Abstract
The determinants of the unexplained component of the male-female wage difference in Sweden are examined. The hypothesis that there is a compensating wage differential associated with undesirable job characteristics is tested. Hedonic wage equations are estimated using data collected in the 1981 Swedish Level of Living Survey. The estimation takes into consideration possible sample selection bias due to the non-randomness of employment and the potential endogeneity of job choice variables in wage offer equations. Overall, the analysis indicates that compensating wage differentials are not a major determinant of the male-female wage difference in Sweden.