The objective of this study was to explain the variation in cesarean section rates among hospitals (obstetrical departments) in Sweden, and to discuss its potential economic consequences. Using data from The Swedish Medical Birth Registry 1991, we made a cross-sectional study of the cesarean section rate at the departmental level. We identified some 20 determinants, demand-related as well as supply-related. A general model including all these regressors was specified. After reducing this model, we were able to explain about one-quarter of the variation. We conclude that the large variation in cesarean section rates indicates inefficiency, due mainly to overutilization, but perhaps also underutilization. It is difficult to calculate the economic consequences or the welfare loss to society. We estimated an additional cost for unnecessary cesarean sections of 13-16 million Swedish crowns (SEK) per year.