Based on the argument that there is no realistic and readily available estimate for the
shape of a supply curve in a given study, the realistic approach is to assume the supply shift is
parallel (Rose 1980 p. 837). Under this assumption, functional form of supply and demand is
insignificant and it is appropriate to use local linear approximation.
As stated in the previous chapter, the dynamics of the shift in the supply curve and the
resulting change in the stock of knowledge are important when measuring the consequences of
research investments. Once research produces results, the response in the supply curve is not the
static snapshot that the static model represents. Alston, Norton, and Pardey addressed lags in
research and adoption by separating them into three categories. The idea is that the stock of
knowledge yields a stream of benefits once it is increased and continues into the future until that
knowledge or technology is obsolete. This happens in three stages. The first stage, the research
lag, is a lag between the initial investment of the research and the results of the research. Then,
the development lag is a lag between the results of the research and the development of the
results into useful technology. And the third lag is called the adoption lag that is a lag in the
generation of technology to its implementation in the real world. They postulate that applied
research has shorter lags and basic research has longer lags.
EMPIRICAL RESULTS
The econometric model in this study consists of supply functions for cotton and peanuts that
were estimated using data derived from pooled time-series cross-sectional data for the four states
of Alabama, Florida, Georgia, and South Carolina. The data were mostly published with the
exception of the research expenditure variable, which was collected and provided by Wallace
Huffman from Iowa State University. The prices, quantities, harvested acres, and cost of
10