This post is more than 11 years old

It’s common to debate how “efficient” financial markets are. Broadly speaking, an “efficient” market is one in which the participants are quick to spot profit-making opportunities, so that prices quickly adjust to reflect available information and it’s very difficult for an outsider to “beat the market,” i.e., consistently earn outsized returns. (If one is interested in more details, I recommend searching for discussions of the “efficient-market hypothesis.”)

Not all opportunities to accomplish good are opportunities to make money. Yet I believe that it’s useful to use a similar concept in the world of philanthropy – to speak of a “broad market” in which people are rewarded (though not always financially) for accomplishing good, and in which we therefore have some reason to expect that (a) the “easiest” opportunities for accomplishing good will get funded and carried out without our help; (b) finding opportunities to do good that aren’t already funded will be challenging.

The question of how efficient the broad market is (and how its efficiency varies from domain to domain) seems to me like a crucial one, and it is a question that we are constantly debating and looking for new information on. It’s equivalent to the question of how effective the world as it exists today is at “scooping up opportunities to do good,” and thus how difficult we should expect it to be to find outstanding such opportunities. When we started GiveWell, the seemingly low quality of dialogue around giving led me to expect extremely low efficiency; over the years, I’ve substantially (though not overwhelmingly) raised our estimate of how much “efficiency” already exists.

The remainder of this post will:

• Discuss why I’ve raised my estimate of “broad market efficiency” and what my current working view is.
• Elaborate on why one’s estimate of “broad market efficiency” is important.

First, a point of clarification. I generally view “market efficiency” as a spectrum: the more efficient a particular market is, the higher the level of intensity and intelligence around finding good opportunities, and therefore the more intelligent and dedicated one will need to be in order to consistently “beat the market.” The most efficient markets can be consistently beaten only by the most talented/dedicated players, while the least efficient ones can be beaten with fairly little in the way of talent and dedication. This way of discussing market efficiency contrasts with the “Are markets efficient, yes or no?” approach that some others take to the question.

How efficient is the “broad market?”When we started GiveWell, my basic picture of the world was that

• For-profit markets are relatively efficient. Therefore, if there’s an opportunity to do good that’s also an opportunity to make money, it’s very likely that someone will seize on this opportunity relatively quickly.
• Nonprofits accomplish good as well, but in the nonprofit world, funding almost exclusively chases “good stories” rather than good ideas. Therefore, there ought to be many opportunities to accomplish enormous amounts of good via nonprofit activities with a strong analytical case but no competitively compelling emotional pitch.

I still hold both these views to a degree, but several observations have complicated the picture for me.

• As we recently wrote, it appears that the most proven cost-effective interventions are often able to attract funding from major funders. While I do believe we’ve found some places in which more money is needed to deliver more proven cost-effective interventions, doing so has been far more difficult than I expected.And while I believe there exist good opportunities to fund more research and strengthen health systems (which could lead to more opportunities to fund delivery of proven cost-effective interventions), I also have not seen a large amount of obvious low-hanging fruit (outstanding, unfunded giving opportunities) in these areas. In particular, many of the interventions whose effects are easiest to study are the ones that have already been fairly well studied; to create good research on other interventions could be much more difficult. I believe there are exceptions to this pattern, but that it does hold as a general pattern.
• In 2012, we developed an interest in meta-research that grew out of our experiences reviewing evidence, and we noted at the time that we couldn’t identify any major foundations working in this area (based on our review of major foundations). We saw this as a promising but potentially ignored area. However, the more we investigated it and got involved in it, the more we saw that there are already many movements afoot that can be categorized as “meta-research” – and several major funders that are interested specifically in this area. Furthermore, it often appears to be the case that “meta-research projects” get funded by funders who don’t explicitly focus on “meta-research,” but instead focus on the field in question. (A future update will give more details on this.)The other funders focused on meta-research are roughly as new to the area as we are; we have uncovered some ideas that may qualify as “promising and un-funded”; and I still have the intuition that this is an extremely promising and under-resourced area. But it’s instructive that our best guess for a “promising but ignored” area turned out to have a non-negligible amount of interest from funders, once we looked more closely.
• We’ve also seen various “global catastrophic risks” highlighted to us as causes that are highly likely to be neglected, because the scenarios in question are highly “farfetched” and low-probability, and philanthropic investments are unlikely to pay off in any tangible way. Thus, it seems that funding these causes might require a highly analytical and genuinely altruistic bent. We’ve done some light investigation of global catastrophic risks, and have found some substantial activity on some of the more “farfetched” ones. For example, on asteroid impacts, tens of millions of dollars are being spent on asteroid detection and the expected lives lost due to such impacts has reached a fairly low level. On threats from epidemics, bioterrorism and other biological threats, the Alfred P. Sloan Foundation ran a program (archived link) for several years in this area, and reports having closed the program after seeing substantial increases in government funding (as a side note, we don’t find its causal attribution compelling): 
  “Sloan’s Biosecurity Program has been very successful in bringing attention to the issues and challenges posed by biological threats. When our program began in 2000, the US government funding for strictly biodefense was approximately $50 million. The FY2010 budget is $1.09 billion.”

