When Performance Measurement Goes Wrong in Government
Scandals over falsified performance data compromise efforts for greater accountability. But there are lessons to be learned.
Recent revelations that employees at U.S. Department of Veterans Affairs hospitals falsified performance data on patient-appointment wait times in order to receive larger bonuses have turned the Veterans Health Administration (VHA), once considered at the forefront of federal efforts at performance management, into a national joke. While it is not known how widespread this practice was, it is a reminder that it does not take many bad actors to ruin the reputation of an entire agency.
The VA scandal comes on the heels of a number of high-profile cases of education administrators and teachers cheating (or encouraging cheating) on standardized tests, making it all the more worrying to those of us who advocate reforms centered on performance management. In Atlanta, the school superintendent, as well as 35 teachers and principals, resigned in the wake of a scandal allegedly driven by pressure placed on teachers to falsify test-score results. Educators in cities including Baltimore, Cincinnati, Detroit, Houston, Las Vegas, Los Angeles, Newark, Philadelphia and Washington, D.C., also have been caught up in cheating allegations involving standardized tests.
What is to be made of this? There is an old saying in the performance-measurement community: "What gets measured gets done." If this is true, then the dishonest behavior that has been alleged is perhaps predictable. We cannot assume that the good intentions of system designers will translate into all participants in a system responding with a good-faith effort to do the right thing. It may be, in cases where performance measures are being used to reward and punish, that too many times the adage should read: "What gets measured appears to get done."
So we are left with a familiar dilemma. There is no incentive to misuse a performance measurement system unless the measures are used for something that people care about. But if they are not used for anything that anyone cares about, why bother?
It's tempting to conclude from cases like these that the entire performance-measurement movement is misguided and creates more harm than good. I am not ready, however, to throw the performance-measurement baby out with the foul bathwater. Instead, let me suggest three cautions for those pursuing performance measurement in government:
1. Measure the right things: Whether there are scandals or not, it is important to make sure that the measures you are using are in fact measures of the underlying performance you care about. If what gets measured gets done, then it follows that if you measure the wrong thing the wrong thing will get done. It is particularly important to measure the right things if you are going to use performance data to reward and punish.
2. Remember that reputations are hard to rehabilitate. Just because there are some high-profile cases of cheating does not mean that that is the norm. In education, for example, surely there are many more cases in which local schools are making a good-faith effort to implement performance-measurement systems than where the system is being gamed. Whether the norm or not, however, the public-relations fallout is a real one. In an age where government is not trusted, it doesn't take much to convince people that government agencies and officials are acting cynically for personal gain rather than to serve the public.
3. Trust, but verify. We should not lose sight of the fact that there is a silver lining to these cases: After all, we know about them. The system, considered broadly, worked. In the VHA case, the inspector general uncovered these practices. In the school cases, the alleged wrongdoing was discovered either through internal audits or media investigations. The publicity afforded these cases is likely to give pause to those who might engage in this kind of behavior in the future.
In the end, what these cases underline is the importance of effective oversight of performance-measurement practices. If a performance culture is going to take hold in government, those of us who are advocating these reforms need to be realistic about the measures themselves and the incentives for people to misuse them.