Introduction¹

Invention patents are accorded particular importance in the pharmaceutical industry – a role both complicated and enhanced by the adoption of the TRIPS (Trade-Related Intellectual Property Rights) agreement under the Treaty of Marrakesh in 1995. Before TRIPS was adopted, many nations did not grant patents on pharmaceuticals, and especially on pharmaceutical products. TRIPS established deadlines for the universal granting of pharmaceutical product patents. Among 32 mostly high- and medium-income nations surveyed by Edson Kondo (1994: 62), 13 – Argentina, Brazil, Chile, Colombia, Ecuador, Greece, India, Mexico, Peru, Portugal, Spain, Thailand, and Venezuela – did not grant pharmaceutical product patents as of 1990. By 1996, the most highly developed World Trade Organization (WTO) members were expected to begin granting such patents, and the least developed nations were required to comply by 2006 (later extended to 2016). (International agency terminology favors using “least developed” to characterize “low-income” nations. In the context used here, “low-income” is more descriptive, even if not always politically correct.)

Patents are significant in pharmaceutical research and development for two main reasons.² First, focusing usually on a precisely defined chemical or biological molecule and its uses, they provide particularly clear and unambiguous property rights, less muddled by the ambiguities that plague mechanical, electrical, and communications apparatus patents. But second, numerous surveys have shown that having patent protection is particularly important in decisions to invest R&D resources to discover and prove the therapeutic efficacy of pharmaceutical molecules. In the first of several such studies, Taylor and Silberston (1973) asked industrial interviewees in England what negative impact on their R&D expenditures a legal regime of requiring world-wide compulsory patent licensing at “reasonable” royalties would have.³ The weighted average reduction for all industries was projected to be 8%, but pharmaceuticals was the outlier, with expected R&D cuts of 65%. There are two main reasons for this difference. For one, pharmaceutical R&D is particularly expensive, requiring discovery and testing outlays measured in the hundreds of millions of dollars to achieve a single new marketable molecule during the early years of the twenty-first century (see Scherer 2014). But second, much of the relevant expenditure is for what is best described as information, assumed to approximate unusually closely to a pure public good. Once a promising new molecule is synthesized and patented, its structure is there for all to see once patent applications are published. Moreover, much of the R&D that follows discovery is devoted to clinical trials that show whether the molecule is effective or not in alleviating disease. When regulatory authorities have validated that the molecule is indeed considered safe and efficacious, others could, absent patent or other regulatory protection, synthesize the identical drug for a few millions of dollars and enter competitive (generic) production. Although the original R&D pioneer would still enjoy reputational “first mover” advantages for its efforts (see Bond and Lean 1979), these might be insufficient to deter entry and allow recovery on average of sunk research, development, and testing costs. If so, incentives for original discovery could fail.

Thus, patents are unusually important in pharmaceuticals. The remaining questions are, how widespread geographically must the scope of patents be to maintain the incentives required to sustain a vigorous pace of discovery and testing? What tradeoffs are embraced when only some nations offer pharmaceutical patents? And since historically the nations with low to modest per capita incomes have also been among those not granting pharmaceutical product patent rights, how might the extension of those rights affect the availability of new medicines?

These questions have wide applicability, but they were particularly pertinent as the TRIPS agreement was being negotiated and then adopted during the early 1990s. Two phenomena gave the question unusual saliency then. For one, multinational pharmaceutical companies took a leading position in the advocacy, first in the United States and then on a united US-Japanese-European front, to secure a TRIPS requirement that patent protection no longer be denied on pharmaceutical products (see Ryan 1998). But second, these demands for increased patent protection coincided with a time when the world fully recognized a deadly new health scourge: HIV/AIDS. After the nature of the disease was first understood in the early 1980s, numerous new drugs were developed to combat it. Nearly all were protected by patents and sold at prices far in excess of their production costs. The key question arose, how should a firm that has a patent-protected monopoly on some life-saving therapy price that therapy across diverse world jurisdictions, some rich like the United States and Europe, some desperately poor (as in Africa, from which HIV/AIDS spread)?

