Friday, July 17, 2009

Book: Challenging the Performance Movement

Radin, Beryl. 2006. Challenging the Performance Movement. Washington DC: Georgetown University Press.

I have been meaning to read this for a while and took the opportunity of working on an old project on performance standards to finally do so. Beryl Radin is apparently a grand old person in the public management world. Public management is a different but overlapping intellectual community from public policy or economics.

The book is a critique of performance management, the bipartisan spawn of intellectual movements sometimes referred to as "Reinventing Government" or the "New Public Management". I am in print as being a critic of aspects of this endeavor as well: see my papers with Carolyn Heinrich and Jim Heckman or with Burt Barnow. Nonetheless, the book is a very different one than I would have written.

Much of the book is, at some level blindingly simple. If you are familiar at all with the literature on "personnel economics" or more broadly on what economists call principal-agent problems, then you will know that it is a bad idea to measure performance using statistics over which the agent whose performance is being evaluated has little control. Our neighbors to the north, for example, have considered using the national unemployment rate as a performance measure for Canada's version of the Department of Labor. This is silly because the fraction of the variance in the unemployment rate accounted for by the agency is approximately zero.

The analog to this situation in the US is performance measures for block grant programs. Block grant programs are programs where the feds simply hand money to the states in lump sums to be used for certain fairly general purposes. It makes little sense to make the federal bureaucrats responsible for what the states spend the money on when they essentially have no control over it.

Another commonplace from the principal-agent literature is that if you want your agent to do two things well, but measure and reward their performance based on only one, then ... well, the reader can probably figure out what happens.

The book also makes very clear that a one-size fits all performance management system, which is what we essentially have under the "Government Performance and Results Act" is a bad idea. Think, to use one of Radin's examples, about NSF grants and related programs. Here the impacts are very long term and very hard to measure. Compare this to, say, SSA, where one can have a quite useful set of simple performance measures based on the timeliness and accuracy with which social security payments are delivered.

The book is very much not an economics book. My sense is that Radin is largely innocent of the economics literature. When she raises the issues mentioned above she does not frame them as an economist would. Other things that economists like to emphasize, like the distinction between outcome levels and impacts on outcomes relative to a counterfactual, are missing here as well, to the detriment of the argument.

I also found the book tough going at times and it occasionally veers of into things, such as alternative theories of intelligence, that seemed to me extraneous. At the same time, it was a useful introduction to me how public management people think about the issues and it describes the basic problems with performance management as currently implemented in a way that non-economists, particularly non-economists who would not like economics if they learned about it, would find congenial.

Bottom line: recommended, but only for afficianados. General readers should turn to James Q. Wilson's classic Bureaucracy book.