The Federal Trade Commission (FTC) recently announced that it would sue to block Amgen’s proposed $27.8 billion acquisition of Horizon Therapeutics. The challenge represents a landmark in the history of pharmaceutical-industry antitrust enforcement, as the industry has largely been given license to engage in permissive mergers and acquisitions of smaller companies without challenge.
In Part One, I reviewed the basic structure and function of the pharmaceutical industry, as well as the theory of harm that the FTC is bringing. In this part, I take a much deeper dive into the economic literature to determine whether the FTC’s theory of harm is likely to hold up in court and whether the commission has picked the right forum in which to bring its claims.
The Economics of Loyalty Discounts
What are the economics behind the FTC’s theory of harm? Ultimately, the theory comes from a strand of the economics literature that analyzes exclusive deals and loyalty discounts.
As a brief review, the FTC is challenging the deal on grounds that Amgen might potentially engage in a practice known as “bundled rebating” after completing its purchase of Horizon. The commission believes that Amgen will be able to offer discounts on its existing portfolio of blockbuster products in order to secure quasi-exclusive “preferred placement” on insurance-company formularies, thereby blocking potential rivals to Horizon’s new rare-disease drugs from entering the market. The economic theories most relevant to these practices are those that study practices known as “exclusive dealing” and “loyalty discounts.”
An exclusive deal is a conditional-pricing practice in which a firm and a buyer sign a contract specifying that 100% of the buyer’s purchases will be made from a specific seller, usually in exchange for a discount on the per-unit price. A loyalty discount is a less extreme form of exclusive deal in which the share of purchases that must be shared between the two parties is not 100%.
As is perhaps obvious, it should be noted that the practices of exclusive dealing and loyalty discounts are not identical to the bundled-rebate practice that the FTC suggests a combined Amgen-Horizon could engage in. For example, both exclusive deals and loyalty discounts are conditional-pricing practices that deal with only one market at a time, markedly different than the FTC’s case.
The literature on exclusive dealing and loyalty discounts is nonetheless useful, as many conditional-pricing practices have similar effects on consumer welfare, even if the details of the practices may differ. Moreover, the legal literature and the FTC report I mentioned in part one both note that courts should treat bundled discounts as exclusive deals, with similar requirements to prove illegality.
No matter the specific details of the conditional-pricing practice, for there to be any consumer harm, the practice has to stand up to the analysis of exclusive dealing that Robert Bork offered in his 1978 book “The Antitrust Paradox.” Bork pointed out that, in the presence of a second firm that could enter a market, it would be impossible for an existing incumbent to profitably exclude the entrant. It might be able to sign an exclusive contract, but to induce the buyer to forgo the benefit of competition, it would have to offer the buyer a payment large enough to offset the price increases it could expect from forgoing competitive entry. Another way to pose this critique is to ask: “why would the seller willingly agree to sign a contract that robs it of the price reductions it could expect from entry?”
Bork’s argument was intended to analyze true exclusive deals, but it’s easily applicable to bundled rebates, as well. Why would a pharmacy benefit manager (PBM) voluntarily agree to place Horizon’s drugs Krystexxa and Tepezza in a preferred section of the formulary and exclude the benefits of competition for those products if it did not receive rebates that fully offset those losses? There may, indeed, be exclusion—to the distaste of the rivals to Krystexxa and Tepezza—but that doesn’t matter if the PBM has been fully compensated by equivalent-magnitude rebates on other products.
Economists have offered many models since 1978 in an attempt to prove that exclusion that harms consumers and is profitable for sellers is, indeed, possible. One feature they all appear to have in common is that, to get around the Bork critique, they include some form of “contracting externality” that shifts the harm from exclusion onto others who are not parties to the exclusive contract. The models most relevant to the Amgen case are ones that incorporate a first-mover advantage, because Krystexxa and Tepezza both are the first drugs to treat their given conditions.
The first potentially relevant model was originally published by Eric B. Rasmussen, J. Mark Ramseyer, & John S. Wiley Jr. in 1991 and expanded upon in 2000 by Ilya R. Segal and Michael D. Whinston. It posits both a firm with a first-mover advantage and the presence of significant scale economies for the potential entrant. Due to the presence of scale economies, the entrant in this scenario would only join the market if there sufficient remaining customers for it to break even after the exclusivity clause has been signed. Therefore, signing an exclusive contract—in this case, a contract to make Krystexxa and/or Tepezza the preferred treatment in a particular PBM’s formulary—could make it unprofitable for the new firm to enter, stranding the PBMs not party to the exclusive contract with monopoly prices. Here, PBMs are faced with a collective-action problem; it would be better for everyone if no one signed an exclusive clause, but it’s better for each individual firm to sign one.
Is this model a plausible explanation of reality? In at least one way, yes! The most stringent requirement—the presence of scale economies—is clearly met. Developing a competitor to Tepezza or Krystexxa would require significant upfront investment by any firm, while the marginal manufacturing and distribution costs, while not insignificant, are small in comparison.
