Measuring marginal willingness to pay using conjoint analysis and developing benefit transfer functions in various Asian cities

We need a consistent methodology to measure the co-benefits of climate change mitigation across Asian countries. This study chose a strategy of modifying the Japan-specific life-cycle impact assessment method based on endpoint modeling (LIME) for wider application across countries. LIME has two dimensions. First, it is an environmental science that links the cause-and-effect chain. Second, it is an environmental valuation that weighs four endpoint damages in monetary terms through a conjoint analysis that is derived from an Internet-based questionnaire survey. This article describes the modification of the methodology for application of the conjoint analysis to weigh environmental impacts. We approached the investigation as follows. First, we conducted Internet surveys to measure marginal willingness-to-pay (MWTP). We used a sample of 112 respondents in their 20 s to 40 s, divided equally between men and women, in 11 cities across China, India, and Southeast Asia. The results obtained showed clear statistical significance and were comparable across the cities. Second, we attempted to develop functions (called benefit transfer functions) to simplify the measured MWTP in order to apply it across different Asian countries. The functions were derived through a stepwise meta-analytic method, a type of multiple regression analysis whose independent variable was MWTP and dependent variables were attributes of both respondents and surveyed cities. The functions showed that coal consumption and percentage of nature reserve were dependent variables. Then, the MWTPs estimated from the functions were compared with the measured MWTP for transfer error, which is calculated by the absolute value of the difference between the estimated value and the measured value divided by the latter. The transfer error was below 50% in about 90% of the 44 results (a combination of four endpoints and 11 cities), implying that the developed functions were statistically significant.