Jump to ContentJump to Main Navigation
Understanding Poverty$

Abhijit Vinayak Banerjee, Roland Bénabou, and Dilip Mookherjee

Print publication date: 2006

Print ISBN-13: 9780195305197

Published to Oxford Scholarship Online: September 2006

DOI: 10.1093/0195305191.001.0001

Show Summary Details
Page of

PRINTED FROM OXFORD SCHOLARSHIP ONLINE (www.oxfordscholarship.com). (c) Copyright Oxford University Press, 2019. All Rights Reserved. An individual user may print out a PDF of a single chapter of a monograph in OSO for personal use. date: 21 October 2019

Public Policies to Stimulate Development of Vaccines for Neglected Diseases

Public Policies to Stimulate Development of Vaccines for Neglected Diseases

(p.319) 21 Public Policies to Stimulate Development of Vaccines for Neglected Diseases
Understanding Poverty

Michael Kremer

Oxford University Press

Abstract and Keywords

This essay focuses on the impact of intellectual property rights (IPRs) on low-and middle-income countries’ health care. There are two different reasons why poor countries may not have access to needed vaccines and drugs. In the case of global diseases, such as diabetes or cancer, patents may hinder the diffusion of pharmaceuticals. In the case of neglected or tropical diseases, such as malaria, tuberculosis, and leishmaniasis, the corresponding vaccines or drugs are not developed because of low profitability due to the poverty of potential customers. The important role of compulsory licensing for low- and middle-income countries is discussed.

Keywords:   compulsory licensing, IPR, global diseases, tropical diseases, prize mechanism, global social contract

In 2003, 5 million people were newly infected with HIV, 3 million of whom live in sub-Saharan Africa (UNAIDS 2004). Worldwide, almost 3 million people died of HIV/AIDS in 2003 (UNAIDS 2004). In 2002, 300 million clinical cases of malaria resulted in over 1 million deaths, and almost 2 million people died of tuberculosis (WHO 2002a, 2002b). Almost all of these deaths occurred in developing countries (WHO 2000b, 2000c). Yet relative to these enormous numbers, very little research is directed toward these diseases. Potential developers of vaccines appropriate for poor countries fear that they will not be able to sell enough of their product at a sufficient price to recoup their research and development (R&D) investments. This is both because these diseases primarily affect poor countries, and because vaccine markets are severely distorted. This essay examines the reasons for underinvestment in R&D and the potential of various public policies, including committing in advance to purchase needed products, if and when they are developed, to address the problem.

The first section of this essay provides background information on malaria, HIV, and tuberculosis, and discusses the dearth of R&D investments in vaccines for these diseases. The second section discusses distortions in the markets for vaccines and vaccine research. The third section examines the potential roles of what are called “push” and “pull” programs in encouraging research and improving access to vaccines once they are developed. The fourth section compares alternative “pull” programs. The fifth section discusses how a purchase commitment (one type of “pull” program) could be made credible and outlines a possible process for determining vaccine eligibility. The sixth section discusses the appropriate size of the commitment (p.320) as well as the cost effectiveness of the program. The seventh section argues that purchase commitments are most needed and would be easiest to implement for vaccines, but that the approach could be adapted for other products needed by developing countries. The eighth section discusses how national governments, international organizations, and private foundations could participate in a purchase commitment program.


The burden of infectious disease is huge and is concentrated in poor countries. More than 42 million people are infected with HIV worldwide, and over 95% of them live in developing countries (UNAIDS 2004). Almost all cases of malaria are in developing countries, and almost 90% are in Africa (WHO 2000c). More than 98% of deaths from tuberculosis occur in developing countries (WHO 2000b).

The spread of resistance poses a threat to developed as well as developing countries (WHO 1997): resistance to the major drugs used for treating malaria and for providing short-term protection to travelers is spreading (Cowman 1995), and up to 17% of tuberculosis infections are resistant to all five major antitubercular drugs (WHO 1997). The existing BCG (Bacillus of Calmette-Guérin) vaccine, which is widely distributed, provides short-run, imperfect protection against tuberculosis, but a more effective vaccine providing longer-term protection is lacking.1 There are currently no vaccines for malaria and AIDS.

However, vaccines have proved effective against many other infectious diseases, and provide the best hope for long-run, sustainable solutions to malaria, tuberculosis, and HIV/AIDS. The potential of vaccines is illustrated most vividly by the success of the smallpox vaccination program, which led to the eradication of the disease in 1980. Currently, about 70% of children in low-income countries receive a standard package of cheap, off-patent vaccines through the World Health Organization (WHO)'s Expanded Programme on Immunization (EPI), and these vaccines are estimated to save 3 million lives worldwide each year (Kim-Farley 1992).2

Little research is oriented toward tropical diseases. Although the difficulty of developing vaccines against HIV/AIDS, tuberculosis, and malaria may have contributed to the reluctance among firms to invest in the necessary research, it is probably not the main reason, since many scientists are optimistic about the long-run scientific prospects for vaccine development.3 A much more plausible explanation for the dearth of R&D is that the potential market for these vaccines is very small.4 Pecoul et al. (1999) report that of the 1,233 drugs licensed worldwide between 1975 and 1997, only 13 were for tropical diseases. Of these, five came from veterinary research, two were modifications of existing medicines, and two were produced for the U.S. military. Only four were developed by commercial pharmaceutical firms specifically for tropical diseases of humans.5 The private sector in particular performs re (p.321) markably little research on the diseases of poor, tropical countries. According to the WHO (1996), 50% of global health research and development in 1992 was undertaken by private industry, but less than 5% of that was devoted to diseases specific to less-developed countries.