To be sure, until the requirements of TRIPS are implemented by all nations in 2016 (or later, if there are further extensions), world-wide patent protection of new pharmaceuticals will be absent. But with some important exceptions that will be addressed later, the reach of pharmaceutical patents is extensive, in part because many nations offering no local patent protection lack the internal capabilities to produce new drugs and therefore are at least partly dependent upon sources in patent-granting jurisdictions.

Pricing New Drugs Under Monopoly Conditions

We begin with a normative question: What pricing strategies would one expect for a multinational drug company selling its products in a cross-section of markets directly or indirectly protected from competition by patents?

Figures 21.1a and 21.1b provide an introductory perspective. They assume that a particular patented drug is sold under essentially monopolistic conditions in two nations: wealthy nation A (e.g., the firm’s home nation or similar nations) and low-income nation B.⁴ Assuming for convenience similar population counts between the two nations, the demand curves for the drug nevertheless differ significantly because of an income effect that shifts the less affluent nation’s demand curve downward and toward the origin. Only at very low prices is the quantity demanded in the low-income nation similar to that of the high-income nation. How should the drug provider set its prices? A well-established tradition in economics suggests that it should engage in what is often called Ramsey-Baumol-Bradford price discrimination (Baumol and Bradford 1970). Assuming identical marginal costs in both nations of $18 per prescription delivered and no leakage of product from the low-price to the high-price nation (i.e., no arbitrage or parallel trade), the firm maximizes its profits decentrally by setting a high price of say $58 per standard prescription unit in the rich nation and a low price of $26 per unit in the low-income nation. In this way it realizes a surplus of revenues over marginal costs of approximately $17.6 million per month from sales in the high-income nation and $1.6 million in the low-income nation. If on the other hand it were to levy the profit-maximizing rich-nation price of $58 per Rx in the low-income nation, it would under the assumed demand conditions sell nothing at all there, foregoing a contribution to profits (and to the recovery of its research and development investments) of $1.6 million per month. The logic is compelling.⁵ Successful price discrimination (also called differential pricing, tiered pricing, or equity pricing) is more profitable when feasible than a single-price strategy. It also has another compelling feature reflecting other well-recognized advantages of price discrimination. When a $26 price is set in the low-income nation, consumers with ability and willingness to pay more than that price realize a “consumers’ surplus,” shown in Figure 21.1b by the dot-shaded triangle above the $26 price line and below the demand curve. Under the near-linearity assumptions made, which probably underestimate the willingness to pay of especially affluent and needy consumers, setting prices well above the low-income nation’s profit-maximizing discriminatory price causes consumers to forego a consumers’ surplus of approximately $1.1 million per month.

Recognizing these advantages, a conference convened in Høsbjør, Norway during April 2001 by important but unusual partners – the World Health Organization (WHO) and the WTO – concluded: “Differential pricing can and should play an important role in ensuring access to existing essential drugs at affordable prices, especially in poor countries. In doing so, it could help to reconcile affordability with incentives for research and development” (Watal 2002: 218).

The Extent of Differential Pricing

Do pharmaceutical firms actually differentiate their prices in nations of diverse affluence, as suggested by Figures 21.1a and 21.1b? Several studies suggest that there is some such tendency, although the evidence is mixed and the reasons for price variation are much more complex than a pricing-to-reflect-income hypothesis can support.

