The Evolution of Supply and Demand in Markets for Generic Drugs

Abstract

Policy Points.

• Much concern about generic drug markets has emerged in recent policy debates.

• Important changes in regulations, the structure of purchasing, and the length of the drug supply chain have affected generic drug markets.

• Effective price competition remains the rule in generic markets for large‐selling drugs. Smaller markets and those for injectable products often have less price competition and are more susceptible to supply disruptions.

Context

The image of generic drugs as a commodity sold in competitive markets is an oversimplification, as evidenced by increasing accounts of price spikes, sustained high price‐cost margins, and market disruptions. The mismatch between the canonical economic model of generic drug markets and reality motivated our empirical project.

Methods

To explore recent changes in those factors impacting the supply and demand for generic drugs, we studied, from a variety of sources, the data on price, competition, supply disruptions and recalls, changes to the supply chain, and buy‐side concentration. We examined quarterly data through 2018 for a cohort of 77 molecules that lost patent protection during the so‐called patent cliff between 2010 and 2013.

Findings

On the supply side, we found that for large‐market oral solids, generic entry and price declines were consistent with past studies showing a significant number of market entrants and substantial reductions in the average price of a molecule. In smaller markets for oral solids and injectable products, we observed fewer entrants, higher rates of exit, smaller price reductions, and, in some cases, considerable price instability. The number of reported shortages increased across all generic market types over time, with the rate of shortage increases especially pronounced in markets for injectable products. The number of product recalls also rose over our study period. Although we did not estimate causal effects, we did find several changes in the market environment for generic drugs that may contribute to these phenomena. The demand side for generics has become more concentrated. Supply chains rely more on producers outside the United States (particularly from China and India). Contracting practices have undergone changes that may inhibit competition in product supply. FDA regulatory scruitiny has increased.

Conclusions

Competition in generic drug markets varies widely by market size and product form. Recent changes in demand‐side market structure imply more downward pressure on prices stemming from buy‐side concentration. The FDA's greater regulatory oversight puts upward pressure on costs, and the lengthening of the supply chain increases production uncertainty for producers. Demand and supply‐side changes point to further market instabilities across all generic markets due to producers’ changing economic position.

Generic drugs contain the same active ingredient as and are therapeutically equivalent to their branded counterpart. American purchasers spend about $103 billion per year on generic drugs. Upon approval by the Food and Drug Administration (FDA), generic drugs are regarded as nearly perfect substitutes for the high‐priced branded products they copy. The FDA ensures that generics “can be substituted with the full expectation that the substituted product will produce the same clinical effect and safety profile as the prescribed product.”¹ The FDA codes drugs eligible for substitution as “AB‐rated.”

Generic drugs are generally inexpensive to produce, and in most cases, competition among generic manufacturers reduces prices much below what buyers would pay for the branded product. For example, we estimated, using data from IQVIA, that in 2011 when generic versions of atorvastatin came to market, buyers spent 20 cents per pill instead of the $4 per pill they spent on the brand Lipitor, leading to an estimated annual savings of approximately $2.9 billion on this one drug alone. In aggregate, generic drugs are estimated to have saved buyers $293 billion in 2018 and $2 trillion over the preceding 10 years.² State and federal regulations, as well as health insurer procurement and coverage practices, lead to the rapid and nearly complete substitution of generic products when they become available. IQVIA reports that in 2018, 90% of all drugs in the United States were dispensed as generics, and when confined to drugs for which a generic was available, 97% were dispensed as generics.³

Generic drugs are often referred to as “commodities,” connoting a market with large numbers of buyers and sellers and with prices approximating the cost of production and distribution.⁴ The image of a competitive market for a generic “commodity” is, however, an oversimplification. From 2004 to 2016, 40% of generic drug markets, albeit mostly smaller markets, were supplied by a single manufacturer.⁵ Some evidence suggests that competition may be becoming less robust; that is, the rate of generic manufacturer exit has increased relative to the rate of entry.⁶ An analysis of Medicaid data showed that 8.2% of generic prescriptions experienced price increases of between 100% and 500% between 2013 and 2014.⁷ The Inspector General of the US Department of Health and Human Services determined that between 2005 and 2014, 22% of generic drugs had price increases that exceeded the consumer price index.⁸ Over roughly the same period, 2006 to 2013, the American Association of Retired Persons (AARP) found that 15% of drugs had significant price increases and that those drugs with large price increases were supplied by several specific manufacturers.⁹ The number of reported product recalls and shortages grew by around 400% from 2007 to 2013.¹⁰ According to one count, recent years have seen between 140 to 190 new shortages per year.¹¹

