Noble Experiments

ong before anyone talked of reinventing government, many people had come to regard the federal personnel system as antiquated, rigid and unresponsive. Responsible managers believed that the rules severely restricted their ability to offer the incentives and awards necessary to attract and retain the highest quality workforce possible. So, nearly two decades ago, some analysts conceived the idea of improving personnel management through a program of small-scale "live" experiments, with successful innovations to be applied governmentwide. To the surprise of many, personnel officials managed to incorporate this concept, albeit with significant limitations, into a major piece of personnel legislation. Although sparingly used, the program has directly affected thousands of employees, has indirectly influenced the employment conditions of many thousands more and has once again become the subject of serious legislative debate.

The experiments-or demonstration projects, as they are designated in the statute-have been valuable, says Carol Okin, the Office of Personnel Management's associate director for merit systems oversight and effectiveness. "They stimulated a new way of thinking about human resources management and provided a channel for creativity in solving problems." Okin identifies several demonstration projects in the 1980s as "precursors" to the bonus and allowance features of the 1990 Federal Employees Pay Comparability Act (FEPCA). She also points out that other agencies have adapted the Navy's pioneering demonstration of pay-banding as an alternative to the traditional grade structure. OPM initiated changes in the law governing demonstration projects that would stimulate more and larger experiments and make successful changes permanent.

But the idea of using small tests to find the one best way to fix a problem governmentwide needs adjusting itself, according to Ronald Sanders, director of Syracuse University's Maxwell Center for Advanced Public Management and former Defense Department civilian personnel director. Agency problems differ so much and environments change so rapidly that agencies "need the ability to test and prototype for themselves or perhaps a few similar organizations," he says. "To do this effectively, they need more flexibility than the current demonstration project requirements of grand scientific design, second-guessing and extended evaluation allow." Sanders' vision of alternative agency personnel systems within a broad framework of fundamental governmentwide principles is explained in a recent Brookings Institution study, "Civil Service Reform: Building a Government That Works," which he co-authored.

The Carter Administration's version of civil service reform was the original vehicle for demonstration projects. A 1977 task force helping to lay the groundwork noted-in language equally applicable today-that far-reaching changes "involve high risk and will generate opposition from special interest groups." Anticipating the president and Congress would balk in the face of such opposition, the task force suggested "experimentation is more likely to be acceptable. It provides opportunity to time-test new concepts, modify them, and, as their feasibility is demonstrated, extend them incrementally throughout the federal establishment."

Buying into this approach, the 1978 Civil Service Reform Act gave the Office of Personnel Management authority to conduct or supervise projects "to determine whether a specified change in personnel management policies or procedures would result in improved federal personnel management." OPM could exempt agencies from most laws and regulations, except those applying to leave, benefits, merit principles, equal opportunity and limitations on political activity. Employee unions were given negotiation and consultation rights on matters affecting their bargaining units. Demonstration projects were limited to 5 years and 5,000 employees, and no more than 10 were to be active at one time.

Let the Projects Begin

Navy installations at China Lake and San Diego, Calif., were authorized in 1980 to conduct the first demonstration project. In a drive to improve recruitment and retention of high quality employees and enhance the powers of managers, the project simplified the job classification system (via pay-banding, or broad-banding, as it is also known), tied pay to performance, and increased flexibility for starting salaries. Nancy Crawford, a Navy psychologist who has worked on the project since 1983 and has been its coordinator since 1988, thinks the experiment is a success. "Local management has provided strong support and OPM evaluators have been excellent and cooperative. They have shown willingness to work around regulatory barriers when necessary."

But Crawford says the project pointed to flaws in the CSRA provisions: The five-year limit may be insufficient for completing and evaluating the demonstration, and no means short of legislation exists for making some personnel changes permanent. "You're left with the question of what to do at the end of five years if you want to continue what you're demonstrating. In our case, it took a great deal of work through the legislative process to get our extensions and, since we're still a demonstration project after 16 years, our status remains somewhat unclear," Crawford says. After twice granting the Navy project 5-year extensions and lifting the 5,000-employee limit to its current 12,000-plus, Congress authorized an indefinite extension in 1994.

The Navy's experimentation with pay-banding and other pay flexibilities was adapted in demonstration projects conducted by the Commerce Department's National Institute of Standards and Technology (NIST) and the Air Logistics Center at McClellan Air Force Base, Calif. Closely followed by the personnel community, the progress of the pioneering work by these agencies led several other departments to embark on such demonstrations, and ultimately influenced the drafting of the recruitment, retention, and relocation bonus and allowance features of the 1990 FEPCA. Pay-banding, adopted in June 1989 under separate legislative authority by the General Accounting Office and now fairly widespread in the private sector, is frequently recommended by study groups as a solution to bureaucratic impediments to better human resource management.

Government experiments and other studies, however, have been unable to demonstrate that pay-banding can be instituted without cost escalation, although agency project staffs and OPM evaluators differ on the significance of the data. The budget-pressed Congress is unlikely to broadly approve pay-banding as long as it's associated with cost increases.

The NIST demonstration at headquarters in Gaithersburg, Md., and the laboratories there and in Boulder, Colo., offers an example of another type congressional involvement. A blue-ribbon panel headed by former Deputy Defense Secretary and Silicon Valley pioneer David Packard had studied federal laboratories in the early 1980s and recommended a new and more flexible personnel system for their employees. When the idea failed to make headway governmentwide, an appropriations bill was used as a vehicle for legislating such a system for only NIST, formerly the National Bureau of Standards. The House Post Office and Civil Service Committee modified the legislation to instead direct a demonstration project in cooperation with OPM, and the experiment began in 1988.

