Tracy Chou for The 2004 Moffatt Prize in Economics

The Drug Price Competition and Patent Term Restoration Act of 1984, also known as the Waxman-Hatch Act, was established to restore effective patent terms, which had eroded substantially over the years due to more complicated and time-consuming approval procedures, and improve generic competition by eliminating the entry barrier of duplicative testing required for generic substitutes. Under the new legislation, generic entrants only needed to submit an Abbreviated New Drug Application (ANDA) that demonstrates the bioequivalence of its drug to the original. The innovators were able to recoup part of their loss caused by patent regulatory delay after 1962. The introduction of Waxman-Hatch act, accompanied by the expiration of patents on many high-sales-volume brand name drugs, has altered the competitive dynamics of the pharmaceutical marketplace.

During 1980s the share of generic prescription drugs sold by retail pharmacies roughly doubled. However, the prices for brand name drugs rose in response to generic entry. Such a price effect diverges from the standard analysis of dominant-firm-competitive-fringe model, and entirely negates the purpose of competition. Many economists have been intensely studying the market structure of pharmaceutical industry in attempt to explain this surprising phenomenon. My obligation in this paper is to gather results from past researches, and modify the standard dominant-firm-competitive-fringe model with calculations from Caves et al. (1991), and mainly from Frank & Salkever (1992) in order to better represent the case.

Caves, Whinston and Hurwitz (1991) studied thirty drugs that went off patent protection between 1976 and 1987. Of the thirty drugs, the patents of seven expired before 1980, sixteen between 1980 and 1984, and seven after 1984. The data in Table 1 covers several major therapeutic drug classes including, cardiovascular (11), psychotherapeutic (7), systemic anti-infective (4), diabetes therapy (4), antiarthritics (2), diuretics (1), and antispasmodics (1). The bottom columns of table 1 provide the average cumulative number of generic ANDAs approved during a given number of years after the patents expired, and the market share of generic products in the same period.

Figure 1: Pharmaceutical Prices, Labor Costs, and Material Input Costs

Table1: Average Cumulative Number of Approved Generic Producers by Years Patent Expired and Number of Years after Patent Expiration

Grabowski and Vernon (1992) later investigated the effect of generic entry on prices for 18 high-sales-volume pharmaceutical products that were first exposed to generic competition during the years of 1983 through 1987. For each drug, the authors examined prices prior to entry and prices one year following generic entry. They concluded that increasing the number of generic entrants had a strong downward effect on generic prices relative to brand-name prices; in particular, the generic-brand name price ratio was estimated to fall from 0.599 to 0.201 as the number of generic entrants increases from 1 to 20.

Table 2: Price Ratio

G = Generic, B =Brand name

Figure 2: Graphic Interpretation of Table 2

Year

To concur with Grabowski and Vernon's results, Frank and Salkever (1997) found that brand name prices increased more quickly than in markets without generic entry. Their explanation was that as the price-sensitive consumers switch to generic drugs, demand for the original brand-name drug declines and becomes less sensitive to price. Thus, the price of a brand name drug can theoretically rise more quickly over time than it would have without generic competition.

3.1. Background

Frank and Salkever developed a model using past empirical data, which theorizes an apparent correlation between generic entry and brand name price levels. In their analysis, the demand side of the market for a particular drug is differentiated by two types of consumers: those who face high switching costs and exhibit strong preferences for the brand name drug, and those who are more price-sensitive and exhibit negligible switching costs. The insensitive component might consist of the patients of physicians in fee-for-service practices; those consumers pay high insurance premiums each period, and enjoy a complete coverage on most of the medical resources. And the sensitive component consists of enrollees in HMOs and similar health insurance services. In this model, consumers are risk-averse and tend to reuse the medicines that suit them with complete safety instead of taking a gamble on trying new drugs. Entry of generics will capture a large part of the price sensitive buyers, and leave all price insensitive buyers to purchase brand-name products. This causes the brand name producers' demand function to shift inward and to become less elastic, which allows the profit-maximizing brand name firms to raise its price.

