A brief look at current debates about pharmacy benefit managers


Pharmacy benefit managers (PBMs) are entities that administer prescription drug insurance benefits. Their key functions include negotiating prices with drug manufacturers and pharmacies, establishing drug formularies and pharmacy networks, and processing drug claims. PBMs are currently attracting considerable critical attention from policymakers. Multiple congressional committees have recently reported out legislation related to PBMs, and there will likely be efforts to reconcile these bills this fall.

This analysis offers a brief overview of PBMs’ role in the prescription drug marketplace and key current debates related to PBMs. An overarching message is that while there are problems in the market for PBM services, they likely have modest effects on the overall affordability of prescription drugs. Consistent with this, while some PBM reforms currently being considered are worthwhile, achieving large reductions in prescription drug costs will require approaches that look beyond PBMs per se. In particular:

  1. Much recent discussion related to PBMs has focused on PBMs’ retention of rebates paid by drug manufacturers and use of so-called pharmacy “spread” pricing. However, restricting these specific practices, as some of the bills currently under consideration in Congress would do, is unlikely to save much money for payers since PBMs could likely extract revenue from payers in other ways. It might even backfire by weakening PBMs’ incentives to aggressively negotiate prices.
  2. The market for PBM services is highly concentrated, with three firms controlling 79% of the market, which almost certainly gives PBMs market power they can use to earn excessive profits. Greater competition could reduce these profits, and recent PBM transparency proposals may help with this, albeit only to a modest degree. Nevertheless, even eliminating all PBM profits would only reduce total drug-related spending by several percentage points. Achieving larger spending reductions would likely require more fundamental market changes such as changing intellectual property protections for drugs or changing how drug prices are regulated.
  3. Cost-sharing for prescription drugs can cause patients to forgo necessary drugs and partially unravels the financial protection that health insurance aims to provide. But contrary to common arguments, the fact that cost-sharing is often calculated using point-of-sale prices that exclude some PBM-negotiated discounts (especially rebates) may not meaningfully increase patients’ overall cost-sharing burdens since consumers and employers can generally choose among plans with more and less generous cost-sharing designs. (This practice does erode the federal premium subsidies available in Medicare Part D and the individual market). Rather, where patients face excessive cost-sharing, the main cause is likely deeper market or regulatory failures, such as consumer difficulties in choosing insurance plans and adverse selection. Addressing those issues is likely to require solutions targeted at those broader problems—such as directly regulating how much cost-sharing insurance plans can impose, subsidizing more generous coverage, or improving risk adjustment systems—not solutions specific to rebates or PBMs.
What is a PBM?

A PBM is an entity that administers a health insurance plan’s prescription drug benefit. One core function of a PBM is to negotiate drug prices with manufacturers; when negotiating prices, a PBM generally offers a drug a place on the plan’s “formulary” (which specifies which drugs the plan covers and on what terms and, thus, determines how much enrollees use the drug) in exchange for paying a price below the manufacturer’s “list price.” PBMs also negotiate with pharmacies, generally offering the pharmacy a place in the plan’s network (which increases how many of the plan’s enrollees use the pharmacy) in exchange for accepting specified prices to dispense drugs. And PBMs perform administrative functions, notably processing pharmacy claims. Most of these functions (e.g., setting coverage terms, negotiating prices, establishing networks, and processing claims) parallel functions that insurers perform for non-drug benefits, but PBMs’ specialized knowledge of drug markets may allow them to perform them more effectively.

Today, PBMs have become increasingly tightly integrated with health insurers. The largest insurers (including Aetna, Cigna, Elevance, Humana, United Healthcare, and many non-profit Blue Cross Blue Shield plans) generally own PBMs or are part of companies that own PBMs. This is a relatively recent development. As of 2018, two large independent PBMs—Express Scripts (which merged with Cigna) and CVS Caremark (which merged with Aetna)—together controlled about half of the market.

This landscape shapes how PBM services are sold. In many cases, PBM services are just an integral part of the combined company’s insurance products, including fully insured insurance coverage (coverage under which the insurer is liable for enrollees’ claims spending) and third-party administrator services sold to self-insured employer plans (plans under which the employer is liable for enrollees’ claims spending). However, some PBMs also sell their services directly to self-insured employers or to insurers that lack their own PBMs (or that have PBMs that lack needed capabilities). Additionally, because self-insured plans account for around two-thirds of enrollment in employer plans, it is still frequently the case that the PBM and the ultimate payer are different entities even where PBMs and insurers are integrated.