It was the HIV/AIDS crisis of the 1990s, with its highest incidence in low-income African and Asian nations, that focused attention most intensely on the determinants of national price differences. A particularly impressive effort was mounted by Stéphane Lucchini et al. (2003). Supplementing data collection efforts by international healthcare agencies, the investigators went into the field to obtain point-of-entry transaction prices for seven anti-retroviral drugs in 13 sub-Saharan African nations plus Brazil. The most striking result of their study was a series of charts tracking the dramatic price decreases observed between 1997 and 2001 – on average from their regression equations (in their Table 3), by 82%, with an especially sharp decline in 2001. Among the variables they tested for explanatory power was GDP per capita in the target nations. Consistent with the differential pricing hypothesis, they found a positive relationship between prices and GDP per capita, but the relevant coefficient was not statistically significant. Its role, however, might have been blurred by the fact that their sample focused almost exclusively on low-income nations. Among other findings, prices tended to be systematically higher in sample nations with drug product patent protection and lower for drugs allowing significant generic competition.

Another study analyzing the prices of AIDS drugs in 18 intermediate and low-income nations, with benchmarking to prices in the much richer United States, was conducted by Scherer and Watal (2002a). Data were obtained from IMS, a pharmaceutical industry data compiler, on wholesale prices, usually charged to retailers, for 15 AIDS anti-retrovirals across the years 1995–1999. The sample differed from that of Lucchini et al. by emphasizing South and Central American nations, mostly with low to intermediate average GNP per capita, and only two sub-Saharan African nations. The main analyses focused on the ratio of observed prices in sample nations as compared to the wholesale prices of the same drugs in the United States. The central question was, were prices in the low- and medium-income nations systematically lower than those in the wealthy United States? The evidence of Ramsey-like discriminatory pricing was weak. On average, prices in the sample nations were 85% of reported US wholesale prices, which tended in turn to be 15–25% higher than actual US transaction prices. There was a weak positive correlation (+0.127) between the ratio of sample to US prices and GNP per capita in the surveyed nations, declining to essentially zero in the last covered year. In an unreported analysis of the original data, the US prices to which the sample nations’ prices were compared were found in most (but not all) cases to have fallen substantially between 1995 and 1999, mirroring the similar price decreases observed by Lucchini et al. (2003) in sub-Saharan African nations.

Keith Maskus and Matthias Ganslandt (2002) approached the price discrimination question from a different perspective.⁶ See also Danzon and Cho (2000), analyzing mainly high-income nations. The focus of Maskus and Ganslandt was on 20 typically well-known brand-name drugs, explicitly excluding AIDS anti-retrovirals, in 14 nations, of which only five fell into the low- or medium-income per capita category. They found only weak support for the income-correlated differential pricing hypothesis. Some lower-income nations did benefit from reduced average prices, but others paid even more than in the wealthier nations. Variables influencing the prices at which nations were supplied included the existence of national price controls, drug makers’ strategies in targeting diverse income groups within national markets, the extent of competitive pressure from imports of the same drugs from lower-price jurisdictions, the way procurement efforts were organized, and the competitive efficiency of nations’ wholesale distribution channels.

Reasons for the Paucity of Differential Pricing

There are several reasons why pharmaceutical prices have not been systematically adjusted to different nations’ income per capita, reflecting ability to pay, under Ramsey-type pricing strategies.

One that has received particularly intense attention is so-called “parallel trade,” that is, arbitrage transactions by middlemen who acquire pharmaceuticals (or other products) at sharply discounted prices in low-income nations and divert them from their intended recipients to the consumers of rich nations, reaping, needless to say, an arbitrageur’s profit by doing so. During the 1990s, European Community authorities actively encouraged parallel trade in pharmaceuticals, hoping that the redirection of supplies from low-price Community member nations (most notably, Spain) to nations in which prices were higher would help perfect the Common Market (see Brittan 1992). A decade later, however, it was recognized that a policy suitable for encouraging parallel trade among rich nations was inappropriate to meeting the needs of consumers in low-income nations. The consensus report of the April 2001 WHO-WTO conference concluded that: “Markets for differentially priced drugs need to be tightly segmented to prevent leakage of differentially priced drugs to high-income markets. … High-income countries may need additional legal authority in order to prevent the import of products marketed elsewhere at differential prices” (Watal 2002: 226–227).