The model assumes that the brand name drug producer behaves as a Stackelberg price leader, in which the leader obtains first mover advantage from patent protection in the first period. The brand name drug producer becomes the dominant firm that makes its profit-maximizing pricing decisions taking into account the generic reaction in the market. Producers of generic substitutes are viewed as fringe firms playing a Nash-Bertrand non-cooperative game, and take the brand name producer's price as given.

3.2. A simple Market Segmentation Model

The profit-maximizing brand name producers face the segmented demand curve: the loyal customers (D_L), and the cross-price-sensitive customers (D_S), who will switch to generic substitutes in the second period. The brand name producer's demand function is:

Q_b = D_L(P_b) + D_S (P_b, P_g) (1)

Where Q_b is the brand name quantity demanded, P_b and P_g are the brand name and generic prices. The brand-name demand response consists of the direct effect on the two segments of the demand function plus an indirect effect of price reaction function determined by the sub-market for generics. For given values of n, as the number of generic firm in the market, [D_L(P_b) + D_S (P_b, P_g)] will be viewed as the reduced-form demand curve for brand name drugs. And the demand function for the n identical generic firms is D_G(P_g, P_b), where P_g = P_g^*(n, P_b), P_b = P_b(n, w), and w is a vector of input price. Cost is a function of Q_b, thus brand name firm's profit function becomes:

_{(P ´ Qb) = Revenue - Cost}

p = P_b{D_L(P_b) + D_s [P_b, P_g^*(n, P_b)]} - C(Q_b) (2)

For given value of n, D_L(P_b) + D_s [P_b, P_g^*(n, P_b)] is the reduced-form demand curve for the brand name producer. Then, the brand name firm maximizes profit with respect to P_b, this yield the first order condition

_{D in DL + D in DS + Cross price effect = ¶Qb/¶Pb Markup}

¶p/¶P_b = 0 = [dD_L/dP_b + ¶D_S/¶P_b + (¶D_S/¶P_g)(¶P_g^*/¶P_b)][P_b - (dC/dQ_b)] + Q_b (3)

Note that the first term summarizes the demand response to change in P_b, which must be negative due to the downward sloping assumption of demand curves, where price and quantity have inverse relationship.

Figure 3: Period 1

Brand name drug producers do not face entry in the beginning, so the first period graph looks just like the classic monopoly model, where a brand name firm produces at MC = MR. The market segmentation line demarcates the loyal consumer (D_L), who occupy the top half of the demand curve due to their higher willingness to pay, from the price-sensitive consumers (D_S) that reside in the lower bottom. The segmentation line does not interfere with profit maximization at this stage, the brand name firm continue to price at P_b as indicated in figure 3.

The effect of generic entry on brand name price (dP_n/dn) can be isolated by mapping the total differentiation of (3) first, which is the following, with SOC equals to ¶²p/¶P_b² < 0 to ensure the concavity of profit function when maximized subject to price:

0 = [dP_b (SOC) + dn (P_b - (dC/dQ_b)] (4)

[(¶²D_s/¶P_b¶P_b)(¶P_g^*/¶P_b) + (¶²D_s/¶P_g²)(¶P_g^*/¶n)(¶P_g^*/¶P_b) +

(¶D_S/¶P_g)(¶²P_g^*/¶P_b¶n) + dn(¶D_S/¶P_g)(¶P_g^*/¶n)] -

[dn (d²C/dQ_b²) (¶D_S/¶P_g)(¶P_g^*/¶n)] [dD_L/dP_b + ¶D_s/¶P_b + (¶D_S/¶P_b)]

Then rearrange (4) in order to examine the condition of market entry on name brand price (dP_b/dn). Express dP_b/dn as following:

_{Markup > 0}

dP_b/dn ={[P_b - (dC/dQ_b)] (5)

_{Effect of generic entry on the slope of the reduced form demand curve}

[(¶²D_S/¶P_b¶P_g)(¶P_g^*/¶n) + (¶²D_s/¶P_g²)(¶P_g^*/¶n)(¶P_g^*/¶P_b) + (¶D_S/¶P_g)(¶²P_g^*/¶P_b¶n)]

