Statement by
Mark B. McClellan, MD, Ph.D.
Centers for Medicre and Medicaid Services
Department of Health and Human Services

Tax Examption for Hospitals and Federal Payment for Uncompensated Care
Committee on Ways and Means

Thursday, May 26, 2005

Chairman Thomas, Representative Rangel, distinguished Committee members, thank you for inviting me to testify today about the tax exempt status of many of our nation's hospitals and the way in which the Centers for Medicare & Medicaid Services (CMS) assists hospitals who provide uncompensated care.

As you are aware, our nation's hospitals frequently treat patients who do not have the ability to pay, or who can only pay a portion of their bill. This is one of the many ways in which a nonprofit hospital can promote the health of the community it serves. The Federal government and state governments have granted non-profit, tax preferred status to hospitals that operate for the benefit of the community. Today's hearing primarily seeks to review what we know about the value of the uncompensated and under-compensated care provided by these non-profit hospitals, and the tax benefits and other support they receive. Although the Committee is focusing on the issue of tax exempt status for hospitals, which is within the purview of the Treasury Department and the Internal Revenue Services, it might also want to review current policies that exist to assist hospitals that provide uncompensated care and to consider whether funds used in those efforts are providing care in the most efficient and effective manner possible. To address this issue, a number of related questions are relevant, including the extent to which quality, costs, and behavior of non-profit hospitals differ from for-profit and public hospitals.

Current Research
Some believe that the financial rewards inherent in for-profit ownership might provide incentives for hospitals to contain costs and respond effectively to patients' needs - for the same reasons that free markets work in the economy at large. Others believe that, because it is difficult for patients and society to evaluate the quality of health care, the opportunity to earn profits might lead hospitals to take advantage of patients or otherwise "cut corners."

These differences are at the root of several important policy debates. Should non-profits' exemption from taxes, access to tax-exempt bonds, and ability to solicit tax-deductible charitable contributions be preserved or be limited? To what extent does tax status correlate with hospital performance or better patient outcomes? To assist the Committee with its deliberations on these questions, we have compiled findings from relevant research.

Economists and health policy scholars have conducted numerous studies to better understand the relationship among for-profit, non-profit and public hospitals. Prior to my government service, as an academic researcher, my colleagues and I analyzed data on the medical expenditures, mortality, and rates of cardiac complications of elderly Medicare beneficiaries hospitalized for new heart attacks between 1985 and 1996. We found that geographic areas with for-profit hospitals have approximately 2.4 percent lower levels of hospital expenditures per patient as areas without for-profit hospitals but virtually the same patient health outcomes.(1)

Areas with for-profits have both lower labor and lower capital costs. When an area's elderly population declines, for-profit hospitals eliminate unneeded beds quickly, whereas non-profits eliminate them much more slowly. (Interestingly, public hospitals are almost as responsive to population declines as for-profits.)

These effects are a combination of direct effects of for-profits on their own patients' costs and of "spillover" effects on neighboring non-profits' behavior. That is, the neighboring non-profits also start lowering their costs. The bulk of the 2.4 percent savings is achieved when the for-profit presence increases from near zero to only a small fraction of admissions in the area. Direct effects of for-profits on their own patients' costs cannot by themselves account for the savings we observe.

Our study is only one piece of a larger puzzle. We evaluated the effects of ownership on only one facet of health care—productivity, or the quality and cost of care for individual patients. Other studies find that ownership may affect important social and economic outcomes, such as hospitals' propensity to exploit Medicare's complex regulated price system, or the volume and quality of care for uninsured patients. Also, we examined only one illness and one patient population; the effects may be different in other settings. Finally, although our measures of health outcomes cover the common adverse outcomes that matter to patients (serious enough complications to cause readmission to the hospital), they may fail to capture fully all of the health consequences of differences in ownership. Other studies have addressed these additional issues.

