Determinants of unemployment duration

Abstract EN

This study uses a set of dates on unemployment, employment, mobility, marriage, and birth, from the 2006 and 2012 rounds of the Egyptian Labor Market Panel Survey, to construct a cross-section of first-unemployment spells, and to measure selected individual-specific attributes (age, residency, cohort, and marital status) at the time of the spell.

After conducting a 2006-2012 comparison of Kaplan-Meier estimates of the probability of staying unemployed, along several characteristics, the 2012 data is employed to estimate a discrete hazard regression model to quantify the effect of these characteristics on the duration of unemployment, after controlling for gender, educational level, as well as father’s education and occupation.

It is found that unemployment duration was longer among women and persons with secondary and higher education.

Individuals who entered the labor market as adolescents experienced longer spells, in comparison to older youth.

Father’s education and unemployment duration were negatively associated for men, but were unrelated among women.

Men whose fathers had no or unskilled jobs exited unemployment faster, with no recorded effect of father’s occupation on the likelihood of leaving unemployment among women.

Regional variations in duration were more pronounced for women than for men.

Marriage was associated with longer spells of unemployment, maybe due to stronger support from (rather than to) immediate family (spouse and children).

The duration of unemployment increased over time for entrants of the same age.

The baseline hazard was flat for women, and hump-shaped for men, suggesting a time dependent duration for the latter group.

JEL Classification : J64, C41