There are, however, at least three complications to this simple story.
The first is whether an agreement to place a drug in a formulary’s preferred tier truly acts as an “exclusive” deal. Does it? Unfortunately, we don’t have great empirical studies to quantify the effects of moving a drug from a preferred tier (with little-to-no co-payment from patients) to a less-preferred tier (with a much more significant out-of-pocket obligation for the patient). The data we do have suggests that demotion from a formulary’s preferred tiers may precipitate a large drop in volume, but not nearly large enough to constitute a near-exclusive clause.
The best of these studies is an ongoing one from Kate Ho & Robin Lee, who directly estimate that moving a branded drug from a preferred to non-preferred tier drops the expected market share of that drug for the given PBM by roughly 10%, and that these numbers are consistent with the observed rebates in the industry. This is likely an underestimate for the specific drugs at the center of the FTC’s case, given that the Ho & Lee study focus on a drug class (statins) where generic competition exists and the number of branded options is quite large. Contracts that banned a PBM from including a competitor’s drug on the formulary at all (true exclusive deals), on the other hand, cause a much larger drop in drug volume, on the order of roughly 70%. But this evidence suggests that the effect of securing preferred placement is much smaller than a true exclusive deal.
A second important consideration is the market structure of the buyers—in this case, the PBM industry. The core reason for profitable exclusion in this case is because of coordination failures among the PBMs. The PBM market is highly concentrated and symmetric, however, with each of the “large” PBM having roughly equal market share. While this is often treated as problematic in antitrust analysis, it also makes it more likely that the PBMs could monitor the deals their rivals make with pharmaceutical manufacturers and thereby mitigate the coordination failure.
It’s also important to consider the effects of downstream competition. In the basic model, the buyers are the end-users. But in reality, PBMs also compete for the business of health insurers. Taking this competition into account requires acknowledging two forces at play. First, there is the important question of how much of the discount or rebate will the PBM get to keep for itself. Typically, it would be considered bad if PBMs kept rebates for themselves, rather than pass those discounts along to insurers (and, ultimately, patients) in the form of lower premiums. In this case, however, if the PBMs weren’t able to keep any of the benefits of a lower-priced drug, they’d be happy to exclude the entrant and keep their input prices high.
For example, let’s assume the price of one of Horizon’s drugs with an exclusive clause (inclusive of rebates) would be $700 per dose, and the price of a new entrant’s drug would be $300 per dose. PBMs only care if they get to keep some of the $400 per-dose difference for themselves. If they pass along every drop of savings to their customers, they have no incentive to fight hard for the $400 per-dose deal.
What does our knowledge about the PBM industry tell us about their pass-through rate? Unfortunately, this data is, once again, hard to come by in a notoriously opaque industry. We can, however, glean from the considerable outrage directed at PBMs in various fora that the pass-through rate is not 100%. Moreover, there is some—albeit hard to verify—data that suggests they keep about 10 to 15% of the rebates they generate. In this case, the ability of PBMs to keep the savings they negotiate suggests that they have incentives to prevent exclusion.
In summary, while the core requirements of an exclusive deal—the presence of scale economies—are met for the FTC’s case against Amgen, there are also alternative considerations that suggest PBMs may be able to avoid the problems of exclusive deals. The real issue with the FTC’s case, however, can be found if we take a broader look at the method that the commission has chosen to challenge this specific practice.
Why Challenge Bundled Rebates in a Merger?
The FTC’s concern with Amgen’s bundled-rebate practices, and with the broader practice of using rebates to secure preferred placement in formularies, is warranted. Economic theory supports the FTC’s fear that such practices could exclude rising competitors to drugs, hurting consumers in the process.
One question remains unanswered, however: why challenge this conduct in the context of a merger?
The problems with challenging bundled rebates in the context of merger enforcement are laid bare when examining another recent example: Pfizer’s $40 billion acquisition earlier this year of Seagen. Seagen is a maker of a new drug class called “antibody-drug conjugates,” which are used to treat cancer. Like Amgen, Pfizer is a large, diversified pharmaceutical company with multiple existing franchises. Seagen, like Horizon, is a much smaller biopharmaceutical company with a strong pipeline of drugs in development, and one to two core drugs on the market.
Sound familiar? As with the Amgen-Horizon deal, the FTC was presented with the purchase of a small pharmaceutical company that has approvals in narrow indications by a well-diversified pharmaceutical giant with products in much broader indications. And while the specific products that Seagen manufactures are cancer medications, rather than medicines for thyroid disease or gout, if we consider the relevant industry characteristics and participants, the two deals are nearly identical.
Why, then, is the Pfizer-Seagen merger not being challenged? The same concern in the Amgen acquisition—that Pfizer could sign contracts to offer rebates on its current blockbusters in order to exclude future rivals to Seagen’s products—applies here too. In fact, there at least five other deals that have similar characteristics still pending from this year alone!
