摘要：Intellectual Property Rights (IPR) regimes for plant breeding are generally justified on the basis that they encourage innovation. Introduction of IPR regimes for plant varieties in the United States has led to increased concentration, but it is less clear whether IPRs have promoted useful innovation, as measured by productivity of available corn hybrids. There are difficulties in finding a satisfactory measure of innovation in plant breeding, and in this paper we propose a procedure. Results from the annual corn hybrid trials conducted by 11 US universities over the 20 years from 1990 to 2009, at 365 separate locations in the 11 states, have been collated. This set of unbalanced panel data for grain corn hybrid trials has been used in a fixed effects model to estimate a production function for corn and the contribution to yield of the genetic characteristics of the corn hybrids. The Hausman Taylor estimator is then used to separate out the contribution of GM traits. Because the data are experimental, the production function can be interpreted as representing the technological frontier. The cross section is made up of the corn hybrids that were submitted for trial over the period. The fixed or unobserved time invariant effects represent the part of production which can be attributed to the characteristics of a particular hybrid. This is taken to be the contribution of the "genetics" of each hybrid to yield, and the maximum fixed or unobserved effect in any one year can be considered to represent the "frontier" of genetic contribution to increased yield.
