As public managers decide how to allocate funds, performance is a key consideration. Carla Flink, assistant professor in the AU School of Public Affairs, recently published a paper in the Journal of Public Administration and Theory that chronicles how budgets can rise and fall based on how well an organization is delivering on its promises.

“For the most part in government, things stay the same,” said Flink. “Policy change is incremental, but occasionally you see large shifts in public policy or innovation. Budgets are another way that we can think about policy change. If you want to enact new changes, money needs to be shifted around to fund a program. If a program is not as productive, you may take away funds.”

Flink wanted to examine how financial resources are altered in response to performance changes at public organizations. She analyzed data from hundreds of schools in Texas districts from 1993 to 2010, looking closely at how performance gaps affected the magnitude and direction of budgetary changes.

Her study tracked funding levels and how that connected to school test scores over time. Flink discovered that as performance slips, the probability of incremental budgetary changes declines and medium positive budgetary changes increase.

“The performance budgeting ideal doesn’t hold here,” said Flink. “The better schools are performing, the fewer funds they are getting toward instructional expenditures. The worse they are performing, the more they are getting.”

Flink said the results may appear somewhat counterintuitive, but schools are not run like businesses. There is a moralistic component to consider. Managers and policymakers may be responding to problems by investing more money in troubled areas.

The findings of the study show that performance matters in policy decisions. And it’s not just performance in one year but how a public organization is improving — or not — that impacts funding, according to Flink. Although education is a unique policy area with a high level of public awareness, the results could be tested to see if the theory applies in other policy contexts.