One reason for the paucity of research on vaccines for malaria, tuberculosis, and strains of HIV common in Africa is simply that the countries affected by these diseases are poor, and cannot afford to pay much for vaccines. If this were the only reason, however, there would be no particular reason to target aid expenditures to vaccines or vaccine research, rather than to other goods needed in poor countries, such as food, or public goods such as roads. However, distortions in the research market destroy incentives for private firms to conduct research that would be cost-effective for society as a whole, even by the stringent cost-effectiveness standards used to evaluate health interventions in poor countries. Moreover, distortions in the markets for vaccines lead them to be underconsumed even relative to the incomes of the poor.

The private returns for developing products to fight diseases of developing countries are likely to be a tiny fraction of the social returns to these products. Economists have estimated that the social returns to research and development are typically twice the realized returns to private developers (Nadiri 1993; Mansfield et al. 1977). To take an example, consider a hypothetical malaria vaccine. A standard way to assess the cost-effectiveness of a health intervention is the cost per disability-adjusted life year (DALY) saved, which takes into account not only the lives lost through disease but also the number of years of disability caused (Murray and Lopez 1996a, 1996b). A common cost-effectiveness threshold for health interventions in the poorest countries is $100 per DALY.6 At this threshold, a malaria vaccine would be cost-effective even at a price of $45 per immunized person (Glennerster and Kremer 2001) but, based on the historical record of vaccine prices, the developer of a malaria vaccine would be lucky to receive payments of one tenth or one twentieth of that amount. With the expectation of such prices, private developers lack incentives to pursue research on socially valuable projects.

At least two other factors also contribute to the reluctance of pharmaceutical firms to invest in R&D on diseases that primarily affect developing countries: first, intellectual property protection is often lacking, implying that the potential revenue from product sales is far smaller than the sum of customers' willingness to pay; and second, there is a tendency for governments to force down prices after firms have sunk their R&D costs.

Facing a trade-off between providing access to critical medicines and rigorously enforcing patent protection, many developing countries have historically provided little protection for intellectual property rights. There may be some advantages to this decision, but it reduces incentives for R&D on prod (p.322) ucts needed in these countries. Pharmaceutical R&D is what economists call a “global public good,” meaning that it benefits individuals who do not themselves finance the R&D expenditures or commit to protect intellectual property rights. The global public good problem implies each country has an incentive to “free-ride” on research financed by the governments of other countries or induced by other countries' greater commitment to protect intellectual property rights. Small countries, such as Uganda, can assume that individually their actions will have little effect on total research incentives. However, if all African countries act this way, there will be little incentive for firms to invest in developing a malaria vaccine. It is not clear whether the World Trade Organization (WTO) will lead to effective intellectual property rights enforcement in developing countries, especially since the Doha Agreements explicitly left open several provisions that provide potential escape hatches. For instance, countries can impose compulsory licensing in national emergencies, the definition of which is deliberately not stated (WTO 2001). Recent events suggest that national emergencies may be interpreted liberally, and whatever the overall benefits of these decisions, there is little doubt that pharmaceutical developers see the weakening of patent protection as a precedent that could be used to obtain vaccines and drugs at low prices in the future.

The high initial R&D costs, combined with the low costs of producing additional vaccines once the R&D has been completed, create what economists call a “time consistency” problem for governments. Once pharmaceutical firms have made their R&D investments and vaccines have been developed, governments have an incentive to set prices at or near the cost of producing additional vaccines (what economists call the “marginal cost”). After vaccine developers have invested the R&D necessary to develop vaccines, governments are in a strong bargaining position because they are often the primary purchasers of vaccines, as well as the regulators of pharmaceuticals and enforcers of intellectual property rights. If firms anticipate low prices, they will be reluctant to invest. In repeated interactions between nations and pharmaceutical producers, this time consistency problem could potentially be overcome through building credible reputations of governments.7

These global public-good and time consistency problems are exacerbated by political conditions in many developing countries that make vaccines a low priority. In particular, since vaccines deliver a widely distributed benefit, they tend to receive less political support than expenditures that benefit more concentrated and politically organized groups, such as salaries for health workers.

Overcoming the global public-good and time consistency problems requires creating incentives that both encourage new pharmaceutical development and provide the poor with access to these new pharmaceuticals once they have been developed. Because of the free-riding problems facing individual countries, solutions will have to come from entities with broader man (p.323) dates—such as international organizations, bilateral aid programs, or private foundations.


Programs to encourage R&D can take two broad forms. “Push” programs subsidize research inputs—for example, through grants to researchers or R&D tax credits. “Pull” programs reward research outputs—for example, by committing in advance to purchase a specified amount of a desired product at a specified price.

While push programs are important for basic scientific research, they also often encounter a number of problems. Because funders cannot perfectly monitor the actions of grant recipients, the latter may have incentives to devote effort to pursuing general scientific research or preparing their next grant application rather than focusing on development of the desired product. In contrast, under a pull program researchers will not receive payment unless a usable product is delivered, so they have incentives to focus on developing the desired product.

