By: Bernard Chao
Introduction
In December 2025, the US Centers for Medicare & Medicaid Services (CMS) published proposed rules for two mandatory models that link Medicare drug reimbursement to international benchmark prices. The GLOBE model (Part B) would require manufacturers of high-cost single-source drugs to rebate the difference between US prices and a benchmark derived from international prices among 19 economically comparable countries. The GUARD model (Part D) applies a similar mechanism to a subset of Part D drugs, benchmarked against the same 19 countries. If finalized, GLOBE takes effect in October 2026 and GUARD in January 2027.
The stated objective is to ensure that US Medicare does not pay more than peer nations for the same therapies. The models are the most concrete implementation of most-favored-nation (MFN) pricing attempted by a US administration. As a threshold matter, the models depend on CMS’s ability to identify what reference countries actually pay for drugs. Because pharmaceutical pricing increasingly relies on confidential rebates, this is far from assured. Even assuming the US can overcome this information problem, the policy creates a set of incentives whose consequences for reference-basket countries deserve close examination. This commentary identifies two principal risks: upward price pressure in reference markets, and a self-defeating dynamic in which manufacturer withdrawal erodes the very benchmarks on which the models depend.
The Incentive Problem
Under both models, a manufacturer’s net price in any reference-basket country becomes a potential ceiling on US reimbursement. A discounted price negotiated by Australia’s Pharmaceutical Benefits Scheme or the UK’s National Institute for Health and Care Excellence would directly reduce the manufacturer’s US revenue on the affected product. Crucially, because US Medicare volume for a given drug typically dwarfs the volume in any single reference country, a price concession in a reference market reduces the US benchmark by the same amount but the resulting revenue loss is a multiple of the concession’s value in the originating market. GLOBE alone covers drugs generating more than US$100 million in annual Medicare spending across oncology, rheumatology, immunology, ophthalmology, and endocrinology.
This asymmetry shapes a rational manufacturer response. A recent analysis in Value in Health found that importing foreign reference prices may not sustainably lower US prices, in part because manufacturers adjust launch and pricing strategies in response. Early policy analysis similarly warns that pharmaceutical companies may delay or forgo launches for reference-basket countries.
Which Countries are Most Exposed?
The 19 GLOBE reference countries were selected on the basis of real GDP per capita (at least 60% of the US level) and annual GDP exceeding US$400 billion (Figure 1). This basket spans a wide economic range. Ireland’s GDP per capita exceeds the US level, though this figure is inflated by multinational corporate profit shifting. At the other end, Japan and Israel sit near the 60% threshold.
Countries near the eligibility boundary face the greatest risk. They offer manufacturers the smallest revenue opportunity relative to the US price liability they create. A 2022 report by the European Federation of Pharmaceutical Industries and Associations found that external reference pricing was a root cause of delayed application and unavailability of innovative medicines across European markets. The GLOBE and GUARD models would substantially increase the stakes of this dynamic.
The Economics of Price Convergence
The current global pharmaceutical market operates through price discrimination: manufacturers charge different prices in different countries, segmented by willingness and ability to pay. If this pricing structure allows manufacturers to serve countries that would not be served at a uniform price, it increases global welfare. On the other hand, if those countries would be served regardless, it merely shifts the burden of paying for drug development between countries.
MFN pricing disrupts this structure by forcing price convergence across markets. The policy assumes convergence will be downward: US prices falling toward international benchmarks. If manufacturers cannot sustain low reference-country prices without eroding US revenue, they face three options: 1) raise prices in reference markets toward US levels; 2) delay or withdraw from reference markets entirely; or 3) absorb the revenue loss. The third is the outcome the policy intends, but it requires manufacturers to accept lower total revenue. Whether a manufacturer withdraws from a given reference market depends on the size of the US revenue loss relative to the profit earned in that market. Where the reference-country margin is large relative to the US exposure, manufacturers will stay. Where it is not, withdrawal or price increases are more likely.
Upward Price Pressure in Reference Countries
The most probable near-term consequence of MFN pricing is not market withdrawal but upward pressure on prices in reference countries. The predictable result is harder bargaining positions during routine renegotiations with reference-country payers, slower negotiations, and resistance to the discounts that these health systems have historically secured.
This is not merely a theoretical concern. Chris Klomp, a senior health official in the Trump administration, has stated publicly that the bilateral MFN deals are intended to raise prices in peer countries, not to lower US prices, telling manufacturers to “price wherever you want” so long as they do not “undercut us in another wealthy country.” The dynamic is particularly concerning because it operates quietly. A manufacturer does not need to exit a market to cause harm. It needs only to hold firm on price during the next contract renewal. The reference-country payer then faces a choice: accept the higher price or lose access to the therapy. For health systems that rely on their negotiating power to manage pharmaceutical budgets, MFN transforms a bilateral negotiation into one where the manufacturer’s reservation price is effectively set by the US market.
