College Labor Shortages in 2018?

By Paul E. Harrington and Andrew M. Sum

The Georgetown Center on Education and the Workforce has engaged in a highly publicized campaign claiming that the nation will face a very substantial deficit of college graduates by 2018 if the American postsecondary system fails to rapidly expand the number of college degrees it awards each year. Indeed, the employment projections developed by Anthony Carnevale and his colleagues at Georgetown University suggest that there will be a shortfall of 3 million college graduates by that year. Such a labor shortage, if it were to actually materialize, could result in an enormous amount of lost production, reduced incomes in the U.S. and a deterioration in our competitive position in the world economy. Firms unable to hire domestic college graduates might shift their production overseas or look to postsecondary institutions abroad for new sources of high-end labor supply to meet the shortfalls predicted by the Georgetown authors.

So what is the evidence of future labor shortages in college labor markets? Should higher education institutions place some bets in terms of organizational structures and resource allocation in response to this projection of a serious labor shortfall? These are important questions since the higher education system has been burned by faulty projections in the past. Perhaps, the most egregious example of this was the “college enrollment crisis” that was forecast by a number of observers in the early 1980s. At that time, some college analysts expected that by the end of the ’80s, postsecondary institutions would face large enrollment shortfalls, as the size of the high school graduate cohort was forecast to decline sharply through the mid 1990s as a consequence of the baby bust generation coming of age. While the number of graduating seniors from the nation’s high schools did indeed decline, no enrollment crisis occurred. Instead, higher education experienced a renaissance from its 1970s doldrums, with increased enrollments and sharp rises in tuition and fees-signaling the effects of a sharp increase in demand for college degrees as the economic gains from completing college rose sharply.

While decline, consolidation and merger were the watchwords of the “enrollment crisis” proponents, colleges and universities in fact prospered over the period when shortfalls in enrollment were expected. While the forecasters got the demographics right, they didn’t account for changes in the nature and magnitude of job growth that favored those with more years of formal schooling. Thus, they missed the rise in college enrollment rates that would take place among high school seniors and the sharp growth in college enrollments among adult women that occurred during that time period; both associated with sharp increases in the earnings advantages of graduating from college.

The Georgetown analysis begins with a rejection of the better-known and well-documented industry and occupational employment projections developed biennially by the U.S. Bureau of Labor Statistics (BLS). They argue that the employment projections of BLS sharply underestimate the future demand for college graduates. They note that when they compare earlier BLS forecasts of employment growth between 1988 and 1998 with projections based on their own method for the same time period, that

“The Bureau under predicted how many workers in the U.S. labor force would have associate’s degrees or better by 19 million. That projection was off by 47 percent. Our methodology for that same period over predicted post secondary educational demand by about 2 million workers or just 4 percent.”

They go on to argue that the BLS underestimates of projected college graduate demand “… encourage a consistent bias against investing in postsecondary education.”

Could BLS, the most objective, impartial and certainly data-rich observer of American labor markets, so grossly underestimate the projected demand for college graduates for such a relatively short time horizon? Our answer to this is no! Instead, after a careful review of their data and methods, we find that the Georgetown authors radically overstate the size of the college labor market and, in the process, ignore perhaps the most pressing problem facing college graduates in the nation today—malemployment. A concept used by Frederick Harbison in his 1973 book titled Human Resources and the Wealth of Nations, malemployment represents the inability of a college graduate to find a job that effectively uses the knowledge, skills and abilities acquired in college and relegates them to employment in low-skill and generally low-wage occupations that don’t utilize college-level proficiencies. Since the skills of the malemployed remain largely unused by employers, they experience considerable wage losses.

Unlike virtually any other analyst of labor market activity, the Georgetown authors define the size of the college labor market as equal to the total number of college graduates that are employed. BLS and most other college labor market analysts define the college labor market as essentially a set of occupations that most often require persons to earn a college degree in order to be fully qualified for employment in that occupations. As a rule of thumb, we could define the college labor market as being composed of professional, technical, managerial and high-level sales occupations (like bond and stock sales representatives or commodity brokers), although BLS uses a much more careful approach than this.

Perspectives on occupations

To understand the difference between the Georgetown and BLS approach, let’s compare data on two occupations: bartender and compensation and benefits manager—jobs that most readers have some familiarity with in either their personal lives or professional capacities. An analysis of the data provided in Table 1 (derived from the U.S. Census Bureau’s American Community Survey) reveals that workers in both occupations have varying levels of formal schooling. But a closer look at the data reveals that most bartenders don’t have any type of college degree (only one in four bartenders report they have graduated from college). In contrast, compensation and benefit managers are much more likely to have finished college. More than six of 10 compensation and benefits managers have a college diploma, and nearly one in five have obtained an advanced academic degree.

