Skip Navigation

How Increased Competition from Generic Drugs Has Affected Prices and Returns in the Pharmaceutical Industry July 1998

The Hatch-Waxman Act helped increase the supply of generic drugs by lowering the cost of getting them approved by the Food and Drug Administration. As a result of that act and structural changes in the demand for prescription drugs, more innovator drugs now face generic competition shortly after their patents expire. They then quickly lose over 40 percent of their market, on average, to generic drugs.

By themselves, the increase in generic market share and the acceleration of generic entry after patent expiration would have substantially reduced the returns from marketing an innovator drug. However, the Hatch-Waxman Act countered part of that effect by providing patent extensions for such drugs, which now average about three years. Those patent extensions offset part of the potential loss. But they do not completely protect the returns of brand-name manufacturers from the dramatic rise in market share for generic drugs.

The analysis in this chapter focuses on changes in patent protection for brand-name drugs as well as on supply-side factors that have boosted generic market share. As noted in Chapter 2, however, demand-side factors, such as the rise of managed care techniques, have also played a role. The Congressional Budget Office's estimate of changes in the returns from marketing a new drug takes those demand-side factors into account only through their contribution to the dramatic growth of generic market share since 1984.

The Hatch-Waxman Act has increased the likelihood that generic copies will become available once the patent on a brand-name drug expires. Before the act (in 1983), only 35 percent of the top-selling drugs no longer under patent had generic copies available.⁽¹⁾ Today, nearly all do.⁽²⁾ At the same time, the share of their market that those drugs lose to generic competitors has also expanded dramatically. In 1980, generic drugs accounted for only around 13 percent of the total quantity of prescriptions sold for multiple-source drugs (excluding antibiotics).⁽³⁾ Fourteen years later, they constituted 58 percent of the total quantity of multiple-source prescriptions dispensed (according to CBO's retail pharmacy data set). Pinpointing how much of that increase resulted solely from the Hatch-Waxman Act, however, is impossible.

For the minority of brand-name drugs that would have experienced generic competition even without the act, the average number of years they are on the market before facing generic competition did not change much. Before 1984, an average of three years elapsed between patent expiration and generic entry. By accelerating the approval process for generic drugs and explicitly permitting them to undergo clinical tests while the innovator drug is still under patent, the Hatch-Waxman Act now enables generic manufacturers to enter a market almost immediately after patent expiration. However, that decline of roughly three years in the average time before generic entry is almost exactly offset by the average increase in patent terms from Hatch-Waxman extensions.

CBO's analysis finds that despite the patent-term extensions and various exclusivity provisions of the Hatch-Waxman Act, the increase in generic market share since 1984 has decreased the total returns from marketing a new drug by about $27 million, on average. (That estimate does not apply to antibiotic drugs, which were not affected by the act.) In this study, the phrase "returns from marketing a new drug" refers to the expected average present discounted value of the total profit stream generated by introducing a new drug onto the market. Previous studies estimated that profit stream at an average of $210 million to $230 million (in 1990 dollars) for drugs introduced in the early 1980s.⁽⁴⁾ Those returns account for production costs but not the cost of research and development, which averaged about $200 million per drug (in 1990 dollars) when capitalized to the date of market introduction. Expressed as a percentage, the $27 million decline in returns equals roughly 12 percent of the total average returns from marketing a new drug. Despite that decline, those expected returns probably continue to cover the costs of developing a drug, on average, including the cost of capital.⁽⁵⁾
 

Changes to the Length of Patents for Brand-Name Drugs

Over the past 14 years, federal legislation--particularly the Hatch-Waxman Act of 1984 and the Uruguay Round Agreements Act of 1994--has altered the patent protection available to pharmaceutical products in the United States (see Table 7). The average length of time between when a brand-name drug enters the market and when its patent expires rose by more than two years--from an average of about nine years before 1984 to 11 to 12 years.⁽⁶⁾ By contrast, the period after that, between when the innovator drug's patent expires and when the first generic copy enters the market, declined from about three years to a few months. After patent expiration, sales of an innovator drug can decline significantly. Between 1984 and 1994, the average market share of generic drugs increased from around 13 percent to 58 percent of prescriptions dispensed for multiple-source drugs (except antibiotics).⁽⁷⁾
 

Determining the extent to which average patent terms have changed under the Hatch-Waxman Act is crucial to assessing whether the returns from marketing a new drug have largely been preserved despite the dramatic rise in generic competition. To that end, CBO analyzed data from the Patent and Trademark Office to evaluate the effect of Hatch-Waxman extensions on the average patent term of an innovator drug.

Patent Extensions Under the Hatch-Waxman Act

The Hatch-Waxman Act allows for patent extensions based on the amount of time a drug spends in the FDA review process. Those extensions now average about three years for new drugs.⁽⁸⁾ Technically, the length of a patent extension equals half of the time spent in clinical testing after the patent is granted, plus all of the time that the FDA spends reviewing the new drug application. (The clinical testing phase starts when the manufacturer files an investigational new drug application, which allows clinical testing in humans to take place.) Those extensions are subject to two limits. First, they cannot exceed five years. And second, they cannot allow the period between product approval and patent expiration to exceed 14 years.

