Why Generic Dispensing Margins (Eventually) Must Fall
October 28, 2010
Why Generic Dispensing Margins (Eventually) Must Fall
- For both PBM mail-order ad retail we estimate that per-Rx generic dispensing margins are roughly $5 higher for generics than brands, and show that this premium results from: 1) payors’ inability to see pharmacies’ generic acquisition costs (uncertainty premium), and 2) payors’ need to ‘bias’ pharmacies in favor of generic dispensing (incentive premium)
- The uncertainty premium fades as payors’ line-of-sight into generic acquisition costs improves: suddenly if HHS publishes acquisition cost data (average manufacturer price or AMP) in January 2011; or if not, then more gradually as cash-strapped states shift Medicaid Rx re-imbursement closer to generic acquisition costs. Alabama is making this shift (and is making generic acquisition costs public), other states likely will follow
- Commercial payors’ need to have pharmacies earn more on generics than brands (i.e. the incentive premium) fades as consumers’ out-of-pocket exposure to brand v. generic differentials grows, and this differential is growing exponentially
- Quite apart from these separate pressures on the uncertainty and incentive premiums, the generic dispensing premium is in its entirety under competitive attack. Wal-Mart / Humana’s Medicare Part-D offering offers a monthly premium that is less than one-half the national average; much of this pricing advantage ties to Wal-Mart’s willingness to forego out-sized generic dispensing premia
- This competitive assault should accelerate. As generic represent an ever larger percentage of prescriptions dispensed, generic dispensing premia grow in importance as an absolute cost of delivering a drug benefit.
- Bullet 2
- Bullet 3
It is generally well known that drug trade (retail, wholesale, PBM) absolute per-prescription margins are larger for generics than for brands; less well understood is why this is the case, and whether this can last.
As generally happens across healthcare, in pharmacy commercial payors tend to adopt the re-imbursement conventions and pricing benchmarks used by government payors. For most retail pharmacy settings, Medicaid conventions are the dominant federal pricing form, and these standards plainly result in higher dispensing margins for generics than for brands.
What Federal Benchmarks Mean to Medicaid Dispensing Margins
Medicaid re-imburses pharmacies at a discount to average wholesale price (AWP), plus a dispensing fee. AWP is set as a function (1.2x) of wholesale acquisition cost (WAC) which, despite its name, does not reflect pharmacy’s true acquisition costs for generic drugs. Generic manufacturers are freely able to provide various discounts and credits to retailers without altering their WAC (and thus either AWP or AWP-based re-imbursement to pharmacy). And, unlike brand manufacturers (who have little or no reason to discount to the trades, as the trades generally cannot afford not to stock a brand) generic manufacturers have every reason to give discounts to the trade, as the trade’s decision of whether to stock a given manufacturer’s generic is all-important. As a result, the spread between AWP-based pharmacy re-imbursement and pharmacy’s actual acquisition costs for generics is substantially larger for generics than for brands (Exhibit 1).
However the longer a given molecule has been generic, the greater the tendency for the spread between Medicaid re-imbursement and pharmacy’s acquisition costs to narrow, as Medicaid administrators establish maximum re-imbursement levels for that molecule. Federal administrators currently set federal upper limits (FULs) as a function of the lowest listed AWP once two or more therapeutically equivalent versions of a molecule are FDA approved; state administrators commonly set less generous limits in the form of maximum allowable costs (MAC’s) that apply to a broader range of drugs than federal FUL’s. Unlike federal administrators who are tethered to AWP, state administrators are freer to set MAC levels based on what they believe real acquisition costs are for pharmacies in their states. Despite this freedom establishing state MAC lists is an imperfect and time-consuming process of never-ending approximations, particularly given the fact that state Medicaid officials with few exceptions cannot readily see pharmacy’s actual acquisition costs.
Thus Medicaid payments to pharmacy for generics can be gamed, and in fact beg gaming. Pharmacies have no motive to accept less payment than is offered for Medicaid dispensing: 1) patient co-pays are small to non-existent, so patients are indifferent to the underlying retail spreads; and, 2) pharmacies cannot expect state Medicaid agencies to shift more volume their way in exchange for a lower spread. More simply, as a fixed price system Medicaid re-imbursement to pharmacy precludes competition among pharmacies, so the greater spreads offered on generics relative to brands are not competed away.
The result is obvious and predictable: Medicaid re-imbursements to pharmacy tend to be larger for newer generics than for older ones. This is in part because newer generics tend to be available from fewer manufacturers and thus also tend to carry higher acquisition costs, but the far more important effect on total price is the tendency of pharmacy to earn dramatically larger spreads on newer generics than on either older generics, or even brands (Exhibit 2). Until Medicaid payments to pharmacy are structured and/or managed differently (more about prospects for this shortly), there’s no reason to expect any change to the trades’ relative profitability of selling generics or brands to Medicaid beneficiaries.