^{< 0 < 0 < 0 < 0 substitutes > 0 > 0 > 0}

+ [(¶D_s/¶P_g)(¶P_g^*/¶n)] -

^{> 0 < 0}_{> 0 < 0 [ ¶Qb/¶Pb < 0 ] > 0}

[(d²C/dQ_b²) (¶D_s/¶P_g)(¶P_g^*/¶n)][dD_L/dP_b + ¶D_s/¶P_b + (¶D_S/¶P_b)]} / -SOC

Since generic firms in the model are by assumption competitive fringe, its price falls and approaches to marginal cost, when the number of generic firms (n) increases, that is (¶P_g^*/¶n) < 0. And (¶D_s/¶P_g) > 0, because generic products are, again, by assumption gross substitutes of the brand name products. So, as Frank & Salkever concluded in their paper that dP_n/dn < 0 unless either 1) entry increases the demand for brand name drugs (¶²D_s/¶P_g² < 0, that is ¶D_s/¶P_g < 0, but ¶P_g/¶n < 0), 2) marginal costs are decreasing for the brand name drugs (d²C/dQ_b² < 0), or 3) entry makes the reduced-form demand curve less elastic (¶²P_g^*/¶P_b¶n > 0, i.e. (¶²P_g/¶P_b²)(dP_b/dn) > 0, and (¶²P_g/¶P_b²) > 0, because (¶P_g/¶P_b) > 0, so we want (dP_b/dn) > 0, too). 1) seems implausible since it would require generic price to rise with entry (¶P_g/¶n > 0) or brand name demand falls when generic prices rises (¶D_s/¶P_g < 0), which implies that the two products are gross compliment in demand. This second possibility was rejected because Frank & Salkever claimed that "little systematic empirical work on the nature of returns to scale has been reported". And if the pharmaceutical industry has decreasing marginal cost, the industry will come close to a natural monopoly, which is seldom the case in reality. This leaves the third possibility: entry makes the reduced-form demand curve steeper, as the most plausible explanation for dP_b/dn > 0. To show this on a graph:

Figure 4: Period 2

As generic firms begin to enter the market in the second period after the patent expires, the generic demand (D_g) takes away part of brand name firm's consumers and causes the demand for brand name products become more inelastic according to Frank & Salkever's calculation, so the price sensitive part of the reduced-form demand curve (D_s) pivots in (D_s'). As a result, the brand name producer maximizes its second period profit with respect to the new marginal revenue curve (MR'). And because now the brand name firm is facing a more inelastic demand than the first period, it will supply at higher price (P_b') to extract consumer surplus.

Frank & Salkever's analysis deviates from the classic dominant-firm-competitive-fringe model because of the clear demand market segmentations that appear in pharmaceutical markets, where each category of drugs have a group of loyal customers with high switching costs, and the first mover advantage provided by patent protection. To fully observe the influence of high switching costs in the market, I need to modify the classic model a bit. If I relax the market-clearing price assumption to allow the dominant firm and competitive fringes to charge different prices; and I impose a uniform marginal cost curve for all, instead of the classic interpretation that the dominant firm obtains its semi-monopolistic position due to efficiency, then we can see the impact of "loyal consumers".

Figure 5: Modified Dominant-Firm-Competitive-Fringe Model

This modified version shows that the price of a brand name drug will definitely decrease upon generic entry, because generic firms will capture the upper part of demand, therefore reducing the consumers' willingness to pay at high prices. Switching costs in the market segmentation model are created mostly by fee-for-service health insurance policies, which provide complete medical coverage once the customers pay a high premium. Consumers that enroll in such insurance policies will inevitably have less incentive to choose a cheaper generic substitute, and their behaviors will be dictated by moral hazard, which results in wasting social resources and impedes market efficiency.

