Drug Makers’ Profits Under Medicare Drug Benefit

How Much Would Drug Makers’ Profits Rise under a Medicare Prescription Drug Benefit? SUMMARY Introduction: In October 2003, we estimated the new profits that the proposed House and Senate Medicare bills offered to prescription drug makers. If $400 billion in new federal spending goes to a Medicare drug benefit starting in 2006, we concluded that drug manufacturers’ profits would rise by $139.2 billion over eight years. Recently, PricewaterhouseCoopers (PwC) disputed that estimate in a critique commissioned by the Pacific Research Institute. This report identifies the main problems with that critique. The PRI/PwC criticisms are mistaken. They do not undermine our analyses of any of the three main aspects of the legislation that we identified as contributing to a steep increase in drug makers' profits. Drug manufacturers’ prices would not be restrained under the new law, we concluded. The critique concurs that private plans now get discounts for employers averaging only 15 percent, but claims, without supporting evidence, that such plans will win Medicare discounts of “around 25 percent.” The volume of drugs sold would rise substantially, we predicted. The PRI/PwC critique vastly underestimates the likely rise in volume when Medicare patients gain coverage, even given this law’s relatively meager benefits. Drug makers’ new costs would be very low, we emphasized. Almost all of drug makers’ revenue from these new prescription sales at high unrestrained prices would be profit, because most of their costs are already covered. The PRI/PwC critique simply ignores the issue of drug makers’ low incremental costs. Our findings last October troubled many. The bill’s inadequate benefits and high costs reflected too much money going to drug makers’ profits, because prices were not controlled. The PRI/PwC critique says virtually nothing about the new Medicare drug benefit itself. But were the PRI/PwC critique accurate, the overwhelming share of money under the bill would apparently replace existing spending on prescription drugs, and very little would buy medications for patients who do not get needed drugs today. A. COST OF MEDICATIONS 1. Ignoring Drug Makers’ Low Incremental Costs: Profits reflect revenues less costs. Yet the PRI/PwC critique entirely ignores half the equation—drug makers’ very low real added costs to manufacture a higher volume of pills for Medicare patients. B. VOLUME OF MEDICATIONS 2. Underestimating the Number of People in Need: PRI/PwC underestimate by onethird (or more) how many Medicare beneficiaries entirely lack prescription drug benefits. They appear unaware of how limited coverage is for many who do have some benefits. ii 3. Underestimating Effect of Coverage and Lower Prices: The PRI/PwC critique uses inappropriate, irrelevant data, ignoring the differences between Medicare and nonMedicare patients. It relies entirely on a price-elasticity estimate from a 30-year-old study enrolling only younger, non-Medicare patients for different benefits in a radically different pharmaceutical climate—an inappropriate source for at least six reasons. We instead documented Medicare patients’ higher need for medications, and differences in use between Medicare patients with and without drug coverage. The drug industry itself has cited studies concluding that lower prices can boost the volume of drug sales enough to maintain the same level of revenue. Numerous observers conclude that the new law will substantially raise drug makers' revenues and profits. PRI/PwC also fails to address the likelihood that any savings from discounts would recycle to buy still more from the drug makers. 4. Extent of Coverage and Circularity of Estimates: The PRI/PwC critique appears to rest heavily on a circular argument. And with today’s huge and growing gaps in drug coverage for people on Medicare, it is simply not credible that, as they claim (without supporting evidence), 58 percent of Medicare beneficiaries’ drug costs were covered by insurance in 2003, leaving only 42 percent to be paid by patients themselves. 5. Failing to Identify the New Rx/Replacement Rx Split: The PRI/PwC critique fails to assess the likely split between new and replacement prescriptions, central to the law’s impact both on industry profits and on patients. PRI/PwC implies, though, that the law will not enable Medicare patients to fill many more prescriptions than they fill now. 6. Hiding Crucial Calculations: Many central calculations in the PRI/PwC critique are not disclosed, especially for crucial estimates of the likely rise in prescription drug use. C. PRICE OF MEDICATIONS 7. Claiming Implausible Discounts and Rebates: PRI/PwC offer only unsupported and inconsistent estimates that private plans will save Medicare more than they now save employers. Yet the law fragments Medicare’s buying power—as drug makers urged. Perhaps the savings claims rest on little-publicized assumptions that plans will use very restrictive formularies, which would likely be unpopular. D. PROFITS 8. Claiming Implausible Overall Effects: The PRI/PwC critique claims that the $400 billion in new Medicare drug spending (equal to about one-fourth of CBO’s projected baseline drug spending for Medicare beneficiaries) could yield a drop in drug makers’ revenues, or at most a very small rise. The claim that drug makers might lose money is attributed only to anonymous industry sources. We describe how PRI/PwC make the $400 billion disappear and fail to follow the trail of their own figures. iii 9. Refusing to Estimate Profits: The PRI/PwC critique fails to address the probable size of drug makers’ new profits, focusing instead on percentage change in revenues. We now note that, over eight years taken together, the industry’s expected baseline profits plus new profits from the Medicare law are likely to be at least $587 billion. We are grateful, however, that the PRI/PwC critique identified one flaw in our original work: The drug makers’ estimated $139 billion in additional profits over eight years would mean a rise of 31 percent (not 38 percent) from the baseline of a cumulative $448 billion in projected profits absent the law. Our calculation had understated the industry’s expected profits after 2006 without the law, and therefore overstated the percentage rise in profits with the law. E. OTHER PROBLEMS IN PRI/PWC CRITIQUE 10. Ignoring Conservatism of HRP Estimates: Our estimates were conservative in several ways. A contingency table gave a range of estimates. PRI/PwC ignores these. 