Only one patent for each newly approved chemical entity is eligible for a Hatch-Waxman extension. If a drug has more than one patent, the manufacturer must choose which will receive the extension. Extensions are usually applied to the patent on a drug's chemical compound (a product patent) or occasionally to a patent on the use of the drug.⁽⁹⁾ Manufacturers must apply for an extension no more than 60 days after the FDA approves a drug for marketing.

For the 51 drugs approved between 1992 and 1995 that have received an extension, the average extension lasted 2.9 years. However, eight of those drugs were subject to a transitional two-year cap because they were undergoing clinical testing when the Hatch-Waxman Act became law. For the 43 drugs not subject to that cap, the average extension lasted 3.0 years (see Table 8).⁽¹⁰⁾ In all, the average patent term remaining after FDA approval for the 51 drugs that received extensions was 11.5 years.
 

Given the length of the clinical testing and NDA approval phases, those extensions would have averaged more than three years were it not for the 14-year cap. A study of the first 65 drugs to receive Hatch-Waxman extensions found that the total extension available under the act's formula, before applying the caps and other restrictions, averaged 4.5 years.⁽¹¹⁾ Almost half of those drugs would have been subject to the 14-year cap had the transitional two-year cap not applied. Similarly, about half of the 43 drugs introduced between 1992 and 1995 that received Hatch-Waxman extensions and were not limited by the transitional cap had their extensions limited by the 14-year cap (see Table 9). Only 10 drugs had their extensions limited by the five-year cap.
 

Not all drugs obtain a Hatch-Waxman extension. The FDA approved a total of 101 drugs containing new chemical compounds between 1992 and 1995, but only half (51) have received a Hatch-Waxman extension so far. Another 12 have an application pending (see Table 10). Of the remaining 38 drugs, 19 had no patent to extend. Fifteen others already had 14 years left under patent when they were approved by the FDA. And four drugs did not apply for an extension, for reasons that could not be determined.
 

Nonpatent Exclusivity Periods Under the Hatch-Waxman Act

In addition to extending patent terms, the act grants special periods of exclusivity in two circumstances (not including some of its transitional features). First, when the FDA approves a new chemical entity, no application for a generic copy is accepted for a minimum of five years. That provision benefits drugs that have no patent, or have a very short remaining patent life when they are approved, because it means that generic manufacturers must wait five years before filing an abbreviated new drug application. Since the approval process for such applications takes more than 30 months, on average, many of those brand-name drugs should actually have six to seven years of exclusivity before they must face generic competition.⁽¹²⁾ In most cases, however, that period is probably too short to fully recover the average costs of drug development.

Second, the act allows the FDA to grant three years of market exclusivity for an NDA (including a supplemental one) if that application requires new clinical investigations. Manufacturers can use NDAs or supplemental NDAs to obtain approval for new dosage forms of an already-approved drug, for a new use, or for marketing the drug over the counter. Those provisions give manufacturers an incentive to continue improving brand-name drugs, and the knowledge about those drugs, after they are on the market.

Manufacturers can also use those provisions to slow generic competition. By introducing a new dosage form just before patent expiration, a manufacturer obtains three years of market exclusivity for the new product under the Hatch-Waxman Act (although generic manufacturers can still copy the original form of the drug). Likewise, if a drug starts being sold over the counter, it enjoys three years of exclusivity before the FDA can accept abbreviated applications for generic over-the-counter versions. The over-the-counter versions of Zantac and Tagamet, for example, have benefitted from that provision. Sometimes, a manufacturer can obtain a separate patent on a new dosage form--particularly an extended-release form. For example, the patent for the active ingredient in Procardia expired in 1991, but the patents for the extended-release version, Procardia XL, do not expire until 2000 or later.⁽¹³⁾

The Effect of Those Changes on the Average Drug

To assess the change in returns from marketing a new drug, analysts need to know the average effect of the Hatch-Waxman Act on all brand-name drugs approved, not just on those that obtain an extension. When the benefits of the act's patent extensions and five-year exclusivity period are averaged over all drugs approved between 1992 and 1995, the average effect is to postpone generic entry by 2.8 years.

CBO calculated that effect as follows. As Table 8 shows, extensions averaged three years for the 43 drugs receiving a Hatch-Waxman extension during that period that were not subject to the transitional two-year cap. Since the transitional cap applies only to drugs in clinical testing in 1984, it will eventually disappear. Therefore, the calculation attributes three years of patent exclusivity to all 51 drugs that received a Hatch-Waxman extension. It also assumes that the 12 drugs with extension applications pending will receive an average extension of three years.

Of the 19 drugs that had no patent to extend, nine were excluded from the calculation because they were "orphan" drugs (those with a potentially small market because of the medical condition they treat), which received seven years of exclusivity under the Orphan Drug Act. The other 10 unpatented drugs were entitled under the Hatch-Waxman Act to five years of exclusivity, during which no generic manufacturer could file an abbreviated application with the FDA. Since it takes at least one year for a generic manufacturer to obtain FDA approval, that exclusivity provision effectively postpones generic entry by at least six years. Thus, the calculation attributes six years of delay in generic entry for those drugs under the act.

The average was taken over the number of new drugs approved between 1992 and 1995, after subtracting the nine orphan drugs and the four drugs that did not apply for an extension but were eligible. Mathematically, the formula is:
 

That average does not take into account the exclusivity periods for new dosage forms. As explained below, CBO accounted for those exclusivity periods in its calculation of returns from marketing by including dosage forms that have no generic versions available in its estimate of average generic market share following patent expiration.

