Uncertainty and Motive in Pharmacy Dispensing Mark-Ups
Richard Evans / Scott Hinds
203.901.1631 / 203.901.1632
email@example.com / firstname.lastname@example.org
November 10, 2010
Uncertainty and Motive in Pharmacy Dispensing Mark-Ups
- Recently we argued incremental (relative to brands) generic dispensing premia of roughly $5 / script ultimately will fall. This note addresses healthy criticisms of that note, particularly: 1) payors already know generic acquisition costs; and, 2) the generic dispensing premium is too small to matter
- We argue that payors know average, but not drug-specific (much less dose– and dosage–form specific) acquisition costs, and that drug-specific knowledge is pre-requisite to efficient purchasing
- The difference between the pharmacy re-imbursement benchmark of average wholesale price (AWP) and actual generic acquisition costs is wide-ranging, and the distribution is near-uniform around the mean. Thus pharmacy dispensing margins on any given drug are equally likely to be well above, or well below, the mean dispensing margin
- If payors set AWP-X such that average generic mark-ups equal average acquisition cost plus a competitively efficient service margin, then pharmacies only want to fill half of the sponsors’ generic prescriptions. Thus AWP-X has to be set above a competitively efficient level
- Under a cost-plus (e.g. average manufacturer price or AMP + X) formula, pharmacy dispensing margin is a constant value, which can be set far more precisely – i.e. in an AMP + X format, pharmacies’ generic dispensing margins can be set equal to a competitively efficient margin
- Separately, clients raise the question of whether the incremental $5 per script generic dispensing margin is large enough to matter. We show that average brand rebates over time have hovered around this value, i.e. in a very real sense the PBM industry was built on $5 per brand prescription. We conclude that the additional $5 dispensing margin paid per generic script is material, particularly as generics become an ever more dominant percentage of total scripts
Last week we published an analysis of generic dispensing margins at pharmacy, our conclusion being that the (roughly $5, Exhibit 1) incremental margin realized by (retail and mail) pharmacies on generics relative to brands ultimately will fall.
We argued that the incremental premium is due to either or a combination of two factors: uncertainty (i.e., payors would reduce the premium but cannot see pharmacies’ generic acquisition costs clearly enough to do so), and/or an informed choice on the part of payors to bias pharmacies in favor of generics, in order to optimize the displacement of brands. And, we argued that the generic dispensing premium eventually will fall, as the result of increased payor visibility into generic acquisition costs, less need to bias pharmacy toward generics as consumers’ out-of-pocket (patent-expired) brand costs rise, and/or simple competition (gains in dispensing share accrued by pharmacies willing to forego larger generic premia).
We’ve heard considerable healthy criticism on these arguments, and recognize two consistent themes in this feedback: 1) that large payors know generic acquisition costs; and 2) that the incremental generic dispensing premium is insufficient to warrant attention, i.e. that the generic dispensing premium is too small to make its reduction / elimination worthwhile.
Clarifying Our Uncertainty Argument
We accept that large payors know average acquisition costs for generics, but we assert that they do not know drug-specific (i.e. NDC by NDC) acquisition costs; and, we assert that it is necessary to know drug-specific acquisition costs in order to exercise purchasing power effectively.
Pricing in pharmacy contracts is generally benchmarked to average wholesale price (AWP). AWP is simply wholesale acquisition cost (WAC) times a conventional multiplier (1.2), i.e. AWP = 1.2 x WAC. As is well known, WAC does not reflect pharmacies’ acquisition costs for generic drugs, as generic manufacturers commonly provide price concessions that are not captured by WAC (or, by extension, AWP). The relationship between WAC and pharmacies’ actual acquisition costs (for which average manufacturer price or AMP is a reasonable proxy) for generics occupies a very broad range (Exhibit 2). And, from summaries that we’ve seen of more recent pharmacy data, we believe the distribution of generic acquisition costs relative to WAC (and thus also AWP) is near-uniform (i.e., there is no evidence that generic acquisition costs are concentrated close to the average; thus a particular drug is as likely to have a very high or very low generic acquisition cost relative to WAC as it is to be close to the average).
