Correctly measuring drug development risk: a public policy imperative

Keywords::

• cost of capital
• financial risk
• innovation
• pharmaceutical R&D
• regulation

Responsible public policy must be informed public policy. This is particularly true with respect to the regulation and governance of the global pharmaceutical industry; an industry that has bestowed – through market-based investment incentives – innovations in medicine that have resulted in staggering societal gains in human health, life expectancy and quality of life. Before addressing the subject of drug development risk – the topic of this article – it is important to draw attention to the socioeconomic value of medical and pharmaceutical innovation. This places our policy imperative, on correctly measuring the risk of drug development, into an appropriate context; this is because it exposes the extraordinarily high cost, in terms of forgone innovation, that bad policy could inadvertently impose on society.

The economic value of innovation

Recent research has uncovered some striking empirical findings on the economic value of medical and pharmaceutical innovation. For example, Yale University (CT, USA) economist William Nordhaus has estimated the value of innovations in medicine during the second half of the 20th Century to be approximately equal to the gains in the economy’s real output, as measured by the Gross Domestic Product (GDP), over the same 50-year period [1]. The value of improvements in health (e.g., life expectancy), unlike the economy’s real productive output of goods and services, is not reflected in national accounting statistics (which, when aggregated, measure economic growth and national income – i.e., GDP). Nordhaus posits a simple, and indeed quite clever, question to demonstrate the intuitive reasonableness of his conclusion, which is based on highly technical research methods:

“You must forgo either the health improvements over the last half-century or the non-health improvements. That is, you must choose either (a) 1950 health conditions and 2000 non-health living standards or (b) 2000 health conditions and 1950 non-health living standards. Which would you choose?” [1].

Nordhaus has observed that choosing the latter or having a difficult time choosing between the two alternatives, implies agreement with his research finding: the value of innovations in medicine in the USA from 1950–2000 approximately equaled the value of GDP gains (national income) from 1950–2000.

As a second example, one that is prospective rather than retrospective, University of Chicago (IL, USA) economists Kevin Murphy and Robert Topel have estimated the socio-economic value of a 10% reduction in the mortality associated with cardiovascular disease and cancer at approximately US$10 trillion [2]. To place this number in perspective, note that the size of the US economy, as measured by the GDP, surpassed the US$10 trillion level a few years ago, in the early 2000s. The societal returns on investment in pharmaceutical R&D, in particular, are estimated to be substantially above their costs; in one econometric study, using data from 1960 to 2001, it was estimated that for every US$1345 invested in pharmaceutical R&D ‘produced’, on average, one additional life-year in the population/economy [3]. This finding, when coupled with the fact that most government cost–benefit and cost–utility analyses assign a value of US$50,000 or higher to a year of life in good health, suggests an unusually high social rate of return to investment in pharmaceutical R&D.

The extraordinary economic value attributable to medical and pharmaceutical innovation, both past and future (potentially), is why public policy affecting drug development investment incentives must be well informed, balanced and transparent. While special interest groups and political agendas are here to stay (which is not a bad thing per se, as we know from public choice theory), making inroads on flawed economic assumptions and analyses, which result in distorted characterizations of the economics of drug development, is imperative. An accurate picture of the economic dynamics of drug development: the costs, risks and returns, and the multitude of ways regulation affects these interrelated factors, is a necessary, but not sufficient, condition for good public policy. In light of the aforementioned research findings on the value of innovation, legislation based on erroneous assumptions, incomplete information or myopic perspectives, could have a deleterious effect on society’s access to future cures and treatments by stifling incentives for innovation.

One of the least understood economic factors affecting drug development, in both the public policy sphere and beyond, is risk.

Drug development risk

Drug development risk exerts an enormous influence on investment behavior in the pharmaceutical industry, yet despite this fact, in most policy debates, risk is perhaps the most misunderstood, misrepresented and mismeasured economic variable that significantly affects the level and direction of drug development and innovation. Indeed, even among some industry experts there is disagreement about the appropriate methods of measuring the risk of R&D.

Technical & financial risk

Most discussions concerning the risk of drug development refer to the probability of an investigational new drug advancing through the different phases of clinical development and gaining regulatory (i.e., US FDA) approval for marketing. In fact, some of the most cited industry statistics are about this type of risk. For example, the Tuft’s Center for the Study of Drug Development regularly publishes data on compound attrition rates by clinical development stage. These data suggest that only one in 5000–10,000 experimental drugs reach the market [4]. This is the technical risk associated with drug development, which is similar in many ways to wildcatting for oil because there are many ‘dry holes’ and only a few ‘gushers’.

It is often heard that pharmaceutical prices and profits are what they are because of this type of drug development risk. This is actually not correct (for several reasons). To be certain, the estimated probabilities of advancing through the different stages of drug development do affect a R&D project’s expected costs and returns, but this type of risk is quite distinct from the type of risk that matters most to investors who finance drug development (i.e., shareholders, bondholders and venture capitalists, among others). The risk that matters to investors is financial risk. It is this type of risk that we observe to be systematically underestimated by assumption or noncontemporary methods.