11. Unquestioning Acceptance of CBO Estimates: The PRI/PwC critique seems to suggest that our disputing CBO estimates is irreverent, or almost uncivilized. Yet the recent CMS cost estimates did not accept the CBO figures on many points. We worry that CBO staff may have faced pressure to develop an estimate consistent with the $400 billion target cost. 12. Ignoring Need for New Solutions and a Peace Treaty: The PRI/PwC critique assumes that today’s high drug prices can and should persist. It ignores the need for and the feasibility of the prescription drug peace treaty we discussed, in which lower prices would benefit drug makers as well as patients and payers. Drug makers would shift from their current business plan of high prices combined with limits on use to a business plan based on lower prices and higher volume, to meet the needs of all Americans. iv Introduction In October 2003, while selected members of a Congressional conference committee prepared their Medicare bill behind closed doors, we estimated the increased profits that the House and Senate bills offered for prescription drug makers. If the law were to contribute $400 billion in new federal spending to a Medicare prescription drug benefit starting in 2006, as Congress proposed, we concluded that drug manufacturers’ profits would rise by $139.2 billion over eight years. Drug makers would retain as profits, we estimated, 61 percent of the Medicare dollars that would be spent to buy additional prescriptions that seniors and people with disabilities currently cannot afford to fill.1 Recently, PricewaterhouseCoopers (PwC) disputed these estimates in a critique commissioned by the Pacific Research Institute.2 The PRI/PwC critique asserted that our analysis “grossly overestimates” the Medicare drug benefit’s impact on pharmaceutical company profits (press release, 4 March 2004). This report identifies the main problems with that critique. In general, PRI/PwC seem to take a posture of anointing or sanctifying their own work. They rely on anonymous industry sources. They do not disclose their calculations. Even as national attention focuses on the disparity between Congressional Budget Office and executive branch actuaries’ cost estimates for the law, PRI/PwC contend that disagreement with CBO’s work constitutes a radical or self-evident failing. They similarly disparage failure to uncritically accept a crucial estimate from a thirty-year-old RAND Institute study—of different drug benefits, enrolling only younger, non-Medicare beneficiaries, and conducted in a radically different pharmaceutical climate. The following response specifically addresses these matters. For example, it describes six reasons why it is inappropriate to rely on the RAND study’s estimate of the priceelasticity of demand for prescription drugs. Unlike PRI/PwC, we have never pretended to know the precise price-elasticity of demand for prescription drugs under the new Medicare prescription drug benefit. Directly relevant evidence is lacking. Instead, in our original report, we documented Medicare beneficiaries’ needs for higher volumes of drugs and differences in use between Medicare patients with and without prescription drug coverage. We also provided a contingency table that sets out the effects on drug makers’ profits of changes in the values of two of the three key factors bearing on those profits under the new law. On completing our October 2003 study, we were dismayed by the Medicare prescription drug plan that Congress was considering, because of its inadequate benefits and its high costs, with too much money going to drug makers’ profits because prices were not controlled. And these traits have left many others similarly dismayed. The PRI/PwC critique seems to focus its energy on a few aspects of our analysis and says virtually nothing about the new Medicare drug benefit itself. Someone who uncritically accepts the PRI/PwC critique might therefore be led to believe that the new benefit is a good one. But that would not be true, even if the PRI/PwC critique were accurate. Were the PRI/PwC critique accurate, the overwhelming share of the money under the bill would apparently go to replacing existing spending on prescription drugs—private insurance, Medicaid, and out of patients’ pockets. (We use the word “apparently” because PRI/PwC never specifies this share.) This means that very little money would go to buy medications for patients who now suffer because they cannot afford needed drugs and don’t get them today. The split of dollars under the new law between buying additional prescriptions or replacing dollars that pay for existing prescriptions is therefore a vital matter for patients who can’t now afford needed medications. That split also is strongly associated with the rise in drug makers’ profits under the new Medicare benefit. Our October 2003 Health Reform Program (HRP) report noted three main reasons why drug manufacturers would likely make enormous windfall profits under the Medicare legislation then about to come before Congress. Though the PRI/PwC critique attempts to dazzle the reader with a flurry of numbers and appearances of welldocumented analysis, its criticisms are mistaken. They do not undermine our analyses of any of the three main aspects of the legislation that we identified as contributing to a steep increase in drug makers' profits. 1. Drug manufacturers’ prices would not be restrained under the new law, we concluded. Congress agreed, at drug industry insistence, to fragment Medicare's purchasing power and also bar Medicare from negotiating drug prices. (The new law would also tend to keep prices high for Medicare patients and others because it retained the ban on importing lower-cost medications from other nations.) Discounts under the new law would likely resemble the 15 percent now prevalent in the private market. That’s because it would rely on multiple private insurance companies and plans, despite their long record of failure to contain drug prices for American families, employers, and Medicare HMO enrollees. (HRP, pp. 1, 5) The PRI/PwC critique (p. 4), while concurring that today’s employer plan discounts average only 15 percent, says Medicare plans will win “discounts of around 25 percent.” PRI/PwC offers no evidence at all to support this claim. 2. The volume of drugs sold would rise substantially, we predicted. Much of the proposed $400 billion Medicare subsidy would likely buy prescriptions not filled today. Unmet need is great. (HRP, p. 3) But the PRI/PwC critique vastly underestimates the likely rise in volume when Medicare patients gain coverage. Among PRI/PwC’s problems: ‰ They underestimate by one-third the number of seniors lacking coverage today. ‰ They use a 30-year old price-elasticity estimate that is inappropriate to the Medicare population, to the benefit structure of the new Medicare bill, and to the 2 nature of today’s pharmaceuticals or the climate in which they are marketed. (This is very much like learning, with reasonable accuracy, the size or climate or existence of life on one of the planets in our solar system and then heroically asserting that the other planets must be essentially very similar—without regard to distance from the sun or other empirical evidence.) ‰ They assume that a huge but never-specified share of the $400 billion simply replaces current spending. ‰ They never offer their own estimates of the split of prescriptions under the Medicare drug benefit between new and old prescriptions. 3. Drug makers’ new costs would be very low, we emphasized. Almost all of drug makers’ revenue from these new prescription sales at high unrestrained prices would be profit. Pre-existing revenues finance their manufacturing, administration, research, marketing, and other costs, so drug makers’ only added costs on the higher volumes of prescription drugs they would sell under the new Medicare benefit would be the “remarkably low real cost of producing [and distributing] the added volumes of pills…” (HRP, pp. i, 5-6) The PRI/PwC critique simply ignores the issue of drug makers’ low incremental costs. While the PRI/PwC critique erred in these three areas, we are grateful to its authors for identifying a flaw in our original work. The main estimates highlighted in our 31 October report were these: • An estimated 61.1 percent of the Medicare dollars that will be spent to buy more prescriptions will remain in the hands of drug makers as added profits. • This windfall means an estimated $139 billion in increased profits over eight years for the world’s most profitable industry. • At $17 billion annually, this means about a 38 percent rise in drug maker profit. In the third point, the 38 percent