The Economic Value of Ecosystem Conservation in Japan: Reduction of elicitation effect by Bid Effect Function Mitsuyasu YABE Kyushu University
Contents • Background and Purpose • Analytical Model • Survey Design and Explanatory Variables • Estimation Results • Conclusion
Background of the Study National Park Aso • Over 18 million people visit and enjoy the view of Aso grassland. • Many valuable flora and fauna were maintained by traditional human activities. • With decline of farmer and the change of farming pattern, the Aso grassland verge to crisis of maintaining
Endangered Species in Aso Grassland Photo by Miura
Purpose of the Study • Estimating the conservation value of Aso Glass Land • Improving the Contingent Valuation Method (CVM): Deceasing Elicitation Effects 1) Starting Point Effects 2) Yea-saying
Elicitation Effects in CV • Dichotomous choice CV is most commonly applied since respondents only need to select “accept” or “not accept” • However, even when the bid is higher than the latent willingness to pay, the respondents tend to “accept” the bid.
Characteristics of the Model • To decrease Elicitation Effects on the WTP →Introducing Bid Effect Function →Applying DC Approach for Five choice
Formulation of Bid Effects • Latent WTP for the irespondent: • The gap between the bid ti and latent WTP: • Bid effect function:
Stated WTP • the stated WTP can be expressed: • The probability that the stated WTP yiis larger than the bid ti:
The Hypothetical Question（1） • “suppose that grassland could be converted to forest and grassland could be lost as grazing and open burning are discontinued in Aso region. • In order to prevent that happens, we set up the “Aso Grassland World Heritage Fund” to register and conserve the grassland • The activity cost of “Aso Grassland World Heritage Fund” is supported by the public contribution.
The hypothetical question（2） • If the fund costs (***) per household per annum, you may pay the amount of money? (Select only one) 1. will pay> “YES” in Model 1 & 2 2. probably will pay > “YES” in Model 3 & 4 3. probably will not pay 4. will not pay 5. don’t know
The Log-likelihood Function in DC-CVM Where di1 and di2 are coded 1 when respondent chose the option and otherwise 0.
Form of Bid Effect Function • If bid effect function is liner function : • We have a relation as follows:
Bid Effect function based on logistic function • Hypotheses: • >>
Summary of Survey Questionnaire • Survey Period： December 1998 • Respondents：Residents of Kumamoto Prefecture • Samples：1000 • Samples used for Analysis：418
Attributes of Survey Respondents • Average Age： 59 years old • Average Income： 5,740,000 yen per ann. (Approx. 52,000 US$) • Conservation Activities of Aso Glass land: Highly Appreciated
Results of Bid Effect Function • Bid effect coefficient was statistically significant • The error term was reduced by more than 70% • Difference between Mean and Median was also reduced
Conservation Value of Aso Grassland • Estimated Value who definitely pay： Mean 1,028 Yen ( = US$ 9.3) • Return rate of this survey:41.8% • Number of households of Kumamoto prefecture:594,197 • Total Conservation Value per Year = 1,028 x 0.418 x 594197 = 255 Million Yen/Year = US$ 2.3 Million/Year
Conclusion • Removal of influence from the bid effect bias enabled a more appropriate WTP estimation • Price Oriented Attribute affected the WTP more than income • The estimated total environmental value was more than the amount of the environmental gross investment at HTB