mn

5%

$ 439 mn

8%

$ 439 mn

9%

$ 439 mn

$ 325 mn

18%

$ 325 mn

18%

$ 325 mn

16%

$ 241 mn

43%

$ 241 mn

31%

$ 241 mn

25%

$ 241 mn

100%

$ 179 mn

49%

$ 179 mn

36%

$ 179 mn

29%

57%

$ 132 mn

42%

$ 132 mn

34%

$ 132 mn

33%

$ 98 mn

32%

$ 98 mn

30%

19%

$ 73 mn

23%

$ 73 mn

11%

$ 54 mn

16%

6%

$ 40 mn

Legend: Peak Sales Probability

4%

Fig. 3.28. Binominal tree with probabilities and market states

Year 7 Approval

Year 8 Approval

Year 10 Market 19% $ 1,458 mn $ 277 mn ($ 97 mn) ($ 40 mn)

Year 11 Market 36% $ 1,458 mn $ 525 mn ($ 184 mn)

Year 12 Market 51% $ 1,458 mn $ 744 mn ($ 260 mn)

Year 13 Market 64% $ 1,458 mn $ 933 mn ($ 327 mn)

Year 14 Market 75% $ 1,458 mn $ 1,093 mn ($ 383 mn)

($ 40 mn)

Year 9 Market 5% $ 1,458 mn $ 73 mn ($ 26 mn) ($ 40 mn)

$ 1,458 mn

$ 1,458 mn

($ 3 mn)

Success Rate Probability

100% 100%

95% 100%

100% 95%

100% 95%

100% 95%

100% 95%

100% 95%

100% 95%

Risk adjusted net CF Discount Net present CF

($ 3 mn) 100% ($ 3 mn) Year 19 Market 100% $ 1,458 mn $ 1,458 mn ($ 510 mn)

($ 40 mn) 88% ($ 35 mn) Year 20 Market 100% $ 1,458 mn $ 1,458 mn ($ 510 mn)

$ 7 mn 77% $ 5 mn Year 21 Market 99% $ 1,458 mn $ 1,443 mn ($ 505 mn)

$ 133 mn 67% $ 90 mn Year 22 Market 97% $ 1,458 mn $ 1,414 mn ($ 495 mn)

$ 324 mn 59% $ 192 mn Year 23 Market 95% $ 1,458 mn $ 1,385 mn ($ 485 mn)

$ 459 mn 52% $ 238 mn Year 24 Market 72% $ 1,458 mn $ 1,050 mn ($ 367 mn)

$ 576 mn 46% $ 263 mn Year 25 Market 53% $ 1,458 mn $ 773 mn ($ 270 mn)

$ 675 mn 40% $ 270 mn Year 26 Market 37% $ 1,458 mn $ 539 mn ($ 189 mn)

Success Rate Probability Risk adjusted net CF Discount Net present CF

100% 95% $ 900 mn 21% $ 187 mn

100% 95% $ 900 mn 18% $ 164 mn

100% 95% $ 891 mn 16% $ 142 mn

100% 95% $ 873 mn 14% $ 122 mn

100% 95% $ 855 mn 12% $ 105 mn

100% 95% $ 648 mn 11% $ 70 mn

100% 95% $ 477 mn 9% $ 45 mn

100% 95% $ 333 mn 8% $ 28 mn

rNPV

$ 2,870 mn

Sales curve Peak Sales Sales revenues Operating expenses R&D Expenses

Sales curve Peak Sales Sales revenues Operating expenses R&D Expenses

Fig. 3.29. rNPV calculation of the uppermost end node

126

Valuation in Life Sciences

Valuation of the end nodes. We now attribute a value to each end node; we calculate the project value as if the project were at the beginning of the approval phase, with the estimated peak sales corresponding to the node. In the most optimistic case, we calculate the value of the project assuming that it reaches peak sales of $ 1,458 mn (field on the top right end). This seems to be an unrealistic case, but bear in mind that this state is achieved only if in each time step the peak sales estimate improved, i.e. with a probability of p6=0.6%. So, the probabilities of the tree already take care of the likelihood of each state. For the valuation of the project in each end node we have to go through all steps of the DCF valuation. The value of the project in the uppermost end node is therefore $ 2,870 mn. This we do for every end node, yielding the following results: Year 1

Year 2

Year 3

Year 4

Year 5

Year 6

Year 7

Phase 1

Phase 2

Phase 2

Phase 3

Phase 3

Phase 3

Approval $ 1'458 mn

$ 1'080 mn $ 2'870 mn $ 800 mn $ 593 mn $ 439 mn $ 325 mn $ 241 mn

$ 439 mn $ 325 mn

$ 241 mn $ 179 mn

$ 439 mn $ 325 mn

$ 799 mn

$ 179 mn

$ 397 mn

$ 241 mn $ 179 mn

$ 132 mn

$ 241 mn

$ 132 mn $ 98 mn

Legend:

$ 800 mn $ 593 mn $ 1'533 mn

$ 132 mn $ 98 mn

$ 175 mn

$ 54 mn

$ 55 mn

$ 73 mn

Peak Sales Value

$ 73 mn $ 40 mn ($ 12 mn)

Fig. 3.30. Project value in the end nodes

Note that the value of the lowest end node is negative. This means that, if we estimate the project at the beginning of the approval phase to only achieve peak sales of $ 40 mn, we are better off abandoning the project, as it does not compensate for all the expenses that are still due. An abandoned project has a value of zero. We therefore replace the negative value with a zero. Working back the tree. We now work back the tree. For this we have to go one time step back from the values we have already calculated. In the cur-

Project Valuation

127

rent situation we have calculated all values for the beginning of the approval phase, we therefore place ourselves now in the last year of phase 3. For each state we have the possibility to reach either the upper or lower state in the following year with the respective probabilities. Phase 3

Approval $ 439 mn

$ 325 mn

$ 799 mn $ 241 mn $ 397 mn

Fig. 3.31. Calculation of the value in one state

By way of example we calculate the value for the state with peak sales $ 325 mn. We know that after one time step the project will be worth either $ 799 mn with a probability of 43% or $ 397 mn with a probability of 57%. However, these probabilities only hold if the phase 3 trial is successful, i.e. in 75% of all cases. During the time step in consideration the company has to pay the last tranche of the phase 3 trial costs, $ 27 mn, and another year elapses, resulting in discounting the future values. Using the formula x, we can compute the value for the current state: t + dt ) + CF = Vt = (1 + r ) Pt ( pVupt + dt + (1 − p)Vdown − dt

(3.13)

(1 + 0.14)−1 ⋅ 0.75 ⋅ (0.43 ⋅ 799 + (1 − 0.43) ⋅ 397 ) − 27 = 347 If the current state is a decision point, then we would also have to correct negative values to zero, corresponding to abandonment of the project in that state. Applying the above formula to all states and working back the tree time step after time step to the rote node, we can fill the tree with project values for each node (cf. Fig. 3.32). We see that in phase 3 in the lowest state (grey field), the company abandons the project and in the approval phase if the peak sales amount to only $ 40 mn. The value of the project in these states is consequently zero. The value of the project at valuation date is $ 3.4 mn. Comparing this with the $ 2.9 mn DCF value we note that the possibility to halt the project if it has a negative value adds 17% of value in this case. Note that decision points are only at the beginning of the phases. In our case the project is abandoned in two decision points, beginning of phase 3 and at the beginning of approval phase in the lowest nodes.

128

Valuation in Life Sciences

Year 1

Year 2

Year 3

Year 4

Year 5

Year 6

Year 7

Phase 1

Phase 2

Phase 2

Phase 3

Phase 3

Phase 3

Approval $ 1'458 mn

$ 1'080 mn $ 2'870 mn $ 800 mn $ 1'357 mn $ 593 mn

$ 800 mn

$ 835 mn

$ 593 mn $ 1'533 mn

$ 439 mn

$ 492 mn

$ 439 mn

$ 705 mn

$ 439 mn

$ 325 mn

$ 112 mn

$ 325 mn

$ 411 mn

$ 325 mn

$ 799 mn

$ 241 mn

$ 50 mn

$ 241 mn

$ 217 mn

$ 241 mn

$ 347 mn

$ 241 mn

$ 3.4 mn

$ 179 mn

$ 37 mn

$ 179 mn

$ 179 mn

$ 179 mn

$ 397 mn

$ 2 mn

$ 132 mn

$ 66 mn

$ 132 mn

$ 151 mn

$ 132 mn

($ 1 mn)

$ 98 mn

$ 51 mn

$ 98 mn

$ 175 mn

$ 73 mn

$ 43 mn

$ 73 mn

($ 16 mn)

$ 54 mn

$ 55 mn

($ 11 mn)

$ 40 mn

Legend: Peak Sales Value

$ 0 mn

$ 0 mn

Fig. 3.32. Solving the tree to the root node

Case Discussion Cash flows size. The determination of the peak sales in the case displays a common problem when valuing early stage projects. At the start of clinical phase one, it is often unclear what the indication of the drug might be if it reaches market. This is especially true in oncology, but also cardiovascular products like Viagra are used for other indications than initially planned. At this stage, we can often only roughly estimate what the drug might do, but not predict how many patients will be treated, and what the dosage and frequency will be. A comprehensive way to determine the sales in this case is to compare the drug to the most similar subgroup of drugs on the market today. In the example, we have simply chosen to look at all cardiovascular drug. We could also have taken a sample of antihypertensives, or betablockers. If we think that the drug we are developing could have the potential to also reach the extremes of the sales distribution of current drugs, we could use the average sales as a reference. If we think that our drug has not the potential to be a blockbuster, but is likely to be an average drug, we should to use the most-likely or the median peak sales as a reference. The number we get out of this approach, the comparable approach, is the sales estimate of the drug as if it were on the market today. This number can be extrapolated to the time of launch and to the time when peak sales should

Project Valuation

129

be reached by estimating a growth rate and by assessing the impact of competitors with drugs currently in development. We will discuss below the problems with applying a growth rate to the sales estimate. Cash flows probability. When we determine the probability of the cash flows we have to make sure not to base the numbers on gut feeling. There is no way in predicting the outcome of clinical trials. We have to base our valuation on average success rates conforming to our drug. Furthermore, we should always use the same source for the success rate when we want to compare the project to others. As we have described above, there are differences in the success rates depending on what sample the publication bases on. Cash flows timing. The tricky part when determining the timing is to be realistic about the project and the sales development. Too often, drug development takes more time than anticipated. In our example, we have used round numbers for the purpose of illustration. These numbers are likely to be at the lower end of drug development duration and are likely to be exceeded. Also the sales curve for an individual drug is hard to predict. We think it is most suitable to use an average sales curve based on historical data as in our case. Obviously, this sales curve will depend on the disease group and classification of the drug (orphan). Dicounting. As the project in our case originates from the company itself and is developed in-house, the company discount rate can be used. If the project were in-licensed or if a particular country risk were to apply, an adjustment of the discount rate could be discussed. Nevertheless, special care is needed not to subjectively manipulate the discount just to reach the desired valuation outcome. In order to compare projects it is important that the valuation approach and the input data are consistent, in particular the discount rate. DCF vs. ROV. Using ROV instead of DCF, we assume that the estimated peak sales can change according to the volatility. The average of the estimated peak sales at the launch nodes is the same as the estimated peak sales in the DCF calculation. We therefore do not change the average predicted sales number in ROV, but we do allow the estimate to go up or down. Obviously, if the sales expectation falls below a certain point, the project is not profitable anymore, i.e. has a negative value. In ROV we model that the management halts the project, once its value becomes negative in the tree. The real option value is then increased by putting the

130

Valuation in Life Sciences

negative value at that node to zero, as the project is abandoned and future losses avoided. If we omit the option to abandon, i.e. if we keep the negative values, then the tree calculation yields exactly the DCF value, as shown in the figure below. This displays the consistency of the two values. The tree has been built with the condition that the expected (or mean) value development of the peak sales estimate equals the one we have supposed in the DCF approach. This was done by the use of the growth rate in the tree parameters u, d, and p. If we now use the same assumptions as DCF implicitly does, namely, that management continues the project except scientific attrition avoids it, then the value calculated with the tree should obviously be the same like the DCF value. It is exactly the ability of the management to react on a change of the situation, in our case a correction of the peak sales estimate due to new information about the product, its competition, and the market environment, that creates the additional value of real options valuation. Year 1 Phase 1

Year 2 Phase 2

Year 3 Phase 2

Year 4 Phase 3

Year 5 Phase 3

Year 6 Phase 3

Year 7 Approval $ 1'458 mn

$ 1'080 mn $ 2'870 mn $ 800 mn $ 1'357 mn $ 593 mn

$ 800 mn

$ 835 mn

$ 593 mn $ 1'533 mn

$ 439 mn

$ 492 mn

$ 439 mn

$ 705 mn

$ 439 mn

$ 325 mn

$ 112 mn

$ 325 mn

$ 411 mn

$ 325 mn

$ 799 mn

$ 241 mn

$ 50 mn

$ 241 mn

$ 217 mn

$ 241 mn

$ 347 mn

$ 241 mn

$ 2.9 mn

$ 179 mn

$ 37 mn

$ 179 mn

$ 179 mn

$ 179 mn

$ 397 mn

$ 1 mn

$ 132 mn

$ 66 mn

$ 132 mn

$ 151 mn

$ 132 mn

$ 98 mn

$ 51 mn

$ 98 mn

$ 175 mn

($ 17 mn)

$ 73 mn

$ 43 mn

$ 73 mn

($ 19 mn)

$ 54 mn

$ 55 mn

($ 16 mn)

$ 40 mn

($ 4 mn) Legend: Peak Sales Value

($ 12 mn)

Fig. 3.33. Real option value without abandonment of negative states

Sensitivity analysis. We now study the impact of various parameters on the valuation and discuss the implications on how to define them. The following five parameters have a very high impact when an early stage project is at the money:

Project Valuation

131

• Peak sales • Growth rate • Margin • Discount rate • Success rates Peak sales. Naturally, peak sales must have an important influence on the value of a project, since they account for all positive cash flows. A 10% change in peak sales, i.e. a change of $ 24 mn, translates into a value change of about $ 4 mn. For a project that is right at the border of profitability as in our example, such a 10% change can lead to a value increase of more than 100%. If however we assume the project to be clearly profitable (or in the money) with annual peak sales of $ 486 mn, the same 10% change of peak sales increases the value only by 19%, or a peak sales increase of $ 49 mn translates into a value increase of $ 9 mn (cf. Table 3.25). The project is sensitive to peak sales, but even very high sales numbers are only partly translated into the absolute value of the project, while the costs for phase 1, as an example, are translated one to one into the value. E.g. an increase of $ 5 mn in clinical phase 1 costs will decrease the value by $ 5 mn. Growth rate. As we have seen, the peak sales are an estimate of what the drug could reach if it were on the market today. Often, people adjust that number by adding a growth rate of 5-10%. The sensitivity analysis shows that a growth rate of only 1% increases the value by 200%, a growth rate of 2% by 500%. We therefore recommend not applying a growth rate at all if we work with the above-described comparable approach to determine the peak sales. It is hard to estimate what influence the industry growth rate has on a single product. The growth can be caused by demographic and pricing dynamics, in which case the growth impacts the sales figures of a single drug. But the growth can also be due to an increase in treatments, which would rather have a negative impact on a single product due to increased competition. Margin. A 10% change in margin provokes exactly the same change as a 10% change in sales. Since we multiply each sales cash flow with the margin, we have exactly the same effect whether we modify the sales estimate or the margin by the same percentage. The following equation illustrates this: SC(i)*(PS*(1+10%))*margin=SC(i)*PS*(margin*(1+10%))

(3.14)

132

Valuation in Life Sciences

Discount rate. Like the growth rate, the discount rate has a tremendous effect on the valuation. A careful and consistent use of the discount rate is recommended and the rate should not be misused to manipulate the valuation result. Success rates. The success rates immediately factor away a large part of the project value. If we look at the current phase, the success rate translates one to one in the valuation. As discussed above, this sensitivity calls for the use of published average success rates and forbids individual speculations about a project’s success. The parameters that influence the valuation most are not observable in reality: The growth rate and the discount rate. Already small subjective changes make valuations arbitrary and incomparable. We therefore recommend using those figures in a standardised way. An interesting observation is that the volatility, contrary to the belief of many critics of ROV, does not have a large impact on the results. If we consult the absolute changes in the example, we see that DCF is more sensitive to alterations of the input parameters. For this we have to understand that the real option value is actually the DCF value of the project plus the value of the flexibility to abandon the project. Real Option = DCF + Flexibility

(3.15)

The flexibility is less valuable the less likely it is that management makes use of it. If now a change leads to an increase of the project value, the value of the flexibility decreases. Hence the total impact of the change is smaller for the real option value. The same is true for an input change that causes a value decrease. The project is then more likely to be abandoned because of economic reasons; the flexibility becomes more important. This increase in flexibility value partly amortises the value loss due to the input change. We will now assess what impact changes of input parameters have on the same project if it is clearly in the money. We use peak sales of $ 486 mn (average sales) instead of $ 241 mn. The valuation is especially sensitive to: • Growth rate • Discount rate All other parameters have only a slight influence on the project as it is now clearly in the money.

Project Valuation

133

Table 3.25. Sensitivity analysis Base scenario Peak Sales +10% Peak Sales -10% Launch Costs +10% Launch Costs -10% Growth Rate 1% Growth Rate 2% Margin +10% Margin -10% Discount +10% (15.4%) Discount -10% (12.6%) Volatility +10% (33%) Volatility -10% (27%) Success Rate C1 +10% (69.3%) Success Rate C1 -10% (56.7%) Costs C1 +10% Costs C1 -10% Success Rate C2 +10% (46.2%) Success Rate C1 -10% (37.8) Costs C2 +10% Costs C2 -10% Success Rate C3 +10% (82.5%) Success Rate C3 -10% (67.5%) Costs C3 +10% Costs C3 -10% Success Rate NDA +10% (100%) Success Rate NDA -10% (85.5%) Costs NDA +10% Costs NDA -10%

DCF ∆ DCF ROV ∆ ROV ∆ DCF-ROV $ 47.88 mn 0% $ 47.90 mn 0% $ 0.02 mn $ 56.82 mn 19% $ 56.83 mn 19% $ 0.01 mn $ 38.95 mn -19% $ 38.97 mn -19% $ 0.02 mn $ 47.07 mn -2% $ 47.08 mn -2% $ 0.01 mn $ 48.70 mn 2% $ 48.71 mn 2% $ 0.01 mn $ 62.75 mn 31% $ 62.80 mn 31% $ 0.05 mn $ 80.12 mn 67% $ 80.28 mn 68% $ 0.16 mn $ 56.82 mn 19% $ 56.83 mn 19% $ 0.01 mn $ 38.95 mn -19% $ 38.97 mn -19% $ 0.02 mn $ 34.24 mn -28% $ 34.25 mn -28% $ 0.01 mn $ 65.04 mn 36% $ 65.06 mn 36% $ 0.02 mn $ 47.88 mn 0% $ 47.90 mn 0% $ 0.02 mn $ 47.88 mn 0% $ 47.90 mn 0% $ 0.02 mn $ 53.57 mn 12% $ 53.59 mn 12% $ 0.02 mn $ 42.20 mn -12% $ 42.21 mn -12% $ 0.01 mn $ 46.98 mn -2% $ 47.00 mn -2% $ 0.02 mn $ 48.78 mn 2% $ 48.80 mn 2% $ 0.02 mn $ 54.71 mn 14% $ 54.73 mn 14% $ 0.02 mn $ 41.06 mn -14% $ 41.07 mn -14% $ 0.01 mn $ 46.74 mn -2% $ 46.76 mn -2% $ 0.02 mn $ 49.03 mn 2% $ 49.04 mn 2% $ 0.01 mn $ 55.97 mn 17% $ 55.99 mn 17% $ 0.02 mn $ 39.79 mn -17% $ 39.81 mn -17% $ 0.02 mn $ 46.62 mn -3% $ 46.64 mn -3% $ 0.02 mn $ 49.15 mn 3% $ 49.16 mn 3% $ 0.01 mn $ 52.32 mn 9% $ 52.34 mn 9% $ 0.02 mn $ 39.45 mn -18% $ 39.46 mn -18% $ 0.01 mn $ 47.86 mn 0% $ 47.87 mn 0% $ 0.01 mn $ 47.91 mn 0% $ 47.93 mn 0% $ 0.02 mn

We see that the difference between DCF and ROV completely vanishes when a project is clearly in the money. The probability that management has to make use of its flexibility to abandon the project is virtually zero, therefore no additional value can be attributed to the project in the real option framework. The larger differences for the growth rate differences stem from slight imprecision of the binomial tree method. In contrast to DCF, ROV illustrates that the project might be abandoned due to a change in the peak sales estimate and sensibilises the management on this.

134

Valuation in Life Sciences

Simulations of Projects

Introduction In this section, we give a short introduction to simulations and how to use them in Excel. Simulations are a very useful tool to appreciate the risk profile of a project or a company, but they are technically much more demanding than standard valuation. In order to exploit the full power of simulations we must write a code in a programming language. This is beyond the scope of this book; but we give an outlook to what can be done with simulations without programming. Simulations are models of reality, imaginations of how it could be. While in the real world we only see one version of a random event – a dice shows only one number – we can simulate all possible outcomes. When simulating rolling a dice 1,000 times, we can see that on average one out of six times we get a one. But we can even retrieve much more out of the simulations. With what probability does an even number follow a three, or how often do we get two consecutive fours? In life sciences, we have two different sorts of random events. First and most importantly, we do not know in advance the outcome of the trials. We can only estimate the chances of the trial to be successful; these are typically the success rates. Second, the peak sales estimate and later the peak sales themselves change over time. In reality, each trial will be either a success or a failure, and if the product reaches commercialisation, it will achieve some peak sales. While in reality we observe the results only a posteriori, we anticipate the results when we simulate. We generate randomly trial results and peak sales. Knowing the project’s fate in advance, we can deduce what cash flows will occur at what time, e.g. if the project passes all phases up to phase 2 but fails in phase 3. We then would have to account for all phase costs until phase 3. Since in this scenario we would actually incur these expenditures, we have to account for them in full, without adjusting them for their probability. If we now discount all cash flows and sum them up, we obtain the value of this scenario. In the case of a phase 3 failure, it is a considerable loss, we spend the phase costs and the trial fails. Repeating this method many times, we will get scenarios where the project fails in the first phase, but also scenarios where the project reaches commercialisation. When correctly simulating the trial outcomes, the quantity of commercialisation scenarios corresponds exactly to the success rate that the project reaches the market.

Project Valuation

135

Analysing the simulation results, we get a better idea of the project’s risk profile. The results allow retrieving data like • What is the probability of an IRR better than 30%? • What is the scenario that will only be outperformed in 10% of all cases? • What is the scenario we can outperform in 50% of all cases? • What is the worst case scenario? When running simulations on a portfolio or company basis, we can also check how well the pipeline is diversified; if it has some cluster risks, or how far the company value can possibly drop in case of the realisation of a very adverse scenario. We can also calculate a risk measure called value at risk (VaR) out of the simulation results. VaR(α) is the value drop which happens with a probability of α. In the portfolio simulation displayed in the figure below, the average value amounts to $ 2,860 mn. In 10% of all scenarios the value is lower than $ 2,150 mn. Consequently, the VaR(10%) amounts to $ 710 mn. We can also interpret the VaR like that: If we assume the mean value as fair, then the VaR corresponds to the negative mispricing that is exceeded in 10%.

VaR(10%)

Fig. 3.34. Value at risk for a pipeline value (1,000 simulations, calculated with ri:val7)

In the continuation, we will explain how we can generate random variables and how we can use them in simulations. 7

ri:val is a valuation software developed by the authors and specially designed for the biopharmaceutical industry.

136

Valuation in Life Sciences

Theory Excel offers a function rand() that generates random numbers in the interval (0,1). Each number in this interval is equally likely to come up. If we now want to generate the success of a clinical trial with success rate P, we need a random variable that has a value of 1 (success) with probability P and a value of 0 (failure) with probability (1-P). Using the following formula, we precisely get such a variable.

= if (rand () < P,1,0)

(3.16)

The cell with this formula takes the value 1 if the randomly generated number rand() is less than P. In the other cases, the cell takes the value 0. The formula also ensures that the simulations on average comply with the success rate for the phases. The more simulations we run the closer they approach the success rates. The simulation of peak sales estimates is more complicated. While clinical success is an instantaneous random variable, the peak sales estimate underlies a dynamic stochastic process; it can increase or decrease, as modeled in the binomial tree. Therefore, the simulation of a peak sales estimate at time of launch does not only depend on the uncertainty, i.e. the volatility σ, but also on the time until launch T-t. The following formula gives the value for a simulated peak sales estimate at time T, S(T), given the current estimate S(t) (with the growth rate µ). S(T)=S(t)*exp((µ-σ^2/2)*(T-t)+σ*(T-t)^(0.5)*normsinv(rand()))

(3.17)

Using these two methods to generate clinical success and peak sales estimates at launch we can set up an Excel sheet calculating the project value. Instead of calculating the probability of the cash flows, we simulate whether the project is still alive (1) at the time of the cash flow or not (0). If the project is alive, the cash flow occurs in this specific scenario, if not, it does not. By multiplying the cash flow with the cell that indicates whether the project is still alive (value 1 or 0) we get the cash flow we have to consider in this scenario. We then discount the cash flows and calculate the net present value. We now set up an NPV calculation using randomly generated numbers determining clinical success and peak sales as described. Automating the calculation we can generate a large number of scenarios and retrieve each time the NPV for that scenario. Implementation in Excel Valuation of one generic scenario. The table below illustrates how to set up an Excel sheet that values a scenario according to the simulated result.

Project Valuation

137

Table 3.26. Excel sheet with formulae for simulation A

B

1 2 3 4 5 6

Year Phase Success rate Success Project alive? Peak sales

1 Approval 85% =if(rand()
7 8 9 10 11 12 13 14 15

Sales Curve Revenues Operating Expenses Investments CF Real CF Discount NpCF NPV

2 =-B10 =B11*B5 1 =B12*B13 =sum(B14:T14)

C 2 Market 100% =if(rand()
We assume a project that is submitted for regulatory approval. We only display the first two years, the numbers for the other years being similarly calculated. In the row 4 we simulate the clinical success of the phases. For this we generate a random number and compare it with the success rate. In row 5 we then evaluate whether the project is still alive. This happens by multiplying the state of the previous year by the survival of the previous year. As soon as the project fails, all cells in row 5 left of that failure become zero. In cell C6 we simulate the peak sales for this scenario. As time step we used one year, the difference between the start of valuation and the time we want to know the peak sales for. The numbers µ and σ are supposed to be referenced in the Excel sheet. In the continuation we can assume the peak sales to be definite, i.e. they do not change any more. In row 8 and 9 we then use the simulated peak sales to deduce the revenues and the operating costs. All other cells are like in a standard DCF valuation. Note that we do not adjust the cash flows by their probability. A simulation is one possible version of reality, and in this version the cash flows actually happen. On average these cash flows occur with a percentage of all simulations that correspond to their probability. The simulations respect this way the probability (success rate) on average.

138

Valuation in Life Sciences

When pressing the button F9 (for Calc Now) the values of all cells get recalculated, above all the random numbers change. All values that are a result of these randomly generated numbers now adjust. Already using F9 we get an impression how the values can differ. But the real power of simulation lies in the number of recalculations. We will now see how we can calculate 10,000 values for this project. Simulation of Scenarios We now set up a work sheet that calculates the NPV for a randomly generated scenario. In order to automate the generation of scenarios and store the results we make use of the function “table”. First, we must prepare a new worksheet that stores all simulation results. We reference cell B1 with the cell that returned the NPV for the generic scenario in the previous sheet. Second, we numerate the column A from 1 to n, starting with cell A2. n stands for the number of simulations we want to run. Typically n=1,000 or n=10,000, the higher the better. In our example we use n=10 for illustration purposes. Table 3.27. Worksheet “scenario” A 1

B =scenario!B15

2

1

3

2

4

3

5

4

6

5

7

6

8

7

9

8

10

9

11

10

We now select the cell range A1:B11, or in more general terms A1:B(n+1). We then go into the menu Data and choose Table. In the prompted pop-up we reference the column input cell to A1 as displayed in the figure below.

Project Valuation

139

Fig. 3.35. Pop-up data-table

As a result, Excel automatically calculates scenarios for each row we have numbered before and stores the results in the cells B2:B11. The cells now contain the formula {=TABLE(,$A$1)}. We cannot fill in this formula directly; we always have to take the described way via the Table function in the Data menu. Table 3.28. Worksheet “simulations” A 1

B =scenario!B15

2

1

{=TABLE(,$A$1)}

3

2

{=TABLE(,$A$1)}

4

3

{=TABLE(,$A$1)}

5

4

{=TABLE(,$A$1)}

6

5

{=TABLE(,$A$1)}

7

6

{=TABLE(,$A$1)}

8

7

{=TABLE(,$A$1)}

9

8

{=TABLE(,$A$1)}

10

9

{=TABLE(,$A$1)}

11

10

{=TABLE(,$A$1)}

We have now managed to simulate n values of the project. Of course it is also possible to simulate the peak sales of a scenario, or the final phase, or simply all cash flows. For this we first need to calculate these values on the worksheet “scenario”. Second, we must reference the cells in the first row next to B1 in the worksheet “simulations” with the cells we want to simulate in the worksheet “scenario”. Imagine that cell B1 is referenced to the NPV, C1 to the peak sales estimate at launch, and D1 indicates whether the project has actually reached commercialisation in that scenario. We then must select the cell range A1:D11 and use again the function Table out of the menu Data. The worksheet “simulations” now looks the following:

140

Valuation in Life Sciences

Table 3.29. Worksheet “simulations” when simulating NPV, peak sales at launch, and whether product received approval A 1 2 3 4 5 6 7 8 9 10 11

1 2 3 4 5 6 7 8 9 10

B =scenario!B15 {=TABLE(,$A$1)} {=TABLE(,$A$1)} {=TABLE(,$A$1)} {=TABLE(,$A$1)} {=TABLE(,$A$1)} {=TABLE(,$A$1)} {=TABLE(,$A$1)} {=TABLE(,$A$1)} {=TABLE(,$A$1)} {=TABLE(,$A$1)}

C =scenario!C6 {=TABLE(,$A$1)} {=TABLE(,$A$1)} {=TABLE(,$A$1)} {=TABLE(,$A$1)} {=TABLE(,$A$1)} {=TABLE(,$A$1)} {=TABLE(,$A$1)} {=TABLE(,$A$1)} {=TABLE(,$A$1)} {=TABLE(,$A$1)}

D =scenario!C5 {=TABLE(,$A$1)} {=TABLE(,$A$1)} {=TABLE(,$A$1)} {=TABLE(,$A$1)} {=TABLE(,$A$1)} {=TABLE(,$A$1)} {=TABLE(,$A$1)} {=TABLE(,$A$1)} {=TABLE(,$A$1)} {=TABLE(,$A$1)}

Analysis of Simulations We can extract various information out of simulations. First of all, we can calculate the mean, which should correspond to the value of the project. The mean of the simulations however will never exactly match the value of the project as calculated with DCF, because the simulations are still only a random set of scenarios that only converge to the actual DCF value on average if we run an infinity of simulations. Nevertheless, the more simulations we run, the more we can trust the result to be close to the actual value. By grouping the simulation results into bins, i.e. into ranges, we can also see what the most likely outcome is, or if a positive value has realistic chances. Excel offers some powerful functions to retrieve interesting information out of a large data set in the menu Tools, under Analyse Data. We do not enter into more detail here, but present some results obtained with simulations that have proven very useful when valuing companies. Manual 1. Set up spreadsheet that calculates a generic scenario with random test results and peak sales. 2. Simulate n scenarios (usually n=1,000, or n= 10,000) with DataTable. 3. Analyse data set

Project Valuation

141

Simulations and Real Options Up to now, we have described simulations in a DCF framework. In the simulations we abandon projects if they fail in a clinical phase, i.e. if the result is negative. We have not included economically motivated abandonment, the principal idea of real options valuation. Therefore, the mean value of the simulations converges with the DCF value. We now assume that in the simulation the project will also be abandoned if its value drops below zero. We consequently simulate the peak sales up to the first decision point, i.e. the end of the current phase. In order to know if we have to abandon the project, we must know its value. But we don’t. We actually run the simulations to get the value of the project. We would therefore need to run another set of subsimulations for the peak sales estimate we have reached at the first decision point, and this for each simulation. In the sub-simulations we then encounter the same problem when arriving at the second decision point. Clearly, finding the real option value by this way of simulation is not an option.

1st decision point

2nd decision point

Fig. 3.36. Simulations with real options

We can elegantly solve this problem. Before starting the simulation, we must find out, what the minimum value of peak sales is for every decision point, such that the project still has a positive value. We call these values thresholds. They are calculated by putting oneself at the time of the decision point and then looking for the peak sales estimate that sets the project value to zero. The following graph displays how these thresholds influence

142

Valuation in Life Sciences

Fig. 3.37. Thresholds for a preclinical project (calculated with ri:val)

the simulations. The height of the graph displays the frequency of scenarios passing through that point (x-axis being the time, y-axis the peak sales estimate). In the simulations we then have to implement the real options idea in the following way: Table 3.30. Implementation of thresholds in worksheet “scenario” A 1 Year 2 Phase 3 Success rate 4 Success

B

C

1 2 Preclinical Phase 1 70% 59%

D 3 Phase 2 (first year) 100%

=if(rand()< =if(rand()
=C6*exp((µ+σ^2/2)*(D1C1)+σ*sqrt(D1-C1)* normsinv(rand())) 198

7 Alive?

=C7*C4*if(D5>D6,1,0)

1

=B7*B4*if(C5>C6,1,0)

=if(rand()
Project Valuation

143

In row 5 we simulate for every year a new peak sales estimate. Just like in the tree, the peak sales estimate can change every year. Row 6 indicates the beforehand calculated thresholds. In row 7 we then decide, whether the project is continued. For this, the project must have been alive the previous year, the previous year must have been survived, and the current peak sales estimate must be higher than the threshold. Note that for years without decision the threshold is zero. The rest of the simulation follows the same lines as in the DCF framework. But when simulating in the real options framework, the mean of the simulations approaches the real option value. Sensitivity Analysis and Sensitivity-Simulations

In valuation many parameters are unknown. We are unsure about the projected sales revenues, we do not know how expensive the trials will be, or if we calculate with the right set of success rates. A valuation may well include more than 20 input parameters, and if any of those is only a vague estimate the confidence in the final result might be jeopardised. We therefore want to know a lower and an upper bound for the value, knowing that the calculated value only represents one possible value, depending on the chosen set of assumptions, provided the particular set of parameters is correct. One method to address this deficiency, i.e. the dependence on the chosen value of the input parameters, is a sensitivity analysis, which points out which parameters need special diligence. We have seen previously that the discount rate and the growth rate are parameters that have a huge impact on the value. We therefore need to justify these parameters with special care. Performing an extended sensitivity analysis, we can also calculate the value presuming two or more parameters fluctuate around our initial assumptions. One popular method is to simulate the assumptions. Each parameter – discount rate, peak sales, trial costs, duration, success rates – underlie some predefined variation. As a result we have a distribution of the value and we can determine confidence intervals, e.g. saying that the value lies with a likelihood of 80% between US$ 15 mn and US$ 25 mn. Such information is very useful as it gives a feeling for the precision of our assessment. In the following we will investigate in a bit more detail how these sensitivity-simulations are performed and what they actually mean. Of course, one cannot simply let a parameter vary a little. For the simulations we need to know the range the parameter is allowed to vary in. We

144

Valuation in Life Sciences

Fig. 3.38. 2,000 sensitivity-simulations for a phase I project with 80% confidence interval (calculated with ri:val)

have to define a distribution for each parameter. We either say that it is distributed normally, log-normally, uniformly, or even according to another probability law. Most people have problems quantifying probability, likelihood, odds, or chances. Much more so when they must estimate a distribution. If we then run a simulation based on these distributions we get a distribution of the complete value. From here we then extract the interesting and assuring confidence intervals. But how much confidence can we have in numbers relying on probability distributions that people are chronically wrongly assessing? These confidence intervals provide a pseudo-security, trying to quantify the uncertainty of the estimates of the input parameters with something that is much more complicated to estimate, namely probability distributions and give a wrong sense of security. Allowing a wrongly chosen input parameter to further fluctuate will not lead to a better understanding of the value. As a second drawback, next to the mentioned difficulty to assess probability distributions, we must add the fact, that estimation errors of parameters are often linked to each other. Typically, estimations are too optimistic, and this results in too high success rates, too short timelines, and too high peak sales assumptions. In most cases these errors occur all at once. Of course, also the vice versa is possible, but even then all assumptions are estimated too conservatively at once. Rarely some parameters are optimistically chosen and others conservatively. However, by letting vary all input parameters independently from each other this cluster effect (all optimistic or all conservative) gets lost, leading to too narrow confidence intervals. In order to account for this effect correctly we would have to define not only each parameter’s probability distribution, but also the joint distribution of all parameters, i.e. their correlations between each other. It is already difficult to cope with the standard notion of probability in the

Project Valuation

145

Fig. 3.39. 2,000 simulation of phase I project with 10% worst and 10% best case (calculated with ri:val)

sense of success rates. The estimation of the probability distribution of parameter estimation (the first drawback) is in our opinion only possible for a very limited set of parameters where we have enough data at hand, e.g. for sales estimates or costs. But the estimation of correlation of estimation errors seems quite an impossible task to do. We are fully aware that the arguments listed here against sensitivitysimulations are pretty much the same like the arguments raised against real options. Also in real option valuation it is not certain that the underlying behaves according to a geometric Brownian motion. Nevertheless real options have their merit because they assume at least some uncertainty and the volatility is scaled according to the risk perceived in the estimation of the peak sales. Although the precise quantification might still not be precise, it is a progress with respect to DCF, because we have to consider future options in the project plan. Sensitivity-simulations call attention to over-reliance on a set of parameters, very much the same as real options do with respect to a predefined development plan. This is certainly a positive point and cannot be stressed enough. Sensitivity analysis, however, provides the same advantages, but without offering a pseudo-security by wrapping the estimation errors in a much more error-prone confidence interval. The described sensitivity-simulations differ fundamentally from the previously treated simulations of a project. In a simulation of a project we assume different scenarios of reality, i.e. of what can happen. Once the project fails in the first phase, once it reaches commercialisation. Once the sales develop better than assumed, once worse. The result of these simulations is the risk profile of the project or the company, i.e. what a company or an investor can expect and with what probability. Sensitivity-simulations do not offer an appreciation of the risk profile but rather of the faith in the chosen assumptions. Obviously, these are two different concepts.

146

Valuation in Life Sciences

License Contract Valuation In the following chapter we will learn how to value license contracts and how the value of different contract structures is captured. In the theoretical part and in the case studies we treat license contracts between two counterparties and then analyse sublicensing agreements involving three parties. Fundamentals of Licensing

Structure of License Contracts Most of the large players in life sciences owe their success to few products in the past. Revenues from these products have allowed them building large R&D centres to generate new innovative projects. This has often worked well. The best ideas however are not necessarily born in these laboratories, but all over the world. Nevertheless, once these ideas reach a mature stage, they need a machinery that first develops, later markets and sells them. The big players in the industry have these development and marketing machines. Licensing between innovators and big players gives new technologies and products better commercial prospects thanks to access to large development and marketing know-how. On the other hand, licensing fuels the marketing machines of the big players. In an innovation driven industry like life sciences, licensing is therefore a logical business model. Furthermore, some companies do not have the necessary financials to develop and launch a project, while others do have the money, but not the projects. Reasons for licensing and technology transfer are8: 1. Forming alliances with partners that can progress the development of the technology to take it to market. 2. Forming alliances with partners with manufacturing capability. 3. Forming alliances with partners with marketing and distribution capability. 4. Exploitation in a different field of application. 5. No commercial capability. 6. Access to innovative projects.

8

“Licensing and Technology Transfer in the Pharmaceutical Industry“. Philip Mendes, Partner, Innovation Law, Brisbane.

License Contract Valuation

147

License contracts allow the licensor securitising at least a part of the project value. The licensor can bypass upcoming liquidity bottlenecks and reduce considerably the financial risk profile. The licensee on the other hand buys a new project, even if the payment modalities seem a little odd. License contracts need to distribute the risk to the license and the licensor and need to take into account the financial preferences of both sides. In order to take account of the different preferences, there is a multitude of possible license contract clauses and structures. Outlined below are some examples: Upfront payment: An upfront payment takes place at the beginning of a license contract in form of cash or equity. The licensor profits from immediate cash and reduces or even eliminates its downside risk. The licensee on the other hand reduces its obligations (royalties) in case the product proves successful. Milestones: A mailestone is a cash payment related to achieving certain milestones in research, development, or marketing, e.g. to successfully pass clinical phase 2 or reach the first $ 100 mn sales. Milestones incentive the licensor for good results. The licensee reduces the risk of spending a lot of money on a product that does not work by linking the payments to positive results. Due to information asymmetry the licensee prefers such staged payments to a one-time payment. . R&D funding: The licensee takes over all or just a part of the licensor’s costs for the development of the licensed product. Payments can be in form of FTE funding, lump sum payments or infrastructure financing. The licensor can build up further research capabilities and advance the company to a higher development stage. The licensee ensures that the project advances according to his plans and can reduce the overall price for the license. Redemption: The licensee refunds sunk R&D costs to the licensor. Redemption is similar to R&D funding with the important difference, that redemption is usually linked to a milestone, e.g. positive results. Redemption is similar to milestone payments from a financial modelling perspective with the exception that the amount is not exactly known in advance. Royalties: The licensor participates in the net revenues of the out-licensed product with a certain percentage. This percentage can be fixed or tied to the amount of sales revenues, i.e. have multiple levels. In practice, a licensor could receive 5% of the annual net sales <$ 100 mn, 6.5% for annual net sales between $ 100-250 mn and 7.5% for annual net sales > $ 250 mn. If the annual net sales amount to $ 300 mn, then the licensor receives 5%

148

Valuation in Life Sciences

of the first $ 100 mn, 6.5% of the next $ 150 mn, and 7.5% of the remaining $ 50 mn. However, it has also been seen that the contract foresees in such a case the payment of 7.5% on all $ 350 mn. In the industry, this is denoted as royalty stacking. Some contracts fix the royalty rate until a certain amount of cumulative sales is achieved, e.g. 4% for the first $ 200 mn, no matter how long this takes. Royalty payments can have a fixed floor or ceiling. They can also depend on markets or indications. An important royalty structure for academia and research institutes is the reach-through royalty. These are paid not on the sales of the licensed product or technology but instead, on products, which derive from the licensed product or technology. Alternatively royalties can be paid on a per unit or per use basis, however, this is relatively uncommon in life sciences. Royalties allow the licensor participating in the commercial success of the product. The licensee shares the commercial risk with the licensor and reduces upfront and milestone payments. Stacked royalties are an elegant way to account for the licensor’s optimistic view of the project. Typically, the licensor claims that the product will have an extraordinary commercial success. The licensee rewards the licensor with attractive royalties if the product excels and limits payments if the product underperforms through a reduced royalty rate. License fees: These are yearly payments, fixed or in relation to the development phase. License fees are common with academia, but do not play an important role in biotech-pharma contracts. These payments are suitable for the licensing of a technology, machinery, or instrumentation where the life of the license is not predefined. The licensor has a desire for safe fixed cash flows and wants to avoid any downside risk, while the licensee thinks there is an upside potential. Guaranteed fixed minimum/maximum annual payments: Annual cash payments of a fixed maximum or minimum cash payment, mostly on a yearly basis. The licensee prevents this way that the licensor profits excessively from the upside potential. Equity: The licensee compensates the licensor by giving equity instead of cash for the license. Alternatively, the licensee signs the obligation to buy equity from the licensor at a certain time during the contract. Equity can be in form of preferred stocks, common stocks, options, or convertible debt and might contain anti-dilution clauses. Traditionally, universities are paid by spin-off/start-up companies in form of equity because these lack cash. The university this way diversifies its license revenues that do not anymore solely depend on the success of the licensed product, but on the whole

License Contract Valuation

149

portfolio of the licensee. Equity investments of the licensor avoid fastidious fund raising of the licensee and tighten the partnership between the two counterparties. Equity investments are also a good instrument to bypass taxes. When receiving a large upfront payment, the licensor might have to pay taxes, because he suddenly made a profit. An equity investment at an overvalued price is a good way to incur an upfront payment, without having to account for it as income. Well understood, the overvalued price is intentionally a part of the agreement between the two parties. Right of first look: The licensee buys the right to have the first look at the pipeline of the licensor. This way the licensee keeps the first mover advantage in a technology he believes in. Call-back option: The licensor has the right to license back a product from the licensee at a later stage. A recent example of this is the licensing back of the rennin inhibitor Aliskiren, which was developed by Speedel and called back by Novartis after phase 2 trials (see box). Typically, the project initially does not fit into the licensor’s pipeline, but wants to keep the back door open in case of success. Sublicensing: Sublicensing occurs if an in-licensed product is out-licensed at a later time to a third party. This contract structure is common with projects biotech in-licenses from academia and out-licenses at a later development stage to pharma. The first licensor does not want to impede his licensee from sublicensing the project due to inflexible deal- terms. Therefore he participates directly in the sublicensing terms with a participation rate. This participation can be a percentage, with floor and ceiling, and maybe linked to either milestone payments or royalties. Anti-stacking: Anti-stacking is a protection of the licensee against possible patent infringement claims from third parties. In case of such a claim, the license payments to the licensor reduce by the third party claim. Nevertheless, contracts often include a lower limit. Anti-stacking allows the licensee to insure potential damage that arises due to non-completeness of patents. The license contract was closed assuming that the underlying patents provide complete protection for the resulting product. If this is not the case, then it is only fair that the licensor does not receive the full amount of agreed license terms. Co-development: licensor and licensee jointly develop a project sharing costs or the actual development work. The licensor wants to strengthen its marketing capabilities and the licensee wants to reduce its costs of introducing a new product.

