BRAND LOYALTY, ENTRY, AND PRICE
`COMPETITION IN PHARMACEUTICALS
`
`AFTER THE 1984 DRUG ACT*
`
`HENRY G. GRABOWSKI and JOHN M. VERNON
`Duke University
`
`IN 1984, Congress enacted a new law that greatly affected the economics
`of the pharmaceutical industry in the United States. It has been character-
`ized as the most important legislation affecting competition in the phar-
`maceutical industry since the 1962 Kefauver-Harris Amendments to the
`Food and Drug Act. This 1984 law, known as the Drug Price Competition
`and Patent Term Restoration Act (hereinafter the 1984 Act), facilitated
`the entry of generic drug products after patent expiration while it also
`restored part of the patent life lost during the premarket regulatory pro-
`cess for new introductions.‘
`
`Market entry by generics was relatively limited prior to 1984 because
`of costly Food and Drug Administration (FDA) requirements that had to
`be met by the imitative products. That is, generic drugs often would have
`to duplicate many of the pioneer’s tests to gain market approval after
`patent expiration. As a result of the 1984 law, generic products need only
`demonstrate bioequivalence to the pioneer’s brand, and generic entry has
`increased significantly. This has provided a body of very interesting data
`to analyze the pattern of entry and the pricing strategies followed by the
`entrants and incumbents.
`
`In this article, we make use of data covering the sales and prices of
`the pioneer and generic products for eighteen drug products, generally
`over the time period 1984—88. A number of issues are examined. First,
`
`" Earlier versions of this article were presented at the Second World Congress on Health
`Economics. the Applied Econometric Association Meetings. the American Economic Asso-
`ciation Meetings. and Workshops at Duke University and University of California. Los
`Angeles. We benefited from the comments from several participants at these meetings as
`well as those from an anonymous referee.
`' For an early analysis of this law, see Henry Grabowski & John Vernon. Longer Patents
`for Lower Imitation Barriers: The 1984 Drug Act, 76 Am. Econ. Rev. Papers & Proc. I95
`( 1986).
`
`[Journal of Law & Economics, vol. XXXV (October l992)]
`© [992 by The University of Chicago. All rights reserved. 0022—2186/92/3502-000650l50
`
`331
`
`000001
`
`Exhibit 1066
`
`ARGENTUM
`
`Exhibit 1066
`ARGENTUM
`IPR2017-01053
`
`IPR2017-01053
`
`000001
`
`

`

`332
`
`THE JOURNAL or LAW AND ECONOMICS ‘
`
`We characterize the typical pattern of generic prices over the early years
`postentry, the pricing responses by pioneers, and the resulting market
`shares of the two parties. The current strong interest in theoretical model—
`ing of entry-deterring strategies makes this empirical evidence particu-
`larly timely.
`Another issue that we examine through regression analysis is the rela-
`tionship between generic entry in a drug category and the perceived
`profitability of that category. In addition, we consider the structure of
`prices and market shares among generic firms in the initial period after
`entry to see whether the lowest-price firms capture the largest shares.
`The article is organized as follows. Section I provides some back-
`ground on the pharmaceutical industry. It also reports on the data that
`we gathered and describes and interprets the stylized facts of pricing
`patterns and market-share results. Section II examines the pioneer's pric-
`ing response to entry in greater detail. Section III is concerned with the
`determinants of generic entry and the general structure of generic prices
`and market shares. Section IV provides a brief summary and concluding
`comments .
`
`I.
`
`PRICE COMPETITION THROUGH GENERIC ENTRY
`
`A. Background
`
`Pharmaceuticals have often been cited in the literature as a particularly
`extreme example of first-mover pricing advantages. Theoretical models
`by Richard Schmalensee2 and Cecilia Conrad3 explain this phenomenon.
`and F. M. Scherer and David Ross4 have provided a recent survey of the
`empirical literature for the pharmaceutical industry. They conclude that,
`“under conditions like those found in pharmaceuticals, first movers have
`natural product differentiation advantages that permit them to charge high
`prices and retain substantial market shares.”5
`Historically, the strong brand loyalty in pharmaceuticals for innovative
`brands over generic competitors has been rooted in several institutional
`considerations. First, physicians generally gained experience with a new
`drug during its period of patent exclusivity. When patents expired and
`
`2 Richard Schmalensee. Product Differentiation Advantages of Pioneering Brands. 72
`Am. Econ. Rev. 349 (I982).
`’ Cecilia A. Conrad. The Advantages of Being First and Competition between Finns. I
`Int‘l. J. Indus. Org. 353 (I983).
`‘ F. M. Scherer & David Ross. Industrial Market Structure and Economic Performance
`(3d ed. B90).
`5 Id. at 592.
`
`000002
`
`000002
`
`