Because push programs pay for research inputs rather than results, decisions must be made about where to commit funds before any product is actually developed—and the authority for these decisions often lies in the hands of administrators who frequently rely on advice from those with vested interests in the decisions. Research administrators or their ultimate employers—the public and their elected officials—may not be able to determine which research projects in response to certain diseases are worth pursuing, nor which diseases should be targeted. Decision makers may therefore wind up financing ideas with only a minute probability of success or, worse, failing to fund promising research because they do not have confidence that its backers are presenting objective information on its prospects. In contrast, under a pull program, in which developers are rewarded only if they successfully produce the desired product, there is a strong incentive for firms considering research investments to assess prospects for success realistically. For instance, if a tuberculosis vaccine is feasible, but a malaria vaccine is not, developers will pursue the tuberculosis vaccine under a pull program.

When governments allocate research spending up front, they may also base decisions partly on political, rather than scientific, considerations. For example, there may be political pressure to allocate research expenditures to particular regions or countries, developing countries in particular. With pull programs, in contrast, the sponsors promise to pay for a viable vaccine wherever it is developed. In addition, even if push programs select research projects that appear to be appropriate at the outset but ultimately prove not to be, the original judgments are unlikely to be revised. If results on a particular research project that initially appear promising later turn out to be disappointing, a private firm is likely to shut the project down. A publicly (p.324) funded entity, on the other hand, may acquire its own bureaucratic momentum, which can lead governments to throw good money after bad.

A centralized push program may prevent a situation where private firms competing for a patent inefficiently duplicate each other's activities. However, in the case of research on malaria, tuberculosis, and even HIV/AIDS vaccines, the world is far from a situation in which developers are overinvesting in R&D or inefficiently duplicating each other's work. Moreover, while decentralization may lead to some duplication of effort, it also ensures that mistakes by a single decision maker will not block progress toward a vaccine.

The problems that plague push programs are illustrated by the U.S. Agency for International Development's (USAID) 1980s program to develop a malaria vaccine. During the USAID program, external evaluators suggested that additional funding should not be provided to two of the three research teams. However, USAID provided substantial new resources to all three teams and was sufficiently confident that vaccines would be developed that it even arranged to purchase monkeys for testing a vaccine. Two of the three researchers diverted grant funds into their private accounts and were later indicted for theft and criminal conspiracy. The project director received kickbacks from the contract to purchase monkeys and eventually pleaded guilty to accepting an illegal gratuity, filing false tax returns, and making false statements. Before the indictments, the agency claimed that there had been a “major breakthrough in the development of a vaccine against the most deadly form of malaria in human beings. The vaccine should be ready for use around the world, especially in developing countries, within five years” (Desowitz 1991, p. 255). That was in 1984. Today, the world is still waiting for a malaria vaccine. By the end of the project, USAID had spent $60 million but had obtained few results. While the example is extreme, it vividly illustrates the problems with push programs.

As an alternative to push programs that directly finance research, some have proposed R&D tax credits targeted to private research on vaccines needed by developing countries. However, such tax credits are subject to similar problems. Firms would have an incentive to relabel as much of their R&D as possible to have it be eligible for the targeted credit. For example, if there were an R&D tax credit for a malaria vaccine, researchers might focus on a vaccine that would likely provide only temporary protection suitable for travelers and military personnel (who spend only short times in developing countries) but not for residents of these areas. To take another example, firms would have every incentive to state that work on an adjuvant intended for an ineligible vaccine was actually for a malaria vaccine, so as to claim a tax credit. Finally, R&D tax credits will not improve access to vaccines once they are developed.

In contrast, under pull programs, the public pays nothing unless a viable vaccine is developed. Pull programs give researchers incentives to self-select projects with a reasonable chance of yielding a viable product and to focus on developing a marketable vaccine. Several historical precedents, such as (p.325) the Orphan Drug Act, suggest that pull-like mechanisms can be effective tools for spurring product development.8 Moreover, appropriately designed pull programs can help ensure that if new vaccines are developed, they will reach those who need them. For example, if developed countries or private foundations committed to purchase malaria vaccine at $15 per immunized person, if and when it was developed, they could then make it available to developing countries either for free or in return for a modest copayment.

A key limitation of pull programs is that they require specifying the output in advance. A pull program could not have been used to encourage the development of the Post-It Note or the graphical user interface, because these products could not have been adequately described before they were invented. Similarly, pull programs may not work well to encourage basic research, because it is typically difficult to specify the desired results of such research in advance. Simply rewarding the development of applied products is not a good way to stimulate basic research, since a program that tied rewards to the development of a specific product would encourage researchers to keep their results private as long as possible, in order to have an advantage in the next stage of research. A key objective of basic research is to provide information to other researchers rather than to develop products, and grant-funded academics and scientists in government laboratories have career incentives to publish their results quickly. In contrast to unanticipated inventions or to basic research, it is comparatively easier to define what is meant by a safe and efficacious vaccine, especially since existing institutions, such as the U.S. Food and Drug Administration (FDA) or its European counterpart, the European Agency for the Evaluation of Medicinal Products (EMEA), are already charged with making these determinations.

Both push and pull approaches have important roles in encouraging R&D on products needed by developing countries. While both push and pull incentives are already in place for pharmaceutical products needed in high-income countries, the world lacks a pull system for diseases that primarily affect low-income countries.


Pull programs that reward successful vaccine research can take several different forms, including commitments to purchase vaccines, patent buyouts, and extensions of patent rights on other products. Given the huge disparities between private and social returns to research, it is likely that any program that is committed to providing compensation to developers of vaccines needed by poor countries would be an improvement on the status quo. However, purchase commitments are likely the most attractive option for a pull program to encourage such R&D.

