Routine-biased technical change can fail: Evidence from France

The paper studies the determinants of wage differentials over time within jobs in France, detailing the contribution of different set of explanatory factors by means of a Recentred Influence Function, to estimate the effect of a set of covariates at different point of the wage distribution. We simultaneously test the contribution of tasks performed by workers and organisational methods at the firm level, labour market institutions and individual characteristics. We do so by exploiting a unique database at the worker level, the French Enquête Complémentaire Emploi: Conditions de travail, between 2005 and 2016, which covers also monthly wages. Main findings support the hypothesis according to which wages differentials along the wage distribution are almost entirely explained by contractual and work arrangements rather than tasks and organisational practices. Overall evidence run against the main argument of the Routine Bias Technical Change hypothesis.