Jim worked in a Defense Department office with an employee whose lack of productivity was matched only by his hostile attitude. Eventually, a good manager with the patience of Job, a mastery of detail to match, and the help of higher management took the time to record, day by day, the offender’s record of non-work. After developing improvement plans for the employee and thoroughly documenting his failure to meet them for many months, the incompetent worker was actually fired. Then, the employee appealed before the Merit Systems Protection Board (MSPB). Two years into the appeal, when it looked as if the government would finally win, MSPB threw out the case when one of the team, moving to dismiss the employee, made an offhand remark that “even the people in his neighborhood association think he’s unstable.” The employee was reinstated with back pay. The agency went through the process all over again. By this time Jim had become boss. With the benefits of hindsight and existing records, the final try took only one more year of work!

Jim’s predicament is faced by thousands of hard-working federal employees who must suffer a small number of lazy, incompetent, and, occasionally, dangerous co-workers. For the past three years, I taught high level federal managers at the Federal Executive Institute (FEI). Despite the stereotypes about bureaucrats, the vast majority of federal managers are capable people who take pride in their work. In dedication and smarts, the government bureaucrats I’ve worked with are more than a match for the college professors I taught alongside during 15 years in the academy. Unfortunately, these good people are thoroughly frustrated by a personnel system which forces them to work alongside (and get the same raises as) a small number of turkeys.

One of my games in the civil service was to ask a lunch table full of federal managers whether it was possible to fire low-performing employees. Save for the ever optimistic personnel specialists, the usual consensus was that it was possible, but hardly ever worth the effort. My informal focus groups mirrored public employee surveys.

While career civil servants and political appointees do not always see eye to eye, mail surveys I conducted in the mid-1990s found that each side did agree that the federal personnel system is broken, at least when it comes to separating non-performers. Eighty-eight percent of Clinton political appointees and 83 percent of career managers agreed that “personnel rules make it too difficult to fire personnel”—more than half of each group strongly agreed. This dovetails with the findings of University of Georgia political scientist Hal Rainey. Reviewing years of survey data, Rainey reports that “[r]oughly 90 percent of the public managers agreed that their organization’s personnel rules make it hard to fire poor managers and hard to reward good managers with higher pay, while 90 percent of the business managers disagreed.”

Since courts have ruled that federal employees have property rights in their jobs, they can only be terminated after lengthy due process and multiple venues of appeal. As a result, managers who decide to use official means to deal with turkeys may face the prospect of spending all their time managing that one person, to the detriment of the rest of the office. As one of my informants, a manager in a regional office, recalled of his one (eventually successful) effort to separate an employee who was both unproductive and breaking the law: “As the year went on, it took more and more of my time and became all consuming. I had to spend a lot of time explaining why I was doing this.” Similarly, when asked if he had ever fired a non-performer, one manager told me that he had no intention of becoming like the one person in his agency who had fired several people and “wears that as a badge of honor, but he has no time to do his real work!”

Not only is firing a government employee time-consuming for managers, it is also dangerous. Managers who move against problem employees take serious risks. As Carolyn Ban details in How Do Public Managers Manage?, low performers facing personnel actions need not go gentle into that good night. They can make life difficult for months or years. They can take their case to the Merit Systems Protection Board (MSPB). They can file a grievance if they are covered by a union contract. Or if they belong to a protected class (by sex, race, age, or handicap), they can file an Equal Employment Opportunity complaint. Some bring their cases to the Office of the Special Counsel, claiming that they are being fired or otherwise harassed because they are whistle-blowers.

The problem of “low performing whistle-blowers” is particularly vexing for many federal managers, who, on the word of a single employee, can be subjected to prolonged investigations worthy of an independent counsel. Imagine having colleagues and subordinates questioned for months about whether you have ever used the long-distance line for personal calls or have padded the expense account. While the investigations go on, little work gets done and communication between co-workers stops, since everyone is afraid that a random remark could lead to a grand jury. Even when the boss is exonerated, a bad reputation can linger for years. And of course, if all else fails, the employee can simply sue his or her manager.

