Games between stakeholders and the payment for ecological services: evidence from the Wuxijiang River reservoir area in China

Research region

The Wuxijiang River reservoir was chosen as the research region. Its administrative region mainly involves the four townships of Hunan, Jucun, Lingyang, and Huangtankou, in the Qujiang district of Quzhou, as shown in Fig. 1. Its position in China is shown in Fig. 2. The watershed system of Wuxijiang River is shown in Fig. 3.

The Wuxijiang River watershed has reservoirs in Hunan and Huangtankou Townships, a primary tributary in Qujiang district, and a water diversion project in Wuxijiang River. The Wuxijiang River watershed has high ecological value and is a good source of quality fresh water. It not only has high forest coverage and abundant biological diversity, but also has a state-level wetland park. However, the Wuxijiang River watershed is facing ecological and environmental problems that are associated with a relatively fragile ecosystem, impacts and pollution from livestock and poultry breeding, negative environmental impacts from tourism and aquaculture, industrial pollution, and farm run-off.

Data sources

Data were derived from public statistics and data collection through a household survey of 12 villages in four townships and 15 blocks on the Wuxijiang River in December 2015 by the project group. At least 30 households were randomly selected from each village for survey. The project group issued 385 surveys, of which 383 were returned and three were invalid. In total, there were 380 valid surveys. The contents of the survey included information about the economic conditions of farmers and householders, suggestions by the local government for ways to improve the ecological protection of the water source, and historical data. The urban community survey on respondents’ WTP for water received a total of 552 valid surveys. The contents included respondents’ personal information, family water use, their understanding of ecological protection of water sources, WTP for water, and the mode of payment.

Research method

There are many methods to estimate the amount of compensation that should be granted for impacted ecological resources such as water (Liu, Xu & Zhang, 2006; Wunder, Engel & Pagiola, 2008; Wang, Wang & Liu, 2007; Li & Chen, 2003; Liu, Liu & Xu, 2014; Wei & Shen, 2011; Xie, Xi & Huang, 2014; Xiao et al., 2015). In order to effectively focus on the direct protection of water and contributors, the CVM (Ciriacy-Wantrup, 1947; Xu et al., 2003; Zhang & Zhao, 2007; Qiao, Cheng & Liu, 2016) was chosen for this research as a reference basis for the ecological compensation standard.

Analysis of stakeholders of ecological compensation in the water source area of the Wuxijiang River

Generally, the impacted research subjects analyzed are the government, industry, and the residents (Lami, Masetti & Neri, 2016). As the first government protected water source in Zhejiang province, Zhejiang provincial government and local government have implemented nearly 20 years of ecological protection policy. The Wuxijiang River power plant is the only remaining large industry. Most other industries have been closed or moved. To quantify the ecological compensation of the water source, we solve for the amount of gambling of the enterprises, government, and residents are willing to engage in. The Wuxijiang River power plant pays the water resources fees and the reservoir funds. There are no direct relationships between industry, local governments, and water sources. Therefore, the relationship between industry, government and residents is relatively simple. The only conflict of interest between the government and industry which needs to be coordinated is that of water use and supply. Therefore, the ecological compensation entities for Wuxijiang River are mainly the local governments and the residents.

The primary impacts on residents in the water source area

There are three levels of water source protection; the core area, the secondary core area, and the water’s edge. Because water resources development occupies a large amount of land, regional environmental protection can lead to limitations in further developing industry, thus reducing the livelihood opportunities for local residents, which negatively affects economic development. The interests of residents in water areas need to be further determined, and require government departments to create effective coordination programs.

The primary impacts on governments at all levels in Quzhou city, Qujiang district and township governments

There is a question as to which level of government is most impacted by claims of compensation throughout the watershed. This study argues that ecological compensation is mainly to internalize the relevant production costs so as to optimize ecosystem service function. For the ecological compensation of Wuxijiang River, the external costs are the investment cost of ecological protection and construction. Therefore, the government level most relevant to the target of ecological compensation is related to the compensation degree of the opportunity cost of development. Wuxijiang River is a tributary on the upper reaches of the source of the Qiantangjiang River, and development costs can be compensated in Quzhou city. With Quzhou city’s focus on ecological protection and construction, its GDP value will not have a great impact on Zhejiang Province. The loss of opportunity for the development of the Wuxijiang River’s water source protection area is fully shared across the area under the jurisdiction of the Quzhou city government (Yang, Cai & Zhang, 2017).

Payment for gambling on environmental services in the Wuxijiang River source area

The main stakeholders’ benefit for playing the game with Wuxijiang River water resources includes the benefit games between the local government and the central government, local governments at all levels, and residents and the government.

