Universities are the main suppliers of higher order skills, where many young (and increasingly not so young) people go to expand their opportunities in the labor market. Historically, universities have been remarkably good at this; they have steadily guaranteed more and better paid employment. However, more recently the need for skills has changed dramatically. Digitization, automation and the rise of new forms of employment have meant that many new jobs have emerged and old jobs have disappeared or have now changed and demand different skills.
In this project, we have used Natural Language Processing techniques to assess the level of alignment between the skills supplied by institutions and their labour markets. As universities act as principal “suppliers” of skills, they are a natural focal point for an analysis on the alignment between the supply and the demand of skills. Our research can cast a light on why more people are unable to successfully access highly paid jobs that are in high demand but go unfilled.
Collaborating closely with data scientists at IBM Research, we have generated the highest resolution dataset to date on the curriculum and skills students are receiving in their tertiary education programs and how these compare with the demands of the labor market. For this report, we mobilized systematic data on 13 million job postings and over 500,000 syllabi from undergraduate degrees in three European countries.
Our analysis sheds light on one aspect of the challenge of providing higher education fit for the 21st Century’s labor market. The skills that universities teach at the undergraduate level are but one aspect of the training universities provide, but they have proven a valuable source of information. Our analysis of unique datasets reveals a number of important facts about what universities train students in and how it compares to what labor markets demand.
We studied the “skill intensity” of the universities in our sample, in other words, the average number of skills contained in a course offering, and adjusted the raw measure of skill intensity by the level of demand for each skill to develop a weighted skill intensity.
After looking at the variation in skill intensity (weighted and unweighted) between institutions and found that institutions vary greatly in their skills intensity.
Some universities are up to eight times better aligned with labor markets than others, placing some students at a significant disadvantage, which they probably did not realize when they began their studies.
The finding of such diversity across institutions suggests there are two challenges for institutions. One is including more training in skills as opposed to focusing on other nonskill specific content in courses, such as disciplinary knowledge, which is less directly valued by employers. The second one is that universities also differ on how well they select the skills that they do train for. That is, beyond the number of skills contained in their courses, how aligned those skills are to the national job market.
We also found that private and newer universities tend to have greater skill intensity, on average in our sample. For each of these institutional characteristics, there are good reasons to expect substantial differences. Private universities rely directly on student tuition. Different countries may have different expectations of the role of universities, universities in a given country may have converged over time given that they are the most natural competitors, and they may be under different regulatory regimes. Older institutions may rely on their reputation or may have more inertia in adapting the content of their courses than newer institutions.
Private institutions train on average in about twice as many skills per course than public institutions.
That newer institutions are more aligned with labor market demands is fairly unsurprising given the pace of transformation in older universities. Institutional inertia and path dependence are germane to any organization. Universities and other educational institutions are perhaps particularly subject to such challenges. They are often more consensus-based than private firms. They also make decisions, such as on staffing with very long-term views, often making “tenured” or life appointments. Similarly they are major builders of new permanent campus infrastructure, often difficult to repurpose. Finally, the content of their courses is decided many years before the graduates of those courses have contact with labor markets. A typical setting is that any curricular reform takes two or three years to be implemented, and then students take four or more years to graduate from a program with the new curriculum. Newer institutions have therefore less burden of such institutional inertia.
Institutions created after the 1970s have more than three times as many skills per course as those created at any point before.
The report also reveals that the type of skills that drive differences in skill intensity across universities are general skills, and not skills specific to any field (such as STEM skills or business-specific kills). These general skills include communication, teamwork, and problem-solving. Of the different degree programs, courses related to business contain more skills.
Looking at the implications of this finding, for institutions that want to increase their degree of alignment with the labor market, this would appear to be good news: differences in skill content can largely be attributed to differences in general skills, which any university could boost. Training in communication skills, teamwork, text analysis, project management, can perhaps be more readily embedded in curricula, than skills that are specific to a field.
The use of large datasets on job markets as well as the content of university programs enables many possibilities that we are only beginning to explore. Through this project, we hope to contribute to increasing our collective understanding of the skills gap between university training and the labor market.
To read the full study, please download the report above.
Research Team
Research Team
In partnership with
In partnership with
