The "Fear" of Performance Management

In a discussion with a colleague recently the topic of why more governments don't have an active performance management program came up. I will admit that the discussion was more speculative than scientific, but we generally agreed that many jurisdictions and agencies likely don't implement performance management programs out of fear of both what they might find as well as how the data that is reported might be used against them (the remainder of non-practitioners likely have no idea what it means!). There are stories of early meetings of CompStat in New York city in which supervisors were skewered based on crime stats in their area (although later accounts suggested a softening in the tone), and perhaps this is what government managers reference when thinking about reasons not to further performance management programs. But anyone focused on that aspect is ignoring the second part of that story which is the potentially positive impact of programs like CompStat. 

New York city has one of the lowest violent crime rates and the lowest property crime rate of large cities in the US. This is in stark contrast to the late 80s when it had one of the highest. The rates dropped precipitously throughout the 1990s, around the same time that CompStat came into being. While this relationship may very well be spurious, I imagine that the new management style in the police department had at least SOME effect on crime outcomes. I use CompStat as an isolated case of one agency's efforts, but it would be an interesting exercise to look at cities and states with advanced performance management programs (Atlanta, Albaquerque, and North Carolina spring to mind) and analyze the short and long term impacts of those programs both socially and politically. The reason this might be helpful is that it would be informative to cities who "fear" such programs in demonstrating the long term value of these programs. Certainly there are painful short term realizations of inefficiencies that would be made from better data and analysis but several case studies on improved outputs might put those fears to rest.