Next year, the GS system—its management philosophy and administrative framework—will be 100 years old.
The genesis of the federal pay system was a 1920 report that concluded government needed a "modern classification of positions to serve as a basis for just standardization of compensation." Three years later, Congress created what is today the General Schedule salary system. It remained largely unchanged, except for the creation of the Senior Executive Service, until 1990 and passage of the Federal Employee Pay Comparability Act.
Next year, the GS system—its management philosophy and administrative framework—will be 100 years old. Over the three decades since the enactment of FEPCA it’s become more and more fragmented. Program control is now spread over an amazingly disjointed array of stakeholders.
Today, any notion that federal employees are paid fairly and consistently under a unified salary system is a myth. Anyone arguing that federal salaries are too low or too high is grossly oversimplifying reality. Federal pay is best described as a hodgepodge of systems. Few salaries have been evaluated to assess the impact on agency performance.
For virtually every employer, payroll is the largest controllable expense. For government, the Office of Personnel Management's website, Fedscope, shows the GS system cash payroll is roughly $120 billion, based on 2018 data for almost 1.5 million employees. The total governmentwide payroll, including benefits, has been estimated at $290 billion. And that does not include compensation for contractors paid to do federal work. The core question is: Are these funds expended in a manner that contributes to good government?
The Fragmentation Keeps Increasing
The fragmentation started with the demonstration projects permitted under the Civil Service Reform Act of 1978. That was followed by the Financial Institutions Reform, Recovery, and Enforcement Act of 1989 that allowed the financial regulatory agencies to establish their own pay programs. Among congressional agencies there may be others but at least the Government Accountability Office and the Congressional Budget office have their own pay systems. Additionally, the Federal Wage System includes hundreds of locally controlled pay programs.
The big break from a unified system was passage of FEPCA in 1990 and the introduction of locality pay. Currently, if my count is accurate, there are 52 locality pay schedules (plus the “Rest-of-U.S.” schedule), and each year the Federal Salary Council is asked to evaluate proposals to add to the list. It's no secret that’s an effective strategy to win salary increases.
Equally important is the provision in FEPCA that permits OPM to adopt “higher rates of basic pay—special rates—for a group or category of . . . positions in one or more geographic areas to address existing or likely significant handicaps in recruiting or retaining well-qualified employees.” The provision has been largely ignored in discussions of federal pay. That provision was prompted by a concern in 1990 that government would have trouble competing in the future for patient care specialists.
When FEPCA was drafted the intent (as one of the drafters) was to permit creation of separate pay systems like that created under Title 38, but it's been interpreted differently. Today there are hundreds of narrowly focused pay schedules.
OPM’s website currently lists 155 job series with employees paid special rates in 63 agencies and 668 locations. The Washington metropolitan area alone has 15 special rate schedules over and above the locality pay schedule. Areas with special rates are as small as Adams, North Dakota, a town with a population of 127. The website is silent on how many employees are paid a special rate and at what grade and step, which makes it impossible to evaluate agency use of the special rate authority.
The special rate tables have a number of surprises. One is a list of 58 miscellaneous nonexempt support job series including sales store clerks, messengers and clerk typists. These are not jobs where high demand is driving up salaries. OPM plans to eliminate several. It’s also surprising to find essential healthcare occupations: nurse, nurse practitioner, practical nurse, physician assistant, and pharmacist. That suggests the Title 38 program needs to be reconfigured.
An added aspect of what FEPCA authorized is that each of the special rate schedules should be assessed and adjusted each year in response to market-driven pay increases. It’s not clear that’s happening.
There is no evidence agencies follow generally accepted salary management practices. Overall, the use of special rates is out of control. This would not be accepted in business.
The special rates provision was included in FEPCA because in 1990 the recruiting and retention of patient care specialists was a concern. Today, its cybersecurity—a field that did not exist in 1990. It's all too clear the GS system is not responsive to the changing demands on government and the need to attract and retain tomorrow’s skills.
More Questions Than Answers
Government’s piecemeal salary system exists in a vacuum; political leaders from both parties rarely express concern. In other sectors, compensation systems have undergone steady, evolutionary change but the GS system remains on paper immutable.
But the reality, substantiated by the fragmentation, confirms it is not meeting the needs of government.
On a series of evaluative questions, the GS system fails. On several key points, OPM has failed to undertake the analyses common with well managed programs.
- It’s generally believed some employees are overpaid while others are underpaid but until a true market analysis is completed that debate will continue. Today, government does not know which jobs are overpaid. That analysis is badly needed.
- It’s understood that some jobs are over graded. The facts are unknown, however. The problem has been largely ignored for years and presumably grown progressively worse. (A separate audit is needed to assess the magnitude of the over grading.)
- Starting salary level is seen as an issue in recruiting, affecting the number and quality of applicants. The facts are unknown. Assessing the impact of noncompetitive starting salaries would be a valuable study.
- A basic textbook step—identifying or defining the employer’s competing for talent—has never been considered. That step is basic to the FWS administration and the agencies with independent pay systems. The agencies created under the 1989 Financial Institutions Reform, Recovery, and Enforcement Act, for example, focus on financial industry data.
- In the mid-1990s the Bureau of Labor Statistics changed the basis for the survey used to assess the GS pay gap and thereby imposed a new pay philosophy. OPM’s response was limited to adding statistical “bandaids” to fill in holes in the data but the impact has never been fully assessed.
- The validity of the annual OPM/BLS analysis of salary survey data has not been evaluated for years. It’s never been audited to determine if it is consistent with accepted theoretical precepts.
- The pattern for promoting new hires to the top of career ladders increases salaries faster than in other sectors. It's not clear those increases benefit government.
- Salary level is a core factor influencing an employee's interest in becoming a supervisor but the rigid GS system forces agencies to ignore that issue.
- The GS program is based on century old thinking and has not been evaluated in decades.
The drafters of the 1978 Civil Service Reform Act four decades ago were aware the GS system did not meet government’s needs. The China Lake demonstration project in 1980 confirmed there is a better answer. That was a decade before corporations started replacing traditional salary programs. It was 30 years ago that OPM concluded a uniform, national salary system was not a viable answer for the future. The proliferating special rates confirm change is needed to enable government to compete for talent. With the heavy retirements and the acknowledged problems recruiting new graduates, the staffing problems are likely to get worse until the GS system is replaced.