A Way Out

arta Cehelsky spent a decade walking a political tightrope at the National Science Foundation. Last summer she fell off, losing her job as executive officer to the National Science Board, NSF's oversight body. She departed soon after the board elected a new chair and NSF Director Rita Colwell decided it was time for a change. But Cehelsky, a former professor of comparative politics, isn't peddling pencils on the street corner. Instead, NSF is paying her to work at the Inter-American Development Bank in Washington, where she's helping bank officials add science and technology to their lending portfolio. Her assignment was made possible by the 1970 Intergovernmental Personnel Act (IPA), which allows people to temporarily join the federal government and lets government workers remain on their agencies' payrolls while taking jobs elsewhere.

According to a 2001 analysis by the General Accounting Office, NSF is by far the heaviest user of the so-called external IPA program. And it is perhaps the only agency whose list of external IPA participants includes many senior managers. NSF also makes liberal use of the IPA to hire people temporarily from outside government to ensure that its staff remains on the cutting-edge of science. "It's an excellent program when it's used properly," says Rosslyn Kleeman, a former GAO official now serving as distinguished executive in residence at The George Washington University in Washington. But external IPA assign- ments are vulnerable to abuse, adds Kleeman.

Indeed, GAO's September 2001 report (GAO-01-1016) on external IPAs was prompted by NSF's 1999 decision to grant a three-month IPA assignment to education chief Luther Williams after he was fined $20,000 for violating rules against accepting payment for outside activities connected to his job. (Williams later retired from the government and now directs educational programs at the Missouri Botanical Gardens in Saint Louis.) GAO concluded that NSF was following IPA rules. But the investigation and report revived a chronic concern among federal human resources officials that the IPA provides an irresistible opportunity for agencies to sideline employees they no longer want. The list of senior managers who have left NSF in recent years through external IPA assignments includes the heads of research directorates, the public and legislative affairs office, the information management organization, and the office of the inspector general.

Former NSF Director Neal Lane acknowledges that the program gives agencies the chance to move out people who no longer fit into their plans. "Often it's people who want to try something else," says Lane, now a university professor at Rice University in Houston. "And they don't often come back. I wouldn't call it an outplacement service, but I don't think that it's bad if those people decide that they can serve the government in another way, or use their talents better outside government."

The GAO report cited NSF as "one of the most active users of the [IPA] program in recent years," reporting that 13 to 22 NSF employees were on external assignments at any given time between 1996 and 2001. NSF with 1,300 employees, far surpasses any other agency in per capita IPA usage. It even holds its own with such bureaucratic behemoths as the Army, with 220,000 employees, and the Health and Human Services Department, with 65,000. In 2000, the Army had 14 employees on IPA assignments, while HHS had 13. More recent figures show the pattern holding steady at NSF. In April 2002, 23 employees-nearly 2 percent of the agency's workforce-were toiling somewhere other than at their desks inside the foundation's Arlington, Va., headquarters.

NSF's high rate is unusual for federal agencies, which pay a premium for external IPAs. Indeed, GAO calculated that the average external IPA assignment costs NSF $207,000 a year. Agencies must cover regular salary and the cost of a replacement for each employee on an IPA assignment. The host institution typically pays a small percentage of the cost. As a result, says Kleeman, simple economics restricts the number of external IPA assignments. "Budgets are so tight," she says, "that most agencies simply can't afford to pay double."

Under IPA rules, employees are supposed to pay back relocation, travel and incidental expenses. But the rules also give agencies discretion to waive those payments. GAO found NSF routinely used waivers. At NSF, more than 40 percent of those on IPA details leave the agency for good before they have completed the required payback period. In only one of seven instances did NSF collect any money it was due.

GAO also found much greater IPA participation among high-ranking, experienced workers at NSF than at other agencies. The average external IPA assignee at NSF has 15 years of experience, GAO said in its report, and nearly one in three is a member of the Senior Executive Service. That top-heavy distribution is unusual: Few executives choose to go on IPA assignments, says Kleeman, because their jobs "may not be there when they come back."

The Environmental Protection Agency is the only federal agency that approaches NSF's enthusiasm for the program. GAO found that 108 of EPA's nearly18,000 staffers-or 0.6 percent-were on external IPAs in 2000, far short of NSF's 2 percent. Moreover, the profile of the typical IPA assignee at EPA is quite different. Most are dispatched to help local and state governments deal with the agency's regulatory mandates, a direct extension of the agency's mission. And none is at the top of the pay scale. "We've never sent out anybody at the SES level. They're just too valuable a resource and too hard to replace," says EPA human resources official Raymond Hall.