What Federal Benchmarks mean to Commercial Dispensing Margins
Unlike Medicaid, commercial payors are not restricted by federal benchmarks and conventions, but adopt the benchmarks as reference points for obvious reasons of convenience. Commercial payors represent 84 percent of payments to US retail pharmacy, as compared to 8 percent each for Medicaid and cash, meaning commercial payors’ actions ultimately are far more relevant to pharmacies’ dispensing margins.
Even though commercial payors are free to shake themselves loose from the artifacts embedded in federal re-imbursement for generics, they haven’t — we see the same general pattern of relative generic v. brand profitability for commercial payors as we see for Medicaid (Exhibit 3).
This begs the question of why commercial payors pay (mail and retail) pharmacies more to dispense generics than brands. Framing things a bit more broadly we can think of four reasons retail spreads to commercial payors would be different for generics and brands:
1) inventory costs for generics are lower than for brands (if we assume sufficient competitive pressure on pharmacies this would tend to lower generic spreads);
2) the trades’ negotiating leverage over generic manufacturers is greater than for brands (this is certainly true, but would only raise pharmacies’ spreads if we assume limited competitive pressures on pharmacies);
3) competitive pressures on pharmacies may be limited by payors’ inability to see, or even reasonably estimate, pharmacies’ acquisition costs for generics; and,
4) commercial payors, even if able to clearly see pharmacies’ generic acquisition costs, might choose to offer somewhat greater dispensing spreads on generics in order to bias the channel away from brands.
Using the spreads pharmacies earn on brand dispensing, we can control for the effect of competitive pressures from commercial payors (3). Since commercial payors have very clear line-of-sight into pharmacies’ acquisition costs for brands, and because the cost (net of differential inventory costs) for dispensing a brand or generic are effectively the same, we can reasonably assume that pharmacies’ (inventory cost adjusted) brand dispensing margins reflect a competitive equilibrium – i.e. brand dispensing spreads are an accurate / competitively driven pricing of pharmacies’ services, which reflects the balance of power between commercial payors and (mail and retail) pharmacies.
We compared pharmacy spreads on (single-source) brands and generics, and adjusted these for inventory costs (which tend to broaden the gap between generic and retail re-imbursement, but by trivial amounts even at inventory turn-over and WACC assumptions well above the ones we used). In sales to commercial payors, and after adjusting for carrying costs, retail pharmacies are retaining roughly $5.07 more per generic prescription dispensed than for brands (Exhibit 4). PBM’s, when dispensing through owned-mail facilities, retain roughly $16.25 more per generic prescription than per brand prescription (for ‘normal’ mail prescriptions of roughly 90 days’ supply); or roughly $5.42 more per generic prescription than per brand if we normalize mail prescription sizes to retail levels (x/3).
Referring back to our reasons brand and generic spreads might differ, and having controlled for both carrying costs and competitive pressures on pharmacy margins, this $5 excess margin should in theory be the sum of two values: 1) the economic cost to commercial payors of being unable to see, or at least to closely estimate, pharmacies’ generic acquisition costs; and, 2) the premium commercial payors offer to ensure that pharmacies prefer to dispense generics instead of brands.
Deciding which of these two factors (uncertainty v. incentive) is larger is inherently difficult and subjective, and ultimately may be un-necessary. More critical is whether these factors fade (as we believe they eventually will — more about this in a moment). The uncertainty surrounding generic acquisition costs is considerable, which is at least consistent with the argument that the economic cost of this uncertainty is high: generic prices are fast moving, particularly for newer generics; and, the relationship between actual prices of generics and published benchmarks occupies a very broad range (Exhibit 5). Also, we struggle to believe that pharmacies would refuse to fill with generics instead of brands for something less than a $5 incremental profit per script. The average pharmacy fills nearly 45,000 generic prescriptions for commercial payors each year; for argument’s sake one might assume that a $2 spread (average $90k / year) would be sufficient to warrant an average pharmacy having a stable and methodical preference for generics over brands.
On the other hand, we plainly see the utility of having pharmacies biased in favor of dispensing generics, and so also believe that some portion of the generic dispensing premium is an intentional inducement to pharmacies to prefer generics. Commercial payors (unlike Medicaid) cannot consistently demand that generics be dispensed when available; accordingly, because generics are far less costly to payors than brands (Exhibit 3, line 2), it makes obvious sense for payors to ‘tilt’ pharmacies in favor of generics.
