For colleges, employers, and policymakers concerned with sustainable work force development, particularly in regard to career-oriented degree programs that prepare students for fairly specific occupations, ensuring that the ongoing supply of workers matches up with long-term demand projections is vital to preventing labor shortages and skill gaps.
Here’s an example: The U.S. Bureau of Labor Statistics (BLS) projects that, from 2014 to 2024, the number of accounting jobs in the United States will rise from about 1.33 million to 1.42 million—an increase of about 142,000. At the same time, about 356,000 accounting jobs are projected to require a replacement due to retirements or workers otherwise permanently exiting the profession. Altogether, between job growth and replacement needs, the BLS projects that about a half million accounting jobs—likely including a breadth of entry-, mid-, and senior-level positions—will need to be filled from 2014 to 2024.1 Thus, the question is: Will colleges produce enough graduates and will employers internally prepare and externally recruit enough accountants to fill these positions at various rungs of the career ladder?
The way that projections like this define “replacement needs” begs the question: To what extent do graduates of various career-oriented degree programs leave or even outright evade the professions for which they were prepared? As technical occupations and their education requirements becoming increasingly specialized, waves of workers permanently leaving these occupations could create vacuums in the work force. Thus far, the literature has been mostly silent on this issue and, given the limitations of the available data, this is unsurprising. Despite this, it would be valuable to explore what data do exist to glean even some rudimentary insight. Therefore, this article will explore potential patterns of work force egress for graduates of four-year career-oriented degree programs: accounting, computer science, education, and engineering.
Data, Methodology, and Limitations
The “gold standard” for such a research premise would be a series of longitudinal surveys that capture the educational and employment experiences of respondents over the course of their lifetimes and include multiple generational cohorts, because cultural, organizational, and economic shifts could have generationally specific effects on respondents’ experiences. However, as no such data exist that meet these criteria, we will use the available historical cross-sectional data—the version of the American Community Survey (ACS) data published by the University of Minnesota – Minnesota Population Center.2 (The original version of the ACS is published by the U.S. Census Bureau.)
In an effort to mitigate idiosyncrasy, multicollinearity, and spuriousness, the sample that will be examined in this article is quite specific and includes college graduates who were identified as having met all of the following five criteria: 1) their highest completed level of education was a bachelor’s or master’s degree; 2) they were employed in what the Federal Reserve Bank of New York designates as an occupation requiring at least a bachelor’s degree to enter,3 3) they “usually” worked in that occupation on a full-time basis (i.e., at least 35 hours per week); 4) they worked for wages or, in other words, were not self-employed; and 5) they were not concurrently enrolled in school. In addition, because of the highly specific (and specialized) careers paths for graduates of doctoral and professional degree programs, any respondents who held such a degree were excluded from these analyses. Other exclusions include those who held a bachelor’s degree (or two separate bachelor’s degrees) in two distinct fields of study and those employed in nominally managerial positions because it is unclear to what extent these manages’ work duties were actually distinct from “the trade.”
In addition, ACS data for the years 2009 through 2015 are used here as 2009 was the first year that this survey collected data on respondents’ undergraduate field of study and 2015 was the most recent year for which a full 12 months’ worth of data were available. Therefore, the percentages included hereafter represent seven years’ worth of data in a single instant. Keep in mind that the ACS does not (yet) include information on respondents’ graduate field of study and, therefore, any cause-and-effect inferences related to master’s degrees are fairly circumstantial. Altogether, given these highly specific parameters, the statistics in this article should not be interpreted as definitive benchmarks, but rather as tools to indicate how these populations have behaved and may behave in the future.
Employment for accounting graduates is remarkably monolithic. Among 21- to 60-year-old accounting graduates, 75.9 percent were employed as an accountant (or auditor) and 11 percent were employed in other non-administrative business management occupation. Altogether, 86.9 percent were employed in a business management occupation,4 of whom nine out of 10 were accountants. Among the minority of accounting graduates notemployed in business management occupations (13.1 percent), most were employed as either computer specialists5 or teachers (4.2 percent and 3.9 percent, respectively).