Most studies have found little difference in the community benefits provided by for-profit versus non-profit hospitals, where community benefits are defined to include uncompensated care and the provision of unprofitable or non-reimbursable services.(2) Indeed, some studies find that non-profits actually treat fewer indigent patients than do for-profits.(3) There is some evidence that public hospitals that convert to for-profit status reduce the amount of uncompensated care that they supply. However, public hospitals that convert supply much lower levels of uncompensated care pre-conversion than public hospitals that do not convert.(4)

I worked jointly on a study that compared patient outcomes in for-profit and non-profit hospitals between 1984 and 1994 using a new method for estimating differences across hospitals that yields much more accurate estimates of hospital quality than previously available. While we found that, on average, for-profit hospitals have higher mortality among elderly patients with heart disease, much of the difference appears to be associated with the location of for-profit hospitals. We noted in our paper for the National Bureau of Economic Research, "Within specific markets, for-profit ownership appears, if anything, to be associated with better quality care. Moreover, the small average difference in mortality between for-profit and non-profit hospitals masks an enormous amount of variation in mortality within each of these ownership types." In other words, hospital-specific factors besides ownership were much more important influences on hospital performance than ownership alone.(5) A later study by Yu-Chu Shen examined the effect of hospital ownership type on patient outcomes after treatment for acute myocardial infarction. Shen found that for-profit and government hospitals have higher incidence of adverse outcomes than non-profit hospitals.(6)

The debate over the effects of for-profit ownership of hospitals must reflect a range of policy considerations. But with expenditures on hospital care running into the hundreds of billions of dollars each year, the productivity benefits of free markets and competition deserve careful consideration.

Uncompensated Care Provided by Hospitals
CMS does collect data on uncompensated care through the hospital cost reports, but we do not have data on the value of the tax exemptions that non-profit hospitals receive. Much of the data that CMS collects in hospital cost reports are tied to payment and the Medicare trust fund. CMS ensures that these data are rigorously audited. Data on uncompensated care, however, do not impact Medicare payment, and hence, are not audited. Further, the most complete data collection instrument in the cost report is relatively new, and hospitals are still becoming accustomed to reporting uncompensated care data in a standard format. Thus, these data are somewhat less reliable than those used to make payments. CMS staff has engaged in discussions with the Medicare Payment Advisory Commission (MedPAC) and the Government Accountability Office (GAO) concerning this issue and is considering ways to make the data on uncompensated care more reliable and useable. However, at this time, it is not possible for CMS to make the sorts of comparisons that are needed to answer the primary question of today's hearing. Congress has, however, mandated that CMS make certain payments to hospitals in recognition of their role in providing uncompensated or under-compensated care and I would like to describe those payments here today.

Before detailing these payment mechanisms, it is important to review the requirements of CMS' regulations with respect to the uninsured, or underinsured. The provider payment rules for Medicare and Medicaid in no way restrict the ability of hospitals and other providers to offer free or discounted care to patients who do not have coverage under these two programs, or to offer relief to Medicare and Medicaid beneficiaries who simply cannot afford to fulfill their responsibility for co-payments and deductibles.

Nearly two years ago, CMS was approached by a number of hospitals requesting guidance concerning whether it was permissible to discount charges to low-income, uninsured, or underinsured patients. In December of 2003, then-Secretary Thompson received a letter from the American Hospital Association (AHA) that alleged that Medicare program rules, as well as restrictions imposed by statutory authorities within the jurisdiction of the Health and Human Services (HHS) Office of Inspector General, hindered the ability of hospitals to provide discounts to low-income patients or to patients who were medically indigent. Secretary Thompson responded to the AHA letter in February 2004.

There are three central ideas addressed in the guidance provided by HHS to the hospital community:

Medicare billing requirements do not prevent discounts to uninsured patients as long as:

  • Full charges, not discounted charges, are reported on the cost report.
  • Accounts and records are maintained in a manner that would be necessary for any business.

Medicare indigency requirements do not prevent discounting to uninsured patients.

  • Providers may make indigency (including medical indigency) determinations using their customary methods.
  • In order to protect all patients and the Medicare program, the methods used in determining indigency for non-Medicare patients should be similar to those used for Medicare patients.
  • Indigency should be supported by documentation (good business practices would dictate that).
  • Indigence should be determined on a patient-by-patient basis because financial need is specific to each patient.
  • Medicare does not reimburse the bad debts of non-Medicare patients.
  • Once indigence is determined, collection is no longer undertaken with regard to the patient for the forgiven amount.

Medicare does not require providers to be aggressive in their collection of accounts. Medicare rules state that:

  • Efforts to collect from non-Medicare patients must be similar to the efforts to collect from Medicare patients. Medicare wants parity in the treatment of Medicare and non-Medicare patients to protect the program and all patients, not just our beneficiaries.
  • Efforts to collect on accounts should be more than a token effort. Rather, they should be proactive efforts that would be used by any prudent business.

