Bob Deller: Controlling performance
- By Bob Deller
- Apr 28, 2003
Responsibility for performance must go deeper than administering programs
It's one thing to measure outcomes based on performance and yet another to influence variables that contribute to those outcomes.
Some agencies have been outstanding in demonstrating that their performance produces results. And they've been rewarded with spending authorization.
But many requirements found in the mission of an organization are outside its control. Or the agency hasn't found a way to control them.
For example, take an agency that provides grants to state, local or tribal governments in return for implementing federally mandated programs.
Health care, safety, education, law enforcement and transportation are among programs that have local implementation and depend to a large degree on federal funds.
We see performance measurement progress made at the federal level.
But what about the segments of the programs under the control of the federal grant recipients?
We are blessed with statistics that report our children's improvements in school; statistics show reduced crimes; traffic safety achievement is at an all-time high; and so on.
But do federal agencies really know why improvements happen?
Aside from a possible Hawthorn effect in some service areas brought about by changes in system implementation, agencies really do not control variables that contribute to success or failure. Their performance metrics are defined in terms of what their administrative programs are responsible for.
Under the best of circumstances, a federal agency that contributes funding for locally implemented programs works from two sets of information.
One is macro'the outcome of national programs, a favorite rhetoric of the administration. Highways are safer, children are better educated, senior citizens have better health.
The second is micro'data collected by local organizations that measure variables that influence outcomes. These data are received as much as two years late by federal managers, and therefore are not representative of current outcomes. Without being able to control these variables, it would be too late to make necessary corrections.
Measuring outcomes is important. Unless federal agencies are equipped with methods that measure and influence the independent variables that contribute to outcomes, agency managers can't be as responsible for budget dollars as they should be.
Bob Deller analyzes the federal IT market. E-mail him email@example.com