Figure 1 displays, among accounting graduates broken out by age and level of education, the percentage employed as an accountant. Figure 2 displays, for the same cross-sections, the percentage employed in any business management occupation (including accountants).
Figure 1 illustrates that, for bachelor’s degree holders, the very high likelihood of being employed as an accountant differed very little by age. If anything, older bachelor’s degree holders were less likely (although still very likely) to be employed as an accountant in slight increments. For master’s degree holders, however, that likelihood did differ considerably by age. In markedly larger increments, older master’s degree holders were less likely (but, again, still very likely) to be employed as an accountant. In addition, Figure 2 illustrates a similar pattern in regard to business management occupations in general although, for graduates at either level of education, the vast majority of those employed in business management occupations, at any age, were still represented by accountants.
Altogether, these observations suggest that, given their highly compartmentalized skillset, accounting graduates without a master’s degree have extremely narrow employment options both at the outset of their professional career and further down the road. On the other hand, options for those with a master’s degree are somewhat more diverse; it stands to reason that accountants (especially older accountants) who feel pigeon-holed in their careers may pursue a master’s degree—most likely an M.B.A.—to transition out of accountancy and into other business management roles, or even out of the business management world altogether. In the end, though, employment outcomes for accounting graduates at both levels of education are highly monolithic and, therefore, even gradual patterns of egress should not be a cause for concern.
The employment of computer science graduates was very similar to that of accounting graduates. Among 21- to 60-year-old computer science graduates, 78.4 percent were employed as computer specialists and, among the remainder (21.6 percent), most were employed as mathematical specialists,6 engineers, or teachers (7.2 percent, 4.6 percent, and 3.6 percent, respectively).
Figure 3 displays, among computer science graduates broken out by age and level of education, the percentage employed as a computer specialist. The figure illustrates a pattern very similar to that observed for accounting graduates. Across age, bachelor’s degree holders remained monolithically employed, while master’s degree holders were somewhat more diversely employed. Altogether, these observations suggest that, as was the case with accounting graduates, employment options for computer science graduates withouta master’s degree are fairly limited, both upon graduation and throughout their lives. In addition, a master’s degree is a means by which computer science graduates can exit the computer/IT work force and enter other lines of work, particularly those in which they can still apply their advanced technical and mathematical “know-how.”
Compared to their accounting and computer science peers, education graduates were even more monolithically employed. Among 21- to 60-year-old education graduates, nine out of 10 were employed as teachers (89.4 percent); the minority notemployed as teachers (10.6 percent) were not concentrated in any other occupation in particular.
Figure 4 displays, among education graduates broken out by age and level of education, the percentage employed as a teacher. Interestingly, this figure illustrates a pattern opposite to that of the other three fields analyzed here. Among education graduates, it was master’s degree holders who were monolithically employed across age, while older bachelor’s degree holders, in slight increments, were less likely (although still very likely) to be employed as a teacher. Altogether, these observations suggest that employment options for education graduates are extremely narrow and that those who leave the teaching profession are notdoing so because there is any particular demand for their skillsets in other industries. In addition, since a master’s degree is very often a requirement of teachers’ ongoing professional development, earning one is a sign that teachers are fully committed to the profession.
The engineering sector is very much concerned with the potential for labor shortages and skill gaps. These worries arise from comparisons of projections for degree completions and job openings7 (although there some debate over whether such fears are warranted8).
Often, the possibility that engineering graduates might be leaving the engineering work force isn’t considered. In fact, among 21- to 60-year-old engineering graduates (excluding computer engineering graduates9), only 61.5 percent were employed as engineers—a majority, yes, but a far cry from the super majorities seen for accounting, computer science, and education graduates and their respective target occupations. Among the more than one-third of engineering graduates (38.5 percent) not employed as engineers, most were employed as computer specialists and a small, but notable, minority were employed as mathematical specialists (19.6 percent and 6.5 percent, respectively).
Figures 5 and 6 display, among engineering graduates broken out by age and level of education, the percentage employed as an engineer, as a computer specialist, as a mathematical specialist, or in any other occupation; Figure 5 displays these data for bachelor’s degree holders, while Figure 6 displays these data for master’s degree holders.