The Self-Defeating Dynamic
A more extreme manufacturer response is to delay or withdraw from reference markets entirely. This creates an additional problem beyond lost access: the GLOBE and GUARD models depend on the existence of negotiated prices in reference-basket countries. If manufacturers withdraw, CMS loses the price data it needs to calculate international benchmarks. Fewer reference prices mean a less effective benchmark, which means less downward pressure on US prices. The policy designed to lower costs may, by changing manufacturer behavior, weaken its own mechanism.
This is not a distant hypothetical. Manufacturers already engage in strategic launch sequencing, and the academic literature on international reference pricing documents price convergence effects and launch delays in countries that serve as reference markets. Bristol-Myers Squibb has announced that it will charge the same price for its new schizophrenia drug in the United Kingdom as in the United States and will not launch in the UK if the National Health Service insists on a discount. The GLOBE and GUARD models raise the stakes by making the financial penalty for low foreign prices explicit and substantial. Separately, the Trump administration has negotiated bilateral MFN agreements with individual manufacturers, in which companies commit to pricing concessions in exchange for tariff relief. These bilateral deals reinforce the same incentive structure through a different legal mechanism. The temporal gap between these deals and actual market launches means that MFN pricing commitments may be difficult to verify.
Conclusion
CMS is not unaware of these dynamics. The proposed rules include provisions for recalculating benchmarks from whatever reference-country price data remains available, and for defaulting to the standard rebate when no benchmark can be identified. However, these mechanisms address a symptom rather than the underlying incentive. A fallback that substitutes missing prices with the next-lowest available reference price does not resolve the strategic incentive to withdraw. It merely shifts the pressure to the next country in the basket. There is also a measurement problem. The default benchmarking method (Method I) relies on commercially available list and invoice prices, yet confidential discounts negotiated between manufacturers and reference-country payers routinely reduce actual transaction prices by 20 to 60 percent below published levels. Because Method I cannot observe these rebates, it will systematically overstate reference-country prices. The alternative (Method II) would use actual net prices, but submission is voluntary, and manufacturers will submit only when doing so raises the benchmark in their favor. The models therefore face a fundamental information asymmetry: the benchmarks may not reflect the prices that MFN is designed to match.
The GLOBE and GUARD models are still at the proposed-rule stage, and the international health policy community has a narrow window to assess these dynamics before the models take effect. The most likely consequence for reference-basket countries is not dramatic market exit but a quieter erosion of their bargaining position, resulting in higher prices and tighter pharmaceutical budgets. The more extreme scenario of manufacturer withdrawal compounds this harm by additionally undermining the models’ own benchmarking mechanism. Several open questions warrant attention: Should reference-basket countries coordinate their negotiating response? Should CMS build in minimum-basket-size safeguards that trigger alternative pricing mechanisms? And should the models index to a composite measure rather than the lowest individual price? These are problems for policymakers in the reference-basket countries, for CMS as it finalizes the rules, and for the broader global health community.
Acknowledgements
The author used Claude (Anthropic, version Claude Opus 4, 2026) as a research and drafting aid during the preparation of this manuscript and to create Figure 1. All analytical judgements, policy interpretations, and final content were reviewed and approved by the author, who takes full responsibility for the work.
About the Author:
Bernard Chao is a professor of law and the Maxine Kurtz Research Scholar at the University of Denver Sturm College of Law, where he serves as co-director of the Intellectual Property and Technology Program. His scholarship spans several substantive areas, including patent law, privacy law, remedies, and jury decision-making. He frequently applies empirical methods and engages in comparative legal analysis. Recent works have addressed the differences in U.S. and European patent law as applied to drugs and data protection laws. Professor Chao’s writings have been recognized with the Samsung-Stanford Patent Prize and included in West/Thomson’s annual Intellectual Property Law and Patent Law Reviews. He has authored several amicus briefs to the U.S. Supreme Court in collaboration with the Harvard Cyberlaw Clinic and the Electronic Frontier Foundation, among others. His articles have appeared in both leading law review and peer review journals including the California Law Review, the New England Journal of Medicine and the American Business Law Journal.
Declaration of Interests:
Bernard Chao has no conflicts of interest in the topics addressed in this commentary.
Submit your guest blog entry to be published on The Source Blog here!
Leave A Comment