`

`BRAND LOYALTY, ENTRY, AND COMPETITION
`
`333
`
`cheaper substitutes became available, many physicians were insensitive
`to the lower-price opportunities and continued to prescribe the brand
`name product. Furthermore, state antisubstitution laws instituted in the
`19505 and 19605 prohibited pharmacists from substituting the cheaper
`generic brands.‘5
`In addition, there were supply-side constraints on generic entry. Under
`the 1962 Kefauver-Harris Amendments, generic products could not rely
`on the safety and efficacy evidence submitted by pioneers for post-1962
`drug introductions. Unless the relevant data were publicly available in
`the scientific literature, a generic competitor had to duplicate many of
`the pioneer's tests to gain approval. Our earlier article found this had a
`major inhibiting effect on the speed and degree of entry by generic prod-
`ucts.7 An exception was in the antibiotics category, where entry did not
`require duplicative safety and efficacy testing. In contrast to other thera-
`peutic categories, antibiotics has been characterized by rapid entry, sig-
`nificant generic prescriptions, and. strong price competition.“
`An analysis by Meir Statman illustrates the historical brand-loyalty
`advantages possessed by pioneering brand vis-a-vis their generic en-
`trants.9 Statman examined twelve (nonantibiotic) drug compounds whose
`patents expired during the I970s. He found only marginal market-share
`gains by generic entrants, even several years after patent expiration. For
`example, the average pioneer's market share (in revenues) two years
`after patent expiration was 99 percent. (We should note that,
`in eight
`cases, the market share was 100 percent, indicating that no entry took
`place.) Similar results also emerged from other analyses. with only major
`antibiotic products systematically deviating from this observed pattern.
`Since the mid-1970s. however, there have been a number of institu-
`
`tional changes in pharmaceuticals that should increase the degree of price
`sensitivity. First, the state antisubstitution laws have been universally
`repealed. The new substitution laws allow pharmacists to dispense lower-
`priced generics in the place of prescribed brands (subject to physician
`override provisions that vary from state to state). In addition, third-party
`payers such as Medicaid have instituted requirements limiting reimburse—
`ments to generic levels. These programs have been imitated by many
`private insurers. The growth of managed health programs and health
`
`‘ See Henry Grabowski & John Vernon, Substitution Laws and Innovation in the Phar-
`maceutical Industry, 43 L. & Contemp. Probs. 43 (1979).
`7 Grabowski & Vernon, supra note 1, at 195.
`' David Schwartzman, Innovation in the Pharmaceutical Industry (1976).
`’ Meir Statman, The Effect of Patent Expiration on the Market Position of Drugs. in
`Dmgs and Health: Economic Issues and Policy Objectives (Robert B. Helms ed. 1981).
`
`000003
`
`000003
`
`