The demonstration used a simplified and delegated classification and pay system to address hiring, classification and compensation problems, but it also broke new ground with a pay-for-performance system which uses only critical performance elements and restricts pay increases to those scoring above a stipulated level on a 100-point scale. The experiment first employed five performance levels, but was modified to two levels in 1990. While the system has the overwhelming approval of supervisors and employees, who are surveyed every two years, the mid-course change points to an unresolved issue. More scientifically inclined observers believe conditions should be kept constant throughout the term of a demonstration project to assure the validity of results. Others say time is needlessly wasted when elements in need of refinement are left unchanged and useable results are unnecessarily delayed.

Allen Cassady, chief of NIST's Demonstration Project Office, recalls much debate on "tinkering" during a demonstration. But he believes the process has worked well. OPM facilitated exemptions that were needed, and the personnel system changes overall have had "a very positive impact on NIST personnel," Cassady says. He, too, laments the lack of a mechanism for making successful changes permanent, noting that Congress had to step in and extend the NIST project indefinitely in March.

Has any project ended when its five years were up? In 1993, 15 years after permission for demonstration projects was granted, the McClellan Air Force Base project known as Pacer/Share finished on schedule despite a major reorganization which brought the Defense Logistics Agency in as a co-sponsor of the project. Pacer/Share tested gainsharing, based on organizationwide quality and productivity measurements which replaced individual ratings. Employees earned productivity payments of $1,924 per employee based on cost savings during the project.

However, organizational performance did not improve significantly and the personnel system changes produced mixed results. Nevertheless, some successful aspects of the experiment might be adopted at McClellan, and cooperation between management and employee organizations improved dramatically after their close partnership in designing and coordinating the project.

The first project to terminate, albeit after an OPM-granted extension for evaluation, was a Federal Aviation Administration test of a recruitment method for restaffing after the 1981 air traffic controllers strike. The agency developed a four-year airway science curriculum in cooperation with a university association, but it produced only 41 hires after seven years and was shut down.

Evaluating the Process

A December 1992 Merit Systems Protection Board (MSPB) report and an OPM "Retrospective" the following year reviewed demonstration authority. Both found positive results, with MSPB asserting the experiments "have helped lead to beneficial changes in federal personnel management," and OPM saying they "have had a significant impact." But with only a half dozen projects launched since 1978, both reports concluded that OPM needed a stronger leadership role in identifying concepts appropriate for testing, soliciting sponsors and streamlining the cumbersome design and approval process. Both reports said OPM should encourage replication of demonstrations in several agencies to better determine validity for the general government population.

Another common recommendation was that statutory limits, such as the 5,000-employee ceiling and the prohibition on demonstrations involving benefits and leave, should be lifted. Interestingly, the House civil service subcommittee's 1996 civil service reform bill originally contained language modifying both provisions, but the ban on benefits experiments was reinstated after employee organizations complained. An authorization of up to 15 demonstrations, with no more than five covering more than 5,000 employees each, remained in the bill, which was passed by the House on July 25.

The reports differed on the best means for changing personnel rules. MSPB recommended OPM vigorously pursue appropriations for agencies willing to sponsor demonstrations, but OPM favored developing other vehicles to promote permanent flexibility and delegation of authorities. One such mechanism, authorizing OPM to waive regulations and approve "alternative systems" for selected agencies, was included in the Clinton Administration's human resources reform legislation submitted last year, but ran into opposition in the House.

A significantly larger role for OPM seems unrealistic in light of staffing pressures on the agency. Barbara Swanson heads a six-person demonstration projects team within Okin's organization. As recently as a year ago, 11 employees comprised the project development and evaluation teams.

Other Roads

There is ample evidence of other methods of testing rule changes. The FBI was authorized by legislation to conduct and evaluate a demonstration of the effects of retention allowances and relocation bonuses on severe staffing problems in its New York office between 1988 and 1993. OPM and the FBI jointly submitted the required annual reports to Congress, which concluded that the major objectives had been met.

More recently, the fiscal 1996 Transportation Department appropriation gave the FAA permission to create a new personnel system, as well as acquisition reforms. Agency task forces went to work streamlining hiring and promotion procedures, standardizing job descriptions, structuring gainsharing programs, and revising grievance, awards, downsizing and other processes.

In fact, the individual agency legislative route may become common under the "performance-based organization (PBO)" concept advocated by the National Performance Review. The PBO approach urges agencies to seek congressional relief from legislative, regulatory and systemic constraints in return for more precise accountability for performance and program results.

Demonstration projects remain the appropriate mechanism for PBOs and for other agencies who need exemptions from personnel regulations. Okin, mindful of the strain of downsizing and tight budgets, says her staff is committed to "help agencies make innovations outside the formalities of the demonstration project process whenever possible." In addition to providing consulting services, her staff has published a "Template of Personnel Flexibilities for Use by Agencies Selected for Conversion to Performance-Based Organizations."

Agencies struggling to remake themselves in the turbulent 1990s often find themselves handicapped by personnel system restrictions. For them, a liberalized demonstration project process may be a tool worth considering.