washingtonpost.com

By Stephen Barr

Wednesday, January 7, 2004; Page B02

The devil is in the details. Yes, it's a cliche, but it applies perfectly to the pay-for-performance plans proliferating inside the Bush administration.

Rank-and-file federal employees in two large departments -- Defense and Homeland Security -- will likely move to new pay systems that emphasize performance in the next two years. Federal executives, who are no longer guaranteed an annual pay raise, await new agency-based systems that are supposed to put some rigor into how their performance is evaluated.

He sums up the challenge with a question that involves four questions: "Who gets to decide who gets a pay boost of how much for what kind of performance?

The first "who," Behn writes, is the boss -- the person who is supposed to know how each worker has performed. But, in reality, workers have multiple bosses, Behn says.

Government employees work on teams and special projects "about which the official 'boss' may have little knowledge. Often a large number of other people (including peers and subordinates) have some detailed knowledge -- and quite different perspectives -- about how any individual has performed over a diversity of tasks. Which of them gets a say in the decision?" Behn asks.

The second "who" is the person getting the raise. Too often, government limits the number of performance-based raises to 20 percent or 25 percent of employees. But, Behn asks, "what if the supervisor has recruited a high-performing team? Should only 20 percent be eligible? Does this suggest that, if you want to win a pay boost, you

should choose to work on a low-performing team?"

Regardless, Behn writes, if only 20 percent can receive a performance raise, that means 80 percent "will be automatically labeled losers." And the 80 percent, he says, end up being depressed. "This is motivational?"

The third issue is how much to raise pay. Behn recounts how one state set aside bonus money for the top 20 percent of employees. But the government faces budget constraints and usually cannot guarantee that funding for a performance system will remain stable over time. In Behn's example, the payout amounted to an extra $33.33 per month, before taxes. "What impact do you think this performance bonus had on performance?" he asks.

The fourth challenge involves deciding what counts as good performance, Behn says. "Different people do different tasks. So how does the boss (or the pay-for-performance committee) compare those who picked the apples with those who picked the oranges?" he asks.

Behn also asks how bosses intend to deal with team performance -- a common way that the government performs its work. "Teams are essential when the task requires many different talents: orange pickers, ladder holders, box packers, box carriers. Sure, the actual pickers are more important than the ladder holders. But if the orange-picking team's performance improved significantly, is this solely because of the star pickers or is it because the team's box packers learned how to be both quicker and more careful?" he writes.

More important, Behn suggests, money is not the only or the most important motivator for people who take up public service.

Most people do not choose to work in government "to maximize their income," he writes. If they find they need more money, they can probably obtain it more quickly "by taking a second, weekend job at the local mall" rather than by striving for "the modest sum awarded in most public-sector pay-for-performance schemes."

Although Behn is not opposed to the principle of paying people for their performance, he emphasizes that in the government, the details matter. "To create a pay-for-performance system that actually motivates -- that does not demotivate everyone -- you have to get a lot of the details very, very right," he writes.

"And with the limitations that we citizens impose on our government -- limitations on both expenditures and on perceptions -- government has a very hard time getting the pay-for-performance details even close to right."