150

Valuation in Life Sciences

Co-marketing: Both parties separately market the licensed product under different trademarks. Co-promotion: Both parties jointly market a product under the same name. The different clauses allow adjusting any license contracts in terms of: 1. Technical risk 2. Commercial risk 3. Liquidity preferences 4. Diverging sales assumptions (…) Speedel partners with big pharma companies in early technical and clinical development. We in-license compounds at the point where data from animal models prove of interest, but before clinical trials in humans have begun. We then perform all the early technical and clinical development, through Phase II of development, where proof-of-concept in man is established. Speedel takes on all the financial risk itself throughout this development programme. The pharma company usually retains a “call back” option, so if Speedel is successful in establishing proof-of-concept in man, the pharma company may license the compound back and proceed to full development and commercialization, activities that big pharma companies perform exceptionally well. For example, in September 2002, Novartis exercised its call-back option from Speedel for Aliskiren (SPP100), an orally active renin inhibitor targeted for cardiovascular indications. In November 2003 Novartis announced that it was moving Aliskiren into Phase III. Speedel had completed Phase I and the first ever successful Phase II development programme for this innovative class of molecule after in-licensing it from Novartis in 1999. Speedel relieves the pharma company of the full burden of risk at this critical juncture in the development cycle, utilizing its own human, operational, and financial resources. The pharma company concomitantly avoids early stage risk and enjoys downstream opportunity; for its part, Speedel receives milestone and royalty payments. In cases where Speedel is successful, and for whatever reason, the big pharma partner (licensor) chooses not to call back the compound, Speedel is free to continue development and commercialization or to offer the asset to another third party licensee. In this scenario, Speedel represents a new source of late-stage compounds for the industry. For example, in May 2003

License Contract Valuation

151

Speedel exercised its option to license from Roche all rights to an endothelin A antagonist (SPP 301) and Roche relinquished its call-back right, although it retains its rights to future royalty revenues. Having completed Phase I and IIa studies, (…) Source: The Wallstreet Reporter Online Academic institutions, medtech, pharma, and biotech have different motivations for licensing. Universities: Universities are not in the position to enter into cost intensive R&D, e.g. clinical trials for medical device or new drugs. Academia is the premier source of innovative advances in the life sciences industry and has an interest in capitalising on this. Most large universities today have a technology transfer office, taking care of licensing the research findings. Licensing partners can either be pharma, biotech, or an academic spinout / start-up company. While biotech and pharma will agree to license contracts with remuneration in milestones, royalties and license fees, spinouts and start-up companies often cannot. Short in cash, these companies will provide the university with an equity stake in the company. As academia is licensing at a very early stage, it is well possible that the licensee will sublicense the project/technology at a later time. In this case, the university needs to define suitable sublicensing parameters ensuring adequate participation in the upside potential of the deal. The preference of academia is to generate a steady income stream of royalties, license fees and milestones while keeping the technical and commercial risk at a minimum. Biotech: Biotech is acting on two fronts. First it in-licenses, often from academia, and second, it out-licenses to pharma or large biotech. There are numerous biotech companies solely built on in-licensed findings from academia, not only at the start but also later in order to fuel the pipeline along the growth of the company. Biotech, at any time in its life, is short of cash. Licensing is therefore an opportunity to fill the cash reserves. While the contracts usually have large upfront and milestone payments, they also contain royalties. So, biotech reduces its technical and commercial risk and at the same time keeps an upside by giving away its projects prior to the cost intensive final clinical and marketing phases. Pharma: Pharma combines the expertise in clinical trials with a large marketing and commercial capability. To keep the machinery running, pharmaceutical companies need to replace existing products loosing patent pro-

152

Valuation in Life Sciences

tection with new innovative products. Thanks to the financing structure of the biotech industry, pharma has a wide field of innovative products to look at and license in. While liquidity is not an issue, pharma needs to reduce the technical risk, often due to imperfect information at in-licensing, by tying the payments to the achievement of certain milestones. Medtech: In medtech the situation is similar to the one we have just described for the drug development industry. Due to shorter development times and cheaper trials sublicensing is more seldom. However, the motivations of academia, small, and big players remain the same. The Value Share Principle in Licensing When negotiating a license contract, we have to judge what a fair license contract is. A common present approach is to define how to split the value of the project between the two parties depending on the phase. The numbers for drug development value sharing are9: Table 3.31. Value share figures Licensor

Licensee

Preclinical

10%-20%

80%-90%

IND

20%-40%

60%-80%

Phase IIb/III

40%-60%

40%-60%

Approval

60%-80%

20%-40%

The more advanced the product is, the higher is the share the licensor gets, because he has borne more risk. We will discuss in a separate chapter how to approach licensing negotiations and the problem with the value sharing method. For illustrative purposes, we will use the method in the following examples on license contract valuation. Valuation of License Contracts: DCF

Like with projects, to value a license contract we first define the cash flows, adjust them with their probability to occur, discount them with the discount rate of the licensor or the licensee, and sum them up. 9

Source: Pharmaventure, Recap, Burrill.

License Contract Valuation

153

Cash Flows In addition to the cash flows related to R&D and marketing expenses, we now need to add upfront payments, license fees, milestones, and launch costs. We will value the license contract from both perspectives, the licensee’s and the licensor’s. Time. We again have to define the timing of the different cash flows. License fees will be due on an annual basis, while milestones will be triggered at the beginning of new development stages after successful accomplishment of the antecedent phase. Size. When we consider the size of the cash flows, we need to account for the perspective of valuation. For the licensee, all license payments to the licensor are expenses. Special care is needed when calculating the royalties. First, it needs to be defined on what basis the royalties are calculated; usually it is on net sales. Then we need to consider the case of stacked royalties, e.g., we have royalties of 5% on net sales up to $ 100 mn, and 7.5% if the net sales exceed $ 100 mn. In an example, we assume now net sales of $ 120 mn. Depending on the contract structure, we now either receive 5% * $ 100 mn +7.5% * $20 mn, or we apply the royalty rate of 7.5% on the entire amount of $ 120 mn. This difference is often neglected, even if it considerably influences the valuation. The peak sales need to be adjusted to the commercial potential of the licensee. We know that pharma has more marketing power than biotech and is more likely to reach higher sales. This especially matters for general care products. But not only will pharma better commericalise the drug, it usually also performs the clinical trials for a project in several indications at the same time, while biotech cannot do that due to financial constraints. Pharma launches the different indications earlier than biotech and has a steeper sales ramp. Probability to occur. The probabilities need to be chosen in accordance with the party being in charge of the relevant phase. If we know that pharma has other success rates than biotech, we need to use the corresponding success rates for all phases pharma is in charge of. If the license contract is split into various indications then the estimation of probabilities becomes delicate. Published success rates relate to compounds. As soon as one indication succeeds, the compound is counted as having passed the phase, no matter how many indications failed before. Success rates for indications are therefore lower. We will elaborate on this in more detail in the chapter about multi-indication projects.

154

Valuation in Life Sciences

Discount Rate The value of a license to the in-licensing pharma company should be calculated with the discount rate corresponding to this company. Pharma conducts the trials, decides on continuing and abandoning, and markets the drug as if it were its own. However, pharma suffers from information asymmetry, i.e. it does not exactly know all the flaws of the project. This might justify a slight spread that comes on top of pharma’s discount rate. From the outlicensing perspective the question of the discount rate is trickier. On the one hand, the project is now controlled by a more stable company not subject to risks that can typically break a small company’s neck like liquidity, nondiversification, or lack of experience. This also justifies the use of pharma’s discount rate even for the out-licensing biotech company. Nevertheless, biotech’s shareholders will still ask for a higher cost of equity than for a pharmaceutical company. What really happens is that biotech’s risk profile changes due to the license. The positive cash flows come much earlier and are more certain. On the other hand, biotech can only cash in a fraction of the sales by royalties. The prospect of less imminent losses and earlier profits reduces the discount rate of biotech. When applying the value share principle, we should however use pharma’s discount rate. Pharma’s information asymmetry is already sufficiently compensated by the fact that they do not have to pay the full price, but only the mentioned share. Biotech on the other side has due to the lower discount rate access to more value in milestones. It is well possible that the upfront payment is already a large part of the project’s value if valued with biotech’s discount rate, leaving not much for the other milestones and royalties. If calculated with pharma’s discount rate, the value of biotech’s share is much larger, allowing also reasonable milestone and royalty payments. The above-mentioned value share principle supposes a project value using pharma’s discount rate. But in the end the counterparties’ needs and negotiation skills decide on the license terms and the discount rate representing a reasonable value share. In the case where pharma licenses a project to biotech because it does not really fit in its portfolio, or because it does not see too much value in it, or when an academic institution licenses a project to a biotech, then the above method is hard to justify. The discount rate would now be much higher than the company’s standard discount rate. Biotech has a higher discount rate than pharma and academic institutions. When valuing the project with the licensee’s, hence biotech’s discount rate, and applying on top of that the value share principle, the upfront payment must be ridiculously small to allow some milestones and royalties. This hap-

License Contract Valuation

155

pens simply because the project has already a much smaller value to biotech than to its counterparties because of its high discount rate. In reality, even academic institutions apply the value share principle assuming a pharma discount rate. Valuation Using the discounted cash flows method, we must perform the same steps as with project valuation twice, for the licensor, and for the licensee: 1. Adjust each cash flow by its probability to occur. This means that we multiply each cash flow with its probability P. 2. Discount each cash flow back to the present or to the date we want to know the value for. 3. Sum up all probability adjusted and discounted cash flows to the net present value. As described above, only the cash flows change with respect to a normal project valuation. The licensee has additional expenses, namely the license payments. The licensor on the other hand is free from all further costs (assuming a full license without any co-development) and starts to earn money with the project immediately. The following case illustrates how to value standard license contracts between two parties. The emphasis is put on learning how to modify different deal parameters without influencing the value of the contract. Case Licensing with Two Parties Mid-Sized Pharma has a new CEO, who has decided that the company needs to focus its R&D activities. After reviewing the pipeline, the CEO thinks that Mid-Size Pharma excels best by keeping and building up its cancer portfolio, and by out-licensing all projects in other indications. Therefore, Dr. Project Manager’s project will be out-licensed, preferably to a global pharmaceutical company. Mrs. License Officer is in charge of finding an attractive counterparty and closing the license deal. In order not to loose time she contacts the strategic alliances and licensing teams of several large pharmaceutical companies. While these are studying the technical dossier, Mrs. License Officer wants to prepare for the negotiations. For the calculation of the project value, she studies Dr. Project Manager’s spreadsheet. Studying recent deals, she finds out that the average value share Mid-Sized Pharma

156

Valuation in Life Sciences

should receive is 20-40% of the whole deal value. A crucial point she is not sure about is what discount rate to use in the valuation. She therefore has an initial talk with the licensing manager of the most promising buyer, BigPharma. BigPharma internally uses a discount rate of 9% to value the pipeline. Mrs. License Officer and her colleague from BigPharma agree that Mid-Sized Pharma reduces its risk associated with the project by outlicensing and BigPharma enters new risks through information asymmetry in the deal. They consider a value share of 40%-60% a good deal for both sides and agree to use a discount rate of 11% (40%-60% weighted average of 14% and 9%) to discount the license deal. Mrs. License Officer is now ready to define the different deal structures that give a value split of 40%-60% based on the following input parameters: Table 3.32. Project parameters Clinical Phase 1 Clinical Phase 2 Clinical Phase 3 Approval Phase Costs Success Rate Length Launch Costs Peak Sales Discount Rate

$ 9 mn 63% 1 year $ 120 mn $ 241 mn 11%

$ 22 mn 42% 2 years

$ 80 mn 75% 3 years

$ 2.7 mn 95% 2 years

After trying several deal terms, Mrs. License Officer comes up with the following three propositions. Table 3.33. License parameters case study 2 parties Emphasis on

Upfront

Milestones

Royalties

Upfront payment: Milestone entry into clinical phase 2: Milestone entry into clinical phase 3: Milestone entry into approval phase: Launch of the drug: Royalty rate: Value share Midsize Pharma Value share BigPharma

$ 3.78 mn $ 1 mn $ 1.5 mn $ 5.5 mn $ 10 mn 3% 40% 60%

$ 1.5 mn $ 2 mn $ 4 mn $ 11 mn $ 18 mn 3% 40% 60%

$ 1.5 mn $ 1 mn $ 1.5 mn $ 5.5 mn $ 10 mn 5.5% 40% 60%

License Contract Valuation

157

Solution of Case Study: DCF The value of the project amounts to $ 22 mn. The value can also be calculated using the same spreadsheet as for case study 1 by setting the discount rate to 11%. Determination of cash flows. For the in-licensing company – BigPharma – the license payments are additional expenses to the project. In the DCF calculation it is therefore sufficient, to add the license payments to the expenses of the project. The license payments include upfront, milestones, and royalties. While upfront and milestone payments have a fixed size, royalty payments are a percentage of the sales revenues. In year I the royalty payment amounts to: Royalties(i)=SalesRevenues(i)*RoyaltyRate

(3.18)

Note that if the contract includes royalty stacking then the royalty rate depends on the level of the sales revenues. Technically, royalties correspond to a reduction of the licensee’s margin. Assuming a royalty rate of 5.5%, like in the license contract with emphasis on royalty, we receive for year 9: Royalties(9)=SalesRevenues(9)*RoyaltyRate=12*5.5%=$ 0.7 mn With this we perform the same DCF calculation as before by increasing the costs at the beginning of each phase by the upfront and milestone payments, and by reducing the margin by the royalty rate. However, beware that royalties are only due until expiry of patent protection. Afterwards BigPhara does not pay any further royalty to the licensor. For Mid-Sized Pharma, the only relevant cash flows are the license payments from BigPharma. The DCF value to Mid-Sized Pharma does not depend anymore on the costs of the project, still assuming that all development and marketing costs are entirely paid by BigPharma. In our example, we have only changed the discount rate. All other input parameters have remained the same. In practice when valuation a license, the licensor has to consider the increased capabilities of the licensee, e.g. better R&D or a strong marketing department. This would request to increase the peak sales or to adjust the launch costs. An in-depth discussion on this can be found in the section on negotiation. Probabilities of cash flows. We must adjust all cash flows by their probability to occur. This gives us the formula for the risk adjusted net cash flows (rnCF) for the licensee:

158

Valuation in Life Sciences

(3.19)

rnCF(i)=Prob(i)*(SalesRev(i)-OpExp(i)-R&DExp(i)-LicPay(i)) In year 9 this corresponds to: rnCF(9)=Prob(9)*(SalesRev(9)-OpExp(9)-R&DExp(9)-LicPay(9)) =19%*(12-4-40-11)=19%*(-43)=-$8 mn

While for the licensee the license payments are expenses, they are revenues for the licensor. The licensor’s risk adjusted net cash flows are therefore: (3.20)

rnCF(i)=Prob(i)*LicPay(i) In year 9 we have for the licensor: rnCF(9)=Prob(9)*LicPay(9)=19%*10.7=$ 2.0 mn

Discounting of cash flows. Finally, we have to discount the risk adjusted net cash flows in order to get the risk adjusted net present cash flows. These summed up then give us the risk adjusted net present value (rNPV) for the licensee or the licensor. The following table displays the cash flows for each year to BigPharma according to the contract that puts an emphasis on royalties. Note that MidSized Pharma receives only royalties as long as patent protection last, i.e. until the year 23.

Pharma

Year 1

Year 2

Year 3

Year 4

Year 5

Year 6

Year 7

Year 8

Year 9

Phase 1

Phase 2

Phase 2

Phase 3

Phase 3

Phase 3

Approval

Approval

Market

Sales curve Peak Sales

19% $ 241 mn

$ 241 mn

$ 241 mn

$ 241 mn

$ 241 mn

$ 241 mn

$ 241 mn

$ 241 mn

$ 241 mn ($ 4 mn)

($ 16 mn)

($ 9 mn)

($ 11 mn)

($ 11 mn)

($ 27 mn)

($ 27 mn)

($ 27 mn)

($ 3 mn)

($ 40 mn)

($ 40 mn)

($ 40 mn)

$ 12 mn

Operating expenses Licence Expenses

Market

$ 241 mn

Sales revenues R&D Expenses

Year 10 5%

($ 2 mn)

($ 1 mn)

($ 2 mn)

($ 6 mn)

($ 11 mn)

$ 46 mn

($ 3 mn)

Year 11

Year 12

Year 13

Year 14

Year 15

Year 16

Year 17

Year 18

Year 19

Year 20

Market

Market

Market

Market

Market

Market

Market

Market

Market

Market

Sales curve

36%

Peak Sales

$ 241 mn

$ 241 mn

$ 241 mn

$ 241 mn

$ 241 mn

$ 241 mn

$ 241 mn

$ 241 mn

$ 241 mn

$ 241 mn

$ 87 mn

$ 123 mn

$ 154 mn

$ 181 mn

$ 202 mn

$ 219 mn

$ 231 mn

$ 239 mn

$ 241 mn

$ 241 mn

($ 30 mn)

($ 43 mn)

($ 54 mn)

($ 63 mn)

($ 71 mn)

($ 77 mn)

($ 81 mn)

($ 84 mn)

($ 84 mn)

($ 84 mn)

Sales revenues Operating expenses

51%

64%

75%

84%

91%

96%

99%

100%

100%

R&D Expenses Licence Expenses

Sales curve

($ 5 mn)

($ 7 mn)

($ 8 mn)

($ 10 mn)

($ 11 mn)

($ 12 mn)

($ 13 mn)

($ 13 mn)

($ 13 mn)

($ 13 mn)

Year 21

Year 22

Year 23

Year 24

Year 25

Year 26

Year 27

Year 28

Year 29

Year 30

Market

Market

Market

Market

Market

Market

Market

Market

Market

Market

53%

37%

24%

13%

6%

1%

Peak Sales

$ 241 mn

$ 241 mn

$ 241 mn

$ 241 mn

$ 241 mn

$ 241 mn

$ 241 mn

$ 241 mn

$ 241 mn

$ 241 mn

Sales revenues

$ 239 mn

$ 234 mn

$ 229 mn

$ 174 mn

$ 128 mn

$ 89 mn

$ 58 mn

$ 31 mn

$ 14 mn

$ 2 mn

($ 84 mn)

($ 82 mn)

($ 80 mn)

($ 61 mn)

($ 45 mn)

($ 31 mn)

($ 20 mn)

($ 11 mn)

($ 5 mn)

($ 1 mn)

($ 13 mn)

($ 13 mn)

Operating expenses

99%

97%

95%

72%

R&D Expenses Licence Expenses

Fig. 3.40. DCF calculation for licensor

License Contract Valuation

159

The following table shows the DCF calculation for BigPharma. The value of the in-licensed project to BigPharma, using a discount rate of 11%, amounts to $ 13.2 mn. Comparing this with the value without licensing payments, $ 22.0 mn, we see that BigPharma receives exactly 60% of the project value. Length (y) 1 2 3 2

Phase 1 Phase 2 Phase 3 Approval Launch costs Peak Sales Growth Rate Operating Margin

Costs $ 9 mn $ 22 mn $ 80 mn $ 3 mn

Suc. Rate 63% 42% 75% 95%

Deal Terms Upfront Phase 2 Phase 3 NDA Approval Royalties

$ 120 mn $ 241 mn 0% 65% Pharma Mid-Sized Pharma 11% 14%

Discount Rate

Pharma Sales curve Peak Sales Sales revenues Operating expenses R&D Expenses Licence Expenses

Year 1 Phase 1

Year 2 Phase 2

$ 2 mn $ 1 mn $ 2 mn $ 6 mn $ 10 mn 5,5% -2,010221601

Year 3 Phase 2

Year 4 Phase 3

Year 5 Phase 3

Year 6 Phase 3

Year 7 Approval

Year 8 Approval

Year 9 Market 5% $ 241 mn $ 12 mn ($ 4 mn) ($ 40 mn) ($ 11 mn)

Year 10 Market 19% $ 241 mn $ 46 mn ($ 16 mn) ($ 40 mn) ($ 3 mn)

$ 241 mn

$ 241 mn

$ 241 mn

$ 241 mn

$ 241 mn

$ 241 mn

$ 241 mn

$ 241 mn

($ 9 mn) ($ 2 mn)

($ 11 mn) ($ 1 mn)

($ 11 mn)

($ 27 mn) ($ 2 mn)

($ 27 mn)

($ 27 mn)

($ 3 mn) ($ 6 mn)

($ 40 mn)

Success Rate Probability Risk adjusted net CF Discount Net present CF

63% 100% ($ 11 mn) 100% ($ 11 mn)

100% 63% ($ 8 mn) 90% ($ 7 mn)

42% 63% ($ 7 mn) 81% ($ 6 mn)

100% 26% ($ 7 mn) 73% ($ 5 mn)

100% 26% ($ 7 mn) 66% ($ 5 mn)

75% 26% ($ 7 mn) 59% ($ 4 mn)

100% 20% ($ 2 mn) 53% ($ 1 mn)

95% 20% ($ 8 mn) 48% ($ 4 mn)

100% 19% ($ 8 mn) 43% ($ 4 mn)

100% 19% ($ 2 mn) 39% ($ 1 mn)

rNPV IRR

$ 13,2 mn 13,1%

88% $ 1 mn 90% $ 1 mn

77% $ 0 mn 81% $ 0 mn

67% $ 0 mn 73% $ 0 mn

59% $ 0 mn 66% $ 0 mn

52% $ 0 mn 59% $ 0 mn

46% $ 0 mn 53% $ 1 mn

40% $ 0 mn 48% $ 0 mn

35% $ 1 mn 43% $ 1 mn

31% $ 0 mn 39% $ 0 mn

Mid-Sized Pharma Own Discount Net present CF Pharma Discount Net present CF Value 1 Value 2

100% $ 2 mn 100% $ 2 mn $ 7 mn $ 9 mn

Fig. 3.41. rNPV calculation for BigPharma

Using the same discount rate, the licensor’s value must be exactly the difference, because the following equation holds: CF(licensee)=CF(project)-CF(licensor)

(3.21)

This means that the cash flows for BigPharma are exactly the cash flows that belong to the project, and in addition BigPharma has to transfer the license payments to Mid-Sized Pharma, which are exactly all the cash flows Mid-Sized Pharma receives. And since all other parameters remain the same as in the normal project valuation without the license contract, especially the probabilities of the cash flows, we can deduce that in the DCF framework the value of the project can be split up in the licensee’s part and the licensor’s part, if all are calculated with the same discount rate. rNPV(project)=rNPV(licensee)+rNPV(licensor)

(3.22)

160

Valuation in Life Sciences

$ 8.80 mn

$ 13.18 mn

BigPharma

Mid-Sized Pharma

Fig. 3.42. Value share licensee-licensor

This means that the rNPV value of a project that is not licensed corresponds to the sum of value shares of the licensor and the licensee if the same input parameter is applied (success rates, duration, peak sales, discount rate). Modifying the deal terms, i.e. upfront payments, milestone payments and royalty rates, we can find several deal structures that fit into the 60%-40% value split. This is exactly, what Mrs. License Officer has done. The valuation of a license with DCF is the same as the valuation of a project. If the cash flows are carefully considered in the calculations, rNPV is a simple and straightforward way to value a license. Valuation of License Contracts: Real Options

We have learned in the previous section how to value license contracts with DCF. The differences between project and license contract valuation are the adjustment of the input parameters to the licensee’s capabilities, and the implementation of contract related cash flows. The adjustment of parameters concerns mainly the peak sales estimate, the development costs, and the discount rate. The contract related cash flows encompass upfront payment, milestones payments, and royalties. The remaining valuation part stays exactly the same: adjusting the cash flows with probabilities, discounting them back, and summing them up to the rNPV. As in project valuation, ROV models the uncertainty of the project’s economic success. According to its profitability, the project is continued or abandoned. But in a license contract we have two parties. And the project must be profitable for both. For the licensor this reasoning is quickly concluded: Instead of facing large investments with the perspective of uncer-

License Contract Valuation

161

tain income, the licensor securitised its project value and thereafter receives only positive cash flows. Since no investment is due anymore, the project will always be profitable for the licensor, no matter what happens. For the licensee however, the situation is more similar to the one discussed in project valuation. At the start of each phase, he must balance expenses and potential income to decide whether he still wants to pursue the project. Therefore it is possible that the licensee abandons a project, although this is against the interests of the licensor. While the consideration of the option to abandon adds value to the licensee, it deprives the licensor of safe earnings. In the case of abandonment, all rights on the project fall back to the licensor. Obviously, he can try to license the project to another partner, but this will probably have the same reasons not to continue the project as the first licensee. The flexibility of management to react to changes of the project’s market potential creates normally value. In the case of licensing this is only true for the licensee. The option approach yields lower values for the licensor, because some scenarios with positive cash flows are wiped out against the interest of the licensor. The reason for this interesting circumstance is the control over the project that is transferred from the licensor to the licensee with the contract. This control enables management to enforce decisions. The licensor is now short in control. For the valuation itself, the change of control has a significant impact. In order to know the value for the licensor we must know whether the licensee continues the project or not. We therefore must value the project also for the licensee. Real Options Valuation of License Contracts 1. Span a binomial tree for the peak sales estimate 2. Value the project in each leaf of the tree for the licensee 3. Work back the tree to the root for the licensee. For each time step 1. Take expectation of up- and down-step 2. Apply success rate if necessary 3. Discount for the elapsed time step 4. Add cash flows 5. Decide if it is a decision point 4. Value the project in each leaf of the tree for the licensor 5. Put value to zero if licensee abandons project in that state 6. Work back the tree to the root for the licensor. For each time step

162

Valuation in Life Sciences

1. 2. 3. 4. 5.

Take expectation of up- and down-step Apply success rate if necessary Discount for the elapsed time step Add cash flows Put value to zero if licensee abandons project in that state

⎧⎪ ⎫⎪ − dt t + dt ) Vt licencee = max ⎨(1 + r ) Pt ( pVupt + dt + (1 − p )Vdown + CFt ,0⎬

N

⎪⎩ 3 ⎪ 2 4 1 ⎭

(3.23)

5

⎛ ⎞ (3.24) − dt t + dt ) Vt licencor = ⎜ (1 + r ) Pt ( pVupt + dt + (1 − p)Vdown + CFt ⎟ ⋅ I V licencee ≠ 0

N ⎟ ⎜ t

2 3 4 1 ⎝ ⎠ 5

with I V licencee ≠ 0 t

⎧⎪1 if Vt licencee ≠ 0 =⎨ ⎪⎩0 else

Usually, DCF yields higher values for the licensor than real options valuation. This must not always be the case. If the license deal includes royalty stacking, it is possible that the real option value is higher than the DCF value. Assume for this an easy example. A project is clearly in the money and will not be abandoned. Peak sales estimates are at $ 400 mn with no growth, and the royalty structure envisions 10% for all sales under $ 500 mn and 20% for sales exceeding this barrier. If we value this deal with DCF we will earn exactly 10% royalties. With real options valuation on the other hand, the tree might also display scenarios with peak sales that are higher than $ 500 mn. On average, we will therefore earn more than 10% royalties with under real options valuation. Value for the licensee. At the beginning of each phase we must account for the upfront and later for the milestone payments, during commercialisation until patent expiry we must reduce the operating profit by the royalty payments. Using this set of cash flows, we can apply the same procedure to value an in-licensed project as for a self-conducted project. Actually, the license payments just increase the costs and later derogate revenues.

License Contract Valuation

163

Value for the licensor. Once we have worked back the tree for the licensee we know exactly in which states the licensee abandons the project. We can now work back the tree also for the licensor, but considering that in these states the value is truncated to zero. Solution Case Study: ROV As in case study 1, we first span the tree. Second, we calculate the value of the project in all end nodes for the licensee. Third, we work back the tree for the licensee. In addition, we then value the project in the end nodes for the licensor, and finally fifth, work back the tree for the licensor.

Sales curve Peak Sales Sales revenues Operating expenses R&D Expenses Licence Expenses

Year 7 Approval

Year 9 Year 10 Market Market 5% 19% $ 40 mn $ 40 mn $ 40 mn $ 40 mn $ 2 mn $ 8 mn ($ 1 mn) ($ 3 mn) ($ 3 mn) ($ 40 mn) ($ 40 mn) ($ 40 mn) ($ 6 mn) ($ 10 mn) ($ 0 mn)

Year 11 Market 36% $ 40 mn $ 14 mn ($ 5 mn)

Year 12 Market 51% $ 40 mn $ 20 mn ($ 7 mn)

($ 1 mn)

($ 1 mn) ($ 1 mn)

($ 2 mn)

($ 2 mn)

($ 2 mn)

Success Rate Probability Risk adjusted net CF Discount Net present CF

100% 95% 100% 100% 100% 100% 95% 95% ($ 8 mn) ($ 40 mn) ($ 46 mn) ($ 34 mn) 100% 90% 81% 73% ($ 8 mn) ($ 36 mn) ($ 38 mn) ($ 25 mn)

100% 95% $ 8 mn 66% $ 5 mn

100% 100% 95% 95% $ 11 mn $ 14 mn 59% 53% $ 7 mn $ 8 mn

100% 95% $ 17 mn 48% $ 8 mn

100% 95% $ 19 mn 43% $ 8 mn

100% 95% $ 20 mn 39% $ 8 mn

Year 25 Market 53% $ 40 mn $ 21 mn ($ 7 mn)

Year 26 Market 37% $ 40 mn $ 15 mn ($ 5 mn)

Year 27 Market 24% $ 40 mn $ 10 mn ($ 3 mn)

Year 28 Market 13% $ 40 mn $ 5 mn ($ 2 mn)

100% 100% 95% 95% $ 18 mn $ 13 mn 17% 15% $ 3 mn $ 2 mn

100% 95% $ 9 mn 14% $ 1 mn

100% 95% $ 6 mn 12% $ 1 mn

100% 95% $ 3 mn 11% $ 0 mn

Sales curve Peak Sales Sales revenues Operating expenses R&D Expenses Licence Expenses Success Rate Probability Risk adjusted net CF Discount Net present CF rNPV

Year 8 Approval

Year 19 Year 20 Year 21 Year 22 Year 23 Year 24 Market Market Market Market Market Market 100% 100% 99% 97% 95% 72% $ 40 mn $ 40 mn $ 40 mn $ 40 mn $ 40 mn $ 40 mn $ 40 mn $ 40 mn $ 39 mn $ 39 mn $ 38 mn $ 29 mn ($ 14 mn) ($ 14 mn) ($ 14 mn) ($ 14 mn) ($ 13 mn) ($ 10 mn) ($ 2 mn)

($ 2 mn)

($ 2 mn)

($ 2 mn)

100% 95% $ 23 mn 29% $ 6 mn

100% 95% $ 23 mn 26% $ 6 mn

100% 95% $ 22 mn 23% $ 5 mn

100% 95% $ 22 mn 21% $ 5 mn

100% 95% $ 23 mn 19% $ 4 mn

Year 13 Year 14 Year 15 Year 16 Market Market Market Market 64% 75% 84% 91% $ 40 mn $ 40 mn $ 40 mn $ 40 mn $ 25 mn $ 30 mn $ 33 mn $ 36 mn ($ 9 mn) ($ 10 mn) ($ 12 mn) ($ 13 mn)

($ 13.7 mn)

Fig. 3.43. Value of the lowest end node

Construction of the binomial tree: The binomial tree is constructed in exactly the same way as in the case study on project valuation. Valuation of end nodes for licensee: We now value the project with DCF for every end node as if it were at the beginning of approval phase, using the peak sales estimate of the end node. The table below displays the calculation for the lowest end node. The final value of the project, if it reaches the approval phase with peak sales of $ 40 mn using the royalty weighted license contract amounts to $ (13.7) mn. The licensee would therefore decide to abandon the project. Calculating the values for all end nodes we get:

164

Valuation in Life Sciences Year 1 Phase 1

Year 2

Year 3

Year 4

Year 5

Phase 2 Phase 2 Phase 3 Phase 3

Year 6 Phase 3

Year 7 Approval $ 1'458 mn

$ 1'080 mn $ 3'433 mn $ 800 mn $ 593 mn $ 439 mn $ 325 mn $ 241 mn

$ 593 mn $ 1'634 mn $ 439 mn

$ 325 mn $ 241 mn

$ 179 mn

$ 325 mn

$ 957 mn

$ 179 mn

$ 475 mn

$ 241 mn

$ 132 mn $ 98 mn

Peak Sales

$ 439 mn

$ 241 mn $ 179 mn

$ 132 mn Legend:

$ 800 mn

$ 132 mn $ 98 mn

$ 211 mn

$ 54 mn

$ 66 mn

$ 73 mn

$ 73 mn

Value

$ 40 mn ($ 14 mn)

Fig. 3.44. Licensee’s value of the end states

In the continuation we set the value of the lowest node to zero, representing the halt of the project in that state. Working back the tree as licensee: There is no difference in the procedure to get back to the root node compared to common project valuation, with the exception that we must now also consider the milestone payments and at the root node the upfront payment. The final values are shown in Fig. 3.45. We see that the licensee would abandon the project in two nodes, in the one we have already discussed at the start of approval phase, and in the lowest at phase 3. We can retrieve the value of the license contract to the licensee from the root node, $ 13.3 mn. This compares to $ 13.2 mn with DCF (actual difference amounts to $ 0.16 mn). We see that real options valuation makes only a small difference. This is due to the lower discount rate than in case study 1. The project is with a discount rate of 11% clearer in the money than with 14%. Valuation of end nodes for licensor: With DCF we value the project for the licensor in each end node, i.e. as if the project is at the beginning of the approval phase, with peak sales corresponding to the end node. This yields the situation shown in Fig. 3.46.

License Contract Valuation Year 1 Phase 1

Year 2

Year 3

Year 4

Phase 2 Phase 2 Phase 3

Year 5

Year 6

Year 7

Phase 3

Phase 3

Approval

165

$ 1'458 mn $ 1'080 mn $ 3'433 mn $ 800 mn $ 1'672 mn $ 593 mn $ 1'066 mn

$ 800 mn

$ 593 mn $ 1'834 mn

$ 439 mn $ 656 mn

$ 439 mn

$ 872 mn

$ 439 mn

$ 325 mn $ 160 mn $ 325 mn

$ 532 mn

$ 325 mn

$ 957 mn

$ 241 mn $ 80 mn $ 241 mn $ 299 mn

$ 241 mn

$ 433 mn

$ 241 mn

$ 13.3 mn $ 179 mn $ 60 mn $ 179 mn

$ 239 mn

$ 179 mn

$ 475 mn

$ 14 mn $ 132 mn $ 103 mn

$ 132 mn

$ 192 mn

$ 132 mn

$ 6 mn

$ 78 mn

$ 98 mn

$ 211 mn

$ 73 mn

$ 60 mn

$ 73 mn

($ 8 mn)

$ 54 mn

$ 66 mn

($ 8 mn)

$ 40 mn

Legend: Peak Sales Value

$ 98 mn $ 0 mn

$ 0 mn

Fig. 3.45. Licensee’s value of the root node

Note that in the lowest node BigPharma abandons the project. This means that Mid-Sized Pharma does not receive any milestone payments and royalties, although they would be worth $ 21 mn. The value must be reduced to $ 0 mn. Working back the tree as licensor: We now work back the tree using the cash flows of the licensor, i.e. milestone payments. At a node of a decision point we then must check, whether the licensee would continue the project in that state, otherwise we must set the licensor’s value to zero. This procedure explains, why real options valuation yields a lower value for the licensor than DCF. The same nodes are put to zero as in the tree for the licensee. The final value for the licensor amounts to $ 8.1 mn compared to $ 8.8 mn. While with DCF we had an exact value share of 40%-60%, the same deal, calculated with real options valuation, corresponds now to a value share of 37.7%-62.3%. Mrs. License Officer notices that the same deals, once calculated with DCF, once with real options valuation, do not have the same value share. Under the real option perspective Mid-Sized Pharma does not get what BigPharma promised, namely a 40% share in the project. Mrs. License Officer therefore modifies the deal terms such that the value share fits again.

166

Valuation in Life Sciences Year 1 Phase 1

Year 2

Year 3

Year 4

Year 5

Phase 2 Phase 2 Phase 3 Phase 3

Year 6 Phase 3

Year 7 Approval $ 1'458 mn

$ 1'080 mn $ 800 mn $ 593 mn $ 439 mn $ 325 mn $ 241 mn

$ 241 mn $ 179 mn

$ 800 mn $ 593 mn

$ 171 mn

$ 325 mn

$ 100 mn

$ 439 mn $ 325 mn

$ 439 mn

$ 241 mn $ 179 mn

$ 132 mn

$ 241 mn $ 179 mn

$ 132 mn $ 98 mn

Legend:

$ 61 mn $ 132 mn

$ 98 mn $ 73 mn

Peak Sales

$ 300 mn

$ 39 mn $ 73 mn

$ 54 mn

Value

$ 28 mn $ 40 mn $ 21 mn

Fig. 3.46. Value of the end nodes for the licensor Year 1 Phase 1

Year 2

Year 3

Year 4

Year 5

Phase 2 Phase 2 Phase 3 Phase 3

Year 6 Phase 3

Year 7 Approval $ 1'458 mn

$ 1'080 mn

$ 300 mn

$ 800 mn

$ 153 mn

$ 800 mn

$ 593 mn $ 104 mn

$ 593 mn

$ 171 mn

$ 439 mn $ 73 mn $ 439 mn

$ 88 mn

$ 439 mn

$ 325 mn $ 21 mn $ 325 mn $ 61 mn

$ 325 mn

$ 100 mn

$ 241 mn $ 16 mn $ 241 mn $ 44 mn $ 241 mn

$ 52 mn

$ 241 mn

$ 8.1 mn $ 179 mn $ 13 mn $ 179 mn $ 37 mn

$ 179 mn

$ 61 mn

$ 8 mn $ 132 mn $ 28 mn $ 132 mn

$ 33 mn

$ 132 mn

$ 5 mn $ 98 mn $ 24 mn

$ 98 mn

$ 39 mn

$ 0 mn

$ 73 mn

$ 22 mn

$ 73 mn

$ 13 mn

$ 54 mn

$ 28 mn

$ 8 mn

$ 40 mn

Legend: Peak Sales Value

$ 0 mn

Fig. 3.47. Licensor’s value of the root node

License Contract Valuation

167

Table 3.34. Adjusted ROV deal terms Emphasis on

Upfront

Milestones

Royalties

Value share Mid-Sized Pharma DCF Value share BigPharma DCF Upfront payment: Milestone entry into clinical phase 2 Milestone entry into clinical phase 3 Milestone entry into approval phase Launch of the drug Royalty rate Value share Midsize Pharma Value share BigPharma Difference to DCF

39.9% 60.1% $ 3.8 mn $ 1 mn $ 1.5 mn $ 5.5 mn $ 10 mn 3% 40% 60% $ 0.02 mn

37% 63% $ 1.5 mn $ 2 mn $ 4 mn $ 11 mn $ 27.5 mn 3% 40% 60% $ 9.5 mn

37.7% 62.3% $ 1.5 mn $ 1 mn $ 1.5 mn $ 5.5 mn $ 10 mn 6.1% 40% 60% 0.6%

While the upfront deal makes barely any difference, the changes for the milestones and royalty deals look attractive. In order to receive again 40% of the project, the two parties need to raise the approval milestone by $ 9.5 mn in the milestone deal, or lift the royalty rate to 6.1%. While a $ 20,000 increase of the upfront payment is hardly recognizable, the changes in milestones and royalties could easily fund an early phase of a future project. Mrs. License Officer, after considering the different license contract structures, decides to aim at a deal with emphasis on the milestones valued with real options. Discussion The case illustrates the important points when preparing a license contract valuation. The choice of the input parameters, especially the peak sales and the discount rate, is not trivial. All input parameters should reflect the characteristics of the licensing partner once the project is in his hands. Big pharma has the capabilities to get the maximum out of a project, while earlystage companies often have to start from scratch, i.e. building-up drug development and marketing expertise. This should be considered in the cost, duration, and success of the project. We might even assume that pharma has higher success rates in clinical trials, but there is no sound data set to adjust the drug development parameters to the license partner. Therefore, we use standard data as outlined in the chapter on fundamentals in life science. When valuing a license contract we need to consider the risk profile of the licensee. It is evident that if big pharma acquires a license, the default

168

Valuation in Life Sciences

risk is minimal and the financial strength of big pharma ensures that the project will not be hindered due to lack of cash. On the other hand, big pharma is stringent in stopping projects that do not pass the hurdle rate, i.e. are not profitable enough. One third of all clinical stage projects will be terminated due to this reason. This of course is not in the interest of the licensor, who after licensing has only positive cash flows. It is therefore a trade-off to enter a license with a large partner. It adds development and marketing power to the project, thereby pushing its success; on the other hand it adds the risk that if certain criteria are not met the project is halted even if it there are no concerns related to efficacy or safety. This would never happen in a small biotech company, because without the project the company would lose its right of being. To split up the project value between the two parties, it is necessary to value both the licensee’s and the licensor’s part with the same discount rate. This discount rate lies somewhere in-between the two usually applied discount rates. A weighted average of the two discount rates according to the value split is a reasonable rate to agree on. The value split method will be discussed in further detail in the section about negotiation further down. Not only the choice of the right input parameters is critical, but also the choice of the valuation method. DCF and real options valuation yield different valuation outcomes in our case. Even if at first sight it seems to be negligible, it makes a major difference in late stage license payments. While a $ 20’000 upfront payment increase is marginal, a $ 9.5 mn milestone increase at launch makes a difference. DCF can well be used to value license contracts, but does not consider the aspect that the licensee might give up a project. Real options are closer to reality. The only additional input parameter is the volatility. While negotiating a license, it is advisable to value the contract with both methods. A difference can be used to negotiate better license terms for the licensor. Diatos SA, a Paris-based privately held biopharmaceutical company, and Medarex, Inc. (Nasdaq: MEDX) have announced that Diatos has licensed from Medarex the exclusive European rights to develop and commercialize Super-Leu-Dox, a potential new cancer treatment. Under the terms of the agreement, Diatos expects to develop and commercialize Super-Leu-Dox, which is currently in late pre-clinical development, in Europe. In return, Medarex has an option, exercisable before Super-Leu-Dox reaches Phase III development, to co-develop and co-commercialize the product with Diatos in Europe. If Medarex <<

Advanced Topics in Project and License Contract Valuation

169

exercises its option, the parties will share equally all development and commercialization costs and future potential revenues from sales and sublicensing of Super-Leu-Dox in Europe. If Medarex elects not to codevelop and co-commercialize the product in Europe, Diatos will have the right to develop and commercialize Super-Leu-Dox in Europe and will be responsible for all development activities and costs. In that event, Medarex will receive royalties on any commercial sales of the product. Under the terms of the agreement, Medarex retains all rights to Super-Leu-Dox in non-European countries. Each party has the right to use pre-clinical and clinical data generated by the other party for developing Super-Leu-Dox in their respective territories. (…) Source: World Biotech Online

Advanced Topics in Project and License Contract Valuation Early-Stage Licensing with Possible Sublicensing

Research institutions and academia license their projects already at early stages, even before preclinical testing. Their licensees are often biotech companies that take the projects from research into preclinical and early clinical development. These companies mostly lack the possibilities to conduct expensive large-scale clinical trials. Therefore, they hand the project over to pharmaceutical companies, which then take the project through clinical phase 3 to the market. The biotech companies find themselves between academia and pharma and are linked with both via license contracts. Once both license contracts are in place, biotech just earns the differential of the two contracts. This means that it receives milestone payments or royalties from pharma and has to pass a part of it on to academia. For academia, this business model has two important implications: First, in early stage the potential of a project is very uncertain. Large deal parameters are hard to justify, especially if the potential licensees are chronically short of cash. Second, as soon as biotech enters into negotiations with pharma, the license contract between academia and biotech serves as a lower barrier. If by that time research revealed that the compound could only be used in a an indication with limited market potential, the original contract between academia and biotech might already be too high, impeding the sublicensing to a potent partner.

170

Valuation in Life Sciences License

Academia

Biotech Sublicense

Pharma Fig. 3.48. Licensing and sublicensing

The industry came up with a solution for structuring early stage license contracts such that first, academia has an upside potential even though their licensees are normally financially weak biotechs, and second, biotech is not encumbered by high deal terms in their sublicensing negotiations: As soon as the project is sublicensed to a third party, academia takes a share in the new license terms. Typically, the license contract includes a clause that says that the licensor is entitled to the sublicense deal terms by a certain percentage, depending on the stage the sublicense contract was concluded. Until that moment, the biotech company must redeem academia according to the base scenario of the contract. The base scenario is a normal license contract between the two parties, assuming that the licensee himself takes the product to the market. These sublicensing parameters allow academia offering relatively cheap license contracts, as it has an upside potential when its licensee sublicenses the project at better terms. On the other hand, it also bears the risk to be worse off if the project turns out to be not as attractive as anticipated. Academia accepts to share parts of the financial risk in exchange of a considerable upside potential. Table 3.35. Sublicensing parameters Sublicensed in

Precl.

CP1

CP2

Clinical phase 2 Clinical phase 3 Approval phase Commercialisation

Appr.

Comm.

50% share

Preclinical phase Clinical phase 1

CP3

Base sc.

40% share

Base scenario Base scenario Base scenario Base scenario

30% share 20% share 15% share 10% share

Advanced Topics in Project and License Contract Valuation

171

Biotech on the other hand is free of any constraints when negotiating with a potential partner. If the original contract between academia and biotech envisioned 4% royalties, but pharma is only ready to pay 3.5%, this would normally be a deadlock. Biotech could not even cover the royalties it has to pay to academia. However, since academia now does not insist on the original terms but agrees to just participate in the new deal terms by, let’s say 50%, 3.5% royalties is still acceptable for biotech. Academia then receives 1.75% of the royalties paid by pharma to biotech and profits from their commercial capability. Academia is then still better off by accepting smaller deal terms compared to obstructing a deal with inflexible terms and not receiving any further payments at all. Note, that this contract structure implicitly admits that the value of the project or the sublicense contract is not known yet. It can turn out well, in this case academia profits from the upside, but it can also go bad, in this case it is important to save what can be saved and not to obstruct sublicensing negotiations with an unrealistic base scenario. Some institutions add special features to these contracts. With a view to forcing the biotech company not to give too much way on the royalty side, they insist for instance on the original royalty rate. In the above example, the institution could maintain its claim of 4% royalties. The biotech must therefore negotiate higher royalties to the expense of lower upfront and milestone payments. Fixed royalties are typical for large, diversified institutions. They rely mostly on royalty payments, while milestone payments account only for a small fraction of their license income. Others take a percentage of the sublicensing terms, but fix a minimum. This reduces the downside. In order to define fair sublicensing parameters, we must value the contract. We have a set of different possible scenarios, namely that the biotech company launches the product itself (base scenario), or that it out-licenses the project in one of the specified stages. Each of these scenarios has a value. The base scenario corresponds to a normal license contract between academia and biotech and is therefore straightforward to value. As soon as the sublicensing terms become active, i.e. as soon as we have to assume that biotech further sublicenses the project, we must suppose a hypothetical license contract. We put ourselves at the time of sublicensing and figure out the structure of a fair license contract between biotech and pharma for each of the scenarios. This each time corresponds to the situation where Mrs. License Officer has found herself in the previous case; e.g. if biotech sublicenses the product at the end of phase 2, we design a hypothetical contract between pharma and biotech at the entry of phase 3. Having defined all cash flows between pharma and biotech, i.e. the upfront and mile-

172

Valuation in Life Sciences

stone payments, and the royalties, we can now deduce the cash flows for academia stemming from this sublicensing based on the participation rate for sublicensing in phase 3. Having done this, we have now defined the cash flows for all three parties for the single scenario that biotech sublicenses after phase 2. Before sublicensing

After sublicensing

Academia

Academia

Base contract

Participation rate

Biotech

Biotech Sublicense contract

Pharma Fig. 3.49. Participation rate in sublicensing

We repeat that procedure for all possible moments of sublicensing. Having the value for all scenarios does not mean that we know the value of the contract. Finally, we do not know which scenario will happen. The actual value is the probability weighted average of all scenarios. Vcontract =

∑ P(i )V

i

(3.25)

i

With P(i) being the probability of scenario i. If we determine the value of a given contract, then we apply the above probability weighing formula. In this case the worst valued scenario already indicates a lower bound for the value of the license contract (all weight on the worst scenario), and the best valued scenario an upper bound (all weight on the best scenario). Either the strategy of the licensee already displays a clear preference for one scenario, in which case we can assume the value of this specific scenario, or we really must chose a value in the indicated range. However, it is factually impossible to quantify the probabilities of the scenarios. These depend on the financial strength of the biotech company, its strategic focus, its investors’ moods, and potential sublicensee’s appetite. Either we endeavour despite of that to quantify the probabilities, or, if still in negotiation, we can use an elegant solution for this problem: academia chooses a contract

Advanced Topics in Project and License Contract Valuation

173

structure that is independent of these probabilities. This is typically the case if all scenarios have the same value, i.e. if biotech conducts and launches the product on its own, or licenses it at any stage. It makes sense that academia chooses the sublicensing terms in a way that the value of the contract does not depend on the biotech company’s strategy or fate. This method then corresponds to negotiating the value of the base contract and then to determine all sublicensing terms such that the corresponding scenarios all give the same value. This technique should be applied by technology transfer managers of universities and research institutions. Theory DCF

Valuation of a license contract with participation in sublicensing terms requires the valuation of each sublicensing scenario. To quantify these scenarios we assume license contracts at each possible stage of sublicensing. Knowing these license contracts, we can deduce the cash flows of all parties and value the contract scenarios. According to the purpose of the valuation we then either can set the sublicense participations such that all scenarios have the same value, or we can weigh the scenarios in order to get the value of the contract. Sublicensing Case: DCF Mr. Tech Transfer is in charge of licensing a molecule from its research institute to the industry. The molecule has just passed the lead validation phase and should now be taken to lead optimisation. There is a strong rationale for the compound being used as a CNS drug. A young biotech Table 3.36. Project characteristics Lead Opt. Preclinical Phase 1 Phase 2 Costs Success Rate Length Launch costs

$ 0.8 mn 70% 1 year $ 300 mn Market Peak Sales $ 650 mn Margin 65% Discount Rate 15%

$ 2 mn 70% 2 years

$ 4 mn 66% 1 year

Phase 3

Approv.

