Hospital surgical volume has been shown to correlate with short-term outcomes after cancer surgery, but the relationship between volume and cost of care is unclear. We sought to characterize variation in payments for cancer surgery and assess the relationship between hospital volume and payments. Using 2000 to 2007 Surveillance, Epidemiology, and End Results-Medicare data, we assessed risk-adjusted 30-day episode Medicare payments for elderly patients undergoing one of six procedures for resection of cancer. Payments for the index hospitalization, readmissions, physician services, emergency room visits, and postdischarge ancillary care were analyzed, as were data on 30-day mortality and complications. The analysis included 31,191 colectomies, 2,670 cystectomies, 1,514 pancreatectomies, 2,607 proctectomies, 12,228 prostatectomies, and 10,151 pulmonary lobectomies. There was substantial variation in cost; differences between the first and third terciles of cost varied from 27% for cystectomy to 40% for colectomy. The majority of variation (66% to 82%) was attributable to payments for the index admission rather than readmissions or physician services. There were no meaningful associations between total risk-adjusted payments and hospital volume. Surgical mortality was low, but complication rates ranged from 10% (prostatectomy) to 56% (lobectomy). Complication rates were not correlated with hospital volume, but occurrence of complications was associated with 47% to 70% higher costs. We found substantial variation in Medicare payments for these six cancer procedures. Cost was strongly associated with postoperative complications and primarily driven by differences in the cost of the index hospitalization. Efforts to prevent and cost-effectively manage complications are more likely to reduce costs than volume-based referral of cancer surgery alone.