PBMs also often operate their own pharmacies. Essentially all major PBMs operate their own mail-order pharmacies and specialty pharmacies (pharmacies that specialize in drugs that are high-cost or complex to dispense, which vary widely, but include drugs that treat some cancers, inflammatory conditions like rheumatoid arthritis, and viral infections like HIV and Hepatitis C). PBMs generally do not operate their own retail pharmacies, with CVS Caremark being a notable exception.

What are drug rebates, and what would happen if PBMs had to pass them along to payers?

When a PBM negotiates a discount off a drug manufacturer’s list price, that discount often takes the form of a “rebate,” an after-the-fact payment by the manufacturer. Rebates constituted around one-fifth of gross-of-rebate spending on prescription drugs in commercial insurance plans in 2019.

PBMs’ contracts with payers (including both insurers offering fully insured coverage and self-insured employers) sometimes allow them to retain a portion of rebates rather than passing them all through to the payer; PBMs retained an estimated  9% of rebates, on average, in 2016. This has spurred proposals to require PBMs to pass all rebates along to the ultimate payer, including a recent proposal advanced by the Senate Committee on Health, Education, Labor, and Pensions that would require passthrough of all rebates that PBMs receive under employer-sponsored plans.

Changing where rebates accrue generally only matters where the PBM and payer are distinct entities. But even there, it is unlikely to meaningfully benefit payers. While passthrough would reduce payers’ net drug spending, it would commensurately increase PBMs’ net cost of serving payers, probably leading PBMs to demand other concessions from payers (such as larger administrative fees). And precisely because payers get more value from PBMs’ services when rebates are passed through, payers would likely accede to PBMs’ demands, erasing their direct savings from lower net drug spending. In short, because this type of policy would just bar a specific contract structure—without changing the parties’ underlying bargaining positions—it is doubtful it would meaningfully change how much money changes hands.

There are a couple of caveats to the conclusion that this type of policy would have little effect on payers:

  • Effects on underlying drug spending: Changing who receives rebates could change how effectively PBMs manage underlying drug spending and, in turn, payers’ costs. Notably, passing through rebates eliminates PBMs’ incentives to prefer drugs with larger rebates to drugs with lower net-of-rebate prices when constructing formularies, which would tend to reduce net drug spending. On the other hand, passing through rebates also reduces the PBM’s incentives to negotiate aggressively with manufacturers, which would tend to increase net drug spending.

    We are unaware of any empirical evidence on which of these effects is larger in practice. However, economic theory suggests that PBMs and payers will choose to share rebates in a way that minimizes underlying drug spending since, if they did not, they could adopt a different approach and split the savings. If that logic is correct, then forcing PBMs and payers to change how they share rebates would increase spending. (Whether increasing or decreasing drug spending would ultimately be a good thing for society depends on other considerations, notably whether—and how much—increases in manufacturers’ expected revenues spur valuable innovation.)

  • Effects of reducing variation in contract terms: Requiring rebates to be passed through to payers would also eliminate one way in which PBM-payer contracts can vary. This could, in principle, make it easier for payers to compare contracts and thereby help payers negotiate better terms. We examine these effects later when we discuss proposals to increase the transparency of PBM-payer contracting; our general conclusion is that these effects are likely modest, albeit not zero.

An additional caveat is that even if requiring rebates to be passed through did not benefit payers, interactions with medical loss ratio (MLR) requirements might reduce premiums for the customers of payers that sell fully insured plans. (MLR requirements compel insurers to spend a minimum share of premium revenue on claims or face some type of penalty; they apply to individual and group market plans, Medicare Part D and Medicare Advantage plans, and many Medicaid managed care plans.) In particular, the discussion above suggests that requiring PBMs to pass through rebates might reduce insurers’ net claims spending while increasing the administrative fees they pay to PBMs, which would tend to reduce insurers’ MLRs. If this made MLR requirements bind, insurers might reduce premiums in response. Importantly, however, this outcome is far from certain. It would not occur if the MLR requirement did not bind, either because the insurer started out far above the required standard or because the insurer had other options to increase its reported MLR. Insurers might also reduce their MLRs by managing claims spending less aggressively rather than reducing premiums. Additionally, if PBMs responded to a requirement to pass through rebates in some way other than increasing administrative fees (like increasing pharmacy spreads, which are discussed in the next section), then insurers’ MLRs might not increase in the first place.

