Do Human Capital Comparisons of Private and Public Sector Compensation Levels Have Value?

The views expressed are those of the author and do not necessarily reflect the views of ASPA as an organization.

By Howard Risher
June 9, 2017

The Congressional Budget Office (CBO) recently released a study comparing federal and private sector compensation. It is one if many research studies extending back to the early 1970s based on human capital theory. The analysis shows:

• Federal workers with a bachelor’s degree earned 5 percent more than their counterparts in the private sector;
• Those with no more than a high school education earned 34 percent more; and
• By contrast, federal workers with a professional degree or doctorate earned 24 percent less.

The study also compared the cost of federal and private sector benefits, which when added to the salary estimates increased the gap. The total compensation gap for high school graduates, for example, increased to 93 percent.

The conservative think tanks (Heritage and American Enterprise Institute) have periodically completed similar studies, with similar conclusions. They contend the analyses support the need to reduce government pay and benefits.

Similar studies have compared state and local compensation levels with pay in the private sector. The conclusions have been very similar.

Each year the US Office of Personnel Management (OPM) working with the Bureau of Labor Statistics (BLS) completes a comparison of federal white collar salaries with supposedly similar jobs. Their analyses show federal salaries are 34 percent below private sector levels. They also rely heavily on inferential statistics.

Both the CBO and OPM/BLS analyses are based on methods and databases that differ dramatically from the common, textbook approach used in other sectors as well as state governments which complete so-called market analyses. When the CBO study was discussed recently at a congressional hearing, the chair, Representative Mark Meadows, concluded he wanted the “facts.”

His request is telling. It highlights an inherent problem with multivariate analyses. They do not generate facts. Plus, the methods go well beyond what students learn in basic statistics courses. The switch to multidimensional space makes eyes glaze over quickly.

It is relevant here that a highly regarded salary survey in the Washington, D.C. area (based on data for over 300,000 employees) provides job by job comparisons of federal pay that fail to support either the CBO or OPM/BLS conclusions. The survey provides the local pay ‘facts’ for over 500 common jobs.

Gary Becker’s classic 1964 book, Human Capital, introduced the framework for assessing the linkage of education and income. Similar analyses supported the arguments for pay equity starting in the late 1970s. The first comparative study of federal and private sector pay was Sharon Smith’s dissertation in the 1970s. Her articles prompted others to study the issue. The basic methodology has become a staple for academic research.

But not one of the studies provides the “facts” to satisfy Congressman Meadows. For public policy decisions, the CBO conclusions are not practical answers. Government is not going to reduce the pay of all employees with only a high school diploma by 34 percent. The OPM/BLS conclusions have been ignored by Presidents for more than two decades. Today the conservatives in Congress would ridicule a proposal to increase federal salaries by even 5 percent.

An underlying problem with the CBO studies (and other similar analyses) is the database, the Current Population Survey (CPS). It was never planned for compensation analyses. In using the data, researchers effectively ignore universally accepted thinking in salary management. One that jumps out is that all undergraduate or graduate degrees are equal. The researchers certainly know that a degree in any field of engineering commands a higher salary than one in history or, I hesitate to say it, one in political science. The analyses also assume age is an indicator of work experience.

In some studies, occupation is included but the CPS reports only very broad groups (e.g., “Life Sciences, physical and engineering occupations”). That includes job families that attract very different salary offers from employers. Other groups also combine ‘apples-and-oranges.’

One of the most important problems is the reliance on samples drawn from the total U.S. population. The non-federal workforce exceeds 140 million. The number of business “establishments” is roughly 7.7 million but 5.7 million have less than 10 employees. The number with 500 or more account for one quarter of one percent. Research shows larger companies pay higher salaries which should attract better qualified applicants. The pay levels in the many mom-and-pop businesses are not relevant to government.

A similar problem affects the relevance of the benefit cost comparisons. 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. If the comparison was limited to the benefits provided by Fortune 500 companies, the conclusion would be very different.

Employers in every other sector complete comparative analyses of pay every year or two. They rely on “benchmark” salary surveys, with pay reported using a simple format very similar to that used with faculty and administrative salaries in higher education. The summaries rely on descriptive statistics—verifiable facts—which are readily understood.

A core question forgotten by CBO is the understanding of the relevant labor markets. That is the reason industry associations sponsor annual surveys. There are possibly more than 100 hospital survey conducted annually. Hospital labor markets are largely local so naturally they want to understand the compensation levels in their local area. This practical labor market ‘theory’ dominates salary planning in other sectors.

This certainly does not to deny the value of human capital theory. But when researchers ignore core policy issues, base their analyses on simplistic models, or rely on data unsuited to the purpose, the results cannot be trusted. In this case the debate has serious implications for careers and family budgets. The headline of a Washington Post headline as this is written includes the phrase “overly generous compensation.” That is attributable to a poorly conceived analysis. We need to do better.

---

Author: Howard Risher has 40 plus years of experience as a consultant to clients in every sector. He has a BA in psychology from Penn State and an MBA and Ph.D. from Wharton. He is the co-author with Bill Wilder of the new book, It’s Time for High Performance Government: Winning Strategies to Engage and Energize and the Public Sector Workforce. You can reach him at [email protected].

(No Ratings Yet)
Loading...