Is the social origin pay gap bigger than we thought? Identifying and acknowledging workers with undefined social origins in survey data

This paper investigates whether recent empirical studies have underestimated the social origin pay gap by omitting respondents with undefined social origins. Specifically, individuals that were not assigned a social origin because the identity of their parental household was not clear, nobody was earning in the household, or the occupational identity of the main wage-earner could not be identified. Data from the UK Quarterly Labour Force Survey is analysed to establish the prevalence of undefined social origins and the extent to which the socioeconomic characteristics of those with undefined social origins are different from those who can be identified using the SOC Classification. We then examine how omitting these groups affects estimates of social origin pay gaps. The results show that 11% of the working age population are from undefined social origins and that the labour market outcomes of these people are on average much worse than those with defined social origins. Results show that omitting these respondents underestimates the range of the social origin pay gap and the number of people affected. This highlights that there is a further effect of parental association with the labour market or not clearly belonging to a household which profoundly affect the life outcomes of a substantial share of the working age population