Grant givers need better track records

Giving grants, by midjourney

There’s recently been a blizzard of capital gracing the world of philanthropy and experimental research funding. And with this glut of capital flowing into the space is the increasing professionalisation and the need for talent. Whether that’s EA in general or the myriad foundations sprouting out in particular, there’s a demand for talent to staff these organisations and to run them well.

The new organisations are often structured in order to not be bogged down by the needs for legitimacy and analysis paralysis that a lot of the older philanthropic organisations have. They are more comparable to venture capital firms, fully fascinated by the power law and aiming to ensure that projects with potentially huge outcomes don’t lose funding because the proven probability of success is small and/or illegible.

And as grant organisations have started embracing the power law world, there is an inevitable consequence. Which is the fact that the VC model of grantmaking should inherently be more forgiving of variance than they are today.

The paradox of teaching entrepreneurship is that such a formula (for innovation) cannot exist; because every innovation is new and unique, no authority can prescribe in concrete terms how to be more innovative. Indeed, the single most powerful pattern I have noticed is that successful people find value in unexpected places, and they do this by thinking about business from first principles instead of formulas.

Much of VC is open minded on the startups it funds, with the sole criteria being that it returns the fund. And therefore they have to be quite open minded in terms of what they fund. Being opinionated can quickly become suboptimal, especially because what creates success is unknown at the earlier stages. But this only works because what is or isn’t suboptimal can prove itself easily.

Which brings us to the bigger problem. This is that the VC model of grantmaking doesn’t have the most important feature of VC investing - that of an objective return metric that makes the reward calculable and therefore the risk worth it.

The benefit of high risk investing like venture capital is that since the outcomes are all measurable, you can easily draw upon the powers of the power law distribution to enjoy the returns.

This implies two very strange rules for VCs. First, only invest in companies that have the potential to return the value of the entire fund. This is a scary rule, because it eliminates the vast majority of possible investments. (Even quite successful companies usually succeed on a more humble scale.) This leads to rule number two: because rule number one is so restrictive, there can’t be any other rules.

But what’s the equivalent for the new grant making institutions?

I wrote about Thiel Fellowships where the outcomes were extraordinary. And yes, in this case Thiel himself does get a fair amount of credit. Deservedly so. But he wasn’t alone, right? Just like Founders Fund isn’t all about Thiel, but have other great investors who forged their own paths, the Fellowships weren’t about just Thiel either.

Some of the alumni like Danielle Strachman and Michael Gibson went on to create the 1517 fund, but this is the illegible benefits helping them out. If you think this is fine and enough, imagine what would’ve been the case were the Fellowships were 5x larger! Danielle and Michael would have had difficulty finding attributable investments that they can point to as the ones they sourced, diligenced, and invested in, with nary a clear ROI to show how good they are at their job!

What happens when a junior associate who worked her ass off for a couple years tries to go do something else, but their expertise remains unclear?

I think a major problem with grant giving is exactly this. That even if the titular head gets the credit, it doesn’t trickle down easily. Which makes hiring, retention, and career development much much harder.

But it should! It’s just as hard to source Stripe at seed stage as it is to find and fund Vitalik. Harder even. So why isn’t this given higher credence?

When Emergent Ventures does its incredible human capital investments, Tyler Cowen is rightfully judged as the talent behind it. But were he to expand the team, that team wouldn’t be able to get the same attributable benefit of having done well.

When Fast Grants (also Tyler, but also Patrick Collison) funded the major interventions of the fight against Covid, amongst others, the same problem applied.

This needs to change. If one were to scale up philanthropy, grant-giving, or indeed any method of capital allocation, its only fair that the capital allocators who work towards it have the recognition of their efforts in a relatively unambiguous fashion.

This stands especially true today, now that we have true behemoths in the grant-giving arena trying to shake things up. Whether that’s EA as a whole, the FTX foundations new grants, or Open Phil scaling up its giving, they all have to contend with this issue.

Why do they have to contend with the issue? Because all these organisations are reliant on the same hiring principles that a venture capital firm would need to do. Find those people who are able to identify non-consensus but right projects (or people). Find those people who are able to invest, in other words.

People take jobs because they either agree with the cause and the impact its likely to have, or the compensation structure appeals to them. In the case of venture capital sometimes it coincides. In the case of someone doing grants for Open Phil its mostly the former. But the skills aren’t all that different. What’s different is that the skills you learn in one comes with a credible track record while in the other it comes without one¹.

So ... how can we do this?

To recognise the fact that someone did help find, source or award a grant to a worthy person/ organisation/ cause is to reward their ability to select someone. Maybe we should create an immutable record of people’s selections, done by each organisation, in order to help the teams working on these have a proven record of the stuff they did within the org.

Did you fund a malaria nets program in [insert country] that had a 10x efficacy? Great! Did you help fund a career development grant to someone who ended up the top cello player in the US? Also great. Did you help fund an organisation which saved several thousand lives by helping evacuate people from Ukraine? Great. Did you fund an organisation that made progress (measured any way you like) in the push against AGI x-risk? Great!

In all these instances the end results are fundamentally illegible, and in many cases incomparable. But the skill to identify and fund projects which have such results are not. Those skills are very much the same as what Buffett calls his capital allocation skill.

The ultimate utility for something like this might end up being similar to the “create a portfolio of your work” end of things. This is pretty common in the more artistic realms, from painting to architecture, and increasingly common in coding (if only via github).

It’s also what’s been thought of as the basis for the replace linkedin or replace CV moves in web3, where ideally the things you learn will reflect in an immutable record of some sort that is both more granular and will follow you around, getting updated over time. The learn to earn program starts to go in this direction.

Both of which are fine and good, as ideas go. The difficulty of parsing someone’s experience, whether educational through transcripts or professional through CV, has been about the surfeit of information vs the pure signalling effect. If someone says they did computer science at MIT, even without knowing all the courses and projects and hackathons I get a sense of their calibre. Additional information here is likely to Berkson’s Paradox me rather than lead me to the right answer.

But capital allocation is an area where this is directly relevant. Picking winners is the whole job. Which means having a record of what you picked would be immensely valuable. Whether that’s a token provided by the grant-giving-organisation itself, or a secondary evaluation, it would be immensely valuable for those working in these jobs. After all, if someone wanted to move back and forth between markets it would be a highly legible symbol of their prowess, and we should encourage that!

And if someone wants to go meta and use this essay as the reason to build the tool to build the legibility to help talent blossom, I’m happy to take the credit.

1. There of course are large organisations which have had tremendous success through actual effort. The efforts of Gates Foundation to combat malaria is a case in point, or their efforts to distribute polio vaccines effectively.

But these are not grant givers, they are entire organisations. Though they too have an issue with the outcome not being directly attributable to some people's efforts, this is not nearly as easily solvable as the capital application question.

For these jobs are more akin to corporate jobs. Within a corporate while many people drive crucial components of any project, their individual impacts are not nearly so easily disaggregated.