Patent buyouts are economically similar to purchase commitments, but purchasing products provides a clearer link between payments and product quality. For example, suppose that a vaccine received regulatory approval, (p.326) but was later found to have side effects. This was the case with the Wyeth-Ayerst rotavirus vaccine, which was withdrawn from the U.S. market following evidence that it causes intussusception in rare cases. If a patent buyout had been made at the date of regulatory approval, it might be difficult to recover the money. Vaccine purchases, on the other hand, could be suspended as soon as there was evidence of unacceptable side effects. Though in theory patent buyouts lead to free competition in manufacturing, because biologicals are difficult to produce, a patent buyout might leave the developer with an effective monopoly due to trade secrets, even without the patent. In this case, the public would effectively pay twice: once for the patent and again for the vaccine.

Another proposed design for a pull program is to reward developers with extensions of patents on other pharmaceuticals. This approach would inefficiently and inequitably place the entire burden of financing development on patients who need these other pharmaceuticals. For example, giving a patent extension on Prozac for developing an HIV vaccine could prevent some people from getting needed treatment for depression.


The effectiveness of a purchase commitment at inducing new research depends crucially on its credibility to investors and its design. Potential developers must believe that once they have sunk money into producing a vaccine, it will be purchased at a price that covers their risk-adjusted costs of research as well as their manufacturing costs. Courts have held that similar public commitments to reward contest winners or to purchase specified goods constitute legally binding contracts, and that the decisions of independent parties appointed in advance to adjudicate such programs are binding. For example, in the 1960s the U.S. government pledged to purchase, at a minimum price, domestically produced manganese. After the world price of the commodity fell, the General Services Administration (GSA), the U.S. agency in charge of administering the program, attempted to renege, but U.S. courts forced the GSA to honor the commitment (Morantz and Sloane 2001).

The more binding the commitment is, the stronger the incentives for potential developers. In general, there is a trade-off between flexibility and credibly committing to pay for a desired product. While general eligibility and pricing rules could be set out, some degree of discretion in interpreting these rules would be needed once candidate products have been developed and tested. Delegating decisions regarding eligibility and pricing to a committee that included some members who had worked in the industry and insulating the adjudicators from political pressure through long terms of service could increase potential developers' confidence that the committee would not impose unreasonable conditions after they developed a vaccine.

(p.327) The eligibility conditions set for candidate products will also be a key determinant of the success of a purchase commitment in encouraging research. A purchase commitment would, for instance, need to minimize the possibility that misspecified eligibility and pricing rules divert research incentives away from appropriate products. For example, it would be important to make clear that the commitment would not cover a hypothetical malaria vaccine that interfered with the development of natural immunity and provided only temporary protection. At the same time, it is important not to set specifications so stringently that they would discourage pharmaceutical firms from following promising leads. For example, it would be a mistake to require a vaccine that achieved 90% efficacy against all strains of the disease, since in this case potential vaccine developers might not pursue a candidate vaccine that would be likely to yield 99% protection against most strains, but only 85% protection against others, even if this candidate vaccine were the best available research opportunity. If developers had sufficient confidence in the program adjudicators, some flexibility in the technical requirements would not substantially decrease research incentives.

The eligibility conditions would likely include some minimal technical requirements that could outline specific characteristics an eligible vaccine must have. Determination of vaccine eligibility might also require either clearance by a regulatory agency, such as the U.S. FDA, or a waiver of regulatory approval in developed countries for products that would pass a cost-benefit analysis for use in developing but not developed countries. Some of the key technical issues that would need to be considered in determining vaccine eligibility and pricing include

  1. Vaccine efficacy, or the reduction in disease incidence among those receiving the vaccine.

  2. Efficacy might vary in different circumstances. For example, a vaccine could potentially be better suited to some geographic areas than others.

  3. The number of doses required, the efficacy of the vaccine if an incomplete course is given, and the ages at which doses must be taken. If too many doses are required, fewer people will bring their children in to receive the full course of immunization. If the vaccine can be given along with vaccines that are already widely administered, delivery will be much cheaper.

  4. Vaccine side effects. Side effects also could differ for different subpopulations.

  5. The time over which the vaccine provides protection, and whether booster shots could extend this period.

  6. What level of rigor would be required in the field trials. For example, how many separate studies in different regions would need to be conducted to assess efficacy against different varieties of the disease?

  7. (p.328)
  8. The extent to which vaccines would lose their effectiveness over time. Presumably, some ongoing monitoring of vaccine effectiveness in the field would be required, and if it appears that resistance to the vaccine is spreading, vaccine purchases would have to be reassessed.

Products that meet the technical requirements could then be subject to a market test: nations wishing to purchase vaccines might be required to provide a modest copayment tied to their per capita income, so that countries would have an incentive to carefully investigate whether candidate vaccines are appropriate for their local conditions. This provision would also help to assure that limited donor funds are allocated well and would increase incentives for developers by increasing the payment offered to the successful developer. On the other hand, a copayment requirement could reduce the confidence of potential vaccine developers in the program by increasing the uncertainty of future vaccine demand in poor countries. A purchase commitment could also include a system of bonus payments for products that exceed the minimum technical requirements.


After suitable eligibility requirements have been set and credibility has been established, the key determinant of research incentives for potential vaccine developers will be the total discounted revenue generated by a product. It is very expensive to conduct R&D, but once research is complete, it is typically fairly cheap to produce additional doses. For a fixed amount of total revenue, product developers will therefore be almost as happy to produce a high volume at a low price as a low volume at a high price.

