much bigger in developing than in developed countries? Such explanations may provide useful
insights into how underinvestment gaps could be reduced.
The model to estimate the underinvestment gap assumes full information and economic
rationality. In reality, however, information by which R&D projects are selected is usually incomplete
and of a rather uncertain nature, while the selection itself may adhere to economic rationality only
partially. Both factors make that governments are unable to pursue the full economic potential of
agricultural R&D in an effective way.
Due to lack of information, governments have only an incomplete and vague idea of the
economic potential of (agricultural) R&D investment. Even ex post, we have to content ourselves
with rather imprecise estimations of the economic impact of past R&D investments. In that sense we
are very much groping in the dark, not only ex ante but also ex post.
Some of the explanations of the underinvestment gap in agricultural R&D suggested in the
literature are based on the notion that the data used in the rate-of-return calculations is systematically
biased. Oehmke (1986), for example, argued that rigidities in the budget process prevent an
immediate response to new agricultural R&D investment opportunities. He captured this in a model
by basing R&D investment decisions on prices and quantities of period T-1 rather than period T. On
the assumption that both variables increase over time, the R&D benefits across all projects will be
underestimated ex ante. This results in a persistent underestimation of the position of the R&D
opportunity curve causing underinvestment in agricultural R&D.
Another argument put forward by Fox (1985) is that the costs of public agricultural R&D are
systematically underestimated. His argument is based on a widely accepted notion in the public-
finance literature, namely that the social opportunity cost of a dollar of government spending is larger
than a dollar. There are direct costs to collecting taxes, but what is more important is that taxes
introduce distortions in factor and product markets that create deadweight economic losses. Ballard,
Shoven, and Whalley (1985), for example, estimated the marginal welfare costs or excess burden for
the US in the range of $0.17 to $0.56 per US dollar tax income. Fox (1985) assumed an excess burden
of $0.30 for every US dollar of tax income, while the US Federal Government advises since 1992 to
use $0.25 in rate-of-return calculations of Federal Government projects (Office of Management and
21