In finance, modern portfolio theory unequivocally demonstrates that the technical risk associated with the high attrition rates in drug development can be completely diversified away. These unique, or firm-specific, risks do not matter to investors holding the market portfolio. Instead, what matters to the financiers of drug development is the systematic, or market, risk of a project (or firm) with the market as a whole (market portfolio), which is essentially the covariation of expected project returns with the market (portfolio) and perhaps other types of systematic risk. The greater the systematic risk of a project or company’s stock, the higher the required expected returns must be to for financing to be forthcoming. This is the well-known risk–return trade-off in portfolio theory.

For investments over long-time horizons, such as drug development projects, financial risk may be amplified by something called leverage, which results in the financial risk being the greatest during the early stages of drug development. For all of these reasons, financial risk has a very significant effect on R&D investment levels and individual project progression decisions, which in turn affects the rate of pharmaceutical innovation. As we will discuss, the systematic underestimation of this type of risk, in both policy analyses and policy formulation, can have an unintended, chilling effect on the incentives for R&D investment.

Underestimating financial risk: consequences for policy formulation

In a series of published academic articles, US congressional testimonies and hearings, we have presented evidence of a systematic bias in the assumptions about, and the measurement of, the financial risk of drug development [5–8,101]. Specifically, we have found that the underestimation of financial risk leads to a mischaracterization of the economic costs of, and investment returns to, drug development. Moreover, because the investment horizon for drug development costs is long (12–15 years) and the horizon for investment returns even longer (>20 years post-product launch), the bias in financial risk measurement tends to significantly exacerbate these mischaracterizations.

A recent example of this was the legislative process through which the appropriate length of data exclusivity for originator biologic products was determined in the USA. The analyses upon which several hearings were based depended critically on assumptions about the level of financial risk associated with biologic R&D. Several papers relied upon in the legislative process assumed the financial risk of biotechnology R&D, known as the opportunity cost of capital, to be a real rate of 10% [Kotlikoff L, Unpublished Data; Brill A, Unpublished Data]. However, by using empirical research based on state-of-the-art models of risk and return and the most current financial data from the biotechnology industry, we estimated financial risk to be a real 13.25%. This contemporary and more methodologically appropriate measure of financial risk, suggested that a much longer period of data exclusivity would be necessary for innovator biotechnology firms to recoup average R&D costs.

In a related line of research, again using contemporary models and data, we updated the Tuft’s Center (MA, USA) estimate of the cost of drug development [5]. The effect of using more accurate measures of financial risk resulted in the cost of drug development increasing from US$802 million (2000 US$) to almost US$1 billion. This increase resulted from running the DiMasi et al. model with only a modestly higher estimate of financial risk; this was because of the long-time horizons of drug development, which increases the influence of financial risk (cost of capital) [4]. Over long-time horizons financial risk matters more than it does over short-time horizons. Drug development is, by definition, a long-time horizon investment proposition.

To provide an example of how sensitive drug development can be to policy initiatives, in one of our recent articles published in the Journal of Financial and Quantitative Analysis[6], we documented that the stock returns (adjusted for general market movements) of the most R&D-intensive pharmaceutical companies in 1992–1993 (when the Clinton Administration’s Health Security Act, which contained provisions to impose price controls on breakthrough new drugs, was being debated) were -90%. The link between stock prices and R&D investment is a direct one, both theoretically and empirically, as described in the article, and the hypersensitivity of stock prices to the proposed legislation in the Health Security Act emphasizes the considerable impact public policy can have on the private incentives to innovate in the pharmaceutical industry.

We have discussed a few examples of how biased measures of financial risk can generate misleading economic pictures of the drug development landscape. The systematic bias towards underestimating financial risk in policy analyses and debate has the potential to do much harm to social welfare, especially in light of the following two facts: the economic data on the costs and returns to drug development, which are debated widely in the policy arena, are extremely sensitive to assumptions about the level of financial risk; and the evidence on the substantial productivity of pharmaceutical R&D, and value of its subsequent innovative output, implies that the cost to society of poorly-informed (or systematically biased) policy analyses and formulation could be astonishingly high. Whatever direction future policies take the pharmaceutical industry in, it is imperative that careful consideration be given to R&D investment incentives, and that realistic measurements of (or assumptions about) the financial risk of drug development are employed.

Disclaimer

The views expressed in this article are of those of the authors and not necessarily those of the institutions with which the authors are affiliated.

Financial & competing interests disclosure

The authors have no relevant affiliations or financial involvement with any organization or entity with a financial interest in or financial conflict with the subject matter or materials discussed in the manuscript. This includes employment, consultancies, honoraria, stock ownership or options, expert testimony, grants or patents received or pending, or royalties.

No writing assistance was utilized in the production of this manuscript.