figure reflected a $17 billion rise over prescription drug makers’ projected profits of $45.3 billion in 2006 without the new law. But that $139 billion over eight years would represent a rise of 31 percent from the industry’s eight-year cumulative $448 billion in projected profits in the absence of the law. (We had compared the $139 billion to a denominator that understated the industry’s expected profits after 2006 without the law, so the result overstated the percentage rise in profits for the law’s first eight years.) It does appear that the PRI/PwC critique, after identifying a problem with our original calculation, actually mirrored our error in their own work. Their calculation of the percentage rise in profit was too low. They suggest that our profit estimates (which they dispute) would mean a 26 percent rise (p. 7). Their undisclosed calculations appear to have used the high end of the eight-year period for the baseline, the opposite of our own error in using the first year. 3 Further, since we wrote in October, the 2003 profit reports from Fortune have been released. They show a slight drop in profits for the ten largest drug makers.3 If the industry’s profits before 2006 drop slightly or grow more slowly than the six percent average annual rise that we have projected, the percentage rise in profits owing to the new Medicare benefit would be substantially greater than the 31 percent increase now projected. Please note that our study was conservative in that it estimated only the higher profits arising from the new $400 billion in federal financing under the law—the new profits that drug makers will win on increased volume under the Medicare law. We therefore excluded profits on beneficiaries’ insurance premiums or on their co-payments. The $139 billion in new profit is only part of the total profits that drug makers will garner on patients covered by Medicare. The total profits on Medicare patients’ medications will of course be substantially larger. They will include both the new profits and existing profits on current purchases of Medicare patients’ prescription drugs. The Medicare benefit will take over paying for some current purchases of Medicare patients’ drugs. So, as we calculated, drug makers’ discounts under the law will result in some reduction in the profits from existing sales. As discussed below, however, the PRI/PwC critique appears to assume that the law will result overwhelmingly in such replacement sales. We did not attempt to develop our own estimate of the cost of the Medicare law or its drug benefit. Rather, we took as a given the widely-cited Congressional Budget Office estimate that the Medicare prescription drug bills prepared by the House and Senate would each cost approximately $400 billion over 10 years (including eight years of Medicare drug benefits). Our report was written before evidence surfaced publicly of higher CMS estimates of the cost of the law. The PRI/PwC critique, though, tries to simply wave away the higher CMS estimate that the law will cost $534 billion, or about one-third more than CBO estimated. There are many reasons to suspect that the CMS estimate is more accurate—or even low itself—and therefore that drug makers’ revenues and profits will be still higher. (Some portions of the higher estimate do reflect additional sums going to HMOs and other parties under the final version of the law.) As we prepared our original report, however, Congressional leadership had announced their intent to write a law that would not incur more than $400 billion in net federal cost. In our analysis, we treated the $400 billion as a commitment by the federal government, and worked to estimate where that new $400 billion in federally-financed drug spending for Medicare beneficiaries would actually go. The PRI/PwC critique asserts that drug makers will not win close to the $139 billion in new profits that we projected, but they do not offer an alternative figure. They offer fragmentary criticisms and alternative calculations, but never put the pieces together into a coherent whole. Among the essential calculations that they ignore are these: 4 • they never address drug makers’ incremental costs to make more pills, so they cannot estimate profits arising from the law (revenue minus costs); • they focus on price-elasticity but never show how this translates into dollar estimates of the overall rise in spending; and • they never offer their own conclusions about the split between old and new business paid for under the new law. If a substantial portion of CBO’s projected new $400 billion in federal spending is not going to profits on new prescriptions, where would all that new money go? The PRI/PwC critique simply doesn’t explain. In their critique, the $400 billion appears to be an incredible vanishing expenditure. As shown later, PRI/PwC think it doesn’t even visibly boost drug makers’ total revenue. As well, please consider that: • The PRI/PwC critique suggests that we underestimated discounts and therefore overestimated the prices that manufacturers will be paid, so higher prices can’t account for the extra money. • The PRI/PwC critique does not dispute our estimate of the low actual cost of manufacturing and distributing the additional pills to fill new prescriptions. • Compared to our analysis, however, the PRI/PwC critique assumes that a much smaller share of the new money will go for increased volume, and a much bigger share will take over the cost of paying for existing drugs. This means that PRI/PwC clearly assert that the law will do much less than we projected it would to bring prescription drugs to patients who cannot afford them today. Proponents of the law therefore should view our analysis, critical of the bill as it was, as more optimistic than is PRI/PwC’s critique. PRI/PwC suggest that the law will mean relatively little rise in new volume, while we suggest that the $400 billion—despite the law’s limited benefit, unnecessarily high patient cost-sharing, and high drug prices—will enable more patients to fill prescriptions that otherwise would go unfilled. Our report assumes that any discounts—whether on drugs paid for by the plans or drugs that patients purchase while in the $2,850 hole of no coverage (from $2,250 in drug costs to $5,100 in drug costs)—will be passed on to patients as lower prices, enabling patients to afford to fill additional prescriptions. (If HMOs or insurors win large discounts, however, they might keep some of the money and use it to boost their profits, as discussed below.) The PRI/PwC critique thus falsely claims that we ignored the benefit of lower prices for patients’ out-of-pocket purchases. (p.7) Worse, although they themselves then say discounts will be passed on to patients as lower prices or premiums (p. 7), they fail to 5 follow the money and its impact on drug makers’ revenues and profits. They never show where they think the savings from discounts would actually go. PRI/PwC concedes (footnote xxviii) that “additional discounts lead to additional induced demand” but they never provide a clear, explicit, or integrated display of their own calculations of either induced demand or drug makers’ revenue. We contend that, under the new benefit, dollars saved through discounting flow back to drug makers when patients use the saved money to purchase needed medications. Any discounts greater than the 15 percent allowed for in our analysis would, we expect, go to offset some of the effect of the law’s high cost-sharing. And that recycled money for new prescriptions again flows to the drug makers—who incur little cost on providing the added