In an apparent response to the conference sponsors’ urging, European national customs officers soon thereafter executed a well-publicized confiscation of drugs reimported from low-income nations back into European markets.⁷ More systematic policies were subsequently adopted to encourage “tiered pricing” and to discourage the parallel importation of low-priced drugs into Europe from less developed nations (see European Commission 2002). Among other things, it was recognized that low-price shipments to the least developed nations should have distinctive packaging.

A possibly more serious obstacle to the differential pricing of drugs comes from the formal price control systems some wealthy nations use to prevent the makers of patented drugs from exploiting their full pricing power. This is the so-called system of “external reference pricing,” under which national price regulators set the maximum allowable price for a drug to approximate the level of prices observed in several “reference” nations. Intricate international reference price networks have been mapped (see Watal 2002: 229, reproducing a chart by Ed Schoonveld). The available evidence suggests that the least developed nations have seldom been included among the reference nations. However, this could change, and if the low prices offered under income-correlated differential pricing were incorporated in reference-based price control systems, the rich-nation price-reducing effect could substantially impair multinational pharmaceutical companies’ incentives to offer drugs in low-income nations at prices approaching marginal cost.

This disincentive could also work without formal reference pricing control systems. If drug X is sold at $65 per 30 capsules in a rich nation and only $15 in a low-income nation, procurement authorities in the rich nation are likely to say to the supplier, “See here. You’re only charging $15 for this drug elsewhere. You’re discriminating against us. We demand that you reduce your prices here too.” The multinational pharmaceutical manufacturer may be forced to concede, in which case, its incentive to offer the drug at low prices in less developed nations is undermined.

Apart from the linkages among rich nations in systems of pharmaceutical pricing, the segmentation of consumers within a particular low-income nation market can also lead to conscious drug manufacturer strategies favoring relatively high prices. Income inequality is pervasive, within rich nations and poor. Many low-income nations (South Africa comes to mind) can with some oversimplification be divided into two groups: a rich minority often employed by multinational enterprises at high wages and, very importantly, enjoying generous health insurance; and the vast majority who toil at low wages and lack formal health insurance plans. Figure 21.2 illustrates the elemental economic theory. It shows a national market in which consumers can be segmented into two groups: sub-market 1, comprising the rich, well-insured consumers (corresponding to the aggregate of consumers in Figure 21.1a) and another sub-market 2 in which are clustered the less affluent and poorly-insured potential consumers.⁸

In this case, the combined demand curve for both classes of consumers is kinked, with demand in the less affluent sub-market D₂ being added to demand in sub-market 1 only below the “choke price” of $35 per unit in the lower-income sub-market. If the supplier is able to engage in price discrimination across the two sub-markets, its profit-maximizing strategy must include ensuring that the marginal revenues in the two sub-markets are equalized. Otherwise a reallocation of output from the lower-MR to the higher-MR sub-market is profitable. Under the assumed circumstances, the supplier will compute its combined marginal revenue function (with a kink where the two sub-market MR functions intersect). It then sets a price of $58 in the rich segment and $32 in the low-income segment. However, within a single national entity, arbitrage is much more likely than across geographically separated nations, and so the effort at differential pricing may fail. In that case, the pharmaceutical supplier faces two alternative single-price options. It can try to serve all customers at a price of $32, or it can confine its sales to the rich consumers comprising sub-market 1 at $58. The surplus of revenues over marginal costs is approximately $4.6 million per month if both consumer classes are served at a uniform price. If sales are directed only toward affluent consumers, profits are $6.6 million per month. If arbitrage is probable and successful market segmentation unlikely, the producer will be inclined to charge a high price and exclude low-income consumers, whose “choke price” of $35 precludes their participation in the high-price sub-market. Among other things, the high-price strategy sacrifices substantial consumer surpluses for both rich and poor consumers even as it maximizes producer profits.

