1. Failures of the Market
   1. Investors cannot appropriate all the benefits of research.
   2. R&D is risky, and investors are risk averse.
   3. There are perverse incentives for R&D on "adverse effects" research.
   4. There are socially important research projects that do not result in commercial products.
   5. Markets do not serve poor populations.
2. Failures of Intellectual Property Rights.
   1. Property Rights systems are difficult and costly to administer.
   2. Secrecy undermines scientific progress.
   3. Races to be "first" are inefficient.
   4. Pricing above post invention marginal cost is economically inefficient.
   5. Strong property rights can lead to ethically unacceptable outcomes.
   6. Strong property rights stifle research.
   7. Complex property rights, bargaining costs and blocking of commercial development.

1. Governments (or other donors) can set research priorities, and direct R&D investments in areas that are not attractive to private investors.
2. Governments (or other donors) can require that R&D enter the public domain, be publicly disclosed or meet other public interest conditions.
3. If the R&D creates intellectual property rights, governments (or other donors) can and do retain rights in inventions, as a condition of funding.

1. Tax credits on R&D.
2. Extensions of exclusive rights as a reward for R&D.
   1. As a reward for pediatric studies on drugs.
   2. Transferable patent extensions as a reward for R&D on neglected diseases.
3. Purchase funds.
4. Prizes.
5. Orphan Drug market exclusivity.
   1. Oxandrolone
   2. 1999 US orphan drug approvals
   3. Data exclusivity

1. Miscellaneous US government investment mandates.
2. Senator Specter's R&D funds.
3. Mandates tied to extensions of data exclusivity.
4. Other country approaches.
5. Amsterdam statement to WTO member states on access to medicines.
6. The advantages of mandates or strong linkage.
   1. Governments can determine by policy the aggregate level of R&D funding.
   2. Governments can determine the composition of R&D funding.
   3. It is possible to increase access to medicines and increasing R&D at the same time.
   4. The system can be as transparent as policy makers wish.
   5. Mixed funding models are possible.

Nearly every effort to lower prices of medicines or other medical technologies eventually leads to debates over the need to provide economic incentives to fund health care R&D. Efforts to provide lower prices, shorter patent terms, compulsory licensing, reform the orphan drug act, or countless other initiatives are confronted with the inevitable question: will this reduce R&D efforts for new therapies, and if so, do we want the lower levels of health care R&D?

This, however is only the most superficial level of the debate. Market incentives for health care R&D are not efficient. This paper explores the mechanisms available to governments (and donors) to enhance R&D efforts.

1. Failures of the Market.

   A few of the many relevant market failures are:

   1. Investors cannot appropriate all the benefits of research. This happens for several reasons. There are many research "spillovers" that benefit others. Early stages of R&D in particular create information that is not protected by patents or other rights, and indeed, the significance of which may not be appreciated by the researcher, or by anyone for a while. Since investors won't obtain property rights to all of the benefits, the market under invests in such research.

   2. R&D is risky, and investors are risk averse. It is difficult to share information about R&D programs with investors, for fear the information will be leaked to competitors.

   3. There are perverse incentives for R&D on "adverse effects" research. Research that finds adverse effects of medicines, or identifies populations that should not use products, results in negative private returns to right owners, and thus too little private investment.

   4. There are socially important research projects that do not result in commercial products. Research on appropriate technologies or delivery systems for the rural poor, the relationship between health and diet, environmental health risks, or many other important areas that do not result in commercial "products" will not attract private investors.

   5. Markets do not serve poor populations. For many tropical illnesses and other diseases that affect that poor, low income levels deter private investors. However, there is global support for public investments in health care R&D for neglected diseases, for humanitarian reasons as well as self interest.

   Intellectual property laws are efforts to deal with the first market failure -- the failure to appropriate the benefits of the research, typically by assigning property rights. And while intellectual property rights do permit the investors to appropriate some of the benefits, they also create a host of other problems.

2. Failures of Intellectual Property Rights.

   For example:

   1. Property Rights systems are difficult and costly to administer. The intellectual property system is arbitrary, expensive to manage, and prone to all sorts of errors. The "noise" in the system creates risk for investors, and the costs of litigating IP claims consumes both time and managerial talent.