The Effect of the Uruguay Round Agreements Act

Ten years after the Hatch-Waxman Act, another piece of legislation, the Uruguay Round Agreements Act of 1994 (URAA), affected patent terms for brand-name drugs. That act changed the length of U.S. patents on all types of inventions to 20 years from the date of application rather than 17 years from the date the patent is granted. That change has had only a very small effect on the average "effective" patent term--the time between FDA approval and patent expiration--for drugs patented after June 8, 1995 (most of which have yet to be introduced on the market). Drugs already patented by June 8, 1995, may benefit from the change as their manufacturers can choose between the 17-year and 20-year terms and still obtain a Hatch-Waxman extension.⁽¹⁴⁾

So-called patent pendency periods (the time between applying for a patent and receiving it) vary considerably among drugs. Of the 100 top-selling drugs in 1996, 45 were granted patent-term extensions under the Hatch-Waxman Act. CBO found that the patent pendency period for those 45 drugs averaged 3.3 years.⁽¹⁵⁾ That implies that the new 20-years-from-filing term should have a slightly negative effect for drugs patented after June 8, 1995. The URAA's effect on patent terms interacts with the rules in the Hatch-Waxman Act used to calculate extensions. On net, CBO estimates, those 45 drugs would have lost an average of almost four months of patent life if the 20-years-from-filing term was applied universally.⁽¹⁶⁾

Companies can file a provisional patent application that establishes priority for their invention but does not start the patent-term clock. They must then file a full application within one year.⁽¹⁷⁾ If companies take advantage of that provisional application, the negative effect of the 20-years-from-filing term could be slightly offset. Firms may also change their behavior in other ways that could speed up the time between patent application and patent grant. For those reasons, CBO assumed in calculating the change in returns from marketing that the URAA had no net impact on effective patent terms.

Some patents that were about to expire under the 17-year term had their expiration dates postponed under the 20-year term established by the URAA. For those patents, a transitional feature in the act allows generic manufacturers to enter a market after the 17-year term expires if the generic manufacturer had already undertaken a substantial investment.⁽¹⁸⁾ However, because of complications in the way the URAA interacts with the Hatch-Waxman Act, that transitional feature does not apply to pharmaceutical products.⁽¹⁹⁾ Some Members tried during the 104th Congress to pass legislation allowing earlier generic entry in the pharmaceutical market in cases in which substantial investment had already been made, but that effort was unsuccessful.
 

Changes to the Approval Process for Generic Drugs

The Hatch-Waxman Act made two key changes that allow generic manufacturers to obtain FDA approval more quickly once the patent on an innovator drug has expired. First, it established an abbreviated approval process for generic copies of innovator drugs that were approved after 1962. Second, it allowed generic manufacturers to conduct the tests required for FDA approval before the innovator drug's patent expired. Those changes shortened the average time between patent expiration and generic entry for top-selling drugs from three or four years to less than three months. That acceleration of generic entry helps consumers by making lower-cost drugs available more quickly. It also roughly offsets the average 2.8-year delay in generic entry provided by the patent-term extensions and exclusivity provisions in the Hatch-Waxman Act.

Before the act took effect, the FDA had two types of application processes for approving generic copies of innovator drugs. When copying an innovator drug that had been approved before October 1962, the generic manufacturer had only to demonstrate bioequivalence through clinical tests. When copying an innovator drug approved after 1962, the generic manufacturer also had to demonstrate safety and efficacy. The tests necessary to demonstrate a drug's bioequivalence are much less costly than those required to prove its safety and efficacy.⁽²⁰⁾ In some instances, the FDA accepted a literature review of published reports in lieu of safety and efficacy tests; such applications were called "paper NDAs."⁽²¹⁾ However, in many cases, sufficient evidence was not available in published reports.⁽²²⁾ After the first generic copy of a drug was approved, subsequent applications by generic manufacturers could more easily substitute a literature review for safety and efficacy tests.

In the case of antibiotics, the distinction between pre- and post-1962 drugs did not exist. An abbreviated process for approving generic antibiotics, which required clinical tests to show only bioequivalence, applied to all antibiotic drugs approved under section 507 of the Federal Food, Drug, and Cosmetic Act. Since an abbreviated approval process for generics already existed, such antibiotics were not included in the Hatch-Waxman provisions and were not eligible for patent-term extensions under the act. However, the Food and Drug Administration Modernization Act of 1997 made antibiotic drugs eligible for Hatch-Waxman extensions, thus increasing the returns from their development.

In essence, the Hatch-Waxman Act extended the abbreviated process for approving antibiotics (as well as generic copies of innovator drugs approved before 1962) to all generic drugs. Generic manufacturers now file an abbreviated new drug application, which requires that they perform clinical tests only to demonstrate that their drug is bioequivalent to a drug with an approved NDA that is already on the market. The FDA relies on the safety and effectiveness determination for that original drug when approving the generic copy.

