Assessing program performance is harder than one might think.
It's been 15 years since Congress passed the Government Performance and Results Act, a law whose sponsors hoped would help establish a government guided by goals and measured by results. Sen. William Roth, R-Del., said on the Senate floor in 1993: "For much too long, this focus on program performance and results has been missing from the federal government. Agencies and managers are expected to follow proper procedures and spend their funds in an appropriate manner-to cross all the t's and dot all the i's-but rarely are their programs held accountable for achieving measurable results toward any pre-established goals. Is it any wonder, then, that program performance suffers, and that public frustration with government increases?"
Roth's successors could simply cut and paste his explanation into their own floor speeches today. Measuring government performance has proved a lot harder than it sounds.
The most recent case comes from Head Start, a pre-kindergarten program for low-income toddlers. In December, Congress passed a reauthorization law that killed the most ambitious measurement experiment launched in the history of that initiative. The youngsters were tested for a couple of years on several skills, including knowledge of the alphabet and counting. The experiment's demise points to several of the problems that results-based government has encountered since 1993.
First, the notion of measuring the performance of toddlers generated massive controversy. Critics charged that 4-year-olds do not make good test subjects; their attention spans, willingness to participate and mood swings could easily sway test results. Across government, program managers who have attempted to establish goals and measures have run into similar controversy over the validity of the goals and the way they are measured.
Second, program administrators at the local level revolted. Head Start is a grant program, with the federal government distributing money to local, nonprofit grantees that perform the service. The grantees complained bitterly that the testing system would eventually be used against them and that it was an undue burden on their time and resources. Since much of the government's work is done by third parties, that kind of tension is common when government tries to measure results.
Third, Congress is in charge, not program managers. Both Democrats and Republicans in the House and the Senate lined up against the Head Start testing program, and they used the power of the legislative pen to kill it. And lawmakers have axed other measurement systems that generated controversy among their constituents.
The Head Start testing system indeed could have been flawed. Rather than dismissing the notion of measurement, Congress in the reauthorization law asked a panel to consider different ways to assess the effectiveness of Head Start in preparing low-income kids for school and fostering good health.
But with the strategic plans and annual performance reports required by the Government Performance and Results Act largely gathering dust, with hundreds of federal programs operating without clear benchmarks, and with some seeing their measurement systems dissolve in controversy, the primary lesson of efforts to assess government results is not what was expected a decade and a half ago. You're damned if you do, and you're still damned if you don't.
Brian Friel covered management and human resources at Government Executive for six years and is now a National Journal staff correspondent.