The past decade has seen rapid market concentration in the United States hospital industry (Cooper et al., 2019; Fulton, 2017). A large literature examines how concentration affects consumers with private insurance and Medicare. However, there is scant evidence on how decreasing market competition affects low-income Americans, particularly those covered by Medicaid. Medicaid typically reimburses health care providers at lower rates than do private insurers, resulting in large gaps in care between Medicaid and other patients. Whether decreased competition reduces or exacerbates inequalities in use and outcomes is theoretically ambiguous. Concentration, which is driven by mergers among other factors, allow hospitals to gain market power and command higher reimbursement rates from private insurers (Gaynor et al., 2015; Dranove and Lindrooth, 2003; Vogt and Town, 2006; Capps and Dranove, 2004; Gaynor and Town, 2011; Haas-Wilson and Garmon, 2011; Vita and Sacher, 2001; Dafny et al., 2019). This could lead to larger gaps between private and Medicaid payment rates, and a renewed focus on privately insured patients at the expense of Medicaid patients. Alternatively, hospitals might use surplus revenue to cross-subsidize care for low-income Medicaid populations (David et al., 2014). Finally, economies of scale and efficiencies resulting from hospital consolidation could lead to expanded and improved care for all populations.

To investigate this question, we estimate the effects of hospital market concentration on hospital-based inpatient care for Medicaid patients in New York State from 2006-2012. New York State is an ideal setting for this analysis because it has a large Medicaid population and has experienced substantial consolidation (Huckman, 2006).

We measure effects of concentration on the number and proportion of hospitals’ inpatient stays for Medicaid patients. Our analyses focus on inpatient volume because inpatient care accounts for a large proportion of Medicaid spending and access to care is a key policy concern for low-income patients. Among other factors, barriers to access stem from providers’ limited financial incentive to treat Medicaid patients. Prices for Medicaid patients are largely tied to state-regulated fee-for-service reimbursement rates while prior research finds that increases in hospital market concentration lead to rising private insurance prices. This difference means that concentration could further disincentivize providers from serving Medicaid patients by widening the reimbursement rate gap between Medicaid and private payers.

We first construct a hospital-specific Herfindahl-Hirschmann Index (HHI)-based measure of market concentration. Because hospitals can operate across multiple service and geographic markets with varying levels of competition, our hospital-specific measure of concentration accounts for each hospital's geographic and service mix (Capps et al., 2017). To test our hypothesis that hospitals’ greater bargaining power with private insurers could impact Medicaid volumes, our measure of concentration is based on admissions for privately insured patients. We conduct falsifications using Medicare-based HHI.

We find that controlling for fixed hospital characteristics, an increase in hospital-specific HHI is associated with a relative decrease in the number of Medicaid patients a hospital admits: a 1% increase in HHI leads to a 0.6% relative decrease in the number of Medicaid patients admitted by a hospital. Decreases were observed in labor and delivery (birth) admissions, the most common reason for admission in Medicaid, and to a lesser extent in non-birth admissions. Because several factors could contribute to changes in market concentration, we construct an alternate measure, which isolates variation in HHI over time to only that caused by hospital mergers or acquisitions. We find evidence that mergers lead to hospital-level decreases in birth admissions for Medicaid beneficiaries, though not for non-birth admissions.

We test whether hospital-level decreases in our main analysis reflect redistribution of patients across hospitals or overall declines in the likelihood of a Medicaid admission at any hospital. We find very small decreases in total admissions for non-birth admissions. In heterogeneity analyses, we find that increased market concentration is associated with declines in Medicaid volume for non-profit hospitals but increased volume for public hospitals, suggesting that public hospitals may absorb Medicaid volume in response to increasing market concentration.

We investigate a potential mechanism by which hospitals in more concentrated markets could engage in screening out Medicaid admissions. For birth admissions, we find that an increase in market concentration is associated with a within-hospital decrease in attending physicians who serve higher shares of Medicaid patients. This association suggests that hospitals may achieve reductions in Medicaid volumes via changes in the types of physicians granted admitting privileges. However, we note that the same empirical pattern could also be consistent with the physicians who serve high shares of Medicaid patients selecting out of hospitals in more concentrated markets.

Our results indicate that increasing hospital market concentration affects health care for Medicaid patients and suggests the need for a more comprehensive and inclusive definition of the consumer in antitrust review of hospital merger cases, beyond the privately insured market segments traditionally analyzed. In particular, our analyses support previous findings that non-profit hospitals do not use surplus from greater bargaining power to invest in safety net populations. Moreover, decreasing market competition may increase strain on public hospitals through a shift in Medicaid admissions.

Our work relates to several strands of research. We contribute to the literature on hospital consolidation and competition by empirically investigating impacts on volumes, with a focus on low-income populations insured by Medicaid. A large body of evidence demonstrates that lack of competition in the health care delivery system increases insurer-negotiated prices for health care services as providers gain bargaining power (Cooper et al., 2019; Vogt and Town, 2006). Increasing prices paid by private insurers may widen the gap in hospitals’ profitability from admitting privately insured patients versus patients insured by payers with lower reimbursement rates, like Medicaid. To our knowledge, there has not been previous examination of the effects of market concentration and competition on health care use by Medicaid patients, despite the fact that state Medicaid programs cover about a fifth of the population nationally (Kaiser Family Foundation, 2018; Torio and Moore, 2016).

More generally, there is little evidence on the distributional impacts of heightened merger activity. Economists have debated whether or not corporate mergers exacerbate inequality and differentially harm low-income individuals in the economy writ large. These concerns are particularly salient in health care, given our reliance on hospitals to serve patients with varying levels of insurance generosity (Furman and Orszag, 2015; Roberts, 2012). The work examining impacts of increased market concentration on inequality has examined wage suppression as a mechanism (Prager and Schmitt, 2019). We contribute to this work by examining how changes in market competition as a result of hospital mergers may impact disparities in hospital care use.

Finally, research focused on multi-payer health care markets has found evidence that changes in demand (for example, due to insurance expansions) often affect payer mix (Kolstad and Kowalski, 2012; Joynt et al., 2013; Glied and Hong, 2018). However, it is unclear how providers screen out patients. In this work, we investigate mechanisms by which hospitals may achieve reduced Medicaid admissions. We find evidence suggesting that hospitals may limit labor and delivery admissions by withholding admitting privileges for physicians serving high shares of Medicaid patients, though we note that other factors such as changes in physician preferences could also explain reduced admissions among Medicaid-focused physicians.

The paper proceeds as follows. Section 2 discusses relevant theoretical frameworks. Section 3 describes our data sources and sample construction. Section 4 presents our main analyses. Section 5 presents analyses on effects by hospital ownership and section 6 presents analyses on admitting privileges as a potential mechanism underlying the main effects. Section 7 concludes.