Thus when a plan sponsor makes a decision on where to fix ‘X’, that sponsor has to recognize that fixing ‘X’ too low would mean that pharmacies would lose money on some portion of the generics dispensed to the sponsor’s beneficiaries. Accordingly, the sponsor has to fix ‘X’ at a level that allows pharmacies to earn a profit on most generics. If ‘X’ were set such that AWP – X was equivalent to average pharmacy acquisition costs plus a competitive service margin, then pharmacies would lose money on (and presumably would choose not to fill) half of the sponsors’ generic prescriptions. As a result, AWP – X% must be set above the competitively efficient service margin.
A simple analogy might be flying over a mountain range in heavy clouds, without (other than an altimeter) instruments … knowing the average height of the range’s peaks is useless. In this extreme case, we simply want to know the elevation of the highest peak in the range, and will fly (at least) at this elevation plus a margin of error.
If, on the other hand, reliable sources of generic acquisition costs are available, then payors can and should specify pharmacy re-imbursement as a function of acquisition costs … e.g. AMP plus a margin. In this case, the amount paid to pharmacy always varies with the true underlying acquisition cost, meaning the payor can precisely set the level of mark-up pharmacies receive – regardless of the dispersion in those acquisition costs. In this cost-plus pricing format, pharmacy dispensing margins can be set at a competitively efficient level. In flight navigating terms, paying acquisition costs plus a margin is the equivalent of terrain following radar.
Thus to our minds it’s not a matter of whether greater transparency into generic acquisition costs can reduce generic dispensing mark-ups, it’s instead simply a matter of whether, and when, sufficient acquisition cost details become available. As we argued last week, this detail might come about suddenly if the Secretary of HHS makes AMP data publicly available; or, more gradually, as states’ publicly available surveys of actual acquisition costs (e.g. Alabama) grow to sufficient sample sizes.
Does $5 per Script Really Matter?
Yes. It’s not hyperbole to argue that the PBM industry was built on $5 per (brand) prescription. FTC’s 2005 PBM report (on ’02 and ’03 data) showed average brand manufacturer payments to PBMs of $5.78 (Exhibit 3); Medco’s ’02 through ’09 SEC filings disclose average rebates growing from $3.46 to $5.98 (Exhibit 4); and, a survey of employer-sponsors of PBM benefits estimates average brand rebates at retail of roughly $3 to $8 (Exhibit 5).
We acknowledge the non-trivial administrative efficiencies PBMs bring to sponsors, but to our minds the rebate and fees-driven +/- $5 in savings per prescription is very clearly the main reason sponsors employ PBMs. We find the popular argument that PBMs manage utilization far from compelling; the change in utilization for PBM beneficiaries is not meaningfully different (in fact PBM trend is a bit higher) than the national average (Exhibit 6) – thus our bias that PBM clients are (and have been) on-board largely for the purpose of accessing PBMs’ roughly $5 / script purchasing leverage.
It follows that $5 per script plainly matters; and, as generic prescriptions increasingly outnumber brands (Exhibit 7), it follows that reducing generic dispensing margins becomes a feasible, even necessary, means of providing drug benefits at competitive monthly premium levels.
As much as we’d love to time the point at which generic dispensing margins fall – and thus also PBM and retail gross margins – we simply cannot, the primary source of uncertainty being whether and when generic acquisition costs come into plain view. What we can say, with some conviction, is that the generic dispensing premium is inherently unstable – if it doesn’t fall as a result of greater clarity, ultimately it falls either for lack of need, or because of competition. Our conviction that generic dispensing premia are unstable frames our investment thesis on the drug trades, where we sense an almost permanent / structural reliance on generic dispensing premia as a source of gross margin expansion. Accordingly we believe out-year estimates are simply too high for PBMs (MHS, ESRX, CVS), drug retail (WAG, RAD, and again CVS), and drug wholesale (CVS, ABC, MCK).
- As is required by January 2011 per the language of the Patient Protection and Affordable Care Act (PPACA), though it’s unclear whether the Secretary can or will following the PPACA language so quickly
- These are total payments, including administrative fees as well as rebates
- These are self-reported employer data, and are likely to be skewed by both limited sample size and varied definitions of the rebate calculation among respondents