Testing Long-Run Productivity Models for the Canadian and U.S. Agricultural Sectors

This paper discusses a portion of our work linking data on the agriculture sector in the United States and Canada. The purpose of this work is to explore the evolution of gains in agricultural productivity in the two countries during the post-WWII period. Comparable data has been developed for each country and a series of tests have been applied about the nature of the long-run production sector. These tests are designed to evaluate the alternate possible structures of shifts in the long-run technology over time.There is considerable evidence in both countries that the long-run shifts have been Hicks Neutral in models that use gross, not net, output measures. The reverse is true for the net output models. The use of the conventional net output measures is strongly rejected. However there is evidence, in both countries, in support of the hypothesis that separability of a type that is similar to, but weaker, than real-value added is not rejected.