The research system beyond universities: Mapping the broader labour market for PhDs

TOBIAS DALBERG (UPPSALA University)

“How many PhDs does the world need?” a recent Op-ed asked somewhat provocatively as it highlighted the fact that PhDs outnumber jobs in academia. It is an important question, because a doctoral degree is generally considered a highly specialized training. Rather than answering that question directly, this post asks a different one: what does the labour market for doctoral degree holders actually look like, and what do we mean when we talk about a “research system”? These questions matter  whenever questions arise about the need for PhDs or, as in the case of our current project, the career prospects of scholars in the formative stages of their career, whatever those careers may eventually contain.

Based on available data, one striking pattern stands out: higher education is not the primary destination for most PhDs. Across OECD countries, only around 20–30 percent of doctoral degree holders work in higher education, a share that has remained relatively stable over time. In countries like Sweden and the Netherlands, roughly 70 percent work outside universities. This immediately complicates the common assumption that a PhD is primarily a pathway into academia.

What does the labor market for PhDs look like? Put differently: in which occupations do we typically find people with a doctoral degree? Using readily available data on the working population of Sweden (SCB), we find PhDs in almost all occupations, except among ships’ deck crews and street and market vendors. That doesn’t mean that all occupations should be considered part of a labor market for doctoral degree holders. To get at that, we need to limit our scope to those occupations where most PhDs end up or those occupations where PhDs are overrepresented compared to their overall presence in the labor force. We can think of the labour market as roughly divided into three segments: occupations where PhDs are overrepresented, occupations where we find most PhDs, and the intersection of the two—where PhDs are both numerous and overrepresented.

^{Note: Calculations based on all employees aged 16–69, 2020–2024.
Source: Statistics Sweden}

For example, there are quite few PhDs who go into complementary medicine therapy and associate professionals, but these few PhDs are still overrepresented in this occupational group. At the other end, PhDs may be slightly underrepresented among insurance advisers and sales and purchasing agents, but since this is a large group, a significant share of all PhDs are still found in this group. Above all, we find PhDs among teachers in higher education. But importantly, we also find them across a wide range of public sector and knowledge-intensive roles – suggesting that the “research system” extends well beyond universities in practice, even if not always in how we talk about it.

Our Freedom to Research project focuses on how variations in funding regimes may impact early-career academics (ECAs) in the social sciences and humanities and how these conditions shape not just career prospects but (perhaps more so) their academic freedoms. This includes not only whether people stay in academia, but when and how they leave it – and what kinds of employment they move into (for example, whether positions are full-time, stable, or aligned with their training). There are many testimonies to the hardships faced by the ECAs, alongside attempts to understand why (Gorska 2026). Others frame leaving academia as a potentially fulfilling transition (Caterine 2020). There are certainly troubling factors pushing people out of higher education and research careers, such as the fixed term situation generating uncertainty. At the same time, there is also an existential dimension to this: what does it mean to spend years becoming a researcher, only to discover that stable institutional space for that identity is limited at best?

One way of framing this is through sociologist C. Wright Mills’ distinction between “personal troubles” and “public issues.” Much of the discussion on early-career academics risks remaining at the level of personal troubles, i.e., accounts of insecurity, mobility, and difficult choices. But seen in the aggregate, these experiences form patterns. What is needed is precisely the kind of sociological imagination Mills called for: an effort to situate these experiences within broader historical circumstances and labour market structures.

When studying the situation for ECAs, however, it is important to be mindful about some common biases among those of us who spent an awful long time studying at university and then try to remain there. We tend to believe that
 A) higher education is an obviously attractive thing for everyone to attain, and that
 B) remaining within higher education after a PhD is even more attractive.

To assume higher education is the most attractive career path could in fact be contributing to the problem: no need to improve anything that people yearn for, right? But the data suggest a more complex picture: many PhDs build careers elsewhere, and the attractiveness of academia cannot simply be taken for granted.

Among specialist occupations, the relatively weak correspondence between the PhD degree and its number one destination (academia) is quite a different story from more tightly regulated professions. Take air traffic controllers for example: They have specialised training programs with strict caps targeting a 1:1 correspondence between number of graduates and number of openings. They have international certification, global labor market, and labor shortages are more common than surpluses. While higher education is indeed the largest subsector hiring PhDs, it seems to take in around 20-30 percent of all PhDs, a share that seem fairly stable over time, as we shall see in just a moment. If we are striving for an air traffic control type of situation, we would need about a third of what we are producing today (except in Portugal, where the share is unusually high).

This already suggests that we may be drawing the boundaries of the “research system” too narrowly, by implicitly equating it with higher education institutions.

My main point is that — unless you work in Portugal, of course — higher education institutions aren’t as important to doctoral degree holders as one might think. On a personal level, they are of course important, but in the aggregate, the options become visible. And yet, employment practices within academia – especially the reliance on fixed-term contracts – often seem to assume an effectively endless supply of candidates willing to remain within it. This is not just about describing the labour market. When there are more PhDs than stable academic positions, bargaining power shifts to employers, making it easier to maintain short-term contracts and precarious working conditions rather than improving them..

According to data in an OECD report, postdoctoral researchers make up the majority of the research work force in many countries, and the vast majority are on fixed term contracts. This raises a basic question: is “employment” in academia equivalent to stable, full-time work, or does it often mean something more fragmented and temporary?

It is of course difficult to compare in detail what has changed, but a project on the outcomes of research education in Sweden found that most PhDs in history, economic history, human geography and political science were hired outside the universities, in schools, archives, libraries, museums, public administration (Blom & Pikwer 1976). Among graduates from 1961-1970, 29 percent were employed at universities. Half a century later, the share is… 27.8 percent! In other words, the overall structure appears remarkably stable.

In the 1960s, this broader research ecosystem was more readily acknowledged. Today, there may be a tendency to rhetorically centre universities, even as the empirical reality remains more dispersed. Looking at contemporary lists of certified administering organisations among Swedish research funders, however, we still find many of the same types of organisations—libraries, museums, and other public institutions—formally recognized as recipients and managers of research funding. This suggests not so much a transformation as a narrowing of how we talk about the system.

The implication is two-fold. First, analytically, we need to broaden our understanding of the research system beyond universities. Second, normatively and politically, this broader perspective can either be used to support diverse career paths – or to justify pushing people out of academia without improving conditions within it.

In other words, the research ecosystem may be broader than we assume – but how we define it has real consequences for how academic careers are organised and governed.

The research system has long extended beyond universities, even if academic careers are still organised as if it did not.

Plus ça change, plus c’est la même chose.

References

Blom, Conny, & Pikwer, Birgitta (1976). Vem blev forskare och vad blev forskaren? : En studie av de forskarutbildade i ämnena historia, statskunskap, kulturgeografi och ekonomisk historia i Uppsala, Lund, Stockholm och Göteborg under tidsperioden 1890-1970. Lund: Historiska institutionen, Lunds universitet.

Caterine, Christopher L. (2020). Leaving Academia: A Practical Guide. Princeton: Princeton University Press.

Gòrska, Anna M. (2026). Quitting Academia: Why Women and Men Leave and What it Says About Higher Education. Cheltenham: Edward Elgar Publishing.

Kwon, Diana (2025). How many PhDs does the world need? Doctoral graduates vastly outnumber jobs in academia, Nature 643, 16-17 (2025)

Mills, C. Wright (1959). The Sociological Imagination. Oxford: Oxford University Press.

OECD. (2021). Reducing the precarity of academic research careers. OECD Science, Technology and Industry Policy Papers. https://doi.org/10.1787/0f8bd468-en