$ 15 mn $ 60 mn $ 2 mn 46% 62% 78% 2 years 3 years 2 years

174

Valuation in Life Sciences

100% 80% 60% 40% 20%

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24

0%

Fig. 3.50. Sales curve Table 3.37. Licensing parameters Upfront payment Lead Opt. Preclinical Phase 1 Phase 2 Phase 3 Approval Phase Launch

Milestone Payments

Royalties

$ 0.2 mn $ 0.3 mn $ 0.8 mn $ 1.5 mn $ 3 mn $ 4 mn

3%

$ 0.1 mn

company that wants to broaden its pipeline is interested in the molecule. Mr. Tech Transfer therefore prepares for the negotiations. In a first step, he values the project with standard industry parameters as outlined in table 3.36. Using a standard sales curve as displayed and a discount rate of 15%, Mr. Tech Transfer comes up with a net present value of $ 1.4 mn. Mr. Tech Transfer aims at a value split of 12%-14% and designs a contract that corresponds to these figures. This license contract has a value of $ 1.9 mn, or 14% of the project value. The final contract that Mr. Tech Transfer has in mind uses these terms as base scenario and then specifies for each possible stage of sublicensing a participation rate. Mr. Tech Transfer now has defined the base contract, i.e. the contract with biotech assuming no sublicensing, and tries to calibrate the sublicense participation rates in a way corresponding to a value of $ 1.9 mn for each possible sublicensing scenario. The final contract should include potential sublicensing at the beginning of preclinical phase, clinical phase 1, 2, and 3, and approval phase. Mr. Tech Transfer now calculates for each scenario based, on the value share figures we have outlined earlier, the

Advanced Topics in Project and License Contract Valuation

175

sublicense terms. For the sublicense contract where biotech out-licenses to pharma in preclinical phase for instance he needs to know upfront, milestone payments, and royalties that would correspond to a value share of 20%-80% for the same project, assuming that it is already in preclinical phase. Mr. Tech Transfer designs the sublicense contracts on the basis of a discount rate of 12%, assuming that the sublicensee will be a mid-sized pharmaceutical company. He then comes up with the following terms for the different phases of sublicensing: Table 3.38. Sublicense contracts biotech-pharma In $ mn Moment of Sublicense

Upfront Milestone Payments Royalties Phase 1 Phase 2 Phase 3 NDA Launch

Preclinical Phase 1 Phase 2 Phase 3 Approval Phase

1.3 3.4 11.2 46.0 134.0

2.5

3.8 6.8

5.0 10.1 22.4

6.3 13.5 33.5 92.0

7.5 16.9 44.7 137.9 268.1

4.7% 6.1% 8.8% 13.8% 28.5%

Mr. Tech Transfer now values each sublicensing scenario as if it were a contract on its own. Instead of using the milestones and royalties of the base scenario, he assumes a given participation rate in the sublicensing terms once the project is sublicensed. The table below represents the numbers he must use for the valuation. The light grey fields correspond to the base contract, the dark grey fields to the sublicensing participation. We see that the payments from the sublicense deals are linked with the participation rate. Table 3.39. Base contract and sublicensing participation Lead Precl. CP 1

CP 2

CP 3

NDA

Launch

Royalties

xp*6.3

xp*7.5

xp*4.7%

Precl.

0.1

xp*1.3 xp*2.5 xp*3.8

xp*5.0

Phase 1

0.1

0.2

x1*3.4 x1*6.8

x1*10.1 x1*13.5

x1*16.9

x1*6.1%

Phase 2

0.1

0.2

0.3

x2*11.2 x2*22.4 x2*33.5

x2*44.7

x2*8.8%

Phase 3

0.1

0.2

0.3

0.8

x3*46.0 x3*92.0

x3*137.9 x3*13.8%

Approval 0.1

0.2

0.3

0.8

1.5

xA*134.0 xA*268.1 xA*28.5%

176

Valuation in Life Sciences

Assuming a dummy participation rate, Mr. Tech Transfer can value each scenario. He then modifies the participation rate for each scenario in a way that the scenario value equals exactly the claimed $ 1.9 mn of the base contract. This he can do either by playing around with the participation rates, or by using the goal seek method in Excel (see exercises). He comes up with the following participation rates: Table 3.40. Participation rates Moment of Sublicense

Participation rate

Preclinical

38%

Phase 1

28%

Phase 2

19%

Phase 3

11%

Approval Phase

7%

These participation rates mean that, for instance if the licensee decides to sublicense in phase 3, Mr. Tech Transfer’s institute would receive 11% of all sublicense terms, i.e. milestones of 11%*46.0=$ 5.1 mn for phase 3, 11%*92.0 =$ 10.1 mn for approval phase, 11%*137.9=$ 15.2 mn at launch, and 11%*13.8%=1.5% royalties. Discussion

Once the project is sublicensed, the license revenues for academia change according to the sublicense terms and the participation rate. This can lead to substantially higher payments if biotech manages to reach a good deal, but can also result in a reduction of future revenues due to a poor sublicense deal. As seen in the previous case, even average deals can change the size of specific payments considerably, e.g. 3% royalties compared to 1.5% if the project is sublicensed in phase 3. It is possible, to gear these changes with specific clauses in the contract. A research institute for instance can insist on its royalty claim and only apply the participation rate to upfront and milestone payments. This however means that biotech must at least negotiate this royalty rate in the sublicense contract, otherwise it would have to pay more royalties to academia than it receives from pharma. This narrows biotech’s leeway in the negotiation and might lead to no deal at all. The participation rate has been introduced precisely to avoid such deadlocks in negotiations.

Advanced Topics in Project and License Contract Valuation

177

Ultimately, a poor license contract is still better than none. Remember that for academia all cash flows stemming from the license deals are positive. If biotech must halt the project because it cannot sublicense it to a third party, then academia does not receive any license payments anymore. An attenuated version of the fix royalty claim is the introduction of a minimum royalty rate. Academia participates in the sublicense deal at the agreed rate, but receives at least the minimum royalty rate. With this, the possibility of a deadlock in sublicense negotiations still exists, but it is less likely, because biotech is given a bit more leeway now. Of course, the same clauses can also apply to milestone payments. How does this translate into the valuation? If academia insists on its royalty claim from the base scenario, then the participation rate refers just to the milestone payments, and only the cash flows related to sublicense milestones are linked with the participation rate. The light grey fields in the table are the licence, the dark grey the sublicense parameters. Table 3.41. License and sublicense parameters Lead. Precl

CP 1

CP 2

CP 3

NDA

Launch

Royalties

xp*6.3

xp*7.5

3%

Precl.

0.1

xp*1.3 xp*2.5 xp*3.8

xp*5.0

Phase 1

0.1

0.2

x1*3.4 x1*6.8

x1*10.1 x1*13.5

x1*16.9

3%

Phase 2

0.1

0.2

0.3

x2*11.2 x2*22.4 x2*33.5

x2*44.7

3%

Phase 3

0.1

0.2

0.3

0.8

x3*46.0 x3*92.0

x3*137.9 3%

Approval 0.1

0.2

0.3

0.8

1.5

xA*134.0 xA*268.1 3%

If we use a minimum royalty rate, then the actual royalty rate would be whichever is higher, either the minimum royalty rate or the participation rate times the royalty rate of the sublicense contract. For all these different contract forms, the DCF calculation works with fixed cash flows. Either the cash flows are really predefined, as is the case in the base scenario, where the milestones and royalty rate are clearly specified. Or we apply a participation rate to the license terms of the sublicense contracts. These however are assumed to be predefined as well. In the valuation it makes no difference whether we value a contract scenario with participation rates, or a contract scenario with predefined terms that equal the sublicense terms times the participation rate. The cash flows are equal. In reality these two contract structures are not the same. The partici-

178

Valuation in Life Sciences

pation rates have explicitly been introduced to make a difference, that is to say to offer biotech more chances to sublicense the compound and to have at the same time also an upside in case the sublicense deal turns out to be better than anticipated. Both motives are value enhancing, but not considered in the DCF calculation. Both motives deal with the uncertainty about future sublicensing scenarios. We do not know yet how well biotech manages to sublicense the project. Nevertheless, the rigidity of DCF forces us to assume hypothetical sublicense contracts, contrary to the very concept of participation rates and neglects that the predicted deal scenario is highly uncertain On these grounds, DCF has severe shortcomings when applied to the valuation of license contracts with sublicensing. We will elucidate how real options cope with this problem in the next section. Theory Real Options

In the previous discussion about DCF valuation of contracts including sublicense participation rates we have noticed two important drawbacks of the DCF method: 1. DCF does not account for the possibility of deadlocks, i.e. the possibility that no viable sublicense contract can be closed. 2. DCF uses predefined sublicense contracts and consequently cannot quantify the upside academia has due to proportional participation also in above average sublicense deals. The first point deals with the option of the sublicensee not to enter a deal. The second point addresses the uncertainty of the future sublicense contracts; its terms might be better or worse than anticipated. Both points call for the use of real options valuation. First, the option to decline a deal lies at the heart of the real options concept. Second, the trigger of the decision to enter a deal or not depends on the conditions, i.e. the sublicense contract terms, and these depend on the quality of the project. This is exactly what we allow to fluctuate in real options valuation and more precisely in the binomial tree; the quality of the project, measured in monetary units. In order to value such a license contract with real options, we must therefore first, model the sublicense contract in function of the quality of the project, i.e. in function of the peak sales, and second, model the decisions of the sublicensee.

Advanced Topics in Project and License Contract Valuation

179

Generic license contracts. As we have seen before in the section about license contracts, these contracts can be defined by the discount rate and the value share each party gets. Upfront payment, milestone payments, and royalties can however be arranged in many ways to reach the same value share. Hence, if we want to value a sublicensing scenario, we must assume a generic license contract. This can be defined by the following parameters:

1. Applied discount rate 2. Value share licensor-licensee 3. Value split upfront-milestones-royalties 4. Relative weights between milestones The discount rate is necessary to define the project value that is to be shared between the two parties. With the value share we get the licensor’s piece. A value repartition between upfront payments, milestone payments, and royalties allows the determination of how much each term contributes to the value and allows to adjust the risk profile of the contract. This then directly determines the upfront payment, and the royalty rate can be deduced. For the milestones, we further need the relative weights between them. The calculation of a generic contract is shown in the example below. Assume the following generic license contract in clinical phase 3: Table 3.42. Generic contract Discount rate

12%

Licensor-licensee value share

50%-50%

Upfront-Milestones-Royalties

30%-30%-40%

Milestone weights (approval, launch)

1:2

Assume that the project valued with 12% yields a value of $ 600 mn. The licensor captures 50% of this, i.e. $ 300 mn. 30% of this value stem from the upfront payment, 30% form the outstanding milestone payments, and the remaining 40% from the royalties. This means that the upfront payment equals $ 90 mn. The sum of the risk adjusted and discounted milestones equals $ 90 mn as well, while the risk adjusted and discounted royalty streams equal $ 120 mn. From this we can deduce the milestones using the following development:

180

Valuation in Life Sciences

Vmilestones = p3 (1 + r ) 3 n + p3 papproval (1 + r ) −t

− t 3 − t approval

2n

(3.26)

This formula indicates the rNPV of all milestone payments, i.e. the part of the license contract value that can be attributed to the milestone payments.

n=

p3 (1 + r )

−t3

Vmilestones −t −t + 2 p3 papproval (1 + r ) 3 approval

(3.27)

Using the $ 90 mn as Vmilestones, 12% for the discount r, and the corresponding success rates (60% for phase 3 and 90% for approval phase) and durations (three years for phase 3 and one year for approval phase), we can deduce x, the milestone unit. In this specific case we get: n=

60%(1 + 12% ) n = $ 80.8 mn

−3

$ 90 mn −5 + 2 ⋅ 60%90%(1 + 12% )

The relative milestone weights tell us that the approval phase milestone equals $ 81 mn and the launch milestone $ 162 mn, two times the unit. If now academia receives 20% of all cash flows stemming from a sublicensing deal in phase 3, it receives 20% of the upfront payment at the beginning of phase 3, i.e. $ 18 mn, 20% of $ 81 mn at the beginning of approval phase, i.e. $ 16 mn, and 20% of $ 162 mn at launch of the compound, i.e. $ 32 mn. For the royalties we use a similar development: Vroyalties = p3 papproval

∑ (1 + r )

− t 3 − t approval − i

roy ⋅ salesi

(3.28)

i

This formula shows how the rNPV of all royalty streams depends on the royalty rate roy. To deduce this rate we must isolate roy:

roy =

(3.29)

Vroyalties

p3 papproval

∑ (1 + r )

− t 3 − t approval − i

salesi

i

The above formula yields the royalty rate that corresponds to a rNPV of all royalty streams of Vroyalties, in our case $ 120 mn. We can deduce the license contract terms from the above given generic structure. Assuming a generic structure for each stage, i.e. from preclinical through to commercialisation, we are able to deduce all required cash flows we for the valuation.

Advanced Topics in Project and License Contract Valuation

181

Let us assume Mr. Tech Transfer’s project at the beginning of phase 3. The targeted generic license structure looks as follows: Table 3.43. Generic contract for phase 3 Moment of Sublicense

Value Share

MilestonesRoyalties

Phase 3 NDA Launch

Phase 3

50%-50%

50%-50%

1

2

3

The project should be split into two equal parts (we assume a joint discount rate of 12%), half of its value should be contributed by milestone payments (assuming the upfront payment as the phase 3 milestone payment), and the relationship between the milestone payments of phase 3, NDA, and Launch is 1:2:3. The value of the entire project amounts to $ 497.7 mn. The licensee should get 50% thereof, i.e. $ 248.8 mn. Milestones make up 50%; the other 50% are contributed by royalty payments. Assuming n to be the milestone for phase 3, the milestone at NDA would amount to 2n, and the launch milestone to 3n. The rNPV of the milestone payments can be written as: n=

1 + 2 p3 (1 + r )

−t3

Vmilestones −t −t + 3 p3 papproval (1 + r ) 3 approval

(3.30)

Using the actual numbers this gives the following solution: 50%$248.8mn −3 −5 1 + 2 ⋅ 62%(1 + 12% ) + 3 ⋅ 62% ⋅ 78%(1 + 12% ) n = $46.0mn n=

With this we can deduct the size of the milestones for phase 3 ($ 46.0 mn), NDA ($ 92.0), and launch ($ 137.9 mn). For the royalty rate we have to proceed in a similar way: (3.31) Vroyalties roy = − t 3 − t approval − i p3 papproval (1 + r ) salesi

∑ i

50% $ 248.8 mn roy = = 13.8% 62%78% $ 1,864 mn

182

Valuation in Life Sciences

Note that the deduction of the license terms as described corresponds to a DCF calculation. Of course, this could also be done by using a real options approach, but the calculation would be much more complex. For this exercise, that is to say to know approximately the terms of the sublicense contract, this approach is more illustrative. As a second remark, we must add that the above-mentioned method is only one possibility to define a generic contract structure. As we will see in the chapter about negotiations, the contract terms do not necessarily have to be defined by a value share method, but can also be determined by some other rationale. Valuation of scenarios. As with DCF, we must first value each possible sublicensing scenario. The valuation of a specific sublicensing scenario comprises however some special features that have to be addressed. First, we cannot simply span a tree up to approval phase and then work back to the root node as we have done for project and license valuation. Place yourself in an end node at approval phase like in the figure. In order to value the project in that node, we must know the terms of the sublicense contract that is valid in this node, i.e. the milestone payments and royalties, otherwise we cannot determine the cash flows necessary for the valuation. The sublicense contract is determined at the time of sublicensing (e.g. start of phase 1 for the scenario with sublicensing in phase 1), in the “sublicense nodes”. The figure displays that an end node is accessible from various sublicense nodes. Hence, the sublicense terms in the end node depend not only on the end node, but also on the path that lead to this end node, or better on the sublicense node that has been passed on that path. It is therefore not possible to value the tree from the end nodes, because we do not know the sublicense terms. However, we do know how to value license contracts from the previous section, and a sublicense is not any different from a license contract with respect to valuation. According to this, we can value the sublicense contract at the sublicense nodes as displayed in the figure. We can therefore attribute a value to each sublicense node with the algorithm outlined in the previous section on license contracts. As a consequence, we only need to span the tree from the date of valuation up to the time of sublicensing. We can then determine, due to the assumed generic structure of the sublicense contract, the upfront and milestone payments and royalties for the valuation of the sublicense contract in each sublicense node. For the valuation of the sublicense contract we

Advanced Topics in Project and License Contract Valuation

Base Scenario

Sublicence Contract

Determination of Sublicence Contract

Approval Phase

Fig. 3.51. Various paths leading to the same end node

Base Scenario

Sublicence Contract

Fig. 3.52. Nodes that can have the same sublicense contract

183

184

Valuation in Life Sciences

then need to span a tree from the time of sublicensing to approval phase in each sublicense node. This obviously increases the complexity of the valuation considerably. As a second particularity, license contracts with sublicensing involve not only two, but three parties, namely academia, biotech, and pharma. We therefore have to value the project for three parties. Although we might only be interested in the value to academia, we must value the project from the view of biotech and pharma as well, because they are in control of the project from some time and consequently decide whether to continue or abandon the project. For the valuation of the sublicense contract in the sublicense nodes, we must value the project for pharma in order to know whether the project is still alive or abandoned because of economic reasons. While valuing the project for pharma, we must use the sublicense payments paid to biotech and the academias participation thereof to value the sublicense contract for biotech and academia. As a result, the sublicense valuation provides three values, one for pharma, one for biotech, and one for academia. Then we can work back the tree from the sublicense nodes

Manual for Valuation of License Contracts with Sublicensing

1. Define license contract as if biotech commercialises the project (base scenario) 2. Define possible points of sublicensing (sublicensing scenarios) 3. Value each sublicensing scenario a. Span tree until moment of sublicensing b. Generate for each end node a possible sublicense contract c. Value the sublicense contract for each pharma, biotech, and academia using ROV for license contracts d. Set the project values at the end nodes to the values of the sublicense contract for academia and biotech respectively e. Work back the tree to the root node using ROV for license contracts, assuming the license terms of the base scenario f. Choose participation rates such that sublicensing scenarios and base scenario have the same values If we value an already signed contract, then we just have to value all scenarios (point 3) and probability-weigh them.

Advanced Topics in Project and License Contract Valuation

185

back to the root node using the known algorithm for real options valuation of license contracts. In this part of the tree, the license payments correspond to the ones of the base scenario. Valuation of the contract. With the algorithm described above we manage to value each scenario of the license contract between academia and biotech. As with DCF, we should now aim at calibrating the participation rates in a way that all scenarios have the same value like the base scenario. If the contract is already signed and the participation rates therefore are fixed, we deal again with the same problem as in DCF, we must quantify the probabilities of each scenario. One possibility to bypass the somehow arbitrary weighing of scenarios is the following: Biotech chooses the scenario that represents the greatest value to it. This argument uses the same logic real options are based on, maximisation of value. Although this reasoning is correct, we should consider that these scenarios have very different risk profiles. While early sublicensing offers biotech revenues in the near future, it also limits considerably the upside. The more risk biotech is ready to bear, i.e. the further it takes the compound on its own, the higher the upside. In reality value maximisation is just one criterion in strategic planning, aspects like risk minimisation, focus on core business, or avoidance of dilution might indicate other strategic options. Sublicensing Case: ROV Mr. Tech Transfer feels a bit disturbed that his DCF valuation does not make a difference between a standard license contract with predefined terms and a license contract with participation rates. He is convinced, that his institute as well as the partnering biotech company profits from the introduction of participation rates by facilitating sublicensing negotiations. But his DCF calculation does not visualise this effect. He therefore decides to value the same project with real options. Using exactly the same input parameters like in the DCF calculation and a volatility of 30%, Mr. Tech Transfer first works out a base license contract between his institute and the biotech company. Again, he aims at a 14% value share, using a discount rate of 15%. The contract is calculated like any standard license contract between two parties. Details can be found in the first part of the licensing chapter. Mr. Tech Transfer comes up with the following base scenario:

186

Valuation in Life Sciences Table 3.44. License parameters with real options Upfront payment Milestone Payments Royalties ∆ to DCF

Lead Opt.

$ 0.1 mn

$ 0 mn

Preclinical

$ 0.25 mn

$ 0.05 mn

Phase 1

$ 0.45 mn

$ 0.15 mn

Phase 2

$ 1 mn

$ 0.2 mn

Phase 3

$ 1.5 mn

$ 0 mn

Approval Phase

$ 3 mn

$ 0 mn

Launch

$ 4 mn

3.25%

0.25%

The value of this license deal amounts to $ 2.0 mn for academia, and $ 12.2 mn for biotech, corresponding to a value split of 14%-86%. Mr. Tech Transfer notes satisfied that the use of real options valuation helps him justifying higher deal terms in the base contract. He then goes on and defines the points of sublicensing. As in the DCF case these points are start of preclinical development, start of clinical phases 1-3, approval phase and at commercialisation of the product. He now defines what the generic sublicense contracts should look like at the respective points of sublicensing. For consistency, he chooses generic contracts that correspond to the contracts he has used in the DCF calculation. The numbers in the phase rows represent the weights of the milestones relative to one another: Table 3.45. Generic sublicense contracts with real options Moment of Value Sublicense Share

Milestones- Precl. Phase Phase Phase NDA Launch 1 2 3 Royalties

Preclinical

20%-80% 50%-50%

Phase 1

25%-75% 50%-50%

Phase 2

35%-65% 50%-50%

Phase 3

50%-50% 50%-50%

Approval Phase

70%-30% 34%-66%

1

2

3

4

5

6

1

2

3

4

5

1

2

3

4

1

2

3

1

2

Advanced Topics in Project and License Contract Valuation

187

Using a peak sales estimate of $ 650 mn and a joint discount rate of 12%, these generic license contracts provide exactly the sublicense terms as used in the DCF case. In the next step, Mr. Tech Transfer undertakes the valuation of the first scenario, i.e. the scenario where the biotech company sublicenses the project already at the beginning of preclinical phase. Therefore, he first spans the tree until the start of preclinical phase, i.e. over a time period of one year. Using a time step of one year in the tree, this corresponds to one time step. For each sublicense node Mr. Tech Transfer now calculates a license contract corresponding to the generic structure he has previously defined. He uses the same formulae as described above. US$ 877 mn

43% US$ 650 mn 57%

US$ 482 mn

Lead Optimisation

Preclinical Phase

Fig. 3.53. Going one time step forward

He must therefore first calculate the value of the project at the corresponding nodes using the DCF method. The value of the project, assuming peak sales of $ 877 mn, a discount rate of 12% and that lead optimisation is successfully passed, amounts to $ 77.9 mn, and to $ 30.9 mn assuming peak sales of $ 482 mn in the lower node. From this value biotech is expected to capture 20%, i.e. $ 15.6 mn or $ 6.2 mn. 50% thereof are allocated to royalties, and 50% to milestones. This corresponds to a royalty rate for biotech of 5.3% or 3.8%, respectively. Mr. Tech Transfer then calculates the upfront and milestone payments in the way as explained above in the section about generic license contracts. He receives the following results:

188

Valuation in Life Sciences Table 3.46. Upfront milestone and royalty payments

Sublicense Node Royalties Precl. Phase 1 Phase 2 Phase 3 NDA Launch $ 877 mn $ 482 mn

5.3% 3.8%

1.9 0.8

3.8 1.5

5.8 2.3

7.7 3.0

9.6 3.8

11.5 4.6

Mr. Tech Transfer notices that without the use of a participation rate, the biotech company would not be able to enter a license contract if the peak sales estimate deteriorates, i.e. if the project is at the lower sublicense node. The offered sublicense terms are all below the base scenario and consequently, the biotech company would have to pay more to Mr. Tech Transfer’s institute than it would receive from the pharmaceutical company, resulting in a net loss. The biotech company would therefore prefer not to sublicense the project, i.e. to let the project die. Mr. Tech Transfer’s institute wouldn’t earn any further license payments at all that way, despite the project having passed all previous phases. Using a participation fee, the institute and the biotech company would split the sublicense revenues between each other, leaving both a profit. So, how should this participation rate be chosen such that the scenario of sublicensing in preclinical phase equals the base scenario in terms of value? For this Mr. Tech Transfer performs a real options valuation of the sublicense contract for the sublicensor (the biotech company) and the sublicensee (the pharmaceutical company). Although he is not interested in the value of the sublicense contract to the pharmaceutical company, he needs to know in which states it abandons the project, as in these states the biotech company does not receive and further sublicense payments. Mr. Tech Transfer receives the following value for the share biotech receives in the sublicense contract with pharma: Table 3.47. Biotech’s share in the sublicense contract Sublicense Node $ 877 mn $ 482 mn

Value Sublicensor @ 15% $ 12.1 mn Sublicensor @ 15% $ 4.5 mn

The value of the sublicense contract at the time of sublicensing will then be split between the institute and the biotech company according to the agreed participation rate xprecl. This means that in the upper sublicense node the project would have a value of xprecl*$ 12.1 mn for Mr. Tech Transfer’s institute, and xprecl*$ 4.5 mn in the lower sublicense node. For

Advanced Topics in Project and License Contract Valuation

189

the biotech company the values amount to (1- xprecl)*$ 12.1 mn and (1xprecl)*$ 12.1 mn respectively. To simplify Mr. Tech Transfer has assumed that both, the biotech company and the institute, use a discount rate of 15%. Otherwise we would have to calculate the value of the sublicense contract for the sublicensor with the discount rate of the institute, and the institute would get xprecl thereof, and once with the discount rate of the biotech company, and the biotech company would get (1-xprecl) thereof. Table 3.48. Value distribution in sublicense nodes Sublicense Nodes $ 877 mn

$ 482 mn

Value Biotech

(1 − x )⋅12.1

Institute

x precl ⋅ 12.1

precl

Biotech

(1 − x )⋅ 4.5

Institute

x precl ⋅ 4.5

precl

Mr. Tech Transfer knows now the values for biotech and the institute in all sublicense nodes. He now works back the tree to the root node, which in this case is just one time step. For the biotech company the value is defined by the following calculation: 1 [p( 1 − x )⋅ 12.1 + (1 − p )(1 − x )⋅ 4.5]P − 0 .8 − 0 .1

N 1+ r N N precl

precl

precl 2

1

N

4

(3.32)

5

3

The numbers in the equation correspond to 1. Expectation of future values at the end of the time step 2. Success rate for current phase 3. Discounting for one time step 4. Costs for current phase 5. License payments due to the institute In a similar way Mr. Tech Transfer calculates the value for the institute: 1 [px12.1+ (1 − p )x 4.5]N P + 0 .1 N

1+ r precl

precl

1

precl 2

N

(3.33)

5

3

Trying several values for xprecl Mr. Tech Transfer comes up with the following table:

190

Valuation in Life Sciences Table 3.49. Participation rates and corresponding values

Participation Rate

Value for Institute

Value for Biotech

36% 38% 40% 42% 44%

$ 1.79 mn $ 1.89 mn $ 1.98 mn $ 2.08 mn $ 2.17 mn

$ 2.11 mn $ 2.02 mn $ 1.92 mn $ 1.83 mn $ 1.74 mn

A participation rate of 40% would yield about the targeted $ 2.0 mn from the base scenario, corresponding to a 49%-51% value split. Clearly, the scenario of sublicensing in preclinical phase does not provide biotech the original 86%-14% value split, but only a 49%-51% value split. However, since biotech contributes only the conduct of lead optimisation in this scenario, a far lower participation in the value is justified. Mr. Tech Transfer now undertakes the valuation of all other scenarios. For each possible sublicense node he designs a contract that complies with generic structure. 9,672 7,165 5,308 3,932 2,913 2,158 1,599 1,184 877 650

1,599

877 650

482

2,913 2,158

1,184

1,599

877

482

877

482

877 650

482 357

264 196

1,599 1,184

650

357 264

2,913 2,158

1,184

650

357

5,308 3,932

482 357

264 196

145

264 196

145 107

145 107

80

80 59 44

Lead

Preclinical

Phase 1

Phase 2

Fig. 3.54. Binomial tree

He receives the following intermediary results:

Phase 3

NDA

Advanced Topics in Project and License Contract Valuation

191

Table 3.50. Sublicense contracts for each sublicense node Scenario Sublicense node P1 P2 Phase 1 Phase 1 Phase 1 Phase 1 Phase 2 Phase 2 Phase 2 Phase 2 Phase 2 Phase 3 Phase 3 Phase 3 Phase 3 Phase 3 Phase 3 Phase 3 NDA NDA NDA NDA NDA NDA NDA NDA NDA NDA

265 482 877 1599 196 357 650 1184 2158 107 196 357 650 1184 2158 3932 44 80 145 265 482 877 1599 2913 5308 9672

P3

NDA Launch Roy. rate Value @ 15%

0.5 0.9 1.4 1.8 2.1 4.2 6.3 8.4 5.1 10.2 15.3 20.4 10.6 21.2 31.8 42.4 - 0.2 0.4 0.7 - 4.1 8.2 12.3 - 11.2 22.4 33.5 - 24.1 48.1 72.2 - 47.6 95.2 142.7 0 0 4.6 9.2 - 19.3 38.6 - 46.0 92.0 - 94.6 189.2 - 183.3 366.7 - 344.9 689.9 0 0 14.4 42.8 94.2 187.8 358.9 670.3 - 1237.9 - 2272.0

2.3 10.5 25.5 52.9 0.9 16.4 44.7 96.3 190.3 0 13.9 57.9 137.9 283.9 550.0 1034.8 0 0 28.7 85.6 188.4 375.7 717.8 1340.6 2475.7 4544.1

2.0% 5.1% 6.8% 7.7% 0.6% 5.9% 8.8% 10.5% 11.3% 0% 4.6% 10.5% 13.8% 15.6% 16.5% 17.1% 0% 0% 13.7% 22.3% 27.0% 29.6% 31.0% 31.8% 32.3% 32.5%

2.0 11.4 28.4 59.1 0.7 15.4 45.8 100.4 198.3 0 17.6 78.8 194.7 403.9 787.7 1485.3 0 0 49.7 172.1 446.4 919.9 1783.6 3332.4 6197.5 11368.2

Using the values of the sublicense contracts to the sublicensor, Mr. Tech Transfer can now fill the values for his institute and the biotech company at the sublicense nodes for each scenario. Working back the tree as if it were a standard license contract using the base scenario terms, he calculates the value of the contract for the root node. The only task that remains is to adjust the participation rate in a way, that the institute’s value equals $ 2.0 mn. Finally, Mr. Tech Transfer finds the following participation rates:

192

Valuation in Life Sciences Table 3.51. Comparison of DCF and ROV participation rates Moment of Sublicense

Participation Rate (ROV)

Participation Rate (DCF)

Preclinical

40%

38%

Phase 1

29%

28%

Phase 2 Phase 3

20% 12%

19% 11%

9%

7%

Approval Phase

Mr. Tech Transfer is happy that this long exercise pays off. In the license negotiations the real options valuation does not only provide him with arguments for better base scenario terms, but also suggests higher participation rates. Discussion The valuation of early stage licensing contracts with participation fees certainly belongs to the most difficult tasks in financial life sciences. The example with Mr. Tech Transfer shows important advantages over the common DCF method. First, the use of real options allows the licensor (the institute) to negotiate higher license terms, maintaining the agreed value split. The base contract in the example is considerably more favourable for the institute if negotiated with real options. Second, we note that real options completely capture the idea of the participation rate. As seen in the example, in each sublicense node the milestones and royalties change, while DCF just calculates with one set of license parameters as if these were agreed on from the beginning. In addition, it is not a big deal to implement in the model a clause saying that the licensee, i.e. the institute, receives a minimum royalty rate. This might occur following a licensee (biotech) asking for smaller participation rates. In order to compensate the value loss due to a decrease of the participation rate, the institute might insist on a minimum royalty rate. While in the real option model this clause becomes immediately visible, it might have no effect in the DCF model. Imagine the discussed scenario when sublicensing at the start of preclinical phase. The hypothetical license contract in the DCF model assumes a sublicense royalty rate of 4.7%. In the real options model we have received in the upper sublicense node a royalty rate of 5.3% and in the lower sublicense node a royalty rate of 3.8%. Assuming a participation rate of 40%

Advanced Topics in Project and License Contract Valuation

193

this would result in an effective royalty rate of 1.88% in the DCF model and 2.12% and 1.52% in the real options model. The biotech company offers you a minimum royalty rate of 1.8% and asks you how much you could accommodate with the participation rate. In the DCF model a minimum rate of 1.8% has absolutely no effect, because it is still lower than the assumed scenario of 1.88%. Consequently, the valuation does not indicate any leeway in return of the offered minimum royalty rate. This clearly doesn’t make sense. In the real options model on the other hand, the institute profits from the minimum royalty rate in the lower sublicense node. It would raise the royalty form 1.52% to 1.8%. In return, Mr. Tech Transfer can cede a little of the participation rate. In this particular example the minimum royalty rate of 1.8% offsets a lowering of the participation rate by 2% to 38%. For the valuation of these complicated contract structures, real options valuation is useful in many ways, not mainly because it includes the option to abandon the project at some point. The main argument for real options is the very nature of the contracts that explicitly assumes that at time of signing the fate of the economic potential of the project and the ability of the biotech company to sublicense it is not known yet. Real options valuation makes use of generic sublicense contracts allowing us to adjust the participation rate to the time and potential of sublicensing. DCF on the other hand has to use hypothetical license contracts ignoring any possible change of economic potential and its influence on the sublicensing contracts. Exactly this contradicts the idea of the participation rate. Real options valuation reflects precisely this idea, that we do not know yet what sublicense contract can be achieved. All contract features that consider the uncertainty with respect to product performance or sublicense deal terms can be quantified. This does not only refer to various possible sublicense contracts, but also stacked royalties, or as seen above to minimum royalties. The large set of future scenarios requires an exact and extensive financial model. The two license partners have a lot of options to design the contract according to their preferences. Some might prefer larger milestone payments to the cost of lower royalties, while others want to put more weight on the royalty side. A flexible financial model is the perfect tool to check the influence of any change in the terms. As we have seen in the example of Mr. Tech Transfer however, this is a challenging task. But considering the ease during negotiation, the flexibility to quickly evaluate the counterparty’s propositions, the gained confidence, and the potential to negotiate higher license terms more than justify a sound model, even if it

194

Valuation in Life Sciences

costs a considerable amount of work. When valuing early-stage license contracts with real options we reach the limits of the usability of spreadsheets. Nevertheless, the valuation can be elegantly programmed in Excel using VBA. Simulation of License Contracts

The simulation of license contracts allows accounting for stacked royalties. Contrary to a standard DCF calculation, in the simulations the peak sales can also take higher values than expected. This may lead to higher royalty payments than anticipated because of royalty stacking in the license contract. Imagine estimated peak sales of $ 400 mn and a royalty structure of 5% if annual sales are lower than $ 500 mn and 8% for the part exceeding $ 500 mn. In a DCF valuation we would always have to calculate with 5%, while in a simulation some scenarios are likely to have peak sales higher than $ 500 mn; in these scenarios the higher royalty rate of 8% applies. Simulations also give a good idea about the likelihood of cash flows as displayed in the figure below.

Fig. 3.55. Simulation of a phase 3 license contract (1,000 simulations calculated with ri:val)

We have simulated a phase 3 license contract. We see that in 35% we have to content ourselves with the upfront payment of $ 60 mn, this is in line with a phase 3 success rate of 65%. In very few cases we get in addition to the upfront payment also an NDA milestone before the project fails. This scenario corresponds to the second pillar in the figure. The other pillars then correspond to the scenarios when the project reaches commercialisation. According to the market potential of the compound the company earns more or less royalties.

Advanced Topics in Project and License Contract Valuation

195

Discovery and Preclinical License Deals: What’s Realistic?

We have performed an analysis of 22 recent early-stage license deals where Avance was involved in the structuring and/or valuation. The aim of the analysis was to find out what deal terms are realistic for preclinical and discovery deals, and to see if there is a clear pattern in the distribution of the milestone payments. Furthermore we wanted to see which parameters could be used as a guideline in finding the right deal terms when preparing an initial term sheet or when countering an offer. The deals that were included in the analysis are across different disease areas, including NCE and biologicals. All of them are pure drug development deals. Diagnostics were not included. The review of the 22 deals comprises a detailed analysis of all deal terms, including co-development, R&D funding, equity participation, or option payments, and a subsequent valuation of the deal at time of deal closing. For the valuation we have used ri:val®, a valuation solution specifically designed to value life science projects and licenses. The calculations are performed with a risk adjusted Net Present Value method, i.e. discounted cash flows using success rates. The input parameters were either provided by the companies themselves, or defined to our best knowledge. The sales were forecasted for each drug individually. The average upfront payments for discovery deals were US$ 0.9 mn, with the minimum being US$ 0 mn, and the maximum being US$ 5.5 mn. For preclinical deals the average upfront payment was US$ 1.9 mn. The lowest upfront for preclinical deals was US$ 0.1 mn, the highest US$ 10 mn. The relative weights of the milestone payments are displayed in Fig. 3.56. The figure shows a clear pattern for both discovery and preclinical deals. This pattern can be used when preparing a term sheet to define the milestone payments or when interpreting a news release where only the upfront payment and the total amount of milestone payments are revealed. Royalties included in most of the contracts a tiered structure. On average the royalty rate was about 4% for discovery deals and 6% for preclinical deals. As there is no clear consensus on how to define if a license deal is good or not, we have analysed the 22 deals in respect to different parameters that are used in the industry: Value share, return on investment (ROI) for the licensee, and absolute deal value. The value share principle is based on the notion that the value of the project is split between the licensor and the licensee. The licensor’s share grows bigger the later in the development the deal is closed. Fig. 4.49 displays the value share for discovery and preclinical deals.

196

Valuation in Life Sciences

38,3

40,0 35,0 30,0 25,0 20,0

18,4

15,0 10,0 1,0 5,0

1,9

0,9

1,0

6,2

4,1

0,7

1,3

2,3

Phase I

Phase II

Phase III

Preclinical Discovery

8,3

4,0

0,0 Upfront

Preclinical

NDA

Market

Fig. 3.56. Relative weight of payments 100,0% 90,0% 80,0% 70,0% 60,0%

82,4%

79,9%

17,6%

20,1%

50,0% 40,0% 30,0% 20,0% 10,0% 0,0% Discovery

Preclinical Licensor

Licensee

Fig. 3.57. Value share of discovery and preclinical deals

The value share for discovery deals is 17.6% for the licensor, and 20.1% for preclinical deals. These results are consistent with data previously reported (by ReCap). For late stage deals, the share of the licensor is in the range of 40% – 60%. A surprising finding was the consistency of the ROI in the range of 20% for most of the 22 deals analysed. The ROI was calculated only for the licensee, as the licensor in most cases has no further expenses and therefore

Advanced Topics in Project and License Contract Valuation

197

no return on investment. The average ROI for discovery deals was 20.5% and 20.8% for preclinical deals. The total deal value, i.e. the value of all future cash flows adjusted for their likelihood by their success rates and discounted back to the date of valuation, is for discovery deals US$ 40.5 mn, and US$ 55.1 mn for preclinical deals. All values are calculated using a uniform discount rate of 12%. The above figures can be used as guidelines when preparing term sheets and negotiating license deals. Particularly useful are the figures on value share and return on investment. The disadvantage of this approach is that it requires a valuation, even a valuation of both parties, which in many cases is not possible due to a lack of time or skills. Nevertheless, the analysis of the 22 deals showed clearly that the best deals were closed by the companies that had in-depth valuation skills at hand. We strongly recommend to seriously prepare every potential license deal; it is a key value driver, especially for young companies. How to Read a Deal

Press releases announcing license agreements in life science in general disclose little information on how much money flows from the inlicensing to the out-licensing company. Usually, the amount of the upfront payment and the total sum of potential milestones – if all are triggered – are disclosed. In some cases the companies also mention whether the royalty payments are one or two digits high. As an outsider it is not trivial to get an understanding of how these deals are financially structured. Nevertheless, having an idea about the deal terms is essential when valuing a company that depends on its license revenues or when benchmarking an own deal. The milestone weights from Fig. 3.56 allow extrapolating the size of future milestones knowing the upfront payment. If in a preclinical deal the upfront payment has been stated with US$ 4 mn, then this would mean that the NDA milestone would amount to four times the upfront payment, i.e. US$ 16 mn. In theory, all milestones should sum up to the total amount of potential milestones as announced in the press release. For a preclinical deal this would mean that the total amount of potential milestones should be 0.7+1.3+2.3+4.0+18.4=26.7 times the upfront payment. Often these numbers do not match exactly. First, the displayed milestone weights represent an industry average. Second, companies tend to artificially increase

198

Valuation in Life Sciences

the launch milestones, and to some lesser degree also the approval milestones. They add significant milestones for the approval of a second or third indication, even if such an extension is unlikely. Of course, this increases the milestone figure excessively and makes it somewhat less representative. In some deals backup molecules are included which double the potential amount of milestones to be paid. Another method to reach an impressive headline is to include sales related milestones. The in-licensing company must pay a milestone payment if it exceeded for the first time e.g. US$ 200 mn sales on a cumulative (or sometimes annual) basis. These payments are linked to well-defined goals, i.e. sales figures. But in fact the figures are royalties disguised as milestones, which help increasing the total milestone figures. We will now see how we can apply the knowledge of these milestone weights to a press release of a deal. “Jerusalem, Israel, and Lund, Sweden, June 14, 2004: TEVA AND ACTIVE BIOTECH ANNOUNCE AGREEMENT FOR LAQUINIMOD POTENTIAL ORAL THERAPY FOR MULTIPLE SCLEROSIS. (…) Teva has agreed to make an upfront payment of US$ 5 million to Active Biotech and to conduct and fund the further clinical development of laquinimod. The contract between the two companies also calls for Teva to make payments to Active Biotech upon the achievement of various milestones, which include sales targets. If such milestones were all to be met, payments to Active Biotech would aggregate to US$ 92 million. Active Biotech will also receive tiered double-digit royalties on sales of the product, once marketed. (…)” (source: www.activebiotech.com) In addition to the US$ 5 mn Active Biotech receives another US$ 87 mn if all milestones will be reached. Laquinimod was in phase II at the time of licensing. If we assume the weights displayed in figure 3.56 and assume the phase II milestone to be an upfront payment, then the subsequent milestones are weighted as displayed in table 3.52. The data sets for preclinical and discovery deals are slightly different, we can therefore derive two sets of milestone weights; but they remain pretty similar:

Advanced Topics in Project and License Contract Valuation

199

Table 3.52. Milestone weights for phase 2 Dataset

Lead Opt. Preclinical

Upfront Phase 3 Filing Approval Total

1.0 1.5 2.0 9.3 13.8

1.0 1.7 3.0 13.8 19.5

In the press release the ratio upfront/total milestones is 5/92=1/17.4. Applying more or less the weighting arising from the preclinical dataset we therefore obtain the following deal structure for the TEVA-Active Biotech deal: Table 3.53. Forecasted milestones Milestones

Calculation US$ mn

Upfront Phase 3 Filing Approval Total

5*1.0 5*1.7 5*3.0 remainder 92.0

5.0 8.5 15.0 63.5 92.0

In 2007 Teva initiated the phase III: “… net sales for the corresponding period in the preceding year included an initial milestone payment of SEK 51.2 M from Teva Pharmaceutical Industries (Teva) relating to laquinimod …” (source: Active Biotech, year-end report, January-December 2007) Active Biotech received another milestone of about US$ 10 mn upon reaching phase III. This is pretty much in line with the predictions from from the above tables. We can slightly adjust the milestones predictions, because we now also know the phase III milestone. We can also assume that the launch milestone will be about US$ 30 mn (distributed over different geographical areas) and the rest will be paid upon reaching sales targets. To our experience launch milestones are only excessively larger than the previous milestones if several indications or commercial milestones are involved. We can expect that the milestones figure of US$ 92 mn is not arti-

200

Valuation in Life Sciences

ficially blown up because no company mentioned any other indication than multiple sclerosis. Excluding other indications we assume that about half of the remaining milestones, i.e. US$ 32 mn, will be paid as launch milestones. The new, revised deal forecast is presented in table 3.52. Table 3.54. Forecasted milestones (2) Milestones

US$ mn

Upfront Phase 3 Filing

5 10 15

Approval

30

Sales Milestones

32

Total

92.0

We still do not know the royalty structure. The royalties are said to be double-digit. In order to divine the royalty rate (disregarding the tiered structure for a moment) we refer to the companies’ press releases that stated the multiple sclerosis market to be 4.2 bn in 2004 with TEVA being a major player commercialising Copaxone® with 2003 annual sales of US$ 720 mn (in 2007 TEVA reported Copaxone® sales of US$ 1,713 mn, excluding sanofi aventis’ Copaxone® sales). We can therefore assume that Laquinimod will reach blockbuster status, above all because of its oral administration. Assuming US$ 1 billion sales and 15% royalties we get a 67%-33% value share between TEVA and Active Biotech (calculated at 12% discount rate). If the sales potential is US$ 2 billion then the value share becomes 77%-23%. A phase 2 deal is said to have a value share between 25% and 35% percent, depending on the industry standard. A tiered royalty structure of 15% up to maybe 20% seems therefore realistic. In review US$ 5 mn upfront seem little for a potential blockbuster drug. One reason is certainly the low success rates for CNS drugs in phase II, which reduce the value of such compounds significantly. A second reason might have been Active Biotech’s will to retain as much upside as possible, putting more weight to sales milestones and royalties. This method of estimating future deal terms is essential when valuing companies whose revenues stem essentially from license agreements. For an analyst it is virtually impossible to value such a company without having an idea of the deal terms. The outlined method should serve as a guideline to divine such deal terms.

Advanced Topics in Project and License Contract Valuation

201

The Virtual Company Model

In the following we will explain in detail a new way to define deal terms in early stage licensing. Biotech, and especially academia face the challenge of how to structure deals where not only one company but sequentially two companies might license the IP rights. As we have already outlined previously in the chapter on licensing, the standard way to determine license terms is to follow the value share principle. This approach often falls short in justifying why the value of the rights to an asset should be shared and not entirely remunerated. The approach also does not help to structure license contracts where the inlicensing company might out-license the IP to a third company in the future and has to cede sublicensing payments to the originator, i.e. the outlicensing organisation, usually a university, a research institution, an inventor, or a biotech. We have built a simple, yet comprehensive model for this that can be used to define not only the license contract with the in-licensing party, but also to define the sublicensing terms, also called the participation rate, in case of sub-licensing. The Model: Building a Virtual CompanyOne should apply the virtual company model in all those cases where we beforehand think the IP could be sublicensed and where these sublicense payments are of vital importance to us. Typically this will matter for technology transfer offices where substantial payments often only occur once a compound is sublicensed to big pharma. We assume that the compound to be out-licensed is wholly owned by a virtual company as the sole asset. We will call the company VIRTUALCOMP. Before the license contract becomes effective, VIRTUALCOMP is wholly owned by the licensor, in the continuation called Academia. The main consequence of a license agreement is that the licensee adds a project to his pipeline and the licensor receives money in exchange for the project. A license is therefore very similar to a sale, with the difference that the price for the asset is not paid in full upfront but only in staged payments. Nevertheless, the licensee should still pay the full price if he wants to acquire the full rights to the asset – if the licensee would simply buy the asset instead of licensing it, he would also have to pay the full price and not only a fraction of it. The licensee, in this case Biotech, will therefore subsequently buy shares of VIRTUALCOMP to buy rights in the asset, starting with an upfront payment, then milestone payments,

202

Valuation in Life Sciences

and finally royalty payments, until in the end he owns all of VIRTUALCOMP’s shares. The license contact becomes effective with the upfront payment, which the licensor, in our case Biotech, pays. With this upfront payment Biotech buys some shares of VIRTUALCOMP from Academia, and consequently in the asset it is interested in. Biotech now has already an initial stake in VIRTUALCOMP. We assume that VIRTUALCOMP conducts the development of the compound and that Biotechshould capitalise VIRTUALCOMP for this purpose. The funding of the first development phase would be equal to a series A investment round with Biotech as the sole investor. In exchange for the investment, Biotech receives shares, which correspond to a capital increase of VIRTUALCOMP, i.e. to newly issued shares. As in reality, Biotech, which would in the real world conduct and finance the development on its own, de facto finances the development. If the compound successfully passes this phase, then another milestone payment is due. Again, this milestone payment corresponds to an acquisition of VIRTUALCOMP shares. But now the value of the shares has changed. The shares are more valuable as the success rate of this phase is not included in the valuation anymore and the compound is closer to the market, i.e. the cash flows are discounted less. To fund the next phase, VIRTUALCOMP conducts another financing round, a Series B round. Again, Biotech acts as the sole investor. These modifications of shareholdership with acquisition of shares due to milestone payments and dilution due to capitalisation for phase costs continue until the compound reaches the market. Once the compound is on the market and the launch milestone is paid, we have to set the royalty rate in a way that Biotech buys the remaining shares of VIRTUALCOMP until the end of the license contract, usually until the expiry of the patent. This guarantees that Biotech buys 100% of VIRTUALCOMP, i.e. pays the full price for the compound to Academia. Another way to interpret the royalty rate is to see it as dividend payments. Biotech and Academia both hold shares in VIRTUALCOMP. We can assume that all yearly profits will be immediately paid out to the shareholders in form of dividends. Academia can claim the portion of these dividends according to its share in VIRTUALCOMP. In the following we will illustrate how to apply this method to the problem of sublicensing.

