Managing Risk in Drug Development
The Case for FDA Swaps and Annuities
tomas j. philipson is a senior fellow at the Milken Institute, the Daniel Levin professor of public policy at the University of Chicago, director of the health care program at the university’s Becker Friedman Institute and a founder of Precision Health Economics LLC.
Illustrations by John Ueland
Published July 27, 2015
Profit in the global health care industry is highly concentrated in the United States, the largest market for pharmaceuticals and medical devices in the world. As a result, most health care innovators around the globe seek approval from the U.S. Food and Drug Administration. However, as is well known, the costs of gaining this approval are formidable. The average development cost for each drug and biologic (medicine derived from living organisms) is $1.2 billion, typically spread over about a decade of R&D, testing and review. Even more dispiriting, roughly seven out of eight candidates for approval never reach market.
It's widely agreed that rigorous testing is essential to ensure public safety. Plainly, though, the process creates substantial uncertainty for investors. I believe that hedging tools similar to those routinely used in financial markets could improve the risk-return trade-off, attracting investors to a field that would benefit from access to more private capital.
Running the FDA Gauntlet
In seeking FDA approval, innovators face two risks. First, they may never earn a penny if the FDA concludes that the innovation in question isn't up to U.S. standards of safety and efficacy. Second, the time-consuming process of regulatory review reduces investors' return because approval may be unexpectedly delayed by months or years, commensurately shortening the period in which the owner of the patent has market exclusivity. I propose creating financial instruments that would help to manage these risks by encouraging outside investors to share them with the innovators.
The FDA's approval process can be charted by several milestones:
- Filing an initial drug application;
- Safety testing in Phase I;
- Further safety and dosage tests in Phase II;
- Larger-scale efficacy testing in Phase III;
- Final approval of the new drug application, which is filed after all the evidence has been generated and assessed.
Drugs can be rejected at any stage – even after successful completion of Phase III testing. (A similar process governs approvals of biologics and medical devices.)
I consider two sorts of financial derivatives that could be used to hedge the risk of non-approval or unanticipated delays in approval. The first I call an FDA swap, which in many ways imitates the form and function of the credit default swaps already widely used to hedge credit risk in bond markets. The second I call an FDA annuity, which hedges against approval delays by paying the investors an agreed-upon sum during the life of the testing process.
The swaps would work as insurance against non-approval. The basic idea is straightforward. The buyer of the swap contract pays a monthly premium to the seller. If the drug is not approved (or if the product isn't resubmitted for consideration to the next stage in the approval process) by the maturity date of the contract, the buyer is paid an amount specified in the contract.
The devil is, as usual, in the details. The swap contract would also need to specify what would happen if there were a change in ownership of the product and whether the seller of the swap would gain possession of the buyer's intellectual property in the event of non-approval. Contracts, moreover, would need to precisely specify what constitutes non-approval. They would also need to specify how much would be paid, and when, if testing were prolonged. But these issues are hardly unique to FDA swaps. They are faced (and surmounted) every day in the creation of derivative contracts that are used to manage risks ranging from interest-rate variation to extreme weather.
FDA swaps would insure innovators against the risk of non-approval. By contrast, FDA annuities would serve as insurance against the risk that success in the approval process would take longer than expected. For small-molecule drugs (most often, drugs taken orally), the first three FDA phases of clinical development are estimated to take an average of about 24 months, 30 months and 42 months, respectively, with the final approval averaging an additional year.
However, as the figure to the right suggests, there is large variance, and thus developer risk, in approval times. The graph shows the share of products that have not yet been decided upon by the FDA, before and after passage of the Prescription Drug User Fee Act of 1992 (PDUFA I) and the amended Act of 1997 (PDFUA II), which permitted the FDA to charge fees that cover the cost of speeding the review process.
The risks inherent in these regulatory "survival curves" could be insured by annuity-like instruments, just as human survival times are insured by lifetime income annuities. For even if a product ultimately gains FDA approval, the statistical tails on approval times can create enormous variability in the rate of return on the developer's investment.
Consider the consequences of a delay of, say, six months for a blockbuster drug with likely monthly earnings of $100 million. The $600 million in lost earnings can never be recovered, because the patent expiration date remains unchanged. Note, moreover, that the losses are in near-term revenue, which is more heavily weighted in rate-of-return calculations because future income flows must be discounted to adjust for the opportunity cost of the investors' capital.
FDA annuities would insure investors against delays in the process, assuming the product ultimately received approval. The buyer of the annuity would make monthly payments to the seller for an agreed-upon period (or, alternatively, buy the annuity upfront with a single payment). After this accumulation period ended, the annuity would click on, just like a standard lifetime income annuity.
One Size Needn't Fit All
Given the many milestones along the way to a regulatory decision at which non-approval or delays can take place, one would expect that swaps and annuities would be tailored to specific testing and evaluation stages. For example, swaps might be written only for the phase a product is entering, or for non-approval in any of the remaining phases until approval.
Swaps might also be created for whole baskets of compounds being reviewed by the FDA. One could imagine, for example, that the entire pipeline of a given manufacturer might be hedged, not unlike the way credit default swaps are sometimes written on indexes or specific baskets of bonds. Consider, too, that securitized default obligations, similar to credit default obligations, could be created for these baskets of products.
If We Built It…
These instruments mimic many instruments used successfully outside of health care. Credit default swaps are like the proposed FDA swaps because lenders' defaults are similar to FDA non-approval. They were introduced in the early 1990s, and their use has blossomed in the decades since because they are so attractive for both hedging and speculating. In 2012, their reported aggregate notional value exceeded $25 trillion (yes, trillion) worldwide.