`

`334
`
`THE JOURNAL or LAW AND ECONOMICS
`
`maintenance organizations (HMOs) have also encouraged generic utili-
`zation.
`
`Alison Masson and Robert Steiner have performed an analysis of the
`initial period after the new state substitution laws.I0 They found that these
`new laws have indeed increased the market’s price sensitivity and the
`amount of generic usage. Nevertheless, they still observe a surprising
`degree of brand loyalty for pioneering brands. In their 1980 sample of
`forty-five drugs, they found that generics accounted for only 23.3 percent
`of prescriptions. Two recent related studies also _find that brand-name
`products retain large market shares relative to lower-priced genen'c en-
`trants.” Both provide an analysis of some of the factors underlying'this
`phenomenon.
`None of the existing literature, however, has focused on the post-1984
`period.'2 As noted above, the passage of the 1984 Drug Price Competition
`Act was a key event in facilitating the entry of generic-drug manufactur-
`ers into the marketplace. As a result of the 1984 Act, generic firms now
`enter the market much more rapidly after patent expiration and enter in
`abundant numbers. The entry situation for all pharmaceuticals now
`closely resembles what formerly held only for antibiotics. Hence, an anal-
`ysis of competition between pioneer and generic products in the post-1984
`period appears warranted.
`
`B. Data Samples
`
`In order to examine the effect of this entry on market prices and shares.
`we have assembled data for a sample of eighteen major products first
`
`to Alison Masson & Robert L. Steiner. Generic Substitution and Prescription Drug Prices:
`Economic Effects of State Drug Product Selection Laws (Stafi' report. Federal Trade Com-
`mission. Bureau of Economics [985).
`“ Mark A. Hurwitz & Richard E. Caves. Persuasion or Information? Promotion and the
`Shares of Brand Name and Generic Pharmaceuticals. 31 J. Law & Econ. 299 ([988); and
`Richard E. Caves. Michael D. Whinston. & Mark A. Hurwitz. Patent Expiration. Entry.
`and Competition in the US Pharmaceutical Industry: An Exploratory Analysis. Brookings
`Papers on Economic Activity: Microeconomics (1991).
`'2 The study by Caves, Whinston & Hurwitz. supra note ll. examined a sample of thirty
`drugs with patent expiration between [976 and 1987. There is a significant overlap with our
`sample. but several of the drugs in their sample experienced initial competition prior to the
`passage of the Drug Price Competition and Patent Restoration Act in September 1984. They
`also employ pooled samples throughout their analysis. thereby generally averaging the
`experiences of the 19705 and early l9805 chronicled by prior researchers (Statman, supra
`note 9; Masson & Steiner. supra note I0). Their very different findings on the post-I984
`period are reported below. There is no real attempt to isolate the effect of the I984 Act
`beyond a general trend variable that is utilized to capture various structural changes during
`the 19805. This variable is insignificant for the drug-store market where the lion's share of
`the sales for the drugs in our sample are made.
`
`000004
`
`000004
`
`

`

`BRAND LOYALTY, ENTRY, AND COMPETITION
`
`335
`
`exposed to generic competition over the period 1983-87. The criterion
`that we used to include a product on this list was fifty million dollars or
`more in sales at the time of patent expiration. Pharmaceutical sales are
`highly skewed in sales value.” Given this, generic competition will be
`directed especially toward the products that have achieved the largest
`markets.
`There were eighteen products that satisified the threshold criterion of
`fifty million dollars in sales at the time of patent eXpiration.” For each
`of the eighteen products, we identified the most popular dosage sizes from
`the data audits of IMS America Inc.” These data sources also provide an
`estimate of total sales of each dosage size by manufacturers and whole-
`salers to drugstore and hospital outlets. Sales are expressed in both dol-
`lars and physical units. Using this information, we computed prices at
`various time intervals for the pioneer products, the generic entrants, and
`the overall market. All of these prices represent the average cost per unit
`paid by drugstores and hospitals for the most frequently consumed dos-
`age size of each product. We also computed market shares for the pioneer
`and generic products in units at the same points in time after entry.
`
`C. Descriptive Statistics
`
`The general pattern is that generic products enter at a significant dis-
`count to the pioneering product with which they compete. Moreover, in
`generic prices, there is a strong downward price dynamic over time. By
`contrast. the prices of the pioneering brands remain higher than their
`generic competitors and actually increase in nominal terms in the time
`period after entry. Average market price,“ however, decreases over time
`as the lower-priced generic products achieve significant gains in market
`share.
`
`Table l provides a summary of our findings for the eighteen drugs. The
`first row indicates that the average market price declined by a little more
`than 10 percent per year in the first two years after generic entry. The
`second row shows that the average pioneer price index rose 7 percent in
`the first year after entry and an additional 4 percent in the second year
`
`'3 Henry Grabowski & John Vernon, A New Look at the Returns and Risks to Pharma-
`ceutical R&D. 36 Mgmt. Sci. 804 (1990).
`" For a list of the products and the date of generic entry, see Table Al in. the Appendix.
`As also shown in Table 1. drugs with patent expiration afier 1984 had entry within the same
`year as or the year immediately after the patent expiration.
`" lMS America Inc.. U.S. Drug Store and Hospital Sales (1983—87) (hereinaner IMS).
`" Average market price refers to total dollars for pioneers and generics divided by total
`units.
`
`000005
`
`000005
`
`