This implies that, at least as a first approximation, prices should be set per person immunized or treated, not per dose. There is little reason to pay more per person immunized if more doses are required to provide immunity than if a single dose is required. In fact, the vaccine is more valuable if only a single dose is required to provide immunity, since this reduces delivery costs and is likely to increase patient compliance. Moreover, the purchase program would not save money by excluding large countries from coverage, or excluding countries if vaccination is cost-effective at the marginal cost of production but not at the average price paid under the program. Such exclusions would be a false economy. Because potential developers will need a fixed amount of revenue to induce them to conduct research, if fewer doses are purchased, the price per person immunized will need to be greater in order to induce the same amount of research.9

The total market promised by a purchase commitment should be large enough to induce substantial effort by vaccine developers, but less than the social value of the vaccine. Because potential developers know that their research may fail, in order to have incentives to conduct work on needed vaccines, they must expect to more than cover their research expenses if they (p.329) succeed. For example, if potential biotechnology investors expect that a candidate product has one chance in ten of succeeding, they will require at least a tenfold return on their investment in the case of success to make the investment worthwhile.

The larger the market offered under a purchase commitment, the more firms will enter the field, the more research leads each firm will pursue, and the faster a product will be developed. Given the enormous burden of diseases such as malaria, tuberculosis, and HIV/AIDS, it is important to provide sufficient incentive for many researchers to enter the field and to induce major pharmaceutical firms to pursue several potential leads simultaneously, so that products can be developed quickly. There is little risk that payments made as a result of a purchase commitment could exceed the cost of saving the equivalent number of lives using existing approaches.

Prior work by the author and others suggests that an annual market of $500 million or more in years with peak sales is needed to motivate substantial research (Kettler 1999; Berndt et al. 2005; Mercer Management Consulting 1998). Berndt et al. (2005) estimate that the total vaccine purchase commitment size that would be necessary to create a “market” comparable to markets of existing pharmaceutical products would be approximately $3.1 billion (in 2004 dollars).

A condition of participation in the program could be agreement among developing firms to license the products to producers in developing countries after a certain pre-set number of doses of the vaccine have been sold. Potential developers are likely to heavily discount sales after, say, ten years, in part because in more normal circumstances developers are likely to assume that competing products are apt to emerge after ten years in any case, thus driving down prices of their product.10

It is useful to consider the cost-effectiveness of a commitment to guarantee a price of $15 (in 2004 dollars) for the first 200 million people immunized against malaria in exchange for a commitment that, after the initial purchases, the price would drop to $1 per person—which is still more than the current price of many EPI vaccines. The exact cost-effectiveness depends on a variety of assumptions.11 But to get an idea of the magnitudes, consider a case in which (a) the contract covers all countries with a per capita income of less than $1,000 per year with sufficient disease prevalence to make vaccination worthwhile; (b) countries adopt the vaccine over seven years at rates consistent with those of the EPI program; (c) the vaccine is deliverable with the EPI vaccines; and (d) the vaccine is 60% effective and protects against infection for ten years. Under this fairly conservative set of assumptions, the cost—including incremental delivery costs—per discounted DALY saved over a fifty-year horizon would be about $15 (in 2004 dollars). Under similar assumptions, a commitment to guarantee a price of $15 for the first 200 million people immunized against HIV would cost about $17 per DALY saved, and a commitment for a tuberculosis vaccine would cost about $25 per DALY saved.12

(p.330) It is thus clear that purchases under a vaccine commitment would save more lives than almost any alternative use of funds.13 A commitment of 3 billion 2004 dollars in net present value of sales (as would be generated by the scenario described in the previous paragraph) would certainly be appropriate. The larger the commitment, the more firms will enter the search for a vaccine, and the faster a vaccine is likely to be developed. Since malaria kills 3,000 people every day, erring on the side of parsimony does not seem wise.


This essay focuses on purchase commitments for vaccines for malaria, HIV, and tuberculosis. Potentially, purchase commitments could be used to encourage research not only on vaccines, but also on other techniques for fighting disease, including drugs, diagnostic devices, and insecticides against the mosquitoes that transmit malaria. Given a sufficient budget, it might be appropriate to commit in advance to purchase vaccines or drugs developed against other diseases that primarily affect developing countries. However, if funding is tightly limited, it may be appropriate to target the most deadly diseases. Table 21.1 shows the number of deaths caused annually by various diseases for which vaccines are needed.

Including some easier-to-develop vaccines in a purchase commitment for a range of diseases, perhaps maintaining separate funds (or making separate financial commitments) for different diseases, would enable the program to build up a reputation for fair play and for fulfilling promises. It also may be useful to first experiment with purchase commitments for a few vaccines or drugs and then consider modifying or extending the program based on the resulting experience. Of course, it is likely to require time and experimentation to refine this new tool, just as it took time for institutions such as the patent system or the peer review process to evolve into their current forms. The institutions that today are integral in supporting our systems of innovation required both time and trial-and-error to develop. The first step in developing price guarantees as a tool for encouraging R&D would be to try the system in a few cases where current R&D incentives are inadequate and where the pull approach seems well suited to fill the gap.


