In the analysis of worker absenteeism, a number of studies have sought to quantify the effect of the remuneration contract on worker behaviour. The study discussed here uses daily panel data gathered from an industrial firm to analyse these effects. It is argued that in using the firm's administrative data rather than survey data the modelling of unobserved heterogeneity is of particular importance. This paper therefore employs a flexible bivariate form for the unobserved heterogeneity in a random effects model, and particular asymmetries are found in the effects of contracted hours and sick pay costs. (C) 2002 Elsevier Science B.V. All rights reserved.
- bivariate heterogeneity