Advanced Topics in Project and License Contract Valuation

203

Case Study Sublicensing For our case study we assume the following situation. VIRTUALCOMP owns all the rights to its only asset, a compound in preclinical development. The parameters of the compound and VIRTUALCOMP are summarised in the following tables. The license is only valid until patent expiry. Therefore, all payments to buy the rights to the compound must take place before the expiry. Table 3.55. Assumptions for R&D phases

Preclinical Phase 1 Phase 2 Phase 3 Review

Duration

Success Rate

Costs (US$ mn)

1 year 1 year 2 years 3 years 1 year

65% 75% 35% 72% 88%

2 2.5 5 30 2

Table 3.56. Assumptions for commercialization Parameter

Value

Launch costs Operating margin Peak sales Market growth Discount rate Patent protection after launch

US$ 50 mn 65% US$ 350 mn 0% 20% 12 years

100% 90% 80% 70% 60% 50% 40% 30% 20% 10%

patent protected

unprotected

Fig. 3.58. Sales curve

21

19

17

15

13

11

9

7

5

3

1

0%

204

Valuation in Life Sciences

VIRTUALCOMP today has a value of US$ 4.8 mn, when valued using the assumptions outlined above. Now the licensee pays an upfront payment of US$ 1.5 mn and in exchange gets 32% of VIRTUALCOMP’s shares. The funding of the preclinical development corresponds to the series A funding round at the pre-money valuation of US$ 4.8 mn. After the financing of the US$ 2 mn for the preclinical phase, the licensee holds 52% of the shares. The funding of the preclinical phase has diluted the licensor. The structure of the license contract is displayed below in the table. Using these figures we can calculate the dynamics of the relative ownership as depicted in the figures below: Table 3.57. License terms Terms

Payment ($ mn)

Up-front Entry into Phase I Entry into Phase II Entry into Phase III NDA Launch Royalties

1.5 1.5 3 5 5 13 3% Value for Biotech

La un ch

ND A

III Ph as e

II Ph as e

Ph as e

I

700 600 500 400 300 200 100 0 Pr ec lin ica l

$ mn

Value for Academia

Fig. 3.59. Participation values in VIRTUALCOMP. The first value jump corresponds to the resolution of uncertainty of the development phase, the second to the financing round

Advanced Topics in Project and License Contract Valuation Value for Academia

205

Value for Biotech

100% 80% $ mn

60% 40% 20%

La un ch

ND A

III Ph as e

II Ph as e

I Ph as e

Pr ec lin ica l

0%

Fig. 3.60. Relative ownership of the Academia and Biotech in VIRTUALCOMP. The first change of ownership in a phase refers to milestone payments (purchase of shares), the second to the dilution due to the licensee’s capital contribution.

Table 3.58. Value development of VIRTUALCOMP’s shares for Academia and Biotech Phase

Transaction Milestone/ Value to Value to investment ($ mn) Academia ($ mn) Biotech ($ mn)

Preclinical Up-front

1.5

3.3

1.5

2

3.3

3.5

Milestone

1.5

4.5

8.0

Costs

2.5

4.5

10.5

Milestone

3

4.8

21.2

Costs

5

4.8

26.2

Milestone

5

12.4

99.3

Costs

30

12.4

129.3

Milestone

5

29.1

359.6

Costs

2

29.1

361.6

Milestone

12

28.1

523.6

Costs

50

28.1

573.6

Costs Phase I Phase II Phase III NDA Launch

206

Valuation in Life Sciences

Once the drug reaches market, the Biotech owns already 95% of the VIRTUALCOMP. Academia has realized already most of the value of its stake in VIRTUALCOMP. Biotechhas to buy for the remaining 5% by paying royalties. If we assume that the operating margin, i.e. the margin after marketing, production and sales costs, we can calculate the royalty rate: 65%*5%=3% Negotiation The higher the value of the compound at negotiation of the contract, the more expensive are the shares of VIRTUALCOMP. The value of VIRTUALCOMP has to exceed the value of the compound if it would be self developed by the out-licensing company, otherwise it would not make sense to give the compound away. The licensor, in our case Academia, usually gains value by out-licensing the project by getting access to the licensee’s superior capabilities in preclinical and clinical development, regulatory issues, and marketing and commercialization. In the case of an academic institution being the licensor, a license is often the only way to move the project forward. This, of course weakens the bargaining position considerably. The cost of capital, i.e. the discount rate we apply, is usually in the range of 12-25%. The IRR of the licensee, as we have found in a recent study, usually falls into this range for preclinical deals and discovery. The cost of capital of VIRTUALCOMP corresponds to the IRR of the licensee. Sublicensing As we have already mentioned, it is common that intellectual property, which is licensed at early stages will be out-licensed again later to a third party, typically to big pharma, once proof of concept in man is reached. In case of sublicensing the remaining license terms from the initial contract can be replaced, but do not have to, by sublicensing terms. The sublicensing terms are often a share of the license terms, which the first licensee, i.e. Biotech, receives from the second licensee, i.e. big pharma. This share is called the participation rate, as it is the participation of the first licensor, Academia, in the terms of the new sublicense contract. The replacement of the original deal terms of the first license contract by a participation in the new sub-licensing terms has two advantages over the initial license terms. On the one hand, if the potential of the project turns out to be much higher than anticipated, then Biotech can negotiate a

Advanced Topics in Project and License Contract Valuation

207

good sub-license contract based on the new, revised sales potential. Academia profits from this at the time of the first deal closing not yet anticipated potential via the participation in the sub-license contract. On the other hand, if the project is performing worse than expected and Biotech still would have to complay with the originally agreed milestone and royalty payments to Academia, Biotech might face a deadlock; the sub-license terms must be at least as high as the original license terms, otherwise Biotech could not enter a sub-license agreement. By softening this liability, Academia allows Biotech also sub-licensing the project to worse than originally expected terms. This is still to the advantage of Academia; finally a share of a small license payment is better than none at all. Table 3.59. Participation rates for VIRTUALCOMP Phase Phase 1 Phase 2 Phase 3 Review Commercialisation

Participation rate 36% 19% 11% 7% 5%

To illustrate this, we assume that Biotech, after successfully completing a phase 2 clinical trial, out-licenses the compound to big pharma. Academia would then receive 11% of the up-front payment instead of the initially agreed US$ 5 mn milestone for the commencement of phase 3. The 11% share in the new contract would then also replace all further terms. Discussion The virtual company model is an alternative and well-understandable alternative to the value share principle and an essential tool for the determination of participation rates. It is also very supportive to move negotiations away from emotions to parameters like discount rate and peak sales. The automatic generation of the participation rates is a major argument to use the method in early stage licensing. We therefore recommend implementing the approach in technology transfer licensing, where negotiations often turn around these participation rates.

208

Valuation in Life Sciences

Negotiating License Deals

Licensing is an integral part of the life sciences industry. Universities do not have the potential to take their inventions through late stage preclinical tests and early stage clinical trials. Biotech companies in-license many of these products and conduct the early stage clinical research. Finally, they most often too lack the financial strength and expertise to bring an invention to market and out-license the product to large multinational pharmaceutical companies. With their financials, they conduct late stage trials and ensure the best commercialisation of the product based on their strong marketing departments. Furthermore, the big pharmaceuticals are not only answering to the financial need of biotech and universities, they themselves are forced to fill their pipeline with new innovative projects. They are dependent on licensing to do so, as the in-house research does not meet that need. Most license contracts take place between two parties, the in- and the out-licensing company. Signing a license deal, the parties first need to evaluate if the license terms are fair and meet their preferences in terms of financials and risk. Both parties need to value the contract in order to position themselves in the negotiation process. The results of the negotiation must benefit both parties. The in-licensing company wants to learn what the maximum price is they can afford for the license, the out-licensing company what the minimal price for the disposition is. These upper and lower boundaries are the leeway of the negotiation. To get the upper boundary of the negotiation, i.e. the highest price big pharma is willing to pay, we have to calculate the project value as if pharma is conducting the project as a self originated project based on its in-house capabilities and capacities. We use the costs, duration, peak sales for pharma and discount at a pharma discount rate between 8% and 12%. We have now calculated the value of the project as if it was pharma’s own project. In the next step, we calculate the lower boundary of the license contract. This corresponds to the project value as if the biotech company conducts the project on its own. We therefore choose the input parameters according to the capabilities of biotech. The peak sales are lower, as biotech does not dispose of a marketing department comparable to big pharma. Furthermore, biotech often takes a risk reducing approach to the clinical trials. It first conducts one single indication to get the proof of concept, and then in case of success sequentially launches the trials for further indications. Therefore, we assume that the sales revenues grow slower than pharma’s. The project is discounted at the biotech’s discount rate, typically above 14%. We could also adjust the success rates of the clinical trials if we can quantify the difference reasonably.

Advanced Topics in Project and License Contract Valuation

209

Value ($)

Biotech

Leeway

Pharma

Fig. 3.61. Leeway in license negotiations

If the terms of the license contract render a value for biotech below the lower boundary, the company is better off conducting the project on its own. On the other hand, if the value is above the upper boundary, pharma should not sign the contract, as it would buy a project at a price exceeding its value. Of course, both parties need to value independently the project’s upper and lower boundary to know the acceptable range for the negotiation. The final contract value should be between the upper and the lower boundary. Nevertheless, pharma is sometimes willing to pay more than the upper boundary due to strategic considerations, and biotech sometimes wants to get rid of a project it could not finance, to avoid new fundraising rounds. If there are several bidders for the project, e.g. due to a new and innovative technology, biotech can try to get more than the lower boundary price. Biotech should also be allowed to profit from the capabilities of the pharmaceutical partner. Pharma can try to argue to pay less than the upper boundary as it brings all its expertise into the project and leverages the value and because it suffers from information asymmetry. Pharma never knows all facts about the project it licenses. If biotech wants to reduce the discount due to this information risk, it can agree to set more weight on royalty payments and therefore shares more risk with pharma. The final terms should ultimately reflect the parties’ need for cash and risk preferences. As we have seen in the section on licensing, there is an alternative way to negotiate license contracts based on the value share principle. The inand out-licensing companies agree to share the value of the project. The share of the out-licensing company increases the more advanced the development. The argument goes that the licensee takes large risks, due to high investments and attrition rates. The licensee wants to be rewarded for

210

Valuation in Life Sciences

this risk. Nevertheless, if we value the project based on risk adjusted NPV or real options, we account for the risk with the success rates and the discount rate, and for the investments with the cash flows. We would therefore reward the pharma company twice to take the risk. The value share method has its strength when used to compare different deals with each other. This can even be applied to deals of different values, as we look at a relative figure with the value share method. Negotiation Case (First Part) 100% 80% 60% 40% 20% 0% 1

2

3

4

5

6

7

8

Biotech

9

10

11

12

13

14

15

16

Pharma

Fig. 3.62. Sales curves for biotech and pharma

A biotech company has just successfully reached proof-of-concept in man, i.e. finished clinical phase IIa testing, and considers to out-license the compound or to conduct all development and to market it on its own. It wants to take this strategic decision based on a clear number. It calculates the value of the project if it does everything in-house. The inputs respect the characteristics of the company in terms of financial strength and marketing expertise. The company then also calculates the value of the project if the main candidate for the licensing conducts the project on its own. The drug development costs and the sales ramp is higher if pharma conducts the project. The discount rate for biotech is 17%, for pharma 11%. The calculation, based on the parameters from the table below gives us the upper and the lower boundary of the negotiation. Biotech should keep the project and continue developing it on its own if the value of the deal drops below $ 52 mn. Conversely, if the value of the license mounts above $ 203 mn, big pharma should walk away. At the upper boundary, the return on investment for pharma corresponds to its discount rate.

Advanced Topics in Project and License Contract Valuation

211

Table 3.60. Input parameters for biotech and pharma

Phase 3 Duration Cost Success Rate NDA Duration Cost Success Rate Market Launch costs Peak Sales Discount rate rNPV

Biotech

Pharma

3 years $50 mn 65%

3 years $60 mn 65%

1 year $3 mn 88%

1 year $3 mn 88%

$120 mn $250 mn 17%

$120 mn $250 mn 11%

53 mn

203 mn

We now consider a license contract where biotech receives the following payments from pharma: Table 3.61. Negotiation case contract Upfront Milestone clinical phase 3 completion Milestone NDA approval Royalty rate

$ 20 mn $ 40 mn $ 80 mn 17%

The license contract has a value of $ 105 for biotech, and -$ 150 mn for pharma. The project therefore increases in value for biotech by $ 53 mn, and for pharma by $ 53 mn as well, once they enter into the licensing agreement. The difference of value of the license contract arises out of the different discount rates for biotech and for pharma. The figure below pictures these numbers: (1) is the original value for biotech and the lower boundary, (2) is the value increase for biotech trough licensing, (3) the difference arising out of the discount rates, and (4) the value increase for pharma.

212

Valuation in Life Sciences

Value ($)

4 3 2

1

Biotech

Pharma

Fig. 3.63. Decomposition of value

As biotech and pharma do not use the same discount rate, we cannot properly distribute the value between them a portion of the value (3) is not attributable. This also explains why license contracts are not communicated in terms of NPV, but rather in terms of cash flows, e.g. $ 20 mn upfront, further $ 120 mn in milestones, and undisclosed royalties. If biotech prefers to have higher upfront payments, the difference in discount reduces. If biotech does not urgently need cash it wants to receive delayed payments in form of milestones and royalties, leveraging the value of the project in case of success. We can refine the valuation approach by incorporating the value of the biotech’s option to self-conduct the project. We can argue that once biotech brings its first compound through late stage clinical trials and builds up an own marketing department, all subsequent projects are more valuable due to the increased expertise and capabilities of biotech and moreover due to a decreased discount rate. Even if the real option to become a fully integrated biopharmaceutical company is not evident in the value of the lead project (the value of the license contract might be higher than if biotech conducts the project on its own), it can pay off for the following projects. These benefit from the new properties of the biotech company, i.e. expertise, already established distribution network and steeper sales ramp, and lower discount rate. The lower boundary for the negotiation then increases and in some cases, biotech might even be better advised to conduct the project on its own. We can also argue that biotech profits from the cash it receives from the license deal and from the reduced risk. This might as well be factored in a reduced discount rate, and a consequent value increase.

Advanced Topics in Project and License Contract Valuation

213

Value ($)

A

Biotech

Leeway

Pharma

Fig. 3.64. Increased lower limit for biotech due to strategic value (A)

We should also not forget to mention the aspect of control. Once biotech out-licenses the project it has lost the control over it. Pharma could delay development and thus reduce the value of the project. Also, we should keep in mind that milestones are non-dilutive cash. Case (Second Part) We now want to put a number on the biotech’s strategic option to conduct the development on its own and to become a fully integrated drug development company. We will contrast the value of the biotech pipeline as it is now, with a discount rate corresponding to the company’s maturity, with the value of the pipeline if the company becomes as well a drug selling company. We assume that if the drug is marketed, we can reduce the biotech’s discount rate from 17% to 15%. The value for the other projects in the pipeline, Project-B, Project-C, and Project-D, changes due to the increased expertise of the company. This increase only takes place if the biotech company is able to launch the first project, the probability being 65%*88%=57.2% The strategic value for the subsequent project then amounts to 57.2%*(9.1+12.6+7.8)=$ 16.9 mn. The lower boundary therefore rises from $ 53 mn to $ 70 mn. If the biotech would have been offered a license contract worth less than $ 70 mn it might have opted to develop the compound on its own. We can conclude that the above discussed negotiation paradigm gives a sound basis for fair negotiations, both, for the licensee and the licensor. To be successful in negotiations, you have to know your leeway and your options. The discussed approach requires some work, but provides security on fixing the deal terms.

214

Valuation in Life Sciences Table 3.62. Value and value increase Project

Project-B

Project-C

Project-D

Phase Value @ 17% Value @ 15% Value Increase

Phase 1 $ 7.9 mn $ 17.0 mn $ 9.1 mn

Phase 1 $ 16.8 mn $ 29.4 mn $ 12.6 mn

Preclinical $ 8.1 mn $ 15.9 mn $ 7.8 mn

A further application of the method, namely strategic decision making, should be considered when evaluation the option to license. The ultimate goal for any company is to increase its risk adjusted value. The method therefore is much more than a simple negotiation framework. Bridging Different Assumptions in Licensing Negotiations During license negotiations we often encounter the situation where the inand out-licensing company have different views about the potential of the project. The out-licensing company usually believes that the sales potential is enormous, the success rates above average, and the timelines very short. The in-licensing company of course does not want to concede, as it might have to pay more for the license contract. So the license is negotiated and the one with the better negotiation skills has to make less concessions. Differing assumptions however can also be used for each side’s advantage by structuring the license contract according to them without impairing the value in of either side. Assumptions often differ regarding: • Peak sales • Sales uptake curve • Success rates • Costs • Time of development • … If the parties cannot agree on a set of assumptions the value share principle is difficult to apply, as the results are not the same at the opposing sides of the negotiation table. One party will always feel disadvantaged.

Advanced Topics in Project and License Contract Valuation

215

Table 3.63. Term sheet for a preclinical license Term Upfront Milestone IND Milestone phase II Milestone phase III Milestone phase filing Milestone phase approval Royalty

US$ mn 2 1 2 5 8 25 6%

In the following we will show how we can use differing assumptions to improve the value of a license contract for both parties. Both case studies start with the following term sheet for a license in preclinical stage. A press release, would typically state that it is a deal with an upfront payment of US$ 2 mn, further potential payments of US$ 41 mn, and undisclosed royalty payments. Case Study: Different Sales Assumptions The out-licensing company thinks that the drug, once on market, could achieve peak sales of US$ 600 mn. The in-licensing party however considers this as too optimistic and calculates with US$ 450 mn. Furthermore the in-licensing party thinks that the uptake of sales won’t be as fast as predicted by the licensor.

Fig. 3.65. Licensor expects higher peak sales and a faster sales ramp (figure produced with ri:val)

216

Valuation in Life Sciences

The two parties could agree to change the original term sheet in a way that each party profits if its assumptions turn out to be accurate. The licensor therefore agrees to lower the royalty rate from 6% to 5% for the first annual US$ 300 mn. But the licensor wants a rate of 8% for sales up to US$ 500 mn and a rate of 10% for sales above this figure (which in the licensee’s eyes would never happen anyway). This way if the sales are really above US$ 450 mn the licensor will get more in return of giving away royalties for low payments. Furthermore the out-licensing company proposes the inclusion of sales milestones in the term sheet: After having sold US$ 150 mn the licensee has to pay US$ 20 mn to the licensor. If the sales ramp is fast this payment happens earlier, and the discounted value of this milestone is higher. This way, if the sales ramp is steeper, the value of the sales milestone is higher as it is discounted less. Table 3.64. Modified term sheet Term Value Licensor at 12% Value Licensee at 12%

Original term sheet 9.9 (8.7)

New term sheet 10.4 (8.4)

As a result of these changes the in-licensing company can reduce the license costs by US$ 0.3 mn, because it has to pay less royalties using its own set of assumptions The value, in this case negative, increases to (US$ 8.4 mn). The licensor on the other hand also increases the value of the contract, because he receives more royalties, if his set of assumptions is closer to the truth. In the negotiation the licensee will also have some trouble to refuse these changes, because he claimed earlier in the negotiations that he does not believe the drug to achieve sales higher than US$ 450 mn, and the licensor’s value increase stems from the royalty payments that exceed these very US$ 450 mn. Case Study: Different Success Rate Assumptions The out-licensing company thinks that the drug has high success rates and that it will reach market with a probability of 27%. The in-licensing party remains conservative and applies an overall success rate of 11%. This difference in success rates often stems form using different sources of published data or of using differently defined success rates, i.e. transition probabilities vs. success rates. As a result of the higher success rates the size of the pie that is to be shared between the two companies, i.e. the value of the project,

Advanced Topics in Project and License Contract Valuation

217

Fig. 3.66. Licensor assumes higher success rates (figure produced with ri:val)

is now larger and the out-licensing company claims higher terms. This however is not justified in the eyes of the in-licensing company. The licensor also assumes some attrition, but to a lesser degree than the licensee. A million dollars in later milestone payments or royalties is more worth in the licensor’s eyes than in the licensee’s, provided they use the same discount rate. But one million dollar upfront payment is worth the same to both, because no assumptions enter in the valuation of that million. The licensee could therefore propose to lower the earlier payments in exchange of increasing the later, thinking the chance of really having to pay them is not as high. Table 3.65. Modified term sheet Term Upfront Milestone IND Milestone phase II Milestone phase III Milestone phase filing Milestone phase approval Royalty

Original

Modified

2 1 2 5 8 25 6%

0.3 0.5 1 3 12 35 8%

This back-loading of the deal again has the consequence of increasing the deal value for both parties, as shown in the table.

218

Valuation in Life Sciences Table 3.66. Modified term sheet Term Value Licensor at 12% Value Licensee at 12%

Original term sheet New term sheet 12.4 12.9 (8.7) (8.4)

Again, the figures for the licensee denote the present value of the costs, i.e. the milestones and royalties. We see that the cost for the licensee is always lower than for the licensor because of more conservative assumptions. This difference is even more accentuated with a term sheet that intentionally focuses on these different assumptions. Another remedy to bridge differing assumptions could be the inclusion of partially fixed royalty payments, i.e. a fixed amount plus relative royalties, or an overall minimum or maximum amount of payments. Some agreements also contain option-like clauses like opt-in or callback options that have similar motivations. Valuing License Contracts with Multiple Indications

Background It is not uncommon that one compound has several possible indications, as is the case with most cancer drugs. The developing company usually has to conduct separate trials for every indication it wants to market. Before phase 2, i.e. in the lab or in phase 1 trials, the indication is not yet an issue and the results of the compound can usually be used for any indication. However, in phase 2 and phase 3 trials the company must prove that the drug is an efficacious treatment for the proposed indication. A drug development project that is still in an earlier stage than phase 2 has therefore the structure shown in Fig. 3.67. The split of the project into several programs sometimes occurs as early as in preclinical stage, sometimes as late as phase 3. Once the trials are conducted for specific indications, each indication of the compound represents a project on its own, we call these indication specific projects subprojects. The joint preceding part up to phase 1 is called the core project. This split into several subprojects has important implications for the valuation. First, the project with just one indication might not be profitable enough to justify its continuation. Expanding the drug into other indications

Advanced Topics in Project and License Contract Valuation Research and Preclinical Phase

Phase 1

Drug Candidiate

Phase 2

Phase 3

219

Approval and Commercialisation

Indication1

Indication 2

Indication 3

Core Project

Subprojects

Fig. 3.67. Subdivision of a project into core and subprojects

can increase its value such that it is worthwhile to develop it further. Typically the values of the subprojects add up. However, if the project is licensed and the license contract contains a tiered royalty structure, the royalty payments of each single indication might fall in a low royalty tier, but if several indications reach commercialization a higher royalty rate is reached. Hence, the simple additivity of subproject values does not work anymore. Some license contracts also include milestones that are triggered after having reached a certain sales figure. Of course, this also varies if only one or several indications are approved. Second, we need to know the success rates not only of the project, but also of the subprojects. Published success rates for phase 2 and 3 refer to compounds and not to indications. A compound counts as having successfully passed the phase if at least one indication has successfully passed phase, no matter whether others fail. This implies that the success rate of indications is lower than the success rate of compounds. Common sense suggests that if one indication passes, the second is more likely to pass, than if the first has failed. This is a concept called correlation. Theory DCF The main concepts of valuation remain the same also in the case of multiindication projects. All cash flows are adjusted by their probability and discounted over the time difference between their happening and the valuation date. In that perspective it makes sense that any additional indication adds

220

Valuation in Life Sciences

value to the project, as long as the indication itself has a positive NPV. The value of the project is simply the sum of the values of the core project and all subprojects. The core project, being composed of exclusively investments, has necessarily a negative NPV (sum of its risk adjusted cash flows). No revenues can be generated in these phases, at least not if it is selfconducted. Nevertheless, these expenses are necessary to enable the subprojects. The core project can be seen like an investment that gives access to the subprojects. We have to adjust the subprojects by their probability; the subprojects are only reached with the probability that all phases up to phase 1 are successfully passed, i.e. that the core project succeeds. From then onwards the subprojects use their own success rates. A discussion of indication specific success rates follows at the end of this section. The mentioned method of dividing the project into parts (the subprojects), valuing these parts, and then summing up the values again works perfectly well as long as the cash flows can be linked to just one of these parts, i.e. as long as the cash flows are indication specific. In license contracts this is not always the case. Royalties often cover the sales of all indications. Sales milestones also refer to sales stemming from all indications. We illustrate this issue with the following example. Example: License Contract for Two Indications BioComp has licensed its cancer project BC-1 to a pharmaceutical company. The development plan foresees the approval of two indications, BCLL and ALL. B-CLL is expected to be approved with a probability of 50%, ALL with a probability of 40%. Upon approval BioComp receives a milestone payment of US$ 10 mn in the case of B-CLL and US$ 5 mn in the case of ALL; both approvals are expected in three years. No earlier milestones are included in the license contract. Royalties are tiered, up to US$ 120 mn a rate of 10% applies, for sales exceeding this threshold a rate of 15% applies. For the sake of simplicity we assume that the pharmaceutical company will immediately reach peak sales, and that annual sales for each indication amount to US$ 100 mn. It is furthermore agreed that after having sold US$ 400 mn BioComp receives a sales milestone of US$ 20 mn. Royalties will be paid as long as patent protection is in place, so for the first 10 years of commercialisation. Mr. Internship has the task to value the license contract. He is told to use a discount rate of 14%. He approaches the problem by listing all cash flows of the license contract by indicating their size, their probability, and their time.

Advanced Topics in Project and License Contract Valuation

221

Table 3.67. Cash flows of license contract Cash Flow

Size

Probability

Time

Launch B-CLL Launch ALL Sales Milestone Royalties

US$ 10 mn US$ 5 mn US$ 20 mn 10%-15%

50% 40% 50%? 40%? 50%? 40%?

Y3 Y3 Y5 or Y7 Y3-Y12

Establishing this overview, Mr. Internship became aware that some cash flows, the sales milestone and the royalties, depend not only on one indication, but on both. If he values each indication separately, where should he put the sales milestone? He cannot account for it in both indications. And if he values each indication on its own, then the second royalty tier would never be reached, although this seems likely. It is therefore not sufficient to know the probability of approval for each indication; he also needs to know the probability of joint approval. He therefore runs three scenarios to find out more about the effect of joint approval. In the first scenario ALL is only approved if B-CLL is approved. So with 40% both indications are approved, and in 10% only B-CLL. In this case the two indications are maximally correlated. The second scenario assumes the contrary; either B-CLL is approved or ALL, but never both. In the remaining 10% none is approved. The third scenario assumes that the two indications are more or less independent, in some cases both are approved, in some B-CLL, and in some ALL. only B-CLL

both

only ALL

none

100% 90% 80% 70% 60% 50% 40% 30% 20% 10% 0% Maximum Correlation

Minimum Correlation

Independent

Fig. 3.68. Scenarios of joint approval

222

Valuation in Life Sciences

Mr. Internship now calculates the values for each possible situation, i.e. if B-CLL, ALL, or both indications are approved. If only one indication is approved, the sales milestone happens only after 4 years of commercialization, so in year 7. If both indications are approved it happens already after two years, i.e. in year 5. As long as only one indication is approved only the lower royalty rate applies, but if both indications reach commercialisation, then US$ 120 mn are redeemed at 10% and the remaining US$ 80 mn at 15%; BioComp then receives US$ 24 mn in royalty payments each year. Mr. Internship now calculates the value of each cash flow for each possible approval scenario. Table 3.68. Values of license contract for joint approval scenarios Cash Flow Launch B-CLL Launch ALL

Only B-CLL

Only ALL

Both

US$ 6.7 mn

-

US$ 6.7 mn

-

US$ 3.4 mn

US$ 3.4 mn

US$ 8.0 mn

US$ 8.0 mn

US$ 10.4 mn

Royalties

US$ 40.1 mn

US$ 40.1 mn

US$ 96.3 mn

Total

US$ 54.8 mn

US$ 51.5 mn US$ 116.8 mn

Sales Milestone

According to the joint approval probabilities, each of these values receives a different weight. For the maximum correlation scenario we get 10%*US$ 54.8 mn plus 40%US$ 116.8 mn and 50% US$ 0 mn, totaling US$ 52.2 mn in value. For the minimum correlation scenario he gets US$ 48.0 mn, and for the independency scenario US$ 50.1 mn. The difference between the two extremes is 9%. Each value corresponds to the initial assumptions of 50% and 40% success rates. But the joint approval probability moves the value up by 9%. The correlation obviously can increase the value of the project as this paragraph has shown. Discussion As shown in the example of Mr. Internship, multi-indication contracts cannot always be fully subdivided into subprojects. As soon as tiered royalties and sales milestones are included in the license contract, the calculation should be based on the assumption of joint probabilities of approval. In this case we must calculate the value of the royalties and sales milestones for each possible approval scenario and then weigh them according

Advanced Topics in Project and License Contract Valuation

223

to the joint approval probabilities, as Mr. Internship did. For self-conducted projects the joint approval probabilities only matter if we include possible synergies in launch, production, and marketing costs. Theory Real Options Taking a real option perspective multi-indication projects require a new concept of tree. We now do not only have one underlying of the option as in the previous models. We now have to assume peak sales for every indication. These fluctuate as usual, but they do not necessarily fluctuate in the same way. If the sales potential for one indication deteriorates, this does not have to mean a decrease of the second indication’s sales potential. We can imagine a competitor launching a drug in the first indication, while the second remains unaffected. Or, the compound performs better in one indication than in the other. However, usually the drug acts in similar ways in the various indications and some changes in sales potential occur simultaneously. This joint behaviour resumes in a positive correlation between the two peak sales estimates (or sales potentials). A common binomial tree cannot deal with two non-identical underlyings, i.e. peak sales estimates. We therefore have to superimpose two trees as displayed in the figure. We use a two-dimensional tree (when visualizing, this becomes three-dimensional, since we have not counted the time axis as a dimension). While moving forward in the tree (in time) the sales potential of the first indication can go up (increase) or down (decrease), and the sales potential of the second indication can go left (increase) or right (decrease). We now have to correctly define the tree parameters, i.e. the step size up, down, right, and left, and the corresponding probabilities. Assume that the two indications are perfectly correlated. Then they should either go jointly up and left, or down and right. On the other hand, if they are independent of each other they should go up and down, left or right, independently on the other indication. This corresponds to the following scenarios: Table 3.69. Joint probabilities Perfectly correlated left right

Independent left

right

up

pul

0

up

pu*pl pu*pr

down

0

pdr

down

pd*pl pd*pr

224

Valuation in Life Sciences

(uAA,dBB) (u A,u B) A B

1st indication (A)

(A0,B0)

(dAA,dBB)

(dAA,uBB)

2nd indication (B)

Fig. 3.69. 2-dimensional binomial tree, one time step

In order to model a joint movement of two correlated underlyings we can use the following set of tree parameters. Table 3.70. Joint probabilities for correlated indications u (uA) d (dA) pu

e µ A ∆t +σ A

∆t

l (uB) e µ B ∆t +σ B

∆t

e µ A ∆t −σ A

∆t

r (dB) e µ B ∆t −σ B

∆t

1− d u−d

pl

1− r l−r

pd

1 − pu

pr

1 − p1

pul

ρ

pur

(1 − ρ ) pu pr

pdl

(1 − ρ ) pd pl

pdr

ρ

pu + pl + (1 − ρ ) p u p l 2

pd + pr + (1 − ρ ) p d p r 2

This set of parameters is a good approximation as long as the two volatilities do not differ too much (0.5*σA<σB<2*σA). We can now model the tree until the end of phase 3. This then corresponds to a construct as displayed in the figure. When placing ourselves in the end nodes, we know in each node the peak sales estimate for both indications. We can therefore value the project assuming that both indications passed all phases. However, it is also possible that only one indication reaches commercialisation. We must therefore

Advanced Topics in Project and License Contract Valuation

225

1st indication (A)

2nd indication (B)

Fig. 3.70. 2-dimensional binomial tree over several time steps

…

n

2

1 0 0

1

2

…

n

Fig. 3.71. 2-dimensional binomial tree with zero-layers

add a column (1st indication) and a row (2nd indication) where only one indication is still alive. In the figure the additional columns are displayed by the wall next to the tree, the additional rows by the floor under the tree. We numerate for each time t step the nodes for the two dimensions S it, j . The nodes S it, 0 correspond to the states where the second indication already failed, in the nodes S 0t , j the first indication failed, and in S 0t , 0 both

226

Valuation in Life Sciences

failed. We call these states building the wall and floor to the multidimensional tree “zero-layers”. Valuing a multi-indication project with real options follows the same lines like standard real options valuation of a project, with some nuances: 1. Span the tree, including the “zero-layers” for each dimension until the last decision point. 2. Attribute a value to all end nodes according the respective peak sales estimates of the indications. 3. Work-back the tree to the root node S10,1 . Spanning the tree. While the concept of spanning the tree has been outlined above, the technical implementation is somewhat tricky. A spreadsheet offers just two dimensions, columns and rows. We can either design the spreadsheet in a way that we put the various time steps of the tree under each other, or we put them on different worksheets. Both methods are too cumbersome. An elegant way is to program the tree in VBA, the programming language that comes along with Excel. We give here a short piece of code explaining how the tree can be programmed. It first calculates all tree parameters, i.e. the steps up, down, left and right. Then it calculates the joint probabilities of the joint steps within the tree. Finally the tree is spanned, i.e. each node is attributed a peak sales estimate for each indication. The zero layers have to be spanned separately. ReDim tree(0 To n, 0 To n + 1, 0 To n + 1, 1 To 2) As Double ‘arguments: time, indication A, indication B, peak sales for A and B ‘n being the number of time increments Dim i As Integer ‘counter Dim j As Integer ‘counter Dim t As Integer ‘counter of time Dim up As Double ‘step up Dim down As Double ‘step down Dim left As Double ‘step left Dim right As Double ‘step right Dim Prob_u As Double ‘probability up Dim Prob_d As Double ‘probability down Dim Prob_l As Double ‘probability left Dim Prob_r As Double ‘probability right Dim Prob_ul As Double ‘probability up and left

Advanced Topics in Project and License Contract Valuation Dim Prob_ur As Double ‘probability up and right Dim Prob_dl As Double ‘probability down and left Dim Prob_dr As Double ‘probability down and right ‘calculation of tree parameters up = Exp(mu_A * dt + sigma_A * dt) down = Exp(mu_A * dt – sigma_A * dt) left = Exp(mu_B * dt + sigma_B * dt) right = Exp(mu_B * dt – sigma_B * dt) ‘mu_A and mu_B are the continuously compounded growth rates ‘sigma_A and sigma_B are the volatilities ‘calculation of single probabilities as intermediary step

Prob_u = (1 – down) / (up – down) Prob_d = 1 – Prob_u Prob_l = (1 – left) / (left – right) Prob_r = 1 – Prob_l ‘calculation of joint probabilities, rho being the correlation Prob_ul = rho * (Prob_u + Prob_l) / 2 + (1 – rho) * pu * pl Prob_ur = (1 – rho) * pu * pr Prob_dl = (1 – rho) * pd * pl Prob_dr = rho * (Prob_d + Prob_r) / 2 + (1 – rho) * pd * pr ‘initialisation of root node tree(0,1,1,.) tree(0, 1, 1, 1) = peak_sales_A tree(0, 1, 1, 2) = peak_sales_B For t = 1 To n ‘node tree(t,1,1,.) tree(t, 1, 1, 1) = tree(t – 1, 1, 1, 1) * down tree(t, 1, 1, 2) = tree(t – 1, 1, 1, 2) * right For i = 2 To t + 1 ‘nodes tree(t,i,1,.) and tree(t,1,i,.) tree(t, i, 1, 1) = tree(t – 1, i – 1, 1, 1) * up tree(t, i, 1, 2) = tree(t, 1, 1, 2) tree(t, 1, i, 2) = tree(t – 1, 1, i – 1, 2) * left tree(t, 1, i, 1) = tree(t, 1, 1, 1) For j = 2 To t + 1 ‘nodes tree(t,i,j,.) tree(t, i, j, 1) = tree(t – 1, i – 1, j – 1, 1) * up tree(t, i, j, 2) = tree(t – 1, i – 1, j – 1, 2) * left Next j Next i

227

228

Valuation in Life Sciences

For i = 1 To t + 1 ‘zero layer tree(t, i, 0, 1) = tree(t, i, 1, 1) tree(t, i, 0, 2) = 0 tree(t, 0, i, 2) = tree(t, 1, i, 2) tree(t, 0, i, 1) = 0 Next i Next t

Attributing values to the end nodes. At the end nodes we can perform a standard DCF calculation of the project according to the peak sales for each indication. If it is a licensed project, the royalties are taken from the joint sales of all indications. Working back the tree. This step is slightly more complicated than in a one-dimensional tree. First, we have to take the expectation not only of two, but of four subsequent scenarios. Second, the success rates and decision points relate sometimes only to one indication while the other is not affected by a success rate or a decision. Third, the decisions are not about continuing the project, but about continuing indications. The application of the success rates and the decision points have to be treated in more detail. Assume you are in the last node of a phase of the first indication. In the next time step, the results reveal whether the indication passes or fails. It passes with the assumed success rate for this phase and indication. In this case we will move to one of the four subsequent nodes of the next time point in the tree. If the trial results are negative, the project is not automatically abandoned, because other indications might still be in testing. The project therefore only moves to a state without the first indication, i.e. to a state of the zero layer. The following formula displays how to calculate the project value when a success rate P1 for the first indication applies. (3.34)

Vi ,t j =

⎛ P1 ( pulVi +t +11, j +1 + purVi +t +11, j + pdlVi ,t +j +11 + pdrVi ,t +j 1 )⎞ ⎜

⎟ 1 ⎜ 2 ⎟ ∆t ⎟ ⎜ + (1 − P1 )( plV0t,+j1+1 + prV0t,+j1 ) ( )

1+ r

⎟ ⎜ 1 3 ⎠ ⎝ 1 2 + CFt + CFt

4

Advanced Topics in Project and License Contract Valuation

229

1. Discount for the time step 2. Expectation if trial successful 3. Expectation if trial not successful 4. Cash flows occurring between t and ∆t The same procedure has already been applied in previous tree calculations with only one indication, but there we could simplify the formula, because the zero layer, i.e. the states of the project where the indication failed, has only zero values. Consequently, step 3 drops out. If trials of both indications end in the same time step, then we even have to further expand the above formula, taking account of cases where both, only the first, or only the second succeeds, or both fail. We will see in the paragraph about success rates how to assess these joint probabilities. The decisions in a multi-indication project are not solely about continuing the project, but even more importantly about selecting the right indications and then continuing the indications. After phase 1 a company has the choice to start trials for several indications. While it could prepare trials for any thinkable indication, it reasonably only selects the indications that add value to the project. This depends on the characteristics of the compound that must be translated into judicious peak sales estimates. After a successful phase 2 trial, the company then does not need to decide on the continuation of the project, but only on the continuation of the indication. In our multidimensional tree such a decision becomes to a choice of the maximum value of the project with the indication, Vi ,t j , and without the indication, V0t, j . The above displayed method can be extended to more than two dimensions. For the adjusting of the joint probabilities of the tree the reader is referred to the literature from quantitative finance (Kaushik 1991,1995). The real options method is especially suited for the calculation of license contracts. Often license contracts for cancer compounds include milestones for different indications, but royalties, most often stacked, on a project basis. Milestones are specified for phases and can vary whether it is the first, second or third indication. Some contracts also make a difference between minor (e.g. hematological cancer) and major (colorectal cancer) indications. The royalties on the other hand are taken from the sales of all indications of the drug. When the royalties are fix, then one can calculate the subprojects separately, using the right probabilities. However, if royalties are stacked, i.e. the royalty rate depends on the sales revenues, then we must know how much is sold in order to determine the royalty rate. Assume two indications, both with sales of $ 200 mn. If the license contract

230

Valuation in Life Sciences

specifies that the licensor receives 15% for sales lower than $ 300 mn and 20% for sales higher than $ 300 mn we have the following problem: The licensor actually receives 20% royalties on $ 400 mn. When valuing the indications separately we do not know whether at that time the indication will be the only one or complemented by another one, possible lifting the joint sales above the $ 300 mn barrier. When valuing the project with a multi-dimensional tree it is each time clear, how large the sales are and how many indications are still alive. Multiple Indications and Success Rates Published success rates refer to compounds and not to indications. For valuation purposes we nevertheless need to know the success rates per indication, and even the joint probabilities, i.e. the probability that none, the first, the second, or both indications are successful. In what follows we offer a simple method to solve the success rate problem. The main challenge is to maintain the project’s overall success rates by calibrating the indications’ success rates. A straightforward approach is to assume that the indication that passes first, passes at the success rate P2. A second indication obviously only passes if another indication has already passed. This second indication is often assumed more likely to pass, because the compound proved viable in the first indication. The conditional success rate for the second indicand tion P22 to pass, given another indication already passed, is therefore higher than the success rate P2. This means that the overall success rate nd for the second indication amounts to P2 ⋅ P22 . If the success rate is 36% and the probability that a second indication passes, given another has already passed, amounts to 70%, then we have the following scenarios for the project: in 36%*70%=25.2% of all cases two indications pass, in 36%*(100%-70%)=10.8% only one, and in 100%-36%=64% no indicand tion passes. The assumption of the conditional success rate P22 for the second indication to pass corresponds to the success rates by indication and correlation ρ : nd

indication phase 2

P

= Pphase 2

P22 + 1 2

P2 ( A and B ) = Pphase 2 P22

nd

(3.35)

(3.36)

Advanced Topics in Project and License Contract Valuation

P2 (only A) = P2 (only B ) =

ρ=

2 Pphase 2 P22

nd

Pphase 2 + Pphase 2 P22

nd

Pphase 2 − Pphase 2 P22

231

nd

(3.37)

2

− Pphase 2 + (1 − Pphase 2 )

Pphase 2 + Pphase 2 P22

(

nd

2 − Pphase 2 + Pphase 2 P22

nd

)

(3.38)

For phase 3 the resulting joint probabilities are:

P3 ( A and B, given A and B enter phase ) = P3 P32

nd

P − P3 P32 P3 (only A,given Aand Benter phase ) = 3 2

nd

P − P3 P32 P3 (only B,given Aand Benter phase ) = 3 2

nd

P3 ( A, given only A enters phase ) = P3

(3.39) (3.40) (3.41)

(3.42)

P3 (B, given only B enters phase ) = P3

With this we can also calculate the probability that both indications survive, only one, or none.

P2 + 3 ( Aand B ) = P2 P22 P3 P32 nd

P2 + 3 (only A) = P2 P22 P2 + 3 (only B ) = P2 P22

nd

nd

P3 − P3 P32 2

nd

P3 − P3 P32 2

nd

(3.43)

nd

nd

+

P2 − P2 P22 P3 2

+

P2 − P2 P22 P3 2

(3.44)

nd

If the formulae are correct, then the probability that at least one indication passes phases 2 and 3 should equal P2*P3, which is effectively the case. We can now value the project either with DCF or with real options valuation. In DCF we have to correctly probability adjust each cash flow of each indication. The royalties can now be calculated by considering every scenario with its probability, i.e. no, one, or two indications reach the market. Using the multi-dimensional tree approach with two indications, this method can be implemented in the following way: We place ourselves in a node right before the end of a phase of indication A, hence the trial results

232

Valuation in Life Sciences

of A are revealed in this time step. If we are in a node of the regular tree (with i≥1 and j≥1), this obviously means that both projects are alive. If indication B has already previously passed the same phase and is now in the next, then indication A would be the second indication to pass that phase. nd We must use P 2 as success rate of indication A. If indication B finishes the same phase only later, then indication A would be the first indication to pass this phase and we must use P as success rate. For nodes in the zero layer (j=0) indication B failed already earlier and indication A would be the first indication to pass this phase anyway. Again, this method can be extended to more than two indications. nd The estimation of the correlation or the conditional probabilities P22 nd and P32 is a difficult task. Data for projects with various indications are scarce and not published in way it is useful for our purposes. We propose to choose the conditional probabilities in a way that the average expected peak sales correspond to sales figures of comparable projects. Valuing Projects with Multiple Markets

It is frequent that drug development companies only maintain a sales force in their home market, American companies in the United States, European companies in Europe. These companies then out-license the commercialization rights for the rest of the world to a global or several local partners. The payoff structure stemming from the market abroad is therefore different. In some cases, the partners already start to jointly develop the compound. In this case, the company that originated the project receives R&D funding and only pays a part of the development costs, or receives upfront and milestone payments. Usually once the companies must apply for approval, the geographic separation starts. As a consequence, the originator receives milestones and royalties. Different geographic markets behave slightly differently, sometimes it even happens that a drug is approved in one country but disapproved in another. Also the commercial potential can change independently due to political or demographic reasons. In DCF we assume peak sales for every market and consider the cash flows from all markets in the calculation. It is possible to assume different success rates for the approval phase. When valuing the project with real options, we have the choice between modeling the markets differently, which is done by multi-dimensional trees, or we can assume the sales potentials in each market to move jointly. This means, if we have an increase in

Technology Valuation and Feed Rate

233

expected peak sales in the European market by 10%, the sales potential in the rest of the world increases by 10% as well. The tree then models the worldwide peak sales of the drug. This can be split into different markets, e.g. 40% USA, 40% Europe, 10% Japan, 10% rest of the world. We can then calculate the value of the project in the end nodes for each market with the corresponding share of global revenues, some commercialized by the company itself, others licensed. We then have to sum up the values and work back the tree as usual.

Technology Valuation and Feed Rate We can classify technology in two categories, first, technology serving as a tool, and second, technology generating projects. The two classes are valued differently. Technology serving as a tool, e.g. for RNAi synthesis, can be used in the own laboratory for research, to provide it to customers on a service basis, or to license it out. The valuation relies mostly on the prediction of the future cash flows derived from service or licensing. These must then be discounted back to valuation date and summed up using standard DCF valuation. Commercial success, competition, legal protection, reinvestment and amplification are to be considered when predicting the future cash flows. Valuing a service company collecting fees, we do not require to use success rates for the valuation the technology generates immediate cash flows as a fee for services model. The valuation of a technology that generates new projects is more challenging. Imagine an apple tree. Each year the apple tree grows apples that can be sold on the market. For this the tree must be watered, dung, cut, and harvested. The apples can be seen as revenues, the labour as costs. A valuation of an apple tree would then include the apple sales and the invested labour of each year. First, the value depends on the kind of apple tree, Granny Smith apples are differently priced than Belle de Boskoop apples. Second, it matters whether the apples are sold directly to customers or to a vendor; again the prices are different. Third, the tree’s yield determines the quantity of apples that can be picked each year. Forth, we must account for the labour required to finally sell the apples. A project generating technology is similar to an apple tree. Its value lies in the projects resulting from it. Like with the apple tree, the nature of the projects is important. Is it a technology generating leads in oncology or does it offer validated

234

Valuation in Life Sciences

targets for cardiovascular diseases? The company’s strategy matters as well. Are the projects taken to market by the company itself or outlicensed at some stage? We must also know the yield of the technology, the feed rate. How many projects does the technology generate each year? Finally, what are the costs to run the technology? The main components of a project generating technology are therefore: 1. Project characteristics (disease area, indications, stage etc.) 2. Strategy of company (self-conducted, licensed) 3. Feed rate 4. Costs Using these four points we can quantify the value of the technology using the formula below. Vtech

(3.45)

⎡ ⎤ ⎢ (kind , strategy ) − Costs ⎥⎥ =− 1+ r) f V Costs + ( ⎢ Ni project

i =1 First year discount ⎢ feed rate project value ⎣ ⎦⎥

∑

−i

Each year, we account for the generated value with the technology, i.e. the value of the generated projects. From this we subtract the costs to run the technology. Note that the costs for the first year are due immediately, while then the costs for each consecutive year coincide with the project output of the previous year. The value of the technology is then the sum of the discounted values of each year. If the feed rate remains constant, the projects are assumed to have all the same value, and the costs are fix, we then can even apply the value for the terminal value.