A vaccine commitment has considerable appeal across the ideological spectrum as a market-oriented mechanism that brings the resources and inventiveness of the private sector to the fight against diseases disproportionately killing some of the world's poorest people. To move forward, it will be necessary for institutions with sufficient resources (such as national governments, international organizations, or private foundations) to launch a legally binding commitment program. A host of policy leaders and organizations (p.331)

Table 21.1 Deaths from Diseases for Which Vaccines Are Needed


Deaths (000)a




















Enterotoxic E. coli



Respiratory syncytial virusc












Chagas disease









Total deaths



Source: Children's Vaccine Initiative, CVI Forum 18, July 1999, p. 6.

aEstimated, World Health Report, WHO, 1999

bA pneumococcus vaccine was just approved for use in the United States, but it needs to be tested in developing countries, and perhaps modified accordingly.

cThe Jordan Report, NIAID, 1998

dR. Bergquist, WHO, personal communication

have endorsed the concept of vaccine commitments. The Clinton administration proposed a pull program for HIV, tuberculosis, and malaria vaccines, and Secretary of the Treasury Larry Summers was a key advocate. Legislation incorporating these provisions was introduced in Congress by Senators Bill Frist and John Kerry and by Representatives Nancy Pelosi and Jennifer Dunn. An Institute of Medicine committee has also recommended pull programs for vaccines in the United States (IOM 2003). The Bush administration's Project Bioshield, intended to improve vaccines and drugs that protect against chemical and biological warfare, uses a spending authority intended to function as a pull program. However, a key weakness of Project Bioshield is that the government is not committed to paying specific prices for new therapies, so developers still run the risk that, after the fact, the government will offer terms that do not cover the fixed costs. The G-8 Group of Nations has also strongly supported the idea of advance purchase commitments. Fol (p.332) lowing its June 2005 Summit, the G-8 asked Italy to study the issue and to come up with more concrete proposals by the end of the year.

In February of 2000, then World Bank president James Wolfensohn, indicated that the institution planned to create a $1 billion fund to help countries purchase specified vaccines if and when they are developed (Financial Times 2000), though the World Bank has not acted on this commitment. The Gates Foundation, with $22 billion in assets and a focus on children's health in developing countries and on vaccines in particular, also would be well placed to forward a vaccine purchase commitment. While continuing to fund its other priorities, such a foundation could simply pledge that if a product were actually developed, the foundation would purchase and distribute it in developing countries.

Any of several organizations—including the World Bank, national governments, and organizations such as the Gates Foundation—thus would have the ability and the opportunity to create a credible purchase commitment to stimulate research on vaccines needed in developing countries. If such a commitment fails to induce the development of the needed vaccines, no funds would be spent. If it succeeds, millions of lives would be saved each year at remarkably low cost.


This chapter draws in part on the deliberations of the Pull Mechanisms Working Group of the Global Health Policy Research Network, a program of the Center for Global Development that is supported by the Bill and Melinda Gates Foundation. I am particularly grateful to Rachel Glennerster and Ruth Levine for extensive discussions, and would also like to thank Daron Acemoglu, Philippe Aghion, Martha Ainsworth, Susan Athey, Amir Attaran, Abhijit Banerjee, Amie Batson, Peter Berman, Ernie Berndt, Nancy Birdsall, David Cutler, Sara Ellison, Sarah England, John Gallup, Gargee Ghosh, Chandresh Harjivan, John Hurvitz, Eugene Kandel, Hannah Kettler, Jenny Lanjouw, Sendhil Mullainathan, Ariel Pakes, Ok Pannenborg, Leighton Reid, Sydney Rosen, Jeff Sachs, Andrew Segal, Raj Shah, Scott Stern, Larry Summers, Wendy Taylor, Jean Tirole, Adrian Towse, David Weber, and Georg Weizsäcker for comments and discussions on these issues. Radu Ban, Marcos Chamon, Andrew Francis, Fabia Gumbau, Amar Hamoudi, Jane Kim, Jean Lee, Ben Olken, Kathy Paur, Margaret Ronald, Courtney Umberger, Heidi Williams, and Alix Peterson Zwane provided excellent research assistance.

(p.333) (p.334)


Bibliography references:

Acemoglu, Daron, and Joshua Linn. “Market Size in Innovation: Theory and Evidence from the Pharmaceutical Industry.” Quarterly Journal of Economics 119 (3) (2004): 1049–1090.

Berndt, Ernst, Rachel Glennerster, Michael Kremer, Jean Lee, Ruth Levine, Georg Weizsacker, and Heidi Williams. “Advance Purchase Commitments for a Malaria Vaccine: Estimating Costs and Effectiveness,” National Bureau of Economic Research, Working Paper #11288, 2005.

Cowman, A. F. “Mechanisms of Drug Resistance in Malaria.” Australia and New Zealand Journal of Medicine 25 (6) (1995): 837–844.

Desowitz, Robert S. The Malaria Capers: More Tales of Parasites and People, Research and Reality. New York: Norton, 1991.

DiMasi, Joseph, et al. “Cost of Innovation in the Pharmaceutical Industry.” Journal of Health Economics 10 (2) (1991): 107–142.

Elliott, Larry, and Mark Atkinson. “Fund to Beat Third World Disease.” Guardian, Feb. 23, 2001. Available online at http://www.guardian.co.uk/globalisation/story/0,,441883,00.html. Financial Times. “Discovering Medicines for the Poor.” Feb. 2, 2000, 7.

Finkelstein. Amy. “Health Policy and Technological Change: Evidence from the Vaccine Industry.” Quarterly Journal of Economics 119 (2) (2004): 527–564.

GAVI (Global Alliance for Vaccines and Immunizations). “Health, Immunization, and Economic Growth. Research Briefing #2: Vaccines Are Cost-Effective. A Summary of Recent Research.” 2004. Available online at http://www.vaccinealliance.org.