To further speed up the process, the Hatch-Waxman Act explicitly allows generic manufacturers to begin those clinical tests before the original drug's patent expires. In most cases, that change lets manufacturers obtain FDA approval and begin selling copies of an innovator drug soon after patent expiration. Prior to the Hatch-Waxman Act, generic testing occasionally occurred before patent expiration; it was subject to legal dispute until the Court of Appeals for the Federal Circuit ruled in 1984 that such tests infringed on the patent of the innovator drug.⁽²³⁾ The Hatch-Waxman Act effectively reversed that decision by stating that generic manufacturers can begin the FDA approval process before patent expiration. By including the patent expiration date in its application, the generic firm makes explicit its intention not to market the new product until after patent expiration. For its part, the FDA will not approve a new generic drug until the innovator's patent has expired (unless the generic applicant successfully challenges that patent in court).⁽²⁴⁾

Before the Hatch-Waxman Act, an average of three to four years elapsed between patent expiration and generic entry. CBO identified 15 cases before 1984 in which one or more generic manufacturers had obtained FDA approval to produce a post-1962 drug by filing a new drug application. For the 11 cases in which a patent expiration date was identified, the average time between patent expiration and generic entry was 3.1 years. In six of those cases, the NDA was applied for before patent expiration. In the other five cases (in which the NDA was applied for after patent expiration), the average time between patent expiration and generic entry was 3.9 years.

For those 15 drugs, generic entry occurred, on average, 1.8 years after the filing of an application. The approval process for those drugs actually took longer than that because before filing an NDA, the generic manufacturers had to research the formulation, contact a chemical manufacturer who could produce the active ingredient, search the literature for preclinical and clinical data, conduct a bioequivalence study, and perhaps demonstrate safety and efficacy as well. Although some of those steps could be taken before patent expiration, the Roche v. Bolar decision required that no clinical tests be conducted until afterward.

As an indication of how much more quickly generic entry occurs since the Hatch-Waxman Act, CBO examined 17 brand-name drugs that lost their patent protection between 1990 and 1993, most of which had annual U.S. sales of $50 million or more. For most of those drugs, generic entry occurred within one or two months of patent expiration, although there were exceptions (see Appendix C for more details).⁽²⁵⁾
 

Effects on the Returns from Marketing a Drug

Makers of innovator drugs were slightly worse off after the Hatch-Waxman Act, largely because many more of their drugs experienced generic competition following patent expiration. The act's provision for extending patent terms merely compensated for the loss of the average three-year delay between patent expiration and generic entry that existed before the act (in cases where generic entry occurred).

Still, those extensions played an important role in protecting the returns from drug companies' research and development. Without them, the rise in generic market share since 1984 would have dramatically lowered the expected returns from marketing a drug and might have caused the pharmaceutical industry to reduce its investment in R&D. In that case, a successful innovator drug would have been likely to lose over 40 percent of its market to generic competitors just after reaching its peak year in sales. If the pre-1984 level of R&D investment was desirable, then the patent extensions benefited society by preserving most of the returns from marketing a new drug.

This study uses as a benchmark the average returns from marketing a new drug in the early 1980s under the modest levels of generic entry that existed then. The analysis estimates how much returns have declined relative to that benchmark because innovator drugs (excluding antibiotics) are losing a larger share of their market to generic competitors after patent expiration. Whether the benchmark level of returns is the best one for society is a separate question, which this study does not address.

When a brand-name drug first comes on the market, its sales revenues are low because its benefits are not yet widely known. As the drug becomes better known through published articles, advertising in medical journals, and detailing, its sales rise and reach their peak by year nine or 10, on average. Both before and after 1984, the average innovator drug had a few years of sales at its peak level before generic manufacturers entered the market.

The Hatch-Waxman Act did not greatly change the average point in a drug's life at which generic entry occurs, because the act's patent-term extensions and five-year exclusivity provision together postponed generic entry by roughly the same amount that the act's streamlined approval process sped it up. Two things that did change after 1984 were the likelihood that generics would become available and the average market share captured by generic drugs. Thus, on net, one would expect returns from marketing a new drug to decline after the Hatch-Waxman Act, because although the timing of generic entry has not changed much, the probability of generic entry and the size of the generic market once entry occurs have grown.

Calculating the Change in Returns

CBO estimated the effect of increased generic competition on the stream of profits generated from the sale of 67 innovator drugs that were introduced in the United States in the early 1980s.⁽²⁶⁾ The data include U.S. sales revenues from 1980 to 1991, covering the first eight to 12 years that those drugs were on the market. The average patent term for the drugs, weighted by sales revenues, was 11 years.

CBO's calculation assumes that the profit stream for an average brand-name drug (excluding antibiotics) would have been the same for the first 11 years with or without the Hatch-Waxman Act. (For more details about the assumptions behind the calculation, see Appendix C.) The total profit stream over the drug's product life is depicted in Figure 7 by the area under the solid curve between year 0 (market introduction) and year 20 (when the drug has become nearly obsolete). The present discounted value of that profit stream, discounted to the date of market introduction, represents the returns from marketing the drug. The negative cash flow before drug introduction represents the investments made in the drug's development. Capitalizing those costs to the date of market introduction brings their total to about $200 million.
 

Figure 7. The Average Profit Stream for a Brand-Name Drug Before and After the Hatch-Waxman Act

SOURCE: Congressional Budget Office.

NOTE: This figure is intended to be illustrative and does not reflect the actual dollar amounts invested in research and development (R&D) or the actual value of profits from drug development.

a. That increased generic competition did not result solely from changes in the Hatch-Waxman Act. Other developments, such as the use of formularies by private-sector health plans to increase generic substitution, also affected the degree to which generic drugs have eroded the profits of off-patent brand-name drugs.