In sum, it appears unlikely that requiring PBMs to pass through rebates would meaningfully reduce payers’ costs, and it could increase them if it undermined PBMs’ incentives to negotiate aggressively with manufacturers. However, interactions with MLR rules might reduce premiums under certain conditions.

What is pharmacy “spread” pricing, and what would happen if it were banned?

The prices that PBMs charge payers for pharmacy claims often differ from the prices that PBMs pay the pharmacies that fulfill those claims. The difference between these prices, which is commonly called the “spread,” is retained by the PBM. We are unaware of research that offers a systematic picture of the use of spread pricing, although some evidence exists for specific payers. Spread pricing is often thought to be more important for generic than brand drugs, something at least one payer-specific study corroborates.

Some recent legislative proposals would limit or ban “spread pricing.” Advocates may hope that this step would save money for payers by reducing the prices they pay for drugs to the lower prices that PBMs pay pharmacies. But like some arguments for requiring PBMs to pass through rebates, this argument misses something important: if PBMs were barred from retaining the “spread,” they could well claw back the money payers would save from lower spreads using other tools, like higher administrative fees.

Also like the rebate proposals, banning spread pricing could affect payers’ costs by changing how the PBM manages underlying drug spending, although it is not obvious how. On the one hand, when PBMs retain the spread, they have strong incentives to be aggressive in negotiating low prices with pharmacies; on the other hand, if the PBM charges the payer different amounts depending on where a prescription is filled, then the pharmacies where the spread is largest may not always be the pharmacies with the lowest negotiated prices, so PBMs that retain the spread may have incentives to favor higher-priced pharmacies over lower-priced ones. As above, we are unaware of empirical evidence on which effect is larger, but it is plausible that PBM-payer contracts already appropriately balance these considerations, in which case banning spread pricing would increase total spending. (Also as above, changing the treatment of spreads could also conceivably benefit payers by reducing how much PBM-payer contracts vary or affect the premiums of fully insured plans through interactions with MLR requirements.)

If banning spread pricing did increase payments to pharmacies, perspectives would differ on whether that was good or bad. While payers and consumers would generally be worse off, pharmacies would generally be better off. Whether society benefited on net would depend in part on whether the prices currently being paid to pharmacies are too high (e.g., because pharmacies wield market power) or too low (e.g., because PBMs wield market power) to ensure appropriate access to and quality of pharmacy services.

Do PBMs earn excessive profits, and, if so, what might be done about it?

While banning spread pricing or requiring PBMs to pass rebates along to payers appears unlikely to meaningfully benefit payers, this does not imply that PBMs are compensated appropriately. Indeed, the market for PBM services is concentrated, with 3 firms serving 79% of the market. This level of concentration likely gives PBMs market power that they can use to extract excessive total compensation from their customers (that is, payment beyond what would cover their costs plus a “normal” profit), whether directly through administrative fees or indirectly by retaining rebates or pharmacy spreads.

Pre-tax operating margins for the three largest PBMs averaged a bit more than 4% of their revenues in 2022. Since PBMs’ revenues encompass both the administrative fees charged to PBMs and payers’ net payments for claims, this implies that even completely eliminating PBMs’ margins would only modestly reduce payers’ drug-related costs. Achieving larger reductions would require reducing the revenue captured by pharmacies or, particularly, manufacturers. Ultimately, the amount of revenue that manufacturers capture depends principally on the extent of intellectual property protections related to drugs and whether and how the prices that drug manufacturers receive are regulated.

