Then, the act of invalidation benefits not only the generic firm who made the investment, but also others who seek to enter the market. Because such public goods tend to be under supplied in a competitive market, Congress created a system to reward the first generic firm to invest in a patent challenge. The reward is given out through a complex process that I summarize here. When a generic firm files an ANDA containing a paragraph IV certification to the FDA, it must directly notify the originator , as well as the other holders of the patents being challenged, about its filing. The originator must then decide within 45 days whether or not to initiate a patent infringement suit. If the originator decides not to sue, then the FDA is allowed to approve the ANDA and the generic may enter the market. If the generic firm is the first to have filed a substantially complete ANDA containing a paragraph IV certification, it is awarded a 180-day exclusivity in the generic market. This means that the FDA is not allowed to approve any other ANDA until 180 days have passed since the first generic product’s commercial launch. If the originator decides to sue the generic entrant, then the FDA is stayed from giving final approval to the ANDA until 30 months have passed or until a court decides that the patent in question is invalid or not infringed, whichever comes sooner. The FDA may review the ANDA in the mean time, indoor vertical garden systems but it can only issue a tentative approval. Thus, the 30-month stay functions as an automatic preliminary injunction against the paragraph IV ANDA applicant. The main possible outcomes of the patent infringement suit between the originator and the paragraph IV applicant are the following: a victory for the generic entrant, a loss for the generic entrant, or a settlement between the two parties.

If the generic applicant wins the patent infringement suit, its ANDA receives final approval from the FDA once the other patents listed in the Orange Book expire. If the generic firm is the first to have filed a substantially complete paragraph IV ANDA, it obtains the right to 180-day exclusivity. The exclusivity period starts when the first to-file generic begins commercial marketing or when a court decides that the patent in question is invalid or not infringed, whichever is earlier. If the generic firm loses the infringement suit for every challenged patent, then its ANDA is not approved until expiration of those patents or until the end of the 30-month stay. Even if the firm is the first-to-file paragraph IV applicant, it is not awarded the 180-day exclusivity, because the right to exclusivity disappears with the expiration of the challenged patents . If the generic and originator firms decide to settle the patent infringement suit, the generic firm’s ANDA is approved only after the 30-month stay. If the generic firm is the first-to-file paragraph IV applicant, it becomes eligible for 180-day exclusivity, which is triggered by the generic product’s commercial launch. The right to 180-day exclusivity is given only to the first-to-file paragraph IV applicant. If the first-to-file applicant loses in patent infringement litigation or otherwise forfeits its right to 180- day exclusivity, the right disappears; it is not rolled over to the next-in-line applicant . If multiple firms file ANDAs with paragraph IV certifications on the same day, and no prior ANDA has been filed, the right to generic exclusivity is shared between those firms. The large profits available from 180-day exclusivities have made generic firms more aggressive in their patent challenges. As Grabowski and Higgins and Graham note, the number of ANDAs containing paragraph IV certifications increased rapidly after the regulatory change: the average number of paragraph IV ANDA filings per year rose from thirteen during 1992-2000 to 94 in the 2001-2008 period.

While this increase partly reflects the greater number of blockbuster drugs going generic in the latter period, observers agree that the regulatory change played a significant role . Table 2.1 presents the share of generic markets that were the subject of one or more paragraph IV ANDA filings in a sample of 128 markets that opened up during 1993-2005. As described more fully in Section 2.5, drug markets were selected for inclusion using the following criteria: the drug product contains only one API; of the set of finished formulations containing the same API, the product is the first to experience generic entry; and there is at least one generic entrant in the market. The propensity of paragraph IV challenges suddenly jumps for markets that experienced first generic entry in 1999. This reflects expectations among generic firms that the FDA would give out more 180-day exclusivities following the 1998 court decisions. The share of generic markets with paragraph IV certifications remains high – at around one-half – in the subsequent years. Grabowski comments that the granting of more 180-day exclusivities has, in some cases, turned the generic entry process into a race to be first. Higgins and Graham note that as a result of more aggressive efforts by generic entrants, ANDA filings have come to take place earlier in a drug’s lifecycle. Indeed, there have been many markets where multiple generic firms filed their paragraph IV ANDAs exactly four years after the approval of the originator’s NDA – that is, on the earliest date allowed by the FDA . Also, Grabowski and Kyle show that drug markets with higher revenue tend to experience generic entry sooner, partly because they tend to be more heavily targeted for paragraph IV challenges. Interestingly, while ANDAs filings are being made increasingly early, Grabowski and Kyle find no evidence that generic product launches are occurring earlier in the drug’s life cycle in markets that opened up more recently.