For years 12 to 20, CBO estimated two revenue paths, one before and one after the Hatch-Waxman Act. The only difference between those two revenue paths is in the amount of sales revenues lost to competition from generic drugs. Sales revenues also decline in later years because of competition from newer, improved brand-name drugs. CBO assumed that decline to be the same before and after 1984. The pre-1984 path assumes that the drug's patent expires at the end of year 11 but that it takes three years for generics to enter the market, consistent with the data for that period. Therefore, profits do not begin to decline because of generic entry until after year 14. But the decline after year 14 is gradual because generic market share was small for nonantibiotic drugs before 1984.

In the post-1984 path, the Hatch-Waxman Act extends patents by 2.8 years. Generics are assumed to enter about a month later and begin taking a large share of the market. For any specific drug, the size of the generic market and whether generic entry occurs at all will vary. The rate at which profits are eroded depends on whether generic entry occurs and, if so, on the size of the generic market. For the average drug, however, profits erode much more rapidly in this case than before the Hatch-Waxman Act because of greater generic competition. In either case, the effect of increased generic entry on the returns from marketing a new drug is less than one might expect because generic entry occurs at the end of a drug's product life, when profits are more heavily discounted (in other words, worth less today because they occur farther in the future).

CBO used the actual stream of sales revenues through year 11 for the 67 innovator drugs it examined as the starting point for its calculation. For the pre-1984 profit stream, it applied a rate of sales erosion after generic entry that was based on a sample of 29 top-selling, multiple-source, nonantibiotic drugs in 1980.⁽²⁷⁾ The erosion rate for the post-1984 case was based on this study's analysis of generic market share in 1993 and 1994.

The total difference between the two profit streams has a present discounted value of $27 million (in 1990 dollars), CBO estimates. In other words, despite the patent extensions and exclusivity provisions in the Hatch-Waxman Act, the growth in generic market share since 1984 has reduced the present discounted value of the returns from marketing a new drug by about $27 million, on average. That figure should be compared with the present discounted value of the total profit stream from marketing an innovator drug throughout its product life, discounted to the date of market introduction, which previous studies have estimated to average $210 million to $230 million for drugs introduced in the 1980s. (Those returns account for production costs but not the capitalized costs of drug development. They include profits from sales abroad, which make up roughly half of total returns.) Expressed as a percentage of those returns, the present discounted value of the returns from marketing a new drug have declined by roughly 12 percent. That result holds true even with modest variations in the assumptions (see the sensitivity analysis in Appendix C).

Grabowski and Vernon and the Office of Technology Assessment estimated that the present discounted value of the returns from marketing a drug exceeded the capitalized costs of R&D by $22 million to $36 million.⁽²⁸⁾ That is, investment in R&D earned a return slightly higher than the cost of capital, on average. The drugs in those studies did not obtain patent-term extensions under the Hatch-Waxman Act because they were introduced before the act was passed. But they did face increased generic competition once their patents expired. On average, therefore, the returns from marketing a new drug would probably still fully cover the capitalized costs of R&D despite the increase in generic sales since 1984. On the margin, however, a few drugs that were barely profitable to develop would no longer be profitable.

Caveats About CBO's Estimate

CBO's estimated change in returns from marketing a new drug accounts for the full impact of increased generic entry since 1984. But it does not account for many changes in the pharmaceutical market that could increase or decrease those returns, such as changes in R&D costs, in technology, or in the overall demand for prescription drugs. Thus, the estimate is only a partial one, which focuses on the effects of the Hatch-Waxman Act and increased generic sales.

Moreover, since the calculation is based on the U.S. sales of drugs during the 1980-1991 period, it does not include the effects of changes in the pharmaceutical market since then (other than increased generic entry). Some of those changes would raise the returns from marketing a new drug; others would lower them. The rise in managed care since 1991 and its impact on the returns from marketing a new drug are considered only through their effect on increased generic market share. The impact of managed care on the volume of drugs purchased or the prices charged by manufacturers has not been considered. In addition, manufacturers selectively offer discounts and rebates on innovator drugs, but those rebates and some of the discounts are not captured by the data on sales revenues, which are based on average invoice prices.

Other factors not included in the estimate could increase the returns from marketing a new drug. For example, the over-65 population, which has a high use of prescription drugs, is growing more rapidly now than it was 10 years ago. In addition, some Medicare beneficiaries are moving into HMOs. Since traditional Medicare does not offer an outpatient drug benefit but many HMOs do, the effect of those moves is to increase prescription drug coverage for the over-65 population.⁽²⁹⁾ As noted in Chapter 2, managed care techniques may also boost the volume of prescription drugs used by people under 65.

In addition, foreign markets for pharmaceutical products will probably continue to grow as the drug-approval process becomes streamlined in Europe and as various countries strengthen their patent-protection rights.⁽³⁰⁾ The Agreement on Trade-Related Aspects of Intellectual Property Rights, which was negotiated in 1994 at the Uruguay Round of the General Agreement on Tariffs and Trade, included provisions to encourage developing countries to strengthen their intellectual property rights, particularly in the areas of agriculture and pharmaceuticals. That agreement provides patented pharmaceutical products with a minimum of five years of exclusivity in a participating developing country.⁽³¹⁾

The net effect of changes not accounted for in CBO's estimate may push the total returns from marketing a new drug in one direction or the other. Overall, however, spending on R&D by brand-name manufacturers has increased as a percentage of their sales revenues--from an average of 14.7 percent in 1983 to 19.4 percent in 1995 (despite the fact that such revenues more than tripled).⁽³²⁾ That increase would seem to indicate that, all factors taken together, the incentive to invest in developing new drugs has remained intact since the Hatch-Waxman Act.