Nevertheless, making PBM markets more competitive could meaningfully benefit consumers, primarily by squeezing PBMs’ profits but potentially also by encouraging PBMs to develop better ways to manage drug benefits. There are a couple of ways that policymakers could try to achieve that:

  • Encourage PBMs to compete more aggressively: One strategy is to encourage existing PBMs to compete more aggressively against each other. One approach that has appeared in several recent Congressional proposals is to require PBMs to disclose more information—including on utilization, gross and net spending, cost-sharing, and formulary construction—to employers they serve. Proposals to require PBMs to pass through rebates or to ban spread pricing (which were discussed above) could also increase transparency by reducing how much PBM-payer contracts vary, albeit perhaps to a more limited degree since they would increase transparency on only one aspect of PBM-payer contracts, rather than many different aspects.

    In principle, greater transparency could help employers negotiate better terms. For example, information on utilization patterns could make it easier for an employer to estimate the costs it would incur under contracts offered by different PBMs, making it easier to “shop” across PBMs and placing greater pressure on PBMs to offer attractive terms. Alternatively, information could help employers better assess how their current PBM contracts compare to “typical” contracts and, thus, how much room there is to press their PBMs for better terms.

    As a practical matter, it is questionable how effective these types of requirements would be; they might not substantially improve employers’ understanding of their PBM contracts, and a better understanding might not significantly improve employers’ leverage. The Congressional Budget Office has estimated that one recent proposal in this vein would reduce PBM revenues by around $900 million per year and that this effect would gradually fade away over time. To put that number in context, the combined revenues of the three largest PBMs exceeded $400 billion in 2022, while their combined operating income totaled around $18 billion.

  • Reduce concentration in PBM markets: Another strategy is to take steps to make the PBM market less concentrated (or keep it from becoming more concentrated). One policy option in this vein is to take a more skeptical view of mergers between rival PBMs. This would clearly involve blocking mergers between large PBMs, similar to what the Department of Justice (DOJ) did several years ago when confronted with proposed mergers between large insurers. But given DOJ’s success in blocking those mergers and the current concentration in the PBM market, large PBMs may be unlikely to try to merge going forward. Thus, the more pressing question may be how anti-trust agencies treat PBMs’ acquisitions of smaller competitors, which might still have large effects on competition in the long run if the acquired firms would have become large over time.

    Reducing concentration in PBM markets could involve tradeoffs. While it would likely reduce PBM profits or increase competitive pressure on PBMs to develop better ways to manage drug benefits, reducing PBMs’ market share could reduce their leverage in negotiations with drug manufacturers or pharmacies, increasing the prices that PBMs negotiate. Those price increases could, in principle, offset the benefits to consumers and employers from lower PBM profits, although On the other hand, if the prices paid to drug manufacturers or pharmacies are currently too low from a social perspective, then price increases due to reduced PBM leverage could be beneficial.

What are the consequences of PBMs’ “vertical” relationships?

Because PBMs are now generally vertically integrated with insurers and pharmacies, concentration in the PBM market could also affect competition in these other markets. For example, insurers that have their own PBMs may opt to charge rivals without PBMs high prices for PBM services, making those insurers less competitive and allowing the integrated insurer to earn larger profits. Companies could also choose to bundle their PBM and insurance services, which could make it hard for insurers without PBMs (or PBMs without insurers) to attract customers, perhaps driving competitors out of the market. Similarly, PBMs could choose to steer business to their own pharmacies to drive other pharmacies out of the market.

PBM-insurer relationships could also help insurers circumvent MLR requirements. For example, an insurer could allow its co-owned PBM to retain rebates or pharmacy spreads. This would increase the insurer’s reported “claims spending”—and, thus, its ability to meet an MLR standard—without affecting the profits of the combined entity (or the PBM’s incentives about how aggressively to manage drug spending). An insurer could achieve something similar by paying high prices to a co-owned pharmacy.

On the other hand, vertical integration can also offer efficiencies. PBM-insurer integration may allow plans to better coordinate the medical and drug portions of a plan’s benefits, care management strategies, and the like; for example, it may allow them to better take account of ways in which increasing drug utilization could reduce downstream medical costs. For fully insured plans, PBM-insurer integration also fully aligns incentives between the PBM and the ultimate payer; in particular, it eliminates the tradeoffs related to how to share rebates or spreads that were discussed earlier in this analysis. Similarly, PBM-pharmacy integration could simplify payment, utilization management, and related activities. Both types of integration can also mitigate “double marginalization” problems that arise when firms at multiple points in the production chain extract profit margins, which can increase prices and inefficiently reduce output.

