How agencies can get the inside scoop on wasteful programs

Steve Kelman is professor of public management at Harvard University’s Kennedy School of Government and former administrator of the Office of Federal Procurement Policy.

Many companies that have a new product or project in the works have adopted a novel way to predict whether the effort is likely to be on-time or on-budget. Rather than relying solely on the project’s manager — because the manager might not be forthcoming with unbiased information — some companies now seek information from all those working on the project.

To ensure that employees have an incentive to (anonymously) give accurate information, each person receives a sum of money to buy “shares” in a “stock” representing a project completion date and/or final cost. The stock pays out to those holding the shares that represent the actual completion date or final cost. The relative prices of the shares reflect the aggregated best judgment of the people working on the project regarding how long it will take and how much it will cost.

That prediction market is an example of crowdsourcing — using the wisdom of crowds to aggregate the judgments of lots of people into one outlook that is often more accurate than the view of an individual expert.

In our current deficit environment, we should, in an analogous way, be soliciting the judgments of the millions of federal employees to locate wasteful government activities ripe for termination.

We could give every civil servant one vote to nominate a wasteful government activity that should be eliminated. Officials could compile a list of the top 100 choices, and then the Office of Management and Budget could more closely evaluate the 50 that would potentially save the most money.

At the end of the process, every civil servant who voted for an activity that was stopped would share from a pool of the first-year savings from eliminating the activity, minus the costs of running the program. Savings in subsequent years would go to deficit reduction.

Depending on the number of people who voted for the winning suggestions and the amount of savings, employees could receive a moderately large amount of money. For example, if 10,000 civil servants voted for a suggestion that saved the government $30 million in the first year and the program cost $10 million to run, each employee’s personal share would be $2,000.

Some activities should be excluded from the program. First of all, substantive policies whose value is under political dispute, such as funds for high-speed rail or agricultural subsidies, should be decided by the president and Congress. Second, government management challenges, such as reducing Medicare fraud or improper payments under benefit programs, cannot simply be turned on and off but instead require significant and ongoing management intervention to accomplish.

Instead, I’m thinking of activities that achieve little toward meeting a program mission but on which significant funds are being spent. Examples include production of nonstatutorily required reports that nobody reads, service contracts that employ lots of people but produce little of value, offices in which a large part of the workforce is significantly underemployed, product purchases that gather dust, nonstatutorily required procedures that cost money for no good reason, and development efforts that are going nowhere.

Allowing employees to take a proportion of savings would require legislation, and some people might balk at the idea of paying feds for their ideas. But keeping 100 percent of the savings on ideas that never get generated yields zero in deficit reduction. Getting a percentage of real savings is a better deal for taxpayers.