Three researchers from MIT’s Laboratory for Financial Engineering, including Andrew Lo, have an intriguing new paper in Nature Biotechnology. They start off by noting that pharmaceutical companies are doing a very bad job at turning R&D expenditures into profits, and suggest an alternative: special-purpose “megafunds” of between $5 billion and $15 billion, which issue debt and equity securities, and then use the proceeds to invest in a diversified portfolio of a couple of hundred biomedical R&D projects.
There are a number of good ideas in here. First is the simple benefits of diversification. A $200 million research project with a 5% chance of success is a gamble. But if you bundle up 50 such research projects, even if there are some pretty strong correlations between them, your probability of success becomes much higher. You won’t make 60 times your initial investment, of course, as you would if you just invested in one project and it worked out. But investors tend to prefer safety to risk: they’re generally happier with a very high chance of getting a 10% return than they are with a very low chance of a 10X return.
On top of that, when researchers are working hard on a project which is going nowhere, there are often many institutional forces at play which keep that project going for longer than makes economic sense. If those researchers are part of a megafund which is developing some much more promising therapies, by contrast, they might be quite eager to move over to one of those teams. Especially when 15% of the profits from any given drug are set aside for the team which developed it.
And for all that the pharmaceutical industry has doubled its R&D spending over the past decade, from $68 billion in 2002 to $127 billion in 2010, there’s now a 20-year backlog of oncology compounds waiting to be investigated, and record numbers of medical and life-sciences PhDs who would love to be able to do those investigations.
Finally, the model of having a pharmaceutical company which funds itself with equity and debt and then which devotes some portion of its cash to R&D — that model doesn’t seem to be working very well these days:
The trend of increasing complexity and risk implies that the traditional financing vehicles of private and public equity are becoming less effective for funding biopharma because the needs and expectations of limited partners and shareholders are becoming less aligned with the new realities of biomedical innovation. The traditional quarterly earnings cycle, real-time pricing, and dispersed ownership of public equities imply constant scrutiny of corporate performance from many different types of shareholders, all pushing senior management toward projects and strategies with clearer and more immediate payoffs, and away from more speculative but potentially transformative science and translational research.
The translational part of the pipeline seems to be especially underfunded these days: while $22 billion was spent on basic research in 2010 and $125 billion was spent on clinical development, less than $7 billion was spent on the “translational efforts” which transform the former into the latter — something the Milken Institute calls “the translational valley of death”. That’s always been an area which appeals more to venture capital than to established pharmaceutical companies, but even the venture capitalists these days are concentrating more on later-stage drug development, not least because the sums of money needed to fund a wide range of translational projects, most of which will fail, are so enormous. Venture capital is all about return on investment, and the up-front investment needed in the biomedical space has been rising steadily, even as it has been falling sharply in technology. You can get monster returns on much lower investments in tech than you can in medicine.
So, what are the MIT people proposing? Basically, a special-purpose vehicle which would act as a “preclinical incubator” for early-stage projects and then shepherd them through to final approval, generating cash along the way by selling assets, if it needed to, to pay creditors. The vehicle would be funded by a mixture of equity and debt: the simple model in the paper has $15 billion in total capital, comprising $6 billion of senior debt yielding 5%, $3 billion of junior debt yielding 8%, and $6 billion of equity. If everything went according to plan, the senior and junior debt would get paid off in full, and the equity would ultimately realize annualized returns in the low double digits. Not exactly a home run, but still a decent uncorrelated long-term return.
The authors see this as a great time to try this kind of thing, because interest rates are so low that bonds yielding 5% and 8% respectively look incredibly attractive, especially if they’re uncorrelated to broader fixed-income risks surrounding the nexus of banks and sovereigns. What’s more, with “venture philanthropy” very trendy these days, it’s possible that the fund could raise a lot of money from foundations which want to invest their money in a way which could help ultimately provide the world with great new drugs which will save millions of lives. Some foundations might even step in to provide a guarantee on the megafund’s debt, which would bring the yields down substantially and allow it to raise even more money up front.
Still, I’m very skeptical that this idea is going to see the light of day, and the main reason is the sheer scale needed. In finance, as in most other areas of life, people want to start off small, and see if something works, before they scale up and go big. And this model, in particular, involves a whole new asset class — or, really, a whole new set of asset classes. Investors don’t like putting billions of dollars into something untried and untested, especially in a world where the returns to R&D spending have been declining steadily for many years.
It’s possible that the problem with R&D returns is a function of constraints that public companies have, and venture capitalists have, but which megafunds wouldn’t have. It’s possible — but it’s far from certain. And there has to be a pretty good chance that a brand-new megafund, making all the mistakes that a brand-new anything makes, will see zero or negative returns on its first go-round.
On top of that, the MIT model has both the senior and the junior bonds paying out coupons from day one, with the senior bonds repaying all their principal in years 3-4, and the junior bonds repaying their principal in years 5-6. That’s a lot of money to pay out in a relatively short amount of time, given that it takes a good 10 years for a drug to go from initial development through to approval. As a result, I fear that the megafund would be forced to sell its prime assets while they were still young, giving the biggest returns to the buyers of the assets rather than to the shareholders in the fund. And then there’s the wind-up date: the fund would have to liquidate after a certain amount of time, no matter how well it was doing, and no matter how crappy the market was for its assets that particular year.
There would also be very little alignment of interests between bondholders and equity holders in the megafund: the shareholders would want to hold on to everything, while the bondholders would want much more conservatism.
It seems to me that the permanent-capital model of a pharmaceutical company with an R&D department is actually better suited to the purpose than a megafund would be. Its “preclinical incubator” can be a permanent thing, rather than just lasting for a few years before being wound down. It can issue highly-rated unsecured debt on the strength of its cashflows from current drugs, and it can spend as much time and money as it takes to develop drugs, rather than being forced to sell half-baked projects at an inopportune time.
And while I’m sure that the MIT researchers did the best simulations they could with their “historical oncology drug-development databases with over 700 compounds in various stages of preclinical and clinical development from 1990 to 2011″, I have very little faith that the past performance of drug-development dollars between 1990 and 2011 is a particularly good guide to what one might expect going forwards. After all, there were two big spikes in biotech returns over that period, the larger one being the dot-com bubble. Since 2001, returns have barely been positive, and I’d love to see the results of the simulations if they used just 2001-2011 data rather than 1990-2011.
In short, while there might be a certain amount of interest out there for a modest initial attempt at this kind of thing, the fact is that a modest initial attempt wouldn’t be good enough. In a world where a successful drug needs over $1 billion in development costs before it comes to market, and where 95% of bright ideas eventually fizzle out and die, a megafund really would need to be mega-sized in order for the diversification benefits to really kick in. And while Andrew Lo is a very good salesman, I don’t think even he could raise $15 billion or $30 billion for this, or any other, experiment.