Labour absenteeism may be detrimental to firms and society because of the economic costs, organizational problems and production cuts that it involves. Although involuntary absenteeism due to accident or illness that prevents workers from performing their work is unavoidable, avoidable voluntary absenteeism may also emerge due to asymmetric information given that neither employers nor doctors have perfect information about workers' health status. Assuming that there is heterogeneity in individual's behaviour and thus some workers are more likely to take sick leave than others due to differences in observable and unobservable characteristics, we specify a Finite Mixture Model to analyse sick leave days per year using a sample of employees from the 2014 European Health Survey in Spain. This specification accounts for unobserved heterogeneity in a discrete way assuming that there are two types of workers even though the data do not allow us to identify which group any individual belongs to. Our results reveal that, although health indicators have the greatest impact on the proportional change in days of absenteeism, there is heterogeneity in sick leave decisions and individual and job characteristics have different effect on the absenteeism of each group.
Keywords: Absenteeism; Finite mixture model; Sick leave; Unobserved heterogeneity.