This may be because the Hatch-Waxman system has had an unintended side effect. As reported by the Federal Trade Commission and Bulow , the system has been used by some originators, somewhat paradoxically, to delay generic entry through the use of so-called “pay-to-delay” settlements. Given that the existence of a patent challenge turns the generic entry process into a race to be first, econometric analysis of generic firm behavior would ideally be based on a model that takes the timing of entry into account. Unfortunately, the data that I use do not contain accurate information on the timing of entry by each generic firm. Also, I do not observe whether or not each ANDA filing contains a paragraph IV certification because this information is not disclosed by the FDA. On the other hand, the FDA publishes a list of drug markets that were the subject of one or more ANDAs containing a paragraph IV certification. Therefore, it is possible to distinguish between paragraph IV markets and non-paragraph IV markets, and to see if firm behavior differs across the two groups. Our interest in this study is in seeing if paragraph IV patent challenges are associated with generic firms’ vertical integration decisions. How might such an association arise? As I argue in Section 2.3, when generic entry involves a race to be first, investments made by upstream API manufacturers tend to become specific to a particular downstream buyer. If contracts between unintegrated upstream suppliers and downstream buyers are incomplete and payoffs are determined through ex post bargaining, plant drying rack this increase in relationship specificity could enhance the role of vertical integration as a way to facilitate investments. In the empirical analysis, I examine whether the occurrence of paragraph IV certification at the market level is associated with higher incidence of vertical integration at the firm level.Before turning to the formal analysis, let us examine the pattern of vertical integration in the generics industry. Figure 2.1 shows how the prevalence of vertical integration at the market level has changed over time. It is based on the sample of 128 markets that opened up between 1993 and 2005. It can be seen that the average number of downstream entrants per market has remained stable at around five. On the other hand, the share of those downstream entrants that are vertically integrated has increased over time. For markets that opened up in the 1993-2000 period, the average share of vertically integrated entrants, as a percentage of the number of downstream entrants, was 8.1 percent. In 2001-2005, the figure rose to 24.1 percent and the difference between the sub-periods is highly significant . The incidence of vertical integration has similarly risen over time. In each of the years from 1993 to 2000, 24.0 percent of the sample markets opening up each year, on average, had one or more vertically integrated entrants. For the years 2001-2005, the average share of markets having any vertically integrated entry was 64.6 percent . An interesting fact about the US generic pharmaceutical industry is that it started off as being vertically separated. When the industry began its growth in the 1980s, finished formulation manufacturers procured most of their API requirements from outside suppliers located in Italy, Israel, and other foreign countries.

This was mainly due to differences in patent protection across countries: while strong patent protection in the US made it difficult for domestic companies to develop APIs before the expiration of originator patents, the weak patent regimes in Italy and other countries at the time allowed firms located there to develop generic APIs early . In addition to these historical origins, the nature of the generics business also made vertical separation a natural outcome. Different downstream manufacturers of generic drugs produce near identical products, because, by definition, they are all bio-equivalent to the original product. Therefore, the APIs manufactured by different upstream firms are also expected to be homogeneous. This implies that in general, investments into API development by an upstream manufacturer are not specific to a particular downstream user. In other words, the investment facilitation effects of vertical integration are unlikely to be important in this industry under normal circumstances. This is analogous to Hart and Tirole’s observation that the efficiency benefits of vertical integration were unlikely to have been strong in the cement and ready-mixed concrete industries during the 1960s when the vertical merger wave took place. Nevertheless, as Figure 2.1 demonstrates, vertical integration has become more prevalent over time in the generics industry. Several possible reasons for this can be found from industry reports. One is that early development and procurement of APIs has become more important to the profitability of downstream manufacturers in recent years, particularly in markets characterized by paragraph IV patent challenges. For example, the annual report of Teva, the industry’s largest firm, describes the motive for vertical integration as follows: “to provide us with early access to high quality active pharmaceutical ingredients and improve our profitability, in addition to further enhancing our R&D capabilities.” . Karwal mentions that “having access to a secure source of API can make a significant difference, particularly relating to difficult-to-develop API, when pursuing a potential Paragraph IV opportunity, and to secure sufficient quantities for development” . Similarly, Burck notes that “Access to API and control of the development and manufacturing process to support patent challenges has often been cited as a reason for backward integration” . These comments suggest that vertical integration allows downstream manufacturers to obtain APIs sooner than they otherwise would, and that this aids them in attaining first-to-file status in paragraph IV markets. This would partly explain why the increased prevalence in vertical integration appears to have followed closely behind the increase in paragraph IV patent challenges. A second possible cause of increased vertical integration pertains to bandwagon effects. A former purchasing executive at Sandoz, one of the largest firms, mentions that firms vertically integrate to “avoid sourcing API from a competitor” . Karwal points out that “Many key API suppliers, especially from India, China and Eastern Europe, are moving up the value chain and decreasing their supply activities, becoming direct competitors in finished form generics” . He suggests that this is one of the factors behind increased backward integration by established downstream manufacturers. In the mid-2000s, traditionally unintegrated US firms in the downstream segment began acquiring API manufacturing assets. Examples include the acquisition of Indian API manufacturers by Mylan and Watson, both large US finished formulation companies. It is important that these actions, by two of the main players of the industry, took place after vertically integrated entry became common.