volume of medications—so they recoup additional profits on the added volume of drug sales. The PRI/PwC critique repeatedly condemns our analysis for differing from CBO estimates on various points. Of course, those who make all the same assumptions that CBO makes would end up where CBO ends up. And the PRI/PwC critique’s rigid adherence to CBO estimates is not surprising, since the authors from PwC are former longtime staff of CBO (see PRI/PwC, p. 13), apparently imbued with CBO’s perspective. PricewaterhouseCoopers describes itself as a firm which provides “business advisory services….to all market sectors in the health care industry.” (PRI/PwC, p. 13) The Pacific Research Institute describes itself as “advancing free-market policy solutions,” stating that it “champions…personal responsibility for all individuals” and “demonstrates why the free market is more effective than the government…” (PRI/PwC, p. 14) The PRI/PwC critique starts by contending that our findings are “well outside the range generally accepted by government and private sector analysts” and those “who track developments in the pharmaceutical industry.” (PRI/PwC, pp. 1,2) Yet the authors have few arguably impartial sources for alternative estimates. They cite just one that tracks industry developments (Merrill Lynch), and otherwise mention only anonymous and published sources from inside the drug industry. (p. 2 and note 4). Although the PRI/PwC critique purports to be careful and objective analysis resting on valid evidence, most of it is actually suspended on the gossamer threads of fragile, uncoordinated, and often irrelevant factoids. As we document below, PRI/PwC build an elaborate structure on a shaky foundation of non-existent or largely irrelevant evidence. The reality probably remains apparent to most observers. The federal government’s commitment of $400 billion in new Medicare subsidies for prescription drugs will not be used overwhelmingly to replace existing undiscounted drug spending. The new federal dollars won’t vanish into thin air either—something which PRI/PwC astonishingly predicts when it asserts that drug makers’ total revenue after the law’s implementation will end up within the range of a 1.0 percent drop and a 3.2 percent rise. 6 Rather, today’s enormous unmet need among patients with no drug coverage or very limited coverage means that a very substantial portion of the new spending will finance new drug purchases and thus boost drug makers’ revenues. Because drug makers’ cost of providing those new medications is so low, even discounted prices will sharply increase their profits. And there is still no evidence that future price discounts will be any larger than today’s discounts. Sadly, the limited discounts and high prices under the new Medicare benefit will mean that this new federal $400 billion will buy far fewer medications than it could— and that far too much of the money will go to benefit drug makers and their stockholders. The main body of this response to the PRI/PwC critique is organized into 12 sections. In each section, we generally first summarize the findings or methods employed in an aspect of our report. We describe the PRI/PwC critique of that aspect and then respond to their critique. The issues in the 12 sections inevitably overlap as we attempt to address the critique’s disjointed elements. We address certain central points in more than one of those interrelated sections, and we hope that some repetition is tolerable in the pursuit of clarity. 7 A. COST OF MEDICATIONS 1. IGNORING DRUG MAKERS’ LOW INCREMENTAL COSTS HRP WINDFALL PROFIT REPORT: Our October 2003 report described three main reasons why drug makers would garner windfall profits under the Medicare legislation then before Congress. Their projected soaring profits reflect unrestrained prices, a growing volume of prescriptions filled, and the “remarkably low real cost of producing the added volumes of pills that Medicare patients need.” (See p. i of our report.) “Drug makers’ costs don’t change very much if they produce and sell more pills,” we noted: [M]arketing, administration, and research should not be expected to change very much when sales to Medicare patients grow. The only added costs for drug makers associated with a new Medicare prescription drug benefit are expected to be in manufacturing and in distribution . . . . Drug makers’ factories are already built, and the ingredients in most prescription drugs are inexpensive. So there will be a very small incremental cost to make the added volume of medications that newly covered Medicare patients will use. Estimates from industry sources are that the actual added cost of making additional pills is only about five percent of the undiscounted full retail price. [See pp. 5-6, and documentation in note 20.] We therefore estimated the incremental manufacturing cost at 5.9 percent of Medicare’s likely discounted retail price. (This reflects our estimate, discussed below, that drug plans are likely to obtain discounts of 15 percent, “so the manufacturing share is 5 divided by 0.85, or 5.9 percentage points.”) Further, we allowed for 2 percent of drug makers’ actual revenues “to cover the added cost of distribution. Manufacturers’ costs of the additional volume of drugs therefore sum to 7.9 percentage points.” Subtracting that from drug makers’ incremental revenues leaves incremental net revenue or added profit. Our sensitivity analysis (pp. 8-9) showed the effect on drug makers’ estimated rise in profits if their incremental manufacturing costs exceed 5.9 percent, ranging up to 15.0 percent of Medicare’s likely retail prices. The $139.2 billion figure for new profits over eight years rested on the estimate that the cost of making any increased volume of medications will be 5.9 percent of payments made at retail. Even if incremental manufacturing costs are nearly triple our main estimates, reaching as high as 15.0 percent of retail prices, as shown in the contingency table in our Exhibit 3 (p. 9), the estimated rise in profits would still be $118.4 billion over eight years. PRI/PwC: The PRI/PwC critique does not dispute our estimate that drug makers’ incremental costs of making more pills are likely to be only about five percent of full retail prices. Indeed, PRI/PwC totally ignore the issue of drug makers’ costs. Thus, they ignore both the main estimate that we offered and our contingency table with 8 alternative estimates for those who make other assumptions about drug makers’ incremental costs. In other writings, however, PRI has concurred that “Actual production costs for the new drugs – after the huge initial investment in R&D – in most cases are relatively low.”4 The profits of a business depend on both revenues and costs, but the PRI/PwC critique has nothing to say about the cost side of the equation. REPLY: Yet drug makers' low actual incremental cost is one of the key elements in the high profits that this law will bring them. Consider a dollar spent at retail to buy new prescriptions, with 25 cents going to pharmacy markup (to cover costs and profit) and wholesalers, and 75 cents going to drug makers. If actual incremental manufacturing cost is 5 cents, a very large share remains for the drug makers’ profit— even if 15 cents returns to the patient or insurer in discounts or rebates. Even if manufacturing costs are double that, the sum available for profit is still high. And, as discussed later, this remains true with discounts and rebates paring another 5-10 cents off the sum paid to drug makers. Windfall profits will be large on any new sales because it will cost drug makers so little