This study presents entirely new econometric evidence on the links between public investments in agricultural R&D and agricultural productivity using state-specific data on U.S. agricultural productivity, and federal and state government investments in agricultural research and extension. Particular attention is paid to the specification of the research lag structure and interstate spillovers and the implications for the estimated returns to research. Using newly constructed sets of panel data, some of which span the twentieth century, the results indicate that the lags linking R&D spending to productivity growth are substantially longer than commonly found or assumed in the prior published work.

The spillover effects of R&D among states are also especially important, such that the national social net benefits form a statea (TM)s agricultural research investments are much greater than the statea (TM)s own net benefits. Specific results are sensitive to unavoidable model specification choices, but the main findings are consistent across all of the models, indicating that the benefits from past public investments in agricultural research have been worth many times more than the costs, that a significant share of the benefits accrue as spillovers, and that the research lags are very long. The evidence is also indicative of a substantial misallocation of public funds to U.S. agricultural R&D. Clearly the total investment has been too little, and the results also suggest that a reallocation of funding among states might raise the national return to investments in agricultural R&D.