Ce document est lié à :
info:eu-repo/semantics/altIdentifier/doi/10.1016/j.labeco.2020.101807
info:eu-repo/semantics/OpenAccess
Pierre Pora et al., « A decomposition of labor earnings growth: Recovering Gaussianity? », HAL SHS (Sciences de l’Homme et de la Société), ID : 10.1016/j.labeco.2020.101807
Recent works have concluded to non-Gaussian features of labor earnings growth. We argue in this paper that it is mainly due to working hours' volatility. Using the non-parametric approach developed by Guvenen et al. ( 2016), we find on French data that labor earnings changes exhibit strong asymmetry as well as high peakedness. However, after decomposing labor earnings growth into wage and working time growth, the log-normality of hourly wages remains a quite plausible assumption since deviations from Gaussianity stem mainly from working time changes. The joint dynamics of hourly wages and working time help explain those deviations which relate most likely to labor supply decisions at the extensive margin.