• unlimited access with print and download
    $ 37 00
  • read full document, no print or download, expires after 72 hours
    $ 4 99
More info
Unlimited access including download and printing, plus availability for reading and annotating in your in your Udini library.
  • Access to this article in your Udini library for 72 hours from purchase.
  • The article will not be available for download or print.
  • Upgrade to the full version of this document at a reduced price.
  • Your trial access payment is credited when purchasing the full version.
Buy
Continue searching

Three essays on pricing and inequality in pharmaceutical market

Dissertation
Author: Jie Chen
Abstract:
My thesis examines the pharmaceutical market from three different angles. The first two chapters examine the optimal pricing strategies from the pharmaceutical companies' and insurance companies' point of view respectively. In the last chapter, I analyze the racial and ethnic disparities in the drug use. Each of these essays is summarized below. The first study provides a model that demonstrates the interplay between quality and product differentiation in determining the optimal pricing strategy. Specifically, higher (lower) quality products will engage in price skimming (penetration) strategies in markets where products are sufficiently differentiated, but will choose a market penetration (skimming) strategy in markets that are less differentiated. I tested this model using a unique database that combines information on drug price and quality for antidepressant drugs during the years 1999-2002. The results indicate that higher quality antidepressants engage in a market penetration strategy, charging initially lower prices that rise over time. The second study examines how pharmaceutical costs are shared among consumers and insurers and how drug quality affects these costs. I provide a model which delineates the tradeoff between paying more for higher quality drugs to reduce future medical costs in determining the optimal copayment strategy for the third party payers. I test the model using two large drug therapeutic classes: brand name antidepressants and non-steroidal anti inflammatory drugs (NSAIDs). These two drug classes are interesting to study because they differ in the degree of variation in product quality. While there is little quality differentiation among the antidepressants studied, quality varies by more among the NSAIDS; hence, quality differences are more readily discernible. The results indicate that consumers' out-of-pocket payments are larger for high quality antidepressants, while insurers pay less for these drugs. In contrast, insurers share the drug cost together with the consumers for higher quality NSAIDs. In the third chapter, I seek to determine the extent to which disparities reflect differences in observable population characteristics versus heterogeneity across racial and ethnic groups in antidepressant drug use employing Blinder-Oaxaca decomposition technique. Using Medical Expenditure Panel Survey (MEPS) data from 1996-2003, I estimate individual out-of-pocket payments, total prescription drug expenditures, drug utilization, the probability of taking generic versus brand name antidepressants, and the share of drugs that are older, lower quality types of antidepressants (e.g., TCAs and MAOIs) for Caucasian, African American, and Hispanic individuals. I find that substantive racial and ethnic disparities exist in all dimensions of antidepressant drug use examined. Observable population characteristics account for most of the differences in drug expenditures, with health insurance and education key factors driving differences in spending. Observable characteristics are also important in explaining racial and ethnic disparities in the probability of purchasing generics and in the quality of antidepressant drugs used. In contrast, differences in total utilization are not well-explained by observable characteristics, and may reflect unobserved differences in knowledge and cultural factors, which tells us that to limit differences in overall antidepressant drug use, policymakers must take into account unobserved heterogeneities, such as discrimination, knowledge, cultural background, etc. Thus, differences in observable characteristics (notably health insurance and education) explain racial and ethnic disparities in expenditures and patterns of use (e.g., brand vs. generic), but not disparities in total utilization.

vi Table of Contents

List of Figures..............................................................................................viii

List of Tables..................................................................................................ix

Acknowledgements ..........................................................................................x

