2002 年 20 巻 4 号 p. 243-252
The purpose of this article is to alleviate omitted variable problem in the hedonic price function, which causes estimator to be biased and inconsistent. Therefore we applied the fixed effect model using '85 and '95 panel data which allows for time-constant unobserved effects. Moreover we improved the form of independent variable and examined its validity by the bootstrap J test. At the end of this article, we constructed a confidence interval of the value of the paddy fields' externality by the bootstrap method. The estimated value is lower than in any previous studies. This shows that unobserved effects significantly affect evaluated value.