3.3 Effect of Generic Entry on Brand Name Advertising

As an extension of the basic model, Frank & Salkever advanced their research in exploring the decline of brand name advertising expenditure, which had presumably been linked to generic entry according to many past empirical evidences. Advertising generally serves two purposes. Most informational advertising aims to improve consumer knowledge, thus increasing demand for a class of pharmaceuticals, or increasing the demand for a specific product if there exist true quality differentials that can be publicized. However, the brand name manufacturers are well aware that if they advertise to increase the aggregate demand for the class of drug, the advertisement might create spill over effect, that is, the brand name advertising might attract consumers for the generic entrants, who would share the benefit without incurring any cost. The brand name firm is basically incapable of increasing the demand specifically for its own product, because generic drugs are by assumption perfect substitutes with bioquivalency approvals. If the brand name manufacturers advertise primarily to differentiate their products from generic substitutes, it may trigger an opposite effect; it may diminish the cross-price sensitivity of its own demand (D_S), and therefore reduce the size of the demand response to entry. In figure 6, notice the new market segmentation line is moved towards the right, because the price sensitivity is being reduced by informational advertising, and now, the price-sensitive segment of consumers becomes less substantial. In fact, these price-sensitive consumers did not cause any changes on MR within the relevant range, because MR' is below the axis. The brand name firm will still maximize its profit with respect to the old marginal revenue curve, but with higher a marginal cost curve generated by advertising expenditure, such as packaging.

Figure 6:Effect of Informational Advertising

Lower levels of advertising reflect drug innovators' concerns about future lower returns from such investments once patents expire and generic entry grows likely.

4. Conclusion

The entry effect on brand name pharmaceutical products in this analysis is based on the assumption that brand name producers follow an optimal behavior principal, in which the firms profit-maximize with demand curves subject to generic entry. The necessary conditions to cause brand name prices to increase simultaneously as advertising expenditures to decrease in response to generic entry are: 1) the entry leads to a substantial decline in price elasticity of reduce-form brand name demand, and 2) market-expanding advertising reduces the cross-price-sensitive segment of brand name demand curve. Such an outcome leads to several observations regarding the pricing pattern in this market. First, it appears that as the number of generic entries increases, the competition among generic drug producers becomes fierce, and eventually forces generic drug prices to approach the market efficient point. Second, increasing generic competition is not necessarily accompanied by lower prices in brand name drugs.

However, the Drug Price Competition and Patent Term Restoration of 1984 did not completely fail its expectations. Evidence has shown a 40 to 50% shift in market share from brand name producers to generic firms, along with a 25 to 30% reduction in generic drug price. That means, even though the price of brand name drugs rises when their generic substitutes are introduced, the average price of the prescription drugs fall, and the trend of name-brand price increase will be attenuated as the cross-price-sensitive segment of the market continues to expand.

Thus, the moral of the story is not to condemn generic competition, but to elucidate switching costs in consumer behavior as a form of market failure caused by imperfect information. Switching costs are incurred when consumers are ignorant of the bioequivalency of generic substitutes. Policy makers should focus on eroding switching costs by educating consumers.

Bibliography:

Aronsson, Thomas, Mats A. Bergmand, Niklas Rudholm. (1997), "The Impact of Generic Competition on Brand Name Market Shares - Evidence from Micro Data", Working Paper Series in Economics and Finance, Department of Economics, University of Ume, Sweden.

Caves, Richard E., Michael D. Whinston, Mark A. Hurwitz, (1992), "Patent Expiration, Entry and Competition in the U.S. Pharmaceutical Industry: An Exploratory Analysis", Brookings Papers on Economic Activity. Microeconomics, Vol. 1991. (1991), pp. 1-48.

Frank, Richard G., David S. Salkever. (1992) "Pricing, Patent Loss and the Market for Pharmaceuticals", Southern Economic Journal, 59, 165-179.

-- (1997), "Generic Entry and the Pricing for Pharmaceuticals", Journal of Economics and Management Strategy, 6(1), 75-90.

Grabowski, Henry, John Vernon, (1992), "Brand Loyalty, Entry and Price Competition in Pharmaceuticals After the 1984 Drug Act ", Journal of Law and Economics, 35, 331-350.

Klemperer, Paul. (1995), "Competition when Consumers have Switching Costs: An Overview with Applications to Industrial Organization, Macroeconomics, and International trade", Review of Economics Studies 62, 515-539

Merino-Castelló, Anna (2000), "The Impact of the Reference Price System on the Pharmaceutical Market: A theoretical Approach", Economics Working Papers, Department of Economics and Business, Universitat Pompeu Fabra.

If you'd like to leave comments about this entry, use the contest feedback form. Make sure to indicate that you are commenting on Tracy Chou's "An Analysis on Waxman-Hatch Act: The Effect of Generic Entry on Pharmaceutical Prices".