[

Vtech = −Costs + f V project − Costs

[fV ]∑ ( ) = 1 + r

−i

i =1

discount

project

]

− (1 + r )Costs (3.46) r

The concept of a feed rate has two enormous advantages. First, it is a well imaginable parameter that can be justified. We will see in the section about company valuation that it can be used to replace terminal value, but without the flaws of the terminal value. Second, the feed rate is useful for portfolio management considerations. Portfolio management aims at building a valuable and sustainable pipeline. If we do not include the technology projects in these considerations, then every pipeline is condemned to dry out. Only with the projects the research

IP Valuation

235

labs will feed into the pipeline and that will be acquired through licensing activities, a useful analysis about the future health of the pipeline can be made. The following figures display the difference of a portfolio view with and without technology projects.

Fig. 3.72. Average pipeline structure of a drug development company without technology projects. Different shadings mean projects in different phases (calculated with ri:val)

Fig. 3.73. Average pipeline structure of the same drug development company including technology projects (calculated with ri:val)

IP Valuation In life sciences it is important that inventions can be protected by patents. Particularly the development of drugs and medical devices requires a lot of investments and time, and is furthermore subject to a high degree of uncertainty. If the inventor cannot protect the resulting product, probably nobody would endeavour to develop a drug or a medical device, because once it is

236

Valuation in Life Sciences

proven to be viable, competitors could jump the train and profit from the work. Others would harvest the fruits of the inventor’s work. It is therefore necessary to allow the inventor a certain timespan, where he can exploit his product on an exclusive basis. Hence, patents, or in more general terms intellectual property (IP), represent the major source of value in life sciences industry. In what follows we discuss how to value intellectual property. Intellectual property comprises intangible assets (non-physical resources or rights to other assets) such as patents, trademarks, and copyrights, protected by law. In life sciences, IP refers mostly to patents, i.e. molecules, technologies, proteins, formulations, etc. A patent is defined as the exclusive right to commercially exploit an invention and the right to sue others for infringement. The invention and the claims tight to it in the patent are protected by IPR. Often there is no clear use of the terms IP and patents. For the purpose of the valuation we have to clearly define if we talk of a single patent or of a basked of patents or IPR allowing commercialisation of a project. Often it is impossible to distinct between a patent and the entire project to commercialise the invention. Some patents however do not offer sufficient protection if not linked to other patents. And if not the complete basket of required patents is in the same hands, the single patents may turn out to be worthless. Projects in life sciences start with certain inventions and know-how that at one point will be patented or already are protected. In contrast to project, license contract, or technology valuation, IP valuation takes place at a very early stage. We often do not even know how the IP will be used. Additional uncertainties are involved; a patent may not be filed at the time of valuation and the granting of the application, as well as the comprising claims is uncertain. Furthermore, due to its early stage, the degree of commercial and technical success is much more volatile. The result of patent valuation is therefore subject to a high degree of uncertainty. Since the life sciences industry makes its living out of exploiting intellectual property, IP/patent valuation naturally matters in a variety of situations: • Fundraising • M&A • Licensing • Lending • Litigation • Accounting

IP Valuation

237

We either want to define how valuable is the IP as part of a company to invest in or to take over. Or we want to fix the price of a license, to use IP as security for lenders, or to calculate the damage inflicted on the holder of the IP. The different situations require that the IP is valued in their context. Fundraising. The value of a start-up company or young biotech company is mainly attributed to its IP. To represent the value of the company, the management has to value the IP in a way corresponding to how the raised funds will be used to exploit the IP. M&A. In the case of an M&A the buyer and the seller will value the company with its IP assets. The two parties will attribute different values to the IP. The seller needs to know what is the value of the IP in his company prior to the M&A. He considers what profits the company could generate by exploiting the IP. The buyer on the other side needs to know what value the IP has once it is integrated in his company. It might be that it generates more value than in the seller’s company by complementing existing IP, or by the buyer’s better ability to exploit it. Licensing. Universities often out-license IP at a stage where it has not yet generated a project but offers the possibility to do so. Often universities do not have the resources to develop the IP further and therefore want to sell it. As the university itself can’t go ahead with the IP, buyers often get it at a very low price. In the case where a company wants to out-license a patent it can’t exploit on its own to a company that has the complementary IP to use the patent, we have to value the patent in this context, i.e. we have to calculate the potential of the patent in the hands of the in-licensing company. As we will see later, this value is the upper limit for the negotiation. IP must always be valued in the right context. Valuing IP for licensing or M&A, we calculate the value of the IP for the buyer and seller in respect to his capabilities. Lending. As with fundraising, lenders want to know where they put their money and what kind of security they have. IP can then be used to collateralise the debt. Litigation. In the case of litigation, the valuation has to produce the value of the damage that was generated by the infringement of IP rights. The holder of the rights wants to be reimbursed for the incurred loss of profits and a possible damage to his brand. Accounting. Valuation of IP is a part of the balance sheet represented in intangible assets and goodwill. Valuation for accounting purposes is per-

238

Valuation in Life Sciences

formed according to national or international accounting standards (USGAAP, IFRS). In order that IP can be valued it should be: • A standalone asset • Licensable in total or in parts • Protected A standalone asset is IP enabling a company to exploit it without requiring further patents. We should not value a patent that on its own cannot be used, i.e. where a concomitant technology is needed. Patents that on their own cannot be exploited do not represent any value to the company. It is of course possible to commercialise a project without it being protected by IPR. The value of the patent in this situation lies in the excess profits arising from protecting the invention. Imagine for instance a patent that protects a product in a certain geographic area, or a patent that covers an improvement of the original product. These patents allow avoiding competition in this geographic area. Or they might cover an improved version of the product and consequently increase and lengthen sales. In the valuation, we have to adjust the cash flows taking into consideration the influence of IPR. The IP might offer different fields of applications, e.g. it might be a technology to be used for cancer drug research or a service technology used for other companies. Obviously, the more applications a patent offers, the more value it has. But for any application that we value, we have to account for the required resources to exploit it, i.e. capital, head count, laboratories. All these factors influence the risk and the cost of capital of the company and need to be considered when valuing IP in the context of the whole firm. IP valuation techniques. There are three main methods used to value IP: 1. Income based (DCF, ROV) 2. Cost based 3. Market based (comparables) Income based approach. The income based approach uses the methods we have outlined above, i.e. DCF or ROV. The value of the IP is composed by the various cash flows generated by it, either in form of projects, licenses, or services.

IP Valuation

239

Cost based approach. The cost based approach works by defining the cost to replace the value of the IP. The disadvantage of this method is that is does not relate to any market value or future profits. This method is generally used in accounting statements. Market based approach. The market based approach looks for comparable IP that has been traded and whose value is known. The approach is widely used and often very helpful when the determination of the input parameters is difficult due to a high degree of uncertainty. Despite being very simple and straightforward, we often encounter the problem that there is no comparable IP, or that the details of license contracts of comparable IP are not disclosed. IP might be acquired for strategic reasons and with a considerable premium. These transactions also cannot serve as comparable. As mentioned above, IP valuation follows the same rules as project and license contract valuation. There are differences that need to be considered in addition to what we have already outlined in the previous chapters. As IP often covers different applications we need to know when we value which application. This is determined by the context of the valuation. If we value IP for fundraising, we will most likely try to value most applications to achieve a higher value. While in licensing, the IP rights are often split in different applications. The buyer will therefore only be ready to pay for the licensed application, not the entire IP entity. In this context, we have to decide in what direction the IP will be exploited and then value it according to the projects or licenses that are generated. However, often IP valuation takes place in a too early stage to know where its development will lead. In life sciences, IP is the basis of drug or medical device development projects, or IP covers a technology that is used for service providing. IP therefore either leads to the generation of new projects, or as a means to provide a service to clients. We have discussed in the chapter on technology valuation that a technology producing projects can be valued using the feed rate. IP covering a technology for service can also be valued with DCF or ROV as we will outline below. We will now discuss the theory of IP valuation using DCF and ROV followed by a case study. The DCF valuation of IP follows the theory outlined above. We first define the applications we need to include into the valuation depending on the goal of the valuation and then go trough the same steps as for the previous DCF valuations.

240

Valuation in Life Sciences

Cash Flows The cash flows for the valuation of IP are again defined by their time, their size, and their probability to occur. They are assessed in the same way as described in the previous chapters. Costs that have to be taken into account are the filing costs for national and international applications and the patent renewal costs. In many cases we will have to assume that the IP or the resulting projects will be out-licensed. We will then suppose a potential license contract based on industry figures. Discount Rate The discount rate depends on the company working on the IP and the anticipated scenario of the IP. If the IP is licensed, the discount rate must reflect the risk of the licensee. Case Study IP Valuation Dr. Postdoc has developed and patented a new technique generating new cytotoxic molecules to fight cancer. As he is not able to exploit the technology in its full potential, he considers licensing it to a large company developing cytotoxics. He now wants to find out how much value the technology might have in the hands of a company with adequate resources. He considers this value to be the upper limit in the negotiations with the buyer. In order to get a better understanding of the potential value of his invention, he first determines the following four points: 1. Project characteristics (disease area, indications, stage etc.) 2. Strategy of company (self-conducted, licensed) 3. Feed rate 4. Costs The molecules will be for the treatment of cancer, mostly late stage advanced cancers. The exact indications at this point are completely in the dark. Dr. Postdoc assumes the following numbers for the typical project generated by the technology (cf. Table 3.71). Dr. Postdoc estimates that the technology can be exploited as long as the patent protection is valid, hence for 16 years. The peak sales are the median sales for cancer drugs, in our case $ 344 mn. The launch costs are about half of a peak sale, for the prototype project $ 150 mn, and include building manufacturing and setting up a marketing campaign. Dr. Postdoc

IP Valuation

241

Table 3.71. Project parameters Lead Preclinical Phase 1 Phase 2 Phase 3 NDA Costs ($ mn) Success Rate Length (years)

2 70% 1

3 70% 1

5 64% 1

15 42% 2

50 65% 3

2 90% 1

calculates with a margin of 65%. The margin accounts for the cost of production and marketing. The sales curve follows the curve outlined in the previous chapters. The technology allows generating new chemical entities that enter lead optimisiation. He assumes that the company will develop and launch the project on its own, i.e. not to license it. With the right capacities, the technology allows to bring one project into lead phase every year. The costs for running the technology consist of headcount, use of infrastructure, material etc. He estimates a feed rate of one project entering lead optimization can be achieved with allocating $ 1.5 mn each year, with the first project entering that phase after one year, with a prior investment of $ 1.5 mn. Solution We first value the prototype project generated by the technology. We calculate the project value with DCF and real options, assuming that in year one and year two we have $ 1.5 mn per year of expenses. Then we can start with the project. The project valuation, based on a volatility of 25%, yields the following result: Table 3.72. Prototype project value Discount DCF

ROV

15%

$ 3.71 mn

$ 2.99 mn

The value for the prototype project is $ 3.0 mn with DCF, and $ 3.7 mn with real options. Assuming that the technology keeps generating one project each year until infinity at the predefined costs we have to apply the following formula: ∞ Vtech =

[fV

project

− (1 + r )Costs r

]

(3.47)

242

Valuation in Life Sciences

However, it is only assumed that the technology generates projects for n years; in this case n equals 16 years. We therefore have to subtract from the value above all projects generated after n years, the exceeding projects. Placing ourselves n years later in the future, the value of all exceeding projects ∞ , because the situation after n years would be would amount to exactly Vtech the same as now: The technology generates each year two projects at the predefined costs until infinity. The value must therefore be the same. And ∞ since the value of all exceeding projects amounts to Vtech in n years from ∞ (1 + r )− n tonow, the present value of the exceeding projects amounts to Vtech day. The value of the technology over the next n years is therefore: Vtech =

[fV

project

]

− (1 + r )Costs ⎛ 1 ⎜⎜1 − n r ⎝ (1 + r )

⎞ ⎟⎟ ⎠

(3.48)

Using Dr. Postdoc’s numbers for the costs ($ 1.5 mn), the feed rate (1), the duration of the technology (16 years), and the prototype project we get: Table 3.73. Technology value Discount DCF

ROV

15%

$ 11.82 mn

$ 7.53 mn

The technology has a value of $ 7.5 with DCF and $ 11.8 with real options. The difference stems from what we have discussed in earlier chapters, namely the option to abandon projects that are not profitable. Discussion We have learned how to value a technology that generates projects over a certain time span, or infinitely. The approach is very useful for the valuation of intellectual property and firms and is simple to handle, once the standard concepts of DCF and ROV are mastered. Technologies can even be valued assuming that the resulting projects are licensed out at one point using the valuation outlined previously. A tricky part of the method is to properly allocate the costs the technology causes. This is not always evident. A further point to stress is the sensitivity of the method to the used discount rate. A high discount rate has a tremendous effect on the value of the technology, as the formula is comparable to a terminal value calculation. In cases where companies are expected to grow over time, as most young biotech companies do, the formula can be adapted

Pipeline Valuation

243

to a changing feed rate. We then have to introduce g for the growth rate (in percent) into the term, resulting in the following expression: Vtech =

[fV

project

]

− (1 + r − g )Costs ⎛ 1 ⎜⎜1 − n r−g ⎝ (1 + r )

⎞ ⎟⎟ ⎠

(3.49)

In the case where we cannot predict a feed rate, but rather the time when certain projects enter the respective phases, we create a timetable for all future projects and value them according to the standard methods. This is more time consuming but gives us more flexibility in our prediction. Our example has shown that valuating with real options can make a sensible difference. The higher value in the case of real options valuation can be used to negotiate better terms in technology licensing.

Pipeline Valuation A pipeline is a basket of projects. The value of a pipeline is therefore the sum of its projects. But the value is only one aspect we are interested in when valuing a pipeline. Other aspects are the risk profile, the distribution of cash flows and the associated liquidity needs, or the structure of the pipeline. A standard project valuation does not offer much insight into these aspects. Simulations however are the perfect tool to analyse a pipeline with respect to risk and development in time. A very important property of simulations is the ability to consider correlations. If we flip two coins and receive a dollar for heads, we expect to win one dollar on average, half a dollar on average by coin. If by some magic way, the second coin always shows the contrary of the first coin (perfect negative correlation between the coins), we would even end up with a safe dollar, because we either win with the first or with the second coin. Valuation just tells us that playing the game is worth 1 dollar, for each coin half a dollar on average. Simulations on the other hand tell us that the first game is risky; in 25% we end up with zero dollars. Since people usually are assumed to be risk averse, one would prefer playing the second game. In a pipeline of a life sciences company we can also observe such correlations, although not as extreme as in the coin example. Two drugs that base on the same mechanism of action are partly affected by the same risks, e.g. COX-2 inhibitors. After the withdrawal of Vioxx by Merck, Pfizer had to withdraw its COX-2 inhibitor Bextra as well, and Novartis temporarily withdrew its COX-2 inhibitor Prexige in the EU.

244

Valuation in Life Sciences

Simulations allow us to include such correlations by simulating common risk factors. This is especially helpful for portfolio management purposes, where companies want to achieve a sound diversification. If several projects depend on the same risk factors, e.g. another company launching a product that competes with several own products at the same time, simulations will visualise this cluster risk. The figures below display a diversified and an undiversified pipeline. The first pipeline is composed of projects in various disease areas, using different technologies and mechanisms of action; the second is composed of similar projects, all using the same mechanism of action. The average value remains the same, but we clearly see that the undiversified pipeline has much lower and higher extreme values. This is a consequence of the correlation between the projects. While an uncorrelated pipeline remains in almost all cases a mix of success and failure, a correlated pipeline is much more a black and white picture, either the mechanism of action is efficacious and safe and all projects pass, or it is not efficacious or safe and no project passes.

Fig. 3.74. Diversified pipeline (calculated with ri:val)

Fig. 3.75. Undiversified pipeline (calculated with ri:val)

Portfolio Management

245

Another very interesting aspect is the simulation of the pipeline structure, i.e. how many projects are in which phase. A portfolio manager can anticipate future gaps in the pipeline and can either advocate for larger resource allocation to in-house research or initiate further in-licensing activities. The figure below illustrates the median case of the pipeline, i.e. 50%best case. In half of the cases the company can count with taking at least two compounds to the market. Nevertheless, within two years the current pipeline of 8 projects will have shrunk to just 4 projects. It should therefore take care of a sustainable pipeline feed.

Fig. 3.76. Median company structure (corresponding to 50%-percentile, calculated with ri:val)

Portfolio Management Introduction

A detailed discussion of portfolio management is beyond the scope of this book. The aim of the following section is to give some ideas on new directions in portfolio management. Portfolio management is inseparably connected to management of biopharmaceutical companies. It forms a company’s future success and position by defining how to use resources. Portfolio management is comparable to asset allocation in banks; it has the same goals: • Investment in the best assets • Alignment of risk and strategy • Maximisation of profits • Diversification of the portfolio

246

Valuation in Life Sciences

• Sustainable growth of revenues • Management of liquidity The task of portfolio management is to decide which projects to continue, which one to in- or out-license, and which ones to abandon. The main difference between financial asset allocation and resource allocation in the life sciences industry are: 1. Access to assets. The assets being projects cannot be bought and sold on an exchange like shares, bonds, or commodities. Usually a company owning a project aims at commercialising it or licence it to a partner company. So at most two companies have access to one project. Compared to financial asset allocation the universe is at the same time considerably smaller. 2. Unity of assets. While in finance we can hold 0.1%, 1%, or 10% of a company, we can only invest in the full project. It is not possible to multiply a good asset. Licensing part of a project, however, is a possibility. 3. Risks. Life sciences assets underlie special forms of risk. Next to common market risk most assets are subject to large attrition rates. Unfortunately, these differences do not allow the standard application of portfolio optimisation strategies known from finance. The elegant models of Markowitz and Sharpe (efficient frontier) are not applicable to assets sharing the above listed properties. It is therefore necessary to look for new ways to manage a life sciences portfolio. One central problem of portfolio management lies at its very root: What should be optimised? Generally, people tend to say that portfolio management is a multidimensional task. Several aspects matter: 1. Value 2. Sustainable revenues 3. Balance between projects in different phases 4. Balance between or focus on disease categories 5. Balance between high and low risk projects All else being equal it is relatively easy to decide between two portfolios that differentiate in only one of the above aspects. Higher values dominate lower values, sustainable revenues dominate volatile revenues, balance is

Portfolio Management

247

preferred to clustering. If already two aspect are different, it becomes much less clear which project selection is preferable, e.g. the one with a higher value, or the one with less risks? Some companies specifically concentrate on one disease area or one mechanism of action, hoping for more efficiency, leading to a higher value at the expense of higher risk. Others prefer a balanced pipeline, sacrificing some of the upside potential. Qualitative Approaches

It is impossible to optimise a portfolio if it is not clear, which aspect to optimise. Portfolio management therefore often remains a rather vague topic, focussing mainly on qualitative aspects like market attractiveness, positioning with respect to competitors, strategic alignment, technical feasibility, or legal and regulatory issues. Qualitative approaches assign a value to each of these aspects and then weigh them against each other in a scoring model. Portfolio management should be objective and not prone to political interests of departments, an often seen problem within larger companies. Therefore, the process of assessing data is of enormous importance. Opposing various properties like remaining patent protection, market potential, risk, competition, or investments allows comparing projects in diagrams like bubble graphs or radar plots. Such graphs are very popular and should further encourage the discussion about pros and cons of the projects. A lot of literature and congresses elaborate on qualitative approaches. As the focus of this book is the quantitative approach to valuation, we will not go into any further details of qualitative portfolio management. In what follows, we take a stand for quantitative approaches.

Fig. 3.77. Bubble graph comparing projects with respect to investments (x-axis), likelihood (y-axis), value (bubble size), and phase (shade), calculated with ri:val

248

Valuation in Life Sciences

Quantitative Approaches

The fundamental idea of quantitative approaches is that any decision should be justifiable by a calculation. If one strategy is claimed to be better than another strategy, this should be reflected somewhere in the calculations. Strategy finding can be a very emotional process where most of the participants trust their gut feeling. Quantitative approaches try to get rid of these subjective gut feelings. Often soft factors or possible industry developments are held responsible for one strategy being superior to another. If these soft factors or industry developments are implemented in the calculation model, then the claimed strategy must yield better values; otherwise it simply is not superior. Building quantitative models on portfolio or company level is a challenging task and requires many assumptions with respect to future market developments, correlations, and resources. These assumptions are the above-mentioned soft factors and it is the responsibility of the management board to define these assumptions. Once they are set, the portfolio management can start finding an optimal resource allocation that respect the assumptions from the management board. Approach 1: Comply with Imposed Restrictions The first approach corresponds to the situation usually the portfolio management department faces: the management board gives some premises on how much resources are provided for project funding, on the focus in disease areas or molecule types (e.g. biologicals vs. chemicals), or on the number of projects to reach market within the next five years. We denote this set of premises with γ. Portfolio management then must find a resource allocation that complies with γ. We denote a resource allocation with ζ. Portfolio management must decide, which projects get funding, which are scaled down, put on hold, licensed, or abandoned. It is possible that for some projects there are different project plans that require different funding and yield different revenues (Sharpe et al. 1997). Some compounds might only be developed for one indication, other indications might be considered once the first proves to be successful. Other projects might be considered for out-licensing to a partner. A special challenge of portfolio management is to find a solution ζ that takes account of the dynamics of drug development. In the next budgeting period the selected projects require different funding because they have moved to another phase. The resource allocation ζ should avoid the creation of bottlenecks in subsequent budgeting periods. If on the other hand

Portfolio Management

249

the management is willing to raise capital in the equity or debt market, then this condition can be relaxed. Obviously, some projects are likely to fail and do consequently not require further funding. Therefore, it might be advisable not to plan for the worst case, but maybe to avoid these bottlenecks for a certain percentage of scenarios (e.g. in 90% of all scenarios the selected portfolio can be financed throughout its life cycle). Portfolio management equally has to consider the entry of new projects thanks to feed rates for each type of project (disease area, biologics vs. chemical, licensed vs. self-originated). The resource allocation problem requires the general market dynamics (feed rate of outside opportunities, market growth and development, competition), the company parameters (current projects, cash balance, feed rate of internally developed products) and the strategy (constraints) as input parameters. The solution can be approached using simulations framed by a much larger concept called artificial evolution. This solution is a resource allocation ζ∗ that complies with γ and optimises a function f. Possible options for the function f could be the value, the 10% value at risk of the pipeline, a value-risk ratio, or a measure of the sustainability of future revenues. Mathematically ζ∗ becomes:

ζ * = arg max{ f (ζ (γ ))} ζ

(3.50)

The equation states that ζ∗ is the one resource allocation out of all resource allocations ζ complying with γ that maximises the function f. We see that the optimal resource allocation ζ∗ is a function of the premises γ. Remember that the premises γ were imposed by the management board. In more common terms γ is the strategy defined by the management board, and the resource allocation ζ∗ simply must be in line with the strategy. Assuming the outlined resource allocation problem as a function of the strategy we can then try to optimise the strategy of the company. Knowing what resource allocation ζ∗(γ) will result from which strategy γ, we can try to find the strategy γ*, i.e. the restrictions for the resource allocation, such that the function f is optimised on a global level.

γ * = arg max{f (ζ * (γ ))} γ

(3.51)

In the outlined approach we have interpreted the resource allocation problem as a Stackelberg game from game theory. Game theory is a mathematical discipline that has already provided excellent solutions in other areas of science and economy and has earned John Nash the Nobel Prize in 1994.

250

Valuation in Life Sciences

Approach 2: Value Maximisation At first sight, value maximisation seems to be a straightforward problem: Simply select all projects with a positive value. Still, this approach neglects the portfolio effect. The diversification between projects leads to a reduced risk profile. This then translates into a lower discount rate, possibly justifying even the development of previously not considered project. Nevertheless, the inclusion of some projects might also lead to cluster risks, increasing the discount rate again. This approach requires a sound model for the discount rate, that correctly takes account of the degree of diversification of the portfolio. Using the proposed framework in the section about discount rates, the value of the pipeline directly depends on the discount rate r, which is a function r(ζ) of the resource allocation ζ which is equivalent with the pipeline. The optimal resource allocation strategy ζ* can then be written as:

ζ * = arg max{value(r (ζ ), ζ )} ζ

(3.52)

The basic idea is that the value eventually should also include the risk of a certain strategy. This way we achieve a single measure by which we can classify all possible portfolios and make a sharp selection. This second approach is a special case of the first approach, the function f being the value function. As a special aspect, the discount rate depends on the resource allocation ζ as well. Optimal Portfolio Structure

In every larger pharmaceutical company management asks about the optimal portfolio structure. How can they make sure that the products that go off patent can be replaced with new, patent protected drugs? This question is a part of broader portfolio management and we display one possible approach on the following pages. The outlined approach does not use any consideration about the sales potential of the projects and should therefore be combined with a more sophisticated valuation approach. Let us first assume for simplicity that all drugs follow the dynamics as exhibited in the table below, i.e. have the same phase durations and success rates. Out of the success rates we can already deduce the probability of a project in a certain phase to get to the market. A phase 3 project for instance needs to pass phase 3 and submission and has therefore a probability of 63%*79% = 50% to reach the market.

Portfolio Management

251

Table 3.74. Project dynamics Phase 1 Phase 2 Phase 3 Submission Success rate

60%

Duration (months) 18

38%

63%

79%

29

42

18

Table 3.75. Probability to reach market Phase 1 Phase 2 Phase 3 Submission Probability to reach market 11%

18%

50%

79%

Table 3.76. Average number of necessary projects Phase 1 Phase 2 Phase 3 Submission # of projects 9.1

5.4

2.0

1.3

From these probabilities we conclude, how many projects in which phase are necessary to get one project to the market. Since a phase 1 project has a probability of 11% to reach the market, we need about 9 (=1/11%) phase 1 projects to get one project to the market on average. These 9 projects become about 5-6 phase 2 projects, or 2 phase 3 projects according to the success rates. But does that mean that a perfectly balanced pipeline should be weighted according to the numbers in the above table. The answer is no, because the projects remain for different times in the phases. If we expect one approval every year, we need not only have 1.3 times as many projects in submission, because these would just lead to the one approval in that year. The projects for next year’s approval should also already show up in the pipeline. Therefore we have to multiply the weights of table 3.76 with the respective length of the phases in years. This leads us to the balanced pipeline structure of figure 3.78. If we assume that a project will have 12 years of patent protected sales, then we can deduct what pipeline should correspond to a certain portfolio of approved drugs. If there are 6 approved drugs still under patent protection, then we can expect about one approval every second year. The corresponding pipeline that is able to replace the current portfolio of drugs has half of the projects displayed in figure 3.78. This means that the company requires a feed rate of 13.6/2/1.5=4.5 phase 1 projects each year.

252

16 14

Valuation in Life Sciences

13.6

13.1

12 10 7.0

8 6 4

1.9

2 0 Phase 1

Phase 2

Phase 3

Submission

Fig. 3.78. Optimal pipeline structure

Of course a pipeline contains projects in various disease areas and the described weighting could be applied to its sub-pipelines in, let’s say, cancer, inflammation, and endocrinology. In a pharmaceutical company it is unrealistic to assume that all projects are self-originated. If the company in-licenses projects in phase 2, then these projects have never appeared in phase 1 in their pipelines. Accordingly the pipeline might look like in figure 3.79. This mixed pipeline requires an average of 6.3 in-licensed projects in phase 2. This then corresponds to a feed rate of 2.6 projects every year (remember, phase 2 projects remain 29 months in phase 2). self-originated

in-licensed

14 12 10

6.3

8 6 4

3.4 7.0

2

6.8 3.6

0.9 1.0

Phase 3

Submission

0 Phase 1

Phase 2

Fig. 3.79. Optimal pipeline structure

Portfolio Management

253

For analysts and managers alike the question of “How lucky has the company been so far” is quite intriguing. Imagine a company with the following pipeline10: 9 8 7 6 5 4 3 2 1 0

8

8

4 3

0 Phase 1

Phase 2

Phase 3

Submission

Market

Fig. 3.80. Current pipeline

Is this a good or a bad pipeline, i.e. has the company been excessively lucky until now or is the perceived attrition about normal? To answer this question, we have to model the pipeline with success rates and timelines in line for oncology projects. Since most of Genentechs projects are monoclonal antibodies we have chosen monoclonal antibody success rates (from Reichert11). Timelines are taken from DiMasi12. Table 3.77. Average number of necessary projects Phase 1 Phase 2 Phase 3 to approval Success rate

75%

Duration (months) 20

10

11

12

60%

80%

29

33+16=49

We have chosen Genentech’s oncology pipeline in 2009. Second and higher indications have not been counted. Reichert, J. et al., „Monoclonal Antibody Successes in the Clinic“, Nature Biotechnology, September 2005. DiMasi, J, et al., „The Cost of Biopharmaceutical R&D: Is Biotech Different?”, Managerial and Decision Economics, 2007.

254

Valuation in Life Sciences

The question is, what is the most likely feed rate for a company exhibiting a pipeline like in figure 3.80? It is clear that randomness is involved in the pipeline progression; first not every project moves to the next phase, and second, in the company’s scenario the success rates are not necessarily followed by the number. For the calculation we must assume a feed rate of preclinical projects, one phase earlier than our pipeline figures. Otherwise the feed rate must be 8 projects divided by 20 months (=4.8 projects/year). We are well aware that the feed of IND projects is also somewhat random, therefore we assume a constant feed rate one phase before the first phase we look at, allowing the preclinical success rate (65%) to act as “randomisation” to the IND feed rate. Furthermore we don’t look at the commercialised projects simply because the exhibited pipeline is a result of an increased R&D activity over the recent years and the number of commercialised projects does not reflect that. Now we can attribute to the current scenario an exact likelihood following the steps as in table 3.78. Table 3.78. Likelihood caluclation Phase 1 Phase 2 Phase 3 Probability (p)

65%

49%

29%

Duration (d)

1.67

2.42

4.08

Projects (n)

8

P(# projects=n) Combinations (C) Total

8 n

3 (feed*d-n)

p *(1-p)

(feed*d)!/((feed*d-n)!*n!)13 P1*C1*P2*C2*P3*C3

With this combinatorial calculation we can then assess the likelihood of different feed rates as displayed in figure 3.81. Apparently a preclinical feed rate of 6 maximises the likelihood of the current pipeline, i.e. the current pipeline is a more realistic scenario if we assume a feed rate of 6 preclinical projects than only 5 or 6.5. What does that mean? 13

Since the factorial function is only available for integer numbers and the feed rate, representing an average number, can also be non-integer, we need to use the Gamma function, an extension of the factorial function to non-integer numbers.

Company and Stock Valuation

255

0.12% 0.10% 0.08% 0.06% 0.04% 0.02%

10

9.6

9.2

8.8

8.4

8

7.6

7.2

6.8

6.4

6

5.6

5.2

4.8

0.00% Feed Rate

Fig. 3.81. Likelihood of current scenario given different preclinical feed rates

If the actual preclinical feed rate of the company is less than 6, then the company currently manages to keep attrition lower than the industry average. One can interpret this as operational excellence, in which case could increase the success rates until the most likely feed rate matches the observed one. This is a method of success rate calibration specific to the company. On the other hand, if we do not believe that some companies can reach higher success rates than others in the long run, this means that the company has just been excessively lucky in the recent years.

Company and Stock Valuation Introduction to Company Valuation

Company valuation – or better: valuation of a company’s equity – is indispensable in different fields of the life sciences industry. We encounter it throughout the entire life cycle of the industry: Fundraising. Any investor wants to know what he gets for his money. As most young companies are private, the company’s share price is unknown or at least opaque. Seed investors, venture capitalists, and private equity investors need to value a company to determine the right share of ownership they are entitled to when investing. Equally, the existing shareholders need to value the company; they want to limit the dilutive effect. In the negotiations the parties agree on a value and allocate the new issued shares accordingly. Consequently, each round corresponds to a valuation. Inves-

256

Valuation in Life Sciences

tors will have different goals during the negotiations, as illustrated in the case at the end of this section. Valuation in fundraising is viewed from two sides, on the one hand new investors value the company where they want to put their money in; On the other hand, the company looking for funds often asks a third party for a valuation or does one on its own. In this constellation, we often face largely diverging views on the value of a company. Company founders and the upper management represent the existing investors and have a personal interest in a high valuation in order to minimise dilution. New investors will try to get the value down, so they get a larger piece of the company with their investment. Reconciling these two views can be very challenging and will often result in one party getting up from the table. Crucial for investors is to have a solid framework to trust their own valuations. The discount rate and the valuation method should be used very consistently over the period of various fund raising rounds; otherwise the valuations become meaningless. Licensing. Licensing payments often include the exchange of equity. As with fundraising, companies are often not listed. There is no official share price. The value of equity needs to be derived by the valuation of the company. M&A. Mergers or acquisitions are common situations where the value of a company is essential. The valuation, in addition to fixing the fair value of equity, can provide information about the value of potential synergies and economies of scale. In the negotiations for the M&A, valuation delivers the lower boundary for the seller and the upper for the buyer. IPO. The issuing price of the share at an initial public offering will obviously derive from the value of the company. As in private rounds, the investors set the price, but in an IPO these investors are the public market players. Analysts. Analysts regularly value private, and mostly public companies. Based on the valuation they recommend buying or selling firms. The following minicase illustrates the different views of investors during a round of financing. Minicase Investors’ Preferences Mr. Venture, Dr. Capital, and Mr. PrivateEquity are investors in the private company BioComp. Today, Mr. Venture holds 70% of BioComp and Dr. Capital the remaining 30%. In the upcoming financing round, Dr. Capital

Company and Stock Valuation

257

and Mr. Private Equity are going to invest $ 20 mn each. All three have to perform a valuation, Mr. Venture in order to defend its right as shareholder, Dr. Capital as shareholder an investor, and Mr. Private Equity as new investor. After preliminary valuations they agree that BioComp’s premoney value lies between $ 40 mn and $ 60 mn. After the financing round, BioComp’s post-money value will therefore be between $ 80 mn and $ 100 mn. Dr. Capital and Mr. Private Equity invest $ 40 mn; the new issued shares must correspond to 40% (if pre-money value $ 60 mn) to 50% (if pre-money value $ 40 mn) of the share capital. What are the respective interests of the three involved parties? Mr. Venture currently holds 70% of the shares, corresponding to a value between $ 28 mn (=0.7*40) and $ 42 mn (=0.7*60). After the financing round, his stake in the company therefore drops to 35% to 42%. Mr. Private Equity contributes $ 20 mn to the new company, corresponding to 20% to 25%. Finally, Dr. Capital already owns 30% of BioComp, corresponding to $ 12 mn to $ 18 mn. With the additional $ 20 mn he invests, this becomes $ 32 mn to $ 38 mn, corresponding to 38% to 40% of the company value. Table 3.79. Investor’s share In $ mn

Pre-Money Contribution Post-Money

BioComp

40

60

Mr. Venture

28

42

Dr. Capital

12

18

Mr. Private Equity

40

80

100

35% 42% 20

40% 38%

20

25% 20%

As we can see in the table, Mr. Venture tries to negotiate a high valuation; the higher BioComp is valued, the less his shares are diluted. Mr. Private Equity clearly favours a low valuation, giving his contribution more weight. For Dr. Capital the situation is a little different; he acts both as existing and new shareholder. As an existing shareholder, he is interested in a high valuation as Mr. Venture. But in order to capitalise on his new investment he also wants a low valuation, like Mr. Private Equity. Since Dr. Capital contributes 50% of the new capital, contrasting his mere 30% of ownership, he prefers a low valuation. Not only that a post-money valuation of only $ 80 mn gives him 2% more ownership, he also becomes the main shareholder of BioComp, overtaking Mr. Venture. If Dr. Capital would contribute less than 30% of the new capital, then he is better off with a high valuation.

258

Valuation in Life Sciences

Theory

Methods and Fundamentals We will now discuss how to value companies and stocks and show the difference in the valuation of public and private companies. For better understanding we will repeat several basics of finance. The balance sheet of a company is formed by assets on the left hand side and equity and liabilities on the right hand side. The balance sheet is a snapshot of the firms accounting value; the value at a certain point in time. Balance Sheet Liabilities Assets Equity

Fig. 3.82. Structure of a balance sheet

The balance sheet shows what the firm owns (left side) and how these assets are financed (right side). The value of the company is the value of its assets. The assets are financed on the one hand by equity, on the other hand by debt (liabilities). In contrast to the balance sheet, the income statement and the statement of cash flows tell us about the firm’s financial activities during a certain period, e.g. one year. The different angles to look at a company, value, income and cash flows, are the source for different methods to value a company. Fig. 3.83 summarises common approaches.

Asset based firm valuation. The asset based valuation approaches mainly use the numbers as represented in the balance sheet. However, the balance sheet serves accounting purposes and does not correspond to a fair valuation of the assets mentioned therein. Asset based approaches are often not suitable for life sciences companies, where most of the value is captured by intellectual property, which is not appropriately represented on the balance sheet.

Company and Stock Valuation

Firm valuation approaches Adjusted book value Liquidation value Cost to create Income based Economic value added Excess earnings method Capitalisation of benefits Cash flow based Discounted cash flows Real options Market based EBIT multiples EBITDA multiples Revenue multiples Asset based

Fig. 3.83. Company valuation methods Table 3.80. Osiris Therapeutics fundamentals (source: NASDAQ) Last Sale Change Net / % Best Bid / Ask 1y Target Est: Today’s High / Low Share Volume 50 Day Avg. Daily Volume Previous Close 52 Wk High / Low Shares Outstanding Market Value P/E Ratio Forward P/E (1yr) Earnings Per Share Annualized Dividend Ex Dividend Date Dividend Payment Date Current Yield Beta

$ 11.10 0.32 2.8% N/A / N/A $ 20.00 $ 11.49 / $ 11.10 7,096 N/A $ 11.42 $ 12.30 / $ 9.84 27,036,000 $ 300,099,600 N/A N/A N/A N/A N/A N/A N/A 0

259

260

Valuation in Life Sciences

Income based firm valuation. These methods deduce the company value from the income statement. As only a fraction of life sciences companies makes profit, income based approaches play a marginal role in early stage life sciences valuation. They are better applicable to late stage companies. Cash flow based valuation. The value of a company can be derived from its future discounted cash flows. We focus on cash flow based methods like DCF and real options, as this approach is the only accounting for the particularities of early stage life sciences companies. Market based valuation. With these methods we can compare different companies. Like the asset and income based methods these approaches are little useful for life sciences companies until these start generating earnings, because earnings and profitability are the main input parameters. Even for publicly traded companies many ratios and measures simply cannot be applied, as shown in Table 3.80. Table 3.80 summarises the publicly available financial figures for Osiris Therapeutics. As the company is not generating any remarkable earnings yet; we cannot use the standard multiples like the P/E ration or earnings per share. In the continuation, we will only treat the cash flow based approach, the best suited approach to value life sciences companies.

Building Blocks of Company Value Life sciences companies depend on the fate of their projects. Their maturity depends on the state of their pipeline, success rates act on projects, and decisions are made on a project basis. This strong project orientation makes it necessary to structure the valuation accordingly. The value of equity therefore is composed of: 1. Projects 2. Technology and IP 3. Cash 4. Other fixed assets 5. Unallocated costs 6. Debt 7. Taxes Projects and technologies are intangible assets, but do not appear at a fair value in the balance sheet. For valuation purposes, we must use the NPV of

Company and Stock Valuation

261

their cash flows, either calculated with DCF or real options. The value for cash and other fixed assets, we can directly take from the balance sheet. Unallocated costs do not appear on the balance sheet, but are necessary to maintain operations; without these expenses the projects could not be realised. Debt and taxes are again NPV’s of cash flows related to these items. The proposed method is a cash flow to equity approach, without accounting subtleties like depreciation or amortization. We therefore have to use the cost of equity as discount rate, even for the debt part. If we want to value the company as an entity, we would have to use the WACC, without deducing the debt. Unallocated costs, debt, and taxes are value diminishing.

Equity Value Taxes

Debt Unallocated Costs Other fixed assets

Cash Technology

Projects

Positive

Negative

Total

Fig. 3.84. Structure of equity value

Projects. Each project represents value. Since characteristics and decisions vary from project to project, it makes sense to value each project separately. Important is the overall analysis (due diligence) of the company’s work, if is well protected by IPR and if does not face the risk of litigation. All assets are valued with respect to their legal security. Technology and IP. The technology of a company is its ability to generate new projects from its IP. Assuming the company value built just by what actually exists, i.e. by the projects and technologies the company owns, the cash flows of the company vanish after the life cycle of the longest lasting project. This would mean that the company would stop existing after that project. Of course, this is not realistic. The company will generate new projects also in the future; after all that is the main task of the manage-

262

Valuation in Life Sciences

ment. Hence, it is legitimate to assume that the company will also have cash flows after that. It is very uncertain to predict cash flows in life sciences; they are subject to a substantial attrition rate. It is much easier to assume that the research department of a company is able to generate a certain output deriving from the company’s IP portfolio or that the licensing office is able to close a certain number of deals each year. This then corresponds to feed rates of projects and licenses as we have introduced them in the section about technologies. While a small biotech company might own a technology enabling it to generate some leads each year for a certain disease group or using a certain mechanism of action, a large pharmaceutical company has the capacity to generate a multitude of new drug candidates in various disease areas and closes a many license contracts each year. For a large pharmaceutical company, we therefore have to assume feed rates for several prototype projects, self-originated or in-licensed. If the company does not provide the feed rate publicly, we can elaborate by comparing the state of the pipeline several years ago, and then track new projects entering a certain phase.

Cash. Of course we must consider the company’s cash and cash equivalents and also the tangible assets. This item does not require any valuation and can be taken directly from the balance sheet. In the view of an investor, however, the cash in a company might not necessarily be worth the same as if it would be right in the investor’s hands. The cash is committed to projects and investments, which might not necessarily be worthwhile in the investor’s eyes. Neither can the cash be immediately extracted out of the company because of administrative hurdles. For these reasons the cash of a company is often only accounted for in a valuation with a certain discount. Furthermore, cash is unproductive as long as it is not invested. Invested money is assumed to accrete at the discount rate, cash only at the interest rate, which is lower. Nevertheless, cash is also some kind of security, since it avoids a company to lose dearly needed management time for fund raising. In the case of larger, already profitable companies cash can be used for the sake of either organic expansion or acquisitions. If such opportunities do not exist, or if the company thinks that its war chest is sufficiently filled it can still use the cash for share buy-back programs or give the earnings back to the shareholders in the form of dividends. However, these measures are usually a sign for lack of alternatives. Other fixed assets. Other fixed assets include plants, property, and machines and are also part of the balance sheet.

Company and Stock Valuation

263

Unallocated costs. A company has costs that cannot be allocated to a specific project or to a technology. Typically, the CEO’s salary, or the office rent is a cost that has to be considered on a company basis. This category can have a very high impact on the value reduction of a biotech company and should be elaborated carefully. This can be challenging depending on how the company is handling its budgets but is absolutely necessary to perform a proper valuation. The more costs can be allocated to project plans, the less the value reducing effect, because they are subject to success rates this way. Debt. All liabilities reduce the value of equity. We must account for all cash flows that serve paying back the debt, i.e. interest payments and payback of the debt’s principal at the end of the maturity. Taxes. On a company basis we also include taxes. Taxes must be paid on earnings after interest payments, in case the company has debt. Young biotech companies do not make profit and therefore do not have to pay taxes. But if we predict cash flows for the future, we must consider tax influence also for those companies. In most countries, the losses can be taken forward; only once the earnings offset the accumulated losses, the company starts to pay taxes. Some governments offer tax credits, i.e. the companies can reclaim parts of their losses, i.e. for R&D funding. This way the companies earn tax credits. Financial Statements As we have already treated in the previous sections the valuation of the components of a life sciences company, we now want to discuss some specific problems of company valuation. We start with financial statements used as data source. Unfortunately, standard financial statements do not allocate revenues and costs to projects and technologies; they summarise them over the entire firm. For the valuation however, we need to break down the cash flows into cash flows allocated to projects, technologies, and cash flows that are unallocated. Table 3.81 below displays the income statement for Biotech Corporation in the year 2XX7. The company has three projects in development, and an income generating technology. Project X3 is licensed out, triggering a milestone payment in 2XX7. We immediately see that the table does not allocate any of the costs to the individual projects of the company. We can now crudely allocate income and expenses to the projects. Tax and SG&A belong to unallocated costs, costs that are not assigned to the projects.

264

Valuation in Life Sciences Table 3.81. Biotech Company income statement Biotech Company Income Statement 2XX7 (in 000’s) Total operating revenues

$0

Gross Profit

$0

Operating Expenses R&D

$28’000

Selling, General, and Administration

$8’000

Operating Income

$(36’000)

Additional Income/Expenses and Tax

$12’000

Earnings before Interest and Tax

$(24’000)

Interest Expense

$1’200

Earnings before Tax

$(25’200)

Minority Interest

$0

Net Income-Cont. Operations

$0

Disc. Operations

$0

Net Income

$(25’200)

Net Income to Common Shareholders

$(25’200)

Table 3.82. Biotech Corporation allocated and unallocted costs Biotech Corporation 2XX7 (in ‘000s) Technology

Pr. X3

Income

$1’000

Expenses

Pr. B2

Pr. L4

Unallocated

Firm

$11’000 $0

$0

$0

$12’000

$0

$0

$0

$0

$0

$0

SG&A

$0

$0

$0

$0

$8’000

$8’000

R&D

$9’000

$0

$11’000

$8’000

$0

$28’000

Interest

$0

$0

$0

$0

$1’200

$1’200

Tax

$0

$0

$0

$0

$0

$0

Net Income

$(8’000)

$11’000 $(11’000) $(8’000) $(9’200)

$(25’200)

Company and Stock Valuation

265

As a second inconvenient, financial statements report either values in a way that serves accounting but not valuation, or they give cash flows only for one year. For the valuation of the company, we theoretically need the cash flows until infinity. In practice, we use the time of the life cycle of its current projects. The feed rates then take care of the remaining cash flows until infinity. Consequently, the financial statements are only little help for the valuation of a company. We have to use forecasts not only for the projects and technologies, but also for the unallocated costs. The financial statement can be used to see whether the forecasted R&D spending or unallocated costs are in line with the company’s past expenditures.