The benefit game between the local government and the central government is reflected in the mismatch of the financial and administrative power controlling the water resource, which shows the trend of financial right above power. The status of central versus local governments differs greatly. The administrative power of the local government is passively increased. The interest’s game between the local governments at all levels is mainly reflected in that the financial and administrative rights do not match between the superiority and inferiority within the administrative hierarchy. If the county government manages directly, while giving the corresponding financial power, the system will be smoother (Song & Liu, 2005). The interest game between the residents and the government in the water source area is mainly caused by the misunderstanding of the compensation methods by residents in water sources. At present, the water source ecological compensation in China is basically government compensation. The mode of compensation is divided into two types: transfusion and hematopoiesis. The former means to give the compensation materials to the residents to sustain their basic lives, and the latter means the compensation materials given can further help them increase income and improve their living standards. The residents tend to prefer the transfusion type, hoping to direct compensation with money. They have low recognition of the hematopoietic compensation type such as policy compensation and industrial compensation (Gen et al., 2010), which leads to the lack of resources, insufficient compensation, and psychological satisfaction, resulting in an objective benefit game.

Estimating compensation amount using CVM-based ecological methods

According to the results of the CVM survey, the ecological compensation standard of Wuxijiang River was determined using conditional value evaluation. The sample area included 15 blocks in the lower reaches of the Wuxijiang River Reservoir Area and the 12 villages in four towns in Wuxijiang River. In the two cases of nonparametric and parametric estimation, we estimated the WTP and WTA for ecological compensation in Wuxijiang River in order to obtain the ecological compensation theory value.

Design of CVM survey

In the design of the survey, WTP and WTA were investigated. The survey on WTP covered 15 blocks in the Wuxijiang River water source. The detailed sample distribution is described in Table 1.

The survey covered three areas. The first area included detailed respondent information, to describe and understand the basic social characteristics of the interviewees and provide the basis for the further analysis of the data. The detailed socio-economic characteristics of the WTP sample is described in Table 2.

The second area was analyzed to better understand the use of water resources, including the price of water in Quzhou, the monthly water consumption of interviewees, and the level of concern about water-related environmental problem of the interviewees. This data was analyzed to understand if the local water quantity and water quality meet the overall demand, as well as to improve the overall understanding of environmental protection in the Wuxijiang River watershed.

A survey of WTA was conducted to investigate the 12 administrative villages with the shortest distance between Wuxijiang River reservoir in the four townships of Lingyang and Jucun townships in the upper reservoir, while Hunan Township and Huangtankou Township in the lower reservoir. The detailed socio-economic characteristics of WTA interviewees is described in Table 3.

Nonparametric estimation of the average WTP

Table 4 shows WTP for water source ecological compensation of the Wuxijiang River. The survey results of WTP were analyzed, and the expectation of the average WTP was computed on the basis of the value and frequency of WTP. The mathematical expectation model of discrete-time variable WTP was used in the computation of WTP (Kong, Xiong & Zhang, 2014; Guan et al., 2016), expressed as: $$E_{\text{WTP}}={\displaystyle \sum _i^m}{B}_i\cdot P_i=46.06\text{ yuan}/\left(\text{year}\cdot \text{household}\right)$$ (1) where B_i is the amount of tender, P_i is the probability of the amount chosen by the interviewee, and m is the number of tenders that can be selected, which is set to 11 in this paper.

According to the investigation, there were 46.2% zero payment wishes in the sample. For this reason, the calculation model of the WTP was corrected. The formula is described as follows: $${E^{\prime }}_{\text{WTP}}=E_{\text{WTP}}^{+}\times \left(1-{\text{μ}}_0^{\text{WTP}}\right)$$ (2) where, ${E^{\prime }}_{\text{WTP}}$ represents the expected value of nonnegative WTP; $E_{\text{WTP}}^{+}$ represents the expected value of positive WTP; and ${\text{μ}}_0^{\text{WTP}}$ represents the ratio of zero payment intention.

It can be calculated that: ${E^{\prime }}_{\text{WTP}}$ is 24.79 yuan/(month·household), that is, 297.48 yuan/(year·household).

If one household unit operates with three individuals, the ${E^{\prime }}_{\text{WTP}}$ is 8.29 yuan/(month·capita); if one household unit operates with four individuals, the ${E^{\prime }}_{\text{WTP}}$ is 6.20 yuan/(month·capita).

