A Better Way To Spur Medical Research And Development
Kremer.2 6/8/00 3:56 PM Page 34
R E S E A R C H A N D D E V E L O P M E N T
The purchase precommitment as a supplement
to patents and government-funded research
A Better Way to Spur Medical
Research and Development
B y R a c h e l G l e n n e r s t e r a n d M i c h a e l K r e m e r
istorically, societies have encour-
ous problems—and each has proved inadequate in spurring
aged research in a variety of ways:
the research needed to develop effective vaccines against hiv,
tuberculosis and malaria.
•
HPatents grant inventors monop-
olies over the goods that are pro-
PATENTS
duced from their ideas.
The Consumer’s View The monopoly pricing of patented
goods prevents some people who need those goods from
•Government directly funds research through such
buying them. This problem is illustrated vividly by the
programs as the National Science Foundation and the
recent dispute between the United States and South Africa
National Aeronautics and Space Administration.
over aids drugs. Up to 20 percent of pregnant women in
South Africa are infected with hiv. aids drugs cost more
• Prizes also have been used to spur research and
than $10,000 annually, well beyond the reach of most
development. For example, in 1714, after a British
South Africans. Alternative, generic versions of the drugs
fleet got lost and struck rocks off England’s coast,
would be much cheaper, but buying these products would
drowning 2,000 sailors, the British government
violate the patent rights of the original drug developers.
established a £20,000 prize for a method of deter-
To enable its citizens to obtain aids drugs more
mining longitude at sea. That prize led to the devel-
cheaply, South Africa is considering legislation to compel
opment of the chronometer.
patent holders to license their discoveries to generic man-
ufacturers and to allow the importing of cheap generic
Today, the United States government spurs research
drugs from countries that do not respect the original
mainly through direct funding and the granting of patents.
patents. Opponents of the legislation argue that if intel-
Both methods are vitally important, but each causes seri-
lectual property rights are not respected, private firms
will lose their incentive to develop new drugs. The Unit-
ed States initially opposed the legislation. However, when
Michael Kremer is a senior fellow at the Brookings Institution, a professor
aids
of economics at Harvard University, and a faculty fellow at the Center for
activists began protesting against Al Gore, who had
International Development, Harvard University. At the time this article
raised the issue as a chair of the U.S.-South Africa Bina-
was written, Rachel Glennerster was a visiting scholar at the Center for
tional Commission, the United States rapidly backed
International Development at Harvard University. She is currently a staff
down. The issues raised by the confrontation are deep.
member at the International Monetary Fund. The views expressed here,
Patents, and the resulting legal monopolies, create incen-
as well as any errors, are the sole responsibility of the authors and do not
tives for research and development that drive medical
necessarily reflect the opinions of the Executive Directors of the IMF, or
other members of the IMF staff.
progress. But under our current institutions, those same
R e g u l a t i o n
34
Vo l u m e 23, N o . 2
Kremer.2 6/8/00 3:56 PM Page 35
patents can sometimes prevent people from obtaining
have a mixed record. To take a few examples, the supersonic
drugs that they need to survive.
transport plane, the Carter administration’s synthetic fuel
program, and the Clinch River Breeder Reactor were spec-
The Inventor’s View Patents nevertheless create insuffi-
tacular failures.
cient incentives for original research because, even with
The government often has difficulty in selecting appro-
patents, inventors do not capture the full benefit of their
priate research projects and in motivating researchers to
inventions. First, as discussed above, some potential pur-
focus on developing viable projects. Researchers applying for
chasers of patented goods are not willing or able to pay
grants have an incentive to present the prospects of suc-
monopoly prices. Second, some of the benefit accrues to
cess in the best possible light to increase their chances of
those consumers who would be willing to pay even more
receiving funding. Research administrators in turn have
than the monopoly price for patented goods. Third, some
incentives to tell their superiors that prospects for success
of the benefit goes to those consumers who buy generic
are bright in order to increase the budgets of their divisions.
products after patents expire. Finally, some of the benefits
Once research grants have been made, researchers may
go to other researchers who draw on the research that led
not focus intently on developing viable products. Many aca-
to patented goods.
demic and government researchers have career incentives and
Many empirical studies suggest that inventors realize no
intellectual interests that orient them to fundamental sci-
more than half the returns to their inventions. Thus, many
ence, whereas the later stages of product development often
beneficial investments in r&d may be
forgone because the prospective
returns are too low.
