Search results

The wage effect of job–education vertical mismatch (i.e. overeducation) has only recently been investigated in the case of Ph.D. holders. The existing contributions rely on ordinary least squares (OLS) estimates that allow measuring the average effect of being mismatched at the mean of the conditional wage distribution.

The authors implement a recentered influence function (RIF) to estimate the overeducation gap along the entire hourly wage distribution and compare Ph.D. holders who are overeducated with those who are not on a specific sample of Ph.D. holders in different fields of study and European Research Council (ERC) categories. Moreover, the authors compare the overeducation gap between graduates working in the academic and non-academic sector.

The results reveal that overeducation hits the wages of those Ph.D. holders who are employed in the academic sector and in non-research and development (R&D) jobs outside of the academic sector, while no penalty exists among those who carry out R&D activities outside the academia. The size of the penalty is higher among those who are in the mid-top of the wage distribution and hold a Social Science and Humanities specialization.

Two policies could reduce the probability of overeducation: (a) a reallocation of Ph.D. grants from low to high demand fields of study and (b) the diffusion of industrial over academic Ph.Ds.

This paper observes the heterogeneity of the overeducation penalty along the wage distribution and according to Ph.D. holders' study field and sector of employment (academic/non-academic).

Using the two waves of the India Time Use Survey, 1998–1999 and 2019, we document a 110-minute (30%) increase in average daily learning time. The largest offsetting decrease was in work time: 61 minutes. The composition of leisure changed, with television rising by 19 minutes, while talking fell by 10 minutes and games by 17 minutes. We then implement a Gelbach decomposition, showing that 68 minutes of the unconditional learning increase are predicted by demographic covariates. Of these predictors the most important are a child's state of residence and usual principal activity, which captures extensive-margin transitions into schooling.