Since the enactment of the Medicare program in 1965, the program's rules have attempted to prevent "cross-subsidization" – in other words, preventing the Medicare program from subsidizing a service that should be paid for by another payor, or preventing another payor from subsidizing a service the Medicare program should be reimbursing. One way that Medicare's regulations do that is to require hospitals to list their stated charges for a service on their cost reports and maintain a uniform charge for a service. To repeat, nothing in CMS regulations prevents a hospital from providing a discount off of that stated charge. But when filing its cost report, the hospital must list its full charges.

Without question, Medicare program rules permit a hospital to provide free care or discount charges to uninsured or underinsured patients. As we noted in our response to the American Hospital Association, "[n]othing in the Centers for Medicare & Medicaid Services' (CMS') regulations, Provider Reimbursement Manual, or Program Instructions prohibit a hospital from offering discounts to any patients, Medicare or non-Medicare, including low-income, uninsured or medically indigent individuals."

Therefore, in reference to the ability of a hospital to develop an indigency policy, it is overstating matters to say that the Medicare program imposes a "restriction" on this. Hospitals – not the federal government – set their own indigency policies and have the discretion and flexibility to define eligibility indicators including income level. This makes sense because a hospital, as a community institution, is in the best position to know what policy best suits the community that it serves.

As I have stated earlier, if a hospital wishes to provide a discount off of its customary charges as part of an indigency policy, it can do so, but it must report the full charge for that service on its Medicare cost report.

Turning to the issue of bad debt, we often hear from hospitals that Medicare somehow "requires" aggressive collection efforts that include seizing a patient's home, use of a bill collector, and other similar tactics. The reality is otherwise. The Medicare program does not require any particular level of collection activity. It does not require that collection activities be "aggressive." It does not require that hospitals seize patients' homes or bank accounts. What the program does require, however, is that if the hospital wants to bill the Medicare program for bad debt related to unpaid deductibles and coinsurance by Medicare beneficiaries, it must use the same level of collection activity as it does to secure collection of debts by non-Medicare patients. For example, if a hospital wants to use a bill collection agency for its bad debts, it cannot turn only non-Medicare patient bills over to that collection agency; rather, the hospital must treat all bad debts the same. But nothing requires a hospital to use a bill collection agency for its bad debts. The principle, again to prevent cross-subsidization, is that collection of Medicare and non-Medicare debts need to be treated similarly.

In addition, a hospital may make an individualized indigency determination for a particular Medicare patient and excuse that patient from any efforts to collect unpaid deductibles and coinsurance. Doing so would not prevent the hospital from collecting bad debt payments from Medicare on those unpaid amounts, provided the hospital treats all indigent patients the same. This is also true if the patient is a dually-eligible Medicare and Medicaid beneficiary. In such a case, the hospital would submit a bill for the unpaid deductible and coinsurance amounts to the state Medicaid plan. If the state Medicaid plan was not liable and denied payment on the account, the hospital could bill the Medicare program for it as a bad debt.

It is also important to note that in very limited circumstances, Medicare payment could be affected by the "lesser of cost-or-charges," or "LCC" principle. This principle was of significant importance in the early years of the program, but is admittedly less so now that most providers are reimbursed on the basis of a prospective payment methodology rather than on the basis of costs. However, where the LCC principle is applicable, a Medicare provider is paid the lesser of its actual costs or its actual charges. Implementing a reduced charge program for uninsured patients could potentially trigger the LCC principle because if a hospital lowered charges for enough patients, a hospital's fiscal intermediary could take the position that a hospital's charges were not its posted, or stated, charges, but rather, the charges applicable to most of its patients who were receiving discounted services. If the FI did take that position, it could then invoke the LCC principle and pay the hospital that lower charge-based amount. Few providers are subject to the principle at all. The only example I am aware of is a pediatric or cancer hospital in its first year of operation, before it becomes subject to the TEFRA methodology, because there are no base year costs upon which to calculate a TEFRA target rate limitation. Other providers, including critical access providers, are not subject to the LCC provision.

Medicare and Medicaid Payments to Hospitals Providing Care to Uninsured Individuals
A recent study calculated the revenue cost of the tax subsidy provided to non-profit hospitals and found that in 1994-95 it amounted to $9.21 billion 2002 dollars, including an exemption from income taxes of $5.43 billion, an exemption from property taxes of $2.01 billion, tax deductibility of donor contributions of $1.34 billion, and tax-exemption of interest paid on debt of $0.43 billion.(7) In addition, Medicaid and Medicare have several payment mechanisms to compensate hospitals for providing care to uninsured individuals. The President's FY 2006 Budget includes proposals to ensure that funds provided to hospitals to reimburse for uncompensated care are used appropriately.

