Is Expansion of Overeducation Cohort-Driven? Evidence from Poland

ABSTRACT

The paper examines factors behind the expansion of overeducation incidence among tertiary educated workers in Poland, which grew by up to 8 percentage points from 2006–2020. It is one of the first attempts in the literature to explain changes in overeducation incidence using age, cohort, and period effects simultaneously. The study finds a strong upward cohort shift in the overeducation risk for workers born after 1970. It suggests that the overeducation expansion in Poland is a cohort-driven phenomenon and it affects more profoundly individuals who received their tertiary education after the collapse of communism.

KEYWORDS:

• Age–period–cohort
• education mismatch
• overeducation
• tertiary education

JEL Classification:

• I21
• J21
• J24

Acknowledgments

I would like to thank Gabriela Grotkowska and Maja Rynko for their valuable comments on this paper.

Disclosure Statement

No potential conflict of interest was reported by the author(s).

Notes

1. The problem is associated with the realized matches approach which identifies the required education based on the actual distribution of educational attainment among workers in a given occupation. An influx of graduates into an occupation which has not required higher education can make the occupation switch its education requirement to higher education purely due to statistical reasons. Hence, overeducation would decrease rather than increase.

2. Battu, Belfield and Sloane (1999) is one of the few exceptions.

3. Vera-Toscano and Meroni (2021) find that Norwegians born between 1955 and 1965 are more likely to be overeducated, while those born in 1975 experience lower overeducation risk. In the Netherlands, individuals born in 1975 are less likely to be overeducated. Whilst in Italy, lower overeducation is found among those born in 1965 and higher among those born in 1950.

4. Vera-Toscano and Meroni report graphs, similar to those in part 5 of this paper, which suggest the existence of positive cohort trends in Italy and the Netherlands, but they do not verify them econometrically.

5. The rationale behind the latter one is that graduating during a recession has a negative long-lasting impact on individuals’ labor market outcomes (Oreopoulos, von Wachter, and Heisz 2012).

6. Besides the realized matches and the indirect subjective approach, there are two other approaches to identify overeducation: the direct subjective approach, in which workers are asked directly whether they feel overeducated in their current job, and the job analysis approach, in which labor market experts assess what is the required level of schooling for each occupation. These two approaches are perceived superior to the realized matches one, but still they are rarely used compared to the realized matches due to data source scarcity.

7. Even when the occupations are aggregated to 2-digit ISCO codes, the number of observations per one occupation in the BKL sample varies from less than ten to over one thousand. Ideally, a more detailed representation of occupations would be preferable to achieve better precision of the results. However, the potential bias of an overeducation rate arising from the aggregation of occupations is probably limited because occupations within the same 2-digit code are usually similar to each other, also in terms of education requirements.

8. Occupations with the number of observations being less than 20 were arbitrary assigned to one of the two categories. There were 8 such occupations. According to the LFS data, these occupations correspond to about 3% of workers with tertiary education in 2020. 7 of them were assigned to the non-university occupations, since they can be characterized as blue-collar occupations or lower rank soldiers. One, high rank soldiers, was assigned to the university occupations.

9. The indirect subjective approach using 50% threshold gives similar results to the realized matches approach using modal schooling. Both methods classify exactly 14 occupations as university occupations. The groupings diverge in case of four occupations. Under the realized matches approach, health associate professionals and information and communications technicians are classified as non-university occupations, whilst business and administration associate professionals and legal, social, cultural and related associate professionals are classified as university occupations. The opposite applies under the indirect subjective approach. The divergent occupations cover 13% of the tertiary educated workers in the sample.

10. The differences might arise due to different approaches (e.g., realized matches, subjective assessment, or job analysis), variations in the level of aggregation of occupations and education levels, or a combination of both. Hence, there is no single true rate of overeducation.

11. At the first glance, this finding is counterintuitive. When the labor market situation is good, one could possibly expect lower overeducation risks as there are many open job opportunities. However, labor demand in low-skill sectors such as construction rises relatively strongly in periods of economic expansion and decreases in recessions. Hence, low-skill labor demand is more procyclical than high-skill labor demand, which might possibly explain the finding.

12. Private universities get on average lower scores from the Polish Accreditation Committee, an independent public institution which evaluates the quality of study programs in Poland, than public universities (Najwyzsza Izba Kontroli [Supreme Audit Office] 2018).

13. Although in logistic regressions I control for fields of studies, still some variation in fields of studies between cohort might be unaddressed due to grouping of study programs into relatively broad categories.

14. This is not in line with findings by Kampelmann and Rycx (2012) or Mahy, Rycx and Vermeylen (2015) according to which overeducation rather improves productivity.

Additional information

Funding

This work was supported by the Polish National Science Centre [grant number DEC-2016/23/N/HS4/03626].

Notes on contributors

Jan Aleksander Baran

Jan Aleksander Baran is PhD candidate at the Faculty of Economic Sciences, University of Warsaw, and economist in Narodowy Bank Polski.