基于陈述性偏好法的三江平原湿地生态保护价值比较

Comparing the value of ecological protection in Sanjiang Plain wetland,Northeast China based on the stated preference method

陈述性偏好法是当前用于评估环境物品非市场价值的主要方法,包括条件价值评估法(CVM)和选择实验(CE).本文利用陈述性偏好法对三江平原湿地的生态保护价值进行评估,基于随机效用理论构建陈述性偏好方法的支付意愿评价模型,求出基于CVM与CE的三江平原湿地生态保护价值的支付意愿(WTP).结果表明:基于CE的平均支付意愿为每年379元,不同选择属性的边际支付意愿分别是水源涵养为每年114.00元,湿地面积为每年72.55元,自然景观为每年59.55元,生物多样性为每年37.09元;CVM的平均支付意愿为每年134元.分析支付意愿的影响因素,并对受访者拒绝支付原因进行探讨.CE方法得到受访者的支付意愿显著高于CVM方法,社会经济属性中受教育程度和个人平均年收入与受访者的支付意愿呈显著正相关,而拒绝支付原因没有显著差异.CE可以进行多属性、多水平分析,求出湿地各选择属性的支付意愿,因此CE比CVM更能揭示受访者的偏好信息,其评估结果更接近于实际.

Stated preference method is usually used to evaluate the non-market value of environ- mental goods which includes contingent valuation method （CVM） and choice experiments （CE）. In this paper, stated preference method was adopted to evaluate the non-market value of Sanjiang Plain wetland. A willingness to pay （WTP） evaluation model of stated preference method was constructed based on the random utility theory. The average WTP of CVM and CE was obtained, respectively. The average WTP elicited by CE was 379 yuan per year, and the marginal WTPs of different selec- tion properties including water conservation, wetland area, natural landscape and biodiversity were 114.00, 72.55, 59.55 and 37.09 yuan per year, respectively. Meanwhile, the average WTP elicited by CVM was 134 yuan per year. The influence of factors on WTP was analyzed and reasons for pro- test responses were discussed. Results showed that the respondents＇ WTP elicited by CE was signi- ficantly higher than that by CVM, and respondents＇ socio-economic attitudes such as level of education and personal annual income had a significant positive impact on respondents＇ WTP. There were no significant difference in the reasons of protest responses between CVM and CE. Besides, re- spondents＇ multiple attributes and multiple levels analysis could be carried out by CE and the WTP of wetland＇ s selection attributes could be calculated. Therefore, CE had the better ability of revealing respondents＇ preference information than CVM and its assessment results were more close to the actual value.