Instapundit points to Bill Frist’s proposals for altering the budgetary process to make it more accountable and controllable. He ask for thoughts on the matter.
I think that every government program should have some kind of metric attached to it which will be used to determine if the program has succeeded or failed.
The legislators would set the metric, the level of the metric that would determine success or failure and the time frame over which the metric would be measured when they authorized and funded the program. If the metric was not measured at the projected levels then the program would be automatically sunsetted.
For example, the success of a program intended to reduce unemployment would be measured by the unemployment rate (perhaps adjusted for GNP). Education programs could be evaluated by test scores. Environmental programs by pollution levels and so on.
How would this help control spending?
First, I think it would force proponents to lay their cards on the table and make concrete prediction for how well they expect a program to work. Other legislators could more easily evaluate the expected cost-to-benefit ratio. Proponents would avoid making pie-in-sky claims for programs because the higher they set expectations, the greater the chance that program won’t hit the target.
Second, programs tend to develop constituencies that have little to do with the effectiveness of the program. Few politicians want to spend political capital killing relatively small programs. Hardwiring a kill switch into the authorization would force ineffective programs into the sunset without having to trade off other programs that might actually be working.
Third, a lot of people seem to vote based on how proponents label the program. Many voters reflexively support programs labeled as “anti-poverty” or “tough on crime” without asking if the actual program does anything positive. Concrete metrics would shift the topic of debate from labels and vague intent to effectiveness.
Fourth, it would introduce something like free-market competition to government programs. Instead of competing against other business programs would compete against the expectations of the legislators who created the program. Programs that did not measure up would die out.
Some might argue that we cannot measure the effectiveness of all programs, but this objection itself raises questions: If proponents cannot define some kind of measurement to judge success or failure ahead of time how can we ever say that the program is working or not? Moreover, if they cannot predict how well a program will work then how well do they really understand the problem they intend the program to ameliorate?
If we truly understand a particular problem to the degree that we can use government power to improve it then we should understand it enough to measure it. We should also acknowledge that new programs are actually experiments and that most experiments fail. We need to have automatic mechanisms for determining which experiments worked and which did not.
Predetermined metrics are not a cure-all but they couldn’t help but improve both the cost and the quality of government programs.