Is what’s measured what matters? Pitfalls with performance measures

By , November 28, 2012

The private sector uses one metric—profit—to determine whether an investment is worthwhile. The public sector, in contrast, cannot determine whether government programs are worth the money by using only one measure. Answering how much we should we spend on public goods—such as health, education, justice, and social services—is a difficult task, and one we face every year.

To insert objective information into subjective (and usually contentious) discussions about budgets, government planners look to performance measures, that is, data on program inputs, outputs, and outcomes. Performance measures can add value to a budget debate, but when defining them, it’s important to avoid a couple of pitfalls.

The first pitfall is not directly relating a performance measure to the program or policy being evaluated. This happens when the causal link between a desired outcome and the program is tenuous and results in a measure that doesn’t effectively gauge the program’s quality. For example, a reduction in the crime rate is an important policy outcome, but a poor performance measure. Many factors affect the crime rate, and although some of these factors can be influenced by police, prosecution, and corrections policy, it would be difficult to attribute a change in the crime rate to a single intervention.

The second pitfall is relying on measures that can easily be gamed, that is, improved without changing the quality of the program’s performance. A typical way this happens is through creaming (as in the “cream of the crop”): selecting participants who are likeliest to succeed, thereby making the program look as if it creates strong outcomes.

It is particularly important to keep these pitfalls in mind when making financial payments based directly on performance measures, whether payments are made through performance incentive funding, a social impact bond, or other “pay-for-success” approaches. The use of performance-based payments is a promising way to promote better outcomes, but the stakes are a lot higher, requiring an investment in research as well as ongoing analysis to ensure that the performance measures accurately assess quality.

We’re focusing on outputs, outcomes, and cost-benefit analysis this month on the blog. We encourage you to leave a comment below or
send us your questions and ideas via Twitter or Facebook, or by e-mailing us at

Leave a Reply

Your email address will not be published. Required fields are marked *

You may use these HTML tags and attributes: <a href="" title=""> <abbr title=""> <acronym title=""> <b> <blockquote cite=""> <cite> <code> <del datetime=""> <em> <i> <q cite=""> <strike> <strong>