to make the additional pills that Medicare patients need— an issue that the PRI/PwC critique entirely ignores. Indeed, the low incremental cost for added volume means that, whatever the rise in revenue for the drug makers, nearly all of it will represent profit. One additional point is worth noting. The industry commonly asserts that high prices for drugs must be maintained because the costs of research are high, an argument that PRI has strenuously and repeatedly emphasized and supported in other writing.5 The PRI/PwC critique does not raise that issue, however. Nor does it address a point noted in our October report (p. 5), that under the new law, Research costs should not increase either. Today, drug makers say they need high prices to sustain research. They may respond to this report by claiming that they would use their vast new profits to conduct more research. But they have never hinted at any plan to vastly expand research when Medicare sends their revenues soaring. And there is no reason to assume that the drug makers should—or could effectively—expand their research budgets tremendously. Few Medicare beneficiaries who seek a prescription drug benefit now expect that one of the benefit’s main aims or consequences is to give drug makers more money that they would promise to use to finance research. Even if those promises were ultimately kept, such a use of scarce public funds should certainly be debated. Any additional substantial growth in drug research might better be used to expand research budgets at the National Institutes of Health. (Already, for example, developing “me-too” drugs consumes a large share—perhaps 40 percent—of industry-financed research.) Further, there is no reason to allow for a large rise in marketing costs. Drug makers will face a rise only in production and distribution costs, both of which our analysis reflects. 9 B. VOLUME OF MEDICATIONS 2. UNDERESTIMATING THE NUMBER OF PEOPLE IN NEED HRP WINDFALL PROFIT REPORT: We estimated that some 40 percent of some 40 million Medicare beneficiaries lacked any prescription drug coverage in 2003, and that millions more were under-insured. (HRP, p. 4) This reflects a well-respected estimate by Laschober that, as of 1999, at any given time, 38 percent of Medicare beneficiaries lacked coverage.6 We used 40 percent because of the steady drop in prescription drug coverage under retiree health plans since 1999. The uncovered share may well worsen substantially further by the time a new Medicare benefit takes hold—owing to cuts in retiree health coverage, Medi-Gap coverage, and Medicare + Choice coverage— but, conservatively, our estimates did not assume that. PRI/PwC: In asserting that we greatly overestimate the rise in use that will result from added coverage, they say "almost 30 percent of Medicare beneficiaries have no coverage." (PRI/PwC, p.4) They cite no source. They then state, “Other beneficiaries, such as those with Medicaid or employersponsored coverage, will increase prescription drug use by less since they already have at least some coverage.” REPLY: PRI/PwC’s assertion that the rise in use will be substantially lower than we project reflects in part their very serious underestimate of the number of people who lack drug coverage today. They also appear unaware of, or intent on downplaying, how skimpy the coverage is for many of those who do have some benefits today. PRI/PwC underestimate by at least one-third the number of Medicare beneficiaries entirely uninsured for prescription drugs. Their 30 percent estimate may reflect, however, an old but fairly widely-cited estimate of the share who lack coverage for an entire year. But that is well below the actual share uninsured at any one time. This is the more useful and appropriate figure for estimating need. (Imagine, for example, a group of ten people, in which five have coverage for the full year, three lack coverage all year, one has coverage only for the first half of the year and one has coverage for the second half. Looking at coverage for the year as a whole and counting the last two people as “covered” would ignore their half-year without protection, while counting them both as “uninsured” would over-estimate the group’s level of need. A cross-sectional estimate for the average day in that year avoids that problem.) Laschober, as just noted, estimated the average share uninsured at any one time at 38some percent for 1999, and the uninsured share has doubtless risen subsequently. (Similarly, a 2001 Towers Perrin analysis concluded that fully 40 percent of people on Medicare lacked prescription drug coverage.7) 10 Elsewhere, PRI itself—and other conservative organizations as well— have asserted that only 24-25 percent of Medicare patients, or even fewer, lack drug coverage today.8 They often do so in documents aiming to dispute the need for a universal Medicare drug benefit and urging a targeted one. Similarly, CBO itself has emphasized an estimate that “about 25 percent of all seniors did not have prescription drug coverage of any kind,” sometimes without stressing or even noting that such figures represent the share who go the entire year without coverage.9 The U.S. Department of Health and Human Services has done the same.10 There is also a tendency to downplay the problem that today’s prescription drug coverage for many people has very inadequate benefits. When discussing insured patients, PRI/PwC mention those in employer plans and Medicaid, some of whom have fairly comprehensive prescription drug coverage; but PRI/PwC entirely ignore patients with Medicare HMOs, which often provide just $500 or $750 in drug benefits annually, and patients in Medigap plans, which require 50 percent coinsurance for drugs. In other writing, as the new law was on the verge of passage, PRI belittled the financial burden of drug costs for Americans, and stated only that, “Drug spending does appear to be a major problem for two percent of Medicare beneficiaries who earn less than twice the poverty level [about $18,000] yet spend $4,000 a year on drugs.”11 They did not explain why, for example, $3,000 in drug costs should not be considered a major burden for those at such income levels or lower, or even for seniors living on $30,000, especially considering the high coinsurance and deductibles that many such chronically-ill Medicare patients must pay to obtain other medical care. Indeed, PRI characterized the new Medicare drug benefit as representing a decision by “Members of Congress…to pander to the entire senior voting block.” PRI stated no concern about the fundamental ways in which the law was shaped to the pharmaceutical industry’s wishes. The failure in the PRI/PwC critique to acknowledge how many Medicare beneficiaries are uninsured today and how many are underserved by their current drug coverage may result from PRI’s apparent stance recommending shallow coverage as desirable for containing costs. In several articles last year, PRI staff criticized co-payment requirements for “shielding consumers from the price differences between drugs,” and urged “a new consumer-driven model that allows consumers to see the full price differences between drugs.” 12 PRI/PwC’s inaccurate estimate of the number of Medicare beneficiaries lacking insurance badly undermines its subsequent estimate that insurance covers 58 percent of beneficiaries’ current prescription drug costs. (Please refer to section 4, on the extent of insurance coverage and the circularity of PRI/PwC estimates.) 