Although patents give potential
Once projects have been started, they acquire
inventors too little incentive to do
original research, they create too
their own bureaucratic and political momentum,
strong an incentive to conduct “me
too” research aimed at designing
making them difficult to shut down, even if the sci-
around existing patents. Suppose,
for example, that a biotech or phar-
entific prospects for success appear dim.
maceutical firm developed a 100-
percent effective, safe, single-dose
aids vaccine. In an ideal world, the
firm would be amply compensated and the world’s vacci-
include activities that are not intellectually interesting. It can
nologists would turn most of their attention to other dead-
be difficult to determine whether a researcher is focusing
ly diseases. However, the patent system creates an incentive
on development of a product, trying to publish an academ-
for other firms to design around the first patent so as to pro-
ic paper, or preparing the next grant application. In con-
duce a competing vaccine and obtain a share of the market.
trast, a private firm pursuing a patent is paid only if it devel-
Sixty percent of patented innovations are imitated within
ops a successful product; thus its incentive is to choose
four years; the average cost of an imitation is typically two-
projects with a realistic chance of success and then to focus
thirds the original cost of an invention. Not only is this use
intently on developing viable products.
of scientific talent socially wasteful, but it reduces incentives
Another problem with direct government financing of
for developers to undertake original research.
r&d is that organized interests (e.g., defense contractors and
Although patents can create too much incentive to
aids activists) may lobby to influence these decisions,
develop substitute products designed around original
diverting research expenditures from objectives that are
patents, they can also block needed improvements that
scientifically meritorious or economically viable. Mem-
draw on the ideas covered by the original patents. For exam-
bers of Congress may support research projects because they
ple, the development of the high-pressure steam engine
are located in member’s districts, not because those projects
was blocked by James Watt’s patent covering all steam
are likely to succeed. Once projects have been started, they
engines; Watt’s steam engine was blocked by a previous
acquire their own bureaucratic and political momentum,
patent until he found a way to invent around it; and Thomas
making them difficult to shut down, even if the scientific
Edison’s improved version of the telegraph was blocked by
prospects for success appear dim.
a prior patent for many years.
In view of the politics of government-funded research,
it is not surprising that empirical studies suggest that the rate
GOVERNMENT RESEARCH AND DEVELOPMENT
of return on publicly financed r&d is much lower than
an alternative to patents as a way of encour aging
that of privately financed r&d.
research is for government to fund research directly,
through such entities as the National Institutes of Health.
PRIZES AND PATENT BUYOUTS
Government clearly has a role in financing basic research,
through the early nineteenth century, prizes were
but government programs to finance commercial r&d
used widely as an alternative to patents and government sub-
R e g u l a t i o n
35
Vo l u m e 23, N o . 2
Kremer.2 6/8/00 3:56 PM Page 36
sidies. For example, when Napoleon needed better ways
Through the buyout of Daguerre’s patent a valuable
to feed his troops, he established a prize that led to the
technology was placed in the public domain and more fully
development of food canning.
used. Further research was spurred because developers of
Prizes also have been used in recent times. In 1959, the
improved lenses or photographic chemicals did not have to
British industrialist Henry Kremer offered a prize of £50,000
worry that their work would be blocked by Daguerre’s
for the first substantial flight of a human-powered airplane.
patent. And because Daguerre’s sale of the patent was vol-
In 1977, Paul MacCready’s Gossamer Condor made histo-
untary, not coerced, incentives for invention were not weak-
ry by flying the one-mile long, figure-8 route required to
ened by setting a precedent for the expropriation of intel-
qualify for the prize. The next year, MacCready won a sub-
lectual property rights.
sequent £100,000 Kremer prize by flying the Gossamer
Of course, prizes can be used only when it is possible
Albatross across the English Channel, entirely under human
to describe the desired invention ahead of time. Who would
power. More recently, a group of electric utilities estab-
have thought of establishing a prize for the Post-it note?
