| The development of Genetically Modified (GM) crop varieties has arguably been the most successful application of agricultural biotechnology research to date. However, the biotechnology is a two-edged sword. Behind the great success, there are also concerns about the negative impact of GM crops. One of the major concerns is the buildup of resistance in the pest population. While refuge policies have been implemented in many countries to manage the development of resistance, most empirical analyses to date have focused only on the United States case. There is very little empirical work that has focused on other countries, especially developing countries.; The overall goal of this dissertation is to analyze, theoretical and empirically, the optimal refuge policy to manage the buildup of the resistance in a developing county. To narrow the scope of the research, I use Bacillus thuringiensis (Bt) cotton in China as a case study. The dissertation consists of three essays.; The first essay creates a framework for analyzing the optimal refuge policy to manage the buildup of the resistance in the pest population. Specifically, I developed and estimated a single resistance bioeconomic model to analyze the optimal control path. The results show that planting non-Bt cotton as refuge is not economic in the short run. In addition, even if planting a refuge might be economic in the long run, it pays to delay the refuge until the buildup of the resistance becomes a real concern.; In the second essay, I extend the single resistance bioeconomic model developed in the first essay into a double resistance model and empirically analyze the optimal refuge policies in northern China. In this model, we assume that the pest population can develop resistance to both Bt cotton and conventional pesticide. To mimic the real cropping system in the cotton production region in northern China, the impact of the natural refuge crops (other host crops of the pest than cotton that are planted adjacent to cotton) on the development of the resistance is considered in the model. The most important finding of my second essay is that there appears to be no need for a policy-mandated refuge policy in China.; In the third essay of my dissertation, I extend the static bioeconomic model of the second essay into a dynamic one. In this essay, I show that even though a dynamic refuge policy provides a smaller production cost than a static policy, the cost saving is not economically significant. In other words, after considering the transaction costs of the refuge policy, planting non-Bt cotton as refuge is probably not economic. Another interesting finding concerns the recovery of the susceptibility to conventional pesticide in a dynamic setting. We find that if 100% Bt cotton is planted without conventional pesticide sprayed, efficiency of the conventional pesticide will recover. Consequently, farmers can use Bt cotton and conventional pesticide, alternatively, to control the pest problem, rather than planting non-Bt cotton annually. |