1. Disproportionate Share Hospital Payments
Since 1986, select hospitals have received payment under the Medicare disproportionate share hospital (DSH) program. The original intent of DSH payments was to reimburse hospitals for increases in their Medicare costs that were associated with treating a large share of low-income patients. Since that time, several changes to the statutory formula have increased the likelihood that DSH payments also compensate hospitals for the costs of treating uninsured patients. Hospitals qualify for Medicare DSH payments if they treat a "disproportionate share" of low-income patients – defined in the statute as the share of a hospital's total inpatient days attributable to Medicare patients who are also eligible for SSI compared to days attributable to all Medicare patients, plus days attributable to patients who are eligible for benefits under Medicaid and also not eligible for Medicare compared to all patients. That ratio, along with consideration of urban/rural status and bed size, plays into a specific formula that yields 16 different categories of hospitals for DSH payment purposes. The payments themselves are a percentage add-on to the Medicare diagnosis related group (DRG) payments used to reimburse hospitals for inpatient services.

The Medicaid program also provides DSH payments. The formulas for establishing those payments vary with each state program, although there are certain categories of hospitals which must be designated as DSH hospitals by state Medicaid plans. States must designate hospitals as eligible for Medicaid DSH payments if they have a low-income utilization rate of 25 percent or more (LIUR is calculated as the sum of the ratio of Medicaid revenues divided by total revenues and the ratio of inpatient charity charges divided by total charges); or their Medicaid utilization rate (Medicaid days divided by total days) is more than one standard deviation above the mean Medicaid utilization rate in the state.

The Medicaid DSH program is also advantageous for states because DSH payments to a hospital under a state plan are not counted in determining whether or not the state has exceeded the Medicaid upper payment limit, thus enabling states to increase payments to other providers participating under their state plan. Preliminary data show that during 2004, Medicare DSH payments amounted to about $8.5 billion, while Federal and State Medicaid DSH payments totaled nearly $17.2 billion.

2. Bad Debt Payments
As I mentioned above, Medicare also reimburses hospitals and certain other providers for the bad debt that arises from treating Medicare beneficiaries who are unable to pay their cost sharing and deductible amounts. Providers who make reasonable efforts to collect Medicare co-payments and deductibles, but are unable to do so, can report those amounts as bad debt. Hospitals are paid for this bad debt at a rate of 70 percent. Other eligible providers receive payments amounting to 100 percent of their bad debt. In FY 2000, the latest year for which we have finalized data, CMS provided $1.03 billion in bad debt payments. For FY 2005, we estimate that bad debt payments will total around $1.6 billion.

3. A Portion of Indirect Medical Education (IME) Payments
As with the DSH, the IME is intended to recognize legitimate variations in hospitals' costs for treating Medicare patients. Academic medical centers that engage in graduate medical education incur higher costs per discharge as a result of their teaching activities. In recognition of that fact, when Congress instituted the inpatient prospective payment system (IPPS), it created add-on payments for these teaching institutions. One of those payment types was meant to cover the indirect costs of providing such education and is referred to as IME. IME payments are based on the estimated relationship between the hospitals' Medicare costs per discharge and their teaching intensity as measured by the ratio of residents to beds. Because Congress was unsure about the ability of the IPPS to fully capture differences in patient severity and other factors that might account for teaching hospitals' higher costs, the Congress required the Secretary to double the empirically estimated IME adjustment. Current law and the most recent data indicate that IME payments are still set at twice the estimated empirical effect of teaching activities on a hospital's cost per discharge. Some say that the difference between the empirical estimate and the current level of IME payments is a subsidy for uncompensated care. Projected spending for IME during 2004 stands at $5.2 billion and that projection rises to $5.7 billion for FY 2005.

The Medicare Payment Advisory Commission (MedPAC) has recommended revising this situation, but at the same time, recognizes that doing so may cause problems for these institutions. Teaching hospitals provide a high level of uncompensated care, amounting to 20 percent in major public teaching hospitals, but only 5 percent in major private teaching institutions. Medicare patients account for only a portion of total patient population, so even increased payments for Medicare services do not necessarily cover costs incurred for all uncompensated care. Nevertheless, the IME payments do provide funds that these institutions can and do use to cover that gap. Whether this approach to funding hospitals that provide uncompensated care actually results in the most healthcare per dollar invested, and is the most appropriate method for targeting those dollars to the uninsured, or indigent, is not clear.