   

   

• We’ve generally struggled to find proposals for projects that are fully fleshed out and immediately compelling, yet un-funded. We believe this is largely because of the dynamics of the philanthropic sector in which proposals tend to be written only after there is strong interest from a funder, but in a less efficient “broad market” we’d have expected to find some strong, exhaustively argued yet un-funded proposals by now.
• In general, in interacting with major foundations (such as the Bill and Melinda Gates Foundation, Hewlett Foundation, CIFF and Open Society Foundations), we’ve encountered many program officers whom we perceive as well-informed, genuinely altruistic, and relatively analytical in their approach. It’s difficult to evaluate the quality of these people’s work, but they certainly don’t fit the stereotype of the “story-oriented donor.”

In general, it seems nearly impossible to find a promising area or idea that is completely ignored (though many may be underfunded relative to other areas and ideas).

It seems to me that there is a real sense in which people are “rewarded” nonfinancially for working on something with high altruistic value – whether through recognition from others, through pride in doing well at the job they’ve been hired for, or of course through altruism – as well as a real sense in which people are drawn toward areas that they perceive as neglected (foundation staff have cited the latter factor to us in many discussions). Even without any particular person taking a bird’s-eye view of the philanthropic world and making analytical, strategic calculations to find the most under-explored causes, an ecosystem with a large number of people who have various drives toward altruism and toward working on what’s neglected can be expected to produce some degree of “broad market efficiency” – a tendency to find and execute on the most outstanding and neglected ideas.

I am not saying that “all the good giving opportunities are taken” or that “philanthropic capital is allocated as well as it could be” – I very much don’t believe that. I believe that a great deal of giving is done with very little thought and goes to causes that are far inferior to the best ones out there; I believe that strategic cause selection is itself a neglected approach, and that it will lead to far more impact than we could achieve otherwise. Even if (as isn’t necessarily the case) the best $10 million worth of funding opportunities for a given cause are already funded, funding the next $10 million down in the most promising causes (as opposed to causes chosen with little reflection) could have enormous value. In addition, I believe that looking at existing giving opportunities doesn’t tell you the full story about a cause’s potential (as discussed previously), and raising the profile of the more important causes could generate outstanding giving opportunities that haven’t yet been surfaced.

What I no longer believe is that there’s any easy way to tell which areas are under-funded. All it takes is one or two idiosyncratic major funders to turn a cause area from under-funded to over-funded or appropriately funded. Thus, the mere fact that a cause is “wonky” (strong from an analytical perspective, but not from a storytelling perspective) or “wacky” (controversial, farfetched or otherwise unappealing to conventionally minded people) doesn’t guarantee that it will be neglected. When it comes to assessing what areas are neglected, there’s no substitute for doing the legwork of figuring out who’s working on them.

Why one’s view of “broad market efficiency” mattersIn theory, it would be ideal to examine all possible giving opportunities and compare the “expected good accomplished” for each. In practice, it’s important to have good rules for prioritizing which areas and projects to investigate further. And having a view on “broad market efficiency” is important for such prioritization. Our view on “broad market efficiency” affects (a) how easy we expect it to be to find good giving opportunities in a given cause (and thus how much effort we expect to put in before having a good idea of what’s available); (b) what sort of claims about giving opportunities we consider highly plausible vs. which will immediately raise our skepticism that a piece of the picture is missing; (c) what sorts of giving opportunities are outstanding enough, in the scheme of things, for us to prioritize them.

If we expected extremely minimal “broad market efficiency,” we might be looking for projects that fit just about all the criteria one could ask for – strong track record, strong upside, strong people, etc. – and we’d deprioritize causes in which these projects didn’t quickly emerge. If we expected extremely strong “broad market efficiency,” we might place high emphasis on our personal interests and experiences, reasoning that these would be the areas in which we’d be most likely to “beat the market.” Our current view is in between the two extremes. Our intuition is that some causes are extremely under-funded relative to others, and we expect strategic cause selection to have major payoffs. (Note that while we have raised our estimate of “broad market efficiency” we’ve also become more confident that strategic cause selection is extremely rare, perhaps even nonexistent, in the philanthropic world.) On the other hand, we expect exploring a given cause to take a good deal of legwork and learning, and we’re inherently suspicious of projects that look “too good to be true.”

Another reason one’s view of “broad market efficiency” matters is it represents one mechanism for what I sometimes term the “fungibility of good” – the idea that making progress on one problem often leads to progress on other problems. For example, reducing the burden of malaria may make it more likely (depending on the degree of broad market efficiency) that other philanthropists shift from addressing malaria to addressing other problems. Thus, even if cause X is more important than cause Y, making progress on cause Y may cause other philanthropists to reallocate their giving away from cause Y and toward cause X, and may thus have some value for cause X.

The “fungibility of good” is one possible contributor to “regression to the mean” in philanthropic opportunities, and one possible justification for placing weight on tractability (what one can accomplish within a cause) and not just importance (the value of progress on a cause).