Effective Altruists Should Be Working Here

If you’re in DC and notice more background discussion of the discount rate on utils than is usual as you sip your vieux carré in Dauphine’s bar this weekend, don’t worry, it is just participants in EA Global, the  jamboree for Effective Altruists (go see my colleague Amanda Glassman speak!).  My big hope for the weekend is that lots of participants fall in love with the place and decide to stay, both because Washington is wonderful—honestly, if the forecast is right, it’s always like this—but also because DC is where EA types belong.

In terms of effective altruism, I would argue the Gates Foundation’s work in creating and funding Gavi, the Vaccine Alliance, was probably one of the very smartest uses of philanthropic resources to save lives. Vaccines are a highly cost-effective health intervention, vaccination rates in many developing countries were low, Gavi distributed a lot of vaccines to developing countries for free, Gavi partners like WHO and UNICEF worked with country governments to deliver them, and vaccination rates went up. I should note that Gates is a major funder of CGD, including my work, but others without that potential conflict agree on effectiveness: while GiveWell doesn’t recommend donating to Gavi, that’s only because it thinks the organization is adequately funded. 

I think the history of Gavi has lessons for effective altruism. First off, the reason it doesn’t need more funding: the Gates Foundation gave about $4 billion to Gavi 2000-20, but that was more than matched by $16.5 billion that came from donor governments. And most of the cost of actually getting shots into arms (doctors, nurses, logistics) fell to recipient governments. Second, countries a little too rich to be eligible for Gavi support saw immunization rates reach pretty much the same levels as the richest countries just below the eligibility threshold who got support. Governments in those richer countries (supported by lower prices that Gavi helped achieve, amongst other indirect effects) matched the effort made by Gavi in poorer countries. The private act of Gates philanthropy was many times magnified and extended by public resources. The altruism was so effective because it leveraged governments. I think that’s likely to be the rule rather than the exception for impact at scale for philanthropic funders, and, if so, it points to working with or in governments as a really important job to maximize impact.

Perhaps that’s especially true for EA folks who aren’t already rich. A few effective altruists have made so much money that perhaps it is pretty much certain ex-post that the best thing they could have done for the planet is earn scads of cash and then give it away to high-impact causes. But this isn’t a reliable strategy. Most Internet startups fail; financial innovation is hardly a guaranteed path to wealth. Ex-ante, an effective altruist should see there are other (perhaps more reliable) ways of directing lots of money at good causes than earning it—not least, directing other people’s money, because there is a lot of that around. McKinsey says global net worth is about $510 trillion while the World Bank’s more expansive measure of wealth (including future earnings from current education stocks) comes to $1,152 trillion.

And what’s interesting about all of that wealth is that the world’s governments have a large stake, and an even larger range of control, on how it is spent. Effective altruists may already have tens of billions of resources, but the world’s governments spend that amount every four hours—about $23 trillion in a normal year (27 percent of an $85 trillion global economy).

And just looking at the financial flows ignores the huge influence of laws and regulations on outcomes. Take just one example: government restrictions on migration worldwide mean the global economy is about half the size it could be. Or look at (arguably) the world’s current biggest really-close-to-fully-existential risk, nuclear weapons—which are everywhere, so far, in the control of government (we think, I hope). On the other hand, the lives saved by the (still very imperfectly enforced) government monopoly on violence over recent centuries probably add up to billions. For all of the regulatory failings that exist, that food is largely safe to eat, drugs to take, air to breathe, water to drink and walls to lick in the United States is something to give thanks (at least in part) to government for...

…Which in turn brings us to another reason to focus on governments: they are, at least in theory, already somewhat aligned with the EA cause. They’re meant to be acting in the common good, and promoting the general Welfare (if, admittedly, at the national level and not caring quite so much about citizens many years hence). In part because of that, they fund a huge chunk of global R&D, alongside global health spending and asteroid research. They operate the world’s biggest cash transfer programs. They create laws and regulations around computer use. Name your EA priority and governments are probably the biggest spenders and major regulators. The money could surely be spent better, the regulations improved to stop a robopocalypse, but that’s a reason to engage more.

Given the scale of global suffering now and unto the future, a collective response is what will have the efficiency, scale, and lasting power to meet the challenge. Effective altruism surely demands working to change collective decision making if that is the most impactful way to make a sustained difference. So, I don’t think people who really want to make the world a better place in the long run can afford to not think about government.

Thankfully, even in the US, government isn’t an utterly intractable mess. New far-from-perfect laws get passed that are sometimes better than old further-from-perfect laws; the same with regulations, the same with budgets. Individual bureaucrats stuffed in tiny cubicles can make meaningful differences to the quality of billions in spending or improve regulations that affect the quality of lives of millions (think of people on Joe Manchin’s staff, or the group of people in the Bush administration who designed PEPFAR). To quote a post about US policy work on the EA Forum website “Being absurdly impactful through US policy work is very feasible, and not even very rare in this community.”

Of course, it is extremely hard to measure how large your individual impact is as a financier of advocacy campaigns to change policies or a government employee trying to get the right thing done, but the collective impact is huge. Again, there may be greater uncertainty in trying to influence policy than simply handing over cash to effective uses, but the history of Gavi alongside randomized controlled trial replication experience and in particular the debate over a recent GiveWell top pick demonstrate, that uncertainty is a matter of degree rather than a discrete difference.

That will be why big players in the Effective Altruism movement years ago moved towards policy and politics—financing campaigns and advocacy.  (Related: the complaint of “not enough policy” in EA is old, as are EA thoughts about doing more policy philanthropy.) And I should say CGD along with a number of other think tanks gets generous support from Open Philanthropy to advance better policymaking. Again, 80,000 hours has a job board for positions it reckons are likely to have a big impact, and there are a lot of jobs that involve influencing policy.

But I’d argue that once you’ve looked at policy as a source of societal progress and regress, especially over the long term, it is hard to look at anything else. As a result, I hope ever-more people in the EA community in town this week decide to work in government, or work on making government work better.


CGD blog posts reflect the views of the authors, drawing on prior research and experience in their areas of expertise. CGD is a nonpartisan, independent organization and does not take institutional positions.