`

`336
`
`THE JOURNAL or LAW AND ECONOMICS
`
`TABLE 1
`
`SUMMARY or GENERIC DRUG FINDINGS
`
`
`Two Years
`One Year
`At Date
`
`of Entry after Entry after Entry
`
`
`Average market price index
`
`Average pioneer price index
`
`Average generic price index
`
`1.0
`
`1.0
`
`1.0
`
`.89
`(.10)
`1.07
`(.07)
`.78
`(.15)
`
`.79
`(.12)
`1.11
`(.10)
`.65
`(.16)
`
`Average ratio of generic price to pioneer
`price
`
`Average generic market share in units
`
`.37
`.46
`.61
`(.13)
`(.14)
`(.11)
`.49
`.35
`.09
`(.11)
`(.12)
`(.09)
`‘ 25.1
`17.2
`N.A.
`Average number of generic suppliers
`
`(3.8) (6.8)
`
`Nora—Each value is an unweighted average of the values for the eighteen drug categories. The price
`indexes take the date-of-entry price as unity. Hence, for example, the average category price two years
`postentry is 79 percent of its value at the date of entry. The market price equals total dollars of sales
`for the leader and generics divided by total units. No attempt has been made to deflate prices for inflation.
`Average generic market share in units at date of entry is the share during the first month of generic
`marketing. Population standard deviations are given in parentheses; N. A. = not applicable.
`
`after entry. At the same time, generic prices fell to 78 percent of their
`initial value at the end of the first year and 65 percent at the end of the
`second year. This steep price decline together with the growing market
`share obtained by the generics is what causes overall market prices to
`decline.
`
`The different behavioral responses of pioneering and generic firms
`causes the gulf in prices between them to increase significantly over time.
`As shown in Table l, generic prices averaged 61 percent of the leader
`price during the first month after entry, and this declined to 46 percent
`one year after entry and then to 37 percent two years after entry.
`The last row in Table 1 shows the movement over time in average
`market share (in units) achieved by the generics. This increases from 9
`percent in the first month after entry to 35 percent one year after entry
`and to 49 percent two years after entry. Furthermore, two years after
`entry, these products average seventeen and twenty-five separate generic
`suppliers, respectively.'7 This is a dramatically higher rate of generic
`
`'7 This refers to the average number of firms supplying products at the wholesale level
`as reflected in the data audits of [MS (r'd.). It should be noted that the average number of
`Abbreviated New Drug Applications (ANDAs) approved by the FDA for these products is
`less than the number of firms marketing the product, so that some firms rely on the ANDAs
`of other firms. This issue is considered further below.
`
`000006
`
`000006
`
`

`

`BRAND LOYALTY, ENTRY, AND COMPETITION
`
`337
`
`penetration than observed by Statman in his study of the 19705.“ As
`noted above, Statman found only marginal gains by the generic entrants
`even several years after patent expiration.
`The results reported in Table 1
`indicate that average market prices
`decline as a result of the significant increase in usage of lower-priced
`generics. In this regard, we computed a descriptive regression of the
`market price index on generic market share, each measured two years
`after generic entry. Letting LMP represent the log of the market price
`index and LGS represent the log of generic market share, the equation
`15
`
`LMP= —.70
`(—11.10)
`
`-—
`
`.6ILGS,
`(—7.54)
`
`where the adjusted R2 = .78, n = l7, and the I-statistics are in paren-
`theses.
`
`The statistical results are strong. There is a negative relationship that
`states that the higher generic penetration is, the lower the market price
`will be. Each 10 percent market-share gain by generics is associated with
`a market-price decline of 6.1 percent two years after entry.
`
`II. THE EFFECT OF GENERIC ENTRY 0N PIONEER PRICES
`
`In this section, we examine the effect of generic entry on the pricing
`pattern of pioneers. The descriptive statistics in Table 1 indicate merely
`that pioneer prices in current dollars rose, on average, by 11 percent
`in the twenty-four months immediately after entry by the first generic
`competitor. Here, we first examine how pioneer prices were changing
`before entry and estimate statistically whether the pattern changed after
`entry. We then compare these results with the patterns for the Producer
`Price Index (PPI) of the therapeutic category to which each of these
`products belongs.
`The data base used for this analysis consists of annual. prices of the
`eighteen pioneer products in drugstores from 1980 to 1990..9 In order to
`have at least five years in the preentry period for all products, the series
`
`" Statman, supra note 9.
`" This analysis focuses on drugstore prices; hospital prices are omitted. Since hospital
`prices are governed by a difl'erent process (contractual bidding), they can follow a difi'erent
`trend line than drugstore prices. Moreover. the observed trend in overall drug prices will
`be influenced by changes in the composition of a firm’s hospital and drugstore sales over
`time. In order to prevent this from confounding changes in the trends associated with
`generic entry. we focus only on the drugstore sector. The dominant component of sales‘for
`the drugs in our sample is through retail pharmacies.
`
`000007
`
`000007
`
`