Not surprisingly, managers who have fired someone describe the process as traumatic. As one of the officials quoted above recalls, “here it is years later, the person has long since left the state, and I still don’t want to talk about that sorry episode. It has an impact upon both the office and the family life of the individuals involved. It’s something you only do once.”

With the deck stacked against them, federal managers tend to avoid using official means of dealing with their turkeys. Phone surveys of managers by the U.S. Office of Personnel Management (OPM), reported in Poor Performers in Government: A Quest for the True Story, found that only 7.5 percent of the managers of low performing employees moved to reassign, demote, or remove them, and 77.8 percent of those managers reported that the efforts had no effect. While OPM gives a rough estimate of around 65,000 poor performers in government, from September 1997 to September 1998, only 159 federal employees were removed by performance based personnel actions, with another 1,693 removed for issues other than performance, such as breaking the law. Federal managers suffer low performers or act informally to improve their work and never use the federal personnel system, or only use it as a last resort.

Not infrequently, federal managers use two traditional means of shedding non-performers. By writing glowing letters of recommendation, a boss can get a turkey promoted to a different office. Fortunately, most civil servants are too ethical to use such tactics, and anyway, you can only do that once or twice before your credibility in the bureaucracy is shot. More typically, bosses place non-performers in “turkey farms,” “dead pools,” or (if it is a single person) “on the shelf.” By quarantining non-performers, a good manager can save the rest of the organization from their influence.

The relative inability to act against a small number of poor performers has the effect of making that small number vexing to managers, who are people more used to solving problems than ignoring them. Managers don’t like using turkey farms, but many feel they have no choice. At least, when downsizing comes, known turkeys make good candidates for reductions-in-force.

The sad part is that the vast majority of federal personnel do a good job. OPM’s Poor Performers in Government report estimates that under 4 percent of federal civil servants are non-performers. One can quibble with OPM’s methodology. I suspect that the real figure is a bit higher, but still, the poor performer problem is not nearly as bad as most Americans think. So why deal with it at all?

Aside from inefficiency, there are four huge costs of the federal government’s inability to kick turkeys off the farm. First, non-performers themselves never get the message that they have to shape up or ship out and never get the chance for a new start. Instead, they often use the system to pursue old grudges for years. Second, good employees like Jim are forced to work alongside, do the work of, and often get the same raises as a small number of turkeys—a real morale killer.

Third, the low performance problem undermines the image and self-image of the bureaucracy. Business people have no tenure and look down on their government cousins who do. Military officers have an “up or out” promotion system—at key points in their careers officers either get promoted or get discharged. This makes the officer corps, at least at higher levels, a turkey-free zone, giving the brass a certain swagger in their dealings with their protected cousins in the civil service.

Most important, a tenured civil service undermines the legitimacy of the bureaucracy in the eyes of the public. After all, very few voters have tenure, so it is hard to tell citizens why their public servants cannot be fired. It is not surprising that college professors, public school teachers, and government bureaucrats have all come under attack in recent years. The very existence of tenure protects a small number of losers and leads the public to suspect the existence of a large number of low performers in government—a suspicion shared by many public managers.

Originally, the federal merit system was set up in 1884 to keep political parties from using government jobs to reward supporters. Patronage was seen as particularly onerous after President Garfield was assassinated by an insane “disappointed office seeker.” Presumably, government would work better if run by technical experts than by political hacks. But on the federal level, at least, the spoils system got a bum rap. Even in the 19th century, a new president and Congress kept most of the incumbent civil servants in place, and only rarely replaced those with special expertise. Politicians have never relished the unpopular task of firing old employees to replace them with political supporters. As politicians have long lamented, each new political appointment provides 10 enemies who themselves wanted the job and one ingrate who got it. Besides, for perfectly sound electoral reasons, politicians cared (and still care) about the efficient management of government. As political scientist Michael Nelson has pointed out, even in the 19th century, more voters sent mail than delivered it. A party that replaces all the mail carriers disrupts service—not a good way to win re-election.

