The purpose of this paper is to propose a nonparametric estimation method of WTP distribution in double-bounded dichotomous choice CVM and to measure the externality of the agricultural production in the Oguraike region of Kyoto Prefecture using this method. I extend the nonparametric method proposed by Kriström in single-bounded dichotomous choice CVM to a double-bounded form and compare the WTP distributions, means and medians estimated by the usual parametric method with those estimated by the nonparametric method.
　As a result, I found that the means estimated by the nonparametric method were much larger than those estimated by the parametric method. This suggests that assuming a parametric WTP distribution has a great risk of underestimating the mean, and the nonparametric method is more effective in estimating the mean. On the other hand, the estimated medians differ only slightly. From this result, I concluded that we obtained a reliable median by the parametric method, and that it was statistically desirable to use the median, which was robust to different estimation methods, as a representative value of WTP distribution.
　Finally, I chose the medians estimated by the nonparametric method as the values. As a result, the values of the disaster prevention and amenity supply services of the agricultural production in the Oguraike region are 4,631 yen and 5,138 yen a year per household, respectively.

View full abstract