It is ultimately an empirical question how the advantages and disadvantages of vertical integration net out. Unfortunately, empirical evidence on these questions is very limited, which makes it hard to offer a confident assessment on whether these types of vertical integration benefit or harm consumers.

Does it matter that cost-sharing is often based on prices that exclude some PBM-negotiated discounts, and how would changing that affect patients?

A common criticism of prescription drug benefit designs is that they require too much patient cost-sharing. This criticism often has merit. While the ostensible goal of cost-sharing is to encourage patients to forgo low-value care, cost-sharing can also cause patients to forgo high-value care. Cost-sharing also increases costs for enrollees who need more care, partially unraveling the financial protection that insurance is supposed to provide. Insulin has been a vivid example of where cost-sharing can go wrong; overuse of insulin is unlikely, but historically many patients faced substantial cost-sharing, which can discourage appropriate use while imposing large costs on people with diabetes.

In debates related to PBMs, it is sometimes argued that one reason patients face high cost-sharing is that cost-sharing is often computed using point-of-sale prices that do not reflect all PBM-negotiated discounts (especially rebates). This has spurred interest in requiring plans to base cost-sharing on net prices.

Basing cost-sharing on net prices would indeed reduce patient cost-sharing burdens (and correspondingly increase payer liabilities and, thus, premiums) if plan designs remained fixed. But it is quite plausible that plan designs would change in ways that would largely offset these effects, at least in the long run.

In particular, consumers (or the employers that choose plans on their behalf) can generally choose among plans with many different cost-sharing parameters (e.g., deductibles and coinsurance rates). Requiring cost-sharing to be computed based on net prices does reduce the cost-sharing burden under a plan with a fixed cost-sharing design, but also increases the associated premium; on balance, the combinations of premium and cost-sharing available to consumers and employers would not meaningfully change. Thus, consumers and employers could—and economic theory suggests would—respond by changing what types of plans they choose to restore the mixture of premiums and cost-sharing they had before.

There are some important caveats to this broad conclusion:

  • Due to the presence of various frictions and imperfections in consumer decision-making, it is probable that premiums and cost-sharing with this type of policy would not exactly mirror those without it, although it is not immediately clear whether these factors would tend to cause cost-sharing to be higher or lower than it is under the status quo. It is particularly likely that some specific enrollees would end up experiencing higher or lower cost-sharing than under the status quo. For example, plans often specify a single coinsurance rate for broad categories of drugs. To the extent that the gap between point-of-sale prices and net prices varies across drugs (which is clearly the case in practice), then shifts in enrollment toward plans that have higher coinsurance rates in response to this policy might offset any reduction in cost-sharing on average but result in less cost-sharing for highly rebated drugs and more cost-sharing for other drugs.
  • In Medicare Part D and the individual market, basing cost-sharing on net prices would have the effect of increasing federal premium subsidies. In both markets, subsidies are based on the premium of a benchmark plan that has fixed cost-sharing parameters (Medicare Part D) or that covers a specified share of point-of-sale claims spending (the individual market), so basing cost-sharing on net prices would increase premiums for benchmark coverage and, in turn, subsidies. This is the main reason why the Trump-era rebate rule would have increased federal spending. Thus, in these settings, basing cost-sharing on net prices would benefit consumers on net, although they could take those benefits as either lower cost-sharing or lower premiums.

    Additionally, in these markets (and the small group market), basing cost-sharing on net prices would have the effect of reducing the maximum amount of cost-sharing a plan can impose. In Medicare Part D, for example, plans cannot impose more cost-sharing than the benchmark plan, and in the individual and small group markets, plans generally must cover at least 60 percent of point-of-sale claims spending. Thus, in these markets, this type of policy could reduce cost-sharing by, in effect, prohibiting some combinations of low premiums and high cost-sharing.

  • Basing cost-sharing on net prices could also put upward pressure on net prices, which would tend to increase premiums, cost-sharing, or both. Notably, this approach could preclude certain types of rebate arrangements (like arrangements based on a drug’s realized volume) and thereby reduce the depth of the discounts manufacturers are willing to offer. Indeed, this a reason that the Congressional Budget Office concluded that the Trump-era rebate rule, which required all manufacturer price concessions to be reflected in point-of-sale prices in Medicare Part D, would have increased net prices. Another potential concern is that this approach could implicitly disclose negotiated rebates, which in some cases might reduce the rebates manufacturers are willing to offer. These concerns could be partially mitigated by basing cost-sharing on the net prices paid in past years or averaging net prices within a therapeutic class.