A complication overlooked in the analysis by Scherer and Watal (2002a) was articulated by Maskus and Ganslandt (2002: 68–69 and 78–79). Especially in the least developed nations, but also in China,⁹ distribution channels linking pharmaceutical manufacturers to eventual consumers are often inefficient and infused with monopoly elements. In many sub-Saharan African nations, exclusive franchises are awarded, often as the result of political corruption, to drug wholesalers and sometimes to private companies that receive drugs at the initial importation node and then relay them to wholesalers. Competition at the drug retailing level is also limited. Given their monopoly positions, middlemen charge extraordinarily high prices for their services, raising the prices ultimately paid by retail consumers. The more monopolistic stages pyramided between the original product source and the consumer, the more the ultimate quantity supplied is likely to be restricted, all else being equal.¹⁰ One consequence is that price concessions offered by manufacturers can be dissipated and hence blurred in the analysis of prices paid by retailers and consumers. This blurring is more likely in analyses that use data sources such as IMS, which focuses its efforts on the prices paid to wholesalers by retailers, than in studies such as that of Lucchini et al. (2003), who ventured into the field to obtain original source pricing data.

The Ultimate Triumph of Differential Pricing

Despite these problems, needy consumers in low-income nations ultimately became the beneficiaries of greatly reduced prices for life-saving drugs and vaccines, especially during the 1990s and the first decade of the twenty-first century. For example, the cost of providing three-drug therapy to persons afflicted with HIV/AIDS in the poorest nations fell from approximately $15,000 for a year’s treatment in 2001 to an average of $127 in 2012.¹¹ These price reductions made possible an increase in the number of low- and middle-income nation citizens treated with anti-retrovirals from the low hundreds of thousands in 2001 to an estimated 10 million in 2012, along with an appreciable decrease in the number of persons dying from AIDS following peak mortality levels in 2005.¹² Similar price reduction trends can be observed for vaccines (beginning in an earlier decade) and drugs targeted inter alia against malaria and tuberculosis.

Several developments led to these changes. For one, increases in the cumulative and ongoing quantity of drugs procured probably led to economies of scale, although detailed recent information on cost-volume relationships are lacking. On the substantial cost reductions achieved in early penicillin production, see Federal Trade Commission (1958: 162–163).

Second, life-saving drugs became available to an increasing extent from government clinics and from international healthcare organizations rather than through private-sector pharmacies, avoiding many of the distribution channel inefficiencies characterized earlier.

Accompanying this change was an increase in the bargaining power of entities purchasing essential drugs from manufacturers. Organizations such as the WHO, UNAIDS (an affiliate of the United Nations), the multinational Global Fund to Fight AIDS, Tuberculosis, and Malaria, UNICEF (at work already in the 1990s to obtain and distribute vaccines at rock-bottom prices), Médecines Sans Frontières, Oxfam, the Gates Foundation, the Clinton Foundation, and PEPFAR (the US President’s Emergency Plan for AIDS Relief) brought to bear both moral and large-purchaser pressure on pharmaceutical manufacturers and then ensured that the purchased drugs reached utilization points expeditiously (see Kapstein and Busby 2013).¹³

By no means all of the dramatic reductions in AIDS drug prices came from price discrimination favoring low-income nations. Moral suasion and large-buyer pressure also had an impact in the rich nations that were home to research-based pharmaceutical companies. In the United States, for example, AZT (original brand name Retrovir), the first drug shown to be effective against AIDS, was sold in 1987 at prices amounting to approximately $10,000 for a year’s treatment of a single patient. A patient advocacy group, the ACT UP (AIDS) coalition, disseminated information on the large disparity between estimated production costs and prices and mounted highly visible protests. Its activity stimulated investigatory hearings in the US Congress, eventually leading to substantial domestic market AZT price reductions. Similar pressures were exerted with respect to the numerous AIDS drugs reaching the market during the ensuing decade and a half.