Introduction ......................................................................................................1 Chapter One ......................................................................................................................7 1.1 Introduction............................................................................................................9 1.2 Previous Work.....................................................................................................13 1.3 The Market for Antidepressant Drugs.................................................................17 1.4 Conceptual Model................................................................................................20 1.4.1 Two Period Model....................................................................................20 1.4.2 Pricing Dynamics......................................................................................27 1.4.3 Summary...................................................................................................28 1.5 Data and Estimation.............................................................................................29 1.5.1 Data...........................................................................................................29 1.5.2 Estimation.................................................................................................31 1.5.3 Antidepressants Revisited.........................................................................34 1.6 Results..................................................................................................................34 1.6.1 Quality and Entry Price.............................................................................34 1.6.2 Predicted Price..........................................................................................36 1.7 Conclusion...........................................................................................................40

vii Chapter Two..................................................................................................51 2.1 Itroduction............................................................................................................53 2.2 Previous work......................................................................................................56 2.3 The Markets for NSAIDS and Antidepressant Drugs .......................................58 2.3.1 NSAIDs.....................................................................................................58 2.3.2 Antidepressant Drugs................................................................................60 2.4 Conceptual Model................................................................................................61 2.5 Data and Estimation.............................................................................................67 2.5.1 Data...........................................................................................................67 2.5.2 Estimation.................................................................................................68 2.6 Results..................................................................................................................71 2.6.1 NSAIDs.....................................................................................................71 2.6.2 Antidepressants.........................................................................................72 2.7 Conclusion...........................................................................................................73

Chapter Three................................................................................................80 3.1 Introduction..........................................................................................................83 3.2 Background and Literature Review.....................................................................86 3.3 Data and Method..................................................................................................89 3.3.1 Data...........................................................................................................89 3.3.2 Method......................................................................................................91 3.4 Results..................................................................................................................95 3.4.1 Descriptive Analysis.................................................................................95 3.4.2 Decomposition Results.............................................................................96 3.5 Discussion..........................................................................................................101

Bibliography................................................................................................110

viii List of Figures

1.1 Predicted Prices of Antidepressants for Each Drug from 0 to 10 Market Years.......49

ix List of Tables

1.1 Antidepressants Summary Statistics.........................................................................41 1.2 Alternative Market Pricing Strategies.......................................................................42 1.3 Antidepressants: Variable Names, Description, and Summary Statistics.................43 1.4 Brand Name Drug Quality and Real Drug Prices with Dependent Variab le............45 1.5 Predicted Entry Real Price of Antidepressants Com pared to the Existing Drugs....47 1.A1 Quality Indicators of Antidepressants.......................................................................55 2.1A Antidepressants Summary Statistics.........................................................................75 2.1B Nsaid Summary Statistics..........................................................................................75 2.2 NSAID & Antidepressants Variable Nam es, Description, and Summary Statistics 76 2.3 NSAIDs & Antidepressants: Dependent Variable....................................................78 2.A1 Quality Indicators of Antidepressants.......................................................................85 2.A2 Quality Indicators of NSAIDs...................................................................................85 3.1 Summary Statistics (Means, standard deviation in the parenthesis).......................105 3.2 Decomposition Results for Caucasians and Hispanics...........................................107 3.3 Decomposition Results for Caucasians and African-Americans............................102 3.4 Summary of Main Factors to Increase the G ap in the Following Five Outcomes..109

Acknowledgments

I would like to thank all the people w ho have helped and inspired me during my doctoral study. I especially would like to express my deep and sincere gratitude to m y advisor, Professor John Rizzo. He sets an example of a world-class researcher for his rigor and passion on research. His perpetual energy and enthusiasm in research always motivates me and will motivate me forever. His wide knowledge and his logical way of thinking have been of great value for me. His understanding, support, and encouraging guidance not only help me finish this thesis but also help me to be a good researcher. I could not have imagined having a better advisor and mentor for my PhD. Deepest gratitude is also due to Professor Debra Dwyer, who always been a constant source of encouragem ent and giving me stimulating suggestions and constructive comments during my graduate study. I also would like to thank Professor Wei Tan, who is abundantly helpful and offers invaluable assistance, support and guidance. I am indebted to my colleagues for providing a stim ulating and fun environment in which to learn and grow. I am especially grateful to Hai Fang, Yang Xie, Hong Liu, Song Gao, Miranda, Yanan, Mir, Partha, and Huan Ni. Last, and most importantly, I wish to thank m y family for their unflagging love and support throughout my life. Specially, I owe my loving thanks to my husband Zongheng. Without his encouragement and understanding, it would have been impossible for me to finish this work. Best wishes to my two-month old son, William. Hope him healthy and happy forever!

