| Farmland preservation policies and programs have been in place at the national, state, and local levels since the 1970s in the United States. Features in landowners’ stated preferences for these programs are critical in evaluating the effectiveness and efficiency in designing and implementing these policy programs.;Quite a few behavioral studies have already been done in this field. However, most of these studies were associated with an implicit assumption—Independent and Identical Distribution (IID) in setting up choice models. This assumption not only eliminates the variation in the personal preference for a farmland preservation program among individuals, but also denies the interdependence among an individual landowner’s preferences for several preservation programs. Technically, this assumption introduces bias and inefficiency into modeling. Without convincible justification, the involvement of IID could lead to biased model estimates and distort the interpretation. Unfortunately, such justification is barely mentioned and/or addressed at all in these studies.;People are different, so do their preferences. Moreover, personal preferences for several preservation programs are simultaneously constrained by one’s social-economic context. With such acknowledgements, both the variation in the personal preference for a farmland program and the interdependences among preferences for several programs are particular of the interest in this research. A modeling framework centering on a system of heteroscedastic probit models is calibrated to achieve unbiasedness and efficiency in implementing and evaluating preservation programs.;A dataset from the state-level urban-fringe landowners’ survey in 2001 and 2002 collected by the research team in American Farmland Trust (AFT) is used in this research. The survey covers fifty metropolitan counties in five states, i.e., Wisconsin, Michigan, New York, Texas, and California. 1617 landowners were contacted either by phone calls or by mail. Farmland preservation programs concerned in the survey includes the agricultural zoning, the Purchase of Development Rights, the Purchase of Conservation Easements in California, and the use-value taxation programs.;The personal preference for a preservation program is expected to be influenced by owners’ social-economic circumstances, farmland characteristics, objectives of land ownership, and local social-economic conditions. Significant findings of related factors in shaping a landowner’s preference for a preservation program from literature are referred for its generalization meanings.;The result shows that the modeling framework based on a system of heteroscedastic probit models succeeds in explicitly capturing variations in and interdependence among personal preferences for several farmland preservation programs. The modeling framework can be generalized regardless of the geographic settings, i.e., state-level and/or county-level. The significance and magnitude of an individual factor effecting on one’s preference for a preservation program, however, cannot be generalized across geographic boundaries, i.e., they are more of localized. Meanwhile, findings of the effect of individual factors are not precisely consistent with those from other studies. Suggestions for effective and efficient policy implementation are derived from findings of effects of individual factors on personal preferences for preservation programs.;The dissertation consists of five chapters. The first two briefly describe the history of farmland preservation programs and the necessity of policy evaluation in the United States. Chapter three mainly puts efforts in describing and comparing features of available discrete choice models in dealing with complicated individual behaviors, and then a desired modeling framework is carefully described in details. Chapter four reports the model calibration process and model estimates with empirical data. Propositions to improve the effectiveness and efficiency of the policy implementation are also delivered here. The final chapter makes conclusion about the generalization of findings and gives hints for the future study. |