lished a $30 million competition for energy-efficient refrig-
erators, which was won by Whirlpool with a line of refrig-
FINDING VACCINES FOR MALARIA,
erators that operated 70 percent more efficiently than federal
TUBERCULOSIS, AND HIV
requirements. Prizes are an attractive way of encouraging
the prospect of patents is not stimulating research
research, because unlike government-funded programs,
commensurate with the social and economic costs of malar-
ia, tuberculosis, and hiv/aids. And
government-funded research to
develop vaccines for those diseases
has so far been unsuccessful. Alter-
Malaria, tuberculosis, and HIV have in the past
native mechanisms (e.g., prizes)
50 years killed several times as many people as all
could work because criteria for a
useful vaccine can, in large part, be
wars. Together, they kill five million people a year,
specified in advance, and there are
institutions, such as the Food and
mostly in developing countries.
Drug Administration (fda), which
are charged with determining
whether vaccines and pharmaceuti-
cals are safe and effective.
they provide strong incentives. Researchers get paid only if
their work succeeds.
The Threat Malaria, tuberculosis, and hiv are the world’s
During the first half of the nineteenth century, when
most deadly communicable diseases. In the past 50 years,
both patents and prizes were used to encourage invention,
those diseases have killed several times as many people as
there was an intriguing case in which a government com-
all wars. Together, they kill five million people a year,
bined the patent and prize systems by buying out a patent.
mostly in developing countries.
In 1837, Louis Jacques Mande Daguerre invented pho-
Malaria kills 1.1 people annually and is particularly
tography. He exhibited images created using his
likely to kill children and pregnant women. Resistance is
Daguerreotype process and offered to sell detailed instruc-
spreading to the major drugs used to treat malaria and to
tions to a single buyer for 200,000 francs or to 100 to 400
provide short-term protection for travelers.
subscribers at 1,000 francs each. Daguerre was not able to
Tuberculosis kills 1.9 million people a year. Although
find a buyer, and the potential of his invention was going
most cases of tuberculosis now occur in developing coun-
unrealized. François Arago, the permanent secretary of the
tries, drug-resistant strains are spreading rapidly, posing a
French Académie des Sciences, argued that it was “indis-
threat to developed countries.
pensable that the government should compensate M.
In 1998, about 2.3 million people died of aids, and 5.8
Daguerre direct, and that France should then nobly give to
million people were newly infected, 70 percent of them in
the whole world this discovery which could contribute so
sub-Saharan Africa.
much to the progress of art and science” (cited by Kenneth
Nelson in “A Thumbnail Sketch of Daguerrotypes”). In
The Promise of Vaccines Vaccines have proved effective
July 1839, the French government purchased the patent
against many other infectious diseases, and in the long
from Daguerre and put the rights in the public domain
run, they are likely to be the most effective and sustainable
(except in England, where the French government allowed
way to fight malaria, tuberculosis, and hiv. The potential
Daguerre’s original patent to remain in force). The inven-
of vaccines is illustrated most vividly by the eradication of
tion was rapidly adopted and improved. Within months,
smallpox in the 1970s. A standard package of cheap, off-
Daguerre’s instruction manual was translated into a dozen
patent vaccines reaches three-quarters of the world’s chil-
languages. Many complementary inventions improved
dren and is estimated to save 3 million lives a year.
the chemistry and lenses used in Daguerre’s process.
It is an open question whether vaccines can be developed
R e g u l a t i o n
36
Vo l u m e 23, N o . 2
Kremer.2 6/8/00 3:56 PM Page 37
against malaria, hiv/aids, and adult tuberculosis, but there
that, even at a cost of $40 per immunized person, vaccines
is reason to be optimistic. Recent research on animals looks
against malaria and hiv would be cost-effective in poor
promising and advances in immunology, biochemistry,
countries. But because private firms would be lucky to
and cloning have given scientists new tools with which to
receive even a tenth of that amount in those countries, they
develop and test vaccines. Genetic sequencing of the organ-
are—unsurprisingly—not rushing to develop vaccines.
isms causing malaria, tuberculosis, and hiv is complete
or far advanced.