1

Introduction Pharmaceuticals account for a significant share of total health care expenditures in the United States. In 2003, Americans spent $179.2 billion on prescription drugs, almost 4.5 times the $40.3 billion spent in 1990. Importantly, the rate of spending growth for prescription medications outpaced other areas of medical care from 1995 to 2003. During the period from 1995 to 2002, pharmaceutical manufacturing was the most profitable industry in the United States (Kaiser 2005). The rapid increase in pharmaceutical spending has generated both interest and controversy over drug pricing. Yet, there is little theoretical work or empirical evidence on pharmaceutical entry-pricing strategies or on the time paths of the prices of new entrants versus incumbents. In my first chapter, I am going to explore the entry pricing strategies of pharmaceutical manufacturers. This study provides a model that demonstrates the interplay between quality and product differentiation in determining the optimal pricing strategy. In particular, higher (lower) quality products will engage in price skimming (market penetration) strategies in markets where products are sufficiently differentiated, but will choose a market penetration (price skimming) strategy in markets that are less differentiated. The reason is that in less differentiated markets, price competition is more important than quality competition for gaining market share and higher-quality firms are better able to garner greater market share through price cuts.

2 I test this model using a unique database that combines information on drug price and quality for antidepressant drugs during the years 1999-2002. Based on quality perceptions by physicians, the antidepressant market is a relatively homogeneous drug class, which consisted largely of selective serotonin reuptake inhibitors (SSRIs). The results indicate that higher quality antidepressants engage in a market penetration strategy, charging initially lower prices that rise over time. In recent years, pharmaceutical costs have risen at double digit rates and have outpaced other medical care services such as physician services and hospital care. In 2005, expenditures for prescription drugs were $200.6 billion, almost five times larger than the $40.3 billion spent in 1990 (Kaiser 2007). There are a number of mechanisms designed to control the rapid growth in pharmaceutical expenditures. Cost sharing is one such mechanism, by forcing consumers to shoulder some fixed (copay) or proportionate (coinsurance) payments for each drug prescription. The most popular method of cost sharing is the use of multiple-tiered copayments. Most previous work has focused on the effects of cost sharing on pharmaceutical utilization. But there is little research on how copayments are determined, and how insurers and consumers share in the cost of pharmaceuticals. My second chapter seeks to bridge these gaps in the literature. To my knowledge, this is the first study to examine how pharmaceutical costs are shared among consumers and insurers. I also consider drug quality as one of the key factors affecting consumer and insurer payments.

3 Although cost sharing may control pharmaceutical costs by increasing consumers’ price sensitivity, it has also been shown to have unintended effects in terms of increasing health care utilization, such as hospital admissions and emergency room visits (Tamblyn 2001, Winkelmann 2004, Gibson, McLaughlin et al. 2001 etc.). Given this tradeoff, how do third party payers decide on the copayment structure that balances the current pharmaceutical costs and future health care utilization when they are offering drugs with different quality levels?

These are the issues I address. This study provides a model which delineates the tradeoff between paying more for higher quality drugs to reduce future medical costs in determining the optimal copayment strategy for the third party payers. In particular, if insurers believe they can save on future medical cost by offering the drug at a lower copayment, they are more likely to do so to encourage consumers to take the drug. Otherwise, they will charge a higher copay to reduce their current pharmaceutical costs. The model is tested using two large drug therapeutic classes: brand name antidepressants and non-steroidal anti inflammatory drugs (NSAIDs). These two drug classes are interesting to study because they differ in the degree of variation in product quality. While there is little quality differentiation among the antidepressants studied, quality varies by more among the NSAIDS; hence, quality differences are more readily discernible. The results indicate that consumers’ out-of-pocket payments are larger for high quality antidepressants, while insurers pay less for these drugs. In contrast, for the higher quality NSAIDs, insurers share the drug cost together with the consumers. These findings suggest that insurers shift the drug costs associated with