The Pfizer-Seagen comparison demonstrates the danger in the FTC’s approach to challenging bundled rebates. Not only is it challenging a merger based on potential conduct before any actual anticompetitive conduct has occurred, but also in a situation where it is impossible to specify narrowly under what circumstances the conduct is likely to occur. In choosing to confront the issue of bundled rebates, the FTC has inadvertently challenged an entire industry’s business model.
This wouldn’t necessarily be an issue if there were no plausible efficiency benefits to the business model itself. But while the FTC is adamant that bundled rebates in pharmaceutical markets have no plausible efficiency justifications, the same is not true for the broader biopharma M&A ecosystem. As noted in part one, we have considerable evidence that the current business model, in which biopharmaceutical firms specialize on specific parts of the clinical-development process, is more efficient than the alternatives.
An alternative approach—notably, challenging the practice of bundled rebates in the context of Section 2 of the Sherman Act—removes the risk of chilling beneficial M&A activity, while preserving (assuming the FTC wins) the deterrent effect on future attempts by pharmaceutical firms to use bundled rebates to exclude competitors anticompetitively. Firms could invest in purchasing attractive new drugs, but also know that, if they attempted to sign potentially exclusive rebate contracts after the fact, the FTC would bring (and likely win) a suit.
It’s also notable that, when similar practices have been challenged in other industries (such as LePage’s Inc. v. 3M), as well as in the biopharmaceutical industry itself, the relevant forum has been Section 2 of the Sherman Act.
Challenging bundled rebates in the context of a Section 2 case has the additional benefit of allowing the FTC to target its choice of case to areas where a victory will result in maximum consumer benefit, in a way that merger enforcement does not. Challenging bundled rebates in the context of a merger constrains the FTC to consideration of harm in markets that are directly implicated by the mergers.
In this case, for example, even if the FTC’s theory of harm was correct, it would only be granting consumer relief for patients in two rare-disease markets that together constitute less than 15,000 patients annually. While all consumer harm, particularly in health-care markets, is undesirable, other markets may represent much more consumer “bang for the buck.” For example, the private antitrust suit that Regeneron filed against Amgen over the drugs Repatha and Praulent (discussed in part one) could apply to a significant chunk of the estimated 40% of Americans with high cholesterol. Moreover, assuming that these practices are relatively common, there could be much larger markets where more harm to consumers is happening that the FTC is simply choosing not to confront.
There’s one last reason to consider confronting the practice of bundled rebates in the pharmaceutical industry through an alternate forum. Simply put, by placing its focus on possible bundling in a merger-enforcement case, the FTC is ignoring areas where possible antitrust victories could be the most valuable: existing products where rebates could hamper biosimilar competition.
Competition between branded drugs—as would occur when a competitor arises to Horizon’s drugs—is important, but it does not tend to lead to much price competition (although such entry could still increase consumer welfare through increased quality or variety). In contrast, in situations where bundled rebates prevent the entry of biosimilars, the potential for price reductions are much larger. When successful, biosimilar uptake can erode prices by 50%, which while can bring enormous benefits to patients. In focusing on acquisitions by large pharmaceutical firms to expand their pipeline of innovative medicines, the FTC is leaving large consumer benefits on the table, while simultaneously increasing the risk that their actions will chill beneficial practices.
The FTC’s suit to block the Amgen-Horizon merger represents a landmark antitrust case. The pharmaceutical industry, previously relatively immune from antitrust suits, is coming under scrutiny for the first time. That increased attention has caused worries among both early-stage venture investors, who often finance the target firms in these deals, as well as in the pharmaceutical industry itself, which relies on M&A activity to refill its pipelines with molecules for later-stage development.
For its part, the FTC adamantly insists that the practices it is challenging—the use of bundled rebates—harm competition and patients. Who is right? Unfortunately, the answer is not so clearcut. The facts of the case give both parties much to stand on.
On the one hand, despite some comments that the FTC’s theory of harm is overly novel, it is, in fact, well-grounded in economic theory. Bundled rebates, when viewed as a form of loyalty discount, can absolutely exclude competitors in an environment such as pharmaceutical development.
On the other hand, the pharmaceutical industry is also right to suggest that the M&A ecosystem has come to represent an important source of new medicines for patients and that damaging that ecosystem is likely to harm patients in the long run.
The tension between these views was not inevitable. As currently formatted, the FTC’s case generates harm by conflating two separate practices: the broad business model in which pharmaceutical firms purchase drugs synthesized by smaller companies, and the use of bundled rebates. The signal sent to society is unclear. Is it bundled rebates that should be viewed skeptically or M&A more generally?
An alternate enforcement route that relies on pursuing bundled-rebate activity under a Section 2 monopolization case would adequately balance the concerns of the FTC and of industry. Such a route could also ensure that the FTC protects consumers from the possible harms of bundled rebates, while preserving the value of the pharmaceutical M&A ecosystem to generate new treatments.