11 3. UNDERESTIMATING EFFECT OF COVERAGE AND LOWER PRICES HRP WINDFALL PROFIT REPORT: Our estimates of response to lower prices rest partly on Poisal's data on the difference in prescription drug spending between Medicare “beneficiaries with and without coverage."13 Those are real and recent data on drug spending for the population of concern, Medicare patients. We also considered evidence on inability to purchase prescribed medications, and on substantial price-elasticity of demand associated with drops in drug prices. PRI/PwC: PRI/PwC assumes that the maximum plausible drug price-elasticity of demand is minus 0.3, the figure they say that CBO uses, one based on RAND research. (PRI/PwC, pp. 5-6, 8) They note that the level of induced demand depends on the level of coverage, but they don't take note that Medicare patients differ from the population as a whole—that Medicare patients are sicker and poorer, so cutting prices might do more to boost their demand than for the average American. They blandly and uncritically assume that the CBO/RAND prescription drug price-elasticity estimate reflects the needs and financial situations of Medicare patients—and the nature of the new Medicare drug benefit itself. In discussing “induced demand,” they criticize our use of Poisal's data, which they describe as “assuming that those without coverage will increase their average spending to the same level as those with comprehensive coverage” (p. 5) PRI/PwC goes on to say that use of these data "ignores any systematic differences in the populations without drug coverage and with full coverage." REPLY: The PRI/PwC critique is an elaborate construction resting on a narrow, shaky foundation—one single 30-year-old estimate of the response to lower drug prices, an estimate that bears little relation to the real world of today’s Medicare patients and the design of the new Medicare prescription drug benefit. Their entire analysis depends on the appropriateness of their assumption (from CBO and RAND) that, for Medicare beneficiaries gaining expanded coverage, a 10 percent drop in out-of-pocket cost would raise drug utilization by 3 percent. a. Price-elasticity of demand PRI/PwC ignores the issue of systematic differences—and specifically, any difference in price-elasticity of demand—between Medicare and non-Medicare patients. PRI/PwC claims that it relies on “accepted research” (PRI/PwC, p. 2). It would be better for them to use appropriate and relevant research. 12 For example, as discussed above, PRI/PwC mistakenly uses data on the proportion uninsured throughout the year when the proportion uninsured at any one time (in other words, on an average day) is the appropriate figure. The research on the proportion uninsured throughout the year may be accepted in some quarters, particularly among those who would like to portray the problem as relatively small, but it is still starkly inapplicable to the question at hand. The RAND Health Insurance Experiment research, on which PRI/PwC relies even more heavily, may be accepted and “well-known,” (PRI/PwC, p. 8) but, unfortunately, it does not appear to be very relevant to gauging patients’ price-elasticity of demand under the new Medicare prescription drug benefit. Even with its restricted benefits and its requirements for patient cost-sharing, the new Medicare law should mean a substantial increase in the number of prescriptions filled. As we noted, a study in eight states, for example, found that nearly one-fourth of seniors surveyed reported that they skipped doses of medication or failed to obtain prescribed drugs because of high costs. Further, a November 2002 Harris poll found that 18 percent of surveyed adults having failed to ask for prescriptions because of their cost, and 22 percent failed to fill a prescription because of the cost, with higher proportions among sicker adults. (For other detailed evidence suggesting great unmet need, see HRP, p. 4.) These figures ignore patients who do not visit the doctor for non-surgical needs because they rightly anticipate they will not be able to afford the medications they expect to be prescribed. Marketing to both patients and physicians remains unrestrained under this the new Medicare drug law, and drug makers and advertisers will try hard to drive up demand for both needed and unneeded drugs. The PRI/PwC critique assumes that the price-elasticity of demand for prescription drugs for Medicare patients will be only minus 0.3 (PRI/PwC, p.5). This is the figure used by CBO, whose work PRI/PwC uncritically accepts throughout its own critique. The PRI/PwC critique tries to wave away the CMS estimate that the law will cost $534 billion, or about one-third more than CBO estimated— which in part apparently reflects higher estimates of the rise in spending per person.14 The estimate of minus 0.3 for price-elasticity of demand for prescription drugs entirely rests on one large RAND study, the Health Insurance Experiment, conducted almost 30 years ago. This is a very shaky foundation on which to build for at least six reasons. 1. Few medications with high annual costs, that could benefit older people, were available at the time of the RAND study—as compared with those available today. Few medications that patients would use daily for many months, many years, or their entire lifetime were then widely available or prescribed. That means it was hard for most patients to spend a great deal of money on medications that provided 13 substantial or sustained benefits. This has already changed enormously. Patients for whom prescription drugs are made more affordable have much more reason now to increase their use of medications. And, with the encouragement of a new Medicare drug benefit, it is likely to change even more in the near future. Please consider the growing availability of very costly, new, and potentially effective medications. Many would be unaffordable without the new Medicare benefit.15 2. Given the prominence of prescription drug advertising on television and other direct marketing to patients, many more patients may be more inclined to use drugs today—and the Medicare law will do nothing to restrain such advertising. 3. The RAND Health Insurance Experiment did not enroll anyone over age 65 or anyone with substantial disabilities—that is, no one resembling the Medicare population. Over the years of the study, a small number of participants entered that age group or developed disabilities, but the study population was nonetheless almost completely unrepresentative of the Medicare population. (As discussed elsewhere, PRI/PwC inaccurately criticize us for not highlighting the possibility that there are systematic differences in likely use rates between Medicare patients who have coverage and lack it today— yet they rest their entire analysis of the rise in prescription drug use on data from a non-Medicare population.) 4. The RAND study used co-insurance of varying percentages to measure the priceelasticity of demand. The study looked at differences in use among participants who had no co-insurance, and others who had to pay 25 percent, 50 percent, and 95 percent. Further, all received cash payments equal to their maximum dollar exposure, which was 5, 10, or 15 percent of income, but capped at $1,000 (roughly $3,000 in today’s dollars). These cash payments were not earmarked for health expenditures—but they were available to offset the out-of-pocket costs of prescription drug and other medical goods and services. There were no separate dollar caps for prescription drugs, so any large medical expenses could exceed the caps, and patients would then face no out-of-pocket costs for drugs. While the RAND Health Insurance Experiment’s design was undoubtedly valuable for many purposes envisaged in the 1970s, it does not resemble the structure of the new Medicare drug benefit, so its usefulness as a guide to patient behavior in that plan is low. 5. In the new Medicare drug program, most enrollees would pay a monthly premium. That would orient them to using the benefit they were paying for, particularly since so many costly and potentially drugs are available and are prescribed by physicians. 