Promising Practices Promising PracticesPromising Practices
A forum for government's best ideas and most innovative leaders.

The Art of Trickle-Down Performance Management


Can federal performance management schemes influence efforts at the local level? Typically, performance management works best in systems where agencies engage in direct service delivery, where leaders have more control over what is going on. But in complex intergovernmental programs, can effective performance management systems be developed and work?

A recent case study of the federal-state-local Child Support Enforcement Program (CSE) suggests it can be done, but it takes time and collaboration across a wide range of stakeholders. The study, by Robert Doar, Dennis Smith and Kinsey Dinan, examines “the experience of a complex federally managed, state-administered program that has already implemented a performance measurement and management approach.” They traced the development of the CSE approach over a 20-year period, starting in 1994 (with initial implementation in 2002). CSE set five key metrics and has been tracking progress over the years. Child support payments, for example, cannot be collected unless an order is established first. The percentage of cases in which states successfully established child support orders rose from 62 percent in 2000 to 82 percent in 2012.

The CSE program was created in 1975 to locate noncustodial parents, establish paternity, file child support orders and enforce those orders by collecting payments from noncustodial parents -- sometimes called “deadbeat dads.” The payments are delivered to the parent caring for a child or shared with the state if the parent is receiving welfare payments. CSE is part of the Health and Human Services Department, but states have primary responsibility for operating the program.

The federal government sets program standards and shoulders about two-thirds of the administrative costs for running the system. The federal arm also provides states incentive bonuses if they are particularly effective in collecting and disbursing payments.

Setting Common Measures

The CSE program was chosen to be one of the first pilots for performance measures in 1994 under the just-passed Government Performance and Results Act. The federal program head at the time, David Gray Ross, said he wanted to create “a consensus-based system of measuring performance, as well as an identified standard or level of performance which would be used to reward above average performance and identify poor performance.” To do this, he followed three steps:

Step 1: Create a vision. The GPRA law required agencies to draft a five-year strategic plan. Federal CSE staff drafted a blueprint and actively sought feedback from state CSE directors. The plan described a vision for the future, with three supporting goals and defined objectives to meet those goals. To develop a cross-state consensus, the original plan did not attempt to define measures. Rather, that was designated as a next step in the process.

Step 2: Develop measures. A cross-state performance measurement working group, which included federal and local representatives, began drafting and debating measures in early 1995. Tentative measures were adopted to test by mid-1996. Ultimately, five measures were designated as key.

Step 3: Design an incentive-based funding system. At the same time the measurement pilots were beginning, a major welfare reform law was passed that called for improvements to the CSE program and required a new performance-based incentive funding formula to be developed. A cross-state working group was formed to create this formula, which resulted in a report in early 1997 detailing five proposed performance measures, which were set by law in 1998. Each year since fiscal 2000, state performance and incentive payments have been calculated based on these five performance measures.

How the New System Worked

At the state level, performance varied widely. New York, for example, embraced the new measures in 1998 and used them to define priorities and strategies that enhanced its performance management system. New York has a state-supervised, locally administered child support enforcement system comprising 58 local departments of social services. Prior to 1998, the state CSE director had already decided the departments should collectively focus on what became one of the five key metrics: the number of support orders achieved in court each month. They set monthly goals, county by county, and the state employees in each county were held accountable for meeting these goals. The new federal management system provided further energy and direction to the state’s approach.

By far the largest of the 58 local departments is New York City’s Human Resources Administration. Its Office of Child Support Enforcement has an active caseload of more than 400,000, and in 2013, collected $735 million in payments (see the Mayor’s Management Report, p. 96). Doar, lead author of the study, is commissioner of HRA. The Office of Child Support Enforcement focused on all five key metrics, with emphasis on one in which the city lagged behind the rest of the state: the percentage of open cases with a court order to collect payments.

To manage improvements in the percentages of cases with court orders, HRA created an extensive reporting structure that enabled staff to detect trends via intermediate process measures. These more detailed operational measures can reveal early warning signs, giving staff a head start on developing operational-level solutions before trends become a problem.

To help with the operational-level management, HRA adapted the police department’s award-winning CompStat approach to conduct monthly “stat” meetings with the borough office directors and key headquarters executive staff. These meetings parse statistical reports prepared by staff and provide a forum for explaining significant increases or decreases in anticipated performance levels and developing corrective strategies.

Lessons Learned

“Implementing performance management within the framework of the federal CSE program’s incentive funding structure has allowed OCSE to significantly enhance its efficiency and effectiveness,” the study’s authors say.

Since 2000, when the metrics were established, nationwide child support enforcement performance has improved against all five measures. In New York State, for example, the percentage of cases in which paternity had been established grew from 80 percent in 2000 to 85 percent in 2013. In the case of New York City, the growth went from 43 percent to 76 percent. 

HRA Commissioner Doar and the head of New York City’s OCSE, Frances Pardus-Abbadessa, concluded that the federal role in performance management was an important factor in improving the city’s child support program. This demonstrates, they say, “that state and local governments can and will organize their operations to achieve performance standards set by the federal government” if the performance system is well-designed and implemented. Good design elements include:

  • A highly collaborative and consensus-oriented approach to developing a vision and performance measures.
  • A focus on a small number of key measures, along with ambitious goals for these measures.
  • The use of rewards more than penalties, with recognition of performance improvement as well as absolute levels of achievement.

The authors believe similar performance management systems could be developed for other federal-state-local programs, such as workforce development and child welfare services. They conclude: “The open question is why it has not been applied more broadly.”

(Image via donskarpo/


John M. Kamensky is a Senior Research Fellow for the IBM Center for the Business of Government. He previously served as deputy director of Vice President Gore's National Partnership for Reinventing Government, a special assistant at the Office of Management and Budget, and as an assistant director at the Government Accountability Office. He is a fellow of the National Academy of Public Administration and received a Masters in Public Affairs from the Lyndon B. Johnson School of Public Affairs at the University of Texas at Austin.

Close [ x ] More from GovExec