The mathematical estimation model was used to obtain the per capita ecological compensation standard for local residents (Yu & Cai, 2015). Q denotes the amount of ecological compensation payment, W_WTP was used to describe the maximum willingness of payment and N to describe the number of people who use running water in the city. Among them, 831,060 people are living in the downtown area, while 126,170 people who use the same water supplies from the Wuxijiang River are living in urban areas. M was used to stand for the population in the four towns of Wuxijiang River water source: Lingyang Town (5,467), Jucun Town (4,063), Huangtankou Town (9,652) and Hunan Town (11,868). All population data came from public statistics released in December of 2015. Then, the upper limit of the ecological compensation standard was given by: $$\begin{array}{l}Q=W_{\text{WTP}}\times N/M=8.29\times 12\times \frac{831,060+126,170}{5,457+4,063+9,652+11,868}\\ =2993.81\text{\hspace{0.17em}yuan}/\left(\text{year}\cdot \text{capita}\right)\end{array}$$ (3)

The lower limit of ecological compensation standard is given by: $$\begin{array}{l}Q=W_{\text{WTP}}\times N/M=6.20\times 12\times \frac{831,060+12,6170}{5,457+4,063+9,652+11,868}\\ =2294.39\text{\hspace{0.17em}yuan}/\left(\text{year}\cdot \text{capita}\right)\end{array}$$ (4)

Nonparametric estimation of the average WTA

Table 5 shows the WTA of water source ecological compensation of Wuxijiang River. The survey results of the WTA are analyzed. The expectation of the average WTA was calculated on the basis of the value of the affordable WTA, and the frequency of the WTA was given by: $$E_{\text{WTA}}^{+}={\displaystyle \sum _i^m}{B}_i\cdot P_i^{+}=6591.95\text{ yuan}/\left(\text{year}·\text{household}\right)$$ (5) where B_i was the amount of tender, P_i was the probability of the amount chosen by the interviewee, and m is set to 48 in this paper.

Through the survey, it was found that there 47.1% of the surveyed population have a zero WTA. Therefore, according to the Spike model (Du et al., 2013) of econometrics, the computation model of WTA is corrected by: $${E^{\prime }}_{\text{WTA}}=E_{\text{WTA}}^{+}\times \left(1-{\text{μ}}_0^{\text{WTA}}\right)$$ (6) where ${E^{\prime }}_{\text{WTP}}$ is the expectation of non-negative WTA, $E_{\text{WTA}}^{+}$ is the expectation of positive WTA, and ${\text{μ}}_0^{\text{WTA}}$ is the rate of zero WTA.

By computation, the ${E^{\prime }}_{\text{WTA}}$ is 3864.48 yuan/(year·household). Most of the residents in this area are farmers, and in this area, if one household includes four persons, ${E^{\prime }}_{\text{WTA}}$ is 3864.48 yuan/(year·household). Then, the average WTA of the local residents in Wuxijiang River was determined.

Parametric estimation of the average WTP

STATA software (Yan, Zhang & Jiang, 2016) was used to analyze the economic factors that affect WTP and WTA in the survey (Xu, Yu & Li, 2015). Regression processing was implemented with stepwise regression analysis and the least squares method (Pan, 2014; Liu & Wang, 2017) to obtain variables with the greatest impact on WTP and WTA. Table 6 shows WTP of the interviewee and the regression results of the related variables. Table 7 shows the WTA of the interviewee and the regression results of the related variables.

By analyzing the regression results, it was found that annual income, age, gender and the amount of environmental concern were the main factors affecting WTP for ecological compensation. There was a positive correlation between the annual income of residents and WTP under consideration for only one influencing factor. The regression coefficient of the annual income was greater than zero, which represented the higher annual income and the greater amount of WTP. The age of the resident was inversely proportional to WTP. The regression coefficient of the annual income was less than zero, which represents older interviewees and a correspondingly smaller WTP. The greater the regression coefficient of the gender, the greater the amount of WTP. WTP of women was assigned to zero, while men were assigned to one. Thus, men had greater WTP. The “concern degree” of the environment was directly proportional to the descriptive statistics. The regression coefficient of the concern degree for the environment was less than zero, which indicated that the greater the value of the descriptive statistics, the greater the WTP.

According to the obtained regression results, the average WTP was obtained by: $$\{\begin{array}{c}\text{ln}{W}_{\text{WTP}}=\text{ε}x+\text{δ}\\ E_{\text{WTP}}=\text{exp}\left(\text{ε}x+\text{σ}/2\right)\end{array}$$ (7)

For the 46.2% of the population in which there was zero payment intention, the above formula needs to be adjusted as: $$\{\begin{array}{c}\text{ln}{W}_{\text{WTP}+1}=\text{ε}x+\text{δ}\\ E_{\text{WTP}+1}=\text{exp}\left(\text{ε}x+\text{σ}/2\right)\end{array}$$ (8)