Glennerster, Rachel, and Michael Kremer. “A Vaccine Purchase Commitment: Cost-Effectiveness Estimates and Pricing Guidelines.” Mimeo, Harvard University, 2001.

G8 Finance Ministers' Conclusions on Development, London, June 10–11, 2005. Available online at http://www.g8.gov.uk/servlet/Front?pagename=OpenMarket/Xcelerate/ShowPage&c=Page&cid=1078995903270&aid=1115146455234.

Herfkens, Eveline. “Strategies for Increasing Access to Essential Drugs: The Need for Global Commitment.” Presentation at the Conference for Increasing Access to Essential Drugs in a Globalised Economy, Amsterdam, Nov. 25–26, 1999.

(p.335) IOM (U.S. Institute of Medicine). Financing Vaccines in the 21st Century: Assuring Access and Availability. Washington, D.C.: National Academy Press, 2004.

Kettler, Hannah E. Updating the Cost of a New Chemical Entity. London: Office of Health Economics, 1999.

Kim-Farley, R., and the Expanded Programme on Immunization Team. “Global Immunization.” Annual Review of Public Health 13 1992: 223–237.

Kremer, Michael. “Creating Markets for New Vaccines—Part I: Rationale.” In Innovation Policy and the Economy. Edited by Adam B. Jaffe, Josh Lerner, and Scott Stern. Cambridge, Mass.: MIT Press, 2001a.

Kremer, Michael. “Creating Markets for New Vaccines—Part II: Design Issues.” In Innovation Policy and the Economy. Edited by Adam B. Jaffe, Josh Lerner, and Scott Stern. Cambridge, Mass.: MIT Press, 2001b.

Mansfield, Edwin, et al. The Production and Application of New Industrial Technology. New York: Norton, 1977.

Mercer Management Consulting. “HIV Vaccine Industry Study, October-December 1998.” World Bank Task Force on Accelerating the Development of an HIV/AIDS Vaccine for Developing Countries, 1998.

Mitchell, Violaine S., Nalini M. Philipose, and Jay P. Sanford, eds. The Children's Vaccine Initiative: Achieving the Vision. Washington, D.C.: National Academy Press, 1993.

Moorthy, Vasee, Michael Good, and Adrian Hill. “Malaria Vaccine Developments.” Lancet 363 2004: 150–156.

Morantz, Alison, and Robert Sloane. “Vaccine Purchase Commitment Contract: Legal Strategies for Ensuring Enforceability.” Harvard University, 2001. Mimeo.

Murray, Christopher J. L., and Alan D. Lopez. Global Health Statistics: A Compendium of Incidence, Prevalence, and Mortality Estimates for Over 200 Conditions. Cambridge, Mass.: Harvard School of Public Health/WHO, 1996a.

Murray, Christopher J. L., and Alan D. Lopez, eds. The Global Burden of Disease: A Comprehensive Assessment of Mortality and Disability from Diseases, Injuries, and Risk Factors in 1990 and Projected to 2020. Cambridge, Mass.: Harvard School of Public Health/WHO, 1996b.

Nadiri, M. Ishaq. “Innovations and Technological Spillovers.” NBER working paper 4423, 1993.

Neumann, Peter J., Eileen Sandberg, Chaim A. Bell, Patricia Stone, and Richard H. Chapman. “Are Pharmaceuticals Cost-Effective? A Review of the Evidence.” Health Affairs (Mar.–Apr. 2000).

Pecoul, Bernard, Pierre Chirac, Patrice Trouiller, and Jacques Pinel. “Access to Essential Drugs in Poor Countries: A Lost Battle?” Journal of the American Medical Association 281 (4) (1999): 361–367.

PhRMA. “PhRMA Industry Profile 2005.” 2005. Available online at http://www.phrma.org/publications/publications/17.03.2005.1142.cfm.

PIU (Performance and Innovation Unit, U.K. Cabinet Office). “Tackling the Diseases of Poverty: Meeting the Okinawa/Millennium Targets for HIV/AIDS, Tuberculosis, and Malaria.” London, May 8, 2001. Available online at http://www.strategy.gov.uk/downloads/su/health/recommendations/default.htm.

Shulman, Sheila, Joseph DiMasi, and K. I. Kaitin. “Patent Term Restoration: The Impact of the Waxman-Hatch Act on New Drugs and Biologics Approved 1984–1995.” Journal of BioLaw and Business 2 (4) (1999): 63–68.

(p.336) UNAIDS. 2004 Report on the Global AIDS Epidemic. Geneva: UNAIDS, 2004.

Wellcome Trust. “An Audit of International Activity in Malaria Research.” 1996. Available online at http://www.wellcome.ac.uk/doc%5Fwtd003221.html.

WHO (World Health Organization). Investing in Health Research and Development: Report of the Ad Hoc Committee on Health Research Relating to Future Intervention Options. Geneva: WHO, 1996.

WHO. Anti-Tuberculosis Drug Resistance in the World. Geneva: WHO, 1997.

WHO. Issues Relating to the Use of BCG in Immunization Programmes. By Paul E. M. Fine et al. Geneva: WHO, 1999.

WHO. Anti-Tuberculosis Drug Resistance in the World: Report No. 2. Geneva: WHO, 2000a. Available online at http://www.who.int/csr/resources/publications/drugresist/WHO_CDS_TB_2000_278/en/.

WHO. Global Tuberculosis Control 2000. Geneva: WHO, 2000b. Available online at http://www.who.int/docstore/gtb/publications/globrep00/index.html.