Government-Funded Research Not the Answer Given the
obstacles to private research into vaccines for malaria,
Obstacles to Private-Sector Research There is little private-
tuberculosis, and hiv, one option would be for the gov-
sector research on vaccines for malaria, tuberculosis, and
ernment to finance research directly. This makes sense for
African strains of hiv. Most of the people who suffer from
basic research. However, direct government financing is
those diseases are poor and cannot afford to spend much
ill-suited to the subsequent development of useful prod-
on vaccines. Most applied aids research is on treatments,
ucts—a task that is much better left to the private sector.
which tend to be more lucrative than vaccines. (aids
In The Malaria Capers, Robert S. Desowitz chronicles the
activists also tend to focus more on lobbying for treat-
sad story of the efforts of the U.S. Agency for Interna-
ments than for vaccines.) The little aids vaccine research
tional Development (usaid) to promote the development
that is conducted is overwhelmingly oriented towards the
of a malaria vaccine. usaid decided in the 1980s to finance
strains of the disease prevalent in
rich countries, not the strains
prevalent in Africa, where most
people are dying.
There is reason to be optimistic about the develop-
Private research on vaccines for
malaria, tuberculosis, and African
ment of vaccines against malaria, HIV/AIDS, and
strains of hiv is limited not only by
the poverty of potential customers,
adult tuberculosis because of recent research and
but also by the limited ability of pri-
vate developers to reap the benefits
advances in immunology, biotechnology, and cloning.
that those vaccines would produce.
Individuals who take vaccines not
only benefit themselves, but also
help break the chain of disease transmission, thus benefiting
three teams seeking a malaria vaccine. One team developed
the rest of the population. Customers would not pay for this
a candidate vaccine, but only two of nine volunteers test-
side benefit of vaccines. In addition, the beneficiaries of vac-
ed were protected from malaria, and the tests indicated that
cines are often children. Keeping children disease free
the vaccine created side effects. Those results, mixed at best,
would allow them eventually to earn enough to compen-
did not prevent usaid from issuing wildly optimistic
sate vaccine developers, but of course, there is no way chil-
statements. In 1984, the agency claimed that there had
dren can sign a contract to compensate vaccine develop-
been a “major breakthrough in the development of a vac-
ers when they become adults. Furthermore, some potential
cine against the most deadly form of malaria in human
customers may be unwilling to pay much for vaccination
beings. The vaccine should be ready for use around the
because they are unaware of its benefits, a problem that is
world, especially in developing countries, within five
particularly acute in poor countries, where many potential
years” (p. 255). Fifteen years later, the world is still waiting
customers are illiterate and may not trust health officials.
for a malaria vaccine.
Moreover, governments often use their power as buy-
Early work by a second team yielded disappointing
ers, their regulatory power, and their power over intellec-
results, but, not surprisingly, the principal investigator
tual property rights to keep vaccine prices low. Many gov-
argued that his approach was still worth pursing and
ernments drive down the price of vaccines and other
requested an additional $2.38 million from usaid. The
pharmaceutical products by limiting intellectual property
expert consultants assigned to review the project recom-
rights and by producing or importing cheap generic versions
mended against funding the research, but James Erick-
of drugs and vaccines. (This practice is widespread in poor
son, usaid’s malaria vaccine project director, told the
countries, but even such countries as Japan, Switzerland, and
usaid Office of Procurement that the expert panel “had
Sweden were not awarding patents on pharmaceuticals as
endorsed the scientific methodology and the exception-
recently as the 1970s.) That strategy avoids the high prices
al qualifications and experience of the researchers” (p.
associated with patents but also discourages r&d.
258). Once the grant was awarded, the principal investi-
As a result of such policies, vaccines used in developing
gator transferred grant funds to his personal account. He
countries typically sell for pennies a dose. Those vaccines
was later indicted for theft.
that cost as much as a dollar or two a dose do not reach most
Although outside evaluations of the third team’s pro-
people in the poorest countries. Rough calculations suggest
posal called it mediocre and unrealistic, Erickson arranged
R e g u l a t i o n
37
Vo l u m e 23, N o . 2
Kremer.2 6/8/00 3:56 PM Page 38
full funding for the project. The principal investigator and
A patent buyout would allow firms to compete freely to
his administrative assistant later were indicted for theft and
manufacture a vaccine, but given the technical complexity
criminal conspiracy for diverting money from the grant to
of manufacturing vaccines and the arduous process of
their personal accounts. Two months before the principal
securing regulatory approval, competition might not be
investigator’s arrest, the Rockefeller Foundation had given
intense even if patents were put in the public domain.
