The difference in the risk-adjusted cesarean rates of mothers who are and who are not privately insured is divided into components attributable to the following four factors: the practice style of the physician attending the birth, the closeness of the physician-mother relationship, individual nonclinical factors, and the direct influence of private insurance on the physician's treatment decision. Estimates from two expansive, complementary data sets indicate that the most of the differential is attributable to the first two factors, particularly the pairing of privately insured mothers with physicians who are inclined to perform cesareans. There is some evidence that these pairings are not incidental but the result of firm (and possibly consumer) choices.