Economists Dean Karlan and John A. List recently made a deal with an unnamed national nonprofit: They would lend the organization their expertise in study design and data analysis if the nonprofit would let them turn its upcoming direct mail fundraising campaign into a big field experiment. Using the campaign as a laboratory, the economists explored whether, why, how, and for whom matching grants work.
Their results were as subtle as their methods: Matching grants do increase donations for some people, but not for the reasons most economists think, they report in the December 2007 issue of the American Economic Review.
“Most fundraising analysis is strikingly void of theory and of careful analysis,” notes Karlan, a professor of economics at Yale University. To add some science to the alchemy, he and List, a professor of economics at the University of Chicago, first developed different versions of a fundraising letter, systematically varying the ratio at which a “concerned fellow member” would match donors’ contributions. In the treatment conditions, letter recipients learned that the fellow member would give the same amount, twice the amount, or three times the amount of their donation. In the control condition, recipients did not read about a matching grant. The nonprofit, a politically liberal civil liberties group, then mailed each of more than 50,000 potential donors one version of the letter.
The researchers found that letter recipients who read about a matching grant were more likely to give and gave more than did recipients in the control no-matching-grant condition. “Contrary to conventional wisdom, however, the match ratio didn’t matter,” says Karlan. People in the 1-to-1, 2-to-1, and 3-to-1 ratio conditions donated equal amounts.
Peering more closely into their data, the authors discovered a further refinement: The matching grant increased giving only in red states, but not in blue states. In other words, blue state residents gave similar amounts in the control and treatment conditions, but red state residents gave significantly more in the treatment condition than in the control condition. (Overall, blue state residents gave more than red state residents.)
The authors do not fully understand why the matching grant worked in the red states but not in the blue states. They do understand, however, the importance of rigorously testing fundraising techniques, as well as of studying the peculiarities of particular donor bases. Karlan urges nonprofits to “hire the type of people who thrive on taking a step back and asking on a theoretical level why people support them, and then on doing careful experimentation and data analysis to test their hypotheses.”
As for his own findings, Karlan speculates that time and place affected how the matching grant resonated with donors. In red states where liberals may feel they are in the minority, a matching grant from an ally might have triggered donors’ latent political identities, he says, and thereby catalyzed them to give more to the organization.
Yet the authors studied only one nonprofit, and so these findings may not apply to other organizations, warns Mal Warwick, founder and chairman of the fundraising consultancy Mal Warwick Associates. “I have never seen anything but an increase in response when a matching gift is offered,” he writes in an e-mail. He recommends that researchers study at least 20 organizations of various kinds before drawing conclusions. He further notes that 3-to-1 or 2-to-1 matching grants are rare: “Most of the time, nonprofits struggle to line up a gift that will enable them to offer a 1-to-1 challenge.”
Nevertheless, the authors’ theoretical point still holds: “Economists have it as wrong as they possibly could when it comes to nonprofit fundraising,” says Karlan. According to economic theory, increasing the ratio of the matching grant should have amplified donations by lowering the “price” of giving. Instead, matching grants of all stripes equally increased donations, but only in red states.
“The decision to donate to a charity is not best modeled by thinking about purchasing public goods,” concludes Karlan. Instead, “many noneconomic factors drive decision making in charitable giving.”
Read more stories by Alana Conner.