4 higher quality onto consumers when there is little perceived quality variation among drug alternatives but share in the costs of higher quality drugs when there is greater perceived variation in drug quality. In the third chapter, I seek to understand whether and to what extent differences in observed characteristics such as income, education, and health insurance, and unobservable heterogeneities account for racial and ethnic disparities in antidepressant drug use. Little is known regarding racial and ethnic disparities in utilization, spending, and types of drugs purchased in the antidepressant market. Identifying the extent and causes of such disparities is important because depression is a common and chronic disease, with patients frequently suffering recurrences or relapses (Thase 1990). Thus, patients who suffer from depression typically engage in ongoing use of antidepressant drugs. Gaskin et al. (2006) and Wang (2006) find evidence suggesting that unobserved heterogeneities may play an important role in explaining the disparity in drug utilization between different racial/ethnic groups. Yet, there is no established literature identifying factors that can explain these disparities and quantifying the effects of these factors on treatment patterns. To help answer these questions, I use the Blinder-Oaxaca decomposition technique (Blinder 1973; Oaxaca 1973). This method has received only limited application in health services research generally, and in studying racial and ethnic disparities in health services use in particular. It is, however, quite useful in that it can not only distinguish how much observed vs. unobserved characteristics affect disparities, but can also determine the importance of individual factors in contributing to disparities.

5 Such information should prove quite valuable from a policy perspective. Particularly, in this chapter, I seek to determine the extent to which disparities reflect differences in observable population characteristics versus heterogeneity across racial and ethnic groups in antidepressant drug use, employing the Blinder-Oaxaca decomposition technique. Using the Medical Expenditure Panel Survey (MEPS) data from 1996- 2003, I estimate individual out-of-pocket payments, total prescription drug expenditures, drug utilization, the probability of taking generic versus brand name antidepressants, and the share of drugs that are older, lower quality types of antidepressants (e.g., TCAs and MAOIs) for Caucasian, African American, and Hispanic individuals. I find that substantive racial and ethnic disparities exist in all dimensions of antidepressant drug use examined. Observable population characteristics account for most of the differences in drug expenditures, with health insurance and education key factors driving differences in spending. Observable characteristics are also important in explaining racial and ethnic disparities in the probability of purchasing generics and in the quality of antidepressant drugs used. In contrast, differences in total utilization are not well-explained by observable characteristics, and may reflect unobserved heterogeneities. This finding suggests that, in order to mitigate differences in overall antidepressant drug use, policymakers must take into account these unobserved factors, such as discrimination, knowledge, cultural factors. Thus, differences in observable characteristics (notably health insurance and education) explain racial and ethnic disparities in expenditures and patterns of use (e.g., brand vs. generic), but not

6 disparities in total utilization. To develop policies to reduce racial and ethnic disparities, it is important to know the relative importance of observed and unobserved parts and identify individual factors that are associated with the differences. Smedley et al (2003) reported that inconsistent treatment can increase overall health care expenditures. This topic is particularly timely now, as the proportion of minorities in the United States is growing substantially (US Bureau of the Census 2004). Understanding the extent and causes of these disparities can lead to more consistent, appropriate and effective policies aimed at their reduction.