The temptation to “politicize” a bureaucracy in search of pork is even less apparent today. In the old days, political campaigns were won by precinct workers who might welcome a federal job. Today, politicians depend on big kickbacks from campaign contributors rather than small ones from government employees. More important, the greater size and expertise of modern government makes it less susceptible to political takeovers. In the old days, it may have made some sense for politicians to hire friends to deliver the mail, but imagine if a modern president hired precinct workers to run the Pentagon and NIH? To think that politicians would raid the civil service, you have to assume that they have an incredible capacity for both venality and stupidity, and the time to exercise both. In fact, modern presidents can hardly handle the 3,000 political appointments they have now. How could they place more? Sure, presidential political appointments have grown in number since 1960, but not nearly enough to match the growth in congressional staffs, interest groups, and reporters—the people appointees deal with on behalf of their agencies. Unless we downsize the rest of the Washington political class—something not likely to happen—we can’t downsize political appointments.

Because of constant scrutiny by opponents, politicians in Washington do not have the sort of vast appointment powers they might have in some states and cities. Presidents are in fact very constrained in who they can appoint to government jobs and how many appointments they can make. The Washington Post test (“How will it look in The Post?”) limits what they can do. Congressional scrutiny limits what they can do. Inter- and intra-party battles limit what they can do. Not surprisingly, the worst abuses under spoils were in states and localities with little political competition, not in a two-party, hyperpluralist Washington.

To its credit, the Clinton administration has at least acknowledged the non-performer problem, and has begun to act on it. In accord with the Poor Performer study, OPM issued a CD-ROM guide to help managers, Addressing and Resolving Poor Performance. This is more than previous administrations dared try, but probably too little, too late. Not surprisingly, as the longtime guardian of the merit system, OPM’s basic inclination is to save a system that should probably be buried. Real change in the merit system requires legislation to simplify procedures, followed by years of culture change inside government, along the lines of the National Performance Review’s reforms of government procurement. The White House considered introducing a bill to overhaul the civil service earlier this year but dropped it in deference to public employee unions, an important constituency in the 2000 presidential primaries.

Real change is occurring on an ad hoc basis, however, in the agencies. Currently, the Federal Aviation Administration and Internal Revenue Service are creating their own alternatives to the traditional merit system. The Pentagon and Department of Housing and Urban Development have floated trial balloons proposing to replace most tenured civil servants with contractors and fixed-term employees who can be separated with relative ease—something that is happening incrementally all over government. Further, the rise of Performance Based Organizations, called for by Vice President Gore’s Reinventing Government reports, is likely to erode tenure by tying organization budgets and staffing levels to results. Indeed, on all levels of government, the reinventing movement is in part a way to use market mechanisms to get around civil service tenure. Declining demands for an organization’s outputs force reductions in force, which in turn push marginal performers out the door.

Many of the most exciting changes are happening in the states. More than 30 states now have serious proposals to reform their civil service systems, but Georgia has gone the farthest. In 1996 Georgia removed tenure from new employees in state agencies. A detailed evaluation of the Georgia experiment does not yet exist, but early work by political scientist Steve Condrey suggests that removing tenure has increased the ability of public managers to manage for accountability. A single Georgia state agency reported terminating nearly 200 employees for cause in the first 20 months after the law came into effect—with no reported challenges or cases of impropriety. Notably, the Georgia civil service reform was developed and pushed through the legislature not by anti-government Republicans but by then Governor Zell Miller, a pro-government New Democrat who sees ending tenure as one way to restore the legitimacy of government.

In short, for the first time in decades, real civil service reform is beginning to happen. For historic context, it took 20 years from 1864 to 1883 to build political support for the original federal merit system, and another 40 years to put it in place. It may take another decade to reform the federal merit system, but tenure now lacks public legitimacy, and the first steps at reform have already been taken. This excites those of us who want to turn “civil service” from an epithet for cumbersome personnel rules into a public service ethic.

Robert Maranto is a visiting scholar at the Curry School of Education at the University of Virginia and co-editor most recently of School Choice in the Real World: Lessons from Arizona Charter Schools.