him a $750,000 research grant, and on the day the investi-
gator was arrested, usaid announced it was giving him an
The Promise of Purchase Commitments Alternatively, the
additional $1.65 million for research.
government (or a private foundation) could make an
By 1986, usaid had spent more than $60 million on its
advance commitment to purchase a certain quantity of a
malaria vaccine efforts, with little to show for it. Never-
vaccine at a certain price, if it were invented. The com-
theless, because usaid believed that there would be many
mitment could take the form of a contractual and bind-
candidate malaria vaccines suitable for testing, it tried to
ing agreement to buy from a prospective vaccine devel-
obtain monkeys as test subjects for those vaccines. Erick-
oper any new vaccine that meets specified criteria (e.g.,
son arranged for a contract to acquire monkeys to go to an
it must be fda-approved and effective at least 80 percent
of the time). The sponsor could
then make the vaccine available to
developing countries in exchange
for modest co-payments.
An advance commitment to purchase a certain quan-
tity of a vaccine at a certain price, if it were invented,
Incentives Unlike direct govern-
ment financing of research, a pur-
could take the form of a contractual agreement to buy
chase commitment allows the pri-
vate sector to decide which
any new vaccine that meets specified criteria.
projects to pursue; that is, research
priorities are not centrally planned
by government, but independently
decided by private firms reacting
associate who paid him a kickback. Erickson eventually
to the market incentive offered by the purchase commit-
pleaded guilty to accepting an illegal gratuity, filing false tax
ment. Pharmaceutical firms and scientists will take the
returns, and making false statements.
risk of investing their money and time only if they
usaid had arranged for independent oversight of the
believe the scientific prospects are worth pursuing. A
project by the American Institute of Biological Science
firm that thinks a vaccine is impossible to produce, given
(aibs). Erickson and the aibs-assigned project manager
current scientific knowledge, will not invest its money in
were lovers.
research and, thus, no government funds will be wasted
Although the usaid project is an extreme example of
on a futile effort. Moreover, a purchase commitment
waste, fraud, and abuse, it illustrates some important points
gives researchers a strong financial incentive to focus on
about government-funded research: First, recipients of gov-
developing a marketable vaccine. Researchers are unlike-
ernment funding have incentives to be overly optimistic.
ly to be distracted by such other pursuits as publishing
Second, government project directors have incentives (aside
academic articles; they will be paid only for producing a
from embezzlement opportunities) to fund unpromising
viable vaccine.
research. Third, because the recipients of government sub-
By agreeing to purchase a large quantity of a vaccine, the
sidies are paid before delivering a product, they may be tempt-
sponsor can purchase it at a reasonable price, while pro-
ed to divert resources away from the search for a vaccine.
viding a sufficient incentive for its development. Because the
cost of r&d represents a large fraction of the cost of pro-
Alternatives to Government R&D There are alternatives to
ducing a vaccine, while the cost of manufacturing addi-
government-directed r&d:
tional doses is usually modest, the total size of the market
(not the price of a dose) will be most important in attract-
• Awarding prizes, such as the Kremer prize to
ing potential developers. Thus, any commitment should
the developer of the Gossamer Condor
specify the total value of the vaccine that will be purchased
•
(i.e., price and quantity). To encourage development of a vac-
Buying out patents in exchange for lump-sum
cine requiring as few doses as possible, the promised price
payment, as in the case of the Daguerreotype
should be set per immunized person, not per dose.
• Committing to purchase a certain quantity of
A purchase commitment should cover as many coun-
vaccine at a certain price.
tries as possible, in order to reduce the price per immunized
person given the total promised market size. Participating
A prize might encourage research, but it would not
countries would provide co-payments, scaled to their
ensure the accessibility of a vaccine once it was developed.
respective per capita incomes. The co-payments would
R e g u l a t i o n
38
Vo l u m e 23, N o . 2
Kremer.2 6/8/00 3:56 PM Page 39
help to ensure the participating countries’ commitment to
could spend 5 percent of its assets annually on grants to
the effort and their satisfaction with a vaccine’s effectiveness,
expand the use of existing vaccines and to fund vaccine
given local needs and conditions
research, while using some of its principal to back a pledge
to purchase and distribute effective new vaccines, if and
Costs and Benefits A rule of thumb in the pharmaceutical
when they are developed.