So, we conclude that the economic cost of uncertainty, and the motive for payors to give pharmacies more profit on generic dispensing, both are large effects, though we have no clear way to determine which is larger.
Where Things Go From Here – Why Generic Dispensing Margins (Eventually) Must Fall
The more important consideration is whether uncertainty and incentive effects – and thus generic dispensing margins — are durable. Ultimately we believe they are not, for three reasons:
- Payors’ visibility into pharmacies’ generic acquisition costs is improving;
- brand v. generic cost differentials to consumers should soon reach the point that consumers routinely insist on generics, which lessens the need for payors to give large generic dispensing margins to pharmacies; and,
- competition – namely the emergence of large pharmacies (e.g. Wal-Mart) who are willing to forego very large generic dispensing margins.
Payors’ Visibility into Generic Acquisition Costs is Improving – Perhaps all at Once
At least two trends are improving payors’ visibility into pharmacies’ generic acquisition costs. First and (potentially) most powerfully, the Patient Protection and Affordable Care Act (PPACA) calls for the Secretary of HHS to publish average manufacturer price (AMP) information for multi-source and generic drugs beginning in January of 2011. Whether this in fact happens is far from certain; the Deficit Reduction Act of 2007 called for similar changes, but HHS’ ability to make AMP public, much less implement an AMP-based re-imbursement formula, was enjoined by the courts as a result of legal challenges by the trades, particularly retail pharmacy. The trades again are saber-rattling with respect to the disclosure of AMP, and it’s simply unclear whether or when HHS will make AMP data available. Second, states have the ability to shift their Medicaid payment benchmarks to surveys of average acquisition cost (AAC), as Alabama recently began doing, and other states appear likely to follow. It is not entirely clear how accurate Alabama’s AAC data will be, as it’s not clear how aggressive or systematic the state will be with respect to various credits and charge-backs that reduce acquisition costs relative to wholesaler list prices, or even wholesale invoices. Alabama’s AAC data will be publicly available, and on the premise that other states follow, we suspect that the economic realities faced by the states ultimately will lead to AAC values that are reasonably accurate – and in plain view of everyone, including commercial payors.
Even if we assume that the HHS Secretary publishes AMP this January or soon thereafter and/or that the states follow Alabama and generally ‘expose’ true generic acquisition costs, there’s still a considerable hurdle—namely PBMs– between public knowledge of generic acquisition costs and a collapse of that portion of the generic dispensing premium that we attribute to payors’ uncertainties over acquisition costs. PBMs profit from larger generic dispensing spreads, and so have every reason to defend those spreads; and, PBMs are payors’ (e.g. employers and HMO’s w/ pharmacy carve-outs) primary ‘agents’ in the commercial relationship with pharmacy. Simple visibility of generic acquisition costs will not be enough; either state Medicaid programs must broadly move towards re-imbursement formulas based on AMP (which presumably leads to the death of AWP, and so forces PBMs to adopt AMP in their commercial contracts); or, payors have to insist on a shift in PBMs’ commercial benchmarks from AWP to either AMP or AAC. Ultimately we believe clarity comes about, but caution that unless the HHS Secretary actually shifts Medicaid reimbursement to an AMP basis (thus ‘obsoleting’ AWP and forcibly changing PBMs’ pricing benchmarks in commercial contracts), eroding the ‘uncertainty’ premium in generic dispensing will take time.
Consumers’ Brand v. Generic Out-of-Pocket Cost Differential is Growing Exponentially
We believe that as consumers move to more restrictive plan designs in a bid to lower premiums, that the cost to the consumer of accepting a brand when a generic is available becomes so great that consumers will rarely accept the brand — which lessens the need for payors to give pharmacies a large spread on generics relative to brands. At present, benefit designs are such that consumers cannot be expected to quickly and reliably move to generics as they become available, thus payors need pharmacies to prefer generics over brands, which is to say that payors need pharmacies to sell consumers on the idea of taking the generic instead of the brand—and for the moment giving pharmacies far larger dispensing margins on generics is a perfectly reasonable way to make this happen. However as brand v. generic cost differentials to the consumer grow–-particularly as brands with generic alternatives increasingly are not covered, we expect consumer preference for generics to reach a point at which it is no longer necessary for payors to provide large generic dispensing premia to pharmacies. Exhibit 6 shows the growing real difference between average generic and average non-preferred brand co-pays over time, as well as the growing real difference between average generic co-pays and average brand retail prices over time. Consumers in plans that cover patent-expired brands as non-preferred are exposed to the lesser differential; consumers in plans that do not cover patent-expired brands are exposed to the larger differential. We’re not immediately able to calculate a weighted average of these two lines, but we do know that consumers are increasingly moving to more restrictive plans that will not cover patent-expired brands, and so believe that the weighted average brand v. generic cost differential faced by consumers is growing exponentially, thus rapidly reducing the need for payors to offer out-sized dispensing margins on generics.