The mathematical model was used to obtain the per capita ecological compensation standard for local residents. In this model, x was used to describe the main factors affecting the willingness of the respondents to pay, annual income, age, sex and ecological environmental concern for the respondents, ε was used to describe the regression coefficient for each factor, and δ was used to describe random perturbation term with a normal distribution of [0, δ/2]. According to the above, lnW_WTP+1 was also a function that follows the normal distribution (Brown et al., 2014), σ represents the standard deviation of the normal distribution function. From the result of the regression, σ = 0.955205, thus; $$\begin{array}{l}{E}_{\text{WTP}}=\exp \left(C+{\text{ε}}_1\overline{x_1}+{\text{ε}}_2\overline{x_2}+{\text{ε}}_3\overline{x_3}+{\text{ε}}_4\overline{x_4}+\text{σ}/2\right)\\ =528.72\text{\hspace{0.17em}yuan}/\left(\text{year}\cdot \text{household}\right)\end{array}$$ (9)

For example, in a family of three, E_WTP is 176.24 yuan/(year·capita), while in a family of four, E_WTP is 132.18 yuan/(year·capita).

The upper limit of the ecological compensation standard was given by: $$\begin{array}{l}Q=W_{\text{WTP}}\times N/M=176.24\times \frac{831,060+126,170}{5,457+4,063+9,652+11,868}\\ =5434.99\text{\hspace{0.17em}yuan}/\left(\text{year}\cdot \text{capita}\right)\end{array}$$ (10)

The lower limit of the ecological compensation standard was given by: $$\begin{array}{l}Q=W_{\text{WTP}}\times N/M=132.18\times \frac{831,060+126,170}{5,457+4,063+9,652+11,868}\\ =4076.25\text{ yuan}/\left(\text{year}\cdot \text{capita}\right)\end{array}$$ (11)

Parametric estimation of the average WTA

Based on the analysis of relevant variables, regression analysis and the least squares method were used to regress the number of WTA with the factors related to the WTA. Based on the results, it can be seen that five variables (age, gender, marital status, whether a respondent was born locally and lived in the village) had the most influence on the respondent’s WTA compensation. Among the above factors, the regression coefficient of the age of the interviewees was greater than zero, indicating that the older the subject, the greater the value of the willingness to be paid. Putting aside the other factors, the older the respondents were, the higher their WTP for the protection of the environment. The respondents’ age had a greater impact on their WTA. Whether the family lived in the main village had a significant impact on their willingness to receive compensation. The regression coefficient was below zero, indicating that the higher the statistical value, the lower the value of the willingness to be paid. The gender regression coefficient of the interviewees was higher than zero. The descriptive statistics for female and male are zero and one, respectively, indicating that men have a stronger WTP than women. The regression coefficients of marital status and whether the respondent was born locally were less than zero, which was positively related to their willingness to receive compensation and had a negative correlation, and had less influence on their willingness to be paid. Whether the married pair was affected by the willingness to receive compensation was very small, and therefore was not tested by a level of significance test.

With analysis of the related variables, regression processing was implemented with stepwise regression analysis and the least square method for WTA values and the related factors. Regression results were used to calculate WTA: $$\{\begin{array}{l}\ln W_{\text{WTA}}=\text{ε}\prime w+\text{δ}\prime \\ E_{\text{WTA}}=\exp \left(\epsilon \prime w+\sigma \prime /2\right)\end{array}$$ (12)

As there was 47.1% of the surveyed population with zero WTA, the above equation was corrected to: $$\{\begin{array}{c}\text{ln}{W}_{\text{WTA}+1}=\text{ε}\prime w+\text{δ}\prime \\ E_{\text{WTA}+1}=\text{exp}\left(\text{ε}\prime w+\text{σ}\prime /2\right)\end{array}$$ (13) where w are the main factors affecting WTA, which are age, gender, marital status, if they were born locally and live in the main village, ε′ is the regression coefficient, and δ′ is a random perturbation with normal distribution of [0,σ′/2] (Rao, Lin & Kong, 2014; Li & Li, 2017). lnW_WTA+1 was also a function with the normal distribution. σ^′ is the standard deviation, and σ^′=4.454884. Then: $$\begin{array}{l}{E}_{\text{WTA}}=\exp \left(C+{{\epsilon }^{\prime }}_1\overline{w_1}+{{\epsilon }^{\prime }}_2\overline{w_2}+{{\epsilon }^{\prime }}_3\overline{w_3}+{{\epsilon }^{\prime }}_4\overline{w_4}+\sigma \prime /2\right)\\ =1514.04\text{\hspace{0.17em}yuan}/\left(\text{year}\cdot \text{household}\right)\end{array}$$ (14)

If one household includes four persons, E_WTA is 378.51 yuan/(yuan/capita·year), which was the annual per capita compensation WTA of local residents in Wuxijiang River.