No one knows whether that amount of investment in R&D is over or under the optimal level.⁽³³⁾ Some people might argue that companies are not investing enough in drug development and that society would be better off if returns from marketing were increased further. Clearly, the avoided surgery and improved quality of life that result from the use of prescription drugs create large benefits for many people. But it is also possible that too many firms invest in the same research projects, and less could be spent on pharmaceutical R&D without significant costs to society.

Other Considerations

CBO's estimate of the average change in returns from marketing a new drug is small relative to the returns earned on highly successful drugs. The reason is that returns from marketing new drugs are highly skewed. The top six drugs in the set of 67 that CBO used in its calculation earned a return of around $1 billion (discounted to the date of market introduction). But only the top 20 earned a return from marketing that exceeded $200 million, roughly the average cost of drug development.⁽³⁴⁾ However, since the cost of developing drugs includes the cost of failures, a drug can be profitable in the sense of covering its own development costs but still not earn enough to cover average development costs (which include the cost of drugs that never made it to market). A company must discover a highly profitable drug from time to time for its average returns from marketing to exceed the average capitalized cost of drug development.

Another factor to consider, which can reduce the impact of lower returns, is the so-called replacement effect. When a manufacturer introduces a new brand-name drug, that drug may erode the sales of similar drugs the company already has on the market. CBO's estimate of the decline in the present discounted value of the returns from marketing a new drug does not consider the dynamic effect of such product replacement. The replacement effect derives from the reduced incentive that companies have to innovate when a new drug will replace a share of the market currently held by one of their other products. (For more details about that effect, see Appendix D.) The rise in generic market share, however, reduces the replacement effect. A firm has less to lose by replacing an older product with a new drug when the patent on the older product is about to expire, since generics will take away a large share of that product's market anyway.

An example is the allergy drug Allegra, introduced in 1996 by Hoechst Marion Roussel, which also sells a competing brand-name drug, Seldane. The two drugs are very similar antihistamines, but Allegra has fewer negative side effects. Because of the replacement effect, Hoechst Marion Roussel had less incentive to introduce Allegra when it would cut into the profits from the sale of Seldane significantly. However, anticipation of generic competition reduced that replacement effect--Allegra was introduced just three years before Seldane's patent was to expire.⁽³⁵⁾

Although the growth of generic competition since 1984 has reduced the returns from innovation overall, the effect of those lower returns on the incentive to innovate will be offset somewhat by a commensurate reduction in the replacement effect. That is, the slightly reduced value of profits at the end of a drug's product life will give firms with existing products a greater incentive to replace them in the market more quickly--as close to patent expiration as possible.

That dynamic effect exists only when pharmaceutical firms continue to invest in developing drugs in therapeutic areas where they are already market leaders. Large firms usually conduct R&D in a variety of therapeutic areas, so the dynamic effect will be greater for some projects and nonexistent for others.⁽³⁶⁾ The operation of the replacement effect reduces--but does not eliminate--the negative impact that the rise in generic market share has on the incentive to invest in developing brand-name drugs.
 

Effects of Proposed Changes to the Hatch-Waxman Act

Some representatives of the pharmaceutical industry would like to modify the Hatch-Waxman Act in various ways to increase the average effective patent term for pharmaceutical products.⁽³⁷⁾ Although lengthening patents would increase profits today for drugs whose patents are expiring, it would not have as large an impact on the incentive to invest in R&D--that is, on the present discounted value of the returns from marketing a new drug. Extending the average effective patent term by one year would increase the present discounted value of those returns by about $12 million.

In contrast, accelerating the FDA review period by one year would have a much greater effect on the present discounted value of the returns from marketing a new drug--a net benefit of about $22 million, on average. Thus, reducing FDA approval times--if it could be done without sacrificing safety concerns--would be much more effective in helping both the drug industry and consumers than would lengthening the patent-protection period.

Some drugs do not benefit from patent-term extensions because they have no patent to extend, or because their patent has already expired (perhaps because the drug lingered in the clinical testing phase). Lengthening the five-year exclusivity period for a new drug (that contains a chemical entity never before approved) would have a sizable impact on the incentive to develop those drugs, because the benefits would be seen relatively early in the drug's product life. Furthermore, the current exclusivity period is probably too short to compensate for the average cost of developing those drugs. Out of the 101 drugs approved between 1992 and 1995, 10 would have benefited from a lengthening of the five-year exclusivity period.
 

Conclusions

The Hatch-Waxman Act eliminated the duplicative testing requirements for manufacturers of generic drugs to obtain FDA approval. That regulatory relief has translated into greater availability of generic drugs and lower average prices to consumers for off-patent drugs. By itself, the doubling of generic market share between 1984 and 1994 would have substantially lowered the returns from marketing new innovator drugs. However, the act also provided patent extensions that postponed the time when an innovator drug would face generic competition.

