Decoding Federal Employee Compensation

Congressman Mark Meadows, R-N.C., took a step forward in leading the hearing on federal employee compensation last week. He asked for the facts. Presumably he meant facts that both sides of this ongoing debate can agree on. The absence of readily understood, verifiable data has perpetuated this contentious debate now for over two decades.

He also said that civil service reform is coming. That undoubtedly means transitioning to pay for performance, although that will require agencies to solve what has been the weakest link in workforce management—employee performance management. Pay for performance will not be accepted if the basis for pay increases—performance ratings—are not seen as valid. Congressman Jim Jordan, R-Ohio, clearly enjoyed ridiculing all the A+’s, A’s and A-’s along with the incredibly low percentage of employees who are terminated.  

His sarcasm is justified. Last year in the Washington area, a highly respected survey conducted by the Human Resource Association of the National Capital Area shows that the typical employer rated only 19 percent of employees at the highest level, and 3 percent at the lowest level. (The 2016 survey had 228 participating employers from every sector encompassing over 300,000 employees.) Federal performance ratings are badly inflated.  More importantly, most federal employees believe their inflated ratings are accurate—it’s the others they don’t trust.

Every employee knows government’s recent pay for performance track record is not good. The policy has consistently worked well in demonstration projects but failed to gain acceptance when Defense rolled out the National Security Personnel System. There were also problems in the Government Accountability Office and in the Securities and Exchange Commission. But several states, counties and municipalities have successfully implemented pay for performance policies—it is possible to do in government. There are critics, of course, but they are largely limited to academia and the unions.  

The philosophy is deeply entrenched and solidly accepted in the business world but there are key differences. First, individual pay is normally confidential. Performance ratings are also confidential.  Second, corporate leadership teams have fortunes riding on company performance. The emphasis on good performance is pervasive in the culture. It’s a different world.

But the job of managing employees is the same in every organization. Money may not be as important for federal employees (that’s a broad generalization) but everyone wants to be treated with respect and valued. Employees want to be recognized. No employer can afford to ignore the best performers.

Good Management Is Essential

Managers are the keys to getting the best out of people. Research shows that managers have more impact on an organization’s performance than any other factor. Government needs to do a better job of developing effective managers. They need support in managing performance, especially in handling poor performers.

Performance management is undergoing a revolution. Experts advocate involving employees in the planning and establishing individual goals for everyone. Every employee should be able to state what she or he expects to accomplish. Frequent feedback and coaching is also important. When managers and employees are involved in the process, they will defend the system to peers and it makes it far more likely the system will be accepted.  

Also important to pay for performance is the belief that employees are fairly compensated. Fair in this context means that employees must believe their pay stacks up appropriately relative to their peers in government (adjusting for performance) and is in line with the broader market for similar jobs.

As a guess, only the analysts who prepare the annual Office of Personnel Management/Bureau of Labor Statistics gap analysis truly believe there is a 34 percent gap between government and civilian workers. The same is likely to be true for the recent Congressional Budget Office analysis. Moreover, no one can read either report and learn if they are fairly paid. Neither analysis relies on information or analytical methods used by any other employer. Neither produces verifiable facts.

That same Washington area survey reports market data for over 500 job titles (or ‘benchmark’ jobs), with federal jobs matched with more than 250 of the benchmarks. A cursory review failed to disclose any pattern that jobs were overpaid or underpaid by the percentages claimed by the Federal Salary Council or the Congressional Budget Office. Here are the median or typical salaries for five, selected to represent broad job families and at different levels in the Washington area from the HRA-NCA Compensation Survey:

• Administrative Assistant III/GS 6: Federal median: $45,792; Market median: $50,900
• Medical Technologist/GS 9: Federal median: $62,338; Market median: $65,600
• Systems Analyst II/GS 11: Federal median: $71,115; Market median: $75,100
• Electrician II/WG 10: Federal median: $60,063; Market median: $59,200
• Staff Nurse III/GS 12: Federal median: $92,987; Market median: $89,600

Possibly more important is data showing that new college graduates are starting at salaries well above GS rates. Other recent reports show the interest in government careers is declining. That should not be ignored.

Those are facts. Descriptive statistics, like means and medians, summarize facts. The analyses of both the BLS and CBO depend on inferential statistics; both rely on small samples and unverifiable conclusions.

This is only one city and one survey. If needed, there are multiple surveys available for Washington. That’s true in virtually every major metropolitan area. Relying on multiple surveys would strengthen the reliability of the data. But the key point is credible surveys are available across the country that would satisfy Congressman Meadows’ request for facts—and at a far lower cost than the BLS surveys.

The CBO analyses is no doubt sound as an academic study but it’s based on data that was never intended to be used for this purpose. The conclusions cannot be used to determine if a job is paid competitively. It went unnoticed in the hearing that the report is silent on locality differentials. It’s also based on simplistic assumptions about age and education. It’s a macro analysis similar to arguing the U.S. unemployment rate is as low as it’s ever been but ignoring the rural areas where factories have closed. Deciding if jobs are properly paid requires a micro, job by job analysis.

The BLS surveys are affected by similar weaknesses. They cannot state how much a job is paid in the labor market or how well a federal job is paid relative to other employers with similar jobs.

The hearing devoted a lot of time to discussing benefits. CBO of course is correct; federal benefits are more generous than the average employer’s. However, less than half of the companies surveyed by BLS provide any form of retirement plan and only 57 percent provide health benefits. Government is not competing for talent with mom-and-pop businesses. Their practices are not relevant.

The Washington area survey confirms the GS system is out of sync in many ways from the practices in larger companies. Labor market trends are well documented, led by increasingly scarce skills. Government needs the pay flexibility to compete for critical talent.