Previous American Growth Project reports have focused on metrics that are relatively easy to calculate. Productivity, for example, is generally defined as the amount of economic output (GDP) per worker – a relatively straightforward measurement once one has the necessary data. In this analysis, however, we turn our attention to an equally essential but more nebulous driver of economic productivity by asking this: What is the skill level of the U.S. economy?
It should come as no surprise that a workforce possessing a large quantity and wide range of skills is preferable to one that does not; the former is able to accomplish more kinds of tasks than the latter with greater competency. However, answering the above question with any degree of accuracy poses numerous challenges. For one thing, the question of defining what constitutes “skilled” labor can be answered in a variety of ways, depending on the context. For another, the act of measuring the level of skills in the workforce is no easy task – due, in part, to these definitional difficulties.
Moreover, the types of skills that are most highly rewarded in the labor market – as well as the types that are most conducive to generating additional economic growth and productivity – are in constant flux. Skill in the use of generative artificial intelligence prompts, for example, is quickly becoming a hot commodity in the current labor market – even with the recent woes in the tech sector – but was unheard of just a few short years ago. Conversely, the skills necessary to perform the work of a telephone switchboard operator or typesetter were in relatively high demand until these systems were automated, thus making these skills impractical for newer generations of job seekers.
In the context of the American Growth Project, we’re also interested in how skills factor into economic development, as well as how different skills are distributed across our microeconomies. We’ve explored how regions and cities across the U.S. specialize in different sectors, creating geographic differences in the demand for certain skills. For example, a workforce with strong competencies in energy and mining would be a boon to the incredibly strong (and high-productivity) oil and gas sector in Midland, Texas, but potentially less so in a finance-heavy area such as New York City.
In an economic context, the term “skills” is a broad term that’s used to refer to a variety of attributes – which creates ambiguity around its definition. “Skilled labor,” however, generally refers to jobs that require a high amount of training, education or experience to perform. This training can come in the form of specific educational credentials (e.g., doctors who require a medical degree) or less formal technical or vocational training (as in the case of coding boot camps or training and apprenticeship programs for computer numerical control machinists). Most importantly for our purposes, the overall skill level in a workforce determines the tasks that workforce can competently perform. As such, it’s directly linked to economic productivity; not only do you need these skills to foster innovation, but the impact of any innovation or technological progress will be severely limited if the workforce lacks the skills to properly use them.
While skills have been studied extensively, there have been few, if any, established methods for measurement (due in part to the definitional ambiguity mentioned above).
Academics have approached measuring or estimating the level of skills in the workforce primarily in two ways. The first uses employer-level survey data to assess whether employers can hire workers with the skills they need to perform a given job. The other is more strongly based in human capital theory and estimates a skills gap as the difference between skills supplied and skills in demand. Skills in demand can be accessed a little more readily (through the Job Openings and Layover Turnover Survey1 or LinkedIn data) than skills in supply. Researchers trying to estimate the amount of skills supplied in the labor market generally use qualifications (e.g., education levels, licensure) as proxies.
Each approach presents a specific set of issues. Issues with the survey approach include employer-level biases that can offer distorted views of the skills needed, as well as the skills present in the labor market. In addition, such surveys can lack key details or variables, and generally hinge on the perceptions of particular employers. Issues with the approach of using a proxy to estimate the level of skills supplied generally entail concerns around the imperfections of these proxies. In particular, education can fail to capture technical skills or the competencies that may actually be employed at a job.
As labor markets continue to redefine themselves in the wake of the COVID-19 pandemic, many employers and commentators are discussing the idea of a “skills gap.” While the term can take on numerous meanings, it’s generally used to refer to the idea that specific talents and competencies needed by employers are not present in the hiring pool of job candidates. In essence, the skills gap can be thought of as a mismatch between the supply and demand sides of the labor market. Subsequently, such gaps can cause economic friction and hamper productivity; employers can’t find the workers they need, and job seekers find themselves ill-suited for the job postings available to them.
The question of whether such a gap exists has been a hotly debated topic for several years. For those arguing in favor of its existence, the high number of job openings, especially in certain sectors such as skilled trades2, is evidence that the U.S. labor pool is insufficiently skilled to meet the needs of employers. Multiple employer surveys also point toward these gaps’ existence; a Wiley survey from January 2023 finds a sharp increase in human resources managers who say they can’t find workers with the requisite skills for their positions.3 By contrast, a 2014 paper from Peter Cappelli finds little evidence in favor of a skills gap and instead attributes skills shortcomings to overeducation (e.g., too many bachelor’s and not enough technical degrees).4 Meanwhile, an in-depth review from the Congressional Research Service finds, after analyzing broad indicators, that a skills mismatch can neither be confirmed nor ruled out.5
This debate demonstrates the difficulty in identifying and measuring the skills gap, since the creation and development of skills involves so many varied and interlocking aspects of the economy – particularly the educational system and the level of technological innovation. Other systemic factors can also cause disconnection between the demand and supply sides of the labor market. For example, the high retirement rates of baby boomers have caused a large number of highly trained and experienced workers to leave the labor pool, creating a relative dearth of skilled labor.6 At the same time, COVID-19 has caused some workers to rethink their job choices (a primary inquiry of our 2023 grand challenge: Workforce Disrupted: Seeking the Labor Market’s Next Equilibrium).7
To summarize, these skills mismatches are a likely combination of the downstream effects of an educational system that’s mismatched to the set of jobs available to workers, rapid technological advancements that have not yet fully developed training systems to match, and demographic and social factors that are decreasing labor supply. As technological change progresses, new inventions will continue to affect and alter the sets of skills most needed by employers.