Table 1: Mean Annual Average Percent Distribution of Employed Persons Ages 25+ by

Selected Occupation and Level of Educational Attainment, 2006, 2007 and 2008 Averages

Educational Attainment Bartender Compensation and Benefit Manager
Less than High School 9.7% 2.3%
High School Only, Diploma or GED 33.2% 12.5%
Some College, no degree 32.7% 21.7%
Associate Degree 8.5% 7.4%
Bachelor’s Degree 14.4% 36.3%
Master’s Degree 1.2% 17.8%
Doctor’s/First Professional Degree 0.3% 2.0%

Source: American Community Survey, 2006-2008 public use files. Tabulations by authors

When we look at existing data on the skill requirements needed to work in either of these occupations, we find similar disparities between bartenders and benefit and compensation managers.

The U.S. Department of Labor, Employment and Training Administration’s O*NET system is a massive database compiled over the past 15 years on occupational skill needs in the U.S. economy. Developed and maintained for the express purpose of understanding the education, training and work experience requirements of different occupations found in the nation’s labor markets, it is designed to inform workforce development, education and training professionals, higher education leaders and the business community about the wide range of skills needs within hundreds of individual occupations in the U.S. labor market. An examination of the O*NET studies of these two occupations reveals sharp differences in education and skills requirements.

O*NET assigns the bartender occupation to job category two, while compensation and benefit managers are assigned to job category four. What the significance of these assignments? O*NET studies of the bartender occupation found that the fundamental educational qualification for employment is a high school diploma and that bartender skills are largely acquired though work experience (although there are training schools including Harvard University, where one can prepare for employment in the occupation). English language and math skills required for work in this field were assigned relatively low values. In contrast O*NET studies of the compensation and benefit manager occupation found that most employers require a college degree for initial qualification for employment along with a considerable amount of work experience directly in the human resource and compensation fields. The English language and math skill requirements for this occupation are considerably higher than those for bartenders and the occupation requires a high degree of specific knowledge of human resource principles and procedures.

Taking the data on the distribution of employment by occupation and the findings from the O*NET studies of the skill requirements of both occupations, it would not be difficult to conclude (as BLS did) that, while the compensation and benefits manager occupation should be considered part of the college labor market, the bartender occupation should not. Indeed, even without these two objective sources of information it would not be hard for the informed reader to conclude from their own experiences that becoming a bartender requires no college experience (other than the usual undergraduate extracurricular experiences of interacting with bartenders) while the gateway to becoming a compensation and benefit manager is by initial completion of a college degree. And there is the rub.

The Georgetown measure of the college labor market includes all employed college graduates, irrespective of the occupation in which they are employed. So for the Georgetown analysts, all the college graduates working as bartenders are part of the college labor market. Indeed, those college grads working in cashier, retail sales, clerical, health aide, moving and transportation occupations, landscape and janitorial services and the like are all part of the college labor market.

BLS analysts disagree. They would not assign any bartender employment to the college labor market because, although one in four bartenders are college graduates, these jobs do not typically utilize the knowledge, skills and abilities acquired in college. Most of us would agree that college graduates working as bartenders are not utilizing their college education. We would regard many, though not all, of these individuals as underutilized with respect to their education or what labor economists refer to as malemployed. Amazingly, in the current labor market environment characterized by a high incidence of malemployment among young college graduates, the Georgetown analysts argue this type of skills underutilization problem simply does not exist. Essentially, the Georgetown approach assumes a world where no under-employment or malemployment of college graduates exists. Indeed, they expressly acknowledge this choice and reject the idea that college graduates could become underutilized or malemployed. The authors explain the discrepancy by arguing that overeducation or underutilization of college skills is non-existent among employed college graduates:

“… BLS’ educational and training requirements data undercount postsecondary degrees by 22 million in 2008. This implies that 22 million workers are overeducated. The overwhelming consensus in the literature contradicts this”.

While reasonable individuals can argue about the specific degree of overeducation or surplus schooling or malemployment of college graduates, it is surely the case that even in times of near full employment substantial numbers of college graduates are malemployed and are unable to effectively utilize the proficiencies associated with a college degree on their jobs. And the labor economic literature has long recognized problems of overeducation in both the U.S. and in Europe.[1]

If malemployment among college graduates simply does not exist, as the Georgetown forecasters argue, then there should be little difference in the earnings among college graduates regardless of whether they were employed in college labor market occupations or not. We examined the issue of malemployment in greater depth using data on annual earnings of employed adults during the 2006 to 2008 period to determine if earnings varied systematically by our measure of college graduate malemployment. Not surprisingly we found very large and statistically significant difference in the annual earnings of college graduates based on their malemployment status. Specifically, we found that:

  • At the associate degree level, those graduates employed in a college labor market occupation had expected annual earnings that were 60% greater than those of high school graduates, while their counterparts who earned an associate degree but were employed in a non-college labor market-related occupation had expected annual earnings that were just 10% higher than those of high school graduates.
  • Among bachelor’s degree recipients, those who worked in college labor market occupations had expected annual earnings that were 88% higher than their high school graduate counterparts, while the earnings premium for those who were not employed in a college labor market occupation was only 15% higher.