Credit default swaps are non-standardized contracts and are thus not traded on exchanges. However, markets for non-standardized FDA swaps and annuities could still prove to be quite liquid, just as markets for non-medical swaps are today. Indeed, one would presume that the stakeholders willing to participate in such markets for corporations engaged in medical R&D – especially portfolio mana-gers, hedge funds and pension funds, as well as pharma companies with competing or complementary drugs whose value could be affected by approval or non-approval – would also have an appetite for FDA risk instruments.
The corporate acquisitions of biotech innovators completed in recent years offer indirect evidence of the potential value of hedging contracts for medical innovations. In deal structures involving so-called contingent value rights, acquirers agree to make additional payments once the acquired companies hit specified regulatory benchmarks. For example, Celgene's $3 billion deal for Abraxis BioScience in 2010 included a contingent-value-rights provision conditioned upon regulatory approval of the cancer drug Abraxane. The terms of the $20 billion sale of Genzyme Corp. to Sanofi-Aventis in 2011 was tied to the performance of Campath, another cancer medication. An interesting aspect of the aforementioned rights is that they were listed separately on exchanges and proved quite liquid in trading.
Regulations to ensure transparency would be needed for both over-the-counter and exchange-traded hedging instruments. But much of the heavy lifting in this regard would likely be done by private enterprise with little official prompting. There are already firms (such as Claravant) emerging in the marketplace that rate the risks associated with pipeline medical products in much the way Moody's rates the bonds underlying credit default swap contracts. In addition, many financial institutions conduct their own surveys of independent experts to better assess development-related FDA risk.
Despite transparency issues, it is important to keep in mind that many markets exhibit substantial liquidity despite asymmetries in information, in which one party to a transaction knows a great deal more than the other. Indeed, issuers of stocks and bonds often know much more about the value of these securities than buyers, yet substantial liquidity exists in most initial public offerings.
FDA swaps and annuities may be particularly valuable for products in late-phase development because the acquisition of companies with early-stage pipeline products by large pharmas essentially serves as non-approval insurance for smaller biotech companies. On the sell side, the small biotech gets a fixed payment even if it fails to deliver valuable products later. In effect, the small biotech purchases insurance by giving up part of its potential profit to the big pharma buyer in exchange for a limited downside in case of non-approval. On the buy side, the big pharma company may acquire portfolios of potential drugs of which, perhaps, one in ten or so compounds succeeds. It thus acts like an insurance company, reducing risk through diversification.
Note, too, that market-making in medical R&D derivatives might assist broader functions. In a liquid market, trading prices serve an important informative role. Speculation may arise because of differences in opinion about risks affecting the value of a financial instrument. In addition, if public pricing of the instruments discussed becomes available, it will be useful for understanding how the market assesses regulatory risks, in much the same way that corporate debt yields compared to Treasury yields offer information about the markets' views on corporate defaults. FDA swap prices would be valuable in predicting future FDA rejection rates for the same reason corporate yields predict defaults. For single-product biotechs, there will be arbitrage opportunities for exploiting any mispricing of regular debt or FDA swaps.
Enter the Nonprofits
The liquidity of these instruments might be enhanced by the infusion of capital from third parties, notably from nonprofit patient groups and foundations that are increasingly taking on the role of equity investors rather than mere donors in the development of new drugs and devices. Such organizations are legally permitted to profit from investments. They differ from for-profit firms, however, in that they must reinvest the gains in ways that further their philanthropic missions. So-called venture philanthropy (in which nonprofits finance for-profits) may be particularly useful in boosting incentives for small firms.
For example, the Cystic Fibrosis Foundation sold the rights to its patented drugs, including Kalydeco, for $3.3 billion. This left the foundation in the enviable position of needing to find other productive ways to fight the disease. One potential approach would be to make an otherwise-illiquid market for medical R&D derivatives liquid by providing funding and eliminating any negative bid-ask spreads between for-profit parties. In other words, third-party nonprofits might choose to subsidize part of a swap or annuity purchase with the goal of creating a viable market.
Such third-party subsidies might even come from less-obvious sources within the biopharmaceutical industry. For example, new regulations might be written to allow pharmas to repatriate foreign earnings without incurring tax liability if the funds were used to eliminate negative bid-ask spreads on FDA swaps.
The Other Side of the Trade
The less variation there is over time in aggregate FDA approval behavior, the more attractive these instruments may be to counterparties. Life insurance offers a useful analogy. A policy has value for individual customers because it removes the financial uncertainty associated with mortality. This is true even though aggregate mortality rates may be certain, making them easy to absorb by counterparties offering the insurance.
Another characteristic of the market for FDA derivatives that increases the potential value to counterparties is how the risk relates to their other holdings. The holy grail of financial-portfolio investment is diversification that allows investors to maximize expected returns for the level of risk they deem acceptable. But true diversification is not easy to manage in an era in which the returns from the big, liquid securities markets are positively correlated. And it is likely to become even harder as the global economy becomes increasingly integrated.
By contrast, it is unlikely that approval behavior on the part of the FDA, driven largely by the biological reaction of new molecules in humans, varies with aggregate economic behavior, such as the business cycle. This lack of correlation with other asset classes implies that FDA derivatives ought to be especially attractive to investors seeking diversification in their financial portfolios. So in the process of reducing the real cost of medical innovation by making it cheaper to manage risk, medical R&D derivatives would offer an attractive way to tame financial portfolio risk to the parties on the other side of the transaction.
It is, of course, important to keep the primary goal here front and center: by making it easier to manage risk, derivatives could provide a much-needed boost to private investment in medical technology. But they could also provide an intriguing bonus in the form of a new resource for financial diversification in a world in which true diversification is becoming ever more elusive.