Our preliminary look at skills levels across our Extended Metropolitan Areas (EMAs) largely relies on educational attainment as a proxy – specifically, the percentage of individuals with a bachelor’s degree or higher. While the research discussed above indicates that this is an imperfect measure, we feel it is a good starting point because – as illustrated below – it is informative of both economic growth and productivity. The map below in Figure 1 gives a broad overview of the geographic distribution of skills via a choropleth map depicting the share of the population ages 25-64 with this level of educational attainment.
The share of the working-age population with bachelor’s or advanced degrees tends to be clustered in large coastal cities, particularly in the northeastern and western United States, and this map in general looks quite similar to our American Growth Projects maps of fastest-growing U.S. cities. To illustrate this point, Figure 2 plots the relationship between the share of the population ages 25-64 within each EMA with a bachelor’s degree and the corresponding EMA GDP growth over the last 10 years.
The scatterplot in Figure 2 above more clearly demonstrates the impact of a skilled workforce on economic growth. While our EMAs exhibit a fair bit of variance, indicating that a bachelor’s degree is not the only factor driving growth, the graph’s relationship is evident – greater education levels are highly correlated with sustained growth. The results from our data reveal that an increase of 10 percentage points in the share of working-age population with at least a bachelor’s degree would increase GDP growth by nearly a full percentage point per year (more specifically, 0.83 percentage points). Putting this into perspective, during this period the U.S. GDP grew an average of 2.1% a year.
While these results are quite powerful, it is important to note the challenges of doing these types of analyses. The tight relationship between education and growth may reflect the propensity of educated people to move to these fast-growing areas, as well as the high cost of living required to live in expensive cities, which educated people can more likely afford. To resolve part of this challenge – growth resulting from more employment – we focus on productivity and one particular driver of economic activity: a given EMA’s share of skilled workers in tech sectors.
Figure 3 plots an EMA’s percentage of its workforce that has a bachelor’s degree or higher and works in the tech industry and the corresponding EMA’s productivity (measured as inflation-adjusted GDP per worker). There is a strong and significant relationship between the two variables; as an EMA increases its share of skilled workers in tech, so too does that EMA’s workforce become more productive. Midland, Texas, in the upper left-hand corner, represents a notable outlier, given its specialization in oil and mining – a rare case of a productive industry that requires relatively less formal education than similarly productive areas – but also an important reminder that our measure of skills needs to be further refined.
In particular, this final graph is essential for understanding how skills factor into economic health. As a highly productive industry, the tech sector is an excellent example of how specific skillsets can bolster economic development. Moreover, educational attainment in tech is an explicit signal of skills acquisition; while less technical degrees are of no less importance, it can be difficult to discern the specific skills that individuals accrue from degrees in these areas.
It also offers a salient point for us to conclude this initial examination of skills and why they matter for economies: Technological progress is often touted as a driver of economic development, but while the role of innovation is certainly important, technology is rendered far less useful if the skills to effectively employ it are not present in the working population. And, more broadly, it matters little how many positive economic characteristics are present in an area or EMA if the population lacks the skills or education necessary to harness them. Skills are thus the crucial link that allows for the realization of economic potential, creating a workforce that is ready and able to power economic growth.
1 Bureau of Labor Statistics. Job Openings and Labor Turnover Survey. BLS. https://www.bls.gov/jlt/.
2 Industrial Skilled Trades. (2022, February 23.) The Skilled Trades Labor Shortage of 2022 (and How to Overcome). https://www.industrialskilledtrades.com/skilled-labor-shortage-2022-1
3 Capranos, D., & Magda, A.J. (2023, January.) Closing the Skills Gap. Wiley. https://universityservices.wiley.com/wp-content/uploads/2023/01/Closing-the-Skills-Gap-2023-Digital-January-2023.pdf
4 Cappelli, P. (2014). Skill Gaps, Skill Shortages and Skill Mismatches: Evidence for the US. NBER Working Paper No. w20382, Available at SSRN: https://ssrn.com/abstract=2482145
5 Donovan, S.A., Stoll, A., Bradley, D.H., & Collins, B. (2022, March 31.) Skills Gaps: A Review of Underlying Concepts and Evidence. Congressional Research Service. https://sgp.fas.org/crs/misc/R47059.pdf
6 Fry, R. (2020, November 9.) More Baby Boomers have retired since COVID-19 began than before. Pew Research Center. https://www.pewresearch.org/short-reads/2020/11/09/the-pace-of-boomer-retirements-has-accelerated-in-the-past-year/
7 Kenan Institute. 2023 Grand Challenge: Workforce Disrupted. https://kenaninstitute.unc.edu/tag/workforce-disrupted/