CBO's analysis has found that the patent extensions available under the Hatch-Waxman Act were not sufficient to fully preserve the returns from marketing new brand-name drugs. The present discounted value of those returns has declined by about 12 percent because of the rise in generic competition. However, that rise has resulted from a variety of demand-side factors as well as from changes in the act itself.

The Hatch-Waxman Act helped increase the opportunity to substitute less expensive generic drugs for more expensive off-patent brand-name drugs. That substitution lowers the average cost of a multiple-source prescription drug. The point in the life of an average drug at which generic entry occurs did not change much under the act, because the average length of a patent extension roughly offsets the average delay between patent expiration and generic entry that existed before 1984. Of course, that specific timing varies significantly from one drug to another. Nevertheless, many purchasers are better off since the act, as most top-selling off-patent brand-name drugs now have generic versions available. And with the lower testing costs required for FDA approval, more generic manufacturers probably find it profitable to enter a given market. Empirical evidence suggests that that puts downward pressure on the average prescription price of generic drugs as well.

Many changes in the pharmaceutical market and in the technology of drug development have affected the returns from marketing a new drug. This study considered only two changes that affect those returns: the increase in generic market share since 1984 and the increase in patent terms under the Hatch-Waxman Act. Changes that were not considered may, taken together, either increase or decrease those returns. Overall, it appears that the incentives for drug companies to innovate have remained intact since the Hatch-Waxman Act; even as sales revenues from innovator drugs have more than tripled, the percentage of those revenues that manufacturers reinvest in R&D has risen from 14.7 percent to 19.4 percent between 1983 and 1995.

1. That figure is based on the top 200 off-patent drugs that year, excluding antibiotics and drugs that were approved before 1962; see Henry Grabowski and John Vernon, "Longer Patents for Lower Imitation Barriers: The 1984 Drug Act," American Economic Review, vol. 76, no. 2 (May 1986), pp. 195-198.

2. For example, in 1994, 95 percent of the off-patent drugs with sales revenues of $40 million or more in CBO's retail pharmacy data set had generic copies available. In that case, off-patent drugs were ones that were not protected by a patent or an exclusivity provision.

3. CBO calculated that average based on 29 nonantibiotic multiple-source drugs that were among the top 100 in U.S. sales, using data from Alison Masson and Robert Steiner, Generic Substitution and Prescription Drug Prices: Economic Effects of State Drug Product Selection Laws (Federal Trade Commission, October 1985), pp. 251-269. See Appendix C of this study for details.

4. See Office of Technology Assessment, Pharmaceutical R&D: Costs, Risks and Rewards (February 1993); and Henry G. Grabowksi and John M. Vernon, "Returns to R&D on New Drug Introductions in the 1980s," Journal of Health Economics, vol. 13, no. 4 (December 1994), pp. 383-406.

5. Ibid. Those two studies found that the present discounted value of the returns from marketing a drug exceeded the capitalized costs of drug development by an average of $22 million to $36 million for drugs introduced in the early 1980s.

6. According to data that CBO obtained from the Patent and Trademark Office, the average patent term remaining after FDA approval was 11.5 years for the 51 drugs approved between 1992 and 1995 that received a Hatch-Waxman extension. For drugs approved between 1978 and 1982, the average patent term remaining was just over nine years, according to Office of Technology Assessment, Pharmaceutical R&D, p. 83.

7. According to CBO's retail pharmacy data set, generic drugs accounted for 36 percent of all retail prescriptions dispensed in 1994 and 58 percent of prescriptions dispensed for multiple-source drugs. Excluding the few multiple-source antibiotic drugs from the data does not particularly affect that average.

8. The average extension for drugs approved before 1992 was less than that because a transitional two-year cap applied to drugs that were in clinical testing when the Hatch-Waxman Act became law. Drugs whose clinical testing began before September 24, 1984, were limited to two years of patent extension, and drugs that were already on the market by that date were not eligible for any patent extensions. However, drugs approved between January 1, 1982, and September 23, 1984, were eligible for 10 years of market exclusivity before an abbreviated new drug application could be submitted to the FDA by a generic manufacturer.

9. A third type of patent, called a process patent, also exists. Since it may not be difficult to formulate a similar compound using a slightly different chemical process, those types of patents do not necessarily prevent generic entry; personal communication by Peter Richardson, chief patent attorney, Pfizer, May 1997.

10. A study by Henry Grabowski and John Vernon found that for about 70 innovator drug products whose patents expired between 1991 and 1993, Hatch-Waxman extensions averaged 2.4 years. Some of those drugs were subject to the transitional two-year cap. See Grabowski and Vernon, "Longer Patents for Increased Generic Competition in the U.S.: The Hatch-Waxman Act After One Decade," PharmacoEconomics (1996).

11. The average clinical testing period for those drugs lasted 5.1 years. After subtracting the time between the beginning of clinical tests and the issuing of the patent, that period came to 3.8 years, half of which is counted when calculating the extension. The average NDA approval phase for those 65 drugs was 2.6 years, for a total average potential extension of 4.5 years. See Alan D. Lourie, "A Review of Recent Patent Term Data," Journal of the Patent and Trademark Office Society (February 1989), pp. 171-176.

12. In 1995 and 1996, an average of 33 to 34 months elapsed between the submission and final approval of abbreviated NDAs; see Department of Health and Human Services, Food and Drug Administration, Justification of Estimates for Appropriations Committees (1997 and 1998).

