Abstract
This paper uses data from the 2007 National Employment Survey to analyse the public-private sector wage gap in Ireland. The purpose of the paper is to highlight the issues around attempting to establish a definitive measurement of the public-private sector wage differential. A number of different approaches are used to estimate the public-private wage gap, and the limitations of these various methods are highlighted. A range of plausible estimates of this wage gap are presented for 2007. In particular, this report investigates the effect of weighting on the estimates as well as exploring the impact that the inclusion of size of enterprise has on the results. These results show that, in general, there is a premium associated with working in the public sector when average estimates are calculated. When quantile regression is used to consider how the premium varies across the distribution of earnings, it was found that the premium was highest at the lower end of the earnings distribution and lowest at the top end. However, in most instances modelled, the public-private wage gap became a discount at the higher end of the earnings distribution. Generally, the exclusion of size of enterprise from the model increased the public-private wage gap by approximately 5 percentage points. Conducting the analysis on weighted data rather than unweighted data tended to increase the size of the wage gap, and this was particularly true for male employees. In this paper, we also analyse a sample that excluded personal and protective services employees. The exclusion of this particular sub-sample had the effect of considerably reducing the public-private wage gap, most noticeably for males. This impact was evident in both the weighted and unweighted analyses. In light of the difficulties highlighted in this report in comparing public and private sector wages, it should be noted that any type of regression or related analysis that attempts to directly compare earnings across the public and private sector is prone to oversimplification.