The imperfect match between the canonical economic competitive model and the reality of generic drug markets is what motivated our analysis. In order to summarize important aspects of economic performance, this article reviews the data on the entry, prices, and quantities in generic drug markets. One challenge we faced is contending with the thousands of generic products currently on the market, varying by form, strength, route of administration; produced by scores of generic manufacturers; and introduced in different time periods. To structure our task, we selected a cohort of 77 molecules that lost patent protection during the so‐called patent cliff between 2010 and 2013, a period during which more than $40 billion in yearly brand‐name drug sales lost patent protection. This sample includes several of the most widely prescribed pharmaceuticals to date, spanning several therapeutic classes, including generic versions of Lipitor, Lexapro, Singulair, and Diovan. Defining our study sample in this way restricts the scope of our analysis to drugs that relatively recently lost their patents, which allows for comment on the numerous demand and supply factors that have changed in the previous decade. Yet, it is important to note that we cannot claim to be able to generalize our results to the state of competition in markets for older generics. When possible, however, we do compare the results for our cohort to those reported for earlier cohorts of generic drugs.

We next discuss the history of the generic drug markets’ regulatory, cost, supply, and demand conditions. Although the importance of regulatory features is generally appreciated, some elements of regulation, such as changes in the rates of facility inspection, deserve more attention. The cost and supply structure of generic markets is special in some ways. Even though generic manufacturers do not invest in research and development years before selling a drug, they do incur one‐time sunk costs and fixed costs (which recur with market participation but are independent of the level of output). Regulatory requirements extend the time between a manufacturer's decision to seek entry and actual entry, thereby introducing additional uncertainty for the prospective entrant. The primary input to a generic drug, the active pharmaceutical ingredient (API), is generally purchased from specialized, high‐volume producers, many of which are located outside the United States. The demand side of markets for generic products has special features as well. Generic manufacturers do not sell to end purchasers (consumers or insurers) but rather distribute their products through the complex drug purchasing and distribution system featuring a small set of drug wholesalers, purchasing groups, and pharmacy benefit managers (PBMs).

Our review of the earlier literature on generic drug supply and pricing comes next, followed by a description of the IQVIA data and the supplemental data we use in this article. Then we present our main findings, highlighting the differences in pricing patterns following generic entry, based on market size. We comment on trends toward more “buy‐side” concentration facing generic manufacturers and call attention to an increase in both shortages and product recalls over time. Finally, we comment on the role of regulatory, supply, and demand factors in understanding the heterogeneity in generic drug market performance.

Background

The markets for generic drugs are highly regulated.⁵, ¹², ¹³ New brand‐name prescription drugs must be approved by the FDA under the extensive clinical testing requirements of a New Drug Application (NDA). In contrast, generic drugs are approved for marketing under an Abbreviated New Drug Application (ANDA) process, a shorter and less costly regulatory review. The FDA approves generic drugs if they have identical active ingredients, dosages, forms, and routes of administration as do the branded reference drugs. This is broadly referred to as “bioequivalence.” Manufacturers must meet Good Manufacturing Practice (GMP) standards and be approved for production of the drug in question. The FDA is responsible for inspecting manufacturing facilities in the United States and abroad.

State regulation and the purchasing and coverage policies of health insurers, often administered by pharmacy benefit managers (PBMs), encourage the substitution of generic for branded products when a generic becomes available.^{14 (p177)} State generic substitution laws allow (and sometimes require) pharmacists to dispense a therapeutically equivalent generic drug in place of a brand‐name drug if the physician does not indicate that the prescription should be “dispensed as written.” PBMs and insurers also encourage substituting generic products via the higher dispensing fees they pay pharmacies, their drug formularies that provide for lower copays for generic products, and by sometimes dropping brand versions from the formulary altogether when a generic version of the drug is available.

The marketing of generic drugs is governed by the Drug Price Competition and Patent Term Restoration Act of 1984, commonly known as the Hatch‐Waxman Act. For most products protected by patents, in order to enter the market a generic version must wait until all relevant patents expire or are found invalid, or the manufacturer risks being sued for patent infringement. The Hatch‐Waxman Act structures a process by which a generic company, by simply declaring its intent to enter, can trigger a patent infringement lawsuit without actual entry, thereby avoiding the risk of the generic having to pay damages to the brand and also facilitating an earlier resolution of patent validity and infringement issues. Consumers benefit if the generic successfully overcomes a weak or uninfringed patent. To encourage generic firms to invest in challenging a brand's patents, the first generic manufacturer to successfully challenge a patent and receive approval to market the generic product may be awarded a 180‐day exclusivity period during which the FDA may not approve another ANDA. This 180‐day period is highly lucrative to the first‐filing generic drugmaker(s) because it can price the drug well above marginal cost when it is the only ANDA‐based competition for the brand‐name drug while still benefiting from state generic substitution laws. After other generics enter, the decline in price reduces the price‐cost margin.

