Omitted variables in the measure of a labour quality index : the case of Spain

Abstract

Traditional measures of labour quality might have the shortcoming of missing some features of the very important increase in labour utilization within European countries. In particular, we explore the case of Spain. Despite showing one of the most important increases in labour quality in the EU according to standard methods, it also offers a negative increase in TFP growth. This paper computes an index of labour quality in Spain between 1988 and 2006 using microdata from the Labour Force Survey and the Structural Earnings Survey-2002 that allows the introduction of all possible interactions in a semi-parametric fashion between gender, age, education, experience in the current job and nationality. Considering those observable characteristics, the index still shows a notable growth at an average annual rate of 0.42 pp. After a period of slight decline (between 1988 and 1992) the index grows continuously until 2006 when it fell again. This is the case because education is, even by considering all possible interactions with other demographic variables, the highest contributor to the quality index’s growth. However, the paper shows the importance of considering changes in average productivities of different socio-demographic groups over time. We include in the analysis two usually omitted variables that help explaining the recent productivity slowdown in Spain: type of occupation held by the individual and unobserved heterogeneity of workers. Both the inclusion of occupation and especially the entry of individuals with below-average productivity levels compared to precedent periods decrease the labour quality growth to an average annual rate of 0.20 pp. Indeed with the addition of these two factors labour quality slightly decreases from 1997 onwards