   2. Secrecy undermines scientific progress. While the patent system is supposed to lead to disclosure of inventions (a putative requirement for getting a patent), a practical effect of patent system is increased secrecy and less sharing of information, in order to maintain proprietary claims and competitive advantages. This is a particularly difficult problem in universities, where academic traditions favor greater sharing of information, but "technology transfer" policies encourage university researchers to mimic the actions of firms.⁽¹⁾

   3. Races to be "first" are inefficient. Because the basis for a patent claim is being "first," there are often highly redundant and wasteful "patent races" to win the legal rights associated with a patent.⁽²⁾

   4. Pricing above post invention marginal cost is economically inefficient. It is inefficient to reward ex ante research efforts with ex post grants of exclusive commercial rights. Inventions, like many other information goods, are sometimes called "quasi public goods." The cost of the good is indivisible, and thus, its marginal cost is zero. However, unlike a pure public good, it is possible, legally, to award rights, and charge for the right to use the invention. However, this results in prices that are too high -- above marginal cost.

   5. Strong property rights can lead to ethically unacceptable outcomes. It is ethically offensive to withhold health care from the poor by pricing inventions far above marginal costs, particularly when the policy is based upon government policy.

   6. Strong property rights stifle research. The proliferation of patents rights has a chilling effect on researchers, who face infringement actions for using patented materials, processes or research tools.⁽³⁾
   7. Complex property rights, bargaining costs and blocking of commercial development. For inventions involving multiple patents held by different parties, there are high transaction costs associated with bargaining over rights, which can lead to blocking of commercial development.⁽⁴⁾

None of these shortcoming by themselves justify the elimination of patents or other aspects of intellectual property system. Even with flaws, the protection of intellectual property creates benefits, and increases R&D levels. However, there are clearly justifications for an active government role in promoting health care research. The pharmaceutical industry is itself a strong supporter of a large public sector role in promoting R&D, backing a wide range of government interventions to support higher levels of R&D.

The purpose of this paper is to explore the relative merits of different public approaches to increase R&D.

1. Governments (or other donors) can set research priorities, and direct R&D investments in areas that are not attractive to private investors.

2. Governments (or other donors) can require that R&D enter the public domain, be publicly disclosed or meet other public interest conditions.

3. If the R&D creates intellectual property rights, governments (or other donors) can and do retain rights in inventions, as a condition of funding.

The International Federation of Pharmaceutical Manufacturers Associations (IFPMA) and its member organizations are frequent proponents of various government subsidies, tax breaks or other financial incentives that are designed to increase private R&D. Among the typical set of incentives are:

1. Tax credits on R&D. In the United States, pharmaceutical companies are eligible to a tax credit for increasing R&D, and for half the costs of expenditures on clinical trials for orphan indications, defined, in the US, for example, as a use for a drug that has a potential US client population of less than 200,000. There are also various proposals for tax credits on vaccine research. Tax credits lower the private cost of doing the R&D, thereby increasing the private returns. Among the criticisms of a general tax credit are that this does not do enough to change the composition of R&D. Targeted credits, like the US Orphan Drug tax credit, are somewhat better at this, but less so than public grants programs. Credits are also criticized on the grounds that the amount of public subsidy to a company or for a particular research endeavor is not transparent -- there is no public information on who gets the credits or what the credits are used for. And also very important, the government does not retain any rights in the research. There is also no "needs test" for these subsidies, and credits are given even when the credit does not influence investment decisions.⁽⁵⁾

2. Extensions of exclusive rights as a reward for R&D. The pharmaceutical industry has proposed in a variety ways that governments reward R&D with extensions of exclusive rights.
   1. As a reward for pediatric studies on drugs. Under Section 505A of the FDA Act,

      "If . . . the Secretary determines that information relating to the use of a new drug in the pediatric population may produce health benefits in that population, the Secretary makes a written request for pediatric studies . . and such studies are completed within any such timeframe,"

      the FDA extends the Orange Book patent exclusivity, Hatch/Waxman data exclusivity and orphan drug marketing exclusivity by six months. Under this program, a company can file even a useless study, and trigger a two month extension of exclusivty, because that is the period to review the application. If the study is accepted as having value, the firm gets the entire six month extention of exclusivity, even when the study was small, and the economic value of the exlusivity extention is large.