Most empirical analyses of prices for generic drugs focus on oral solids (pills or tablets), which account for 93% of generic drugs sold in the United States. Overall, generic prices charged by manufacturers have fallen steadily over time,¹⁵, ¹⁶ despite several recent high‐profile price spikes for some generic products.⁷ Within these overall trends are some important economic distinctions. The overall price declines are largely driven by lower prices for molecules new to generic competition. Prices for older generic drugs have been flat or rising only modestly,¹⁶ and there have also been allegations and some legal settlements related to price fixing among generic manufacturers. For example, a complaint filed by the attorneys generals of 51 states and territories in June 2020 was precipitated by the discovery, during another investigation, of documents allegedly containing explicit agreements among manufacturers to reduce competition and increase prices.¹⁷ In addition, evidence suggests that price reductions at the manufacturer level may not be fully passed through to consumers and purchasers of generic drugs.¹⁶ In addition, consumers’ out‐of‐pocket expenses and insurers’ payments decline less than do the prices charged by manufacturers implying higher margins somewhere in the supply chain.

The regulatory framework and market structure for generic competition have been changing, with potential effects on economic outcomes. The purchasing arrangements for generic drugs have evolved. Consolidation has led to fewer direct purchasers. New organizational forms have emerged, principally the vertically integrated joint ventures between large wholesale drug distributors and major retail drugstore chains. Currently, the three largest joint ventures are Red Oak (Cardinal‐CVS), Walgreen Boots Alliance (with Amerisource), and McKesson OneStop (with Walmart and Express Scripts). These three organizations have been estimated to account for as much as 90% of all US generic drug purchases.¹⁸

The supply chain for generic drugs has been lengthening, extending outside the United States. Currently, only 12% of API manufacturing takes place in this country, with 31% in India and 14% in China. About 37% of final product manufacturing takes place in the United States, figures that are much lower than in the past.¹⁹ The supply chain increasingly features exclusive contracting, in which generic manufacturers contract with or purchase a single API producer, which can preclude the producer from selling to other manufacturers.¹⁹ Exclusive contracts allow the manufacturers to negotiate more favorable prices for API, but they also make the manufacturer more vulnerable to disruptions in supply from its contracted source.

In part related to concerns about quality, the FDA has recently changed its approach to inspecting the manufacturing facilities for generic drugs.²⁰ More enforcement detects more quality problems; higher quality costs more; and stricter enforcement may increase cost along with quality. One count of drug “shortages” found that 62% of shortages stem from quality problems, some uncovered by the FDA's inspections.¹⁹

Earlier Research on Generic Drug Prices and Supply

Most studies of the effects of generic entry on drug prices use proprietary data from IQVIA (formerly IMS Health) on prices paid to wholesalers by retail drugstores, hospitals, clinics, long‐term care pharmacies, and other nonretail outlets to drug manufacturers. As Berndt and colleagues explain, using data from wholesale transactions to study market functioning implicitly assumes that “wholesale price ratios provide reasonable proxies for relative consumer price impacts.”²¹ Reiffen and Ward use an alternative IMS data set (Generic Spectra) of drugs that lost exclusivity in the late 1980s and that contains both wholesale and retail prices.²² Their results show that wholesale prices of generic drugs averaged 69% of the pre‐entry brand price, while retail prices of generic drugs averaged 76% of the pre‐entry brand price. Using a later version of the same data, Grabowski and colleagues studied 54 branded drugs that lost their patent protection between 2006 and 2008.²³ They found that wholesale prices of generic drugs averaged 45% of the brand drug's price one year after generic entry, while retail generic prices averaged 66% of the branded drugs’ prices. For drugs with sales of $750 million or more in the year before generic entry, wholesale and retail generic price ratios were even smaller, 18% and 48%, respectively.