Feed Rate and Terminal Value The feed rate represents the ability of the company to generate or to license projects. The feed rates correspond in a way to the terminal value in standard valuation. Like terminal value, the feed rate is a very sensitive variable that can have a major impact on the value. A feed rate is however a more imaginable parameter than a constant cash flow in 20 years from now. Imagine a company that has one extraordinary product responsible for most of its revenues like in the figure below. If we apply the terminal value principle at a moment when this project still strongly contributes to the company’s revenues (A), this implies that we assume that the company can replace this product by another of the same economic power. We can only use this assumption when there is an indicator that the company can indeed replace the product with a similar one. If on the other hand we apply the terminal value only once the project is off-market (B), then most other projects’ life cycles might have ended as well and the terminal value becomes too low. The terminal value this way might misguide and give a much too high or low value. We therefore propose to use the feed rate approach. A B

Fig. 3.85. Where to start assuming constant cash flows for the terminal value

266

Valuation in Life Sciences

Taxes Taxes are to our great surprise a widely neglected topic in the literature about valuation. The reason might be the relatively simple setting of the problem: Taxes must be paid on a yearly basis as a percentage of the company’s benefit. This short and easy to understand definition gives rise to a widely used short-cut method: The value typically is made up by positive cash flows exceeding the expenditures. Taxes are levied exactly on this difference, which on a yearly basis is called benefit. We therefore can assume that the tax rate can simply be applied to the pre-tax value of the company. The following mathematical development for a profitable company visualises the rationale of this short-cut method:

V post _ tax =

(3.53)

∞

∑ (1 − tax ) ⋅ (Income − Expenditures ) ⋅ (1 + discount )

−i

i

i

=

i =1

∞

(1 − tax ) ⋅ ∑ (Incomei − Expendituresi ) ⋅ (1 + discount )−i = i =1

(1 − tax ) ⋅ V pre _ tax It seems as if taxes are not a big issue. We can calculate the company value without taxes and then simply reduce the value by the tax rate, as suggested by the mathematical development above. This corresponds to paying taxes on each cash flow. The company would pay taxes on all revenues, but would also get back something for its expenditures. But in the development above we have assumed that the company is profitable from the beginning. We know that for a biotech company this is usually not be the case. If a company reaches profitability only in a few years, it can accumulate the losses and set them off against the first benefits. On these first benefits it does not have to pay taxes. Tax payments only start once the accumulated losses are used up. If we would apply the short-cut method, this would mean that the company would receive some payments from the tax authorities for their expenditures until it has reached profitability. Of course, this is not true. Theoretically, these payments happen in the years when the first benefits arrive, i.e. when the accumulated losses are used up. But this “tax income” happens later than the expenditures provoking it. It must therefore be discounted for a longer period of time, reducing this income. The short-cut method therefore underestimates the value reducing effect of taxes. On the other hand, if the company has already accumulated losses from the past, these represent a positive value

Company and Stock Valuation

267

and are called tax assets. Typically, the tax assets are the tax equivalent of the accumulated losses, but discounted over the time period until they can be realised, i.e. when the company is expected to reach profitability. Unfortunately, the case of biotech companies is even more complicated, because a biotech company is not even certain to achieve profitability. Let us assume the simplest case of a one-project-company. The company will only reach profitability once the product is launched (assuming it commercialises the product on its own), and consequently only with the success rate of the project. In this case the short-cut method works even worse. As previously seen, the short-cut method reduces each cash flow by the tax rate. However, the R&D expenditures can only be offset against the sales revenues if these sales revenues effectively happen. This means that the “tax income” would have to be reduced by the full success rate of the project, and not only by the success rate of the source cash flow as in the short-cut method. This means, that again the short-cut method underestimates the value reducing effect of taxation. Of course, the accumulated losses, or tax assets, could be sold in case of failure in form of a valuable shell company, but due to administrative hurdles and transparency issues this is rarely done. But in general we can assume that the expenses can only be offset if the company becomes profitable, and that in all other cases the “tax income” stemming form these expenditures will be lost. This puts entrepreneurs in an unprivileged situation towards tax authorities. In case their companies become profitable they have to give a share of the benefits to the tax authorities. But as long as they are not profitable, they have to take care of themselves. The following example explains this unbalanced situation. Case Taxes For the purpose of illustration we assume a one-product med-tech company that still has to pass a pilot study and a subsequent pivotal study, each lasts one year. For the sake of simplicity we assume that all profits then incur in the third year. Once we assume no success rates at all, and once we assume that the pilot phase is only passed with a probability of 50% and the pivotal phase with a probability of 60%, giving an overall probability of 30%. The pilot phase costs US$ 3 mn, the pivotal phase US$ 15 mn, and the earnings in the commercialisation year are US$ 100 mn. If we assume a discount rate of 15%, a tax rate of 30%, and accumulated losses of US$ 4 mn then we get the results as outlined in Table 3.83.

268

Valuation in Life Sciences Table 3.83. Post-tax value with and without success rates w/o success rates Pre-tax value Pre-tax value*(1-tax) Delay of tax income Tax assets Not realised tax income Final value Value reduction through tax

US$ 59.6 mn US$ 41.7 mn (US$ 0.7 mn) US$ 0.9 mn US$ 41.9 mn 29.7%

with success rates US$ 13.2 mn US$ 9.2 mn (US$ 0.2 mn) US$ 0.3 mn (US$ 1.4 mn) US$ 7.9 mn 40.3%

We see that the value reducing effect of tax for an early-stage company can be important. Instead of 30% as in the case of a profitable company, the displayed company suffers from a value reduction of 40%. This value reduction is due to the asymmetric situation of tax authorities and investors. Tax authorities accept carrying forward losses, but investors must bear the capitalised costs of these losses, and the investors must bear the full attrition risk that can make their investments worthless. The value reducing effect of taxation becomes more important the lower the probability to become profitable for a company. Tax credit systems counteract this value reduction somewhat. While the calculation of an easy example as described above can consider every possible scenario, it becomes too complicated for a company with several projects. For each possible scenario we have to check if and when profitability is achieved. For a small company with 4 projects, each with 4 phases left before reaching commercialisation, this means that we need to analyse 44=256 scenarios. For the calculation of the post-tax value of a company we are therefore left with two alternatives. First, we can approximate the three above described corrections to the short-cut method by assuming an expected date of profitability for the company. This method is relatively fast once it is implemented, but remains an approximation. Second, the posttax value of a company can be simulated. Each simulation represents one possible scenario of the future, and we can exactly replicate the tax payments triggered in that particular scenario. This way we get the exact value for a scenario. Running a sufficient number such that each scenario is represented proportionally in the full sample of simulations, we can then calculate the post-tax value of the company as the mean of all simulations. As a drawback of this method we must mention that either we have to run a lot of

Company and Stock Valuation

269

computationally expensive simulations, or we have to accept a certain degree of randomness in our result. Until now we have assumed a simple tax system with one rate and losses being carried forward indefinitely. The tax rate can vary according to size and sort of income. Drug related income streams can be expected to fall within the highest tax progression range. In some countries tax authorities differentiate between normal and IP related income, i.e. license revenues. Typically IP related income is taxed at a lower rate. Furthermore, some countries offer a tax credit system, reducing the tax burden of R&D intensive businesses such as life sciences. The tax value of a company is far from being a simple issue; and in reality much more important, the value reducing effect of taxation is important. It is therefore difficult to understand why many companies attribute little priority to setting up tax saving structures by incurring their IP related revenues streams in countries with favourable tax systems. An optimal tax structure increases the value of the company considerably and is therefore of great interest to the shareholders of a company. Discount Rate It is not uncommon to use different discount rates to value project and companies. A company in-licensing a new compound might consider the project riskier due to information asymmetry than those developed inhouse and set the hurdle rate higher. But once we value a company as an entity, the projects are part of the portfolio. We discount the portfolio consequently at the company’s discount rate. As outlined in the section about discount rates economists have not yet agreed on what should be included, whether only non-diversifiable risks (CAPM), or also diversifiable risks (MCPM). The reality shows a picture that lies somewhere in between these two standpoints; the discount rate should reflect the risks that stem from the pipeline, the diversification the pipeline provides, and the maturity of the company. If we value the company as an entity we should use the WACC; in case we only value equity – which is more of the case – we use the cost of equity. In case the company has no debt there is no difference between WACC and cost of equity. Debt Equity belongs to the company’s shareholders, debt to its creditors. There are some very important differences between equity and debt. First, shareholders have a say in the company. A share gives them the right for one

270

Valuation in Life Sciences

vote in the shareholder assemblies. Creditors cannot influence the company’s strategy. Second, equity and debt have different payoff profiles. Debt is a fixed amount, that has to be paid back at maturity and earns interests until then. If the company is not able to pay back its debt, it goes bankrupt and the liquidation value belongs to the creditors. Debt has therefore a limited upside, i.e. the maximum when all interest and the full amount are paid to the creditors. This is the normal case. The payoff of the debt only worsens in case of bankruptcy of the company, when the debt value corresponds to what is left of the company. Equity is what shareholders receive, once all debt is paid back – the circumstance that creditors are served before shareholders is called seniority. The payoff profile of equity is symmetric. If the company performs well, then the shareholders fully profit from this upturn, while creditors only receive the fixed amount of the debt. However, if the value of the company drops dramatically, the creditors do not lower their claim, only the shareholder’s equity drops simultaneously in value. And if the company is not able to meet this requirement, the shareholders lose all its rights on the company. These different payoff profiles are illustrated in the figure below.

Equity

Debt Principal of Debt

Company Value

Fig. 3.86. Payoff profile of equity and debt

Only settled and stable companies are creditworthy and can raise debt. The cost of debt then depends on the amount of debt and the company’s rating, i.e. its probability to meet the debt requirements. Private companies often do not have access to the debt market and are consequently completely equity financed. The company value firm without debt then coincides with the value of equity. Debt has an influence on the risk profile of a company’s equity and therefore on its value.

Company and Stock Valuation

271

Value Upside Capital Protection

Share Price Share Price at Issuance

Conversion Price

Fig. 3.87. Payoff profile of a convertible bond

If we need to value the company’s equity, we have to deduct the liabilities from the value of the company. A major implication of debt is leverage. While the value of the company, i.e. the value of the assets is exposed to risks like changing market environment and attrition rates, the value of the debt is only seriously affected by these risks when the company faces bankruptcy. All risks are passed on the shareholders. But the value of equity is smaller than the value of the company, the difference being the debt. Therefore, shareholders bear relatively more risk than the whole company itself. This increased risk is offset by the possibility to achieve a return on investment with more capital, that is to say not only with equity, but equity plus debt. While the shareholders have to pay the creditors the agreed rate, they can keep the excess return for them. This is called leverage. Most young companies are not eligible for debt because they are not creditworthy. The risks stemming from the attrition rates are too high. Next to plain debt that includes interest payments and the paying back of the principal at maturity or in staged tranches, companies have the possibility to issue convertible debt. A convertible bond yields interest as a normal bond, but its holder has the right to convert the bond into a specified number of shares, if the share price passes a certain limit. This limit usually lies considerably above the share price at time of issuing, the difference being the conversion premium. A convertible bond therefore has some debt and some equity features. On one hand it offers capital protection; on the other hand it offers the bondholder an attractive upside potential. This upside potential reduces the yield the issuing company has to offer. The conversion premium offers the company to issue potential shares at a higher price than if it would raise equity. This is in the interest of existing shareholders.

272

Valuation in Life Sciences

When valuing equity in presence of convertible debt, we have to consider that the convertible bond might turn into shares. Consequently the income has to be distributed among more shareholders. The convertible bondholders potentially dilute the existing shareholders. Dilution (Employee Stock Option Plans, Ratchets) Dilution is not only linked to convertible bonds; also employee stock option plans (ESOP) or anti-dilution clauses (ratchets) of some investors provoke dilution. Normally the share price equals the value of equity divided by the number of shares outstanding. ESOP offer the employees of a company the right to buy shares at a fixed price. The shares will be issued at the time of exercise of the employees’ option. At the time of exercise the company therefore issues share at a too low price. The value of equity has is divided among the shareholders, but the participants of an ESOP did not have to contribute proportionally, the difference being justified with their value adding work for the company. The ESOP participants bargain is at the expense of existing shareholders. Unfortunately, at the time of valuation it is not clear yet whether the options will be exercised. In capital rounds investors sometimes insist on protecting its capital by adding anti-dilution clauses to the contract. These clauses are called ratchets and ensure the investor not to be diluted unless the share is valued higher than pre-defined limit. Ratchets can be justified by the information asymmetry between new and existing investors and serve as insurance for new investors. A proper consideration of the dilutive aspects of convertibles, ESOP, and ratchets requires sound knowledge of quantitative finance. We content ourselves with outlining the consequences. Company Valuation Manual

1. Define the building blocks of the company. 2. Define the parameters of each building block. 3. Value each building block (except taxes). 4. Combine the valuations of all building blocks. 5. Calculate the tax value. 6. Sum up the value of all building blocks.

Company and Stock Valuation

273

Interpretation of Valuation Results

Depending on the purpose of a firm valuation, we want to answer different questions. Most often, we want to learn about the company’s value, its risk profile, about the development of its cash balance and its value in the future. Value. The most important question of the valuation is how much value the company or its shares have. Risk profile. In addition to the value of the firm today, we might want to know how this value will behave in the future. We want to learn more about the up- and downside potential of the company. We also want to elaborate different scenarios in the future, e.g. what happens to a company when its lead compound in phase two fails? Simulations play an important role in visualising the risk governing the company value. We refer the reader to the section about simulations for more details. Cash balance. Due to the lengthy and costly process of drug and medical device development, a common question is how long the cash of a company will last, and if and when the company does need to raise further funds. It is again worthwhile to look at particular scenarios, as well as to worst case scenarios. Value development. A very interesting aspect is the value development of the company. From a theoretical standpoint the company value must grow on average at the assumed discount rate because of the following relationship: Vt =

1 (T − t ) Vt (T − t ) Ε[VT ] ⇔ Ε[VT ] = (1 + r ) (1 + r )

(3.54)

Actually, it does not grow exactly at the discounted rate, because not the entire value is assessed as a discounted expectation. Cash only grows at the interest rate offered by the bank. The expected growth rate is therefore a weighted average of the discount rate and the interest rate. The growth at the discount rate is owed to the progression in time. Other factors that influence the value development are: 1. Trial results. Either the uncertainty is resolved and we do not have to use the success rate anymore, or the project must be abandoned. By setting the success rate of a project phase to 100% or 0% we can quantify the value jump the trial results should provoke. In the figure below we can observe the value development of a phase II project

274

Valuation in Life Sciences

over time. We see that the value then grows steadily because the cash flows have to be discounted over a shorter period of time. At completion of a phase the value takes a jump, because the success rate does not have to be taken in account anymore. Overall the project will produce a net balance of US$ 874 mn. These final US$ 874 mn are reduced to a net present value of US$ 7 mn due to discounting and success rates. The graph also displays how much of the value is already realised at each point in time. Since the project initially requires investments the cash balance becomes negative and only the value of the project grows. But once the project is on the market, this value gets transformed into cash, ultimately in US$ 874 mn.

Fig. 3.88. Value development of a phase II project, calculated at a discount rate of 12% (calculated with ri:val)

2. Change in discount rate. The company might enter a new phase of its life cycle, e.g. by diversifying its pipeline. As a consequence, the discount rate falls. This immediately leads to a higher valuation of the company and therefore a higher share price. After this value correction the value then grows at the new discount rate. As we see in the figure below the same project calculated at only 7% produces a value of US$ 40 mn (compared to US$ 7 mn at 12%). The value growth of the project calculated at 7% however is more moderate. At the end the project still only generates US$ 874 mn, the set of assumptions has not changed. But since we start at a higher initial value (US$ 40 mn opposed to US$ 7 mn) the growth to the final US$ 874 mn must be slower. In fact, the value growths now at 7% instead of 12%.

Company and Stock Valuation

275

Fig. 3.89. Value development of a phase II project, calculated at a discount rate of 7% (calculated with ri:val)

Simulation on Company Level

As for portfolio management, simulations provide good insight into the risk profile of a company. They can visualise the company’s degree of diversification and chances of success. We then have to check whether these results are in line with the discount rate we have used. If a company is well diversified a high discount rate is hard to justify. Next to the advantages of considering correlations between projects, company level simulations offer the possibility to correctly account for tax effects. Since each scenario represents one possible version of reality; we just account for the taxes as the company would in that specific scenario. Finally, we average over all scenarios and get the value of the taxes. Case Company Valuation

Mr. Investor has the opportunity to take over a stake in BioComp in a trade sale. The company is specialised in treatments of the cardiovascular system. Currently it has a project in late clinical phase 2, another in clinical phase 1, and two in preclinical phase. On top of that, the management claims to be able to take one project every six months from its research labs into preclinical phase. A brief valuation of the compounds at a discount rate of 17% yields the following results.

276

Valuation in Life Sciences Table 3.84. BioComp project values Project

Phase

Value

BC-1130 BC-1140 BC-1200 BC-1202

Phase 2 Phase 1 Preclinical Preclinical

$ 25 mn $ 11 mn $ 5 mn $ 5 mn

BioComp’s research expenses in addition to the costs allocated to the projects amount to $ 7 mn a year; furthermore it has $ 1.5 mn expenses to maintain operations. The cash position amounts to $ 21 mn, tangible assets are negligible. The tax rate is expected to be 20%, but the CFO of BioComp thinks that tax credits in the near future might lower the effect even down to 16%. Mr. Investor wants to get an idea about BioComp’s value. Furthermore, the company plans an IPO if its lead product passes phase 2. The raised money will then be invested in phase 3 trials. Mr. Investor wants to find out about the potential value increase due to a successful phase 2 trial. Solution Case Company Valuation Mr. Investor has already valued the pipeline. In addition, he must quantify BioComp’s ability to generate new projects, the unallocated costs, and the tax effect. BioComp has a feed rate of 2 preclinical projects a year. Assuming the new projects to be as good as the current ones, Mr. Investor calculates the value of the prototype preclinical projects to be $ 5 mn. The cost of the research delivering these projects is $ 7 mn a year. This leads to the following value for BioComp’s future projects, i.e. for its technology: Vtech =

f ⋅ V project − Costs (1 + r ) r

=

2 ⋅ 5 − 8 ⋅ (1 + 17% ) = US $10.6mn 17%

(3.55)

In a similar way we calculate the value of the unallocated costs: Vunallocated =

∞

Costs

∑ (1 + r )

i

i =0

=

Costs (1 + r ) 1 ⋅ (1 + 17% ) = = US $ 6.9 mn r 17%

(3.56)

With this we have all values we need and we can proceed to the final assessment of the company value.

Company and Stock Valuation

277

Table 3.85. Value of company components BC-1130 BC-1140 BC-1200 BC-1202 Technology Unallocated Costs Total

$ 25 mn $ 11 mn $ 5 mn $ 5 mn $ 10.6 mn $ 6.9 mn $ 49.7 mn

From these $ 49.7 mn we can subtract 16% for the tax value, i.e. $ 8 mn. We then have to add the cash and end up with a company value of $ 62.7 mn. The average growth of the company value, assuming an interest rate of 5% for the cash is: Cash 21 V − Cash 62.7 − 21 rf = 5% = 13% r+ 17% + V 62.7 V 62.7

(3.57)

The more cash is invested, the more this average growth rate approaches the discount rate of 17%. 90 80 70

Company Value

60

Taxes Unallocated Costs

50

Cash

40

Technology

30

Projects

20 10 0 Positive

Negative

Total

Fig. 3.90. Value of the company

The success rate for phase 2 in cardiovascular diseases is 43%. Consequently, if BC-1130 passes clinical phase 2, its value will increase by a factor 1 43% = 2.3 from $ 25 mn to $ 58.1 mn. Taking account of the tax

278

Valuation in Life Sciences

rate this corresponds to a value increase of exactly $ 25 mn on company level, a jump by 80%. In addition, it is reasonable that the company might lower its cost of capital with a project in phase 3 and once on the stock exchange. This would then even increase the value further. On the other hand, a failure would result in loss of a $ 25 mn project. Stock Valuation

Stock valuation refers to the value of a public company, or more precisely of its equity. We follow the same rules like for company valuation explained above, with some minor differences. In contrast to private companies, the market value of public companies is known at every moment. The market capitalisation, i.e. the current stock price times the number of outstanding shares, stands for the market value of equity in the investor’s eyes. Contrary to the shares of private companies the access to information of public companies is limited. Due to insider trading regulations, we only have access to public data. On the other hand, if we look at a private company we are going to invest in, once we sign a confidentiality agreement we can make use of all sources of internal company information. If we do not have unrestricted access to company information, the valuation of outlicensed products poses a special challenge. As many early-stage and midstage biotech companies base their value on these licensed projects, we need to value them. In the case where we do not know the parameters of the license contract, we have to assume generic contracts. We then have to go back to the time when the contract was signed, and either base the deal terms on comparable transactions or we calculate a possible contract based on the value share principle, as outlined in the section on licensing. The main difference, however, is that we trade shares of public companies at ease. Investors can quickly react to new information. If we want to sell the shares of a private company, we have to find a counterparty buying them in a trade sale. This can take months and often hinders investors from reacting to news. At the same time the price of the shares does not have to be negotiated, but is given by the market. We even know the market capitalisation since the IPO. We see how the share price reacts to information and what has happened in the past. This will help us to get a feeling how the share might behave to future news. Indeed, before we can judge if the market capitalization is fair, we should get an impression what future events are already anticipated in the stock price. When valuing public stocks, we want to find out if shares are traded at the right price, if they are overpriced, or if we still have an opportunity to

Company and Stock Valuation

279

invest. Of course, if we assume perfect markets, we would expect that stocks be always traded at the right price. But as we do not have perfect information flow in the market, it is possible to find attractive investment opportunities. We will now discuss the influence of clinical trial results on stock price, and consequently, we will show how to predict possible future share price movements. There are situations where we see dramatic changes in the share price: news on successful or failed trials of lead compounds, the withdrawal of drugs, or successful completion of clinical phases. We illustrate such effects with the example of Actelion’s share price (ATLN, SWX). On April 20th 2001 Actelion has announced that the phase III clinical trial of Veletri, an intravenous dual endothelin receptor antagonist, did not meet its primary objective of significantly improving symptoms in the treatment of acute heart failure (AHF). On the 7th of November 2004 Actelion has announced the further development of Veletri was stopped. The news on the negative results from the Veletri trial triggered a share price plunge from CHF 145 to CHF 55, a decrease of 62% in a single day. At that time, Veletri was the lead compound of Actelion. The market potential, if it were to reach market, was enormous. The loss of the project therefore destroyed two third of the market value of Actelion. The announcement in 2004 to stop development of Veletri for futility caused a share price 250

200 Share Price CHF

Veletri fails in phase 3 trials (2001) 150

100

50

Veletri development stopped (2004) 0 Jan-00

Jan-01

Jan-02

Jan-03

Jan-04

Jan-05

Jan-06

Jan-07

Fig. 3.91. Share price of Actelion (ATLN) since its IPO in April 2000 (source: SWX)

280

Valuation in Life Sciences

drop form CHF 153 to CHF 115, a decrease of 24%. Meanwhile, the development of Actelion’s second lead compound, Tracleer, proved successful. At the time when Actelion halted the development of Veletri, the company value derived not mainly from this product, but from Tracleer. Furthermore Veletri’s value was not the same anymore as in 2001, because the first phase 3 trial has already unveiled some weaknesses, therefore investors have already lowered their expectations with respect to Veletri’s sales potential. The two reactions to bad clinical trial news show how dependent the share price on the clinical development news is. The case of Genentech (DNA, NYSE) displays that positive news can have a significant impact as well. On May 19th 2003 Genentech announced that a phase 3 trial of Avastin plus chemotherapy markedly extended survival of metastatic colorectal cancer patients. The share price skyrocketed by 45%. On March 14th 2004 an interim analysis of phase 3 trial has shown that Avastin plus chemotherapy improves progression-free survival in patients with first-line metastatic breast cancer and on April 14th 2004 an interim analysis of phase 3 trial has shown that Avastin plus chemotherapy extends survival of patients with first-line non-squamous, non-small cell lung cancer. The share price reacted with a 25% and 18% increase, respectively. Of course, we have chosen two prominent examples. 120

Share Price in US$

100 80

Metastatic breast cancer phase 3 results

60 40

Metastatic colorectal cancer phase 3 results

Non-small cell Lung Cancer phase 3 results

20 0 Jan-00

Jan-01

Jan-02

Jan-03

Jan-04

Jan-05

Jan-06

Fig. 3.92. Genentech share price reactions to clinical trial results

Company and Stock Valuation

281

These large movements of the share price are interesting, if they can be anticipated, but they involve a fair amount of risk. Due to the information constraints, it is normally not possible to have enough information to predict a positive or negative outcome of the clinical trials. Above we have seen how to calculate the company value today. If we consider investing in a stock, the calculation of the present value of the company based on the probability adjusted and discounted cash flows helps to judge if the share is priced right. But we often not only want to know the current market capitalisation and the respective share price today; we want to know how the stock will evolve in the future, we want to anticipate the share price’s reaction to expected clinical news flow just as seen in the examples about Actelion and Genentech. We know that on average the share price develops at the discount rate, provided our estimates are correct. But in reality the average scenario never happens. There is no trial passed with 60%; a trial is either a success or a failure. The trial results have an immediate impact on the value of the concerned project. Either the trial is passed and we do not have to multiply anymore by the success rate for this trial, or the compound failed and the project value drops to zero. The average scenario, i.e. multiplying the project value with the success rate, lies somewhere in between success and failure. The value of a stock therefore does not follow the average scenario but one specific outcome. But since we cannot predict which outcome will happen, we have to content ourselves with an expectation, i.e. an average value. As soon as we know more about the scenario that is actually going to happen, we adjust the valuation to this new knowledge. We can elaborate the movement of the stock based on a selection of scenarios, ranging from bad to best-case outcome. To do so we need to value all projects as if they were already one year, two years, and more advanced in their development. This implies that they would also have passed all phases successfully in that period. The projects naturally gain in value with proceeding time, first because earnings come closer and have to be less discounted, second because some investments are sunk and do not have to be considered anymore, and third and most importantly, because of positive trial results. When investigating possible scenarios of a company’s value development, we simply have to determine which projects will still be alive at the moments we want to analyse. The company value must then be calculated with these projects’ value corresponding to that time. And while the project value usually increases with time, the cash balance decreases. We must also keep track of the cash balance and reduce it by the amount spent, corresponding to the pipeline state. Note that once a project starts earning money, e.g. through

282

Valuation in Life Sciences

revenues or milestones, the project value then decreases but conversely the cash balance increases again. In these cases we then have to consider the milestones or revenues as sunk. We will now make this exercise with an early stage biotech company, having the following pipeline today. Table 3.86. Present pipeline of an early stage biotech company Project

Preclinical Phase 1

Phase 2

Phase 3

NDA

Market

1 2 3 4 5 6

We consider investing in the stock and want to know what we can expect in terms of share price development. Table 3.87. Project characteristics Project Mode of conduct 1 2 3 4 5 6

Self-conducted Self-conducted Out-licensed in phase 3 Out-licensed in phase 3 Self-conducted Out-licensed in phase 3

Peak sales $ 575 mn $ 653 mn $ 1,790 mn $ 250 mn $ 1,102 mn $ 375 mn

We consider the following scenarios: 1. Project 1&2 reach commercialisation, all other fail in phase 2. 2. Project 3 reaches commercialisation, all others fail in phase 2. 3. All projects reach commercialisation.

283

20 06 20 08 20 10 20 12 20 14 20 16 20 18 20 20 20 22 20 24 20 26 20 28 20 30 20 32 20 34 20 36 20 38

Company and Stock Valuation

revenues

expenses

taxes

average growth

company value

Fig. 3.93. Project 1 and 2 reach commercialisation

revenues

expenses

average growth

Fig. 3.94. Project 3 reaches commercialisation

8

6

20 3

4

20 3

2

20 3

0

20 3

8

20 3

6

20 2

4

taxes

20 2

2

20 2

0

20 2

8

20 2

6

20 1

4

20 1

2

20 1

0

20 1

8

20 1

20 0

20 0

6

If only project 1 and 2 reach commercialisation, and all other compounds fail in phase 2, the share price will first decline and be smaller than the average growth, i.e. the growth at the discount rate, in our case 16%. In this scenario, we would only profit after 4 years from a solid share price increase.

company value

284

Valuation in Life Sciences

20 06 20 08 20 10 20 12 20 14 20 16 20 18 20 20 20 22 20 24 20 26 20 28 20 30 20 32 20 34 20 36 20 38

In the next scenario, only project 3 passes clinical phase 2. Again all other projects fail in this phase. This bad case scenario does not allow the stock to grow significantly. The increase after four years is due to positive phase 3 results of project 3. Nevertheless, the value drop is not dramatic because, as we can see in the figure, the company has no expenses anymore. In the third scenario, all projects come to market. The share price rapidly grows at a rate much higher than the average growth of 16%. This, of course is a very optimistic scenario, but gives us a feeling what happens in a good case. Of course, for a thorough study of the share, we would not only look at three, but at many more scenarios. Nevertheless, we get a good impression about the possible development of the stock, in an average, a bad and a good case. We can also undertake such an analysis for a private company. But since it is not listed on a stock exchange we cannot immediately profit from an increase in value.

revenues

expenses

taxes

average growth

company value

Fig. 3.95. All projects successfully pass phase 2

Advanced Company Valuation

In the following chapter we will take a much more detailed look at company valuation and discuss important issues inherent to it. We will look at the company valuation from the outside investors’ point of view and from

Company and Stock Valuation

285

the inside, namely the CFO/CEO’s point of view, who represent the already invested shareholders. Aspects that are discussed are: • Value components • Value development • Feed rate, pipeline structure and terminal value • Risk profile and diversification • Cash balance • Tax • Scenario analysis o Cash needs o Value development o Dilution

• Sensitivity analysis o Tax o Unallocated costs

Case Study: Advanced Company Valuation Bio-Tec is a mid stage biotech company with four compounds in development. Bill Dollar, the CFO, prepares the next board meeting. He will present the pillars of the strategy the company might follow in the next months and their financial impact. The board will have to decide on the following issues: • Out-licensing or self-conducting the lead compound. • Cash needs and ways to cash. • Fallback scenarios if lead compound fails. • In-licensing opportunities. • Tax optimisation. In order to better present the different strategic options, Bill prepares a thorough company valuation. He gathers all the information necessary and comes up to the following input parameters:

286

Valuation in Life Sciences Table 3.88. Company parameters Cash US$ 10 mn Tax rate 20% Accumulated losses US$ 7 mn Unallocated costs US$ 1 mn per year Discount rate 17% Table 3.89. Project characteristics

Project

Mode of conduct

BT – 1045 BT – 1201 BT – 1202 BT – 1038

Out-licensed in discovery Self-conducted Self-conducted To be out-licensed in phase 3

Phase

Peak sales Disease Group

Clinical I $ 1,000 mn Preclinical $ 700 mn Preclinical $ 380 mn Clinical II $ 631 mn

Cancer CNS CNS Cancer

The company has a technology platform, which generates new projects. For the company valuation, Bill has elaborated that the basic research team generates each year one project that enters preclinical phase each year. The costs for maintaining the technology are US$ 450,000 per year and the technology will be protected, i.e. can be exploited for another 10 years. Bill now has all the data together and performs a valuation using ri:val and comes up with the following results: Table 3.90. Company valuation Value US$ mn BT – 1045 BT – 1201 BT – 1202 BT – 1038 Technology Cash Bio Tec costs Tech costs Taxes Company

DCF 12.8 7.0 (3.9) 52.3 20.0 10.0 (4.6) (2.1) (46.0)

Real Options 13.1 7 0.4 52.4 20.6 10.0 (4.6) (2.1) (47.2)

45.5

49.6

Company and Stock Valuation

287

After valuing all the parts of the company, Bill wants to know how the value of the company will develop over the next years. The graph below represents the average value development of the company until 2050. On average, BioTec will require US$ 140 mn more in the next 10 years. But out of these US$ 140 mn it manages to generate US$ 4 bn cash until the years 2040.

Fig. 3.96. Average value development of Bio-Tec (calculated with ri:val)

Bill uses simulations to check how the pipeline will evolve in future years on average. He understands that with the feed rate of one project in preclinical phase per year and with the current pipeline, Bio-Tec should manage to get one project on the market.

Fig. 3.97. Simulated pipeline structure (calculated with ri:val)

Once the feed rate of the technology drops to 0, the pipeline is not anymore sustained and the company should replace the existing platform.

288

Valuation in Life Sciences

In the simulations, Bill also has a chance to see the value distribution and the risk profile of the company using the present strategy.

Fig. 3.98. Simulated value of Bio-Tec (calculated with ri:val)

Bill wants to prepare for possible questions during the board meeting and therefore values the company using different assumptions for the strategy. In a first scenario, he supposes that Bio-Tec will not out-license the lead compound BT – 1038, but will conduct it on its own. He is especially interested in understanding the effect on the value and the value development of the company, on the cash needs, and on the changed risk profile. He assumes that if Bio-Tec launches the drug, the sales and the launch costs will only amount to half of what the license partner BigPharma would spend/earn. With these assumptions he comes up to the following values: Table 3.91. Company value self conducted Value US$ mn

DCF

Real Options

Company BT – 1038

25.05 24.7

29.15 24.7

The new valuation does not look favourable to Bill, as the company value is almost reduced by half. But not only the value is now lower, the cash needs are higher as well. Over the next 10 years the company would have to raise US$ 165 mn to finance the development on averaget, opposed to US$ 140 mn in case BT-1038 is out-licensed. Nevertheless Bill Dollar must consider these cash requirements with caution, because they are based on an average scenario. A CFO usually has to prepare his company for reality and not for the average scenario; and in reality costs have to be

Company and Stock Valuation

289

paid in full or not at all, but not just the success rate of it.We will see that the scenario analysis proves very useful for forecasting future cash needs and budgeting. The risk profile has also changed with the modified strategy for BT-1038. The likelihood of a negative company value increases from 55% to 64% while the chances of a higher company value increase only marginally. This also explains the lower value of the self-conducting strategy.

Fig. 3.99. Simulated value of Bio-Tec (BT-1038 selfconducted), calculated with ri:val)

Bill has also been instructed by the board to check if an optimisation of the tax situation of the company is possible and which effect it would have on the value. Bill has found that by adapting the corporate structure by transferring some IP to a subsidiary abroad the tax rate could be as low as 15%. He then performed a valuation to see if this has any effect on the company value today. Surprisingly, he came up with the following figures: Table 4.92. Company value at 15% tax Value US$ mn

DCF

Real Options

Company

57.1

61.4

The negative value of the future cash payments reduces the value by US$ 35 mn today if Bill calculates with 15% tax, whereas it reduces the value by US$ 48 mn when calculated at a 20% tax rate. For the next fundraising round, Bill also wants to give new investors a feeling of where the company value could head. He looks at predefined scenarios, which he thinks are a good, an average, and a bad case. The good case is that BT-1038

290

Valuation in Life Sciences

and BT-1201 reach market and all other projects fail in phase II (which is a conservative assumption, because the projects could also fail earlier, which would save costs). In the average scenario BT-1201 reaches market and all other projects fail in phase II. In the bad case only a project out of the technology succeeds, and all others fail in phase II. The values in the year 2013 for the three scenarios are: Table 3.93. Company development up to 2013 Value US$ mn

DCF

BT-1038 and BT-1201 succeed BT-1201 succeeds CNS Tech 1 succeeds

1,171 198 134

Cash needs 115 192 180

Multiple 26.0x 4.3x 2.9x

The first scenario is not only enormously attractive because of the high value multiple, but also because of the relatively moderate additional cash needs compared to the other two scenarios. The cash needs obviously lead to a dilution of the existing shareholders. Hence, the existing shareholders cannot expect to reach the displayed value multiples, but only a fraction of it, according to the dilution. The degree of dilution can only be quantified of we know at which point in time the company raises how much money. The second scenario requires interestingly more cash, although less products are conducted. But the license revenues from BT-1038 are now missing and cannot cross-fund some of the R&D expenses of Bio-Tec. The third scenario is worse in terms of value development, because the successful project is in an earlier stage than in scenario 2. But this also means on the other hand that the expensive trials happen later, therefore the cash needs are slightly lower. Bill sees that the scenario where the lead compound reaches market is highly attractive to investors, with a multiple of 26 until 2013. Of course, the scenario that the lead compound fails would be a highly unfavourable situation for the company, but nevertheless, the average scenario still promises a multiple of 4 in the same time. Mergers and Acquisitions

In the pharma and biotech industry we often see merger and acquisition transactions (M&A). Although license deal actually also fall in the category of asset acquisition we will focus in this section only on deals involv-

Company and Stock Valuation

291

ing whole companies. The reason for this lies in the brisk demand of new projects to feed big pharma’s pipelines. Often executives also identify potential synergies when they merge two companies. As a merged company some costs do not have to be spent for two units anymore. In the merger of Sanofi-Synthélabo with Aventis in 2004 these synergies have been estimated at EUR 0.5 bn each year, starting two years after the transaction. 7 June 2006: The Boards of Novartis AG (“Novartis”), Novartis Pharma AG (“Novartis Pharma”) and NeuTec Pharma plc (“NeuTec”) announce that they have reached agreement on the terms of a recommended cash offer (the “Offer”) to be made by Lehman Brothers on behalf of Novartis Pharma, a wholly owned subsidiary of Novartis, for the entire issued and to be issued ordinary share capital of NeuTec. The Offer will be £10.50 in cash for each NeuTec Share. The Offer values the existing issued share capital of NeuTec at approximately £305.1 million. The Offer represents a premium of 108.9 per cent to the closing mid-market price of NeuTec Shares of 502.75 pence per share on 5 June 2006 (the last dealing day before NeuTec entered an offer period). (…) Large pharma mergers usually do not provide the shareholders with a big premium. In contrast to this, take-overs of biotech companies by pharmaceutical companies make big headlines with premiums of over 50% to the traded share price. The displayed example of Novartis’ acquisition of Neutec in 2006 exhibited even a 108.8% premium for Neutec’s shareholders. Other examples of large take-over premiums are AstraZeneca’s acquisition of CAT and Medimmune or Merck’s acquisition of Sirna. The M&A activity in the drug development industry has grown that intense that meanwhile private companies do not necessary have to float their company on a public stock exchange via an IPO anymore. Earlier a trade sale has been perceived as a lucky punch offering a biotech company’s investors an attractive way to realise the value of their investments. Meanwhile many investors estimate a trade sale as a plausible exit scenario worthwhile working towards. But how can these huge premiums be justified? Why is a pharmaceutical company ready to pay the double of the price investors attribute to the company, either in a fund raising round or on the public stock exchange? After all, a pharmaceutical company does want to add value with their

292

Valuation in Life Sciences

M&A transactions. But how can they create value if they apparently grossly overpay their acquisitions? The answer to this question is threefold. First, once the target is incorporated in the acquirer, it should also be valued as a part of it. The pharmaceutical company can therefore apply its own cost of capital, which usually lies way below the cost of capital of the target. This value difference alone already justifies a large premium. Second, a pharmaceutical company is mostly more experienced in clinical trials, regulatory affairs, and commercialisation. The value of the asset is worth more in the hands of a pharmaceutical company than in the hands of a biotech. Third, after the incorporation in the acquirer’s organisation, the joint corporate has some synergies. These add to the value of the target as well. These three points added up can justify in most cases a large premium. But the pharmaceutical company, of course, does not want to give all the added value to the shareholders of the target. With the transaction, the pharmaceutical company locks in some value, because it pays less for the assets than they are worth in its hands. This value, however, must first be realised. At the same time also the shareholders of the target company belong to the winners, because they earn a nice premium to the current value of their shares. 160

140 Costs of transaction

120

Premium

US$ mn

100

Value of deal to acquirer Synergies

80

Value added in pharma's hands Difference due to discount at 12%

60

Company at 17% 40

20

0 Value

Value to acquirer

Joint value to acquirer

Net value to acquirer

Price paid

Fig. 3.100. Value composition of a M&A transaction

Company and Stock Valuation

293

Other, Less Serious Approaches

Corporate valuation is not an exact science. While we can observe the price of a stock, we cannot observe the value of a company. It is a subjective measure, depending on the perception of risk and potential in the eyes of the investors. This opens the door to many different valuation approaches. To make it even more complex, we can account for probabilities of drugs to reach a certain development stage, i.e. the success rates. Understanding probabilities is a difficult task for the human mind and we will see in the following that some minds already stumble over much easier concepts. We show some examples – sometimes funny, sometimes even frightening – how valuation should not be performed. We will discuss the different methods with an example company that has only one phase II project with the following parameters: Table 3.94. Assumptions Assumptions Phase 2 Phase 3 Review Sales Company

2 years, US$ 10 mn, 40% 3 years, US$ 50 mn, 60% 1 year, US$ 4 mn, 90% 300 peak sales, 65% op margin 25% tax rate, 50 employees

The standard rNPV valuation at a discount rate of 17% yields a pre-tax value of US$ 39 mn, and a post-tax value of US$ 25 mn. “We use the risk-free rate for discounting in real options (AR 2004, Astra Zeneca)” The concept of risk-free discounting has been adopted in many publications about real options. The origin of this idea is always the similarity between real and financial options. Financial options are financial instruments, with an underlying that is traded and liquid. Consequently it is possible to design a portfolio with the financial option, the underlying asset, and a bond that allows to hedge away all risk (in theory). Therefore we can apply risk-neutral discounting. In drug development this is clearly not the case. The underlying of the option is the potential future sales of the drug. There is no instrument in the market that allows efficient hedging of

294

Valuation in Life Sciences

these sales, which are uncertain in time, size, and event. Companies using real options and risk-free discounting thus run the risk of significantly overvaluing their projects. Our example company would have a pre-tax value of about US$ 187 mn if we use 5% risk free discounting. “If you use success rates, then you have considered already the attrition risk and you should discount with a pharma discount rate (CFO of a listed biotech company)” This would mean that the discount rate should be the same for pharma and biotech. The argument is that both pharma and biotech companies depend on the same underlying market, drug sales. However, there are two main differences between pharma and biotech. The value of a pharma company is mainly made up by products that are already on the market. Biotech companies have their projects still in R&D and their sales potential is not yet observable. The sales projections for biotech companies with their projects in development are highly uncertain, regardless of the success rates, which have to be accounted for on top of that. The second difference is that a pharma company’s value is driven by the next few years’ benefit. These years are certainly easier to estimate than the value relevant years for a biotech company, which lie 15 to 20 years in the future. And if the biotech company wants to get there it has to finance the same period of loss-making R&D. This increased estimation uncertainty and the implicit leverage of mandatory R&D expenditures before reaching profitability lead to a further increase of the discount rate for biotech companies. Finally, we can also observe a general risk-aversion of investors. Investing in a biotech company is like betting on the company’s lead compound, or on its technology. Investing in a pharma company corresponds rather to placing bets on various products. Imagine playing with a biased coin (15% chance for head, 90% chance for tail). If it’s head you win $1,000, otherwise you don’t win anything. How much would you pay to participate? The fair price would be $150, but typically you would say that in most of the cases you will loose anyway, so better don’t put too much at stake. And now compare this game with the following: you play 100 times head or tails with the biased coin, and fro each head you get $10. Again, if all coin flips show head you win $1,000. But as shown in the graph the chance of loosing everything is much lower. The forecastability of the payoff has increased tremendously. While for the first game we were sure to be off from the actual payoff ($0 or $1,000) by a large amount, in the second game we are pretty sure that the final payoff will be somewhere between $100 and $200. In a simulation of 1,000 scenarios (cf. Fig. 3.101)

Company and Stock Valuation

295

the payoff never fell out of the range of $60 to $290. This is quite a difference to the range of the first game, which is $0 to $1,000. As a consequence, in the second game we are ready to pay a price that is closer to the fair price of $150. The first game corresponds to investing in biotech, where you place a bet on one project. The second game corresponds to investing in pharma, where you are not so much dependent of the fate of one single project. 900 800 700 600

num ber of sim ulations

500 400

biotech

300

big pharma

200 100 0 0

100

200

300

400

500

600

700

800

900

1,000

payoff

Fig. 3.101. Simulation of 1,000 biased coin flips

The lack of diversification in a biotech company is usually punished by a higher discount rate. The success rate does not account for the risk aversion. The value of our example project becomes US$ 112 mn if discounted at 9%. “Terminal value is 125% of value (Copeland)” Tom Copeland is not the only author who claims that in R&D intensive companies most of the company value lies in the terminal value, i.e. that the main chunk of value stems from the time period beyond the estimated cash flows. Usually it is assumed that from then onwards the company is able to maintain stable annual cash flows. These stable cash flows are then summarised in the so-called terminal value. If this terminal value is more than 100% this means that the company would be an unprofitable venture if we don’t include the cash flows of the time period we do not analyse in detail. Probably nobody who invests his own money would like to rely on such assumptions. In life sciences terminal value has also another short-

296

Valuation in Life Sciences

coming: terminal value assumes a fair degree of stability of the company. But drug development projects are subject to a large amount of attrition and it is very difficult to estimate when a company will have reached a stable state. A stable state corresponds to a healthy pipeline that is able to compensate for attrition. Will the company really be already that far by then? Even pipelines of big pharma companies do not have this property. Is, e.g., Pfizer’s pipeline strong enough to compensate for the loss of patent protection of Lipitor? We propose an alternative methodology to capture the value of the company’s ongoing business: Add every year new projects to the pipeline. These new projects, more or less standard projects, come from business development (in-licensed projects) or from the labs. We can estimate this feed rate by looking at the past performance of business development and research of the company. This way the rather delicate estimation of the moment when the company reaches a stable state is not an input but rather an output. If we set the time horizon at ten years then the terminal value is US$ 50 mn and the overall project value US$ 39 mn. A surprising match with our initial rNPV valuation, and also surprisingly close to Copeland’s figure (here the terminal value is 129% of the value). However, if we push the time horizon to 12 years then the company value already becomes US$ 49 mn. But the value of the company should not depend on our time horizon. “Value=DCF without success rates (because compounds are not comparable) but discounted at 50%” We often see this valuation practice being used for early stage companies. A Harvard business school article by William Sahlman (“A method for valuing high-risk, long-term investments”, 2003) gave this method the apparent academic justification. The discount rate is decomposed into various components that account for risk, liquidity, added value by the investor, and cash flow adjustment because of uncertain business outcome. While we completely agree with the first two components, i.e. risk premium and liquidity adjustment, we object to the other premiums. The argument for the value added premium goes as follows: The investor adds his experience to improve the performance of the company. This premium can be avoided by estimating the cash flows for the pre-money valuation (which is also pre-investor, i.e. without his valuable input) without the value adding effect of the investor and his network. This is less arbitrary than a premium to the discount rate. The cash flow adjustment should ac-

Company and Stock Valuation

297

count for the attrition. We think it makes much more sense to use success rates instead of a higher discount rate. Success rates are available for any disease area, for any compound type, and for each phase. Success rates allow differentiating between safer and riskier therapeutic areas. A premium to the discount rate lacks a clear quantification. How much should this premium be for a preclinical cancer compound compared to a phase II Alzheimer drug? Finally, cash flows after launch of eth drug have the same probability. It is unclear why nevertheless with every additional year the cash flows should be worth about 50% less. For the sake of transparency and credibility of the valuations we strongly argue to quantify (finally valuation is a quantification exercise) the impact of each effect. But these effects should rather have an impact on the cash flows than on the discount rate whose effect on the overall valuation is difficult to judge. Our example company now has a value of US$ -2.6 mn. In the continuation we exhibit a few even more adventurous valuation methods. Comparables A so-called valuation specialist tried to find the value of a company by comparing it to companies whose price is known (e.g. the market capitalisation) based on the number of PhDs: Companies are comparable if the number of PhDs is about the same. Unfortunately not each technology requires the same amount of research, and not each researcher adds the same value. Although completely meaningless, the number of PhDs seems to be a well observable parameter that our colleagues happily used, because more value-relevant parameters were hard to get. We could also cite other meaningless parameters like P&L, amount of burnt capital, etc. Comparables are a well-accepted method for price finding, but it should use criteria that are indeed value-relevant. Such criteria can be stage of lead compound, same therapeutic area, similar technology, or same business model. “Value=#employees * US$ 1 mn (business development of a private bio-tech company)” This valuation method falls into the same category like the one just above. If it were that easy we would recommend that you hire a few 1,000 employees. This way you boost your value by US$ 1 bn. Our example company would be worth US$ 50 mn.

298

Valuation in Life Sciences

“Post-tax value=(1–tax) * pre-tax value (a European investment bank)” This tax consideration is correct for companies that are profitable (and remain so). But for companies, which are not yet profitable the effect is actually worse, because they can only realise their tax assets once they reach profitability, without interests. And for companies that become profitable only in some cases (like biotech) the value destroying effect of taxes is even more devastating. In the example the post-tax value is US$ 25 mn, only 64% of the pre-tax value. The described method assumes a post-tax value of 75% (=100%-25%) of the pre-tax value, i.e. US$ 29 mn. “Value=invested capital compounded at target IRR (‘valuation specialist’)” This method puts a lot of trust into the management of the company. It assumes that the management could increase the company value at the target IRR since inception regardless of what actually happened. It is clear that the valuation depends on the current situation of the company and its future outlook; the past is irrelevant for the company value. “Value=DCF without success rate * success rate (analyst)” This analyst valued each project without success rates but then felt that he had to adjust for the attrition rate of the projects. Consequently he simply reduced the value by the success rate of the project. This risk-adjustment is wrong. In biotech the value reducing investments are unfortunately more Table 3.95. Assumptions Method rNPV pre-tax rNPV post-tax Real options at risk-free With pharma discount Terminal value With 50% discount 1FTE=US$ 1 mn Pre-tax*(1-tax) Risk-adjustment

Value (US$ mn) 39 25 187 112 39-49 -2.5 50 29 52

Company and Stock Valuation

299

likely than the value adding positive revenues from sales. This means that the value suffers more than just by the success rate if you correctly adjust the cash flows for their risk. The example company would have a value without success rates of US$ 240 mn. Multiplying this with the success rates (*40%*60%*90%) this would lead to a value of US$ 52 mn. To conclude this series of remarkable valuation methods we describe the dynamic P/E ratio. Dynamic P/E Ratio In this article we focus on analyst’s reports on public biotech companies that have not yet reached profitability. These companies often have a market capitalisation below US$ 500 mn and are therefore under the radar screen of most analysts. Nevertheless, some analysts accept the challenge to value unprofitable public biotech companies. Reviewing some of their reports, we have discovered some original valuation models that merit a short discussion. We will review in this article the P/E ratio as a valuation method for non-profitable biotech companies. Analysts usually value profitable companies. A good and fast method to see how expensive a share is is the price-earning ratio (P/E ratio). If we assume a company with stable earnings, we assume that its value would be as in the formula below. ∞

V = ∑ (1+ r ) E = t=1

−t

E r 14,

Fig. 3.102. r = discount rate, E = earnings.