And Finally, Old Faithful (Competition)
The final threat to generic dispensing margins is the simplest, and the most inevitable – competition. In 2006 Wal-Mart famously shook up US pharmacies with its $4 generic program, which meaningfully under-cut average cash prices for generics dispensed at both independent and chain pharmacies. The impact of the $4 program was largely limited to cash-paying (i.e. un-insured) customers, as most customers with insurance would have generic co-pays that were low enough to limit any incentive for that consumer to move their prescriptions to Wal-Mart. On October 1 of this year, Wal-Mart effectively moved into the game of dispensing prescriptions for insured consumers. Wal-Mart and Humana are offering a prescription drug plan (PDP) under Medicare Part-D for $14.80 per month, nationwide, which is less than one-half the average premium charged for PDP’s. Presumably, Wal-Mart / Humana are able to offer a lower premium in large part because Wal-Mart has chosen to forego the very large generic dispensing premia that remain a feature of more traditional PDP plans that rely on independent pharmacy, chain pharmacy, and-or PBM mail-order dispensing.
We believe this eventually must lead to lower generic dispensing premia in the traditional (independent / chain / PBM) channels. Recognize that as generics’ share of total prescriptions dispensed rises (Exhibit 7), generic dispensing premia are multiplied across a broader base of total prescriptions, making the absolute cost burden of paying these premia much larger. And, as patent-protected—and rebate-paying—brands disappear (Exhibit XX), the competition for who can offer the best pharmacy benefit for the least amount of money shifts from the old game of who can get the largest brand rebates, to the new game of who can offer the lowest dispensing cost.
Bear in mind also that as we approach the launch of health insurance exchanges in 2014, and as employers likely shift employees onto the exchanges rather than continuing to provide health benefits directly, that more and more health policies (and as a carve-out, prescription drug policies) will be purchased by cost-sensitive individuals looking for the lowest possible premium. In effect, when the exchanges begin operating, the market for working age drug benefits begins to look more like the market for Medicare Part-D PDP’s – i.e. the market where generic dispensing premia are plainly collapsing.
We’re convinced that generic dispensing premia eventually must fall, though timing is – very unfortunately — unclear. Publication of AMP in January by HHS would meaningfully reduce uncertainty; incorporation of AMP into federal pricing benchmarks would all but eliminate uncertainty. It is not at all clear what HHS plans to do, thus despite the clarity of PPACA language calling for both AMP publication and use of AMP as a federal pricing benchmark (in January), we still cannot rule-out that legal challenges will slow the process, as happened with the implementation of AMP-based changes in the wake of DRA. Even if HHS takes no action, it’s increasingly clear that states (e.g. Alabama) will shift to more accurate acquisition-related Medicaid pricing benchmarks on their own. The fact that Alabama has gotten the ball rolling, and that virtually every state is 1) under intense fiscal pressure and 2) sees Medicaid spending as one of the largest single budget line-items, suggests that commercial payors will have increasingly clear line-of-sight into generic acquisition costs within a fairly short (e.g. 1 to 2 year) period, if not sooner.
The need for commercial payors to offer large generic dispensing premia is falling rapidly; as consumers increasingly choose restrictive plan designs, more and more consumers are falling under plans that simply do not cover brands that have generic equivalents. As a proxy for the general trend to more restrictive plan designs, enrollment in high deductible health plans (HDHP’s) grew from 8 percent of employer-sponsored insurance (ESI) in 2009 to 13 percent in 2010, and we expect commensurate enrollment gains going forward. We would expect the opening of the health insurance exchanges (HIE’s) as early as 2014 to increase enrollment in restrictive plan designs, as we expect that many employees with little current choice of what coverage to buy will wind up on the HIE’s, and once on the HIE’s will have a much broader set of choices – and as a result increasingly will choose less expensive (and more restrictive) plans.
Finally, as generics represent an ever larger share of total prescriptions, generic dispensing premia fall under increasingly intense competitive pressure, since lowering these mark-ups becomes the most potent means of reducing overall drug benefit cost. Wal-Mart / Humana have set this ball rolling in Part-D, and there’s every reason to believe the same process eventually unfolds in drug benefits for those of working age, especially as an increasing proportion of working age individuals and families buy health coverage on HIE’s.