industry is that an annual market of $250 million dollars is
needed to attract strong interest from potential develop-
CONCLUSION
ers. If a commitment of that size were to result in the
the united states currently supports r&d through
development of an effective vaccine, it would yield a high
the granting of patents and government-funded research.
payoff. For example, in 10 years, more than 400 million
It is time to consider supplementing these mechanisms. In
people could be vaccinated against malaria at a per capita
particular, programs to help create markets for malaria,
cost of $7.50. At least 216 million years of life could be
tuberculosis, and aids vaccines could harness the resources
saved at a cost of roughly $14 per year per life saved.
and expertise of the private sector in the fight against the
The benefits of a successful program would probably
world’s worst infectious diseases while avoiding the ineffi-
continue even after the expiration of the purchase com-
ciencies associated with many government programs.
mitment. It is likely that competing vaccines would become
Commitments to buy large quantities of vaccines could
available, driving prices down and making purchases more
lead to the development and delivery of effective vaccines at
affordable for developing countries and donors.
low cost, saving millions of lives. Taxpayers would pay noth-
ing unless and until those vaccines have been developed.
Potential Sponsors Recently, two institutions that tradi-
tionally have taken a centralized, statist, approach to r&d
have begun exploring market-oriented approaches.
World Bank president James Wolfensohn said recently
that his institution plans to create a $1 billion loan fund to
help countries purchase specified vaccines if and when
they are developed (Financial Times, February 2, 2000). It is
not clear whether the Wolfensohn proposal will pass
through the organization’s internal bureaucracy and win
board approval.
The U.S. government, which sponsored the ill-fated
usaid effort to find a malaria vaccine, is now considering
a more market-oriented approach. Private firms, rather than
government bureaucracies, would make research decisions,
knowing that they would be paid only if they develop effec-
tive vaccines. Specifically, the Clinton administration’s bud-
get proposal would match every dollar of qualifying vaccine
r e a d i n g s
sales with a dollar of tax credit, effectively doubling the
•Robert S. Desowitz, The Malaria Capers: Tales of Parasites and
incentive to develop vaccines for neglected diseases. A qual-
People. New York: W. W. Norton, 1991.
ifying vaccine would have to attack an infectious disease that
•Financial Times, “Discovering Medicines for the Poor,”
kills at least one million people a year and would have to be
February 2, 2000, p. 7.
approved by the fda. To qualify for the tax credit, sales
•Karen Hsu. “Making Drugs, Profits, and Doing Some Good.”
would have to be made to approved nonprofit organiza-
The Boston Globe, October 25, 1999, p. F1.
tions or international institutions. The program’s matching
feature could encourage the funding of vaccine purchases by
•Michael Kremer. “Creating Markets for New Vaccines:
Rationale.” 2000 (available at www.nber.org/books/innovation/
nonprofit organizations, international institutions, and the
vaccine1.pdf ).
governments of developing countries. The cost of the pro-
•Michael Kremer. “Creating Markets for New Vaccines: Design
gram would be capped at $1 billion and it would run from
Issues.” 2000 (available at www.nber.org/books/innovation/
2002 through 2010, but it could be extended for 10 years if
vaccine2.pdf ).
no vaccine has been developed in that time.
•Kenneth E. Nelson. “A Thumbnail Sketch of Daguerreo-
Private foundations could also play a major role in cre-
types.” The Daguerrian Society, 1996 (available at
ating markets for new vaccines. Because foundations have
http://java.austinc.edu/dag/resources/history).
more continuity of leadership, they can more easily make
•New York Times. “New Vaccines for the Poor.” March 14, 2000,
credible commitments to purchase new vaccines. (The
p. A22.
Gates Foundation, for example, has $22 billion in assets; one
•Jeffrey Sachs. “Sachs on Development: Helping the World’s
of its main priorities is to provide vaccines for developing
Poorest.” The Economist 352, no. 8132 (1999): 17.
countries.) U.S. law requires private foundations to spend
at least 5 percent of their assets annually. A U.S. foundation
R e g u l a t i o n
39
Vo l u m e 23, N o . 2
