Performance measurement: impossible or essential? Alnoor Ebrahim gets me thinking

Harvard’s Alnoor Ebrahim gave a very nice talk during our our Philanthropy and Nonprofit seminar yesterday; I feel rather bucked in that I lured a number of my planning colleagues to attend and a bunch of my PhD students. Alnoor was one of my favorite colleagues when I was at VT, and his talk was really quite informative. He focused in depth on the extent to which foundations and nonprofits attempt to formulate performance indicators for social phenemena in order to nail down both controls and causality. Many organizations can’t go that deep: they just count clients served or outputs in some way. Still, others do try to measure such abstruse things as child poverty reduction or economic development, and he is attempting to derive theory about what leads organizations to do that, and what organizations–funders or the nonprofits themselves–are really in a position to examine outcomes.

He started us off with a question that I found myself rather fascinated with: is performance measurement impossible or essential? He referred us to some lectures worth listening to by Onora O’Neil for the BBC called “A Matter of Trust.” Alnoor raised her points about the problems with performance measurement, but he never really returned to them as, from there, he focusses in on the state of the practice in the field. The two big concerns about the measurement focus that I gleaned from O’Neil were:

1. It’s possible that our focus on measurement causes us to disregard what is difficult to measure about progress in the social sphere, and thus, the missions and strategies of helping institutions become rather small and constrained. That is, the bean-counting causes us to focus on beans that can be counted, rather than seeking to deal with social issues in all their difficult-to-measure complexity; and

2. There are transactions costs and compliance costs associated with all these performance measures, and the more of our resources we expend on measuring how well we’ve done at delivering service means fewer resources with which to actually deliver service. Making funding contingent on the ability to demonstrate that service is delivered in a particular way may favor those institutions who are good at #1 (cherry picking easy things) and #2.

Neither of which strike me as particularly good for social policy, though you can measure the last one.