      Extensions of marketing exclusivity can be worth a lot to the drug companies. AstraZeneca is reportedly seeking a six month extension for its Losec product, which generates more than $16 million per day worldwide. Prices for Losec in the US market are $3.52 per pill, while generic copies of the drug sell for as low as $.07 to $.17 per pill in some markets. The costs of a pediatric clinical trial may be relatively small. According to a recent CPT survey, the typical cost of outsourcing clinical trials for pharmaceutical drugs are in the range of $2,000 to $7,000 per patient, numbers which are consistent with data from the National Cancer Institute (NCI). At $5,000 per patient, a trial involving 500 patients would cost $2.5 million, or less than 4 hours of Losec sales.⁽⁶⁾

      The US program for pediatric studies would make more sense if the Secretary could negotiate the term of the extension, up to the maximum of six months, and also require disclosures of the costs of the additional studies, in order to facilitate a cost benefit analysis of the program, and to provide some measure of the relative advantages of the government funding its own studies.

   2. Transferable patent extensions as a reward for R&D on neglected diseases. This proposal is a current favorite of the major pharmaceutical companies. As the Losec example above illustrates, companies can protect huge profits by extending, even for a few months, patents on blockbuster drugs. The industry has proposed the award of transferable patent extensions for up to three years, in return for carrying out various R&D programs -- for example, for development of a new drug for tuberculosis. Again, the criticism of patent extensions are that they are expensive, in terms of higher prices, and undoubtably wasteful, from an economic point of view.

      What would make a program like this more interesting would be some type of market mechanisms to get the maximum public health benefits for the minimum cost to the public. Alternative approaches might involve bidding or negotiations on a variable that was linked to the extension of exclusivity. For example, such mechanisms might include bidding for:

      1. The shortest period of exclusivity, in return for a specified R&D outcome,

      2. The least amount of money in sales, before the exclusivity expired, or

      3. The most amount of money invested in qualified R&D projects.

      There is also the issue of the intellectual property rights from the R&D. Under a program of grants or contracts, governments can negotiate or mandate a share of the intellectual property rights in the R&D, or set public interest conditions on issues such as pricing or access, as was done recently, for example, in South Africa for an AIDS vaccine research project.⁽⁷⁾ If the public is in fact paying for the R&D, indeed, paying several times over for the R&D, it is reasonable to ask that the public obtain rights in the research that they have already paid for.

3. Purchase funds. Jeffrey Sachs and others have suggested governments and donors endow huge funds to guarantee purchases of new medicines, solving the lack of purchasing ability market failure. The practical difficulties in such approaches are many, including, for example, determining the criteria for products that could be bought from the pool, or dealing with perverse incentives regarding the timing of product development. A more general problem is that the donors, government or private, deal with the inventions after intellectual property rights are in hand. By not directly funding the pre discovery R&D, there are no mechanisms to assert property rights, post invention. For this reason the purchase fund approach is an extremely expensive way to buy R&D.

4. Prizes. A variation on the purchase fund approach is to create "prize" funds for R&D. Under this type of approach one could imagine using the prize fund to purchase intellectual property rights, to deal with post discovery access issues. However, like the purchase fund approach, it can be an expensive proposition when the fund is dealing with post discovery rights.

5. Orphan Drug market exclusivity. The US and EU orphan drug programs both provide grants of market exclusivity to enhance R&D investments in neglected diseases. When the US Orphan Drug Act (ODA) was first proposed, there were no provisions for marketing exclusivity, and indeed the problem was perceived to be that the markets were so unprofitable that market-exclusivity was a non-issue. Why would exclusivity be important for a market that no one wanted in the first place? However, pharmaceutical company lobbyists successfully made a grant of market exclusivity one of the core features of the US program. Under the US program, the marketing exclusivity applies to both new and old drugs, and gives a firm exclusivity for any indication for which the there are fewer than 200,000 US patients. The US government can grant marketing exclusivity for one indication, such as the BMS version of Taxol for ovarian cancer, and then later approve the same drug for an indication such as breast cancer, where the client population is large, and not subject to marketing exclusivity, and then again for Kaposi's sarcoma, where the client population qualifies for the orphan designation.

