We know how trustees decide how much to pay their CEOs—what factors they consider, and what they want to reward. So why are professors at the University of Connecticut trying to figure this out through statistical analysis? And why do they conclude that tax exemptions should be tied to the amount of charity care and community benefit an organization provides? Jeffrey Kramer and Rexford Santerre, professors at the University of Connecticut, published an interesting study of executive compensation at Connecticut’s hospitals in Inquiry (“Not-for-Profit Hospital CEO Performance and Pay: Some Evidence from Connecticut,” Fall 2010, pp. 242-251). The study covers 35 CEOs at 29 hospitals between 1998 and 2006. Hospitals’ pay practices encourage CEOs to fill beds with privately insured patients instead of Medicare, Medicaid or uninsured patients, the authors conclude and, for that reason, question whether hospitals deserve to be tax-exempt. The question derives from the assumption that tax-exemption is predicated on using what would otherwise be paid as taxes to provide uncompensated care, education, research, and outreach programs, or other community benefits, and the presumption that CEOs should therefore be rewarded more for charitable performance than for economic performance. Both the assumption and the presumption are wrong. The IRS long ago changed the rationale for hospitals’ tax-exemption to promotion of health and access to healthcare for the community as a whole. Regardless, even if tax-exemption were still predicated on providing free care to the poor, that would not mean that CEOs should be rewarded for increasing charity care, bad debt, and losses on Medicare and Medicaid patients. In exchange for tax-exemption, the authors believe, hospitals should pay CEOs for their success at fulfilling the hospitals’ charitable mission—for increasing charitable care and other community benefits. They believe that hospitals may instead allow CEOs to maximize the benefits they obtain from being CEOs, since there are no investors to counterbalance the CEOs’ interests. (Oddly enough, this posits the absurd notion that boards of for-profit firms do a better job governing pay than boards of not-for-profit organizations and the equally absurd notion that shareholders have any ability to constrain executive pay in public companies.) Previous studies, the authors note, demonstrate that pay for hospital CEOs is determined largely by size of the hospital; that there is little difference in total CEO pay between not-for-profit and for-profit hospitals; that there is no clear relationship between total CEO pay and financial performance at not-for-profit hospitals; and that there is no clear relationship between CEO pay and charitable performance. None of this is surprising. It can all be easily explained. Kramer and Santerre undertook their study “to sort out the precise relationship” between pay and performance. They developed a model to determine whether pay is more closely related to charitable performance or to economic performance. They decided to use number of staffed beds and the number of major procedures (CAT, MRI, linear accelerator procedures, and cardiac catheterization procedures) per FTE employee as the indicators of hospital size; operating margin and occupancy rate as indicators of economic performance; and uncompensated care (charity care and bad debt) and Medicare and Medicaid reimbursement, as a percent of operating revenues, as indicators of charitable performance. Their statistical analysis shows that CEO pay is proportional to the number of staffed beds and the occupancy rate, affected only modestly by the intensity of clinical procedures, barely affected by operating margin, slightly reduced by increases in charity care, and significantly reduced by increases in Medicare and Medicaid volume. The authors deduce from this that “not-for-profit hospital CEOs, at least in Connecticut, are motivated by pay incentives to increase the occupancy of privately insured patients at the expense of uncompensated care and public-pay patients”; that “economic performance takes precedence over charitable performance”; that “not-for-profit hospital CEOs may face the . . . incentive to cherry-pick additional privately insured patients and ignore community benefits at the margin.” What a lot of work to identify what any trustee could have told the professors! A few brief interviews with the conscientious trustees who determine executive pay could have told them a lot more, of course. They would have told the professors that of course they and their CEOs prefer privately insured patients, because they lose money on Medicare and Medicaid and on uninsured patients. They would eventually go bankrupt if the balance between privately insured patients and Medicare, Medicaid and uninsured patients got out of kilter. Hospitals couldn’t maintain their mission, they would say, if they preferred and pursued money-losing cases. The idea that they should incent or reward CEOs for increasing uncompensated care or Medicaid volume is preposterous and irresponsible. Then the trustees would explain how they determine CEO’s pay. First they choose a peer group; then decide whether to pay at median, or higher, and decide how much pay to put at risk. They reward good performance, they would say, on clinical quality, patient satisfaction, cost-effectiveness, and operating margin—not occupancy, not unreimbursed care, and not Medicare or Medicaid volume. Yes, they would admit that size and complexity are the principal determinants of pay; but they measure size by revenue and operating expenses, not by staffed beds, since so much of their business is now outpatient care; and they wouldn’t consider intentionally increasing or decreasing pay in relation to uncompensated care or Medicare or Medicaid volume. If Kramer and Santerre then asked a consultant how trustees determine CEO pay, they would say the same thing, except that a consultant would have explained that a study that ignores the boards’ intentions (as expressed in their compensation philosophy), the difference between independent and subsidiary hospitals, and distance from Manhattan couldn’t possibly begin to give an accurate explanation of CEO pay. Their study isn’t very informative. CEO pay is determined largely on the basis of size, the same as it is in any industry, and size is measured by revenues (or expenses), since that is the best way to measure the total volume and value of services provided. CEOs are paid to keep their institutions afloat, under circumstances that make this difficult: Americans expect unlimited access to the best possible care, nurses want higher pay and higher staffing levels, and doctors want hospitals to pay for the newest (and most expensive) technology and drugs, while employers and insurers, Congress and the states want to pay as little as possible. Trustees pay CEOs to manage the hospitals’ business as well as possible and reward them for the very things Americans want from hospitals—patient safety, clinical quality, attentive service, and cost-effectiveness.