Some studies assess the number of generic sellers relationship to factors that might reasonably be expected to affect market structure. The number of generic entrants is positively associated with the size of the market. Using data for 2003 to 2008, Tenn and Wendling found that on average, “small drug markets” attracted 2.7 generic manufacturers, whereas “large drug markets” attracted 5.8 generic manufacturers.²⁴ Grabowski and colleagues estimated that market entry levels were more than 50% greater for drugs with more than $1 billion in sales before generic entry, relative to those with sales of between $100 million and $999 million between 2011 and 2014.²³ They also found that brands retain a lower share of larger markets and that the rate of brand replacement by generic products accelerated over their 16 years of observation.

Many studies have found that a greater number of generic entrants predicts a more rapid decline in the drug's average price following generic entry. Caves, Whinston, and Hurwitz, using data from the late 1980s, reported that generic prices fell to 34% of the branded drugs’ prices as the number of generic competitors rose from 1 to 10. A Congressional Budget Office (CBO) study using data from drugs that lost exclusivity in the late 1980s showed that as the number of generic competitors increased, the prices of generic products fell further relative to the brand price.²⁶ Based on data from the mid‐1990s, Saha and colleagues found that “a lower price ratio is associated with higher generic share and, on average, each additional generic manufacturer induces nearly a 2.3 percent monthly decrease in the value of the generic‐to‐brand price ratio.”²⁷ Using data for 2015 to 2017, the FDA found that drugs with two generic manufacturers had average generic prices that were about half the price of the brand drug and that this price ratio declined further as the number of generic manufacturers increased.²⁸ Olson and Wendling, using data on oral solid markets from the mid‐2000s found that the price of a generic drug with two generic competitors was 13% to 14% below that of a drug with one generic and that the price of a generic drug with three competitors was 31.5% to 48.3% lower.²⁹

More recently, attention has also been directed to supply disruptions and drug “shortages,” a term used differently by different sources. In economics, a shortage exists when buyers are not able to purchase all they desire at the going price, in other words, a condition of “excess demand.” The policy literature on generic drugs also uses the word shortage to include instances in which the quantity supplied is sharply reduced. A supply disruption may lead to a higher price, which clears markets (without excess demand). The literature also sometimes defines a disruption and shortage at the level of the supplier rather than at the level of the market. In other words, a supply interruption by a particular firm may sometimes be counted as a disruption even if it has little impact on the market supply. Finally, the literature also identifies product recalls, a particular form of supply disruption, which often result from regulatory oversight.

The Office of the Assistant Secretary of Policy Evaluation (ASPE) of the US Department of Health and Human Services found shortages in injectable drug markets following a rapid increase in the number of generic drug approvals by the FDA that was not accompanied by an increase in the generic industry's production capacity.³⁰ The natural interpretation is that the generic drug approvals led to a rapid increase in demand for generic drugs with a limited production capacity, resulting in very high rates of capacity utilization, which may then have led to quality problems and shortages. Initially, ASPE viewed this as a “short‐run problem” that would be addressed by competitive responses to market prices. This longer‐term response has not materialized because entry into small injectable markets is low and price spikes have continued in the injectables markets (discussed later). One explanation for continued price hikes, suggested by Woodcock and Wosinska, is that downward price pressures have resulted in underinvestment in production facilities.³¹

Stromberg studied the relationship between greater regulatory oversight and the number of supply disruptions from 2001 to 2013, using manufacturers’ reported shortages from the University of Utah Drug Shortage Database (UUDSD).¹⁰ He found a significant relationship between the number of FDA inspections and citations and the number of supply disruptions for both parenteral and nonparenteral drugs. Stromberg noted that previous work focused on the patterns and causes of shortages in injectable markets, even though similar patterns of shortages have been observed in parenteral and nonparenteral markets, suggesting a common cause of increasing shortages over time across product forms.

Data

We assembled data on generic drug prices and quantities, generic firm entry and exit, supply disruptions, and recalls. Our data on drug prices, quantities, and participating suppliers came from IQVIA, and our supplemental data came from the several sources described here.

IQVIA Data

The National Sales Perspective (NSP) database provided quarterly data from 2009‐Q1 to 2018‐Q2 on generic pharmaceutical retail and nonretail sales. The main unit of analysis was the molecule‐product‐form (e.g., atorvastatin oral solids), which we refer to as a “market.” Dollar sales were based on the average invoice price of purchases from manufacturers or wholesalers. Although rebates are not captured in these data, rebates for generic drugs are much less significant than for brand drugs. The quantity of sales was measured as the extended unit total prescriptions (EU TRx), which counts, depending on the form of the product, the number of pills, tablets, or milliliters of solution sold. The NSP contains information on the product‐form (eg, oral solid or injectable), strength (eg, 10 milligram or 10 milliliter), the manufacturer (eg, Teva or Mylan), and the distribution channel (eg, retail or mail order). The NSP includes dollar sales and volume by molecule, product‐form, strength, manufacturer, and distribution level. We aggregated these into the combined‐molecule‐product form markets. The data included information on the type of product within the molecule category (eg, branded, authorized generic, or generic) of each manufacturer and the loss of exclusivity date for each product (this information was made available by IQVIA).

