Econometric analyses of cross-national secondary data sources:
- European Union Statistics on Income and Living Conditions (EU-SILC) will be the primary source. It includes cross-sectional and longitudinal microdata on income, economic circumstances and socio-demographic characteristics for all EU28 members, Norway and Switzerland.
- Complemented by data from the European Labour Force Survey (EU-LFS), which collects rich microdata on demographic background, employment status and employment characteristics as well as education and training for the population in private households in all EU member states.
Differentiation in the college wage premium will be studied through the use of quantile regression (where effects are differentiated by the unobserved variable effects), by comparing across subgroups (defined by time elapsed since graduation, socio-economic background, demographic characteristics), and by examining wage penalties associated with the under-employment of graduates in non-graduate jobs.
To measure underemployment, we deploy a previously validated statistical method for classifying ‘graduate jobs’. The classifier is based on skills use from work task data drawing on information in the OECD Survey of Adult Skills. We have shown in earlier research that this classifier predicts graduate earnings, job satisfaction and opportunities for career progression better than hitherto existing methods.
Presuming we can find suitable variables that influence the decision to participate in higher education without also directly changing individuals’ earning potentials, we shall investigate conceivable instrumental variables (IV) strategies.
However, IV estimation methods estimate ‘local average treatment effects’ rather than population average treatment effects. Their value may thus be reduced when it comes to studying dispersion. In some data sets, background controls are relatively rich, and these may be able to tell us the extent of any biases in cases where these controls are not present.