Differentially, however, the producers of drugs with patent protection were induced to reduce prices dramatically in less affluent markets not only through appeal to the logic of Ramsey-type price discrimination, but also by the stern discipline of competition. Some nations, we have seen earlier, did not issue patents for pharmaceutical products at the time the TRIPS agreement was signed in 1995. Among these were India and Brazil. Although it was not required to do so at an early date, Brazil extended patent protection for pharmaceutical products shortly after the ratification of TRIPS. India chose not to do so until 2005. Even for Brazil, however, only the newest pharmaceutical products were covered by patents. Older molecules remained patent-free while protection continued in the nations that had earlier coverage. This made it possible for Brazilian companies not bound by patent rights to begin competitive generic supply and (more importantly) for Brazilian health authorities to draw upon Indian suppliers at greatly reduced prices. The multinationals were required either to follow suit or confine themselves to severely reduced market shares when they tried to maintain high prices in no-patent jurisdictions.

The history of India’s role in these developments is particularly interesting. The chronicle begins much earlier in Europe, however.

Until 1978, Italy’s laws expressly excluded pharmaceutical products from patent protection. In 1978, however, the Italian Supreme Court ruled in a law suit brought by both multinational and some Italian pharmaceutical firms that this exclusion was unconstitutional. The Italian parliament passed a law implementing the court’s decision in 1982, and from that time on, new pharmaceutical products tended to receive Italian patent protection. Before that change, Italy was home to the world’s leading export-oriented generic pharmaceutical industry. With no domestic patent barriers, Italian firms were able to supply new drugs competitively in their home market. But in so doing they also enjoyed a first-mover advantage in supplying those products to other nations (including Greece, Spain, and Argentina) that did not grant product patents. As the almost certain consequence of these 1978–1982 legal changes, Italy’s balance of trade in pharmaceuticals turned sharply negative beginning in 1982 as the export advantage of its domestic producers faded, while pharmaceutical R&D in Italy rose at an annual rate below multinational company trends (see Weisburst and Scherer 1995; Challu 1995).

India’s industry rose to fill the gap. It became the world’s leading generic drug supplier. The AIDS crisis amplified that role. As the twenty-first century dawned, anti-retroviral prices in most of the world, including severely impacted African nations, were so high as to be unaffordable for most AIDS sufferers. With the encouragement of William Haddad, president of the US Generic Pharmaceutical Manufacturers Association, Médecins Sans Frontières, and an organization created by US consumer activist Ralph Nader, Indian firms, led by generic specialists Cipla and Ranbaxy, began exporting substantial quantities of anti-retrovirals and other important drugs to Africa and some southeast Asian nations at prices as low as two percent of those quoted by multinational patent holders.¹⁴ Again, the consequence was a dramatic decrease in drug prices for low-income nations and indeed the emergence of even greater cross-nation price differences than those implied by the theory of differential pricing.

Compulsory Patent Licensing

The TRIPS treaty requirement that recalcitrant nations begin issuing patents on pharmaceutical products and other previously excluded subject matter, immediately for so-called developed nations and with delays for less developed nations, was not the sole determinant as to whether generic competition could emerge. The original TRIPS agreement included language permitting national governments to “march in” and negate otherwise binding patent rights through so-called compulsory licensing decrees under certain specified circumstances. Under Article 31, compulsory licensing was permitted when a would-be patent user is unsuccessful within a reasonable period of time in obtaining from a patent holder the right (i.e., a license) to supply competitively a patented invention “on reasonable commercial terms and conditions” – for example, with the payment of what are called (but hard to quantify) “reasonable royalties.” The failed negotiations condition could be waived in cases of national emergency or extreme urgency or for non-commercial public use.¹⁵ Subparagraph 31(k) also allowed compulsory licensing to correct “anti-competitive practices,” which were spelled out more fully in Article 40. The combination of subparagraph 31(k) and Article 40 appears to track in a general way the history of the United States, where, especially between 1939 and 1956 and more sporadically thereafter, tens of thousands of US patents were subjected to compulsory licensing in full or partial settlement of antitrust complaints (see Hollabaugh and Wright 1960).