Supplemental Data

The FDA website contains information on the number and timing of ANDAs filed with the FDA for approval and marketing and the number and timing of recalls in each market. Specifically, the FDA Orange Book lists each ANDA approval date (the Orange Book contains information on the ANDA applicant, which may not be the actual manufacturer or marketer due to outsourcing or other licensing arrangements). The FDA Enforcement Report's archives count the number of recalls associated with each combined molecule and, back to June of 2012, list enforcement actions including the removal of, or corrections made for, a drug meeting the FDA's definition of a recall. Notably, we recorded only those recalls that were national in scope, as opposed to those affecting a single batch that had little effect on the aggregate supply. The report contains information on the initiation date, termination date, and status of each recall at the combined‐molecule level.

Erin Fox of the University of Utah generously provided data on supply disruptions from the UUDSD. The UUDSD tracks supply disruptions reported by industry participants. It defines shortages as “a supply issue that affects how the pharmacy prepares or dispenses a drug product or influences patient care when prescribers must use an alternative agent”³² The UUDSD is frequently used by organizations such as the Government Accountability Office (GAO) to track so‐called shortages in the generic drug market.³³ Each disruption in the UUDSD represents an event reported by the manufacturer and is confirmed by contacting other manufacturers in the market. Reported shortages vary from smaller supply disruptions lasting a few weeks to wide‐scale shortages spanning several years. The resulting report includes the combined‐molecule name, the initiation and termination dates of the supply disruption, the reason for the disruption, and the status of the disruption (active or inactive). In order to address the varied duration and impact of the shortages in the data set, we defined a second, stricter measure of shortage as a shortage reported in the UUDSD that was accompanied by a greater than 10% reduction in prescriptions dispensed in the two quarters following the announcement of a shortage relative to the two preceding quarters.

A large generic drug manufacturer provided us confidential information about the share of purchases made by wholesalers. These shares apply to this manufacturer's sales; they are not market wide. The data provided were for the entire period from 2012 to 2019 as the percentage of total net dollar sales and the percentage of total volume sold.

We obtained the total yearly number of inspections from the FDA's inspections database recording inspections from October 2008 through December 2018. Inspections can take place for different purposes, and similar to Stromberg (2018), we recorded only those pharmaceutical inspections listed as related to drug quality assurance related and performed by the Center for Drug Evaluation and Research (CDER). Inspections were linked to a manufacturer, not to a particular product.

Sample Selection and Variable Definitions

The initial sample of drugs contained 144 molecule‐product‐form combinations, associated with 108 molecules that lost exclusivity between January 2010 and December 2012. We kept only oral solids and injectable forms. Similar to other studies,⁵, ⁶ we did not include markets in which price was not fully observed. We also excluded voriconazole from our analysis owing to its highly erratic price behavior, such as several 10‐ to 20‐fold increases in price over the price before the brand's loss of exclusivity (LOE). The final sample consisted of 88 molecule‐product‐forms and 76 molecules observed continuously for at least 4 quarters before and 16 quarters after the loss of exclusivity.

We calculated the quarterly prices by dividing the total dollar value of sales of all manufacturers in a market by the TRx in a given quarter. Prices and spending are expressed in 2016 dollars based on the Federal Reserve Board's GDP deflator. We examined price movements by indexing the average molecule price after LOE to the average price in the year before LOE, weighted by prepatent sales. Each manufacturer was weighted by the log of its sales in the year before LOE. The manufacturer count is the number of distinct manufacturers with positive sales and positive TRx values in a given quarter. We conducted a sensitivity analysis of the count of manufacturers by removing those not appearing in the FDA's Orange Book that were more likely to be labelers or repackagers and did not show meaningful differences in the results.

For each molecule, qualitative response (0,1) variables indicate whether a shortage or a recall was posted in a given quarter (this is not product specific). Qualitative response variables also indicate the resolution of a shortage or recall in a given quarter. A molecule experienced an ongoing shortage or recall if the cumulative number of initiated events less the number of terminated events was greater than zero. Given that information on recall initiations was unavailable before 2012, the fraction of markets ever experiencing a recall reported is understated in this study.

