Over the next several years, governments at all levels will face more calls for—in fact, in many cases, the necessity for—budget cuts. The question, for these governments and agencies, will not be whether the budget will be cut, but what will be cut and how decisions will be made on which programs to spare and which ones will face the budget ax.

In such an environment, it is worth asking whether the much ballyhooed reform of performance budgeting—which is supposed to promote conversations about what works and what doesn't—can help. This conversation is especially important to have when resources are scarce; if something that works gets cut while something that doesn't gets left alone, this is not a responsible use of public resources. Unfortunately what evidence we have suggests that far too little use is being made of performance budgeting to accomplish that goal of responsible fiscal policy.

For many governments, the easiest way to deal with reducing the budget is to resort to the "equal sacrifice" rule, which employs across-the-board cuts (in Europe, where they like to employ culinary references when possible, this is called "cheese slicer" budgeting). This is often done in the name of fairness, with the justification that each program or agency should make an equivalent contribution to deficit reduction. The more enlightened versions of this regime will exempt some programs (elementary and secondary education, for example; after all, the governor may have campaigned promising to be the "education governor"), while taking bigger chunks out of everyone else.

The temptation to do this is understandable. All that is needed is a spreadsheet that can shrink last year's budget for each agency by, say, 10 percent. It is also predictable for the agencies. But budgeting is about setting priorities, and everything quite simply is not equally important, nor do all programs work equally well. Those (mayors, governors, presidents, legislatures) responsible for making decisions on budget reductions owe it to citizens to make those decisions in a way that will leave the government doing the most good for the resources that it has, even when those resources are diminished.

To what degree is that happening? A recent study by Yilin Hou and his colleagues at the University of Georgia addressed this issue. The study examined the extent to which 11 states—many of which supposedly are among the leaders in performance budgeting—used performance data to inform their budget decisions between 2008 and 2010, when the "Great Recession" was necessitating budget reductions for the first time in almost a decade.

The results of this research are not encouraging to supporters of performance budgeting. While the research found some evidence that some performance information was being used to inform budget decision-making (for example, in Maryland and Louisiana), the predominant story in the case studies was that performance information was "not necessarily very useful as a budget tool in the present fiscal climate." The essential dilemma was illustrated by the state of Utah, which was limited in its ability to reduce funding for low-rated social service programs because the recession necessitated that the state maintain the services provided by those programs and there was a lack of alternatives.

The general conclusion of this study is that performance information played a greater role in these states when the economy was strong than when it was weak. Moreover, the more consistent use of performance information is found in management, rather than in budgeting.

Why is performance information not being used to inform budget decisions during times of budget cutbacks? There are several possibilities. First, a given government or government agency may not yet have good measures of performance for its programs. Second, and even more likely, there may be insufficient knowledge concerning the relationship between decreases (or increases, for that matter) in funding and performance. Finally, performance data may not yet have the credibility with elected officials to allow it to be used in evidence-based decision making.

Regardless of the reason, it would seem that if performance information is ever going to be used for budgeting at a government-wide level, it is most important that this use come when the stakes are highest—that is, when resource scarcity puts established programs, with real constituencies and potential real effects—at risk. If any readers know of places where performance data are influencing decisions about budget reductions, please let us know. Such success stories need to be highlighted and emulated by others.