7 Chapter One Entry Pricing and Product Quality: The Case of Antidepressant Drugs

ABSTRACT The rising prices of pharmaceuticals have generated considerable, and often acrimonious, debate. Yet, there is little theoretical work or empirical evidence on pharmaceutical entry-pricing strategies or on the time paths of the prices of new entrants versus incumbents. This study explores the entry pricing strategies of pharmaceutical manufacturers. Previous theoretical models have considered a “price skimming” strategy, with drugs entering the market at a premier relative to incumbents or a “market penetration” strategy, pricing the drug at a discount in the hope of gaining market share. The selection of appropriate strategy depends upon the degree of product differentiation and/or the nature of repeat purchase arrangements. However, these investigations have not examined the implications of product quality for entry pricing and pricing dynamics. The present study provides a model that demonstrates the interplay between quality and product differentiation in determining the optimal pricing strategy. In particular,

8 higher (lower) quality products will engage in price skimming (market penetration) strategies in markets where products are sufficiently differentiated, but will choose a market penetration (price skimming) strategy in markets that are less differentiated. The reason is that in less differentiated markets, price competition is more important than quality competition for gaining market share and higher-quality firms are better able to garner greater market share through price cuts. I tested this model using a unique database that combines information on drug price and quality for antidepressant drugs during the years 1999-2002. Based on quality perceptions by physicians, the antidepressant market I examine is a relatively homogeneous drug class, which consisted largely of selective serotonin reuptake inhibitors (SSRIs). The results indicate that higher quality antidepressants engage in a market penetration strategy, charging initially lower prices that rise over time.

JEL Classification: I110, L100, L110

9 1.1 INTRODUCTION Pharmaceuticals account for a significant share of total health care expenditures in the United States. In 2003, Americans spent $179.2 billion on prescription drugs, almost 4.5 times the $40.3 billion spent in 1990. Importantly, the rate of spending growth for prescription medications outpaced other areas of medical care from 1995 to 2003. During the period from 1995 to 2002, pharmaceutical manufacturing was the most profitable industry in the United States (Kaiser 2005). The rapid increase in pharmaceutical spending has generated both interest and controversy over drug pricing. Yet, there is little theoretical or empirical evidence on entry pricing strategies or the time paths of the prices of new entrants versus incumbents. This study explores the entry pricing strategies of pharmaceutical manufacturers. Previous work has suggested that market conditions determine whether entrants choose lower penetration prices or higher skimming prices. Specifically, in markets where products are relatively well-differentiated and repeat purchase arrangements are less common, entrants are thought to adopt a market- skimming entry pricing strategy, charging a relatively high initial price. However, in more homogeneous markets and in those where repeat purchases are more common, one might expect a market penetration pricing strategy by an entrant (Eaton and Lipsey 1989; Schmalensee 1982; Dolan and Jeuland 1981; Rao 1984). In such markets, there may be greater uncertainty as to whether the entrant’s drug is a significant improvement. Thus, the manufacturer may adopt a low-price entry strategy to familiarize consumers with the product, reaping the benefits in the form of higher

10 market shares and prices over time. Product quality has also been thought to affect entry pricing. For instance, Dean (1969) argued that higher-quality entrants will engage in a skimming strategy, while entrants that offer only marginal improvements will adopt a market penetration strategy, charging lower initial prices. Previous research seems to suggest that (1) the degree of product differentiation and nature of repeat purchase arrangements affect the decision to engage in skimming or penetration pricing strategies and (2) substantially higher- quality products tend to adopt skimming strategies while modestly better products will engage in market penetration strategies. While earlier theoretical models considered the roles of product differentiation and repeat purchases on pricing strategies, existing models have not explicitly derived the role of quality in affecting entry pricing strategies and subsequent pricing decisions. Moreover, existing studies on the effects of product differentiation and repeat purchase arrangements on entry pricing are fundamentally across- market issues, i.e., they account for alternative entry pricing strategies across markets that are distinguished by their product differentiation and/or the nature of repeat purchases of their products. However, these studies do not tell us if and when firms within a product market might engage in skimming or market penetration pricing strategies. Implicitly, these models assume that all firms within a given market adopt the same entry pricing strategy. Thus, no current theory suggests how firms within a product market choose to engage in skimming or market penetration pricing strategies based on differences in product

