Systemic change or proven programs?

Fund what is proven

Programs that have been studied rigorously, ideally more than once, let givers know what their money does. Evidence protects against waste and against good intentions that backfire, and cheap proven programs save lives today.

Fund systemic change

Charity that works only inside existing institutions leaves untouched the structures that produce suffering, such as trade, debt and political power. Changing them matters more than any single program, even when the results can’t be priced in advance.

Structural work can be evidenced too

Some of the largest documented gains came from structural change, such as taking lead out of gasoline and paint or raising tobacco taxes. Their effects can be measured by other methods when no trial is possible, so givers can hold systemic work to evidence standards instead of choosing between evidence and depth.

Proven programs can become structure

A program proven in trials can be built into government systems and change institutions from the inside. The two approaches are stages of one path, not rivals.

The debate

GiveWell’s criteria state the case for proven programs. It looks for programs that have been studied rigorously and ideally repeatedly, estimates what they achieve per dollar and, since July 2022, recommends only charities it estimates to be at least ten times as cost-effective as cash transfers.

In a 2015 essay in the London Review of Books, Amia Srinivasan made the case from the other side. She argued that effective altruism works within existing institutions instead of questioning them, leaves structural causes of suffering such as trade, debt and labor exploitation unexamined, and judges political change with an expected-value arithmetic that can’t price radical change.

The third position points to evidence. In the United States, children’s average blood lead levels fell about 95 percent between the late 1970s and the 2010s, declines the researchers credit largely to federal regulation. That evidence comes from national health surveys, not a randomized trial, and J-PAL notes that randomization is hard when a policy applies to a whole country. An expert review convened by the International Agency for Research on Cancer found sufficient evidence that higher tobacco taxes cut smoking.

Attribution is the hard part. A World Bank account of the Philippines’ 2012 tobacco and alcohol tax reform credits a coalition of officials, legislators, development partners, medical professionals and civil society activists, and concludes that no single actor could have done it alone.

The fourth position points to programs that moved into government. Teaching at the Right Level was tested by researchers and a nonprofit, then run through government schools in Indian states. The first attempts failed and later versions worked; see the entry.

Where Charitology stands

Charitology’s depth scale speaks directly to this question. It describes where work acts without grading it, so givers can see the mix they fund. The evidence shows that structural change can produce very large, measurable gains, and that credit for any one change is hard to assign. Whether to accept more uncertainty for a chance at deeper change is a value question Charitology doesn’t settle. It treats the measurement gap as the problem it can work on. See measuring the source, the Symptom Trap, systemic change, and the entries on lead and tobacco taxes.

The cruxes

What would have to be true for each side to change its mind:

Whether the effects of advocacy and policy change can be measured well enough to compare with direct programs.
How much of a law’s success any one funder or group can claim.
Whether funding proven programs delays structural change or builds the trust and capacity that make it possible.
How much risk of achieving nothing a giver should accept for a chance at a much larger change.