Results

We cover the key elements of market structure and regulation, as well as the market outcomes in prices, number of suppliers, and supply disruptions.

Market Structure

We begin by describing some of the structural changes in the industry that affected supply and demand conditions. Since 2012, the total number of ANDAs has remained relatively stable at between 1,000 and 1,300 per year.³ The mean time between submission and approval has declined steadily, in part as a result of legislation and policy (the Generic Drug User Fee Amendments known as GDUFA).⁵ In our study cohort, an average of eight ANDAs were filed for each molecule in the first four years after LOE. While many generic drug manufacturers participate in US markets across all products, the number of sellers can be small in particular markets.

Active pharmaceutical ingredient (API) production and final product manufacturing have shifted outside the United States, primarily to Asia.⁵ The share of drugs produced abroad is currently roughly 86% for APIs and 62% for finished products. Partly in response, the FDA has increased the number of inspection staff and inspections. In 2009, the average number of total inspections increased from 1,500 to an average of 1,800 per year over the next decade. Whether this increase has “kept up” with the number of plants producing overseas is unclear. Since GDUFA, the frequency of warning letters issued to generic drug companies has increased by 150%, perhaps putting upward pressure on quality and production costs.

Three large buying groups have been formed through the creation of joint ventures between wholesale druggists and retail pharmacy chains: Boots Alliance (Walgreens, Amerisource‐Bergen, Econdisc; 2013), Red Oak (CVS, Cardinal Health, Target, OmniCare; 2013), and McKesson One Stop (Walmart, RiteAid, McKesson; 2014–2016). Based on data from a generic manufacturer, estimates published by Drug Channels, and UBS analyst reports, the share of purchases by the three large buying groups rose from roughly 59% in 2014 to between 72% and 81% in 2018.

Prices, Entry, and Exit

We categorized products into oral solids and injectables. Within each product form, we stratified them by terciles of the branded products’ dollar sales in the year before LOE by the brand. Table 1 summarizes our 2‐by‐3 categorization. Note that the number of sales in all terciles of injectables was much smaller than the number for oral solids. To understand the impact of recent changes to the supply‐side market structure and federal regulations, we also examined each metric overtime after splitting the sample by the year of LOE (2010, 2011, 2012).

Table 1.

Sales Terciles by Product Form Before Loss of Exclusivity

	Small Market	Medium Market	Large Market
Oral Solids (N = 66)	Less than $125m	$125m to $612m	Greater than $612m
Injectables (N = 23)	Less than $570k	$570k to $2.5m	Greater than $2.5m

Figure 1 shows the number of sellers per quarter in oral‐solid markets for 16 quarters after LOE (the quarter of LOE is designated “Quarter 0” for each market). An average of ten sellers entered in the first year for oral solids in the highest tercile of sales. After two years, the number of sellers leveled off at about 13. For the middle tercile of sales, an average of roughly five firms entered in the first year LOE, and the bottom tercile had fewer than three entrants in the first year.

Figure 1.

Median Number of Oral Solid Molecule Manufacturers by Market Size and Quarter Relative to Loss of Exclusivity [Colour figure can be viewed at wileyonlinelibrary.com]

Market participation in Figure 1 reflects both entry and exit. Market exits tend to begin in the third year after LOE. In those markets with the highest tercile of sales, 21% of the entrants had exited by quarter 15 after LOE. For the smallest markets, this figure was 13%.

Figure 2 shows the price patterns for molecules experiencing generic competition by tercile indexed to the price before LOE. The highest tercile follows a pattern consistent with the previous literature. After one year, the median index price of generic products in the largest markets fell to about 10% to 20% of the brand price before LOE. For the middle tercile of oral solids, the price fell to about 50% of the brand price before LOE. The prices for the smallest markets trended down sluggishly after LOE but, in our data, rose in later years relative to the pre‐LOE brand price .

Figure 2.

Oral Solid Molecule Price Index by Market Size and Quarter Relative to Loss of Exclusivity [Colour figure can be viewed at wileyonlinelibrary.com]

Figure 3 reports the pattern of market participation for injectables after LOE. For all terciles, the rates of participation were well below those for oral solids. Injectables are often more complicated to produce than oral solids, and the dollar size of markets, as noted earlier, is typically much smaller. For the lowest tercile of sales before LOE, the median number of sellers was just 1 to 1.5. For the top tercile, a median of three firms entered during the first year. The number of market participants grew to a median of five at the end of four years.

Figure 3.

