The higher educated, the lower paid: the fixed-term wage penalty within highly educated workers in Italy

ABSTRACT

This paper explores the temporary-permanent wage gap experienced by Italian graduate workers. To evaluate the wage gap along the entire wage distribution and account for heterogeneous effects of temporary employment, we apply the Recentered Influence Function (RIF) regression approach. The results suggest that the monthly net wage earned by Italian highly educated workers having a temporary contract is lower than that of their ‘permanent’ counterparts even after controlling for a plethora of individual and job characteristics. More than 50% of the wage gap is due to unobserved characteristics. The field of study, more than the location of the University, explains this discrimination effect due to contractual arrangement. Although the latter is verified along the entire distribution, major effects arise among high-paid and low-paid jobs depicting a U-shape pattern. Results are robust to self-selection, endogeneity, and gender controls.

KEYWORDS:

• Graduates
• temporary employment
• wage differentials
• RIF-regression
• Instrumental Variable Quantile Regression (IVQE)

JEL CLASSIFICATION:

• J20
• J31
• J40
• I23
• I24

Acknowledgements

We are grateful to SIE conference participants for their helpful comments. Valeria Cirillo acknowledges the support of the Italian National Research Project - PRIN 2017 ‘Regional Policies, Institutions and Cohesion in the South of Italy’ (Project code 2017-4BE543; website www.prin2017-mezzogiorno.unirc.it), financed by the Italian Ministry of Education, University and Scientific Research from 2020 to 2023.

Disclosure statement

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

Notes

1 The monthly net wage, expressed in euro at current prices (2014), is converted into full time equivalent (FTE) (see Venturini and Villosio 2008).

2 For this purpose, one should acknowledge that when performing Oaxaca and Blinder decomposition there is an identification problem in the wage structure effect as referred to by Oaxaca and Ransom (1999). Therefore, the contribution of each dummy variable to the wage structure effect is not invariant to the choice of reference group. In the discussion of results on the unexplained components of the decomposition, we always refer to the base group defined by males graduated in Medicine, in a university located in Northern Italy and working in workplace located in Northern or Central Italy.

3 M^M denotes the inverse Mill’s ratio series expansion that uses the SLS estimates of the participation probability.

4 Table 7 reports the estimates from the first-stage regression showing that the instrumental variable and temporary contracts are strongly correlated.