The exact legal procedures through which compulsory licenses can be issued under TRIPS were left unspecified, and publicity on their actual issuance is sparse. It is known that licenses were issued on at least seven pharmaceutical patents by Thailand, one by Brazil (with other threats settled by voluntary licensing), and one (with others under consideration) by India. The more important function of Article 31 may be to induce price reductions or the issue of voluntary patent licenses without the formal declaration of a compulsory license. In 2001, for example, both the United States and Canada threatened compulsory licensing of Bayer AG’s patent on Ciproflaxin when, it was feared, terrorist activity might trigger an epidemic of otherwise untreatable anthrax. In that case, however, substantial price reductions were forthcoming and formal licensing was averted.

South Africa, with some of the most severe AIDS incidence, implicitly used the provisions of Articles 31 and 40 (plus its own competition laws) to secure licenses without carrying the action all the way to a formal declaration. By 2003, it had become known that a so-called “triple therapy,” including three different anti-retroviral drugs, was the most effective way to treat AIDS. Three-drug therapy was more effective in abetting the frequent mutations that could render individual therapeutic molecules impotent, and combining three molecules in one twice-daily pill was a superior way to ensure daily compliance, again reducing the danger of mutation. But patents covering one of the drugs (AZT) were held by GlaxoSmithKline (successor to Burroughs-Wellcome) and those for two other key ingredients were held by the German firm Boehringer-Ingelheim. The two firms declined to cross-license each other, so no single-pill therapy was available. An action by the South African Competition Commission aided by CPTECH, an offspring of Ralph Nader’s US consumer advocacy organization, induced the firms to offer what were in effect compulsory licenses, first to a South African generic supplier, Aspen Pharmacare, and then to foreign (e.g., Indian) suppliers (see Competition Commission of South Africa 2004).¹⁶ New triple therapies became available at unprecedentedly low prices.

The original text of TRIPS Article 31 specified that the compulsory licenses it authorized had to be “predominantly for the supply of the domestic market.”¹⁷ This posed a difficulty. Many less developed nations lack the technological capability to produce advanced organic pharmaceutical molecules. A license issued only domestically for them would be of little use. For a supplier in India, on the other hand, the “domestic market” language could be a bar to licensing and then exporting in majority quantities to other needy markets. This issue was addressed at the 2001 WTO conference in Doha, leading to a declaration that: “We recognize that WTO members with insufficient or no manufacturing capacities in the pharmaceutical sector could face difficulties in making effective use of compulsory licensing under the TRIPS Agreement.”

A definitive solution was postponed, but in 2003 the WTO formally declared that border-hopping licenses could be issued “in good faith to protect public health.”¹⁸ The WTO’s Doha declaration also stated that “Each Member has the right to grant compulsory licenses and the freedom to determine the grounds upon which such licenses are granted” and that “Each member has the right to determine what constitutes a national emergency or other circumstances of extreme urgency, it being understood that public health crises, including those relating to HIV/AIDS, tuberculosis, malaria and other epidemics, can represent a national emergency or other circumstances of extreme urgency.”

Many of the most advanced pharmaceuticals originate not in company laboratories but through basic research pursued by university scientists. Since universities lack the capabilities needed to manage large-scale clinical tests, manufacture new drugs, and distribute them to users, universities in the United States since at least passage of the Bayh–Dole Act (94 Stat. 3019, 1980) have granted typically exclusive licenses for relevant new drug entity patents to commercial firms. The practice attracted public scrutiny in 2001. Yale University licensed a patent on the anti-retroviral molecule stavudine (d4T) to Bristol-Myers-Squibb, whose high prices on third-world sales of its branded product Zerit stimulated student protests. Yale officials thereupon brought pressure to bear on Bristol-Myers to reduce its third-world prices drastically and to allow other companies to enter into competitive manufacturing overseas. This experience in turn led to a multi-university concord encouraging patent licensing strategies that “make affordable versions [of drugs, vaccines, and medical diagnostic tools] available in resource-limited countries” (see Universities Allied for Essential Medicines 2007).

