Sickness absence tends to be negatively correlated with unemployment rates. In addition to pure health effects, this may be due to moral hazard behavior by workers who are fully insured against income loss during sickness and to physicians who meet demand for medical certificates. Alternatively, it may reflect changes in the composition of the labor force, with more sickness-prone workers entering the labor force in upturns. A panel of Norwegian register data is used to analyze long-term sickness absences. The unemployment rate is shown to be negatively associated with the probability of absence, and with the number of days of sick leave. Restricting the sample to workers who are present in the whole sample period, the negative relationship between absence and unemployment becomes clearer. This indicates that procyclical variations in sickness absence are caused by established workers and not by the composition of the labor force. Copyright © 2005 John Wiley & Sons, Ltd.