Abstract Previous research on over-education primarily focuses on new graduates, and documents age as a linear corrector of over-education. However, labor markets are aging in developing and developed countries. This research revisits the relationship between age and over-education across the total span of a career in 33 country-years and for 31,608 individuals with tertiary education, using two cycles of the PIAAC dataset. The findings show that, (1) Although over-education decreases in the first half of a career parallel to the existing literature, from the age of 45 onwards it starts increasing, contrary to what earlier research would suggest. (2) This effect is documented only for senior workers with tenures shorter than 5 years, suggesting that over-education is a potential scarring effect of job change or job loss in the tail-end of a career. (3) The effect of the skill level of a senior worker is negligible on their risk of over-education. Statistical tests show that the quadratic specification of age is a better fit than linear and cubic specifications. The results are robust across three different measures of over-education, across country-years, and alternative model specifications.
Ömer Şahin (Mon,) studied this question.