If we assume that the market capitalisation of a company, i.e. its price, is the value the market attributes to this company, then the P/E ratio corresponds more or less to the inverse of the discount rate. If the ratio is high, this means that “the market uses a low discount rate”, i.e. the investment is perceived as rather safe. A low P/E ratio points at a high discount rate, i.e. a risky investment. If an investor thinks that the market uses a wrong discount rate, then he can either buy or sell the company’s shares.

14

∞

With a growth factor the equation becomes V = ∑ (1 + r − g ) E = t =1

−t

E . r−g

300

Valuation in Life Sciences

The P/E ratio is so simple and easy to understand that it is widely popular metric amongst analysts. Nevertheless, the P/E ratio has two dramatic disadvantages. It requires the company to be profitable and stable. Biotech companies are mostly not yet profitable and only at the beginning of their planned life cycle. So, the P/E ratio is a bad metric to use for biotech companies, you might think. Have a look at the following valuation of Clinuvel, an Australian company, performed by an analyst at Louis Capital Markets. Clinuvel has one product in phase III trials.

Probability Weighted Sales

Probability

Potential Peak Sales (US$ mn)

Table 3.96. Sales analysis of Clinuvel

Phase 3 PLE EPP

40 25

50% 50%

20 12.5

Phase 2 SCC/AK SU Other TOTAL

240 12 200 517

30% 30% 15%

72 3.6 30 138.1

Clinuvel develops the project for various indications. The analyst estimates for each indication the peak sales and the likelihood that Clinuvel receives marketing approval for this indication. The overall probability adjusted peak sales are $ 138.1 mn. As a next step the analyst calculates the net present value of these probability-adjusted sales, assuming that they only occur in 5 years. In order to get to risk adjusted earnings (=Net in table 3.97) he must reduce the NPV of sales by cost of goods, sales, general and administrative expenses, and taxes. The final $ 30.3 mn represent the risk adjusted net present value of Clinuvel’s earnings in 5 years. On a per share basis this corresponds to (risk adjusted and discounted) earnings per share (EPS) of $ 0.1. At a share price of A$ 1.07 this corresponds to a P/E ratio of 9.1 (including the exchange rate

Company and Stock Valuation

301

Table 3.97. Sales analysis of Clinuvel Parameter

rNPV

At 15% discount 15% ~25%

68.7

Sales COGS SG&A EBIT Tax Net

35%

–10.3 –17.7 46.615 –16.3 30.3

A$/US$). The analyst then compares this to the Industry P/E (BBG World Biotech Index) of 45.05 and concludes that the share is a bargain. This method, the so-called dynamic P/E ratio, is much wider spread than one would believe16. Some analysts do not even use risk-adjusted sales but simply use a much higher discount rate (35%) and factor in the attrition risk this way. But then they all use an industry P/E ratio. These industry P/E ratios differ quite dramatically. Another analyst from RRS, also valuing Clinuvel, uses a 25 P/E ratio. He mentions, that profitable biotech companies trade at 25 to 40 times their earnings. Other analysts use a multiple of 30. 1. Determine representative year regarding revenues (5 to 10 years in the future): T. 2. Assess revenues for that year: RT 3. Discount and probability-adjust earnings: Rt=p*RT*(1+r)–(T–t) 4. Calculate EBIT: EBIT=Rt*(1-COGS-M&S) 5. Calculate earnings: Et=EBIT*(1–tax) 6. Multiply with selected P/E ratio: P=Et*P/E Fig. 3.103. Dynamic P/E ratio

15 16

It is unclear why this is not 40.7. Dr. Christa Bähr, Dr. Markus Manns: „Life Science am Kapitalmarkt, Biotechnologie im Fokus“, publication of DVFA, 2005.

302

Valuation in Life Sciences

P/E Ratio in More Detail Established drug development companies typically exhibit a P/E ratio of about 10-15. Table 3.98. P/E ratios and corresponding discount rates Company Amgen BMS Lilly Merck Pfizer

P/E ratio

Discount rates

9.73 14.28 13.29 11.96 9.29

10.3% 7.0% 7.5% 8.4% 10.8%

We can assume that these companies have reached a stable stage of their business cycle. Therefore we can apply the P/E idea and deduct their discount rates, with some caution. A 30x multiple (read price = 30 x earnings) would therefore correspond to a discount rate of 3.3%. Of course, this does not correspond at all to biotech companies that usually exhibit much higher discount rates than pharma companies.

Price Earnings P/E ratio

Time Break-even

Fig. 3.104. P/E ratio along the business cycle of a company

Figure 3.104 exhibits first, that unprofitable companies do not have any P/E ratio at all (due to lack of earnings). Second, companies that are only at the beginning of profitability have naturally a high P/E ratio, because the

Company and Stock Valuation

303

price is calculated using the future and higher earnings, but are divided by the current and lower earnings. A high P/E ratio is therefore not only a sign of an established and diversified business as is the case for pharma companies, but also for a business that is still in growth. The described risk-adjusted discounted P/E method has some serious flaws. First, it does not account for all the R&D expenses it takes to get to a profitable stage. Second, it is completely unclear that the selected year represents already a stable state of business. Calculating with the expected P/E ratio in 5 years corresponds to taking the terminal value already after 5 years. This is quite a short period of time in an industry where it takes 15 years to commercialise a product that should then stay on the market for another 10 years. Third, this method is unsuited to correctly account for the tax value of the company. Fourth and most important of all, it makes no sense to compare with an industry P/E ratio. The P/E ratio depends on the stage of the company and on the discount rate. Both are particular to the company and the selected year. We can only speculate why analysts use this method. One advantage is that the method is relatively insensitive to the discount rate, one of the most critical parameters. On the other hand the method supposes that the company is comparable to the industry average and uses an at best questionable P/E ratio. It is careless to hope that all the flaws of the method will compensate each other. With risk-adjusted net present value (rNPV) analysts have a method at hand that accounts for all particularities of the company. Fortunately, most analysts make use of rNPV, although the described dynamic P/E method is widely spread.

Exercises

Introduction In this chapter, the reader has the opportunity to apply what he has learned earlier in the book. Step by step, we will guide the reader through project and license valuation, and repeat important details. We then also discuss more advanced topics by way of exercise. The solutions serve as a blue print and comprehension questions help the reader verifying, whether he has properly understood the topics.

Exercises Project EXERCISE In the following exercises, we will use the parameters of the following project that we call EXERCISE: Table 4.1. Parameters of project EXERCISE In $ mn Length Cost Success Rate Peak Sales Margin Growth Rate Volatility Launch Costs Discount

Phase 1 Phase 2 Phase 3

NDA

1 year 2 years 2 years 4 15 45 66% 39% 62%

1.5 years 2 75%

230 55% 1% 25% 100 15%

B. Bogdan, R. Villiger, Valuation in Life Sciences, 3rd ed., DOI 10.1007/978-3-642-10820-4_4, © Springer-Verlag Berlin Heidelberg 2010

306

Exercises

The sales curve is displayed in the following figure. The product will have exclusivity for the 13 first years on the market. 100 80 60 40 20 0 1

5

2

3

4

5

6

7

8

9

10

11

12

13

14

15

16

17

18

10 18 26 34 41 48 54 60 66 72 77 82 86 90 93 96 98 99 100 100100 99 99 97 95 90 72 56 42 33 27 22 17 13 9

19

6

3

20

1

0

Fig. 4.1. Sales curve for project EXERCISE

Exercise 1: Project – Getting Prepared We must set up a spreadsheet with the input parameters. It is important to put all input parameters on a sheet showing all hypothesis. All consecutive calculations must reference on the cells on this hypothesis sheet when making use of the input parameters. Having all input parameters on one sheet allows changing the hypothesis easily. Exercise 2: Project – Growth Rate We have assumed a growth rate for the peak sales estimate. We calculate the peak sales estimate corresponding to each semester from now (t=0) to the end of the project’s life cycle (t= 26.5). We use half-year time steps because the approval phase is estimated to last 1.5 years. As a rule, smaller time steps yield more accurate results, but require a little more work. Exercise 3: Project – DCF Value We now want to calculate the DCF value of project EXERCISE. Consider the use of half-year time steps for the growth rate as seen in exercise 2, for the discount rate, and especially for the sales revenues. Note that the peak sales estimate refers to sales numbers achieved in one year.

Exercises

307

Exercise 4: Project – IRR Usually the value just indicates whether a project is profitable or not. But the value is not a good measure to compare projects between each other. Imagine project A and project B, both having a value of $ 3 mn. Project A requires $ 1 mn investment, and project B requires $ 2 mn investment. Although project A and project B have the same value, project A seems much more attractive, because it requires less investment. The internal rate of return (IRR) is the discount rate that returns a value of zero. This means, that investing the capital a project corresponds to investing the capital in an asset that yields the IRR as annual interest. We now want to know to what return investing in project EXERCISE corresponds. Exercise 5: Project – Binomial Tree We now also want to value the project using real options valuation. For this we need to span the tree that models the uncertainty of the peak sales estimate. Assuming the last decision point being at launch, we construct a tree that lasts from now (t=0) until launch (t=6.5). Again, we use half year time steps. Exercise 6: Project – End States As the next step, we place ourselves in the end states. In an end state the product is ready to be launched. We assume that all uncertainty has been resolved until then and that we therefore can make an accurate estimate of the product’s peak sales. Hence, we know all cash flows and can value the project as if it were at launch, using the peak sales estimate corresponding to the end state we are in. Finally, we do not only need to value the project at launch, we also have to decide whether it makes sense to launch the product. Exercise 7: Project – Working Back the Tree In order to get the value of the project today we must work back the tree. At each node lying one time step before the nodes we have already calculated, we must take the expectation, discount for the time step, adjust if in this time step a phase ends, account for the cash flows of this time step, and take a decision if the nodes are at a decision point. Not in every time step we have to account for all these points; in some time steps no phase ends, in some no cash flows are triggered, or no decision has to be made. The expectation and the discounting however applies to every time step.

308

Exercises

Comprehension Question 1 We have received different values for DCF and real options. Where does this difference come from? Can we spot this difference somewhere in the calculations? Exercise 8: License Contract – Getting Prepared We now assume that the project EXERCISE is licensed to a pharmaceutical company that applies a discount rate of 11%. The license terms are as follows: Table 4.2. License contract for project EXERCISE License contract terms Phase 1 upfront payment

$ 3 mn

Phase 2 milestone payment

$ 4 mn

Phase 3 milestone payment

$ 8 mn

Filing milestone payment

$ 12 mn

Milestone payment after having sold $ 100 mn

$ 25 mn

Royalty rate for annual sales up to $ 150 mn

6%

Royalty rate for annual sales higher than $ 150 mn

9%

The license terms with respect to commercialisation are a little more exotic than what we have previously seen. The milestone for launch is delayed and only paid once the cumulative sales of the drug exceed $ 100 mn. We call this sort of license payment cumulative milestones. It is not clear from the beginning when the milestone is triggered. If the launch is successful and the sales increase fast, then the milestone arrives early. If on the other hand the sales potential is low or the sales curve not very steep, then it takes longer to the payment of the cumulative milestone. The royalty rate depends on the amount of annual sales. Usually, the licensor receives 6% of sales, but if they exceed $ 150 mn – the barrier – then the licensor receives 9% of the exceeding amount. We must now first adapt our input sheet to the license contract.

Exercises

309

Exercise 9: License Contract – DCF Calculating the DCF value of a license contract is not fundamentally different from calculating the project value. Basically, we use milestone payments instead of costs, and royalties instead of sales revenues. Nevertheless, in this case we have to pay special attention to the royalties, because they are stacked, and to the commercialisation milestone, because it is a cumulative milestone. Exercise 10: License Contract – Real Options When calculating the real option value we must, of course, consider the same aspects about cumulative milestones and stacked royalties, but this time for each end node. In addition, we must calculate the value not only for the licensor, but also for the licensee, because the licensee decides about the continuation of the project. Comprehension Question 2 Explain the value differences between DCF and real options. How does the value difference behave with increased peak sales? Try out an initial peak sales estimate of $ 400 mn and explain the observation. Exercise 11: Technology Assume that the company owns a technology that allows bringing every 18 months one project like EXERCISE into clinical phase 1. Project EXERCISE is a project that has only a niche indication, usually peak sales are estimated at $ 310 mn for a standard project. As for the project EXERCISE we assume an annual growth rate for the value of the projects and for the costs as well. The company spends $ 3 mn each year for research and preclinical testing that precedes the clinical phases. The technology can probably be exploited for another ten years. Comprehension Question 3 Put the value of a standard project in relation to the value of the technology. Does it make sense? Exercise 12: Company – Input Parameters After having discussed all items of project, license, and technology valuation in detail, we now want to value a company.

310

Exercises Table 4.3. Company pipeline

Project

Phase

ABC102 Phase 1 (Start) ABC107 Preclinical (6 months left) ABC201 Preclinical (Start) ABC111 Lead Opt. (6 months left) ABC205 Lead Opt. (1 year left)

Disease Group Peak Sales

Strategy

CV

Excellent prec. data Licensed

CV

Too early to tell

Self conducted

Antiinfectives Too early to tell

Self conducted

CV

Too early to tell

Self conducted

Antiinfectives Too early to tell

Self conducted

The company has five projects in cardiovascular diseases and in antiinfectives. The company’s research laboratories focus on CV and antiinfectives. All four projects have been generated by researchers of the company and it is assumed that the research groups are able to sustain the current feed rate. The company aims at eventually commercialising the projects on its own. But it is in negotiations about licensing ABC102 to a pharmaceutical company. This large player showed also some interest in possibly taking over the whole company. In a first step we need to find out the input parameters for the projects, i.e. costs, durations, success rates, peak sales, sales curves, and margin. Exercise 13: Company – Discount Rate To value the company we must use a discount rate that corresponds to the company’s risk profile. The current pipeline is mainly planned to be developed by the company itself. Propose a discount rate. Exercise 14: Company – License Negotiation Initially the company planned to develop all projects on its own, since it is backed by very potent and confident investors. Nevertheless, a large pharmaceutical company approached the company whether it is interested in licensing its lead compound ABC102. The company therefore tries to find out what license terms it can possibly ask for. The targeted structure would be a milestone-royalty ratio of 1:1 and a steadily increas-

Solutions

311

ing milestone structure. The deal must be clearly value enhancing for the company, otherwise it would keep sticking to its initial strategy. What deal terms are possible assuming a discount of 10% for the pharmaceutical company? Exercise 15: Company – Feed Rates When valuing the company we must also consider the technologies. The company has two fields of excellence, CV and antiinfectives. Both base on a patented technology that can be exploited for 13 more years. All in all the company spends about $ 3 mn each year, however it is not exactly clear how these can be allocated to the single technologies and to general expenses. The projects generated by the antiinfectives technology are similar to ABC205. For the prototype project of the CV technology we use a project with peak sales somewhere in between the peaks sales of ABC102 and the peak sales of the other projects. Exercise 16: Company – Value Value the company on a stand-alone basis. Include a dummy corporate tax rate of 15%. The company’s current cash balance is $ 13 mn. More exercises and downloadable spreadsheets can be found on www.avance.ch.

Solutions Exercise 1: Project – Getting Prepared If a spreadsheet is arranged that all cells are referenced to the input cells, then we can easily observe the influence of one parameter by changing its value in the input cell. If we would hard code the numbers into the calculations we would have to change every cell that uses this input parameter, most probably forgetting some.

312

Exercises

Fig. 4.2. Worksheet “input”

Exercise 2: Project – Growth Rate Using the discrete compounding formula we can calculate the grown peak sales estimate for every time step. The following figure displays the formula for cell E7, corresponding to the peak sales estimate 1.5 years from now.

Fig. 4.3. Implementation of growth rate

Exercise 3: Project – DCF Value For the solution we display the spreadsheet once with its formulae, and once with its values.

Fig. 4.4. Worksheet “DCF” with formulae

Solutions

313

Note that the sales are taken account of half-yearly, the peak sales, however, refer to one full year. We therefore have to account for just half of the peak sales (as displayed in cell P8).

Fig. 4.5. Worksheet “DCF” with values

Exercise 4: Project – IRR The IRR is the discount rate that returns a value of zero. We can consequently try to find the IRR simply by modifying the discount rate on the worksheet “input”. We will find a discount rate of about 14.6%. On the other hand, Excel provides a function IRR to calculate the internal rate of return. As in the DCF we have discounted the risk adjusted cash flows, the risk adjusted cash flows in row 11 are the input. When calculating the IRR like this (=IRR(B11:BC11)) we get a value of 7.051%. This is not in line with the 14.6% we have found right before. The reason is that the IRR function calculates the rate for the time step between the cash flows, in this case for 6 months. To get the IRR corresponding to one year the formula would be =(1+IRR(B11:BC11))^2-1, which gives us an IRR of 14.599%. Finally, we could also use the Goal Seek function in the menu Tools as displayed in the figure.

Fig. 4.6. Finding the IRR with the Goal Seek function

314

Exercises

Exercise 5: Project – Binomial Tree In our tree the time steps are half-yearly. The formulae for the binomial tree give us the factor up u and the factor down d.

u = eσ d=

∆t

= e 0.25⋅

0.5

= 1.19

1 1 = = 0.84 u 1.19

With this the tree becomes:

Fig. 4.7. Binomial tree in worksheet “ROV”

Note that for the construction of the tree interestingly the growth rate does not matter. Only when calculating back we use the growth rate as input parameter to the probabilities for step up and step down. Exercise 6: Project – Calculate the End States For each end state we have to assume the corresponding peak sales estimate and calculate with this. When projecting the future revenues we must not forget that the peak sales estimate keeps growing every year by the growth rate. These grown peak sales we then multiply with the sales curve point corresponding to that semester. Note that we have to take just half of

Solutions

315

the peak sales estimate, because we only calculate the cash flow six months and not for one year. These cash flow must then be discounted back to launch (not back to today!). Summing them up and accounting for the launch costs we get the value for the end states. The following figure displays how the value can be calculated in one line.

Fig. 4.8. Valuation of project value for top end node

The figure displays the values of the project in the different end states.

Fig. 4.9. Value of end states

We see that in the three lowest states the value of the drug is negative. In these cases we would not launch the drug. Consequently, we can put these values to zero. A much better option however is to do that automatically by using the function max(.,.). This allows afterwards to modify all input parameters and the spreadsheet automatically calculates whether the project would be continued in that state. Taking the maximum of zero and the calculated value returns the value if it is positive and zero if the value is negative.

316

Exercises

Fig. 4.10. Implementation of option to abandon

Exercise 7: Project – Working Back the Tree The easiest method to work back the tree is to create a formula that applies to each time step and then copy-paste it into all nodes. This formula must therefore consider the following five points: 1. Expectation of up and down step 2. Discount for time step 3. Account for success rate 4. Account for cash flows 5. Take decision if we are in a decision point

Fig. 4.11. Solved tree

Solutions

317

The following figure displays how this concept can be implemented in a spreadsheet. For this we need to specify the success rates and the cash flows for every time step (lines 3 and 4). Of course we also have to calculate the probabilities of going up or down (B9 and B10). The implementation of the decision can be applied to all nodes, because a decision is every time necessary when a cash flow occurs. If no cash flow occurs, the project cannot turn negative, because by definition all future values are positive (otherwise the project would be abandoned and the value consequently turns zero, which is not negative). In the cell D25 we see the formula that has been pasted into each cell of the tree. The bold numbers correspond to the values of the project in the corresponding nodes. Comprehension Question 1 When valuing the project with real options we have increased the value in some nodes from a negative number to zero. This corresponds to a value increase. In the tree, these value increases translate automatically to the value in the root node by taking the expectation of up and down step. Maybe not the subsequent node is concerned by abandonment, but then it is either one of the next nodes, or even a successor of those. The figure displays where in the tree the project is abandoned because of lack of profitability (black nodes). Some nodes (grey nodes) are then in the shade of these abandonment nodes, i.e. they will never be achieved because the project has already been abandoned beforehand.

Fig. 4.12. Abandonment of project in the tree

318

Exercises

If we change the initial peak sales estimate from $ 230 mn to $ 250 mn, then we get a different picture. Now we are more likely to continue the project if it passes phase 2 and NDA.

Fig. 4.13. Abandonment of project if peak sales estimate is $ 250 mn

Exercise 8: License Contract – Getting Prepared We must put all license parameters on the worksheet “input”. It is important also to put in the time of market protection (I10) as well as the licensee’s discount rate (C12). We have also prepared two new worksheets “DCF Lic” and “ROV Lic” for the calculation of the value of the license.

Fig. 4.14. Worksheet “input” for a license contract

Solutions

319

Exercise 9: License Contract – DCF We first approach the problem of the cumulative milestone. It is only triggered after a certain amount of the drug could be sold. We can now simply calculate when this would be the case and find that somewhere between two and two and a half years after launch this limit will be passed. But remember that we want to design the spreadsheet in a way that we can change all input parameters. We therefore have to find a formula that automatically finds out when the cumulative milestone is triggered, independently of the peak sales estimate or the cumulative limit.

Fig. 4.15. Implementation of cumulative milestone

With the formula in the row 10. in the figure cell R10 is displayed, we must check whether the cumulative sales have exceeded the limit, only then the milestone can be triggered. But we can only trigger the milestone if it has not been triggered already beforehand, i.e. if the limit has not been exceeded in the previous time step already. These two conditions are in the AND-function. If both conditions are verified, then we account for the milestone, otherwise the cell has gets a zero-value. For the royalties we must check whether annual sales exceed the barrier. Note that the barrier refers to annual peak sales while in the sheet we calculate with semi-annual sales figures. For simplicity we therefore just divide the barrier by two such that we can compare it directly with semiannual figures. The error resulting from this shortcut is negligible.

Fig. 4.16. Implementation of stacked royalties

When implementing stacked royalties we have to check whether the sales revenues exceed the barrier. If they don’t we simply apply the low royalty rate to the sales revenues. If the sales revenues do exceed the barrier, we

320

Exercises

apply the low rate to an amount corresponding to the barrier and the high rate to the part exceeding the barrier. Finally, we have to consider that the royalties are only received as long as patent protection or market exclusivity is valid. In the formula in cell R11 the if clause return the royalties if the product has still market exclusivity and returns zero otherwise. We do not outline the rest of the calculations since they are similar to the ones already discussed in the case of a self-conducted project. We content ourselves with providing the final result.

Fig. 4.17. DCF calculation of license contract

The DCF value of the license for the licensee is $ 11.1 mn. Exercise 10: License Contract – Real Options Since we must calculate not only the licensor’s but also the licensee’s value at the end nodes we arrange for a little more space when constructing the tree. The calculation of the end nodes then follows the same lines as in the DCF calculation. We provide in the following a table the formulae for the uppermost end node. Table 4.4. Formulae for uppermost end node A

O

P

1 Time

6.5

7

2 Phase

Market

Market

3 Success 100% Rate 4 Costs

–100

13

2,290

100%

=$O13*(1+input!$B$8)^(P$1$O$1)*input!B$15/2

Solutions

321

Table 4.4 (continued) A

O

P

14

=MAX(SUM(P15:BC15) +O$4,0)

15

=IF(O14=0,0,SUM(P16:BC16)) =(P13*input!$B$7-P14)/ (1+input!$C$12)^(P$1-$O$1)

16

=IF(P$1= input!$J$6,SUM(O13:$P13)< input!$J$6),input!$H$6,0)+ IF(P13*2
=P14/(1+input!$B$12)^(P$1-$O$1)

In the column O lie all the end nodes, cell O13 being the peak sales estimate of the uppermost end-node. In the cells left of column O in row 13 we calculate the sales revenues corresponding to the time in row 1. In row 14 we calculate the license payments, i.e. the cumulative milestone plus the royalties. Row 15 then corresponds to the cash flows to the licensee discounted back to the end node, row 16 corresponds to the cash flows to the licensor. In the cell O14 we calculate the value of the license for the licensee, which is the sum of the discounted cash flows plus the launch costs. In O15 we calculate the value for the licensor. Note that the licensor’s value is zero if the licensee decides not to continue the project. The value of the license contract is $ 3.6 mn for the licensee and $ 9.2 for the licensor. Comprehension Question 2 In the figure above, we can see that the licensee refuses to continue the project in the lowest node at the beginning of phase 2 and in the lowest two nodes at launch. In these scenarios the licensor does not receive any license payments anymore, although they have a positive value. This implies that the value for the licensee is lower when calculated with real options than with DCF. We can attribute this difference to the flexibility that was handed over from the licensor to the licensee. The licensor is short in flexibility while the licensee is long in flexibility. If we use a peak sales estimate of $ 400 mn we get a DCF value of $ 13.69 mn and with real options $ 13.75 mn. Surprisingly, the real option

322

Exercises

Fig. 4.18. Calculation of the real options value for licensee and licensor

value is now higher. The real option value is now higher because a second effect now covers the lack of flexibility. In the binomial tree, we have now a lot of scenarios where most of the sales revenues are compensated at the high royalty rate. This finally leads to a higher average royalty rate than when using DCF. Furthermore, the flexibility is not as valuable at $ 400 mn than at $ 230 mn because it is much less probable that the projects gets abandoned.

Solutions

323

Exercise 11: Technology We have to use the same discount rate as for the company, i.e. 15%. This gives us a project value of $ 7.42 mn with DCF and $ 7.47 mn with real options. If the company can put every 18 months one project into clinical phase 1, it can consequently put 0.66 projects per year into clinical phase 1. Using the formula for valuing technologies, we can calculate the value of a technology producing 0.66 projects at a growth rate of 1% over an infinite time horizon. ∞ Vtech =

[fV

project

− (1 + r − g )Costs r−g

] = 0.66 ⋅ 7.4 − (1 + 15% − 1%) ⋅ 3 = $ 10.6 mn 15% − 1%

But since the technology is only exploited over ten years we have to subtract the discounted value of it after ten years. Vtech =

[fV

project

]

− (1 + r − g )Costs ⎛ ⎛ ⎞ 1 ⎞ 1 ⎜⎜1 − ⎟ = 10.6⎜⎜1 − ⎟ = $ 8.0 mn n ⎟ 10 ⎟ r−g ⎝ (1 + r ) ⎠ ⎝ (1 + 15% ) ⎠

The technology has a value of $ 8.0 mn when calculated with the DCF value of the standard project. The real options approach yields a value of $ 8.1 mn. Comprehension Question 3 A value of $ 8 mn seems moderate when compared to the $ 7.4 mn of a project generated by the technology. Nevertheless, it will take 18 months until a new project enters the pipeline, until then the company will spend $ 4.5 mn. The first project therefore has an approximate value of $ 1.7 mn (we must discount the $ 7.4 mn for the 18 months time difference and also part of the costs). This number now compares well to the value of the technology. Exercise 12: Company – Input Parameters Since project ABC102 has excellent preclinical data we are confident in using average sales numbers for this compound. For above average sales numbers it is still a bit early to tell, normally proof of concept studies in man then justify further adjustments towards blockbuster status. For all other projects we use the median numbers of the respective disease group. Not knowing better we assume the other input parameters to correspond

324

Exercises

more or less to industry standards. We use overall numbers mentioned in the chapter about fundamentals in life sciences. The costs of the current phase must be adjusted to the remaining length. Usually we would need to assume higher launch costs for the first compound that reaches commercialisation, accounting for the setting up of a marketing and distribution network. Unfortunately we do not know which project will be the first one; therefore we use a standard assumption of the half of the peak sales estimate as the launch costs. Table 4.5. Parameters of project ABC102 In $ mn

Phase 1

Phase 2

Phase 3

NDA

Length

1.5 years

2 years

2.5 years

1.5 years

Cost

4

10

45

3

Success Rate

63%

43%

76%

84%

Peak Sales

466

Margin

65%

Growth

0%

Launch Costs

233

Table 4.6. Parameters of project ABC107 In $ mn

Preclinical Phase 1

Length

0.5 year

1.5 years 2 years

2.5 years 1.5 years

Cost

1.5

4

10

45

3

Success Rate

65%

63%

43%

76%

84%

Peak Sales

145

Margin

65%

Growth

0%

Launch Costs 73

Phase 2

Phase 3

NDA

Solutions

325

Table 4.7. Parameters of project ABC201 In $ mn

Preclinical Phase 1

Length Cost Success Rate

1 year 3 65%

Peak Sales Margin Growth Launch Costs

265 65% 0% 133

Phase 2 Phase 3

NDA

1.5 years 2 years 2.5 years 1.5 years 4 10 45 3 71% 51% 80% 97%

Table 4.8. Parameters of project ABC111 In $ mn

Lead Opt. Preclinical Phase 1

Length Cost Success Rate

0.5 years 1 year 1.2 3 70% 65%

Peak Sales Margin Growth Launch Costs

145 65% 0% 73

Phase 2 Phase 3

NDA

1.5 years 2 years 2.5 years 1.5 years 4 10 45 3 63% 43% 76% 84%

Table 4.9. Parameters of project ABC205 In $ mn

Lead Opt Preclinical Phase 1

Length Cost Success Rate

1 year 2.5 70%

Peak Sales Margin Growth Launch Costs

265 65% 0% 133

1 year 3 65%

Phase 2 Phase 3

NDA

1.5 years 2 years 2.5 years 1.5 years 4 10 45 3 71% 51% 80% 97%

326

Exercises

$ 145 mn peak sales is very moderate for CV drugs. Many CV drugs report sales below $ 100 mn and there is no clear sign yet that the projects are superior to other approaches. Nevertheless, the promising preclinical results of ABC102 justify peak sales of about $ 300 mn for the prototype project of the CV technology. We will use the following sales curve (quarterly): 1 0.8 0.6 0.4 0.2 0 0

2

4

6

8

10

12

14

16

18

20

22

Fig. 4.19. Sales curve for projects

Exercise 13: Company – Discount Rate The discount rate is a critical input parameter to the valuation of a company. The company has only one project at the beginning of clinical phase 1, all others being in preclinical phase or lead optimization. Our proposed framework in the discount section suggests a discount rate of around 20%. Higher success rates in CV and antiinfectives compared to the average success rates used in the mentioned discount model justify lowering the discount rate to 18%. Success rates have an important influence on the discount rate. Since we have made conservative assumptions for the peak sales we can feel comfortable with this lower success rate. The license deal would allow lowering the discount rate to 17.5%. Exercise 14: Company – License Negotiation Since for our company the deal must be clearly value enhancing we must compare the value of a possible license with the value of the project if developed in-house. We do not consider any aspects like steeper sales curve or better marketing of the pharmaceutical company. First, we need to value the project as in-house project at a discount rate of 18%. To estimate the leeway in the negotiations we should also calculate the same project in the pharmaceutical company’s eyes. The values are $ 26.7 mn for our company and $ 74.2 mn for the pharmaceutical company.

Solutions

327

We then can produce the numbers corresponding to a generic license contract that has a milestones-royalty ratio of 1:1 and milestone weights of 1 for clinical phase 1, 2 for clinical phase 2, 3 for clinical phase 3, 4 for filing and 5 for approval. The minimum contract must correspond to $ 26.7 mn when valued at a discount rate of 17.5%, the maximum contract to $ 74.2 mn at 12%. Table 4.10. Extreme license contracts delimiting the negotiation leeway In $ mn

Minimum Contract

Maximum Contract

4.5 9.1 13.6 18.2 22.7 16.5%

12.1 24.1 36.2 48.3 60.4 25.3%

Upfront payment Phase 2 milestone Phase 3 milestone NDA milestone Launch milestone Royalty rate

The license contract should therefore lie somewhere in between if the premises of the generic license contract should be respected. Especially the royalties seem very high for a phase 1 deal. For the pharmaceutical company it is more advisable to pay a larger upfront payment, reducing all future payments. An alternative would be buying the whole company, which is apparently the option the pharmaceutical company also decided to consider. Exercise 15: Company – Feed Rates We can subdivide the company in two parts, a CV part and an antiinfective part. The CV part is composed by three projects in phase 1, preclinical phase, and lead optimisation as well as a technology. Displayed on a time axis we have the following picture: Table 4.11. CV projects Time

1

2

3

Phase Probability Project

Lead opt 100% ABC111

Preclinical 70% ABC107

Phase 1 45.5% ABC102

328

Exercises Table 4.12. Antiinfective projects Time

1

2

3

Phase Probability Project

Lead opt 100% ABC205

Preclinical 70% ABC201

Phase 1 45.5%

Assuming that ABC111 has just been produced by the technology we have to find the feed rate that is most likely to end up with one project in preclinical and another in phase 1. This leads to the following problem, where pi denotes the probability of phase 1 and ni the number of projects in phase i (the ni describe the current pipeline): ⎧⎪ max ⎨ f ⎪⎩

⎛ ⎜ ⎜ ⎝

∑( ) ∏ f ni

i

i

j =1

⎞ p j ⎟⎟ ⎠

ni

⎛ ⎜1 − ⎜ ⎝

i

∏ j =1

⎞ p j ⎟⎟ ⎠

( f − ni )

(4.1)

⎫⎪ ⎬ ⎪⎭

This method yields a feed rate between one and two for both technologies. Staying conservative we use feed rates of 1. The CV prototype project (with 1 year lead optimization at $ 2.5 mn, $ 330 mn peak sales, and $ 165 mn launch costs) has a value of $ 0.3 mn with DCF at 18%, project ABC205 has a value of $ 2.3 mn with DCF. As an overall technology for the company we receive: Vtech = Vtech =

[f

CV

VCV + f antiVanti − (1 + r )Costs ] ⎛ 1 ⎜⎜1 − n r ⎝ (1 + r )

[1 ⋅ 0.3 + 1 ⋅ 2.3 − 1.18 ⋅ 3] ⎛1 − 0.18

⎜ ⎝

⎞ ⎟⎟ ⎠

1 ⎞ ⎟ = −$ 4.6 mn 1.1813 ⎠

Note that the technology has a negative value. This does not necessarily mean that the technologies do not generate any value, because part of the costs is general and administrative which is also spent to manage the existing five projects. Exercise 16: Company – Value To value the company we must value its parts. Using DCF we get:

Solutions

329

Table 4.13. Value of the company Asset ABC102 ABC107 ABC201 ABC111 ABC205 Technology Sum Tax (15% of sum) Cash Total

Value ($ mn) 26.7 8.2 –2.9 –3.9 2.3 –4.6 25.8 –3.9 13 34.9

Interestingly, the company also conducts negatively valued projects. We have included these in the final value, because the company plans to go ahead with them without caring about their value. If somebody wants to take over the company, these should not be included in the final value, because the new company would not need to continue them if it considers them loss making. This increases the value by $ 6.8 mn, neglecting also a part of the technology the value increases even more. Considering further that the overtaking company can value the same projects at a much lower discount rate, the shareholders could be offered a nice premium on the current share price.

Case Studies

Introduction In this chapter we present four case studies. For each case a template Excel spreadsheet has been prepared. These template spreadsheets can be downloaded on www.avance.ch. The reader can solve the case studies as an extended exercise – a sketch of the solution for the first two cases is added at the end of each case. Each case study addresses a specific topic. The first case, “Creating a Term-Sheet for BT-100”, is about structuring a term sheet for the buy-side. A mid-sized pharmaceutical company is interested in in-licensing a phase I project. The solution requires a well-structured Excel sheet with rNPV valuation. The solution in class takes about 2 hours. In the second case, “Bio-Tec’s Lead Compound”, the student must compare different strategic options for a project. Either the company licenses it now, after the next phase, or not at all. For the solution the student is expected to create an Excel spreadsheet with rNPV and real options. He must compare different options and how these alternatives can be used in negotiating a term sheet. The solution in class takes 3-4 hours. The third case, “Licensing MA-1204” deals with the problem of multiindication projects. The solution in class requires 4 hours. Finally, an interactive case simulates a negotiation between two parties. The solution for this case can be requested from the authors for academic purposes, e.g. use in MBA classes. This negotiation case study requires two days and at least two groups.

Case Study 1 General Midpharm is a medium sized pharmaceutical company with an established sales force in Europe. Midpharm commercialises its drugs in the rest of the world through various partners. Midpharm has its own research labs but

B. Bogdan, R. Villiger, Valuation in Life Sciences, 3rd ed., DOI 10.1007/978-3-642-10820-4_5, © Springer-Verlag Berlin Heidelberg 2010

332

Case Studies

also in-licenses up to 50% of all drug candidates they have in their pipeline. The license agreement may start as early as discovery phase, but usually Midpharm has a preference for products in clinical development. Typically they conclude worldwide license contracts with Midpharm as the sole developer of the product. However, in some cases they agree to codevelop the compound with the licensor, who then usually maintains the commercialisation rights outside of Europe. Licensing and business development is therefore an important part of Midpharm’s business model. Midpharm focuses on different disease categories, amongst them cancer. Dr. Partnering is responsible for cancer opportunities within Midpharm’s licensing department. In this function he visits conferences and also conducts the negotiations between Midpharm and the potential licensor’s. However, he is not involved in the later commercialisation agreements. BT-100 In order to spot in-licensing opportunities Dr. Partnering participates in industry conferences. In one such conference he gets to know Mrs. Business Developer from Bio-Tec. Bio-Tec has a promising CNS technology, as far as Dr. Partnering can judge. However, Bio-Tec still in early stages, with the most advanced drug in phase I. In addition they also have a cancer drug – BT-100 – against Glioblastoma. The CNS technology turned out to be Bio-Tec’s value driver and all cancer activities except this drug in phase I have been abandoned. Dr. Partnering is interested in BT-100 because of three reasons. First, Glioblastoma can only be treated with Temodar until now, and only 15% of all patients react positively on Temodar, so Dr. Partnering thinks that there could be some potential. Temodar’s sales amounted to US$ 703 mn in 2006. Second, Dr. Partnering imagines that the CNS projects could be interesting for Midpharm as well and that in that respect it would be helpful to have already one step in Bio-Tec’s door. Third, he assumes that BioTec would be ready to cease the rights on BT-100 at a reasonable price since it is as a cancer therapeutic not really aligned with their strategy. Due Diligence After the conference Dr. Partnering decides to pursue this opportunity and asks Bio-Tec for further information. After having signed a confidential disclosure agreement Midpharm gains full access to all preclinical data of BT-100 and the currently prepared phase I protocol. Dr. Partnering studies with a team of scientists, clinical trial specialists, statisticians, competitive

Case Study 1

333

Table 5.1. Clinical development data

Costs Duration Success rates

Phase I

Phase II

6 2 64%

15 2 42%

Phase III

Approval

60+60 3+2 3 each 1 each 69%/69% 85%/85%

intelligence professionals, and marketing professionals the data and prospects of BT-100. The team comes up with the following preliminary estimations: Phase I and phase II studies will take quite long because only patients can be recruited that have been unsuccessfully treated with Temodar. After having successfully completed a combination phase III study Midpharm plans also to conduct a comparative study. Midpharm expects to reach peak sales of US$ 300 mn in case only the combination treatment will be approved and another US$ 500 mn if BT-100 will also be approved as first line treatment. The first line treatment study (phase III) will only be initiated if the combination treatment gets approved. So, the probability of the first line treatment study is slightly lower. Midpharm expects to have launch costs of US$ 50 mn for the combination treatment and another US$ 50 mn for the first line treatment. On an ongoing basis Midpharm expects to have an operating margin of 70%, i.e. it will have expenditures of 30% for cost of goods and marketing and sales. The sales are forecasted as displayed in the figure below. 900 800 700

US$ Mio

600 500 400 300 200 100 0

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 Combination

First Line

Fig. 5.1. Expected sales curve

334

Case Studies

The graph displays the different sales curves as percentage of peak sales for Midpharm and Bio-Tec. Midpharm is expected to achieve peak sales of US$ 300 mn with the combination treatment and US$ 800 if also the first line treatment gets approved. Table 5.2. Generic phase I license contract Phase I deal Upfront 1st dose phase II 1st dose phase III Filing Approval Royalties Value share

1 1 2 3 10 4%-15% 25% vs. 75%

A phase I deal (closed after preclinical results) usually displays a value share of 25%-75% for licensor and licensee. Upfront and milestone payments are set in a way that their relative size behave like the indicated weights. The filing milestone therefore is three times the size of the upfront payment. The approval milestone can be split by country, sometimes also by indication. These rules can however be modified according to the preferences of the contract partners. Dr. Partnering asks his colleague Mr. Parkinson, who is responsible for the CNS opportunities, whether Bio-Tec’s CNS portfolio could be of interest for Midpharm. Based on the scarce information available Mr. Parkinson thinks that it looks interesting and worthwhile to track. However, for a substantial investment he would need more data, i.e. the NDA should be extended to the CNS projects as well. But Mr. Parkinson thinks that it might be a very good idea to establish already a partnership with Bio-Tec in order to get easier access to the CNS compounds in a second phase. Entering the Negotiations After having discussed BT-100 at the monthly BD&L board meeting Dr. Partnering receives the mission to go back to Bio-Tec and offer them to license BT-100. Midpharm would like to postpone most license payments to later development and commercialisation when the value of the com-

Case Study 1

335

pound is better justified. The value is usually calculated at the licensee’s cost of capital. Most analysts value Midpharm, a listed company, at a cost of capital of 12%. Table 5.3. License contract with University of Research Payments Preclinical phases

sunk

st

0.1

st

1 dose phase II

0.3

1st dose phase III

0.8

1 dose phase I

Filing

0

Approval

2

Royalties

3%

Dr. Partnering then arranges a meeting with Mrs. Business Developer to learn more about the needs of Bio-Tec and the implications to a possible license contract. Mrs. Business Developer explains that Bio-Tec is currently in another financing round where they want to raise US$ 15 mn, which allows them to cover the company’s operations for the next 30 months. Furthermore, Bio-Tec has to comply also with the terms of its original license contract with the University of Research, where BT-100 originally was discovered. Remembering his discussion with Mr. Parkinson he thinks that Bio-Tec would welcome an equity investment. This would have beneficial effects for both companies. First, Bio-Tec would close at least part of its investment round and could postpone further dilution to a moment when it has a stronger value proposition. Second, offering part of the upfront payment as equity investment at a high valuation, this license income would not be taxed. Third, doing this favour to BioTec and having already a stake in the company, Midpharm would have good access to the CNS technology at a later point in time. Tasks 1. Value BT-100 as if it were Midpharm’s project. 2. Find reasonable deal terms for a phase I deal according to industry standards.

336

Case Studies

a.) Calculate the value share at the joint discount rate of 12%. b.) Calculate the ROI (return on investment) for Midpharm. 3. Adapt the deal terms assuming that Midpharm will also make a US$ 4 mn Equity investment. 4. How could you modify the valuation assumptions in order to get a reasonable, but still low-balled first offer? Case Study 1: Solutions The case study, the template spreadsheet, and the solution spreadsheet can be requested by e-mail from the authors. We have not outlined the solutions in details so that the cases can be used for the instruction in class. In the continuation present just a sketch of the solution as presented in the downloadable spreadsheet. Value of BT-100 as if It Were in Midpharm’s Hands For the negotiations we need to know what is the value of the project in Midpharm’s hands. This represents the upper albeit unrealistic limit in the negotiations. With a standard rNPV calculation we receive with a discount rate of 12% a value of US$ 77.70 mn and a, IRR of 25.9%. Reasonable Deal Terms Assuming a 12% discount rate and a 25%-75% value share for a phase I deal, the table below presents a fair term sheet. Table 5.4. Fair deal terms for a phase I deal Terms

US$ mn

Upfront

2

Milestone phase II

2

Milestone phase III

4

Milestone filing

6

Milestone approval combination/first line Royalties

20/10 8%-12%

Case Study 1

337

The term sheet accounts for one milestone of US$ 10 mn if also the second indication is approved. The royalties are tiered, the first tier going from US$ 0 mn to US$ 200 mn annual sales with a royalty rate of 8%, the second tier ranging from US$ 200 mn to US$ 400 mn annual sales with a royalty rate of 10%, and the third tier for all annual sales above US$ 400 mn and a royalty rate of 12%. The project has an IRR of 22.2% for Midpharm. Reasonable Deal Terms with Equity Investment If Midpharm also takes a share in Bio-Tec’s equity, then the Bio-Tec’s need in upfront cash is reduced and the deal can be back-loaded. The following table presents one such back-loaded deal. Table 5.5. Back-loaded phase I deal Terms Upfront Milestone phase II Milestone phase III Milestone filing Milestone approval combination/first line Royalties

US$ mn 0 1 3 6 20/10 10%-13%

The royalty tiers this time are 10% up to US$ 350 mn, 11% up to US$ 500 mn and 13% above. Modifying Assumptions In order to negotiate, Midpharm would like to send Bio-Tec a low-balled but still defendable offer. For this purpose Midpharm could lower the value of the project while maintaining a 25% value share and hence declaring the offer as fair. A lower value can be achieved by different means, typically one opts for the difficult to estimate parameters and chooses a conservative estimate. Cost of capital, peak sales, or sales ramp typically fall into this category. For success rates often various probability sets are available from different sources. Indication specific success rates and correlations or joint approval probabilities are quantifiable and offer room for subjective estimates.

338

Case Studies

Either side has to be prepared that the negotiation partner chooses parameters in their own favour. In this case the inclusion of performance dependent payments are a remedy to reconcile different assumptions. Indication and market specific milestones, milestone payments dependent on how much has been sold, or tiered royalties are popular terms that allow aligning different assumptions.

Case Study 2 General Bio-Tec is a young biotech company, founded six years ago by Dr. CEO. Dr. CEO, while still working as a postdoc at the university, discovered a new method to identify a special class of molecules that act on the central nervous system. Dr. CEO, supported by the tech transfer office, spun-out his technology from the university. In exchange, the university now holds 5% of Bio-Tec’s shares. The technology allows Bio-Tec to identify drug candidates acting on the CNS. At the moment, Bio-Tec holds a promising pipeline of three projects, two of them already being studied in clinical trials. In addition, several other leads are still in discovery, and the company expects to be able to put each year one compound into regulatory preclinical phase. The company currently has EUR 16 mn cash at hand, which should be sufficient to take BT-100 to proof of concept in man, i.e. end of phase IIa, BT – 101 through phase I, and apply for an IND for BT-102 within the next 24 months. Table 5.6. Pipeline of Bio-Tec Project

Phase

Indication

BT-100 BT-101 BT-102

Phase I/IIa Phase I Preclinical

Glioblastoma Parkinson Not disclosed

In addition, we can expect that Bio-Tec can generate each year a new preclinical project.