14 We simply do not believe it is plausible that a new Medicare drug benefit that includes 75 percent coverage (that is, a 25 percent co-insurance) on drug spending at levels from $250 to $2,250 yearly will have a price elasticity of demand remotely close to minus 0.3. As described by PRI/PwC, this price-elasticity of demand of minus 0.3 would mean that a drop in out-of-pocket cost of 10 percent would mean a rise in drug utilization of only 3 percent. (In the promised Medicare drug benefit, each dollar of out-of-pocket spending within this $2,000-wide band buys four dollars worth of medication because each such dollar mobilizes three dollars of insurance coverage.) Again, that calculation rested on the RAND Health Insurance Experiment, whose design, enrollment, and prescription drug environment did not resemble that of a Medicare drug benefit taking effect in 2006. An example may be helpful. As a 70-year-old Medicare patient, imagine that yesterday you had no insurance and must pay for all medications out-of-pocket—a reality that applies to roughly 40 percent of Medicare beneficiaries. Suppose you had two prescriptions, each for $100 monthly, which your doctor says you must take throughout the year, for a total out-of-pocket cost of $2,400. You decided that you just couldn’t afford the bill. Now imagine that, today, if you pay a $420 per year premium and a $250 deductible, you enjoy 75 percent insurance for the next $2,000 of your drug costs this year. Once you pay the insurance premium, you are oriented toward trying to get your money’s worth. You fill both of the prescriptions. After the premium, your total outof-pocket spending is $787.50. But the total rise in your prescription drug spending is $2400.00. You go from using none to using $2400.00 of medications. The PRI/PwC critique seems to suggest that, in this example, a subsidy of 67.2 percent of the $2400 cost of the medications would yield a typical rise in use of drugs of only 22.4 percent. That doesn’t seem to make sense. (But this small rise in use is consistent with PRI/PWC’s later-stated small rise in the share of drug purchases covered by insurance from fully 58 percent to 70 percent. See PRI/PwC, p.6.) 6. Economists and actuaries who try to project use of very large new programs from the experience with narrow empirical studies, extrapolation from other populations, and the like, very often miss the mark badly. Recall the explosive growth in use of hospital, physician, and nursing home services in the wake of the passage of Medicare and Medicaid in 1965—growth far beyond what the best actuaries had predicted at the time. Recall the explosive growth in use of many health services in the United Kingdom after the introduction of the National Health Service in 1947. In these six ways, the PRI/PwC critique rests on narrow and old evidence about a very different benefit in a radically different pharmaceutical climate, while ignoring the history 15 of under-estimation of actual use under new health care programs. There is little reason to credit the PRI/PwC critique. What PRI/PwC has done by uncritically accepting RAND’s minus 0.3 price-elasticity of demand is very much like learning, with reasonable accuracy, the size or climate or existence of water or of life on one of the planets in our solar system, and then heroically asserting that the other planets must be very similar in all important respects—without regard to distance from the sun or any other actual empirical evidence. Unlike PRI/PwC, we have never pretended to know the precise price-elasticity of demand for prescription drugs under the new Medicare prescription drug benefit. Directly relevant evidence is lacking. Instead, in our original report, we documented Medicare beneficiaries’ needs for higher volumes of drugs and we also provided a contingency table that sets out the effects on drug makers’ profits of changes in the values of two of the three key factors bearing on those profits under the new law. Other predictions of price-elasticity A number of close observers of the drug industry have generally spoken or written in ways that suggest a price-elasticity of demand for medications that exceeds minus 0.3. For example, • Jordan Schreiber, health fund manager at Merrill Lynch, said in August of 2000 that “Even with drug price cuts I think there’s a good chance the pharmaceutical group will actually come out as a net beneficiary as the presently uninsured become customers, albeit less profitable customers.”16 Merrill Lynch was, at this time, apparently anticipating that substantial price cuts—in the range of 40 percent from manufacturer’s price— would accompany a new Medicare drug benefit. 17 Since the new Medicare drug benefit enacted in 2003 actually makes drugs more affordable to patients largely through federal subsidies rather than through price cuts, drug makers will be able to sell more pills at prices close to full retail. Resulting profit growth will be substantial. • David Lipson of IMS Health said in 2001 that “Despite the industry’s initial concern about the effects of such a [Medicare prescription drug] program, the consensus now is that coverage for the elderly will expand the market.”18 Lipson was speaking at a time when anticipated federal subsidies were substantially below those estimated even by CBO for the benefit enacted in 2003. • As an Australian government evaluation has observed, PhRMA claims that “price controls do not reduce pharmaceutical expenditure, citing studies by Redwood in 1993 and Gross in 1994 that lower prices lead to a sufficient increase in the volume of drug sales to maintain the same level of revenue.” 19 16 Specifically, in its Industry Profile for 2000, PhRMA wrote that “A 1993 study by Heinz Redwood and a 1994 study by David Gross comparing international pharmaceutical-spending controls across countries found that while price controls produce lower prices, they do not reduce pharmaceutical expenditures (price times volume) or contain health care costs.”20 Thus, in arguing against price controls in the past, PhRMA has apparently endorsed findings that price controls will not save purchasers money because the price reductions will boost sales volume sufficiently to replace the revenue that would otherwise be lost to lower prices. • Some observers note that expanding coverage is likely to mean the largest jump in sales for those companies that make the costliest prescription drugs, which most uninsured patients cannot afford. "‘People with insurance for drugs not only use more prescriptions, but they also have more expensive prescriptions,’ says Frank Lichtenberg, a Columbia University health economist.”21 Other observers’ estimates of revenue and profit under the new Medicare benefit It is no surprise that numerous observers have concluded that the law will substantially raise drug makers' revenues and profits—though by amounts that vary widely. Consider these observations and estimates by various parties outside, close to, or inside the pharmaceutical industry: • The law "should have a positive impact on industry profits, with anticipated price discounting outweighed by increased volume," at least in the near term, according to a leading Standard and Poor’s analyst.22 • Calling the new law “A Good Bill for Drug Industry," and referring to "Massive New Spending on Drugs,” the editor of The Pink Sheet recently cited several estimates of the anticipated rise in drug maker revenue and profits (including our estimate of windfall profit to the industry, and the three other figures that follow). 