   The ODA market exclusivity can be broader than the rights granted under a patent, and in some cases will block the introduction of products that are protected under different patents⁽⁸⁾. Companies frequently use the ODA marketing exclusivity provisions as a weapon against entry by generic products, for example, as BMS did in the case of the IVAX application to market generic paclitaxel for Kaposi's sarcoma. In this case, BMS and IVAX both applied for ODA marketing exclusivity, but BMS beat the IVAX application by just six days.

   1. Oxandrolone

      The costs of orphan drug exclusivity to consumers can be very high. For example, Oxandrolone is an anti-wasting drug that had been on the market for decades, with a generic price of $.30 per pill. Bio-Technology General (BTG) received an orphan drug designation for the product in 1995, for use in treating wasting of AIDS patients, an old use for a new illness, qualifying as a new indication, and raised the price to $3.75 per pill. The cost of the drug at the higher prices was estimated to be $5,475 to $43,800 per year in one analysis.⁽⁹⁾ BTG responding to criticisms of the higher price by capping the price at $15,000 per year for any patient, but the use of the medicine was greatly reduced by the higher price.

      Given the high cost to consumers of marketing exclusivity, one has to ask about the value of this approach, in terms of incentives for R&D, particularly when the drugs are also eligible for patents and data exclusivity. Investors receive 20 years of exclusive rights under patents when there is an invention, and 5 to 10 years of data exclusivity in the US and Europe, when they invest in data needed for registration of new drugs. The ODA marketing exclusivity is most important when:

      1. The company cannot claim a patentable invention, and

      2. Does not own the rights to the data used for registration of a new product.

      The ODA basically rewards companies that have done very little. Supporters of the ODA say it rewards investors for investments in research associated with new uses of an older drug, or for the start-up costs of marketing a product, but it is important to appreciate the limits of the companies contributions when the drug is not new, and when the company cannot claim a patentable invention.

      A review of the early history of the orphan drug act illustrates the paucity of the new private sector investments in clinical research for orphan drugs. From 1983 to 1993, the total amount claimed under the orphan drug tax credit was $107 million. This represents half of the cost of claimed expenses for human use clinical trials for orphan drugs, or $213 million before the credit. During this period the US FDA gave marketing approval to 93 orphan products, including blockbuster products that generated hundreds of millions in annual sales. The cost of the marketing exclusivity, in terms of higher prices to consumers, was several multiples of the $213 million in private investment in orphan drug clinical trials. For example, Amgen used its orphan drug marketing exclusivity to build a thicket of process and indication patents to bar future competitors for EPO, a drug that now generates more than $4 billion globally in annual sales.

   2. 1999 US orphan drug approvals

      In 1999, there were seven US orphan drug approvals. Of the seven, four have FDA orange book listings for patents, and three do not. None of the seven drugs are marketed by the inventors. The number of patients in clinical trials referenced in the drug approval ranged from 152 to 1,281, with an average of 588 patients. The prices for the orphan products are high, running as much as $72 thousand for temozolomide, a drug sold by Schering-Plough for the treatment of refractory anaplastic astrocytoma. Temozolomide was approved on the basis of a clinical trial involving only 162 patients. Even at $10 thousand per patient, which is 60 percent more than the average for the fy 1999 NIH DCP cancer trials⁽¹⁰⁾, the cost would have been 1.6 million, half of which would have been offset by the US orphan tax credit.

      As flawed as the US ODA is, it has become the model for similar laws, pushed by the industry and some patient groups, in Singapore, Australia, Japan and the European Union.

6. Data Exclusivity. The US, the EU and several other countries have a sui generis form of protection on data rights. The US law was part of a compromise in the 1984 Hatch/Waxman Act, which also included provisions for easier generic drug entry, and provides for five years of exclusivity for data used to support new drug approvals. The European Union has similar laws at the EU and country level, which provide 6 to 10 years of data exclusivity, the longer term originally designed to compensate for a lack of patent protection in Spain and Portugal. The pharmaceutical industry is seeking a harmonization of 10 years for all countries, and they are also trying to extend the rights to any registration data, including data on new indications of older drugs.

   The data exclusivity provisions are irrelevant in cases where there is a patentable invention, and it is also worth noting that for products with small client populations, there is strong marketing exclusivity now under US and EU orphan drug laws, so the protections are for non-inventions that serve large populations.

