One hypothesis I’d like to investigate is that we’ve gotten a little too centralized and monolithic in our funding. That is something to address. The NIH today, budget over $40 billion, the largest funder by far of life sciences in the world, and they fund more than half of their grants through this particular R01 process. Regardless of what you think about that process, and there are arguments that that process is short-term focused, project-based, risk-averse, prone to group think. There’s some good reasons to be worried about those things. Even if you set those aside, should so much of our funding be going through a singular mechanism like that? No matter how good that mechanism is, isn’t it the case that that mechanism might be prone to blind spots?
Patrick Collison suggested in a tweet somewhere, what if we took the funding and split it into ten or fifteen agencies and somehow incentivize them to deliberately pursue heterogeneous approaches? Almost like, I don’t care what you do, just do something different! Don’t all do the same thing. One of the advantages of the for profit world, if you’re going to seek venture capital for a for-profit idea, there’s no one VC that you have to get on board. There’s a dozen or more top prestigious VCs and there’s hundreds that you could go out to that would write you a check. Thousands maybe. There’s all these different ways of funding a venture, and you only have to get one of them to say yes.
That doesn’t mean you have all the diversity in the world. There might be some cultural homogeneity among those VCs, but they are incentivized to spread out in idea-space, in the space of basic premises, and even the space of models and mechanisms for sourcing models, how do they source their deals? In the for profit world, investors are actively incentivized to be contrarian. By the nature of competition for ideas, and competition for deals, they’re incentivized to be right in a space where everybody else is wrong. If you think of a bunch of positively charged particles spreading out in space and not clumping together, there’s an incentive for for-profit investors to do that.
There’s also an incentive in the for profit world for investors to get in early. This is a mechanism we have in for profit investing that we don’t have in non-profit funding, which is you get highly outsized returns for being right early. I don’t think anything comparable exists, or has ever been devised, in the non-profit world.
That was Jason Crawford on the Venture Stories podcast, “Progress Studies in 2020.” Listen to the whole thing, fascinating throughout.
I completely agree with Jason, and think that the problems he suggests are rampant in the DoD. While there is one funding mechanism for more than half of the NIH’s $40 billion, there is just one funding mechanism for the DoD’s $237 billion FY21 request for RDT&E and Procurement. That, of course, all comes through the singular Planning-Programming-Budgeting-Execution (PPBE) process, and it basically bakes in another $230 billion in O&M funding as well. Any criticism of monolithic, risk-averse, project-based funding you can levy against the NIH is probably an order of magnitude larger and more sclerotic in the DoD.
One could potentially say that the DoD budget is really defined at the top levels not of the DoD but by the military components (e.g., Army, Navy, Marines, Air Force, MDA, SOCOM, etc.). That’s true. But it still must follow the exact same PPBE process which requires the exact same sorts of justification (e.g., lifecycle cost estimate, affordability analysis, validated requirements, acquisition strategy). Moreover, it must pass through OSD CAPE, Comptroller, DepSecDef and SecDef, OMB, and then four congressional committees.
Let’s say the Space Force wanted to change it’s funding mechanism so that instead of being project-focused it wanted to be portfolio-focused. Moreover, let’s say in doing so it wants to change the specific projects/programs being funded to move away from legacy 20th century platforms. Sorry. There’s simply no way that could happen. And that’s why the Space Force asked in its retracted May 2020 report that Congress allow it to consolidate program budgets into portfolios. It needed before-the-fact approval. The House Appropriations Committee basically said, take a hike.
Jason also makes the extremely important point that in private investing, there’s a huge incentive to spreading out in the idea-space to tackle projects that no group of bureaucrats would reach a consensus on. Ultimately, those are the ones that have the highest potential for 100x or 1,000x gains. While VCs will take the risk in order to achieve those gains, government officials simply cannot. Even if you could get it funded through the PPBE (which, you could not), inspector generals, the GAO, and any number of oversight agencies will descend upon the project with criticisms that it isn’t 100% likely to succeed. Case in point, the ABMS program.
So there’s zero benefit to investing in the right technology early if you’re a government official. There’s only risk to your career. And so the DoD is stuck with small experiments in the labs, while the only scalable project has a super-high TRL and thus cannot contribute to leaps in capability.
Could we devise a method to get government officials to invest early in technologies and take risks to capture those 100x or 1,000x rewards? Well, as a society we’ve basically decided that government officials shouldn’t make a lot of money. So it can’t be monetary reward, whether that is in pay. Air Force acquisition executive said he wanted to see more models where government takes equity stakes in companies it invests in. But then no individual can financially benefit from success. What other models are there?
Leave a Reply