To the Editor:

As a CEO and former investment banker specializing in securitization, I noted with interest the paper by Andrew Lo and colleagues1 proposing the application of securitization techniques to fund biopharmaceutical R&D. Such an approach promises to both enlarge the investor base willing to back breakthrough research and create a new asset class, thereby increasing diversification across the financial system.

Before embarking on such a path, however, we should all heed Bernard Munos's cautionary note: no financial engineering shall fix the root cause of the industry's failed innovation model2. The creation of an efficient market requires the ability for investors to gauge the quality of the underlying assets to be securitized. Failure to gauge such quality may result in the swift drying up of liquidity, as was observed during the 2008 credit crunch in the market for securitized mortgages.

Furthermore, Lo and colleagues1 provide expectations for returns and the ability to service the debt on the basis of data about past attrition rates, costs and revenues. The payer-driven market's shift away from buying products to buying outcomes will increase commercial risk in a way that may take years to measure and understand. Additionally, default risk is highly sensitive to correlation assumptions between R&D programs, which are notoriously hard to pin down.

For megafunds to become a staple of the biopharmaceutical industry's landscape, a rethink of R&D is necessary.

New approaches based on mathematical modeling of diseases to predict a compound's efficacy over carefully characterized target populations should be encouraged as a way for investors to gauge assets' risk-return profiles. Appropriate methodological tools to measure outcomes in real life, such as the Effect Model Law3 must be implemented. Finally, mathematical models have the potential to vastly expand the field of potential druggable targets, thereby reducing correlation risk and making it possible to reap the benefits of diversification.