Pay for performance back again

The federal government hasn’t had much success with pay-for-performance systems. The hoary General Schedule draws its share of criticism, too, but attempts to replace it have been unsuccessful.

Now there’s a new effort to revive pay for performance, led by Rep. Dennis Ross (R-Fla.), the new chairman of the House Oversight and Government Reform Committee's federal workforce subcommittee. Although he hasn't introduced legislation yet, Ross has begun holding hearings on the topic. On March 9, he brought labor advocates and small-government proponents to the table as witnesses, with Office of Personnel Management Director John Berry on hand to provide the government's perspective.

Ross dubbed his hearing "Are Federal Workers Underpaid?," which might be taken as a straightforward question or an ironic dig.

Berry refuted the potential irony in his testimony. "I have said before that the system is six decades old and could use a re-examination," he said. "But even if the system is not perfect, we must reject misleading uses of data that perpetuates the myth that federal employees are as a whole overcompensated."

James Sherk, senior labor policy analyst at the Heritage Foundation, staked out a position at the opposite end of the field. "The federal pay system unnecessarily inflates federal pay," he wrote in prepared testimony. "My research shows that, on average, federal employees earn hourly wages 22 percent higher than otherwise comparable private-sector workers." Sherk blamed that discrepancy on the GS system and the benefits that make up part of federal employees' total compensation.

Ross faces a political battle, but beyond that, he will have to figure out how to make pay for performance work in the government. While some individual organizations use pay-for-performance approaches, efforts to institute it on a wide scale have faltered. The most prominent recent federal experiment — the National Security Personnel System — failed in 2009 after limping along for a few years.

Just why is pay for performance such a difficult concept for the government? The cynical answer — and one that many proponents of pay cuts and staff reductions appear to believe — is that employees prefer to be rewarded for longevity rather than accomplishments because they're not interested in working hard. But other reasons are more plausible.

First, agency budgets are not particularly fluid. A manager has a certain pool of money available for raises no matter how many people deserve it. If the result of employee evaluations suggested that 70 percent of an organization’s employees merited raises but the funds available could cover only 30 percent, managers would have to withhold increases from people who deserved them.

Another difficulty lies in measuring performance. Government jobs are not always amenable to metrics, leaving managers to essentially guess at how productive a given employee is.

In commenting on a Federal Computer Week article on the topic, one reader wrote, “The root of the issue is there is no true performance measurement system in place. The idea of a pass/fail system is meaningless if there are no measures by which to gauge the effectiveness of the process and those it is intended to measure. Managers are poorly trained in this regard and, in the end, the employees are as well. Until this central problem is fixed, pay for performance is not quantifiably feasible.”

Still another challenge, another reader noted, is favoritism. If managers have discretion in who gets a raise and personally prefer some employees over others, favoritism — whether real or only suspected — can demoralize a team.

“Favoritism under NSPS was rampant, much worse than under the GS system, and would continue…under a new system,” the reader wrote. “From what I saw in my organization under NSPS, actual performance had little, if anything, to do with the evaluations and subsequent pay increases. NSPS was a total failure because it did not provide what it promised for most of the people who worked under it.”

Charges of favoritism don't have to be true to be toxic. The accusation can undermine a manager's authority and demoralize a team. On the other hand, real favoritism is even worse.

“Unfortunately, my years of government experience [have] taught me that pay for performance more often means pay for the boss' buddies,” another reader wrote.

In that reader's experience at the U.S. Postal Service, managers could only select 10 percent of their employees to receive merit raises each year.

"The guy I worked for thus had three performance pay [raises] to hand out, and all three went to his drinking buddies," the reader wrote. "He certainly wasn't going to waste the three he was allocated on someone who didn't drink with him — even if that someone was the only one in the unit who stayed sober long enough to do a halfway decent job.”