FREEING THINKERS

NSF relies on a continuing stream of temporary help, typically from academia, to identify and fund the most exciting science. The foundation usually reimburses universities for the salaries of incoming IPA assignees, whom NSF calls "rotators." The scientists keep their academic status and return to home base in a few years with greater knowledge of their field and of how NSF operates. Of 600 scientists and engineers on NSF's payroll, almost 40 percent are rotators.

That constant infusion of talent creates a work environment more akin to an academic setting than to a government agency. "NSF is like a university," says John Wilkinson, head of the agency's human resources department. In both places, he says, the idea of providing employees with opportunities to recharge their intellectual batteries is well accepted. NSF's IPA assignments are the equivalent of sabbaticals, he says, and are awarded only after the employee makes a persuasive case for their value. Because NSF doesn't have its own labs, Wilkinson adds, external assignments offer NSF program managers rare opportunities to engage in research. "We are interested in furthering science, and that's done out there, not in here," he says. The exact mechanism doesn't matter, Wilkinson adds, noting that NSF also designates people as detailees to other federal agencies and runs a smaller program, called Long Term Professional Development, with the same goals. "The IPA program is tailor-made for us because of our links to academia or the nonprofit sector," he says.

Lane, a former provost at Rice, says that as NSF director, he was willing to run the risk that some people might not return. "I accepted their word if they told me that they planned to come back," he says. "But I also knew that they might be looking around [for other jobs]. And there are times when you don't mind if they don't come back."

The summer of 2002 was an especially busy time for senior NSF managers to flee Arlington and settle elsewhere with an IPA or other assignment. In May, physicist Robert Eisenstein, the head of NSF's largest research directorate, flew off to Geneva for a temporary assignment at CERN, Europe's high-energy physics laboratory. In June, NSF's chief information officer, Linda Massaro, took an assignment at the National Defense University to help launch a new e-government training program for federal workers. And in July, National Science Board executive officer Cehelsky went to the Inter-American Development Bank on an IPA. "We lost a person we respected and one who had a lot of experience," says Diana Natalicio, president of the University of Texas, El Paso, and vice chair of the science board.

All three had gotten on the wrong side of the NSF director, according to several sources familiar with the circumstances, and all sought a change in scenery as a way to resolve the situation. That's not unheard of in large organizations, says John Palguta, who joined the recently formed Partnership for Public Service after spending 34 years at the Merit Systems Protection Board. "I've seen cases where a director of an agency says, 'Look, I don't want you here and you probably don't want to be here. But I can't fire you. So why don't you find another job, and in a few years we'll see where we stand,'" says Palguta, who retired from MSPB as the head of its personnel research arm. Still, Palguta says that the departure of three top aides from a relatively small agency within a very short time raises questions about its management practices.

Several other top NSF managers also have used the IPA program in recent years to extricate themselves from uncomfortable circumstances. Linda Sundro, NSF's inspector general for almost a decade, took an IPA to study the nonprofit sector shortly after Colwell was nominated to be NSF director in 1998. Sundro moved to the Office of Personnel Management, and in 2001, she left the government to become a consultant. Julia Moore, formerly head of NSF's legislative and public affairs office, has been at the Woodrow Wilson Center, an academic think tank in Washington, since March 2000, writing a book about public attitudes toward genetically engineered crops. A former State Department official, Moore is nearing the end of her three-year IPA assignment.

Other high-level NSF employees have used IPAs to mark time while they assess their professional circumstances. Judith Sunley served almost two years as acting head of NSF's billion-dollar education directorate before Judith Ramaley, a former university president, took the job in the summer of 2001. Sunley, a 22-year NSF veteran, spent 2002 at the State University of New York, Stony Brook, working on a project to improve undergraduate education. "There was a natural break in what I had been doing," Sunley says, "and [the IPA assignment] gave me a chance to think about what I might want to do for the next 10 years after I return to NSF."

Hugh Van Horn, former head of the astronomy division, saw a two-year IPA as a chance to work on a long-delayed textbook and a monograph on white dwarf stars. The 64-year-old astronomer returned to NSF's materials science division last summer to oversee large research facilities that the foundation supports. But Van Horn says that he plans to retire once he has completed his two-year payback period.