Case Study 2

339

Strategy Bio-Tec’s most advanced project, BT-100, has just entered phase I/IIa trials. Dr. CEO assumes that after having shown proof of concept in man after this phase I/IIa study, the optimal time has come to partner the compound with a pharmaceutical company. Dr. CEO acknowledges that large-scale clinical trials are not Bio-Tec’s core competence and assumes that a large pharmaceutical company is better placed to conduct a phase III trial. Also, Dr. CEO believes that the commercialisation of BT-100 requires an enormous effort and a set-up of a large sales department within Bio-Tec. A license deal would be beneficial to the company in three ways: 1. A license deal would provide Bio-Tec with valuable cash, potentially avoiding further dilution of existing investors. 2. A license deal allows Bio-Tec to securitize a part of BT-100’s value, mitigating the risk of phase II & III failure or non-approval. 3. License deals often mean that renowned industry players approve the concept of the drug, which has a value enhancing effect on the whole company. Midpharm Enters the Scene On an industry conference Mrs. Business Developer from Bio-Tec got approached by Midpharm, a mid-sized pharmaceutical company specialised in the cancer market. Midpharm has shown interest in licensing BT-100. After having exchanged confidential disclosure agreements Midpharm started a due diligence of BT-100’s preclinical results and of the current phase I/IIa setup. Mrs. Business Developer then received a letter from Midpharm, saying that they seriously consider in-licensing BT-100. Attached Midpharm sent a term sheet where they offered to license BT-100. The main points of the term sheet are displayed in the table below. Table 5.7. Term sheet from Midpharm Milestone payments

In EUR mn

Upfront (cash) 1st dose phase IIb 1st dose phase III Submitting NDA Marketing approval

1 1 3 5 6

340

Case Studies

In addition, Midpharm pays 4% of net sales as royalties to Bio-Tec. Mrs. Business Developer immediately heads for Dr. CEO’s office, presenting him the term sheet. Evaluating the Options Dr. CEO took a look at the term sheet and commented that these terms will hardly please Bio-Tec’s investors. On the other hand it seems pretty clear that this term sheet is meant as a low-balled starting point for negotiations. Although Dr. CEO has not considered licensing BT-100 right now, this might be an attractive way to take the company forward. Especially because Dr. CEO considers that the core competence of the company is not in cancer, but in CNS. He could therefore use the cash to bring the company’s core project BT-102 forward. Dr. CEO decides to consider Midpharm’s term sheet and asks Mrs. Business Developer to provide him with the necessary analysis to respond to Midpharm. In particular, Dr. CEO wants to know: 1. What is BT-100 worth if Bio-Tec decides to commercialise it on its own? 2. What can Bio-Tec expect to achieve in a license contract after positive phase II data (phase III deal)? 3. What deal terms would be reasonable for a phase I/IIa deal? 4. What deal terms should Bio-Tec get to reach a similar result to a phase IIb deal? 5. What counteroffer should Bio-Tec make? 6. Are there other reasonable options? Market Research Mrs. Business Developer immediately starts working on these six questions. First, she establishes the data for the clinical trials, including development costs, timelines, and success rates. Mrs. Business Developer assumes that independently of the study sponsor the phase I/IIa trials will cost EUR mn all together. They are estimated to last over full years. She assumes that the success rate for a cancer phase I/IIa trial corresponds to published phase I cancer success rates. According to cancer success rates she estimates phase I/IIa trials to be successful in 64% of all cases. Phase IIb trials would then take three years at a cost of EUR 10 mn and with a success rate of 42%. Phase III trials last another three years at a cost of

Case Study 2

341

Table 5.8. Clinical development data Phase I/IIa Phase IIb Phase III Approval Costs Duration Success rates

4 2 64%

10 2 42%

68 3 69%

2 1 85%

EUR 68 mn and a success rate of 69%. Finally, approval phase will take another 12 months at EUR 2 mn with approval 85% of all cases. Once BT-100 gets approval, Mrs. Business Developer estimates the drug to achieve peak sales of EUR 300 mn worldwide if commercialised by Midpharm. She calculated the number by comparing the properties of BT-100 to the current market leader for the treatment of Glioblastoma, Temodar, which had sales of 703 US$ mn in 2006, taking into account that BT-100 probably will only be a second line treatment. Accounting for the lack of marketing experience she estimates the drug only to achieve EUR 250 mn if marketed by Bio-Tec. In both cases, EUR 50 mn launch costs and an operating margin of 70% (i.e. Cost of goods and marketing and sales account for 30% of net sales) need to be considered. Mrs. Business Developer also thinks that Midpharm can achieve peak sales faster than Bio-Tec. This is displayed in the different sales ramps she uses for the calculation. The sales of BT-100 will be protected by patents for 14 years after launch. Furthermore, Mrs. Business Developer estimates the Glioblastoma market to grow at 0% and the volatility of sales for real options analysis to be 30%. For the calculation Mrs. Business Developer also needs to know the cost of capital of both companies. In the last financing round 2 months ago BioTec was valued at a cost of capital of 18%, and since then no significant event happened. Midpharm is a publicly traded company and analysts use 13% as its cost of capital. The figure displays the different sales curves as percentage of peak sales for Midpharm and Bio-Tec. Midpharm is expected to achieve peak sales of EUR 300 mn, Bio-Tec of EUR 250 mn. Both companies need to spend EUR 50 mn to launch the drug. Operating expenses amount to 10% COGS and 20% marketing and sales, which results in an operating margin of 70%. For the calculation of license contracts Mrs. Business Developer consults industry reports and databases and finds that in general, the deals are arranged such that in phase III licensor and licensee each get 50% of the project value (calculated at the licensee’s cost of capital) For phase II deals

342

Case Studies

US$ mn

100% 90% 80% 70% 60% 50% 40% 30% 20% 10% 0%

Bio-Tec Midpharm

1

3

5

7

9

11

13

15

17

19

Fig. 5.2. Sales curves for Bio-Tec and Midpharm

this value share shifts to 35%-65% for licensor and licensee. For a phase I deal finally, the value share is 25%-75%. Milestone payments are typically increasing and royalties range from 10%-30% for a phase III deal and from 8%-20% for a phase II deal, and from 4%-15% for a phase I deal. Table 5.9. License contracts Phase I deal Phase II deal Phase III deal Upfront

1

1

1st dose phase II

1

-

1st dose phase III

2

1

-

NDA

3

2

1

Approval

4

3

2

Royalties Value share

1

4%-15%

8%-20%

10%-30%

25%-75%

35%-65%

50%-50%

A phase II deal (closed after phase I results) usually displays a value share of 35%-65% for licensor and licensee. Upfront and milestone payments are set in a way that their relative size behave like the indicated weights. The approval milestone therefore is three times the size of the upfront payment. These rules can however be modified according to the preferences of the contract partners. Mrs. Business Developer feels now sufficiently equipped to value the various options.

Case Study 2

343

Task 1. Value BT-100 as if Bio-Tec will ultimately commercialise the project 2. Find reasonable deal terms for a phase II deal 3. Find reasonable deal terms for an immediate phase I deal 4. Design a counteroffer to Midpharm 5. What do you propose Dr. CEO to do? Case Study 2: Solutions The case study, the template spreadsheet, and the solution spreadsheet can be requested from the authors. In the continuation we present a short description of the solution. Value of BT-100 as if Bio-Tec Would Commercialise It We want to value BT-100 from Bio-Tec’s perspective in order to get a lower limit for the negotiations. In this case we must use Bio-Tec’s discount rate and also apply the peak sales estimate elaborated for Bio-Tec in connection with the appropriate sales curve. With this we get a rNPV of US$ 6.07 mn and a real options value of US$ 6.51 mn. The value difference between rNPV and real options is 7%. Slight value differences might occur because of different conventions when to expect sales and expenses (at the beginning or at the end of a year). Reasonable Deal Terms for a Phase II Deal The following deal terms give a value share of 35%, corresponding to industry standards, calculated at a discount rate of 13%. The value for Midpharm amounts to US$ 57.25 mn, the value for Bio-Tec to US$ 30.76 mn. Due to the advanced stage of the project there is no difference between rNPV ad real options value detectable anymore. For the calculation we used the parameters for a commercialisation by Midpharm.

344

Case Studies Table 5.10. Deal terms for a fair phase II deal Terms

US$ mn

Milestone Phase II Milestone Phase III Milestone Filing Milestone Approval Royalty

7 7 14 21 10%

Midpharm achieves a risk adjusted ROI of 22.3%. Reasonable Deal Terms for a Phase I Deal The current term sheet is worth US$ 34.24 mn for Midpharm and US$ 6.11 mn for Bio-Tec, but calculated at a discount rate of 13%. A fair deal would offer Bio-Tec a 25% share of the whole project value. The following terms comply with these conditions. Table 5.11. Deal terms for a fair phase I deal Terms Upfront Milestone Phase II Milestone Phase III Milestone Filing Milestone Approval Royalty

US$ mn 2 2 5 8 10 6%

Under these terms the value for Midpharm amounts to US$ 30.25 mn and for Bio-Tec to US$ 10.09 mn, corresponding to a 25% value share. This term sheet still offers Midpharm an IRR of 21.4%. Counteroffer As a counteroffer Bio-Tec can design a term sheet corresponding to a value share of 40%. The following term sheet displays such a deal.

Case Study 2

345

Table 5.12. Deal terms for a 40% value share in phase I Terms

US$ mn

Upfront Milestone Phase II Milestone Phase III Milestone Filing Milestone Approval Royalty

4 4 8 12 16 8%

The project has with these terms still an IRR of 19.2% for Midpharm. During the negotiations Mrs. Business Developer can stress the still attractive IRR, or can as well include even a growth rate, making the project even more attractive and justifying larger terms. Proposition to Dr. CEO The table below displays Bio-Tec’s various options. Table 5.13. Bio-Tec’s options Option Self-conduct project License with current term-sheet License to reasonable terms License in phase II

Value in Comment US$ mn 6.51 4.09 6.99 6.77

Real options value Real options value discounted at 18% Real options value discounted at 18% Risk adjusted and discounted real options value, accounting for phase I costs, discounted at 18%

The valuation suggests pursuing negotiations with Midpharm in order to achieve reasonable deal term. A license in phase I looks like the most attractive option to Bio-Tec.

346

Case Studies

Case Study 3 General MAB Rx (Monoclonal Antibody Therapeutics) is a biotechnology company specialised in the discovery, production, and development of monoclonal antibodies. MAB Rx’ strategy is to license its projects at one point along the development path. MAB Rx has already closed several license deals. Mr. Deal, the head of business development, is right now in negotiations with Pharma International, a multinational pharmaceutical company for a license deal for the project MA-1204. MA-1204 The project MT-1204 is a monoclonal antibody with potential applications in cancer and in rheumatoid arthritis (RA). During the early phase of the project MAB Rx worked with Research Company. This collaboration included a license to access Research Company’s technology that was crucial for the development of MA-1204. From this license MAB Rx owes license payments to Research Company. Preclinical studies have shown promising activity in breast cancer. MAB Rx has concluded all IND enabling studies and has filed an investigational new drug application (IND) at the FDA. Meanwhile Pharma International has expressed its interest in MA-1204. Recent Deals Mr. Deal conducts a survey on the latest monoclonal antibody deals and is particularly impressed by the top line figures of the Dyax-sanofi aventis agreement, the Regeneron-sanofi aventis agreement, the deal between Genmab and GlaxoSmithKline and the Morphosys-Novartis deal. Feb 12, 2008. Dyax Corp. Enters into Antibody Agreements with sanofi aventis for the Fully Human Monoclonal Antibody DX-2240 and Phage Display Technology – Upfront Fees and Milestone Payments Total up to $500 Million – Dyax Retains Co-development Rights and Profit Sharing for Certain Partner Programs CAMBRIDGE, Mass. – Dyax and sanofi aventis announced today that they have entered into agreements in which sanofi aventis has been granted an exclusive worldwide license for the development and commercialization of the fully human monoclonal antibody DX-2240, as well as a nonexclusive license to Dyax’s proprietary antibody phage display technology.

Case Study 3

347

Under the terms of the two agreements, Dyax is eligible to receive up to $500 million in license fees and milestone payments in the case of full commercial success of the first five antibody candidates, including DX-2240. Dyax will receive $25 million in 2008. In addition, Dyax is eligible to receive royalties based on commercial sales of DX-2240 and other antibodies developed by sanofi aventis. As exclusive licensee, sanofi aventis will be responsible for the ongoing development, commercialization and consolidation of sales of DX-2240. For certain other future antibody product candidates discovered by sanofi aventis, Dyax will retain co-development and profit sharing rights, while sanofi aventis will maintain the leadership in development and commercialization, and book sales worldwide. Source: Dyax press release

Dec 2, 2007. MorphoSys and Novartis Forge New Strategic Alliance to Establish Innovative Therapeutic Antibody Pipeline One of the Industry’s Largest Pharma-Biotech R&D Collaborations MorphoSys AG (Frankfurt Stock Exchange: MOR; Prime Standard Segment) announced today the formation of one of the most comprehensive strategic alliances with Novartis in the discovery and development of biopharmaceuticals. The deal is aimed at establishing a pipeline of innovative drugs, and combines MorphoSys’s and Novartis’s research and development capabilities. Novartis becomes MorphoSys’s preferred collaborator for HuCAL-based drug discovery, allowing MorphoSys to progress to the next stage of its corporate development, which involves a greater focus on drug discovery and development within the Novartis alliance, and proprietary drug development, thereby substantially reducing MorphoSys’s reliance on new or extended fee-for-service discovery deals. The expanded alliance also includes rights to co-detail co-developed products in specific territories through creation of MorphoSys’s own sales force. In addition to programs pursued jointly, Novartis has accelerated its plan to internalize MorphoSys’s leading human antibody technology, HuCAL, at its research sites under the option agreed in the original contract. The structure and the financial terms of the collaboration further create a solid financial footing allowing MorphoSys to accelerate and broaden its proprietary drug development efforts and participate in development activities with Novartis. Transaction Structure Under the agreement, Novartis will make a major long-term commitment to MorphoSys’s HuCAL technology. The collaboration has a term of 10 years. Novartis has the option to prolong the collaboration for a further two years or to conclude the alliance after 7 years in certain limited circumstances. Over the lifetime of the agreement, the parties will engage in approximately double the annual number of therapeutic antibody discovery programs as compared to the previous alliance, encompassing a wide range of diseases. MorphoSys also has options to participate in certain development activities in various programs, with part of the early stage costs being funded by Novartis. Under the co-development options, MorphoSys may elect to participate in these projects through cost and profit sharing with financial participation reflecting its level of investment in the respective programs.

348

Case Studies

Financial Terms Based on a 10-year term, committed total annual payments sum to more than US$ 600 million in technology access, internalization fees and R&D funding, excluding reimbursement of R&D costs related to early stage development activities. Total payments under the agreement, including committed payments and probabilityweighted success-based milestones, contingent upon successful clinical development and market approval of multiple products, could potentially exceed US$ 1 billion, assuming the collaboration successfully runs its maximum term. In addition to these payments, MorphoSys would also be entitled to royalty payments and/or profit sharing on any future product sales. The substantial committed and contingent funding that the new alliance brings will transform MorphoSys’s ability to build value by expanding and accelerating its own drug development activities. In addition, MorphoSys will further build its drug pipeline via the co-development rights provided by the new agreement. MorphoSys remains committed to its ongoing proprietary drug development programs – MOR103 and MOR202 – and to the active therapeutic antibody pipeline built with its other partners. Source: Morphosys press release

Nov 29, 2007. Regeneron Initiates Major Global Collaboration with sanofi aventis to Develop and Commercialize Fully-Human Therapeutic Antibodies Sanofi aventis plans to increase its stake in Regeneron to approximately 19% TARRYTOWN, N.Y. – (BUSINESS WIRE) – Nov 29, 2007 – Regeneron Pharmaceuticals, Inc. (Nasdaq: REGN) and sanofi aventis (Euronext: SAN and NYSE: SNY) announced today that they have entered into a global, strategic collaboration agreement to discover, develop, and commercialize fully-human therapeutic antibodies utilizing Regeneron’s proprietary VelociSuite of technologies (including VelocImmune(R)). Sanofi aventis will also increase its ownership of Regeneron’s outstanding common stock from approximately 4 percent to approximately 19 percent by purchasing 12 million newly issued shares of Regeneron common stock at a price of $26.00 per share, subject to customary closing conditions including antitrust clearance. As part of the research agreement, sanofi aventis will make an $85 million upfront payment to Regeneron and will fund up to $475 million of research over the next five years. Sanofi aventis will have an option to extend the research agreement for up to an additional three years. Sanofi aventis will have the exclusive option to co-develop with Regeneron each drug candidate in the collaboration portfolio. Development costs will be shared between the two companies, with sanofi aventis funding drug candidate development costs up front and Regeneron reimbursing half of the development costs from its share of future profits to the extent they are sufficient for this purpose. The first therapeutic antibody to enter clinical development under the collaboration is an antibody to the Interleukin-6 receptor (IL-6R), which has started clinical trials in rheumatoid arthritis. The second is expected to be an antibody to Delta-like ligand-4 (Dll4), which is currently slated to start its clinical development in 2008.

Case Study 3

349

For any new product successfully developed as part of the collaboration, sanofi aventis will take the lead in commercialization activities and will consolidate the sales. Regeneron will have the right to co-promote any and all collaboration products worldwide. In the United States, profits will be shared equally. Outside the United States, profits will be split on a pre-determined sliding scale with sanofi aventis’ share ranging from 65 percent to 55 percent. In addition, Regeneron will be entitled to receive up to a total of $250 million of sales milestone payments when the collaboration achieves certain aggregate annual ex-U.S. sales levels, starting at $1 billion. Source: Regeneron press release

Dec 19, 2006. GlaxoSmithKline and Genmab Enter Global Agreement for HuMax-CD20 London, UK & Philadelphia, USA and Copenhagen, Denmark; December 19, 2006 – GlaxoSmithKline (GSK) and Genmab A/S (CSE:GEN) today announced a worldwide agreement to co-develop and commercialize HuMax-CD20™ (ofatumumab), a fully human monoclonal antibody in late stage development for CD20 positive B-cell chronic lymphocytic leukemia (B-CLL) and follicular non-Hodgkin’s lymphoma (NHL) and in Phase II for rheumatoid arthritis (RA). Under the terms of the agreement, Genmab will receive a license fee of DKK 582 million (approximately £ 52 million and approximately $102 million*), and GSK will invest DKK 2,033 million (approximately £ 183 million and approximately $357 million) to purchase, 4,471,202 ordinary shares of Genmab. The total potential value of this agreement, in the event of full commercial success, in cancer and various autoimmune and inflammatory diseases, could exceed DKK 12.0 billion (approximately £ 1.1 billion and approximately $2.1 billion), including the initial license fee and equity purchase, milestone payments, totaling DKK 9.0 billion (approximately £ 0.8 billion and approximately $ 1.6 billion) and expected development, commercial manufacturing and commercialization costs. In addition, Genmab will be entitled to receive tiered double digit royalties on global sales of HuMax-CD20. GSK will receive an exclusive worldwide license to HuMax-CD20 as well as any other antibodies with affinity for the CD20 antigen which Genmab may develop. GSK will also have an exclusive option to a CD20 UniBody™ to be developed in collaboration with Genmab. GSK and Genmab will co-develop HuMax-CD20. Genmab will be responsible for development costs until 2008, including costs of the two ongoing late stage oncology studies after which development costs will be shared equally between GSK and Genmab. GSK will be solely responsible for the manufacturing and commercialization of HuMax-CD20. Genmab will have an option to co-promote HuMax-CD20 in a targeted oncology setting in the US and in the Nordic region. Should this be undertaken, Genmab will also have the option co-promote Bexxar™ and Arranon™ in the US and Atriance™ in the relevant countries of the Nordic region. Source: Genmab press release

350

Case Studies

MABRx’s Goals First, senior management of MAB Rx wants to get attention by the international investor community, because MAB Rx’s share price has a mediocre performance since its IPO three years ago. Second, the CFO would also like to seize the opportunity to secure some cash for the operations of the next months or years, so he can delay the next capital increase to times when the share price is higher and current shareholders get less diluted. He mentioned that he would like to get additional US$ 50 mn over the next three years. Third, the CEO thinks that MAB Rx could grow significantly on the back of this license contract. He would like to structure it in a way that Pharma International performs clinical trials, funded by Pharma International, and keeps the rights of MAB Rx’ home market, i.e. North America. This way MAB Rx gains further clinical trial experience and can then become a fully integrated biotechnology company, doing work from research to development, approval, marketing, and sales. It is even thinkable that MAB Rx conducts the early clinical trials, Pharma International advances the funding, and MAB Rx ultimately pays back the advanced sum once the project is on market. Nevertheless, they only want to look at this kind of contract in a second step. The new deal should of course also cover the license payments that MAB Rx owes to Research Company. This license agreement has the following structure: Table 5.14. License contract with Research Company Payment

Amount in Comment US$ mn

Start Phase 1

1.5

For first indication

Start Phase 2

2

For first indication

Start Phase 3

4

For first indication

Start Approval Phase Launch Royalties Commercial Milestone

0

For first indication

10

For first indication

5% 20

Paid by MABRx to Research Company After having sold US$ 900 mn

Case Study 3

351

Deal Structure The deal should include an upfront payment and possibly also an equity investment. In the continuation MAB Rx receives milestone payments when reaching certain goals. The milestones should be split by indication. MAB Rx then receives royalties on net sales. On top of this Mr. Deal would also like to include some sales milestones to increase the total deal size. In a second step Mr. Deal wants to investigate an agreement similar to the Regeneron-sanofi aventis deal. MAB Rx conducts the early clinical development of MA-1204. Pharma International takes over the development of MA-1204 after successful phase II studies. MAB Rx and Pharma International share all development costs and revenues. But Pharma International advances MAB Rx’ development costs, which are offset against the first royalties. After successful clinical phase III MAB Rx leads the approval process in North America, Pharma International in the rest of the world. Pharma International and MAB Rx agree that the North American market and the rest of the world (ROW) have about the same sales potential. Nevertheless, the American market will probably have less approval costs and marketing and sales expenses due to its homogeneity. In view of the revenue and cost split Mab Rx receives royalties on Pharma International’s sales in ROW and pays Pharma International royalties on its North American sales. Table 5.15. Term sheet from Pharma International Payment Upfront Start phase 2 Start phase 2 Start phase 3 Start phase 3 Start approval phase Start approval phase Launch Launch Royalties

Amount in US$ mn Comment 20 10 5 15 10 25

In cash For first indication For second indication For first indication For second indication For first indication

20

For second indication

40 30 10% up to 250 12% up to 750 14% above

For first indication For second indication Paid by Pharma International to MAB Rx

352

Case Studies

Mr. Deal thinks of the possibility of co-development, giving up half of the license payments, i.e. to reduce the proposed terms by half, but keeping 50% of the rights, i.e. to split all future expenses and revenues 50:50 (except the license payments), and structuring the rest of the agreement according to the lines above. Term Sheet The team from Pharma International agreed on the structure of the contract. They propose to come up with a term sheet and then also to present the assumptions. Assumptions In view of a productive and friendly partnership, Pharma International agrees to disclose their assumptions that led them to the above terms. The development plan foresees that for each indication phase I takes two years, phase II three years, and phase III another three years. Review takes one year for the cancer indication and two years for Rheumatoid Arthritis. The RA phase I trial starts with a delay of one year. Table 5.16. Cost and success rates assumptions by International Pharma Cancer Costs (US$ mn) Success Rates Phase 1 Phase 2 Phase 3 Review

5.2 15.9 43.1 3.5

77% 59% 67% 85%

Rheumatoid Arthritis Costs (US$ mn) Success Rates 3.1 7.4 30.8 3.5

75% 35% 72% 88%

Pharma International uses published success rates for cancer and analgesics/anaesthetics and own cost estimates for the development. Pharma International calculates with overall launch costs of US$ 100 mn for each indication. These are distributed over 5 years, starting 2 years before launch with US$ 10 mn, then US$ 20 mn, in the launch year US$ 30 mn, the first year after launch another US$ 30 mn, and the second year after launch another US$ 10 mn. Cost of goods sold (COGS) is 10% of revenues. Marketing and sales expenses (other than launch costs) amount to 15% of revenues for the cancer indication and to 20% for Rheumatoid Arthritis. Patent protection can

Case Study 3

353

be extended in a way that finally the cancer indication is protected for 10 years after launch and the Rheumatoid Arthritis indication for 11 years. Peak sales are assumed to be US$ 700 mn for the cancer indication and US$ 500 US$ for RA. For RA Pharma International expects a lower uptake curve as for cancer. Cancer

Rheumatoid Arthritis

100% 90% 80% 70% 60% 50% 40% 30% 20% 10% 0% Y1

Y2

Y3

Y4

Y5

Y6

Y7

Y8

Y9

Y10

Y11

Y12

Y13

Y14

Y15

Y16

Y17

Y18

Y19

Y20

Fig. 5.3. Sales curves

Pharma International says that they have gone with the term sheet to the upper limit of what would be acceptable to them in terms of value and value share (calculated at a 12% discount rate). But they are open to suggestions to modify the term sheet. Going through the different assumptions Mr. Deal notices that there are substantial differences in various assumptions that have an impact on the value of the project: First, Mr. Deal assumes that the RA drug will have the same uptake curve like the cancer drug. Second, Mr. Deal has found some slightly different success rates for cancer and RA drugs, which he prefers to use. They are specifically for monoclonal antibodies. Third, Mr. Deal assumes that if the cancer indication gets through, the RA indication is more likely to get through as well, based on the good results in cancer as a general proof of concept for the compound. Finally, Mr. Deal thinks that the sales scenario is rather conservative, given other monoclonal antibody sales figures. He estimates that the peak sales amount to US$ 1 bn and US$ 750 mn for cancer and RA, respectively.

354

Case Studies Table 5.17. Success rates assumptions by MAB Rx

Phase 1 Phase 2 Phase 3 Review

Cancer

RA

69% 53% 75% 90%

63% 48% 67% 90%

Tasks Mr. Deal first must evaluate the current term sheet. He does this with Pharma International’s set of assumptions as well as with his own set of assumptions. In a second step, Mr. Deal thinks of possible improvements of the term sheet so it meets the requirement of MAB Rx, namely to get US$ 50 mn in the first three years, to get a good headline, i.e. a large sum of possible milestone payments, and to cover the license contract with Research Company. Finally, Mr. Deal wants to extend the contract into a codevelopment agreement as proposed by the CEO. He therefore modifies the terms accordingly. 1. Value the term sheet with Pharma International’s assumptions. 2. Value the term sheet with MAB Rx’ assumptions. 3. Improve the term sheet for MAB Rx without deteriorating it for Pharma International (value & value share). 4. Extend the term sheet into a co-development agreement.

Case Study 4 GlobePharm’s Perspective GlobePharm is a pharmaceutical company with a strong presence in specialist therapeutic areas such as dermatology, women’s health, urology, and ophthalmology. Typically GlobePharm licenses drugs and commercialises them through its sales force. The license agreements can be for already approved products, but GlobePharm’s business model also allows for partnering in late stage development. GlobePharm has recently acquired a cuttingedge manufacturing company and has now the ability to produce a wide range of formulations in-house which effectively allows them to increase their operating margin by producing just at half of the normal costs.

Case Study 4

355

The Opportunity OT-1 On a conference you have learned about Ophthatec, a privately held European ophthalmics company. Ophthatec is financed by a few high net worth individuals and was not really visible until recently. Ophthatec has developed a compound, OT-1, against neo-vascular glaucoma (NVG), an orphan indication. OT-1 is expected to provide an effective treatment option for NVG, which typically results in blindness. There is no specific or satisfactory treatment in place today. OT-1 perfectly fits into GlobePharm’s portfolio as specialty treatment. GlobePharm has already several products in the ophthalmology segment and would like to leverage on its already established sales force in this area. OT-1 has just completed a first clinical phase 1/2 study with 20 patients. Ophthatec’s CSO has presented the results at the conference you have attended. Now, Ophthatec is looking to partner OT-1. In an informal discussion with the CSO you have learned that there are two main reasons for this step. First, Ophthatec’s investors are not familiar with the life sciences business model and have become a little impatient. They are reluctant to invest more before finally seeing some revenues. Second, Ophthatec’s management is very well networked in Europe’s ophthalmologist community, but has no significant reach to the important North American market. Ophthatec’s management, which seems to be heavily invested in the company as well, would like to avoid further fund raising and consecutive dilution by putting the North-American rights to OT-1 for licensing. The license deal should cover all phase 3 trials and should also fund the launch of the drug in Europe. Development of OT-1 According to discussions between Ophthatec and the regulatory authorities you assume one pivotal clinical study should be sufficient. The number of patients is reduced and clinical development shortened (2 years vs. 3 years). This single study, however, must still address specific issues related to safety (as for any other NCE), efficacy and pharmaco-economics. OT-1 is then also expected to have a shortened regulatory review phase. The upcoming pivotal trial is a phase 3 trial. You and the team of GlobePharm believe that the reduced R&D program to date does not offer the same weight as full sized phase 1 and phase 2 trials. You have seen this already several times before and are therefore inclined to apply the success rates that you have elaborated together with an industry data provider.

356

Case Studies Table 5.18. Success rates for OT-1 Phase

Success rate

Clinical Phase 3 Approval

60% 85%

You estimate the costs for the pivotal trial to be in the range of US$ 15 mn and they expect approval 12 months after at a cost of an additional US$ 1 mn in North America and in Europe respectively. The Market for Neo-vascular Glaucoma Reviewing the literature you have found that the incidence of NVG is about 1.5/10,000. You therefore estimate that in Europe the patient population suffering from NVG is about 60,000 and in the US and Canada about 50,000. OT-1 will be a daily treatment over a period of 5 months at about US$ 20-25 per day. In Europe the price is 20% lower. Currently NVG is managed with invasive techniques (surgery) that are often associated with undesirable side effects. The relatively inconvenient formulation as a cream makes you assume that you will reach 35% of all patients despite of the strong rationale for the drug and GlobePharm’s position in the ophthalmic segment. Due to the superior administration and convenience the uptake is expected to be rather fast. Launching the drug will take about US$ 10 mn in North America. Market exclusivity lasts 10 years starting at approval. You assume a standard sales curve, with a sharp decline after patent expiry. 100% 90% 80% 70% 60% 50% 40% 30% 20% 10% 0% 1

2

3

4

5

6

7

8

9

10

11

12

Fig. 5.4. Sales curve

13

14

15

16

17

18

19

20

Case Study 4

357

Negotiation You have presented the opportunity to enter a partnership agreement with Ophthatec to the management board of GlobePharm. They agreed that OT1 would be a good fit with GlobePharm’s product portfolio and initiated a due diligence process. This due diligence revealed that GlobePharm’s recent acquisition of a manufacturing company allows to increase the operating margin from an estimated 65% to an estimated 70% for the OT-1. The clinical data seems to be compelling with the caveat that the phase 1/2 trial contained only 20 patients, but you consider this with your modified success rates for the pivotal trial and approval. All in all the management of GlobePharm wants to pursue this opportunity and entrusts you with the negotiation of the financial terms of the partnership. You have set up a meeting between the BD&L team of GlobePharm and the management team of Ophthatec. Ophthatec has repeated its interest to enter negotiations with you. License Terms The CEO of GlobePharm who worked previously at a multinational pharmaceutical company has set you a target value share of 65%, not willing to close a deal with less than 55% (calculated with GlobePharm’s assumptions, at 12% discount rate). He said that a 60%-40% value share would be typical for this kind of deal. Ophthatec has been clear that it does not want to further dilute its investors with an equity investment. The deal will therefore be based on upfront payment, milestones, and royalties. GlobePharm’ preferences lie clearly on a deal that also shares the risk between the two parties, i.e. that Ophthatec gets rewarded if the product actually reaches market. The upcoming development and launch costs for OT-1 would amount to US$ 26 mn over the next three years in North America. The two parties agreed that GlobePharm finances this. The CFO told you that he prefers not to spend more than US$ 30 mn on the product before knowing that it can be launched. Table 5.19. Costs for OT-1

Clinical Phase 3 Approval

North America

Europe

US 15 mn US 1 mn

– US 1 mn

358

Case Studies

It is Ophthatec’s intention to keep the European rights, but you would also be interested in licensing global rights to OT-1. In case Ophthatec would keep the European rights, this would be a co-development and you would expect Ophthatec to contribute to the trial costs. Ophthatec’s Perspective You are member of the management team of Ophthatec, a privately held European ophthalmics company. You are with other managers a founding member of Ophthatec. Together with a few high net worth individuals you have established Ophthatec four years ago to develop a compound with the generic name OT-1. The investors funded the company with US$ 6.5 mn at that time. Today each of the four founding members still holds 15% of Ophthatec, the remaining 40% belonging to the investors. OT-1 is a new treatment against neo-vascular glaucoma (NVG), an orphan indication. OT-1 is expected to provide an effective treatment option for NVG, which typically results in blindness and for which no specific or satisfactory treatment is available today. Ophthatec’s Situation Until now you have raised US$ 6.5 mn from private investors. This money has been used to acquire the IP of OT-1 from the University of Oeil, to finance the preclinical studies, and to conduct the clinical phase 1/2 trial. Each of the founders has a long-standing relationship with the investors, but the investors are not familiar with the life science business model. The relationship with your investors has become a little delicate because they are impatient to see their investments finally paying off. Recently, you have successfully concluded the phase 1/2 study with 20 patients for OT-1. The next step is to initiate a pivotal study, which then allows you to file OT-1 for approval in Europe and North America. The management team is very well networked in the European ophthalmologist community, but lacks ties to North America. Until now Ophthatec has not been visible as a drug development company, but was rather a hidden company working in its own labs. But the tense situation with the investors and the lack of a North-American network led the management team to the conclusion to explore the possibility of partnering the drug. The idea would be to out-license the NorthAmerican rights. The partner should fund the pivotal study. The upfront

Case Study 4

359

and milestone payments should provide enough financing for the filing and launch of the drug in Europe. Ophthatec’s Business Development As part of Ophthatec’s new strategy to actively look for partners the management team has decided to participate in a couple of conferences to present the clinical data. At such a conference you came across GlobePharm. GlobePharm has a strong presence in specialist therapeutic areas such as dermatology, women’s health, urology, and ophthalmology. Typically GlobePharm licenses the marketing rights to drugs and commercialises them through its sales force. The license agreements can be for already approved products, but GlobePharm’s business model also allows for partnering of drugs in late stage development. GlobePharm was recently in the news with an acquisition of a manufacturing unit, but nothing more has been disclosed. OT-1 perfectly fits into GlobePharm’s portfolio as specialty treatment. GlobePharm has already several products in the ophthalmology segment and could leverage on its already established sales force in this area. Development of OT-1 According to discussions between Ophthatec and the regulatory authorities you assume one pivotal clinical study should be sufficient. The number of patients is reduced and clinical development shortened (2 years vs. 3 years). This single study, however, must still address specific issues related to safety, efficacy and pharmaco-economics. OT-1 is then also expected to have a shortened regulatory review phase. The upcoming pivotal trial is a phase 3 trial. You have found a good reference for success rates for ophthalmic drugs in an issue of Nature Reviews on Drug Discovery, by Kola and Landis: Table 5.20. Success rates in ophthalmology Phase Clinical Phase 1 Clinical Phase 2 Clinical Phase 3 Approval

Success rate 67% 40% 63% 92%

360

Case Studies

You assume that the costs for the pivotal trial are about US$ 15 mn. Approval can be expected within 12 months after filing and costs in total about US$ 1 mn for each market. The Market for Neo-vascular Glaucoma Reviewing the literature you have found that the incidence of NVG is about 1.5/10,000. You therefore estimate that in Europe the patient population suffering from NVG is about 60,000 and in the US and Canada about 50,000. OT-1 will be a daily treatment over a period of 5 months at about US$ 20-25 a day. In Europe the price is expected to be 20% lower. Currently NVG is managed with invasive techniques (surgery) that are often associated with undesirable side effects. You expect to reach in Europe about 45% of all patients thanks to your strong network. You assume that GlobePharm as a key player in that area achieves the same market share in North America. Launching the drug will take about US$ 10 mn in each market, Europe and North-America. Market exclusivity lasts 10 years starting at approval. Ophthatec estimates that the operating margin will be 65%. Since the drug is a niche product you don’t expect generics even for some years after patent expiry. 100% 90% 80% 70% 60% 50% 40% 30% 20% 10% 0% 1

2

3

4

5

6

7

8

9

10

11

12

Fig. 5.5. Sales curve

13

14

15

16

17

18

19

20

Case Study 4

361

Negotiation After the conference GlobePharm has expressed its interest to license OT-1. After a due diligence of the clinical data the GlobePharm business development and licensing team proposed to meet with the Ophthatec management team in order to discuss the financial terms. Ophthatec has some clear constraints regarding the deal. First, your investors are unlikely to provide Ophthatec with more funds and you would like to avoid taking other investors aboard. Second, Ophthatec needs to pay milestones and royalties to the University of Oeil (cf. table 5.21). Third, you expect European launch to cost you about US$ 10 mn, which you would like to finance with the milestones payments received until then. Table 5.21. License terms with University of Oeil Term Start of pivotal trial Approval in first market Royalties

US 1 mn US 3 mn 3%

Table 5.22. Costs for OT-1

Clinical Phase 3 Approval

North America

Europe

US 15 mn US 1 mn

– US 1 mn

Finally, in order to satisfy your investors in the long run, you would need to present them a 30% – 35% return on investment pre-tax if everything goes well; this was the basis for the share allocation when you founded the company four years ago. The investors received 40% of shares at that time, this means that the investors paid US$ 6.5 mn four years ago and receive 40% of all future payments. Of course, if the project fails at some point, you would be happy to be able to pay back at least part of the investors’ investment. Therefore a license deal as early partial securitisation is a welcome option. At the same time you would also like to maximize the value of your own stake in Ophthatec. If you keep the European rights, then GlobePharm would expect you to contribute financially to the trial costs. You hope that you can cover this with the upfront payment or can negotiate that with GlobePharm. Otherwise you could also discuss a global license. Typically, you can expect 30%-50% of the value of OT-1 when calculated at 12% using the value split method.

References

AUTM (2005) AUTM U.S. Licensing Survey: FY 2004. (Survey summary of the AUTM U.S. Licensing Survey: FY 2004) Black, F. and Scholes, M. (1973) “The Pricing of Options and Corporate Liabilities” Journal of Political Economy, 81, 637-654 Burrill&Company (2005) BioCentury Report 2005 CDER (2004) Measures of Pharma Industry’s New Drug Output, 1995-2003 Ernst&Young (2006) Beyond Borders: Global Biotechnology Report 2006 Kaushik, A. (1991) “On the computation of continuous time option prices using discrete approximations”, Journal of financial and quantitative analysis, (26)4 Kaushik, A. (1995) “Option pricing trees”, The Journal of Derivatives, 2(4) Kola, I. and J. Landis (2004) “Can the pharmaceutical industry reduce attrition rates?” Nat Rev Drug Discov 3(8): 711-715 Lehman Brothers (2003) Assessing Patent Risk and Company-specific Exposure to Generics, 2003-2007e Ma, P. and R. Zemmel (2002) “Value of novelty?” Nat Rev Drug Discov. 1(8): 571-572 Markowitz, H. M. (1952) “Portfolio Selection” Journal of Finance 7, 77-91 McNulty, J. and Yeh, T., Schulze, W. and Lubatkin, M. (2002) “What’s your real cost of capital?” Harvard Business Review, October 2002, 114-121 Sharpe, P. and Keelin, T. (1998) “How SmithKline-Beecham makes good resource allocation decisions” Harvard Business Review, March/April 1998, 45-57 Villiger, R. and Bogdan, B. (2005) “Getting real about valuations in biotech” Nature Biotechnology, April 2005, 423-428 Villiger R. and Bogdan, B. (2005) “Valuing Pharma R&D – The Catch-22 of DCF” Journal of Applied Corporate Finance, Spring 2005

B. Bogdan, R. Villiger, Valuation in Life Sciences, 3rd ed., DOI 10.1007/978-3-642-10820-4, © Springer-Verlag Berlin Heidelberg 2010

364

References

Villiger, R. and Bogdan, B. (2005) “What’s biotech really worth?” Scrip Magazine, June 2005, 27-31 Villiger, R. and Bogdan, B. (2006) “Pitfalls of Valuation” Journal of Commercial Biotechnology, April 2006, 175-181

Index

510(k) 71 A Actelion 98, 279 Analysis of simulations 140 Anti-dilution clauses 272 Anti-stacking 149 Approval Phase 70 Arbitrage 60 Asset allocation 246 Astra Zeneca 98 AUTM 2 Avastin 280 Average drug Development Costs 74 Average duration of drug development 76 B Balance sheet 258 Basics of valuation Cash flows 11 Discounted cash flow valuation 29 Discounting 13 Real options 36 Valuation methods 29 Berna Biotech 98 Beta 18, 20 Bextra 243 Binomial tree 125, 163 Peak sales 47 Solving the end nodes 53 Spanning the tree 49

Working back the tree 55 Working back the tree to the root 57 Binomial Tree Probability distribution 53 Biologics New approvals 3 Biotech 3 Biotech companies Sublicensing 169 Black, Fisher 59 Black-Scholes formula 7, 59 Black-Scholes Formula 64 Brownian motion 62 C Calculation of end nodes 47 Calculation of generic conracts 179 Call-back option 149 Cancer drugs Multiple indications 218 Capital asset pricing model 18 CAPM 8, 19, 269 Case Study Sublicensing 203 Cash 262 Cash flow size 12 Cash flow statement 12 Cash flows Discounting 112 Probability 111 Uncertainty 21 Cash flows from financing activities 11 Cash flows from investing activities 11

366

Index

Cash flows from operations 11 Cash flows probability 129 Cash flows size 128 Cash flows timing 129 CE-mark 70 Change in liquidity 12 Clinical phase 1 trials 69 Clinical phase 2 trials 69 Clinical phase 3 trials 70 Co-development 149 Co-marketing 150 Company valuation 255 Accounting value 258 Asset based 258 Building blocks 260 Cash balance 273 Cash flow based 260 cash flows 260 Discount rate 269 Feed rate 265 Financial statements 263 Income based 260 Market based 260 Methods 258 Risk profile 273 Simulations 275 Value 273 Value development 273 Company Valuation Case study 275 Cash 262 Debt 263 Fixed assets 262 Fundraising 255 IPO 256 Licencing 256 M&A 256 Projects 261 Taxes 263 Technology 261 Unallocated costs 263 Company value development Change in discount rate 274 Clinical trial results 273

Comparison of DCF and real options 41 Comparison of participation rates 192 Comparison of real options resolution methods 44 Convertible bond 271 Convertible debt 271 Co-promotion 150 Copyrights 236 Cost 72 Cost of capital 15 COX-2 inhibitors 243 Cumulative Milestone 308, 319 Cumulative Sales 308, 319 D DCF 7, 11, 29, 30, 105, 109 Managment flexibility 37 Sensitivity analysis 130 DCF vs. ROV 129 Debt 270 Payoff profile 270 Decision Tree 32 Development costs 7 Dicounting 129 Dilution 256, 269, 272, 290 DiMasi 72 Discount rate 7, 132, 269 Discounted cash flow valuation 29 Discounting 96 Continuous compounding 14 Discrete compounding 14 Expected return 14 MCPM 97 Risk 14 Theory and Reality 96 Time value 13 Disease category average sales 78 Disease category median sales 78 Diversification 23 Drug development costs Approval 74

Index Clinical phase 1 73 Clinical phase 2 73 Clinical phase 3 73 Lead optimisation 73 Preclinical 73 Drug Development Duration 76 Drug development success rates 75

I

Efficient frontier 246 EMEA 69 Employee stock option 272 Equity 148 Estimation of peak sales 77 Excel Goal Seek 313 IRR 313 Expected Utility Theory 99

Income statement 263 IND 69 Inflation 16 Internal rate of return 31, 307 Investors 256 IP 2 IP valuation Accounting 236 Fundraising 236 Licensing 236 Litigation 236 M&A 236 IP valuation techniques 238 Cost based 238 Income based 238 Market based 238 IRR 135, 307

F

K

Fashion drugs 102 FDA 69, 70 Feed rate 234, 262 most likely 328 Financial option valuation 59 Financial options 38 Fund raising 5

Kaushik 229

E

G Game theory 249 Genentech 98, 280 Generic contract calculation 179 Generic contracts 181 Global biotech financing 3 Growth rate 47, 131 H Hamilton-Jacobi-Bellmann equation 64 Hedging 59 Hit identification 68 Hurdle rate 31

L Lead optimisation 68 Leverage 269, 271 License contract valuation 146 DCF 159 Licensee 159 Licensor 158 Manual for real options 162 Real options 160 Spanning the tree 163 Value share licensee-licensor 160 Working back the tree 163 License contracts Case study real options 163 Case study two parties 155 Cash flows 153 DCF valuation 152 Discount rate 154 Valuation with DCF 155

367

368

Index

License fees 148 License Negotiation 214 Licensing 5, 7 Biotech 151 Pharma 151 Universities 151 Value share 152 Life saving drugs 102 M M&A 5, 7, 290 Margin 111 Market asset disclaimer 64 Market capitalisation 278 Market portfolio 18 Market risk premium 18 Market-derived capital asset pricing model 20 Markowitz 18, 246 MCPM 8, 20, 97, 269 Medical device development 70 510(k) 71 Class I device 71 Class II device 71 Class III device 71 Premarket approval 71 Medimmune 98 Merck 98, 243 Milestones 147 Multi-dimensional trees 224 Multiple indications 218 Correlation of indications 230 DCF 219 Real options 223 Success rates 230 Multiple markets 232 N Nash, John 249 Negotiation Sublicensing 176 Negotiation deadlock 171 Net present value 30

Novartis 243 NPV 30 O Operating expenses 111 Out-licensing Sublicensing 201 P Partial differential equation 44 Patent valuation 236 Patents 236 Peak sales 7, 131 Indication 223 Peak Sales 77, 80 Bottom-up calculation 80 Peak sales estimate 48 Pfizer 243 Pharma 4 Pharma R&D 4 Phase 1V trials 70 Pipeline Correlation 243 Pipeline simulation 243 Pipeline valuation 243 PMA 72 Portfolio management 5, 245, 246 Qualitative approach 247 Quantitative approaches 248 Post-development volatility 102 Post-money valuation 257 Preclinical Testing 68 Pre-commercialisation 102 Prexige 243 Probability of cash flows 13 Project valuation 103, 107 Analysts 103 Business developers 103 Case study DCF 107 Case study real options 123 Cash flows 105 DCF 105 Discount rate 106

Index Investors 103 Portfolio managers 103 Project managers 103 Real option manual 121 Real options 121 R R&D funding 147 Ratchets 272 Reach-through royalties 148 Real option valuation 11 Theory 36 Real options 7 Balck-Scholes formula 59 Binomial tree 45 Binomial tree formulae 49 Binominal tree 124 Compound options 38 Difference to financial options 59 Finite differences 41 Formula 41 Management flexibility 37 Market conditions 37 Option to abandon 38 Option to expand 38 Option to grow 38 Option to switch 38 Options to defer 38 Project valuation 121 Resolution methods 41 Sensitivity analysis 130 Simulations 41 Strategic management tool 40 Trees 41 Valuation of end nodes 126 Volatility 100 Working back the tree 126 Redemption 147 Replicationg portfolio 59 Research institutions Licensing 169 Resource allocation 249

369

Right of first look 149 Risk adjusted net present value 30 Risk Aversion 24 Risk free interest rate 13 Risk profile 135 Risk-neutral probability 62 rNPV 30 Royalties 147 S Sales curve 129 Schering 98 Scholes, Myron 59 Sensitivity analysis Discount rate 131 Growth rate 131 Margin 131 Peak sales 131 Success rates 131 Sharpe 246 Sharpe ratio 18 Simulation Clinical success 136 Portfolio 135 Simulation of license contracts 194 Simulation of scenarios 138 Simulations 134, 275 Company value 135 Excel 136 Peak sales 136 Real options 141 Thresholds 141 Simulations in Excel 134 Stacked royalties 148, 194 Stackelberg game 249 Standard sales curve 109 Stock valuation 278 Scenario analysis 282 Stock Valuation 255 Structure of license contracts 146 Sublicensing 149, 169, 201 Calculation of participation rates 176

370

Index

Case study 203 DCF theory 173 Negotiation 176 Participation rate 175 Real option valuation manual 185 Sublicensing parameters 170 Success rate 7 Success rates 74, 132, 219 Discount 326 T Take-over 310 Target identification 67 Target validation 68 Taxes 266 Technology transfer 5, 207 Sublicensing 207 Technology transfer managers 173 Technology valuation 233 Terminal value 265 Terminal Value 17 The role of valuation 5 Time of cash flows 13 Time steps half-yearly 306 Trademarks 236 Tree for multiple indications 224

Underlying of the binomial tree 47 Universities 2 Upfront payment 147 US Licensing Universities 2 V Valuation Current problems 6 Input parameters 7 Interpretation of results 8 Methods and methodology 7 Valuation basics 11 Valuation in Life Sciences 67 Valuation of financial options 60 Value at risk 135 Value Development Discount Rate 273 Success Rate 273 Value Jump 23 Value sharing 152 Veletri 279 Vioxx 243 Virtual company 201 Volatility 47, 100, 124 Competition 102 General state of the economy 102 Regulation 102

U Uncertainty 21 Market Uncertainty 21 Technical Uncertainty 21

W WACC 15, 16, 261, 269 Worst case scenario 135