Most studies of the gender pay gap use cross-section earnings functions to apply a Oaxaca decomposition into the contributions of differences in characteristics and coefficients. But the accounts that these studies provide of the gender pay gap are often hard to relate to more informal stories told about the sources of women's disadvantage in the labour market. In this paper we show how one can use a minimal amount of panel data to decompose average earnings into the contribution of the average

more »

... starting wage for workers entering paid work from non-employment, average wage growth for those in continuous employment and the fraction of workers entering employment. We use this decomposition to try to identify the source of the pay gap between men and women and the gap between full-time and part-time women using data drawn from the British Household Panel Survey. Comparing men and women we find no significant differences in wage growth whilst in continuous employment so that the source of the gender pay gap comes from the entrant pay gap and the share of entrants. Looking at longer-run changes suggests that we would expect to see a further narrowing of the gap. Comparing full-and part-time women there is no difference in entry pay shares and little difference in wage growth so that the bulk of the differential can be explained in terms of the fact that part-time women are much more likely to be entrants.