Thus all in, we would expect relative (to brands) generic dispensing premia to be all but eliminated by 2015 – 16 at the latest, since by then we will plainly have greater visibility into generic acquisition cost, greater consumer cost sensitivity (and choice) with respect to insurance premiums, and a prescription drug market in which generic dispensing premia are far more important than brand discounts / rebates.
This obviously creates earnings pressure for retailers (WAG, CVS, RAD), PBMs (MHS, ESRX, CVS), and drug wholesalers (ABC, CAH, MCK). Retail pressures are straightforward, as are PBM (mail-order) pressures. PBM mail-order volumes and share of mail-order prescriptions filled as generic vary; using 2009 full-year data, we show that each PBM’s revenue and gross profit exposure in 2009, if generic premiums over brands had not been in place, would have ranged from roughly $390M (ESRX) to roughly $960M (MHS, Exhibit 8). Given the higher proportion of national prescriptions filled at retail, and the higher proportion of retail prescriptions that are generic, retailers’ revenue / gross profit exposure is far greater, ranging from an estimated $835M (RAD) to an estimated $1.9B (WAG, also Exhibit 8).
Finally wholesalers – we recognize that wholesalers, like PBMs and drug retail, directly capture a higher per-unit margin on generics than brands, and so stand to potentially lose margin as generic mark-ups fall at retail. More important than this, we believe, is the relevance of falling generic dispensing margins to wholesalers’ customer base, namely independents. In the current market, generic mark-ups are sufficiently generous that both independents and wholesalers can share what’s available. Going forward, as this margin compresses, independents will still have to pay wholesalers, but chains and mass merchants (who already buy directly from generic manufacturers) won’t have to share margin with a distributor. Accordingly, it’s likely that a reduction in generic dispensing margins re-ignites the erosion of independents’ share of national dispensing by chains and mass merchants, which in turn erodes wholesalers’ customer base.
Among the trades (retailers, PBMs, wholesalers) PBM consensus reflects the most prominent trend of improving gross margins as we head into the 2011-2013 generic wave; and, PBM valuations on
out-year earnings are higher than both drug retail and drug wholesale (Exhibits 9, 10).
- Perversely, federal administrators can see the actual average cost at which pharmacies acquire each generic molecule (average manufacturer price, or AMP) but being tethered to AWP cannot act on this to lower Medicaid re-imbursement to pharmacy. State administrators, who are free to set MAC’s closer to their best estimate of acquisition costs, cannot see AMP’s
- Single-source brands (SSB’s) rarely if ever offer discounts (beyond standard prompt-pay terms) to wholesalers or retailers, since neither of these two channel players can afford not to carry a single-source brand. As a result, wholesale acquisition cost (WAC) for single-source brands tends to very accurately reflect the price at which wholesalers and pharmacies acquire brands. SSB’s do offer rebates to PBM-owned mail-order pharmacies, since the PBM’s formulary can shift share to or from the brand. Even though SSB’s offer rebates to PBM-owned mail-order pharmacies, note that the majority of these rebates flow through to the plan sponsor, rather than to the PBM
- The assumption that commercial payors have leverage over pharmacy outlets seems quite reasonable – the average American consumer has 21 competing pharmacies within a reasonable distance from their home (“Consumer Access to Pharmacies in the United States; SK&A Healthcare Information Systems 2007)
- We assumed 15 days of carry and WACC of 10%
- AMP is a very reasonable reflection of what wholesalers and pharmacies pay to purchase generics, though it does ignore such items as prompt payment discounts (roughly 2%) and fees that a generic manufacturer might pay to a wholesaler or retailer to offset that entities costs associated with carrying and distributing the drug. Clearly those fees will become larger, but it’s very unlikely they’ll be able to replace all of the margin that currently exists between AWP-X% re-imbursement and actual acquisition costs
- Risks to timing of the exchanges is non-trivial, see: “What a Republican House Means for Health Reform”, Sector & Sovereign Research LLC, October 4, 2010
- See: “Why the Average Employer Will Drop Health Insurance in 2014”, Sector & Sovereign Research LLC, May 26, 2010
- To be clear, most working-age folks are likely to buy their prescription coverage as part of their larger health insurance coverage, whereas many Medicare beneficiaries buy a stand-alone PDP. This is because of MLR restrictions – HMO’s are incentivized to carve-out benefits (e.g. prescription drugs), since payments to providers are treated as a medical cost for the purpose of calculating MLR. Nevertheless, the contribution of the drug carve-out to the total insurance premium likely is sufficient to warrant HMO’s doing everything possible to lower the cost of the drug-benefit – and in a post-brand world, everything possible begins with reducing generic dispensing premia
- Patient Protection and Affordable Care Act
- Deficit Reduction Act of 2007