`

`338
`
`THE JOURNAL or LAW AND ECONOMICS
`
`began in 1978 for one product and in 1979 for four products. We also
`obtained the PPI for the relevant therapeutic categories for this period.20
`
`A . Regression Analysis
`
`We first estimated regressions of each product's price on time and
`allowed for a sl0pe change in the postentry period. Nominal prices were
`deflated by the gross national product implicit price defiator. The specifi-
`cation is given below:
`
`log(Pricc) = a0 + (II Time + azDumT.
`
`Time is the year; DumT is (time — T") if Time is greater than T" and
`zero otherwise; 1* is the year in which generic entry occurs. The variable
`DumT is constructed to insure that the time-trend segments join in T".
`Hence, the annual rate of price change is al in the preentry period and
`(al + (12) in the postentry period.
`The estimated coefficient for the rate of price increase, (1., is positive
`and statistically significant for seventeen of the eighteen products.“ The
`other product’s estimated coefficient for al is negative but not statistically
`significant. The average rate of price increases across all products derived
`from these regressions is 8.4 percent.
`The prevailing pattern for the slope-change coefficient in the postentry
`period is negative. In particular, the estimated coefficient for a2 takes on
`this negative sign for fourteen of the eighteen products.22 Only two of
`these negative coefficients, however, are statistically significant at 10 per-
`cent confidence intervals or better. Hence, the conclusion at this stage is
`that only two products (Haldol and Keflex) qualify as possible cases in
`which generic entry moderates the pioneer’s rate of price increase.
`The next step is to compare the patterns for these two products with
`the pattern for the relevant therapeutic category’s PPI. The comparison
`
`I” The study by Caves, Whinston. & Hurwitz, supra note ll. used a pooled sample
`spanning both the pre- and post-1984 period to estimate the effects of entry on the prices
`of brand-name pioneers. They found a small but statistically significant negative effect. They
`did not attempt to look at differences across products or differentiate between behavior in
`the pre- and post-I984 Act.
`2' While we do not report the actual eighteen regressions here, interested readers can
`obtain tables of the regressions from the authors. We should note that the one case with a
`negative coeflicient was Motrin, which experienced competition from the introduction of
`an over-the-counter version during this period.
`21 Two of the four products with positive coefficients were also statistically significant
`indicating that these products actually showed a price-rate increase in the postentry period.
`These products, however. increased their prices from a base-period rate in the preentry
`period that was well below the mean for the average rate of increase in their therapeutic
`class. Hence. this increase is subject to alternative interpretations.
`
`000008
`
`000008
`
`

`

`BRAND LOYALTY, ENTRY, AND COMPETITION
`
`339
`
`is necessary since a reduced rate of price increase in the postentry period
`may be due to a general competitive change affecting the therapeutic
`category as a whole, and it is desirable to try to distinguish this possibility
`from the specific generic-entry effect.
`For each product, a regression for the therapeutic category PPI (with
`T" corresponding to the product) was estimated. Haldol‘s rate of increase
`was [4.5 percent, falling to 2.5 percent postentry. The deflated PPI for
`psychotherapeutics had rates of 10.8 percent preentry and 9.0 percent
`postentry. Since Haldol's relative decline was substantially greater than
`its therapeutic category, we concluded that generic entry led to the mod-
`eration in Haldol‘s rate of price increase. We concluded the same for
`Keflex. Its rate of price increase was 6.6 percent preentry and only 4.4
`percent postentry. By comparison,
`its
`therapeutic category (anti-
`infectives) had a rate of price increase of 4.3 percent both pre- and post-
`entry.
`
`Overall, the effect of generic entry on pioneer pricing is not very sig-
`nificant in economic terms. There is no indication, for example, that any
`pioneers decreased their nominal prices in response to the much-lower
`generic prices or attempted any kind of entry-deterring price strategies.
`Recall from Table 1 that the ratio of generic to pioneer price averages
`.37 two years postentry. Furthermore, the highest ratio for any drug in
`the sample is only .58. Also, the two products that exhibited a statistically
`significant decrease from prior trends continued to increase their real
`prices by at least 2.5 percent per year.
`
`B.
`
`Interpretation of Findings
`
`Rather than attempting to match the prices of the lower-priced gener~
`ics, the originators have continued to increase prices at an average rate
`that exceeds general inflation. It may seem surprising that, in the face of
`the high rates of generic entry, these firms have generally chosen not to
`decrease prices. The literature contains some examples of incumbent
`firms that respond to entry by cutting prices and other examples where
`they keep prices unchanged. It is, perhaps, the most common view, how-
`ever, that the incumbents will be forced to lower their prices if entrants '
`enter with especially low prices.
`The rationality of pharmaceutical companies not matching lower ge-
`neric prices can be understood in terms of a segmented-market model.23
`
`21 A recent theoretical analysis using a segmented market model of pharmaceutical pricing
`is presented in Richard G. Frank & David Salkever, Pricing, Patent Loss, and the Market
`for Pharmaceuticals (Working Paper No. 3803, Nat'l Bur. Economic Research, August
`I991).
`
`000009
`
`000009
`
`