Median Number of Injectable Molecule Manufacturers by Market Size and Quarter Relative to Loss of Exclusivity [Colour figure can be viewed at wileyonlinelibrary.com]

Exit is quantitatively important in markets for injectables. In contrast to oral solids, 13% of the large market entrants had exited by the end of four years after LOE. More than half (57%) of the entrants in small markets had exited by the end of the fourth year after exclusivity.

Figure 4 shows the indexes of median prices by tercile. Prices in the top tercile, with an average three entrants in the first year, fell steadily to roughly 80% of the price before LOE at the end of the first year to about 35% of the brand price at the end of the second year. In contrast, there is no evidence for price declines for the two smaller‐sales terciles for injectables. The lowest tercile shows several sharp, temporary, upward movements in the indexed price for that group. This may be due to the greater number of supply interruptions among a relatively small number of manufacturers.

Figure 4.

Median Price of Injectable Molecules by Market Size and Quarter Relative to Loss of Exclusivity [Colour figure can be viewed at wileyonlinelibrary.com]

Shortages and Recalls

Shortages were reported in in 24 of the 66 oral solid markets included in our data. Sixteen percent of these shortages were associated with price reductions higher than 10% in the quantity prescribed. Between 2010 and 2014, roughly half the medium and large oral solid markets experienced a shortage of some form. The percentage of drugs experiencing a shortage grew more rapidly between 2012 and 2014 than in the previous two years. For small markets, the likelihood of a shortage was considerably lower, at less than 30% of products.

The rate of national product recalls grew sharply over the six‐year period we observed, extending to more than 60% of products in the medium and large markets by the end of our observation period in 2017. The rate of recalls for the small markets for oral solids was about a third of that for the two larger sets of product markets.

The patterns of shortage and recall were quite different for injectable products. Sixteen of 23 markets experienced a shortage over the study period. Nineteen percent of these shortages were associated with reductions of more than 10% in the quantity prescribed. In medium and small markets, nearly 80% of products had experienced shortages by the end of our observation period, whereas fewer than 50% of the large market products had experienced a shortage, which is more in line with our findings for oral solids.

When examining the trends in recalls for injectable products, we found that all three groupings of market size showed large increases in recalls over time. The recall rates for medium and small markets grew rapidly after the first 10 quarters of observation, reaching 65%. In the larger markets, the growth in recall rates was more gradual.

Supply and Prices by Year of Loss of Exclusivity

Figures 5a and 5b show the trends in the number of sellers and prices relative to the prices before LOE based on the year of LOE for oral solid drugs. The 2010, 2011, and 2012 cohorts include 19, 20, and 27 molecules, respectively. Supply consistently trended upward following LOE for about 10 quarters for the three cohorts. Prices decline steadily for the first 15 quarters after. Price movements became more erratic in later quarters, especially for the 2010 cohort. This may in part be due to the smaller number of molecules and the increases in exits in later quarters. The prices continued to drop for the 2011 and 2012 cohorts, albeit not smoothly. At the end of our observation period, prices appear to have reached 50% to 80% of the prices before LOE. Again, this may be due to the declining numbers in the cohorts.

Figure 5a.

Total Number of Manufacturers of Oral Solid Molecules by Year of Loss of Exclusivity [Colour figure can be viewed at wileyonlinelibrary.com]

Figure 5b.

Price Index of Oral Solid Molecules by Year of Loss of Exclusivity

Small numbers of injectables in our data prevented us from reporting reliable results stratified by entry cohort.

Discussion

The data on market entry and price declines for oral solids were largely consistent with the data of previous studies, showing significant numbers of market entries soon after LOE and substantial reductions in the average price of a molecule. These results are in line with the competitive model interpretation of the market for generic drugs. The picture is not so simple, however, even for the oral solids. Smaller markets for oral solids see fewer entrants. While this finding is also similar to those of previous works, it implies caution in applying the concept of a competitive model too widely in the generic market. We also observed upward drifts in prices after several years. This may indicate that the overall pattern of consistent aggregate generic price declines is largely the result of the declines in prices stemming from new products being subject to generic competition.

In addition, markets for injectables do not square easily with the competitive model. We observed fewer entries, smaller price reductions, and, in some cases, considerable price instability. The high rate of exits of small market injectable products implies that few manufacturers remain in these markets after four years. One possible cause for this is “too much” entry in relation to market size due to the time lag between filing an ANDA indicating the intention to enter and the actual entry.