These two figures illustrate a curious pattern across age that was not seen among accounting, computer science, or education graduates. For those three groups, it was the youngest among them who were the most likely to be employed in their respective target occupations, while older graduates were (incrementally) less likely. But, for engineering graduates at both levels of education, the least likely to be employed as engineers were not the oldest, but rather those in the middle of the age range (30- to 44-year-olds). These graduates were, in turn, somewhat more likely to be employed as computer specialists. What might be causing this unique non-linear dip? That answer may lie in one advantage that cross-sectional data have over panel data.
In these data, this group likely completed their bachelor’s degrees from (roughly) the mid-1980s to the mid-2000s, a period of major expansion for the computer/IT market that is “bookended” by the dawn of the microcomputer revolution and the Great Recession. This suggests that engineering graduates whose (presumed) entry into the work force coincided with the proliferation of the computer/IT sector were more drawn into computer-specialty occupations and, consequently, away from traditional engineering occupations. At the same time, older engineering graduates, whose (again, presumed) entry into the work force was further removed from the computer revolution, were less likely to work as computer specialists and more likely to work as engineers. In addition, younger engineering graduates, who in these data (presumably) entered the work force during more economically tumultuous times, may have opted to stick in the more stable engineering sector.
A Shot in the Dark
Despite the limitations of the available data, the preceding analyses do offer some clues to how longitudinal data might behave and, in turn, allows us to ask better questions and make more targeted recommendations for future research and data collection. One key factor that remains unclear are the more precise mechanisms of choice and timing. For example, for what reasons do accountants choose to leave the accounting profession and at what point in their careers are they most likely to make that decision? Another important factor that likely affects even minor occupational egress is stagnation in wage growth over time, particularly for—although not limited to—STEM occupations.10 Unfortunately, while the ACS and similar surveys collect data on respondents’ income and wages, we do not find those data to be reliable enough to apply in such analyses.
Given the limitations of the available data, we recommend that those interested in researching this topic use a mixed-methods approach that makes use of both cross-sectional and longitudinal data whose findings can complement each other. For bodies collecting data (BLS, Census Bureau, and so forth), we recommend that collecting data on field of study for advanced degrees, employment history by occupation, and some elements that capture choice and intent as they relate to educational and employment history.
1 “Employment Projections.” U.S. Bureau of Labor Statistics.
2 Genadek, Katie; Goeken, Ronald; Grover; Josiah; Ruggles, Steven; and Sobek, Matthew. “Integrated Public Use Microdata Series: Version 6.0.” University of Minnesota: 2015.
3 Abel, Jaison R., Deitz, Richard and Su, Yaqin. “Are Recent College Graduates Finding Good Jobs.” Current Issues. Federal Reserve Bank of New York: January 2014.
4 In this analysis, “non-administrative business management occupations” included, as delineated in the original data, accountants and auditors; actuaries; business and promotion agents; inspectors and compliance officers, outside construction; insurance underwriters; management analysts; operations and systems researchers and analysts; other financial specialists; personnel, HR, training, and labor relations specialists; and purchasing managers, agents and buyers, n.e.c.
5 In this analysis, “computer specialists” included, as delineated in the original data, computer software developers, computer systems analysts, and computer scientists.
6 In this analysis, “mathematical specialists” included, as delineated in the original data, accountants and auditors; actuaries; economists, market researchers, and survey researchers; management analysts; mathematicians and mathematical scientists; operations and systems researchers and analysts; and other financial specialists.
7 Wright Joshua. “The Most In-Demand (And Aging) Engineer Jobs.” Forbes. September 12, 2014.
8 Teitelbaum, Michael S. “The Myth of the Science and Engineering Shortage.” The Atlantic. March 19, 2014.
9 Computer engineering graduates were excluded from this analysis because they are, more so than graduates in other engineering specializations, uniquely positioned to be employed as computer specialists and are therefore likely to skew the results for engineering graduates at-large. This exclusion applies to all subsequent analyses.
10 Carnevale, Anthony P.; Melton, Michelle; and Smith, Nicole. “STEM.” Georgetown University Center on Education and the Workforce:October 2011.