4. Medicaid Waivers
Medicaid waivers allow states to explore new approaches to delivery and payment for health care services. In particular Medicaid section 1115 waivers have been used to develop new mechanisms to provide health insurance for uninsured individuals with a limited income.

The 1993 approval of the Hawaii Quest Demonstration program includes redirected DSH funds in order to provide insurance coverage rather than uncompensated care. Similar initiatives were included in the early TennCare program, New York Partnership, Vermont Health Access Program, Massachusetts MassHealth Program – all broad comprehensive statewide section 1115 demonstration programs that included eligibility expansions to uninsured populations. Also, more recently, CMS has collaborated with the State of Maine and the District of Columbia to reprogram Medicaid DSH funds to provide health insurance instead of uncompensated care. These two states have used demonstration authority to focus on redirecting Medicaid DSH payments to increase the insurance coverage in the states.

The State of Maine currently has a HIFA waiver under the authority of Section 1115 of the Social Security Act, implemented October of 2002, that has allowed the State to expand coverage to childless adults up to 125 percent of the Federal Poverty Limit. Maine chose to use its unspent Medicaid DSH Federal allotment to extend coverage to this population. This program has provided health insurance coverage to an additional 26,000 residents of the State of Maine (as of February 28, 2005).

The District of Columbia currently has a Section 1115 demonstration, implemented February of 2003, to provide primary and preventive health care services to non-disabled, childless adults, between the ages of 50 and 65, with income at or below 50 percent of the Federal Poverty Limit. The funding source for this demonstration is the District of Columbia's Medicaid DSH Federal allotment. This program was implemented with an enrollment cap of up to 2,400 people.

These two programs have expanded insurance coverage to more than 20,000 low-income childless adults. At the start of 2004 there were 20,900 childless adults up to 100% of the Federal poverty level (FPL) insured through the program in Maine and 2,400 childless adults up to 50% of the FPL in the District of Columbia. In both of these programs the participants receive benefit of the full Medicaid benefit package. These examples illustrate that states have options available, under demonstration authority, to move from reimbursing providers for direct care to increasing health insurance coverage.

In the state of Massachusetts, demonstration authority will be used to reimburse for primary care services for the uninsured and encourage the utilization of services that can prevent the need for more costly hospital services for these individuals. Under the MassHealth Section 1115 Demonstration, effective with the extension period beginning July 1, 2005, a Safety Net Care Pool will be established to pay for costs related to providing health care services to the uninsured. The Safety Net Care Pool (SNCP) will be established using a combination of demonstration savings, in addition to the Commonwealth's Medicaid DSH allotment.

5. Increased Funding for Community Health Centers
This Administration has undertaken other initiatives to provide health care services to individuals who otherwise lack access to health insurance or who may be under-insured. Community health centers (CHCs) serve as the "front line" treatment option for low-income uninsured individuals. They provide professional, family-oriented preventive and primary care to low-income individuals within their communities. Typically, about 40 percent of the patients of a community health center are uninsured. A study published in Health Affairs earlier this year illustrated the important role of CHCs as a reliable source of primary and preventive care for a vulnerable population.(8) According to the study, visit rates for uninsured CHC patients, individuals for which a chronic disease condition was "managed", and established CHC patients all increased. Furthermore, CHC patients experienced fewer hospitalizations and emergency room visits for ambulatory care, compared with similar people living in the same areas who seek care elsewhere. The President's Health Centers Initiative, which began in FY 2002, will open or expand 1,200 health center sites to serve another 6.1 million patients by 2006. The FY 2005 appropriation for community health centers exceeded $1.7 billion. The President has set a new goal to open a health center or rural health clinic in every poor county that can support one. The FY 2006 Budget level includes $26 million to open new health center sites in 40 of the Nation's poorest counties and will support 25 planning grants as well. These expansions complement the President's proposals to increase health insurance coverage in private and public insurance programs, to help ensure that all Americans have access to health care. The President's Health Centers Initiatives will broaden the health center safety net and increase access to primary health care for the Nation's underserved populations, thus reducing the amount of uncompensated care that must be provided by our hospitals.

6. Section 1011 of the Medicare Modernization Act (MMA)
Under the Emergency Medical Treatment and Labor Act (EMTALA), hospitals participating in Medicare must medically screen all persons requesting a medical screening examination to determine whether or not the individual is suffering from an emergency medical condition. In addition, if the hospital determines that the individual has an emergency medical condition, it must provide the treatment necessary to stabilize that individual, regardless of payment method or insurance status. As a result, hospital emergency departments treat uninsured or underinsured individuals who cannot pay for the services they receive.