Although the IPA program is intended to prepare employees to be more productive when they return, some NSF program managers have discovered instead that time away from NSF brought disenchantment with previous jobs. Julia Clark went to work in 1997 for Rep. Cynthia McKinney, D-Ga., after NSF ended a teacher preparation program she had been running. After returning briefly to NSF, she took a second IPA assignment to join the staff of Rep. Bob Etheridge, D-N.C., who serves on the House Science Committee. "I really like what I'm doing now, and I'd like to keep doing it, but I don't know if I'll have the chance," says Clark, whose latest congressional tour ended in December.

For Ivan King, who spent 30 years in the military before joining NSF in 1992, a four-year assignment at George Mason University in Fairfax, Va., provided a chance to develop a newfound interest in conflict resolution. "I made a compact with God that, if he gave me a Ph.D., I'd use it to teach people the opposite of fighting and shooting," he says. But King's search for an academic position hasn't panned out. Instead, after trying to land a position at several area universities, King returned to NSF in September and immediately went on a brief paternity leave before picking up a temporary assignment. Both King and Clark say they hope to find something more to their liking than what NSF so far has offered them.

The IPA has become part of a professional odyssey for Jennifer Bond. For many years, Bond oversaw NSF's most visible product, an authoritative biennial statistical compendium of U.S. scientific efforts called "Science and Engineering Indicators." In 2000, Bond decided it might be fun to see the world through the eyes of one of the volume's most important users-Congress. She landed a spot on the staff of Sen. Joseph Lieberman, D-Conn., where her one-year appointment was extended.

Last spring, as her assignment was ending, Bond prepared to return to NSF's science resources statistics division. But her boss had filled her old job, leaving Bond with a newly created position of senior adviser for outreach. Bond decided to take another IPA assignment, this one at the Council of Competitiveness, where she heads up a global science initiative at the Washington-based, business-oriented consortium that monitors U.S. scientific and technical prowess.

NOT AN ESCAPE HATCH

Although NSF has no agencywide policy on stepping off the grants-making treadmill, some directorates strongly encourage their scientific staffs to do so after several years. Two NSF directorates-biology and mathematics/physical sciences-even set aside funds to hire temporary replacements so that staff can take external assignments. "I don't like people who are responsible for spending taxpayer dollars not being close to the science," says Mary Clutter, NSF's biology chief. "So I reserve one [position] a year for people to go on leave, and they compete for the opportunity."

An external IPA assignment provides staff with an ideal way to stay close to science, participants say. "I'm always asking myself, 'What's the next big thing?' '' says Michael Foster, a division head in NSF's computing directorate. Foster has spent more than two years at the Defense Advanced Research Projects Agency (DARPA), which recruited him to set up a program on quantum computing. "Part of the Pentagon's mission is to predict the future, and we're trying to learn which technologies will succeed," says Foster. The project is attempting to aggregate expert opinion on topics by creating electronic trading systems that let participants set market values for various ideas. Foster plans to stay at DARPA until the spring, before returning to a new assignment at NSF. His contribution is so valuable to DARPA that the agency picks up his salary-a rare act of beneficence for a host agency.

Aquatic ecologist Penny Firth spent last year pondering a different sort of challenge-trying to find innovative ways to teach America's children about the environment. She developed online lessons that integrate scientific concepts with real-world events-from the ecological impact of cities to how passenger pigeons helped fertilize the American West-during a one-year stint at the American Association for the Advancement of Science. A 10-year veteran of NSF's biology directorate, Firth tried twice to take leave before getting an IPA, and she's grateful for the chance to try something different. "The workload is very high, and there's a significant burnout factor," says Firth, who returned to NSF in October. "You need a chance to renew yourself."

Firth's boss is not surprised she came back, invigorated and full of new ideas. "Everybody we've sent off has returned," Clutter says about the dozen employees who've taken advantage of the de facto sabbatical program she created after becoming head of the biology directorate in 1989.

Whichever way the assignment is managed, the goal of IPAs is to have trusted employees take important jobs elsewhere for short periods and come back with new skills. The program is not supposed to provide an escape hatch for unhappy employers or disillusioned employees, says Kleeman. An agency should confront the situation directly if things go sour, she says, rather than sidestepping the problem by sending the employee off on an IPA assignment. "A good manager should have the guts to get rid of someone for cause," Kleeman says. "And if the employee hasn't done anything wrong, then the agency shouldn't use the [IPA program] in that way."