by Jason Crawford · May 14, 2020 · 3 min read
At the beginning of April I got obsessed with a new topic: how research is funded. The last several weeks, I’ve been exploring a lot of ideas and projects. In the spirit of working with the garage door up, here’s what I’m doing now.
My current focus is the history and present state of funding for research, especially but not exclusively “basic research”. My goal is to understand how research is funded, why it’s done that way, how the present landscape came to be, and where might be gaps or opportunities to do it better.
A word on terminology. There are many models of innovation, each with their own set of terms for different activities: pure science, applied science; basic research, fundamental research, exploratory research, uncommitted research, bench research, industrial research; discovery, invention, innovation; design, implementation, development, engineering; production, distribution, diffusion. So far, I haven’t come across any model that I love. But here are some basic distinctions I think are important.
First, there is a fundamental distinction between discovery and creation; between the pursuit of knowledge or understanding, and the attempt to make or produce something. Science is in the former category; invention, engineering, and business are in the latter.
Second, I think there is an important practical distinction between activities where enough is known that you can at least roughly predict how long they will take to produce a useful output, vs. those that are shrouded in enough unknowns and uncertainty that you have no idea how long they’ll take or even exactly what they’ll come up with—if anything. The latter is what I’m currently calling “research”, and it includes both science and what I’m thinking of as “invention”.
“Research”, in this definition, is what I think is particularly tricky to fund. It is funded today through both for-profit and nonprofit models, but neither is perfect. Research, by its nature, needs long and unpredictable time horizons. It can be hard to capture the value created from it, especially since a lot of the value is created by downstream applications when the results of the research are shared openly. These properties make it a bad fit for the for-profit model. But the extremely high-risk, high-reward nature of research means that we would ideally have a globally diversified portfolio of bets, which is a strength of the for-profit model.
On top of this fundamental challenge, there are indications that research funding may be in a suboptimal place today for historical or cultural reasons. The NIH, for example, by far the largest funder of health research in the world, has been widely criticized for being slow and risk-averse. The NIH’s budget this year, adjusted for inflation, is lower than it was in 2003. Grant applications, however, have continued to increase; with more applications chasing roughly the same number of research dollars, success rates have fallen from 30–40% in the 1970s to about 20% today. At the same time, grantees are getting older; the median age of a first-time recipient of an R01 grant (the NIH’s most common grant type) rose from about 36 in 1980 to almost 45 in 2010.
Beyond the NIH, there is evidence that health research more broadly is in trouble. The R&D cost to get a drug to market has been exponentially increasing, doubling every nine years since 1950, a phenomenon termed “Eroom’s Law” (“Eroom” being “Moore” backwards). And there is widespread talk of a funding gap in the pharma pipeline in between academic research and clinical trials, referred to in the industry as the “valley of death”.
In the middle of a century pandemic, the problems are even worse. Researchers don’t have time to write the long grant proposals required by government funding agencies; when a project called Fast Grants was launched to provide COVID-19 funding with a lightweight, low-latency process, it got over 4,000 applications in less than a week. Funding models may also be hindering vaccine development and ventilator production.
Funding models are crucial for progress, as of course is research itself. If research is inherently difficult to fund, then funding models for research—for science and invention—may be one of the highest-leverage topics in progress studies.
Here’s what I’m interested in researching and writing about in the near future:
Sources of research funding, both the current state and how we got here. This includes:
Specific fields and how they’re funded. I mentioned biomedical research above and specifically the pharmaceutical pipeline; what other fields are promising today? How is quantum computing research being funded, for instance?
Historical case studies. Just a few examples on top of my mind right now are the invention of the transistor at Bell Labs, that of nylon at DuPont, and virtually the entire career of Pasteur. What are other good ones?
The relationship of science to invention. Recently I’ve been reading about the basic-vs.-applied dichotomy or spectrum, including its intellectual roots going back to Plato; and the “linear model” of innovation. I’ve written about the relationship of science and invention in the past and have more ideas brewing.
(For those of you who were interested in my work on agriculture, my apologies—it’s been pre-empted by this topic, although I plan to return to it at some point.)
Comment on Reddit
Get posts by email:
Don't show again