We have analysed cancer mortality and cancer incidence among unemployed persons identified from the Danish linkage studies based on the 1970 census and the 1986 register-based census. In 1970, 1% of Danish men were unemployed; in 1986, 14% were unemployed. In both periods, unemployed men had an excess cancer mortality of close to 25% when they were followed-up for a five-year period and their mortality was compared with that of all men in the labour force. Unemployed women in the 1970 cohort also had an excess cancer mortality of 25%. Cancer incidence data were not available for the 1986 cohort. For both cohorts, the excess risk came mainly from lung cancer. Survey data from Denmark in the 1980s indicated that unemployed men had a slightly higher smoking prevalence before unemployment than men who continued working, and that unemployment did not increase smoking. It is therefore unlikely that the excess lung cancer risk among unemployed men is explained by differences in smoking habits alone.