The Rationalist Conspiracy

There are a million examples of people not noticing the abundance of money, but one particularly egregious one was posted on Overcoming Bias a few days ago, which I will now proceed to refute. Original text below, in italics.

GiveWell’s charity recommendations – currently Against Malaria Foundation, GiveDirectly and the Schistosomiasis Control Initiative – are generally regarded as the most reliable in their field.

This isn’t wrong, of course, but one should note that, eg. Against Malaria Foundation has only raised about $1.5 million per year. Why is the #1 rated charity so small?  It’s because there’s no shortage of donations. Warren Buffett alone donates a thousand times that amount, and so most useful charities won’t benefit from an additional marginal dollar and won’t make the rankings. There’s a shortage of charities with enough visible human and social capital to be worth donating to.

I imagine many readers here donate to these charities. This makes it all the more surprising that it should be pretty easy to start a charity more effective than any of them.

All you would need to do is found an organisation that fundraises for whoever GiveWell recommends, and raises more than a dollar with each dollar it receives. Is this hard? Probably not. As a general rule, a dollar spent on fundraising seems to raise at least several dollars.

This confuses average return with marginal return. When Saudi Arabia pumps oil out of the ground, it costs them $3 per barrel. They then sell the oil for about $90 per barrel. That’s a 3,000% profit margin! Surely, then, it’s easy as dirt to get rich in oil?

But Saudi Arabia has all the easy oil fields. Those have already been found and claimed. To get rich yourself, you have to find a new oil field – one that’s much harder to find, and much harder to drill. Your cost is probably more like $50 per barrel.

This is the law of diminishing returns, and it’s ubiquitous in economics. The more of X you do, the more difficult the next unit of X becomes. Think of X as a rubber band. You can stretch it a little easily. The next bit is harder. The next bit is harder still. Pretty soon, you’re using your whole strength to keep it taut. The average bit of stretching didn’t need all your strength – if it did, you could never have stretched it in the first place. The average bit was pretty easy. It’s only that last bit that’s hard.

The same goes for fundraising, or pretty much any economic activity. The first bit of fundraising is really effective. The next bit is less effective. And so on. Where do you stop? In a perfect world, you stop where one dollar of fundraising raises one dollar of funds. There’s no point to doing any more. Hence, some outside entity doing more fundraising will not help a charity. For every dollar it spends, it’ll get back less than a dollar, leaving everyone worse off. Even though the average dollar of fundraising is helpful, the marginal dollar is not.

Of course, the world isn’t perfect. What if charities are, say, 50% inefficient? Then, you’d stop fundraising where one dollar raises two dollars of funds. One dollar will be wasted on inefficiency, and the other will be spent effectively.

But again, there’s no point to doing more. If one started a new fundraising charity, it would also be inefficient. For every dollar it spent, it would raise two. But then, it would have to spend one on inefficiency, and we’d be worse off. (In real life, as a new organization, it would always have to spend more on inefficiency, since it would need to pay for its own board and website and accountants and lawyers and whatnot. We’d be even worse off.)

There are many reasons for diminishing returns. Maybe the existing donors are tapped out, and it’s hard to find new ones. Maybe the fundraising manager already has eight employees, and can’t handle more. Maybe they can’t buy more clicks on the relevant Google keywords. Why does this keep happening?

It happens (in large part) because virtually every process has inputs other than money. Suppose you had a million bucks to start a fundraising charity. But who’s going to manage it? Who’s going to recruit the team? Who’s going to oversee it, and prevent people from stealing the money and running off to Latvia? What about accounting? Legal? Advertising? Marketing strategy? These things don’t grow on trees. After a certain point, money stops helping.

It’s a pretty simple and fast multiplier that obviously beats putting your money in the stock market.

The reason there’s no easy way to triple your money in the stock market is exponential growth. If you could triple your money every year, you’d soon have way more money than you could use on one stock play, and your returns would taper off. (And the next person who came along would find you already tapped out that play.) The same holds for charities. Even assuming people most are stupid, and don’t spend more on fundraising when fundraising has 3:1 returns, at least one nonprofit out of several million surely would. That one nonprofit would then grow until it had eaten all the easy fundraising opportunities. (I’m against efficient market arguments generally, but even the most diehard EMH opponents would agree you can’t make billions by just copying what everyone else already does.)

An independent organisation raising money for GiveWell’s top charities should do even better than a typical fundraiser, thanks to:

 – the strength of evidence, which is especially compelling to big donors

As noted earlier, GiveWell’s top charity has an approximate budget of $1.5 million. US charitable spending, by big donors alone, is tens of billions annually. Billions. With a ‘b’. Apparently, evidence isn’t as compelling to donors as it seems.

  – the independent recommendation, which looks particularly credible and removes the perception of any ulterior motive

Virtually every charity is recommended by someone. If it couldn’t even get that much, it would have no donors, and would soon go out of business. There are many, many recommenders, public and private.

 – a willingness to maximise (for example by targeting the wealthy, and focussing on regular or legacy donors)

Virtually every fundraiser already targets the wealthy. Charity balls are attended by the rich, not the poor, or even the middle-class. This isn’t some sort of strange coincidence – it’s a deliberate strategy by nonprofit managers.

 – an intrinsic motivation to do good

Even assuming, for sake of argument, that everyone associated with GiveWell had this “intrinsic desire”, how would anyone know about it? People’s motivations cannot be seen with binoculars. There’s a reason why virtually everyone has gotten scammed, at one point or another – scammers appear trustworthy.

