The university as an internal labour market

Catherine M. Haeck, Frank Verboven, 17 June 2010



During the past decade policymakers have spent considerable effort to reform European universities. Aghion et al. (2008) provide a critical review of recent higher education policies and an agenda for desirable reforms. While European countries have coordinated quite intensively to reform their teaching programs and achieve compatible degrees throughout Europe, there has been little effort to coordinate reforms of the internal organisation of universities. As a result, European universities still show a large diversity in hiring, promotion, and tenure rules. For example, Musselin (2004) compares the French and German systems. Access to a permanent position is relatively early and at a low assistant professor rank in France, whereas it is rather late but at the associate or full professor rank in Germany. In France and several other countries, academic careers tend to develop internally through promotions. In Germany on the other hand it is common to obtain associate or full professorship from another institution.

While the institutional differences across countries have been documented in several studies, little is known about how universities organise their hiring and promotion policies in practice. In recent research (Haeck and Verboven 2010), we provide such detailed evidence based on the personnel records of a large Belgian university followed for over 30 years.

As shown in Figure 1, in the early seventies the university employed almost exclusively male professors with a PhD from the own university. There has been a gradual, but slow, increase in female professors and PhDs from other universities since then. The university has four academic ranks, ranging from assistant professor (rank 1) to full professor (rank 4). In many other countries promotions or tenure decisions are awarded through “tournaments”, but in Belgium the tournaments have two interesting specific features; they are held on an annual basis and decided within the university. The tournament implies that eligible candidates compete for a limited number of slots and winners are selected based on various aspects of performance, including research and teaching.

Figure 1. Evolution of time-invariant variables

Ports of entry and exit

The university shows the existence of a strong port of entry at the lowest rank 1. Furthermore, there is a port of exit at the highest rank 4, unless the professor reaches retirement before reaching the highest rank. To see this, Table 1 shows the university’s hiring and promotion transition matrix. Looking at the entry row, one can see that 74.9% of all new hires are in the lowest rank 1, and only 2.5% of new hires are in the highest rank 4. Furthermore, among the new hires in the lowest rank 1, 83.4% were previously employed as a postdoctoral or other researcher at the university (not shown in the table). The exit column shows that most professors exit at the highest rank 4. Furthermore, up to 73.4% of the exits at ranks 2 and 3 occur when retirement age has already been reached (not shown in the table). The other cells on Table 1 show the promotion dynamics. Demotions never occur, and promotions are typically one-step promotions, as stipulated by administrative rules at the university.

Table 1. The university's hiring and promotion transition matrix

Fraction promoted to:
Current state:
Rank 1
Rank 2
Rank 3
Rank 4
Rank 1
Rank 2
Rank 3
Rank 4


In sum, the university is characterised by a strong and persistent internal labour market. As a result, careers within the university tend to be long. For example, the newly hired professors between 1991-1999 had an average prior employment history at the same university of 8 years, and over 80% were still at the university after 8 years.

Promotion dynamics

Having established the presence of an internal labour market, the next question is on what basis promotions are granted. To address this question, we estimate a hazard rate model of promotion. This model describes the probability of being promoted to a higher rank, given that this has not yet occurred, as a function of the professor’s personal characteristics, his or her research and teaching performance, and prior employment history within the university. The model enables us to assess the relevance of incentive and learning theories of promotion, and see whether administrative rigidities also play a role in promotion dynamics. This gives the following main findings.

First, both research and teaching performance significantly contribute to the promotion probabilities, but there are important differences across the three university groups (exact sciences, medical sciences, and humanities). An increased teaching load of 1 course significantly increases the promotion probability by about 1-2%. An extra publication raises the promotion probability by a comparable amount in exact and medical sciences, but by a much larger amount of 9-16% in humanities. This difference may be due to different publication practices (for example fewer co-authors in humanities), though it may be due to other forces – something that may well be the subject of future research. Note that not only the number but also the impact of publications raises the promotion probability.

Second, there is evidence of fast tracks. Professors who were quickly promoted to earlier ranks are also more likely to be promoted to the next ranks. Models of employer learning about ability would also predict a late-beginner property (Chiappori et al. 1999). Late beginners who are able to catch up have a higher promotion probability than early starters. While the data at first suggest evidence for such a late-beginner property, the evidence disappears once we control for research and teaching performance. Hence, employer learning about ability does not have a major influence on promotion dynamics; past performance is a more important determinant, consistent with the incentive role of tournaments.

Third, there is evidence of administrative rigidities constraining promotion policies. This is best illustrated by the average promotion hazard rates after controlling for personal characteristics, performance, and prior employment history. Figure 2 shows the promotion hazard rates to rank 2, 3 and 4, as a function of the number of years in the current rank. For each rank, the figure shows that promotion hazards are very low during the first two years in rank. They sharply increase in the third year, and then gradually decline in subsequent years. This pattern is consistent with administrative rules at the university, which stipulate that promotion is not allowed during the first two years after previous promotion. This creates a pool of successful researchers who are promoted in the third year, after which there is a return to normal levels in subsequent years.

Figure 2. Promotion hazard rates by rank

Note: Promotion hazard rates by rank estimated using the complete model for a representative individual with average characteristics.


European universities show a diversity of personnel policies, ranging from internal to external labour market organisations. Our case study provides evidence on the workings of an internal labour market, showing how a university rewards performance achievements but at the same time is constrained by administrative rigidities. Nevertheless, policies to remove administrative rigidities cannot be considered in isolation, since they may exist because of other constraints, such as financial constraints that limit the budget of universities in their personnel policies. We hope that the results of our case study will stimulate detailed studies in other countries, and ultimately contribute to improving coordination in personnel policies across Europe as academic labour markets are becoming increasingly integrated.

Authors' note: We are grateful to Georg Kirchsteiger for earlier comments on this column.


Aghion, Philippe, Mathias Dewatripont, Caroline Hoxby, Andreu Mas-Colell and André Sapir (2008), “Higher aspirations: an agenda for reforming European universities”, Bruegel Blueprint 5:59.
Chiappori, Pierre-André, Bernard Salanié and Julie Valentin (1999), “Early starters versus late beginners”, Journal of Political Economy, 107:731-760.
Haeck, Catherine and Frank Verboven (2010), “The internal economics of a university: evidence from personnel data”, CEPR Discussion Paper 7843.
Musselin, Christine (2004), “Towards a European academic labour market? Some lessons drawn from empirical studies on academic mobility”, Higher Education 48: 55-78.


Topics: Education, Labour markets
Tags: education, higher education, research, teaching

Doctoral researcher at the Center for Economic Studies at the University of Leuven
Professor of Economics at the Catholic University of Leuven and CEPR Research Fellow