`

`340
`
`THE JOURNAL or LAW AND ECONOMICS
`
`In particular. it is reasonable to assume that the market is segmented
`between two groups—a price-sensitive group and a group that has strong
`brand loyalties to the incumbent. The strength of brand loyalties is dem-
`onstrated by the fact that, on average, pioneers keep about half their
`market in units despite the fact that generics are roughly one-third the
`price of pioneers (measured two years after entry).24 If the pioneer can
`keep half the market at the unchanged price as opposed to keeping the
`whole market by lowering the price to one-third its original level, simple
`calculations show the wisdom of choosing the unchanged-price strategy.”
`One qualification to these results should be noted: our eighteen drug
`products are all oral prescription drugs sold primarily through retail phar-
`macies to outpatients rather than to hospitals or other institutions. As
`part of a case involving a contested patent, one of the authors examined
`the data on pricing patterns for a few injectable products first experienc-
`ing generic competition in the period 1986-88. Injectables are sold almost
`exclusively to hospitals where the degree of brand loyalty is considerably
`less. In this case, it was found that the more prevalent pattern is for the
`incumbent firm to cut prices in the face of generic competition, although
`not to the lowest level of the generic entrants.26 These contrasting pat-
`terns for outpatient and inpatient drugs show that pricing policies will be
`influenced by the relative sizes of the different market segments. If the
`price-sensitive segment of the outpatient market were to continue to ex-
`pand in relative terms in the case of future patent expirations, then the
`
`2‘ Here. of course, we are referring to the ratio of the wholesale price of generics to
`pioneers—as reported in Table I. In fact, the relevant price ratio to the ultimate consumer
`is the retail—drug-store ratio. For our sample. the retail price ratio of generics to pioneers
`averages 65 percent compared to 45 percent at the wholesale level (measured one year
`postentry). This suggests that drugstores have higher markups for generics than on the
`pioneers. In a companion paper that we are in the process of writing. we have analyzed the
`margins for the generic and pioneer products at retail drugstores. Absolute margins for
`generics are higher than for pioneers, despite their lower cost to the drugstores. This means
`that price reductions at the manufacturer level stimulated by the I984 Act are not fully
`passed on to the ultimate consumers.
`2’ Using our average estimate of marginal cost at the date of entry as 25 percent of price
`(from Section III) and assuming that overall market demand is perfectly inelastic, profits
`will fall by 89 percent with price cutting and by only 50 percent with the price unchanged.
`Of course. this type of calculation is meant merely to illustrate the strong incentives for
`incumbents with significant brand loyalty to pursue a "market-harvesting" rather than a
`price-cutting strategy. A more extensive analysis would examine the effect of the pioneer‘s
`pricing policies on the rate of generic entry. Some results on this issue are presented in
`Section III.
`
`as The selection criteria for this admittedly small sample involved products with annual
`sales of at least ten million dollars (at the time of entry) and no close substitute or modified
`version of the product with patent protection. It was found that four of the five incumbent
`firms decreased their price in the period after generic entry. The average price decline for
`these branded products was 14 percent one year after initial entry. The sample of products
`was calcium leucorovinm, cytarabine, vancomycin, clindamycin. and methlydopa.
`
`000010
`
`000010
`
`

