This paper develops and assesses empirically a simple model of firms’ optimal decision regarding working hours, where productivity varies with hours and where the firm faces quasi-fixed labour costs. Using Belgian firm-level data on production, labour costs, workers, and hours, and focusing on the estimation of elasticities along the isoquant and the isocost, we find evidence of not only declining productivity of hours but also of quasi-fixed labour costs in the range of 20 per cent of total labour costs. The tentative conclusion is that firms facing such costs are enticed to raise working hours, even if this results in lower productivity.
Delmez, F., & Vandenberghe, V. (2018). Long Working Hours Make Us Less Productive but Also Less Costly. Labour. DOI: 10.1111/labr.12128
Previous version: Françoise Delmez & Vincent Vandenberghe, 2017. "Working long hours: less productive but less costly? Firm-level evidence from Belgium," Discussion Papers (IRES - Institut de Recherches Economiques et Sociales) 2017022, Université catholique de Louvain, Institut de Recherches Economiques et Sociales (IRES).
Using an dynamic panel of 15 developed countries over the 1960-2010 period, we compare the recovery path of the employment rate to the recovery path of individual hours of work in the aftermath of recession. We argue that the repartition of total hours of work between the intensive and the extensive margin is, in part, a firm decision based on the cost of additional hours compared to the cost of an additional worker, in presence of fixed costs affecting only the extensive margin (creation of new job position) and not the intensive margin (extra individual hours). Using financial crises as an indicator for an increase in those fixed costs, we show that recoveries subject to higher fixed costs display a stronger recovery of individual hours and a weaker recovery of the employment rate. This can generate an apparent jobless recovery in terms of the employment rate, but the joblessness disappears as soon as one considers individual hours. The results are robust to controlling for the strength of the recovery in terms of GDP growth per capita, for the depth of the preceeding recession and for dynamic panel bias. In conclusion, we argue that considering both margins of employment improves our understanding of jobless recoveries as events displaying a change in the composition of total hours worked between the number of jobs and hours worked per job.
With Mathias Hungerbühler.
We develop a model where workers internalize the impact of their individual effort on the risk of being dismissed during bad times. We implement this mechanism within a simple general equilibrium model with efficiency wages. The resulting increase in effort during recessions matches empirical evidences of productive recessions and allows firm to optimally reduce wage or fire more workers for a given production level. Technically, the model also yields non-symmetric results for positive and negative shocks.