Undocumented immigrants' use of medical services has been a long-standing issue for medical providers, particularly for hospitals located along the U.S.-Mexican border. Section 1011 of the Medicare Modernization Act (MMA) provides $1 billion through 2008 to help hospitals and other emergency providers recoup some of the expenses of providing this critical care to undocumented immigrants. Earlier this month, CMS announced the final implementation plan for hospitals and other providers to being receiving reimbursement under section 1011.

Between Medicare and Medicaid DSH, IME, bad debt payments, and the other mechanisms I have mentioned, the Federal and state governments will provide tens of billions of dollars this year to our hospitals to compensate them for the provision of uncompensated care. In addition to these payments, according to the study I cited earlier, non-profit hospitals will realize several billion dollars more in tax exemptions. As someone who trained as a physician in a teaching hospital, and who has conducted some published research this topic, I am familiar with how often these institutions provide uncompensated care and I know how valuable they can be to a community. However, the question that should be asked is whether the funding mechanisms I have mentioned most effectively target those funds to the programs and settings that provide the best value in terms of the type of care they provide.

Other Administration Initiatives for the Uninsured
The Administration has approached the issue of the uninsured along other lines as well, advocating giving an advanced health coverage tax credit to certain individuals who are receiving a pension from the Pension Benefits Guaranty Corporation or who have become unemployed due to the adverse effects of international trade and are eligible for Trade Adjustment Assistance. This tax credit pays 65% of the premium for qualifying health insurance, including either employer-sponsored "COBRA" coverage or a state-designated private health insurance plan. The Administration's Medicaid waivers, state plan amendments, and HIFA waivers have provided health insurance for 2.6 million people who would have otherwise lacked coverage.

Many of you in Congress voted for and deserve credit for the provisions in the MMA that accelerate adoption of health savings accounts and help make insurance more affordable for millions of Americans. In addition to creating a Medicare prescription drug benefit and providing interim savings and subsidies through Medicare-approved discount cards, this historic legislation allows people to establish health savings accounts (HSAs) in conjunction with affordable, high-deductible major medical coverage. These new products will make health insurance more affordable to businesses large and small, as well as to individuals whose employers do not sponsor coverage.

CMS strives to make sure that the payments we provide are in line with statutory requirements and that they meet the legitimate, data-driven needs of our partnering providers. The programs we administer serve some 80 million Americans and we want to be sure that we are supplying them, and those who serve them, with the best we are capable of. I appreciate the chance to appear before you this morning and look forward to any questions you may have.


1. Kessler, Daniel P. and Mark B. McClellan. "The Effects of Hospital Ownership on Medicaid Productivity." RAND Journal of Economics. Vol 33(3). Autumn 2002: 488-506.

2. Young, Gary J., and Kamal Desai. "Non-profit Hospital Conversions and Community Benefits: New Evidence from Three States." Health Affairs 8 (1999):146-55.

3. Duggan, Mark. "Hospital Market Structure and the Behavior of Not-for-profit Hospitals." RAND Journal of Economics 33 (2002):433-46.

4. Desai, Kamal, Carol V. Lukas, and Gary J. Young. "Public Hospitals: Privatization and Uncompensated Care." Health Affairs 19 (2000):167-72.

5. McClellan, Mark, and Douglas Staiger, "Comparing Hospital Quality at For-Profit and Not-For-Profit Hospitals," Working Paper 7324, National Bureau of Economic Research, August 1999.

6. Shen, Yu-Chu, "The Effect of Hospital Ownership Choice on Patient Outcomes After Treatment for Acute Myocardial Infraction," in The Journal of Economics, vol. 21, 2002, pp. 901-922.

7. Gentry, William M., and John R. Penrod. "The Tax Benefits of Not-for-Profit Hospitals." In David M. Cutler, ed., "The Changing Hospital Industry: Comparing Not-for-profit and For-profit Institutions." Chicago: University of Chicago Press, 2000.

8. O’Malley, Ann S., Forrest, Christopher B, Politzer, Robert M., Wulu, John T. and Shi, Leiyu. "Health Center Trends, 1994-2001: What Do They Portend For the Federal Growth Initiative," Health Affairs 24(2) 2005: 465-472.

Last Revised: July 22, 2005