Against Intellectual Monopoly

medicine you take.

How Steep Is the Trade-off, Then?
    A relatively recent National Bureau of Economic Research paper, by Hughes,
Moore, and Snyder and sponsored byAventis Pharmaceuticals [now Sanofi-
Aventis], attacks directly the costs and the benefits of drug patents.40 They
conclude that if the appropriate rate of interest for discounting the social
benefits of new drugs is greater than about 5 percent, then the social benefit
of eliminating patents is greater than the cost. Because the social benefits
of pharmaceuticals are risky, and indeed in this study they are assumed to
be perfectly correlated with private risk, an appropriate interest rate is the
rate of return in the pharmaceutical industry. Indeed, the interest rates used
for cost-benefit calculations for government projects are usually around 15
percent, which is the same as the rate of return Hughes, Moore, and Snyder
assume for pharmaceutical R&D. This is substantially in excess of 5 percent
for eliminating patents.
    Because Hughes, Moore, and Snyder are among the few who have
attempted to quantify the costs and benefits of intellectual monopoly, it
is worth reviewing their calculations. They assume that demand for pharmaceuticals is linear. From the perspective of cost-benefit analysis, this
assumes that as output expands past the monopoly level, demand falls off at
a constant rate. If demand falls abruptly, then the loss of consumer surplus
is much smaller than would be estimated by a linear demand function, and
we would get a more favorable case for patents. However, there is some
reason to think that demand for pharmaceuticals depends upon income,
and if this is the case, the linear demand assumption is a reasonable one.
Other parameters of the Hughes, Moore, and Snyder model are calibrated
to the data. They assume that 75 percent of pharmaceutical revenue is generated by drugs still under patent, that market exclusivity lasts nine years,
and that the lifetime of a new drug is twenty-five years. They assume that it
will take generic manufacturers one year to enter after innovation. Also on
the basis of data about competition between generic and nongeneric drugs
after patent expiration, they attribute a first-mover advantage to the innovator by assuming that the innovator will be able to charge the monopoly
price and still serve 20 percent of the market. In fact, evidence from India
suggests that it takes closer to four years for generics to enter; and relatively unbiased sources such as the Congressional Budget Office suggest
that market share after the entry of generics is substantially greater than
20 percent.
    Finally, a critical assumption is the connection between producer surplus
and the number of new drugs discovered. That is, higher expectations of profit due to monopoly lead to more pharmaceutical research, and consequently to more drugs. Notice, however, that this effect can be negative,
as the monopolization of existing drugs may also make it harder to discover new drugs, and we saw that this was empirically important in the
history of the chemical industry. Hughes, Moore, and Snyder assume that
the number of new drugs discovered is proportional to producer surplus.
That is, because they estimate that, without a patent, profits are about
25 percent of what they would be with a patent, they assume that there
will be 25 percent as many drugs discovered without patents. Even without
the problem of innovation chains and the cost of inventing around existing
patents discussed earlier, this assumption is very favorable to the patent
system. The number of discoveries is scarcely likely to drop 75 percent if
profits are reduced by 75 percent. On the basis of survey data from industry
interviews (which, in turn, probably understate the number of drugs that
would be developed without patents), a figure of 40 percent would appear
to be closer to the mark. We should also note that our own estimate is
that, without patents, firms would earn closer to 80 percent of what they
earn with patents, rather than 25 percent.41 Despite these apparent biases,
Hughes, Moore, and Snyder still find that, even with an interest rate as low
as 5 percent, the immediate benefit of wider drug availability exceeds the
long-term cost of having fewer new drugs.

    How steep is the trade-off society faces, then? Not too steep, apparently, if
a 5 percent discount rate is high enough for even a pharmaceutical industry