WHO. World Heath Report 2000. Geneva: WHO, 2000c.

WHO. WHO Global Tuberculosis Control Report. Geneva: WHO, 2002a.

WHO. World Heath Report 2002. Geneva: WHO, 2002b.

WHO, Regional Office for South-East Asia. “Prevention of Hepatitis B in India: An Overview” New Delhi: WHO, Aug. 2002.

World Bank. Disease Control Priorities in Developing Countries, edited by Dean T. Jamison et al. New York: Oxford University Press, 1993a.

World Bank. World Development Report 1993: Investing in Health. New York: Oxford University Press, 1993b.

World Bank. World Development Indicators. 2001. CD-ROM.

WTO. (World Trade Organization). Fact Sheet: TRIPS and Pharmaceutical Patents. Geneva: WTO, 2001. Available online at http://www.wto.org/english/tratop_e/trips_e/factsheet_pharm00_e.htm.


(1.)  The BCG vaccine has been much more effective in some trials than others: trials in Britain suggest effectiveness up to 80%, while those in the southern United States and southern India suggest close to zero efficacy (WHO 1999). A widely accepted explanation is that exposure to environmental mycobacteria, often found in warmer climates, reduces the protection provided by BCG.

(2.)  However, more than a quarter of children worldwide and over half of children in some countries do not receive the EPI vaccines, and 3 million lives are lost annually as a result (World Bank 2001). Only a small fraction of children in poor countries receive newer hepatitis B and Haemophilus influenzae b (Hib) vaccines, which are still on patent and hence significantly more expensive at a dollar or two per dose.

(3.)  Recent advances in immunology, biochemistry, and biotechnology have provided new tools for understanding the immune responses to these diseases, and genetic sequencing of HIV and the organisms causing tuberculosis and malaria is complete. Although some scientists are more pessimistic about the prospects for a malaria vaccine, Moorthy et al. (2004) argue that “although exact predictions are not possible, if sufficient funding were mobilized, a deployable, effective malaria vaccine is a realistic medium-term to long-term goal.” Recent results from a GlaxoSmithKline (GSK) vaccine trial in Mozambique offered some evidence that the development of an effective malaria vaccine may be possible: they suggest the vaccine reduced the risk of falling sick from malaria by nearly 30%, and halved the risk of contracting a severe case of malaria. Candidate vaccines have been shown to induce protection against tuberculosis infection and HIV in animal models and, for HIV, to induce immune responses in humans.

(4.)  For example, Africa now generates less than 1% of pharmaceutical sales (PhRMA 2005).

(5.)  Note, however, that the definition of tropical disease used in their assessment was narrow, and that many of the other drugs licensed in this period were useful in both developing and developed countries.

(6.)  Others have suggested using a cutoff equal to a country's per capita GNP (WHO, Regional Office for South-East Asia 2002), and several have noted that the World Bank may use this as a rule of thumb (GAVI 2004; WHO 2000c). For comparison, health interventions are considered cost-effective in the United States at up to 500 to 1,000 times this amount: $50,000–$100,000 per year of life saved (Neumann et al. 2000).

(7.)  Indeed, one reason why developed countries are developed may be that they were able to establish good reputations in a variety of areas, including research incentives. Developed countries typically have more stable governments that are more likely to invest in reputation formation for the long run.

(8.)  On the effect of market size on innovation in the pharmaceutical industry, see Acemoglu and Linn (2004) and Finkelstein (2004). Acemoglu and Linn exploit variations in market size for pharmaceuticals linked to demographic changes, and estimate that a 1% increase in the potential market size for a drug category leads to a 4%–6% increase in the number of new drugs in that category. Finkelstein investigates the private response to health policies that, in attempting to increase immunization rates, also increased the expected profits from new vaccines, and argues that a 1993 Medicare policy helped stimulate the R&D responsible for the approval (in 2003) of the first new flu vaccines since 1978.

(9.)  Excluding countries that would have bought vaccines, in the absence of a program, at prices greater than or equal to the price paid by the program, would, however, increase incentives to develop products.

(10.)  The life of a patent is twenty years. However, a vaccine would reach the market only several years after the date of application for a patent. The effective life of a patent is the number of years remaining from the time that it is first brought to market. Shulman et al. (1999) report that the average effective patent life for new drugs and biologicals is 11.2 years under the Waxman-Hatch Act, which granted extra protection to inventors to partially make up for loss of patent life during regulatory review. Without the act, patent life would be 8.2 years. The act covers the United States only, and there is no reason to believe that developing countries will offer similar patent protection.

(11.)  A more detailed description of our data and assumptions on the burden of disease, fertility, and delivery costs, and the benefits of vaccination can be found in “A Vaccine Price Guarantee: Preliminary Cost-Effectiveness Estimates and Pricing Guidelines.” This is available at http://post.economics.harvard.edu/faculty/kremer/vaccine.html.

(12.)  For HIV, the calculation assumes that ten-year-olds will be vaccinated at a higher cost of delivery than newborns, and also considers a twenty-year vaccine rather than a five-year vaccine. The calculation for tuberculosis also considers a vaccine that protects for twenty years after immunization.

(13.)  Indeed, commitments would remain cost-effective under a wide range of assumptions about vaccine efficacy, level and speed of vaccine adoption, and the amount of money spent for vaccines. On the whole, there is a large range of values under which a vaccine commitment would be sufficient to stimulate substantial research, yet still be extremely cost-effective. See http://post.economics.harvard.edu/faculty/kremer/vaccine.html.