Indeed, we find that at every level of college attainment and across all age groups, large negative earnings impacts were associated with failure to find work in the college labor market. These findings clearly suggest that most of the economic gains to a college degree are strongly associated with the ability to obtain employment in the college labor market in occupations that utilize the knowledge, skills and abilities developed as part of a program of study leading to a college degree. Despite the claims of the Georgetown researchers, the available empirical evidence overwhelmingly supports the view that the personal and social payoffs to a college degree occur largely when graduates have access to jobs within the college labor market and that gains to a college diploma are quite small when graduates are relegated to jobs outside of the college labor market.

This finding implies that the Georgetown researchers dramatically overstate the size of the college labor market by including a substantial amount of employment in low-skill occupations. Paradoxically, by including this low-end employment in their “college demand forecasts,” they actually produce a slower projected growth rate in the demand for college graduates than projections based on our definition of the college labor market, although the absolute size of their projected college graduate demand remains quite exaggerated.

The demand for college graduates would grow more substantially if the U.S. economy can get back on a sustained recovery track and generate jobs in key industrial sectors that hire relatively large numbers of college graduates. Whether a generalized labor shortage of college graduates will emerge or only spot shortages in a few technical areas is simply unknown at this time. The Georgetown study fails to provide any serious evidence of college labor shortages in the future. The study is fatally flawed by its methods and even by its interpretation of fundamental labor market concepts.

As we have shown, the economic gains to earning a college degree can be quite high and, thus, college completion can be a very important determinant of lifetime success in American labor markets. However, the existence of these returns is highly dependent on one’s ability to obtain access to jobs in college labor market occupations. Rather than responding to unsubstantiated claims about future shortages, education leaders should focus on the results of serious research and evaluation and improve higher education efforts to broker new graduates into college labor market jobs, to build stronger relationships with employers to help develop more college labor market jobs for graduates, and assist alumni to get them back into the college labor market when they re-enter the workforce or lose their jobs.

Today, labor market problems especially for our younger college graduates are associated with both relatively high levels of joblessness and especially malemployment. The personal and social costs of these malemployment problems can be quite severe in forms of decreased employment, lost earnings, diminished job satisfaction as well as real output losses to society. The higher education community should seek innovative  strategies to reduce the very real problems of unemployment and malemployment that plague college graduates today instead of relying on unsupported forecasts that their problems will be solved if they simply wait long enough. As the late John Keynes once forecast with 100% certainty, “In the long-run, we are all dead.”

Paul E. Harrington is associate director of the Center for Labor Market Studies at Northeastern University. Andrew M. Sum is the center’s director.


[1]. For a review of the concepts of malemployment and overeducation and their costs to workers in the nation and the world, see: (i) Andrew Sum, Ishwar Khatiwada, Joseph McLaughlin, et. al., The Status of Teens and Young Adults (16-24 Years Old) in the Commonwealth of Massachusetts: Implications for State and Local Youth Development Systems, Center for Labor Market Studies, Northeastern University, Prepared for The Commonwealth Corporation, Boston, MA, April 2009, p. 90; Selected studies of overeducaton include: (ii) Richard R. Verdugo and Naomi Verdugo, “The Impact of Surplus Schooling on Earnings; Some Additional Findings,” Journal of Human Resources, Vol. 24, No. 4, 1989, pp. 629-673; (iii) Wim Groot, “The Incidence of and Returns to Overeducation in the U.K.,” Applied Economics, Vol. 28, pp. 1345-1350.(iv) Russell W. Rumberger, “The Impact of Surplus Schooling on Productivity and Earnings,” The Journal of Human Resources, Vol. 22, No. 1, 1981, pp. 29-50; (ii) Richard R. Verdugo and Naomi Verdugo, “The Impact of Surplus Schooling on Earnings;  Some Additional Findings”, Journal of Human Resources, Vol. 24, No. 4, 1989, pp. 629-673; (v) Nachum Sicherman, “Overeducation in the Labor Market”, The Journal of Labor Economics, Vol. 9, No. 2, 1991; (vi) David Mills, Overeducation and Earnings, Labor Economics Seminar Paper, Northeastern University, 1996; (vii) Paul Harrington and Andrew Sum, The Post College Earnings Experiences of Bachelor Degree Holders in the U.S.:  Estimated Economic Returns to Major Fields of Study, 1998 Conference on Higher Education and Workforce Development, Portland State University, March 1998; (viii) Steve Rubb, “Post-College Schooling, Overeducation, and Hourly Earnings in the U.S.”, Economics of Education, Vol. 11, 2003, Issue 1, pp. 53-70; (ix) Lisa Kahn, The Long-Term Labor Market Consequences of Graduating from College in a Bad Economy, Harvard University, Cambridge, 2006. Santiago Budria and Ana Moro-Esido, “Overeducation and Wages in Europe”, University of Madeira and Granada, January 2007


Leave a Reply

  • (will not be published)

XHTML: You can use these tags: <a href="" title=""> <abbr title=""> <acronym title=""> <b> <blockquote cite=""> <cite> <code> <del datetime=""> <em> <i> <q cite=""> <s> <strike> <strong>