Does 'Excess Supply' Drive Excessive Health Spending? The Case of Certificate-of-Need Laws.

Author:Bailey, James
  1. Introduction

    Certificate-of-need (CON) laws require health care providers to obtain the permission of a state board before opening, expanding into new lines of service, or making large capital expenditures. These laws were passed rapidly between 1964 and 1980 in the hope of restraining the growth of health spending. By 1980, every state but Louisiana had a CON program, and the federal government was pushing states to adopt CON. Since the Medicare payment reform and the end of the federal push for CON in the 1980s, fifteen states have repealed their CON laws.

    Earlier attempts to estimate the effects of CON laws on spending have suffered from a lack of theory--in particular, with regard to how the laws' effect should differ among payers and health services. I model CON laws as a reduction in supply and argue that they will only reduce spending when the demand for health care is price elastic. However, health care is generally estimated to be price inelastic, suggesting that CON laws are likely to backfire and increase total spending on health care for two reasons. First, inelastic demand means CON will increase the price of the services it targets more than it will reduce their use. Second, CON is not completely comprehensive. To the extent that sectors covered and not covered by CON laws are substitutes, CON that succeeds in restraining the use of covered care will increase the demand for, and spending on, uncovered care. I further show that CON likely reduces total welfare in the health care market. I show that results of the simple supply-and-demand model still hold after considering relevant complications such as moral hazard and point out other complications that more sophisticated theoretical work should explore.

  2. Background and Literature Review

    This section discusses the history of certificate-of-need laws and the goals behind their implementation, the laws' effect on quality and access to health care, and the previous literature on CON and spending.

    1. Certificate-of-Need Laws: History and Intentions

      The first CON law was passed by New York in 1964. Other states rapidly followed suit, and twenty-three states had programs in place by 1974. The rapid progress of CON laws was accelerated further when President Gerald Ford signed the National Health Planning and Resources Development Act of 1974 (P.L. 93-641). The law incentivized states to create CON programs, offering funding to those that did and threatening to withhold Medicare and Medicaid funds from those that did not. By 1980, every state except Louisiana had a CON program in place.

      The text of the 1974 federal law promoting CON identifies two main goals for the legislation: promoting equal access to health care and restraining cost growth. The law makes clear its intention of restraining health spending: "The massive infusion of Federal funds into the existing health care system has contributed to inflationary increases in the cost of health care.... Increases in the cost of health care, particularly of hospital stays, have been uncontrollable and inflationary, and there are presently inadequate incentives for the use of appropriate alternative levels of health care, and for the substitution of ambulatory and intermediate care for inpatient hospital care" (P.L. 93-641, Section 2a). The law plans to reduce spending growth through state-led planning: "In recognition of the magnitude of the problems described in subsection (a) and the urgency placed on their solution, it is the purpose of this Act to facilitate the development of recommendations for a national health planning policy, to augment area wide and State planning for health services, manpower, and facilities, and to authorize financial assistance for the development of resources to further that policy" (P.L. 93-641, Section 2b).

      These planning agencies are expected to achieve spending reductions by "preventing unnecessary duplication of health resources" (P.L. 93-641, Section 1513). State CON programs are expected to "provide for review and determination of need prior to the time such services, facilities, and organizations are offered or developed or substantial expenditures are undertaken in preparation for such offering or development, and provide that only those services, facilities, and organizations found to be needed shall be offered or developed in the State" (P.L. 93-641, Section 1523).

      The federal push for CON was repealed in the mid-1980s (P.L. 99-660, Title VII). Most states have been slow to respond, but fifteen have repealed their CON programs, as shown in figure 1. These CON repeals offer an opportunity to study the effect of CON: What happened to health care access, outcomes, and spending in the states that dropped CON compared to those that did not?

    2. Effect of CON on Access and Quality

      In addition to controlling spending on health care, CON legislation had an express goal of improving equality of access to health care, both by inducing providers to supply more indigent care and by limiting "cream skimming," where hospitals take only profitable patients while leaving unprofitable patients to others. One theory of how CON could reduce cream skimming is by restricting the creation of suburban and specialty hospitals in order to protect urban and rural hospitals serving poorer patients (Reinhardt, Reinhardt, and Reinhardt 1987; Stratmann and Russ 2014).

      Another theory is that CON boards could use their power over the approval of new projects to induce providers to offer more indigent care. Zhang (2008) finds this to be the case, estimating that CON laws result in a very slight (.07%) increase in the admission of uninsured patients. However, most of the literature indicates no effect or a negative effect of CON on access to care. Cutler, Huckman, and Kolstad (2010) find that CON increases travel distance for coronary artery bypass graft surgery, and DeLia et al. (2009) find that CON increases racial disparities in care. Stratmann and Russ (2014) find that CON programs do not increase the amount of indigent care provided.

      More recently, CON advocates have argued that CON increases the quality of care by promoting regionalization, moving patients into high-volume facilities that are associated with better health outcomes (Vaughan-Sarrazin et al. 2002). The literature on CON and quality has focused almost entirely on the quality of heart surgery and has indicated that CON may decrease heart surgery mortality (Ho 2006), increase it (Cutler, Huckman, and Kolstad 2010), or have no effect (Popescu, Vaughan-Sarrazin, and Rosenthal 2006). The only papers to examine how CON affects more general outcomes have found that it has no effect (Bailey 2018) or worsens mortality (Shortell and Hughes 1988).

    3. Previous Literature on CON and Spending

      In addition to the goals of improving quality and expanding access to care, CON laws are meant to reduce spending. The empirical literature on how CON laws affect spending has found mixed results, as summarized in table 1 (see Mitchell 2016 for a more thorough review of the literature). These mixed results may stem from the fact that the studies measure different types of spending.

      Conover and Sloan (1998) find that CON reduces spending on acute care by 5 percent but does not reduce overall health expenditures. Hellinger (2009) finds that CON reduces the number of hospital beds by 10 percent and argues based on other literature that this should translate into a 1.8 percent reduction in spending. Grabowski, Ohsfeldt, and Morrisey (2003) find that repealing CON for nursing homes has no effect on Medicaid nursing home spending, and Rivers, Fottler, and Frimpong (2010) find no effect of CON on hospital spending per patient. Lanning, Morrisey, and Ohsfeldt (1991) find that CON fails in its goal, increasing hospital spending by 18.5 percent and total health spending by 12.7 percent.

      What accounts for these differences? There is some variety in the empirical techniques employed by the literature; most papers use fixed-effects estimators, while Hellinger (2009) uses generalized estimating equations and Lanning, Fottler, and Frimpong (1991) use two-stage least squares. Some papers use a binary definition of CON, while others test the effect of CON stringency; moreover, different authors allow different lag times for the introduction or repeal of CON to take effect. But while these differences in specification can lead to...

To continue reading