Cross-Market Theories of Harm to Competition in Healthcare


Economists Ink: A Brief Analysis of Policy and Litigation

View David A. Argue Profile
EI Principal David A. Argue has extensive experience in antitrust healthcare matters, including testimony on behalf of the defendants in St. Luke’s.  He co-authored, with Scott Stein of Sidley Austin, a more detailed article on this topic in the Fall 2015 issue of Antitrust.

One of the bedrock principals of the Merger Guidelines may be nudged further aside with new theories of harm to competition in “cross-market” transactions. These theories, which have captured the attention of antitrust enforcers, attempt to explain how harmful competitive effects can arise from mergers involving firms in separate product or geographic markets. For decades, the Guidelines and court decisions (including the Ninth Circuit’s decision in FTC v. St. Luke’s) have recognized the importance of market definition in competition analysis. But complaints by health plans that hospital systems have increased bargaining strength in multiple markets have recently drawn antitrust enforcers’ attention. An examination of the nascent theories and empirical research shows that they currently do not support increased antitrust enforcement.

Cross-market theories of competitive harm are founded on a concept of economic “linkages” between markets that enable a merger of suppliers selling in different markets to increase market power. In healthcare, these theories have gained the most attention in provider geographic markets, but they have also arisen in product market contexts, as in St. Luke’s. For hospital services, the linkages between geographic markets supposedly arise when an employer purchases health insurance for its employees who reside in distinct geographic areas and who do not consider hospitals outside of their area to be acceptable substitutes. The theories postulate that a merger that creates a cross-market hospital system can create market power because employers choose a single network of hospitals to cover all of their employees, notwithstanding that employees care only about the hospitals in their geographic area. By withholding its hospitals from a network, a hospital system creates “holes” in the network that make it less profitable for a health plan to market its product to cross-market employers. Thus hospital systems ostensibly gain bargaining leverage over the health plan by their ability to withhold hospitals in more than one market simultaneously.

A key hypothesis of the theories is that a health plan’s profits are affected disproportionately by a cross-market hospital system’s refusal to contract with the plan (e.g., the refusal by a two-hospital cross-market system to contract with a health plan is more harmful to the health plan than the refusal of the two hospitals individually). In one version of the theory, each additional hole in a hospital network has a greater incremental effect of reducing the probability of employers’ choosing the network. In another, as the health plan adjusts its premiums in response to network holes, its profits decline at an increasing rate as the number of holes increases.

Numerous assumptions undergird these models, several of which do not withstand scrutiny. In the “Employer Choice” model, the competitive danger arises from merging hospitals into one system that, through cross-market linkages, gains incremental market power over customers that purchase in both markets simultaneously. Notably, purchasers that do not require hospital services in both markets (i.e., single-market health plans or employers) are not subject to the exercise of market power created by cross-market linkages. Because the lack of employee substitution between local markets means that single-market employers are immune to cross-market leverage, cross-market employers can prevent cross-market systems from increasing their bargaining leverage simply by offering single-market network options to each set of employees. As long as employers acquire health insurance through a health plan rather than purchasing services directly from the hospital, the hospital system would not be able to identify cross-market employers and price discriminate against them.

The theories also frame the economic transaction as a purchaser contracting for the option to use a single bundle of hospitals. The implications of that framing, however, conflict with market realities. Substitution in the bundle implies that an employer would accept a less favorable network for employees in one hospital service market if it could get a sufficiently more favorable network in the other market. The employer as the bundle purchaser would view that trade-off as an overall improvement, even though its employees with the less attractive local network would be worse off. This conclusion is at odds with employers’ common practice of treating employees equitably. Further, the employer would likely find it difficult to internalize a trade-off of hospital options among its employees.

The “Employer Choice” model relies on additional assumptions that are not likely to be generally valid. For example, the disproportionate effect of a network hole on the probability of a network’s being chosen by an employer depends on the employer’s having a strong pre-existing preference for that specific health plan. Absent that strong preference, the disproportionate effect is no longer evident and the cross-market linkages disappear.

In the “Health Plan Pricing” version of the theory, a cross-market system that creates network holes by withholding its hospitals causes the health plan to have lower premiums and profits. One of this model’s restrictive assumptions requires the health plan to offer an employer the same premium in all markets, regardless of the strength of the local network. That assumption means that a cross-market hospital system has greater bargaining power by threatening to cause a plan’s premiums and profits to decline across all markets even if it creates a hole in just one market.

Empirical tests of these theories have attempted to capture a cross-market impact attributable to hospitals’ forming systems. Some of these studies have identified higher prices paid to cross-market hospital systems, but they fail to account for other factors that could increase hospital prices without signaling competitive harm. These factors include systems having more sophisticated bargaining teams backed with high-quality contract analyses, the value a single system adds to multiple local networks, improvements in hospital quality, and improved ability to bear contracting risk. In addition, these studies may inaccurately measure the hospital price variable and omit non-price contract terms.

The novel economic theories of cross-market competitive effects currently are insufficient to support an extension of the frontier of antitrust enforcement beyond the bounds of the Merger Guidelines. The models rely on some important assumptions that both limit their applicability and undermine their validity. Likewise, numerous plausible alternative explanations need to be accounted for before the empirical analyses supporting these models are given credence. Absent sound economic theories of cross-market competitive effects, antitrust enforcers should remain skeptical of health plans’ complaints that cross-market hospital system transactions harm competition. More compelling analytical approaches exist in the Guidelines framework and the St. Luke’s appellate decision that require market power to be shown in well-defined markets.