11 quality. Such a model would need to derive not only the relationship between quality and the entry price, but also the evolution of price over time. 1 Moreover, no previous work has examined the potentially important interactions between quality and product differentiation on optimal entry pricing strategies. In particular, no study has considered whether the relationship between product quality and pricing might vary depending upon the degree of product differentiation within a given market. To help bridge these gaps in the literature, I provide a model in which entry pricing and pricing dynamics are determined as a function of product quality. Here, I demonstrate that the decision by manufacturers of higher-quality products to engage in skimming versus market penetration price strategies itself depends upon the degree of product differentiation within the market. In relatively homogeneous markets, market share is quite sensitive to price and a market penetration strategy is beneficial to makers of the higher-quality products. This is consistent with Comanor’s (1986) assertion that price competition is more important when the market is relatively homogeneous. 2 In such cases, our model shows that market share of the higher-

1 For example, to formally derive a skimming strategy, one needs to show conditions for which higher quality leads to a higher entry price and conditions under which that price declines over time. 2 Comanor 1986: “…where sellers of new products have not been able to achieve substantial quality advantages, they rely more on price competition to enter a therapeutic market. Products that embody higher quality, on the other hand, are more distant from the competitive pressures of established products and can be priced at higher levels.”

12 quality products is particularly sensitive to price, so that, for example, higher-quality drugs will set low initial prices to help generate greater market share. In more differentiated markets, firms manufacturing higher-quality products will optimally choose a market skimming strategy. Our model is tested empirically using a nationally representative data set on drug utilization and expenditures combined with a physician survey on the quality attributes of drugs to examine the effect of drug quality on pharmaceutical pricing strategies. Our quality measure consists of a comprehensive physician assessment of drug attributes and it provides an overall index of a drug’s efficacy and side effects. Specifically, I examine the brand name antidepressant drug class during the period 1999 to 2002. Based on physician quality perceptions, the brand name antidepressant market is a rather homogeneous drug class, consisting largely of selective-serotonin reuptake inhibitors (SSRIs). Using the data described above, I find strong evidence that antidepressant drugs adopt a market penetration pricing strategy, that is, higher- quality antidepressants enter at lower prices, but their prices rise over time. Low- quality antidepressants enter at higher prices and their prices decrease over time. These results are consistent with our model, namely, that market share is particularly sensitive to price for antidepressants with relatively high quality. Thus, the optimal strategy for higher-quality antidepressants is to adopt a market penetration pricing strategy. The remainder of this paper is divided into seven parts. Part II summarizes the literature on entry pricing and empirical work as it pertains to the pharmaceutical

13 industry. The antidepressant drug market examined in this study is described in Part III and the conceptual framework is presented in section IV. Part V describes the data and empirical models to be estimated and the results are presented in Part VI. Part VII summarizes the findings and their implications. 1.2. PREVIOUS WORK There are two distinct types of rivals in pharmaceutical markets. First, there is competition from the branded substitutes, which typically embody different chemical entities and enjoy patent protection, and second, there is competition from generic versions of new and existing products (Rao 1984). 3 Unlike different generic versions of the same drug, brand name competitors are differentiated in terms of quality. As I am particularly interested in the role of quality in entry-pricing strategies, our analysis focuses on brand name-drugs. 1.2.1 THEORETICAL STUDIES

Dean (1969) distinguishes between two strategies for pricing new products:

3 In much of this literature, price competition is not considered until patents have expired and generic substitutes enter. For instance, Frank and Salkever (1997) find higher prices following generic entry, while Caves et al. (1991) find a slight decline in a branded drugs’ price after the appearance of generic rivals. Grabowski & Vernon (1992) find the price trends of the original brand name drugs following generic entry differ by drug category. Thus, evidence on the effect of generic competition on the prices charged for the original brand name drugs is mixed.