And will motivation to “do good” make people work harder than self-interest? Every investment banker and bond trader I’ve seen works a lot harder than volunteers doing community service. No one ever snorted coke to keep awake at 4 AM so they could finish digging African wells.

 – the freedom to choose which of the three organisations they promote, depending on who they are talking to.

Virtually every charity already has mechanisms to screen prospective donors based on interests. It does no more good to talk to the 0.3% of the population interested in education/health/poverty, and then talk to them about education/health/poverty, than it does to just pull the 0.1% of the population interested in education and talk about that.

Putting your money into fundraising, rather than just giving it directly, does impose additional costs on the donors you inspire, and may ‘crowd out’ gifts to other charities. However, the logic of giving to GiveWell’s top rated charities is that they make (much) better use of money than most other individuals or organisations. So if you have a fundraising ratio significantly above 1:1, these downsides shouldn’t much matter.

Diverting lots of money from non-GiveWell charities to GiveWell charities would indeed be useful, if money were a limiting factor. But it is not. Even if 0.1% of US donations were to GiveWell charities, these three charities would rapidly have more money than they knew what to do with, and would get taken over by thieves. GiveWell selects for charities where a marginal $10 is useful, not ones where a marginal $10 million is.

You might ask: if fundraising is the best thing to do, why wouldn’t AMF, SCI or GiveDirectly just spend the money you give them on fundraising? My guess is that it’s simply a bad look. If they spend too much on fundraising, it will irrationally scare off their existing and potential donors.

If there are social costs to spending a lot on fundraising – which seems reasonably plausible – this would automatically make an organization which does nothing but fundraising a complete non-starter. If spending, say, 50% of income on fundraising scares off donors, despite a GiveWell top ranking, what would spending 100% of income on fundraising do?

Even if a charity should ideally spend most of its receipts on further fundraising in order to grow more quickly, the option simply isn’t available. The social norm against ‘optimising fundraising‘ is generally helpful, because intense competition between charities for donations would cause ‘rent dissipation’, and less total money would flow to charity recipients. But if your charity actually is much better than other charities, and so it’s good when you ‘take’ their money, this social norm does harm by preventing you from doing so.

This is a sociology version of a group selection argument. Way back in the day, biologists would say that such-and-such adaptation was for “the good of the species”. The problem is that this isn’t a Nash equilibrium. If everyone acts for the “good of the species”, at cost to themselves, each individual can benefit by acting differently. Sooner or later someone will do so, become better off, have more descendants, and their genes will dominate the population. So much for that.

Similarly, social norms do not arise to benefit society. Social norms arise because it benefits each individual actor to obey them. In the case of the ‘fundraising norm’, charities obey it in order to get donations, and donors impose it to prevent their money from being snuck away to buy new Ferraris. These forces will not go away just because a social norm no longer benefits the group. In particular, one can’t just say that one won’t spend the money on Ferraris; everyone already says that. One has to convincingly prove it, which is very difficult. Even if the charity is seemingly well-run, how do donors know you aren’t buying nails from your nail company at a 20x markup, and using that money to buy Ferraris? See, eg., the special “movie accounting” which “proved” Star Wars never made a profit.

So, if you are unlike most donors and are willing to have your money spent on effective fundraising, you can easily increase your impact several times over. Just help GiveWell’s top charities take their fundraising efforts ‘off the books’ by founding or giving to a separate organisation that does it for them.

This logic immediately falls apart, once one takes it a step further. If fundraising offers a 3x multiplier, why not fundraise to fundraise, and get 9x? 27x? 81x? Where does the tower of meta end? The answer, of course, is that doing things everyone is already doing (like generic “fundraising”) never has a 3x return.

This isn’t actually an impractical plan. Starting up a lean and effective fundraising organisation is difficult, but much easier than building a global team to distribute insecticide-treated bed nets.

Speaking as a serial entrepreneur, marketing is usually the hardest part of starting a new business, not the easiest. Marketing, fundamentally, is getting people’s attention, and people’s attention is very scarce because a) everyone wants it, b) the Internet and mass media allows anyone to compete for it, and c) it’s a fundamentally limited resource. Modern logistics makes globally distributing anything pretty easy, for someone dedicated and competent. But you can’t make marketing easy the way you can make logistics easy. If you could, (eg.) by inventing a new kind of advertising, there would suddenly be ten times as much demand for people’s attention as supply, and marketing would be as hard as ever. See also Top Ten Geek Business Myths.

Any bright and energetic person in a rich country who went and received the necessary training would have a decent shot at getting such an organisation off the ground.

There are roughly two million charities in the United States. All of their leaders are in a rich country, and most of them are at least relatively bright and energetic. Out of those two million, GiveWell recommends three. If all you know about a person is that they’re bright and energetic… well, three out of two million ain’t great odds. Even assuming GiveWell is missing dozens of effective charities, despite having spent more than five years searching… what fraction does that give you? 0.1%? 0.01%? This must be pretty hard if so many fail, relative to the abilities of those who try.

In summary: there is already plenty of money. There are plenty of people managing money (banks). There are plenty of people managing the managers’ money (mutual funds). There are plenty of people managing the managers’ managers’ money (funds of hedge funds). There are plenty of people managing the managers’ managers’ managers’ money (hedge funds themselves).

Fundraising is, fundamentally, a way of allocating money. We don’t need more towers of allocation and meta-allocation and meta-meta-allocation and meta-meta-allocation. We need places for the money to go at the end of it all. And those need social and human capital. Social and human capital, not infinite meta.