48 



tural research as an investment creating a capital stock of R&D 

 that has a variety of impacts. 



Second, we have primarily focused on impacts on agricultural 

 productivity. There are other impacts that do exist but we did not 

 primarily focus on them. When we talk about agricultural produc- 

 tivity, we mean the average product of resources under the control 

 of farmers. We provide evidence that for over 100 years in the 

 United States there has been statistically significant impacts of 

 both public and private R&D on agricultural productivity. 



We and other studies have shown a relatively high social rate of 

 return for these investments, on the order of a 40 percent real an- 

 nual rate. We did, however, disaggregate and look more specifically 

 at the livestock sector and the crop sector as well as the aggregate. 

 We also split agricultural research into what we call a 

 pretechnology science and applied science components. Our view of 

 the organization of science is that general science is at the top, 

 pretechnology science is intermediate, and applied science is closer 

 to the real products that farmers and other people might use. 



We found a much higher social rate of return to the 

 pretechnology science research than to the applied science re- 

 search. We also found a much higher rate of return to crop re- 

 search than to livestock research. 



Why might some of these differences occur? We believe that the 

 pretechnology science research, which is more basic science in na- 

 ture, is one that is quite catalytic to bringing real advances in gen- 

 eral science to the applied area and ultimately into useful tech- 

 nologies. If that catalytic role is missing, then applied science even- 

 tually becomes unproductive. This fits in with the problems that 

 the public sector had in taking advantage of the new biotechnology 

 and molecular biology in the 1980's. 



Now, why would plant and livestock research be so different? We 

 argue in our book that the biology of plants and animals are really 

 very different. On average, plants are much more affected by the 

 local geoclimatic conditions than livestock are. For livestock re- 

 search, this means that research benefits frequently spillover quite 

 wide areas, across State boundaries, almost nationally. In the case 

 of crop research, we believe that much of it has very narrow im- 

 pacts, but it has importance for determing the stock of relevant re- 

 search. 



In some livestock research areas like broilers it seems that the 

 stock, when you accumulate expenditures across States, is a na- 

 tional stock. For many plant commodities it is much smaller be- 

 cause you have to divide by a larger number of regions to obtain 

 the relevant stock. We also find that there seems to be stronger 

 science linkages in crops research than in the case of livestock re- 

 search. 



What about the management of agricultural research and the 

 priority setting in the public sector? R&D is a very productive ac- 

 tivity, we believe, using specialized labor and other resources and 

 the facilities and breeding stock are complementary to that process. 

 We do believe that micromanagement of agricultural research is 

 not a particularly productive activity, but there is a role for high- 

 level decisions to be made on priority setting, some at the national 

 level and some at the State level. 