The broader context of supply and demand for generic products has been changing over time. Supply disruptions have been increasing, and the rise in the rate of shortages was especially pronounced in markets for injectable products. But it is important to keep in mind that only a modest percentage of the reported supply interruptions and shortages have a meaningful impact on market supply. These shortages occurred along with higher numbers of product recalls during the observation period. We observed some upward trends in generic prices in the oral solid markets that corresponded to periods when the shortages and recalls were at their highest levels.

Although we have not estimated causal relationships in this article, we do call attention to several changes in the market environment for generic drugs that may contribute to the observed path of prices and suppliers. These are the greater concentration of purchasers of generic drugs, the lengthening of the supply chain (greater reliance on foreign production), and the FDA's enhanced regulatory scrutiny of production facilities. Concentration on the “buy side” reduces margins. Downward pressure on prices tends to drive management to cut costs, sometimes at the expense of maintaining higher levels of quality, which may contribute to more shortages. These findings are consistent with the proposition advanced by Woodcock and Wosinska that these shortages are the result of underinvestment in production capacity. They are also in line with other research. Dave and colleagues identified a correlation between the lowest‐priced generic drugs and the appearance of shortages.³³ The FDA released a report on shortages of generic drugs and their root causes.¹⁹ The report highlights the fact that shortages occur differentially across market segments and contends that quality problems are at the heart of the shortages. While low prices generally benefit consumers, if those low prices come at the cost of more frequent shortages and price spikes from producers’ unstable or unsustainable economic position, consumers may end up being hurt.

The structure of the generic market also needs to be understood in terms of input supply to generic manufacturers. A supply structure with many generic manufacturers, each relying on a small number of sources of API, may not provide consumers much of a benefit from competition if it is the API suppliers that are exercising market power. Integration of the generic manufacturer with the API supplier may threaten market efficiency if such arrangements preclude competitors’ access to upstream resources. A few API suppliers may also increase the risk of supply interruptions when manufacturers lack redundancy in sources of input supply.

Finally, the FDA has increased its inspections of foreign manufacturing facilities, increasing the number of deficiencies identified. This may simply reflect its uncovering previously missed deficiencies, or the quality of production outside the United States may be deteriorating. Either way, if production continues to shift outside the United States, drug shortages will likely become more frequent.⁹

Viewed as a package, these changes in the market structure of supply of inputs, generic manufacturing, and purchasing all put downward pressure on the margins of generic firms. Buy‐side concentration pushes down price, and input‐supply‐side concentration increases costs, all in a context of increasing regulatory scrutiny associated with the more difficult regulatory task of monitoring production quality of overseas suppliers. It is not surprising to see more frequent supply disruptions with this mix of factors.

The timely availability of low‐cost generic drugs is a critical element of US health care policy. Antitrust authorities have worked to maintain timely generic entry by monitoring and challenging the settlements of patent disputes between brand‐name and generic drugs that may include an anticompetitive delay in the entry of generics. If these markets were reliably competitive, the regulatory task of ensuring adequate competition would end with opening markets to generics. The findings in our article, in line with some previous research, indicate that the competitive model does not fit many generic drug markets, and not just because of the market structure of the generic industry but because of the nature of the markets that supply inputs and buy products. Although we did not explore the possibility of price fixing among generic manufacturers, recent allegations regarding agreements to restrain trade imply that traditional antitrust enforcement should remain an element of public policy.

Expanding the scope of regulation outside the industry itself also seems to have merit. The demand and supply factors affecting the generic industry raise quite different issues. The recent consolidation and high market concentration on the buy side of generic markets raises special challenges. Purchasing joint ventures between wholesalers and chain pharmacies may result in lower purchase prices from generic manufacturers that are not passed to the ultimate payers, that is, insurers and consumers. This possible exercise of market power likely deserves scrutiny, as it may involve monopsony power reducing prices to manufacturers and monopoly power raising prices to insurers, taxpayers, and consumers.

Contending with the quality of inputs and the reliability of supply is a major and difficult challenge for US regulatory authorities. Manufacturers currently have very strong market incentives to reduce prices along with the relatively weak penalties for supply disruptions. In the past, manufacturers maintained redundant sources of input supply by contracting with multiple API producers. Exclusive contracts have become increasingly common as a means of driving down the cost of inputs. These contracting arrangements make manufacturers more susceptible to supply disruptions. Policy measures that promote a rebalancing of the incentives for lower prices and reliable supply would likely improve consumers’ welfare.

Funding/Support: The authors gratefully acknowledge support from the Dean's Initiative award program for Innovation Grants in the Basic and Social Sciences at Harvard Medical School.

Acknowledgments: We are grateful for comments from anonymous reviewers on a previous version of this article.