These caveats notwithstanding, the discussion above implies that understanding why patients sometimes face excessive cost-sharing requires looking beyond the gap between point-of-sale and net prices and considering broader insurance market failures. One factor may be difficulties consumers often have in choosing among insurance plans. Notably, consumers sometimes (though not always) overweight premiums relative to cost-sharing. This tendency may drive enrollment toward plans with higher cost-sharing and lower premiums in markets where consumers select plans directly (e.g., the individual market and Medicare) and shape the plans that employers decide to offer to their employees.

Another factor may be adverse selection, the tendency of people who need more care to select more generous plans. In markets where consumers select plans directly, this can cause more generous plans to carry excessive premiums or be entirely unavailable (if the risk adjustment systems that operate in those markets do not fully offset the additional costs that high-risk enrollees bring). Similar dynamics can play out in the group market if employers wish to avoid high-cost enrollees or, perhaps more likely, simply prioritize keeping premiums low for the many employees with lesser health care needs.

These types of broader insurance market failures may shape aspects of plan design beyond cost-sharing. Adverse selection could, for example, lead plans to implement overly stringent prior authorization requirements or adopt overly narrow pharmacy networks. (Of course, choices over these aspects of plan design are likely shaped by other considerations as well. While prior authorization requirements create hassles for patients and providers, they can deter inappropriate utilization. Similarly, narrowing pharmacy networks may strengthen PBMs’ bargaining position in negotiations with pharmacies; as discussed above, network strategy may also be shaped by PBMs’ direct presence in the pharmacy market.)

Addressing these types of problems requires solutions that have little to do with rebates or PBMs per se. One approach would be placing tighter minimum requirements on the level of coverage plans must offer, as Congress recently did with respect to insulin in Medicare Part D or creating (or expanding) subsidies for enrollees who select more generous forms of coverage. Another would be improving the risk adjustment systems that operate in the individual, small group, and Medicare markets.

  • Acknowledgements and disclosures

    We thank Marta Wosińska and Benedic Ippolito for helpful comments. Vani Agarwal and Caitlin Rowley provided excellent research and editorial assistance, respectively. We gratefully acknowledge financial support from Arnold Ventures.

    The Brookings Institution is financed through the support of a diverse array of foundations, corporations, governments, individuals, as well as an endowment. A list of donors can be found in our annual reports published online here. The findings, interpretations, and conclusions in this report are solely those of its author(s) and are not influenced by any donation.

  • Footnotes
    1. Unlike the rest of the effects discussed in this section, this effect of requiring passthrough could arise even when the insurer and PBM are the same entity since it could prevent the insurer from shifting profits into its own PBM.
    2. These estimates were derived from the annual reports of the relevant parent companies: CVS Health (for which we used results reported for its pharmacy services business); Cigna (for which we used results reported for its Evernorth business); and United Healthcare (for which we used results reported for its OptumRx business). In all cases, the reported results also encompass the firm’s mail-order and specialty pharmacies (but not, in the case of CVS Health, its retail pharmacies); for Cigna, the results also include a small amount of revenue for non-pharmacy businesses.
    3. Greater transparency could also help employers police PBM conduct. For example, it could help employers detect instances where a PBM that retains some rebates has elected to give an attractive formulary placement to a highly rebated drug. Interestingly, this could make it more attractive for employers to allow PBMs to retain rebates.
    4. These estimates were also gleaned from the same parent company annual reports described earlier.
    5. Concerns about “raising rivals’ costs” are most relevant to fully insured markets (where PBM and insurance services are sold together), while bundling concerns are more relevant to self-insured markets (where they may not be).
    6. There is also a potential “raising rivals’ costs” concern with respect to pharmacies; namely, PBMs that own pharmacies could set a high price for those services. In practice, this may principally be a concern with respect to CVS Health since major PBMs now generally use their own mail-order and specialty pharmacies. 
    7. Some proposals reported out by the Congressional committees working on PBM legislation would take a much more limited step in this direction by capping cost-sharing at the net price of a drug.