[Cass R. Sunstein] How to test charity effectiveness

Americans are a generous people, contributing about $300 billion to charity each year. But when we make out a check to a charity, how do we know that the recipients are actually getting help? How do we know that our dollars are doing what we want them to do?

Remarkably, we often lack good answers. Fortunately, we are starting to get them. Before long, the answers are likely to make our contributions a lot more effective. They might well end up revolutionizing the charitable sector.

In “Moneyball,” Michael Lewis explored how Billy Beane, general manager of the Oakland Athletics, evaluated baseball players by using statistics, rather than pictures, emotions, intuitions, anecdotes, impressions or even experience. The idea of evaluating performance, with careful attention to the numbers that matter, has gained traction in many domains, including government. Both charities and donors should be playing a version of Moneyball.

Unfortunately, some current statistical measures aren’t especially helpful. For example, it is common to ask what percentage of donations goes to a charity’s administrative expenses. Although not irrelevant, that question can be misleading.

A charity might be using just 1 percent of donations for administration, but if the 99 percent isn’t doing much to help people, it’s a pretty bad charity. And even if a charity is using 15 percent of donations for administration, it might be a terrific investment if the remaining 85 percent is having a big impact on people’s lives.

In terms of evidence-based analysis of charities, the best current resource is probably Give Well, a nonprofit organization founded by people in the hedge-fund industry who wanted to find out how people can do the most good with their money. Give Well focuses on the relationship between aid activities and improved outcomes (for example, better health).

Give Well finds that three kinds of programs tend to work well: direct cash transfers to the poor, treating children for parasites and distributing insecticide-treated nets to prevent malaria.

The evidence to date shows, somewhat surprisingly, that direct cash transfers can be very effective. The idea is that recipients know what is right for them, and that they will choose wisely if they can put the money to their preferred use. Yet many people insist that targeted giving, based on a charity’s judgments about recipients’ most important needs, works best. They fear that if poor people are simply given cash, at least some of them will use the money wastefully or improvidently (perhaps for alcohol or tobacco).

To test the effect of giving cash, it would be best to have a randomized controlled trial, the gold standard in empirical research. One group of people, not given money (or anything else), would be the control, and it would be compared with an otherwise similar population randomly provided with cash, so that researchers can learn the actual effects of receiving money. One of the few such trials in the charitable area (and possibly the only one) was conducted a few months ago by the well-regarded Poverty Action Lab at the Massachusetts Institute of Technology, which specializes in randomized evaluations.

The MIT study involved GiveDirectly, a charity that provides cash transfers via mobile phone to poor households in Uganda and Kenya. (The amounts are often equivalent to a full year’s worth of consumption.) The experiment’s findings are impressive and encouraging.

As a result of the cash transfers, recipients increased their asset holdings by 58 percent, mostly with investments in home improvements and livestock. Business and agricultural income jumped by 28 percent. Expenditures increased significantly on food, education and health care ― but not on gambling, alcohol or tobacco. And there were big reductions in hunger, with a 42 percent cut in the number of days that children went without food, along with significant reductions in stress and depression.

Emphasizing these findings, Give Well ranks GiveDirectly first on its list of top charities. It also singles out two other charities: the Schistosomiasis Control Initiative and Deworm the World Initiative. The former treats people for parasitic worm infections in sub-Saharan Africa; the latter does the same in Kenya and India. Give Well likes these charities largely because of their cost-effectiveness.

The empirical findings are noteworthy, but what is most important is the movement, still in its earliest stages, toward rigorous evaluation of whether and how charities are actually helping people. With hundreds of billions of charitable dollars being given every year, it’s time for Charitable Moneyball. Let the game begin.

By Cass R. Sunstein

Cass R. Sunstein, a Robert Walmsley University professor at Harvard Law School, is a Bloomberg View columnist. He is a former administrator of the White House Office of Information and Regulatory Affairs, the co-author of “Nudge” and author of “Conspiracy Theories and Other Dangerous Ideas,” forthcoming in March 2014. ― Ed.