14 “skimming” pricing and “penetration” pricing. Skimming is the strategy of setting a high introductory price and then lowering it over time, while penetration is the strategy of setting a low price for a new product to gain market share, raising the price thereafter. Dean suggests that skimming may be used to extract the highest willingness to pay among consumers. This strategy may also be adopted when a product has few “close rivals” in its early stages. 4 On the other hand, Dean asserts that penetration is more common when market share is very sensitive to the price or when the products face drastic (potential) competition. This work gives an insightful qualitative analysis, but does not provide a formal model. A number of theoretical studies of pricing strategies followed Dean’s seminal work. Schmalensee (1982) considers the riskiness of new products, suggesting that pioneering products should charge low entry prices, because the manufacturer must persuade customers to try its “ex ante risky product” in order to build its reputation. In addition, late entrants face substantial disadvantages relative to pioneering brands, so they need to set even lower introductory prices. Therefore, the optimal strategy for the later entrants is to differentiate themselves from the predecessor and avoid being perceived as a “me-too” brand. Shapiro (1983) explored the optimal pricing strategy of a firm with an experience

4 That is, price skimming will occur in more differentiated markets. Dean discusses these ideas but does not formalize them in terms of a model. Eaton and Lipsey (1989) make a similar argument in the context of a spatial model of product differentiation.

15 product, in which consumers learn about quality through use of the product itself. He states that firms’ pricing strategies may be categorized according to customers’ initial evaluations of the product: overestimation and underestimation of the product (i.e. optimistic and pessimistic evaluations). If consumers overestimate the quality of the product, the firm will milk its reputation. In this case, the optimal strategy is to set a high introductory price and then decrease the price over time. In the underestimation case, the product’s reputation must be established and the best strategy in this case is to set a low introductory price and raise it over time. Shapiro considered the role of consumers’ information, but not product signaling from the supply side. Bagwell and Riordan (1991) model entry prices as a signal of quality. At the beginning of the period, the firm sets a price and then consumers form their beliefs about product quality on the basis of this signal. This belief is updated each period. As consumers become more informed about the drug’s quality, the price distortion lessens. The investigators argue that the most efficient way for the firm to signal high quality of new products is to charge high introductory prices. The studies by Schmalensse (1982), Shapiro (1983) and Bagwell and Riordan (1991) shed light on how entry pricing may be affected by perceptions of quality (Schmalensee; Shapiro) and by the impact of price as a quality signal (Bagwell and Riordan). However, these studies do not explore the means by which products with different quality levels adopt different entry pricing strategies. Moreover, they do not consider the role of product differentiation on entry pricing.

Dolan and Jeuland (1982) devised the first study to explicitly model drug pricing

16 strategy in an intertemporal framework. They set up a theoretical model and interpret the skimming and penetration strategies according to the nature of demand. They conclude that market penetration pricing is optimal if repeated purchases of products are important and the skimming strategy is optimal if the demand is stable and production costs decrease over time. However, Dolan and Jeuland do not relate skimming or penetration strategies to variations in product quality. 1.2.2 EMPIRICAL STUDIES Reekie (1978) investigated Dean’s entry pricing ideas in the pharmaceutical market by examining the launching prices for 171 new molecular entities (NMEs) introduced into the U.S market between 1958 and 1975. He finds that the prices of new drugs with important therapeutic gains are significantly higher than the existing counterparts and that the prices of these drugs decline over time. In contrast, the prices of imitators are much lower than the existing drugs and these low initial prices are followed by increases in price over time. Lu and Comanor (1998) explore the demand-side determinants for New Chemical Entity (NCE) prices. They examine the pricing strategies of 144 newly patented pharmaceuticals in the United States between 1978 and 1987, finding that the main explanatory elements are the “therapeutic value” of the product and the competition in the market. Their study shows that the introductory price of drugs representing “important therapeutic gains” can be two or three times those of existing products. However, the drugs that largely duplicated existing ones are typically priced at