   Again, programs like this would make more sense if there was a clearer connection between the benefits to society and the company's investments. A starting point would be to routinely require public disclosure of the costs of data collection, and to link the protections to some measure of cost recovery (including opportunities for competitors to share costs) rather than providing a flat 5 to 10 year period of exclusivity.

A different approach than those described above would be government mandates or strong linkage to promote public R&D objectives.

1. Miscellaneous US government investment mandates. Governments can and do mandate firm spending or contributions to spending in a wide range of areas. For example, in the employment area, there have been various proposals for mandated investments in worker training. A recent example is the US H-1B program for temporary workers in the technology sector. Under this program the employer pays a fee of $500 for each employee, which is used for training and scholarships for US citizens. In the US banking sector, there is the "Community Reinvestment Act," which requires US banking institutions to invest in low income neighborhoods. This law was first passed in the 1970s, in response to criticisms that banks were "redlining" poor neighborhoods.

   The government can also raise money from one part of the economy to address entirely different needs. For example, in 1997 Senator Specter sponsored legislation (S. 435) to create to create a "Healthy Children's Trust Fund," to provide funds so that eligible children could get vouchers to purchase state health insurance. The funding for the trust fund was to come from an auction of spectrum for wireless telecommunications.

2. Senator Specter's R&D funds. In early 1995, Senator Specter introduced S. 18, the "Health Care Assurance Act," which, among other things, would have created a "Trust Fund for Medical Treatment Outcomes Research." This R&D fund would have been funded by a tax of .1 percent of the premiums on private health insurance. Later the same year, Senators Hatfield, Harkin, Boxer, Inouye, Simon, Kerrey, Mikulsk and Moynihan sponsored S. 1251, "the `National Fund for Health Research." The money for the fund would have come from an increase in the excise taxes on tobacco products. This bill also would have created a voluntary check-off system, whereby taxpayers could designate $1 of their tax refunds be donated to medical research.

3. Mandates tied to extensions of data exclusivity. In 1997, Senator Specter and others proposed, at the urging of Bristol-Myers Squibb, that the US government's Hatch/Waxman "data exclusivity" protections be extended from 5 to 10 years, in return for the government receiving 3 percent of a drug's revenues to be used for research, and a commitment that the company would spend another 3 percent on R&D.

4. Other country approaches. Other countries have explored similar approaches. In the United Kingdom, the government permits higher reimbursement prices for pharmaceutical when companies have above average R&D expenditures. In several countries, including Canada, the drug companies have negotiated promises for increased R&D levels, in return for changes in public policy. In India, the government has tried to push for minimum levels of R&D investments, but has meet resistance. In Argentina, there are proposals in the domestic industry for a tax on pharmaceutical sales to fund an Argentine R&D program.

5. Amsterdam statement to WTO member states on access to medicines. The joint CPT/HAI/MSF statement from the November 1999 conference on "Increasing Access to Essential Drugs in a Globalised Economy" called for: "Compulsory research obligations, such as requirements that companies reinvest a percentage of pharmaceutical sales into R&D, either directly or through public or private sector R&D programs."

6. The advantages of mandates or strong linkage are many, including:
   1. Governments can determine by policy the aggregate level of R&D funding.

   2. Governments can determine the composition of R&D funding.

   3. It is possible to increase access to medicines and increasing R&D at the same time.

   4. The system can be as transparent as policy makers wish.

   5. Mixed funding models are possible. Governments can decide if R&D funds are invested by governments or companies or a combination of both. Mandates and strong linkage can be use in combination with other approaches, including intellectual property rights and direct public investment via general tax revenue.

1. Corporate Collaborations: Scientists Can Face Publishing Constraint," .the Scientist 13[11]:1, May 24, 1999; S.A. Rosenberg, the New England Journal of Medicine,, 334:392-4, 1996), D. Blumenthal et al., New England Journal of Medicine, 334:368-73, 1996). Rosenberg SA. Secrecy in medical research. N Engl J Med 1996, 334:392-4; Blumenthal D, Campbell EG, Causino N, Louis KS, "Participation of life-science faculty in research relationships with industry," N Engl J Med 1996, 335(23):1734; Steven Benowitz, "Progress Impeded?," the Scientist 10[7]:1, Apr. 01, 1996.