Consequences for Research and Development

Dramatic changes have occurred in the supply of advanced pharmaceutical products since the TRIPS agreement was adopted in 1995. Nations that once chose not to do so are gradually being required to issue pharmaceutical product patents covering bona fide technological achievements. Numerous national and international organizations have been reoriented or formed de novo to bargain for low prices in procuring patented pharmaceuticals directed to low-income nations when suppliers were disinclined for diverse reasons not to engage in differential pricing voluntarily. The TRIPS agreement initially carved out exceptions to its strong pro-patent requirements with compulsory license clauses and then adopted further extensions. All this has led to huge decreases in the prices paid for life-saving drugs (if payment has been required at all) to the world’s least advantaged citizens.

The key remaining question is, how has downward pressure on third-world pharmaceutical product prices affected company incentives to invest in research and the development of new drugs? As a first approximation, measures that limit the profits obtainable through pharmaceutical innovations also reduce incentives to invest in bringing new drugs onto the market (although the advent of generic competition against already marketed drugs also spurs R&D on more advanced replacements). Difficult tradeoffs must be faced. For drugs targeted toward diseases at least as prevalent if not more prevalent in rich nations, providing weaker patent protection and facilitating tougher competition in low-income nations undoubtedly has a small, even if not vanishing, negative impact. A deeper analysis (Scherer 2004) of the tradeoff suggests the following generalizations:

• The smaller is the ratio of attainable potential quasi-rents (roughly, profits before deduction of fixed costs) in low-income nations relative to those in wealthy nations, the stronger is the case for weak patents in low-income nations.
• The more the marginal utility of income diminishes with increased affluence, the stronger is the case on total economic welfare grounds for weak patents in low-income nations.
• The case for weak patents in low-income nations is stronger when R&D is a virtuous rent-seeking process, with R&D expenditures rising competitively to dissipate supra-normal profit opportunities, than when a prospect theory of R&D investment applies (i.e., when some companies hold preferred and relatively imitation-proof ex ante positions in developing a new product).¹⁹
• The case for weak patents in low-income nations is strengthened when indigenous firms build their technological capabilities through imitative (i.e., generic) production that can then serve as a platform for advancing to higher levels of productivity and innovativeness (mirroring the Japanese, South Korean, and most recently, Chinese economic development histories).

A special case of the first condition exists for so-called “tropical diseases” – that is, diseases endemic preponderantly in low-income nations and rare in highly developed nations. Malaria and dengue fever are examples. Then nearly all of the quasi-rents attainable through successful R&D come from low-income nations. The expectation of such rents may be so meager that R&D investments will not be made even under the most favorable market conditions. See Médecins Sans Frontières (2001), which found that among 1393 new drug chemical entities introduced into world markets between 1975 and 1999, only 13–15 were for so-called “tropical diseases.” But in borderline cases, limiting the strength of patent rights is likely to weaken incentives even more. If so, some form of public-spirited philanthropy by private organizations or governments must fill the gap. The gap-filling problem can be addressed through subsidies on either the supply side or the demand side. On the supply side, governments (or even pharmaceutical companies, in their philanthropic modes) can support research, development, and testing of therapies for diseases neglected under normal market incentives. Or on the demand side, advance commitments can be made to purchase at subsidized prices new drugs and vaccines effective against neglected tropical diseases. Levine et al. (2005) propose in detail an “advance market commitment” initiative targeting HIV/AIDS, malaria, and tuberculosis vaccines that was considered favorably by G-8 government leaders at plenary meetings between 2005 and 2008 but never fully implemented. If such philanthropy could be successful, the conflict between technological progress and maximum consumer access to its fruits would be reduced.