The development of a variety of stated preference methods has been an important agenda in the field of environmental valuation during the last decade. It leads to the wide use of those methods for costbenefit analyses. In this article, choice modeling is examined as to whether or not it is reliable for the evaluation of agri-environmental policy measures such as the direct payment and rice terrace partnership program. A choice modeling approach has an advantage in evaluating the multi-attribute environmental goods compared with contingent valuation method. Rice terrace has two different kinds of environmental impacts, i. e., environmental benefits (multifunctionality) and damage. Both environmentally positive and negative impacts should be evaluated simultaneously. Otherwise, the result is considered to be biased. The choice modeling can accommodate a number of attributes. Therefore, it is quite a useful tool to assess the environmental value of multi-attribute goods. Environmental benefits are classified into two categories.
One is the conservation of rural landscapes and ecosystems, and the other is disaster prevention. As for environmental damage, water contamination caused by the excessive application of pesticide and fertilizer is presented.
A mail survey was conducted in July, 2001. Two types of questionnaires were mailed to 400 general households in Kamogawa-city. Samples were randomly chosen from the telephone directory. The difference of those questionnaires appears in an order of attributes and their explanation. Exchanging the attributes between an environmental benefit and damage enables an analysis of an order effect. Their overall response rates were 74.5% and 82.0%, respectively.
This experimental choice technique produced marginal willingness-to-pays for three types of environmental impacts with statistical significance of key variables. Marginal willingness-to-pays for water contamination were at least twice as high as those of two environmental benefits.
An order effect of the questionnaire design was tested using likelihood ratio test and t test. Consequently, an order effect was statistically revealed. This finding suggests the choice modeling is not an ideal tool for environmental valuation. It is also affected by some biases due to respondents inconsistent attitude toward environmental risk.