2. J. Hirshleifer & John G. Riley, The Analytics of Uncertainty and Information, 17 J. ECON. LIT. 1375, 1404 (1979); Partha Dasgupta & Joseph Stiglitz, Industrial Structure and the Nature of Innovative Activity, 90 ECON. J. 266, 279 (1980); Partha Dasgupta & Joseph Stiglitz, Uncertainty, Industrial Structure, and the Speed of R&D, 11 BELL J. ECON. 1, 3 (1980); Pankaj Tandon, Rivalry and the Excessive Allocation of Resources to Research, 14 BELL J. ECON. 152 (1983).

3. Michael A. Heller & Rebecca S. Eisenberg, "Can Patents Deter Innovation? The Anticommons in Biomedical Research," Science, Volume 280 Number 5364, May 1 1998, pp. 698-701; Report of the National Institutes of Health (NIH), Working Group on Research Tools Advisory Committee to the Director, June 4, 1998; Steve Bunk, "Researchers Feel Threatened by Disease Gene Patents," The Scientist 13[20]:7, Oct. 11, 1999; June 4, 1998, Naomi Freundlich, "Will increasingly aggressive licensing terms on research tool patents hurt basic research?", Signals; June 12, 1998, Naomi Freundlich, "Cre-lox controversy divides institutions, prompts NIH panel," Signals; Intellectual Property Rights and the Dissemination of Research Tools in Molecular Biology: Summary of a Workshop Held at the National Academy of Sciences, February 15-16, 1996 (1997); Joint Recommendations of Mouse Genomics and Genetics Subgroup and Mouse Models of Human Cancers Subgroup.

4. The recombinant hepatitis B vaccine, for example, requires fourteen different patents to produce.

5. There is a substantial literature questioning the efficacy of the general R&D tax credit, as well as concerns that firm modify accounting rules to qualify for the credits, without significantly increasing spending. Rachel Griffith, John Van Reenen and Nick Bloom, "Promoting R&D through fiscal incentives: an assessment of the arguments." Institute for Fiscal Studies, Innovation and R&D. August 15, 2000.

6. Based upon a 24 hour day.

7. 5 October, 2000, Pierre Steyn, "SA to pay less for HIV drug than other countries," News24.za.

8. Peter S. Arno, Karen Bonuck, and Michael Davis, 1995, "Rare Diseases, Drug Development, and AIDS: The Impact of the Orphan Drug Act," The Milbank Quarterly, Vol 73,No 2; sThoene, J.G. 1991. Curing the orphan drug. Science, 251:1158_59; Asbury, Carolyn, 1991, "The Orphan Drug Act: The first 7 years." Journal of the American Medical Association 265(7):893_897; James Love, "Comments on the Orphan Drug Act and Government Sponsored Monopolies for Marketing Pharmaceutical Drugs." United States Senate, Committee on the Judiciary, Subcommittee on Antitrust, Monopolies and Business Rights, Anticompetitive Abuse of the Orphan Drug Act: Invitation to High Prices, January 21, 1992, Serial Number J-102-48, pages 259-283; David M. Richardson, "The Orphan Drug Tax Credit: An Inadequate Response to An Ill-Defined Problem," The American Journal of Tax Policy, Vol. 6:135. pp. 135-210, 1987; Patricia J. Kenney, "The Orphan Drug Act--Is it a Barrier to Innovation? Does it Create Unintended Windfalls?" Food Drug Cosmetic Law Journal, Vol. 43, pp 667-679, 1988; James T. O'Reilly, "Orphan Drugs: The Strange Case of 'Baby M,'" Food Drug Cosmetic Law Journal, 42, 516-526; 1987. Donna Brown Grossman, "The Orphan Drug Act: Adoption or Foster Care?" Food Drug Cosmetic Law Journal, 39, 128-151, 1984.

9. 1996, Stephen LeBlanc, Rob Sabados and Don Howard, "Your money or our life, ACT UP's Survey of Unconscionable AIDS Drug Pricing: the story of little orphan Oxandrolone. (http://www.actupgg.org/abs_reps/dpr/gouge.html#Oxandrolone)

10. National Cancer Institute, "DCP Cooperative Group Treatment Trials and Funding, 1993 to 1999." The per patient cost for fy 1999 was $6,202.