Identification of the private-public wage gap

Selection issues represent the main methodological challenge to the analysis of the private-public wage gap. Recent studies apply individual panel data and account for observable and unobservable characteristics including time-varying controls and worker fixed effects. Our starting point is a difference-in-difference model assuming that shift from public to private sector is a treatment. We reject the assumption of parallel wage trends between shifters and public stayers and suggest a new identification strategy where shifters early in the period studied are compared with workers still in the public sector that shift later. The estimates are based on rich register data for high-educated male workers in Norway during 1993-2010. Comparing three years before and after the shift year, the analysis shows that the wage gain from shifting to the private sector is about 10%. The identification strategy indicates that a difference-in-difference model comparing shifters with stayers leads to an overestimation of the wage gap by 20% (corresponding to 2 percentage points in our case). The counterfactual established with a comparison of early with late shifters gives a lower estimate of the additional private sector return.