13. Department of Health and Human Services, Food and Drug Administration, "Approved Drug Products with Therapeutic Equivalence Evaluations," January 31, 1998 (available at http://www.fda.gov).

14. According to a 1996 ruling by the U.S. Circuit Court, products patented before June 8, 1995, that were already into their Hatch-Waxman extension period on that date are not eligible for the new 20-year patent term under the URAA.

15. Based on data on patent pendency periods provided by Pfizer and data on regulatory review periods and patent-term extensions from the Patent and Trademark Office.

16. Henry Grabowski and John Vernon found that the average patent pendency period for 105 drugs approved between 1990 and 1995 that received Hatch-Waxman extensions was 3.8 years. The overall effect of the URAA, when interacted with the Hatch-Waxman extensions, was a loss of 0.34 years. See Grabowski and Vernon, "Effective Patent Life in Pharmaceuticals," International Journal of Technology Management (forthcoming).

17. Title V, section 532(b)(1) of the URAA pertains to provisional applications and the right of priority (see 35 U.S.C. 119(e)(1), 108 Stat. 4985). Section 532(a)(1) defines the new 20-year patent term (see 35 U.S.C. 154(a)(2), 108 Stat. 4984).

18. The generic manufacturer must pay an equitable remuneration to the patent holder (see 35 U.S.C. 154(c)(2) and (3), 108 Stat. 4985).

19. It also does not apply to other products reviewed by the FDA that are eligible for Hatch-Waxman extensions--namely, biological products, food and color additives, and medical devices.

20. Grabowski and Vernon, "Longer Patents for Lower Imitation Barriers."

21. See Donald O. Beers, Generic and Innovator Drugs: A Guide to FDA Approval Requirements, 4th ed. (Englewood Cliffs, N.J.: Aspen Publishers, 1995), pp. 3-59 to 3-71.

22. House Committee on Energy and Commerce, Report on the Drug Price Competition and Patent Term Restoration Act of 1984 (June 21, 1984), pp. 16-17. According to that report, the FDA estimated that sufficient published evidence was not available for 85 percent of all post-1962 drugs.

23. The case was Roche Products, Inc. v. Bolar Pharmaceutical Company, Inc. (733 F. 2d 858 Federal Circuit 1984). See Alan D. Lourie, "Patent Term Restoration," Journal of the Patent Office Society, vol. 66, no. 10 (October 1984), pp. 526-550; and Beers, Generic and Innovator Drugs, pp. 4-75 to 4-77.

24. The process for a generic applicant to challenge an innovator's patent is discussed in 21 U.S.C. 355(j)(2)(A)(vii), paragraph IV, and section 355(j)(5)(B)(iii) of the Federal Food, Drug, and Cosmetic Act of 1938, as amended.

25. The date of generic entry came from Table 1 of Grabowski and Vernon, "Longer Patents for Increased Generic Competition in the U.S."

26. Data on average annual U.S. sales of those drugs were provided by Henry Grabowski of Duke University. The analytical approach is based in part on Grabowksi and Vernon, "Longer Patents for Lower Imitation Barriers."

27. Masson and Steiner, Generic Substitution and Prescription Drug Prices, Appendix A5.

28. Grabowski and Vernon, "Returns to R&D on New Drug Introductions in the 1980s," pp. 383-406; and Office of Technology Assessment, Pharmaceutical R&D.

29. In 1997, 4.5 million out of 38.2 million Medicare beneficiaries were enrolled in an HMO or risk-based health plan. CBO projects that the proportion of Medicare beneficiaries enrolled in such plans will continue to grow. See Congressional Budget Office, The Economic and Budget Outlook: Fiscal Years 1999-2008 (January 1998), Appendix F.

30. Standard & Poor's, Healthcare: Pharmaceuticals, Industry Surveys (New York: Standard & Poor's, August 29, 1996), p. 21.

31. See Dorothy Schrader, Intellectual Property Provisions of the GATT 1994 and the Uruguay Round Agreements Act, CRS Report for Congress 94-302A (Congressional Research Service, September 23, 1996), pp. 36-37.

32. Pharmaceutical Researchers and Manufacturers of America, 1997 Industry Profile (Washington, D.C.: PhRMA, March 1997), p. 57. Those figures equal R&D spending in the United States divided by domestic sales plus exports.

33. See F.M. Scherer, "Pricing, Profits and Technological Progress in the Pharmaceutical Industry," Journal of Economic Perspectives, vol. 7, no. 3 (Summer 1993), p. 111.

34. Grabowski and Vernon, "Returns to R&D on New Drug Introductions in the 1980s," pp. 398-400.

35. Department of Health and Human Services, Food and Drug Administration, Approved Drug Products with Therapeutic Equivalence Evaluations (1997). Seldane's patent expires in April 1999; Allegra was introduced in July 1996.

36. For a discussion of the diversity of R&D projects within a single firm and the benefits of such diversification, see Rebecca Henderson and Ian Cockburn, "Scale, Scope and Spillovers: The Determinants of Research Productivity in Drug Discovery," RAND Journal of Economics, vol. 27, no. 1 (Spring 1996).

37. See testimony at the Senate Judiciary Committee's hearing on the Hatch-Waxman Act on March 5, 1996.

---