23 He touted “over $70 billion in spending through 2007” as one of the “reasons for optimism.” He asked rhetorically, “Why Aren’t Drug Companies Happier?” and answered, “Gloating would fuel backlash.” Thus, in a recent panel discussion, according to Bloomberg News, a Merck lobbyist asserted, “`This prescription drug benefit is not a windfall for Merck….There will be tougher bargaining on prices.’ “ But Princeton economics professor Uwe Reinhardt, in the same discussion, reportedly commented, “`If any CEO of any drug company says they can’t make money on this, I would volunteer to replace that CEO….I cannot believe this isn’t a boon to the drug industry.’” 24 Reinhardt concurred with the estimate of a 9 percent rise in revenue for drug makers and said that "the percentage increase in profit is probably higher...."25 17 • IMS has projected that the Medicare benefit will mean 75-100 million additional prescriptions filled annually in the U.S., an estimate reiterated by Walgreen.26 • AmerisourceBergen has estimated there will be $10-$15 billion in new prescription drug spending in 2006 under this law. 27 What would this mean over the law’s first few years? Assume that the first-year incremental spending is $12.5 billion. Assume also that incremental spending rises by the same percentage each year as CBO’s projected baseline spending on outpatient prescription drugs—increases of 11 or 12 percent annually, through 2013.28 This would mean $26.7 billion in added spending in 2013, and $150.8 billion in new prescription drug spending added cumulatively over the first eight years of the benefit (2006-2013). • Goldman Sachs projected 9 percent incremental growth in drug sales as people gain coverage, which some observers estimated at a $13 billion increase over a year.29 • An alternative estimate, from Merrill Lynch, is that the pharmaceutical "industry could see up to $10bn in extra sales and $4.7bn in additional profits, a rise of 2 per cent, in 2006 when the benefit starts...The study estimates potential sales of $15.4bn in 2010 and $7.4bn in profits." They appear to have assumed that only 25 percent (10 million) of the 40 million Medicare beneficiaries lack coverage, rather than 40 percent. 30 • As a consultant told the Washington Post, “`It couldn't be clearer there is going to be a positive effect overall…The volume will definitely go up, There will be a lot of people who didn't have coverage before who will have it now and a lot of people getting an upgrade in terms of coverage.’ "31 Because the PRI/PwC critique does not provide either a clear table or a text summary that tracks their view of the flow of the dollars, we have tried to understand what they intended. It seems to be something like the following. What if the minus 0.3 priceelasticity of demand means that $300 billion in federal Medicare dollars going to manufacturers under the new law (after retailers and wholesalers take their shares) generates a net rise in drug makers’ revenue that’s as low as $90 billion. Where does the rest of the money go? The PRI/PwC critique seems to assume that almost all of it returns to patients as lower prices (through the higher discount that the critique assumes). A 25 percent price discount would mean that $75 billion of the $300 billion would not remain with the drug makers. Rather, it would be returned to the insurers, HMOs, PPOs, and other entities that provide drug coverage under the new benefit. What would then happen to the money? PRI/PwC don’t clearly explain this. But (a) we have here, in effect, a federal commitment to a $400 billion federal prescription drug subsidy, three-fourths of which will go to manufacturers. Further, (b) 18 Medicare beneficiaries’ need for medications won’t be met fully by that $400 billion. (The PRI/PwC critique assumes that the share of beneficiaries’ drug costs covered by insurance rises to 70 percent, which we consider high, as discussed below.) Therefore, it can be expected that the savings from the discount would be recycled to buy still more medications from the drug makers. And the drug makers earn further windfall profits on these additional purchases. b. Differences in spending between Medicare beneficiaries who were insured and who were uninsured for prescription drugs PRI/PwC blatantly misrepresents in several ways the evidence we use on differences in spending between Medicare beneficiaries who were insured and who were uninsured for drugs. We plainly described Poisal’s data as comparing uninsured Medicare beneficiaries and “beneficiaries with any drug insurance” (HRP, p. 4)—as Poisal put it, “beneficiaries with and without coverage.” Poisal’s study included people on Medicare with all sorts of drug coverage—Medicare HMOs, Medicaid, employer plans, and more, whether with slim or substantial benefits. Further, our calculations with the Poisal data conservatively brought uninsured beneficiaries only to average spending for those insured, not to the level of those with adequate coverage. Yet the PRI/PwC critique repeatedly mis-characterizes the insured group in the Poisal analysis we cited as having "comprehensive” or “full” coverage. (PRI/PwC, p. 5) This suggests either extreme carelessness or deliberate distortion. Our term, “any coverage,” obviously does not mean “comprehensive” coverage. The false claim that we use a comparison with comprehensively-insured Medicare patients is consistent with PRI/PwC’s mistaken assertion that we anticipate an implausibly large rise in spending. The PRI/PwC critique argues that we ignore possible systematic differences between insured and uninsured people on Medicare, and implies that this contributes to our alleged overestimate of new spending. But they never themselves address such differences. (And, as discussed above, they also ignore the even bigger issue of systematic differences between Medicare and non-Medicare patients.) In reality, any difference between Medicare patients who are insured or uninsured for drugs would undermine our estimates only if uninsured people who gain coverage would tend to use fewer medications than people already covered, perhaps owing to better health. The Medicare population uninsured for drugs likely has a bi-modal distribution—some are low-income patients unable to afford coverage (often despite great need), while others uninsured for drugs are healthy or wealthy, or both, and therefore have not bothered to get drug coverage. But there is good reason to think that Medicare 19 beneficiaries who lack drug coverage (and would gain it under the new law) are likely to be poorer than average. And lower income people tend to have worse than average health (implying greater prescription drug needs)—so our use of the average was, again, conservative. Quick calculations from Laschober’s data suggest, for example, that Medicare beneficiaries with drug coverage were almost equally divided between people with incomes above and below $20,000 in 1999, while a noticeably greater number of those without coverage were in the low income group. Laschober and colleagues also found that Medicare beneficiaries between ages 65 and 74 were more likely to have drug coverage than older beneficiaries and than those under 65 with disabilities—though the 65-74 group is doubtless the healthiest.32 20