Investigating Rice Consumers' Preferences for Payment for Ecosystem Services of Sefidrood River

Among the various available tools in the field of natural resources and environmental management, the payment for ecosystem services (PES) is one of the market-based methods that is considered worldwide to protect the environment and ecosystem. PES is an important method for effective management of natural resources and public goods and one of the tools for managing degraded ecosystems and related environmental and economic services. Considering that Sefidrood is considered as the most important and valuable source of agricultural water supply and aquatic environment in Guilan province, and also the water quality of this important river is in a bad and very bad condition, this study was conducted using PES economic tools through payments by rice consumers in Guilan province to rice farmers and thus encouraging them to take environmentally friendly measures (organic agriculture) to reduce pollution of the Sefidrood River. 

This research was conducted using a choice experiment method. In our CE, each PES alternative is described by a set of attributes that include distribution of payments, contract duration, implementing organization, monitoring times, possibility to cancel and payments. First, to investigate the effect of different attributes of PES scheme on rice consumers' willingness to pay and their marginal utility, a conditional logit model was used to compare the results of random parameter logit model and latent class models with a base model. Then, the RPL and LC model was used to further investigate the invisible heterogeneity that exists in the behavior of respondents. The RPL model is an advanced model that allows attributes coefficients to change randomly among respondents. Therefore, instead of estimating a fixed coefficient for each attribute, two coefficients are estimated, which together describe the distribution of heterogeneous preferences of the respondents for this attribute.

To confirm the CL model, the independence of irrelevant alternatives assumption was performed using the Hausman-McFadden test. Given that the value of chi-square statistics has become large and significant, therefore, the CL model is not suitable for investigating the effect of attributes on consumer’s willingness to pay, and more advanced models should be used. For this reason, RPL and LC models are estimated. According to the results of the RPL model, the highest willingness to pay is related to the monitoring times therefor indicating that consumers are willing to pay 1347 Tomans for more monitoring. The amount of willingness to pay for the duration of contract and distribution of payments is equal to 1326 and 914 Tomans, respectively, which indicates if the contracts are short-time and also more payments are made to low-income rice farmers, the willingness to pay will increase to 1326 and 914 Tomans, respectively. Based on the results of the LC model, in the first class, except for the contract duration, all other attributes were not statistically significant. In the second class, the distribution of payments, the contract duration and the monitoring times with a positive sign and the implementing organization with a negative sign are significant. Class membership coefficients for organic rice consumers indicate that the likelihood of being in second class depends significantly on the respondents' age, gender, and level of education.

The results of RPL and LC models confirm the existence of heterogeneity in the preferences of organic rice consumers. Therefore, appropriate methods can be used to differentiate organic products and thus improve the utility of consuming these products. Consumers were also more inclined to have a short-time and high monitoring scheme, this result is not unexpected due to the novelty of the scheme. Therefore, it is recommended to start short-time schemes with high monitoring. Consumers also tended to make more payments to low-income rice farmers, so it is recommended that lower-income rice farmers be given priority in implementing the PES scheme. The results of both model showed that the distribution of payments and monitoring times had the highest priority for consumers in choosing the PES scheme, respectively. Therefore, in order to increase the participation of consumers in such schemes, it is recommended to include these attributes in the schemes. Also, although PES is not designed as a tool to reduce poverty, it can increase the incomes of low-income rice farmers and help their livelihoods. Given that such schemes have not yet been implemented in Iran, it is suggested that in order to increase consumer participation, various levels of attributes should be provided to the respondents.