17 comparable levels. They also examine the time path of drug prices following entry and find that the prices of the important new drugs declined by about 13% on average four years after entry into the market, while the prices of drugs with little or no therapeutic improvement rose by 22% on average. Following Lu and Comanor’s (1998) approach, Ekelund and Persson (2003) review the pricing strategies of 246 NCEs in the Swedish market and compare the results with those in the US market. The Swedish pharmaceutical market is highly regulated, with various forms of price-cap regulations and other regulatory initiatives to limit pharmaceutical costs. Using the methodology delineated in Lu and Comanor (1998), they also find that introductory prices are positively correlated with drug quality. In contrast to the results in the US market, however, they did not find evidence of market penetration pricing strategy. Instead, all prices decrease substantially over time. Perhaps due to sample size considerations, empirical studies of entry pricing in the pharmaceutical industry only consider the effects of quality on pricing across drug product markets. In contrast, I explore the effects of quality on entry prices within a large drug product market – antidepressants. This approach allows us to examine whether and to what extent quality affects entry pricing within a specific drug product class and to examine whether drugs within the same product class adopt different entry pricing strategies. Before turning to our empirical tests, however, the following section describes the antidepressant drug product market. 1.3 THE MARKET FOR ANTIDEPRESSANT DRUGS

Full document contains 133 pages
Abstract: My thesis examines the pharmaceutical market from three different angles. The first two chapters examine the optimal pricing strategies from the pharmaceutical companies' and insurance companies' point of view respectively. In the last chapter, I analyze the racial and ethnic disparities in the drug use. Each of these essays is summarized below. The first study provides a model that demonstrates the interplay between quality and product differentiation in determining the optimal pricing strategy. Specifically, higher (lower) quality products will engage in price skimming (penetration) strategies in markets where products are sufficiently differentiated, but will choose a market penetration (skimming) strategy in markets that are less differentiated. I tested this model using a unique database that combines information on drug price and quality for antidepressant drugs during the years 1999-2002. The results indicate that higher quality antidepressants engage in a market penetration strategy, charging initially lower prices that rise over time. The second study examines how pharmaceutical costs are shared among consumers and insurers and how drug quality affects these costs. I provide a model which delineates the tradeoff between paying more for higher quality drugs to reduce future medical costs in determining the optimal copayment strategy for the third party payers. I test the model using two large drug therapeutic classes: brand name antidepressants and non-steroidal anti inflammatory drugs (NSAIDs). These two drug classes are interesting to study because they differ in the degree of variation in product quality. While there is little quality differentiation among the antidepressants studied, quality varies by more among the NSAIDS; hence, quality differences are more readily discernible. The results indicate that consumers' out-of-pocket payments are larger for high quality antidepressants, while insurers pay less for these drugs. In contrast, insurers share the drug cost together with the consumers for higher quality NSAIDs. In the third chapter, I seek to determine the extent to which disparities reflect differences in observable population characteristics versus heterogeneity across racial and ethnic groups in antidepressant drug use employing Blinder-Oaxaca decomposition technique. Using Medical Expenditure Panel Survey (MEPS) data from 1996-2003, I estimate individual out-of-pocket payments, total prescription drug expenditures, drug utilization, the probability of taking generic versus brand name antidepressants, and the share of drugs that are older, lower quality types of antidepressants (e.g., TCAs and MAOIs) for Caucasian, African American, and Hispanic individuals. I find that substantive racial and ethnic disparities exist in all dimensions of antidepressant drug use examined. Observable population characteristics account for most of the differences in drug expenditures, with health insurance and education key factors driving differences in spending. Observable characteristics are also important in explaining racial and ethnic disparities in the probability of purchasing generics and in the quality of antidepressant drugs used. In contrast, differences in total utilization are not well-explained by observable characteristics, and may reflect unobserved differences in knowledge and cultural factors, which tells us that to limit differences in overall antidepressant drug use, policymakers must take into account unobserved heterogeneities, such as discrimination, knowledge, cultural background, etc. Thus, differences in observable characteristics (notably health insurance and education) explain racial and ethnic disparities in expenditures and patterns of use (e.g., brand vs. generic), but not disparities in total utilization.