Once again, my wireless access disappeared mid-day at Google, so I am posting these thoughts from notes scribbled the old fashioned way - pen and paper.
The breakout sessions were off limits to the press and bloggers. I attended one on measuring development impact, and had a helpful talk afterward with Rachel Glennerster of the MIT Poverty Action Lab (who agreed to let me write about our conversation).
MIT's Poverty Action Lab is doing something most philanthropists or public funders probably would have said was impossible - random trials of social interventions to determine impact. The work is important and should be considered as a public good and public resource for all social investment decisions. Its a startup, so the body of work so far is small but it will grow. As important is the way Glennerster responds to the typical funder question about how/what to measure. I'm paraphrasing here, but when asked how a funder should think about evaluation when they have limited funds and/or are making small grants she responded along these lines:
Not everything should be measured. Not everything needs an impact study. What is important is that we think about the field-wide questions and try to answer those. So not every microfinance program needs to be evaluated, but we need to know what elements of microfinance make what differences. Then we can build new programs that use those principles but can adapt them to the circumstances.
We need good data. We need to connect good data to policy and funding decisions. The lab is doing this with policy briefs, publications, speeches. They also make their data available so you can run your own regressions - should you choose.
Many of the findings she reported were counter intuitive. For example, what is the most important thing to do to keep poor children in school? Provide de-worming programs. Want to lower the rates of pregnancy and/or HIV in adolescent girls in certain African communities? Provide free school uniforms. Want to lower the costs and increase the rates of vaccination delivery in rural India - offer 10cents worth of Dal to anyone who comes to get vaccinated. See more on their website.
As important, in my mind, was what will we need to do to make sure decision makers use these data. As Glennerster acknowledged, humans are not the rational economic beings that she and other economists have long thought we were. Funding institutions are not necessarily more rational. Passion and personal interest and conviction drives a lot of money and social activism. That said, I believe, as the commercial sector continues to build its presence in philanthropy, and business practices and performance-based funding grows in influence for social enterprises, government programs, and hybrid funding streams - the rationalizing power of markets will come to play a greater role in how we provide public goods. This should help good data - its quality will be recognized, its utility seen, and it will be used more.
But if anyone ever tells you that philanthropy and citizen action will be purely rational and data driven - well, I've got some swampland you might be interested in also. The Poverty Action Lab is making an enormous contribution by gathering and analyzing the data - now we've got to use it.