`

`BRAND LOYALTY, ENTRY, AND COMPETITION
`
`34]
`
`pricing policies for incumbent brands are likely to move in the direction
`of those observed for the hospital market.
`Finally,
`it
`is important
`to observe that firms can employ product-
`differentiation strategies to thwart genen‘c competition. In at least four
`cases (Catapres, Inderal, Calan/Isoptin, and Indocin), a significant vari-
`ant of the original product was introduced around the date of generic
`entry. The most successful variation was a “slow—release" dosage form
`of Calan/Isoptin, introduced in December 1986. Within two years. the
`slow-release version accounted for 81 percent of pioneer sales. Hence,
`shifting patients from the original formulation, subject to severe price
`competition, to a new formulation, insulated from price competition, is
`apparently an important strategy that can be employed at least in some
`instances.27 This is an issue that warrants further research.
`
`III.
`
`THE DETERMINANTS OF GENERIC ENTRY AND PRICES
`
`As observed above, generic prices in our sample move on a rapid
`downward dynamic path over time. These markets have apparently been
`characterized by easy entry conditions, at least since 1984. Furthermore,
`over the long run. one could reasonably expect that entry and competition
`would push generic drugs prices down to a level approaching the marginal
`cost of production.
`
`A. Determinants of Entry
`
`In this section, we will examine the factors influencing generic entry
`and prices. In a number of economic models in which a dominant firm
`experiences entry from a competitive fringe, it is assumed that the rate
`of entry will be a function of the difference between the dominant firm's
`price and the potential entrants minimum unit cost.” In other words, the
`rate of entry will increase with the entrant's profit opportunities, but it
`also can be retarded by barriers producing brand loyalty for the dominant
`firm's product.
`We will test whether the degree of entry by the generic firms in our
`eighteen-drug sample can be explained by such a relationship. In particu-
`lar, we estimate the following equation:
`
`Xl,l=l = flflu=0v ELit-=0),
`
`(I)
`
`17 All four modified versions of these products obtained patent protection utilizing the
`new mode of delivery as the basis. In addition. some obtained parallel market-exclusivity
`rights as a new dosage form under the I984 Act. See the US. Food and Drug Administration
`Approved Product List (annual) ("the Orange Book").
`2' See. for example. Darius W. Gaskins. Jr., Dynamic Limit Pricing: Optimal Pricing
`under Threat of Entry. 3 J. Econ. Theory 306 (197]).
`
`000011
`
`000011
`
`

`

`342
`
`where
`
`THE JOURNAL or LAW AND ECONOMICS
`
`X“.1 = number of generic suppliers for product i at the end of first
`year, I = I;
`11“,,0 = profitability to generics of entering the market for product
`i at time of initial entry, t = 0; and
`Film) = a set of brand loyalty entry barrier variables for product i
`at time of initial entry.
`
`is further defined as the percentage
`The profitability variable, 11,._,,0,
`markup of pioneer price over marginal cost at the point of initial entry.
`In symbolic terms one has
`
`1'l'r.i=o = (PPi,r=o " Mci)/PPI,I=Or
`
`(2)
`
`where
`
`PP,"=0 = pioneer's price at time I = 0; and
`MC,- = marginal cost of production for product i.
`
`A key issue is how to measure the marginal cost of each drug (MC) in
`order to compute their profitability. Our basic procedure was to make
`the assumption that competition would eventually force the generic price
`down to marginal cost and to estimate marginal cost as the asymptote
`toward which the generic price was falling?9
`Figure 1 illustrates our approach by plotting the trend in generic prices
`for the drug Lorazepam. Generics began competing with Lorazepam in
`September I985. The average generic-price data is plotted at six-month
`intervals'running from 1985 through mid-1988. Figure 1 also shows a
`regression of the generic price for Lorazepam on the reciprocal of time.
`The estimated equation is
`
`GP, = 5.34 + 175.3 (1/1),
`(2.89)
`(3.91)
`
`(3)
`
`where the adjusted R2 = 0.74. Hence, for Lorazepam, the intercept indi-
`cated a long-run marginal cost of 5.34 cents per pill. This is the value
`that we used for MC for this product.” This estimated value for marginal
`cost is 21 percent of the pioneer’s price at initial entry.
`
`1’ In order to utilize this approach, one needs s