Sage Journals: Discover world-class research

Abstract

To alleviate the deteriorating environment and protect biodiversity, China has implemented a natural forest protection system, demonstrating the importance of sustainable forest management for ecological conservation and socio-economic development, including the complete cessation of commercial logging of natural forests. Financial compensation is adopted to increase farmers’ enthusiasm within the commercial Logging Ban of Natural Forests framework. This study used the contingent valuation method and the Heckman two-stage model to explore farmers’ willingness to participate in the Logging Ban of Natural Forests and the willingness to accept by survey data on 486 farming households. 72% of farmers are willing to join the Logging Ban of Natural Forests. Their willingness to accept is 517.95 yuan/ha per year, higher than the current state subsidy standard (225 yuan/ha per year). The key factors influencing willingness to accept include the education and degree of fragmentation of woodland and village collective willingness to accept. The age of the rural household head, the fragmentation of the forest, and the evaluation of the Logging Ban of Natural Forests policies have inhibited the increase of farmers’ compensation. Farmers’ assessment of the Logging Ban of Natural Forests policy only impacts the medium level of compensation. The age and the degree of forest fragmentation would affect the higher compensation amount. The results from this study suggest more financial sources and increased compensation standards are needed. The government should also strengthen ecological awareness and adopt different compensation standards for other groups to achieve sustainable forestry.

Keywords

Logging ban of natural forests contingent valuation method financial compensation willingness to accept Heckman two-stage model

Introduction

Sustainable development means that the natural environment is in harmony with the interests of socio-economic development and the needs of human beings.¹ A well-functioning ecosystem can provide human beings with necessary food, water, energy, wood, and other raw materials, as well as a series of vital services such as climate maintenance, hydrological regulation, entertainment, and culture, which play an essential role in the sustainable development of human beings.² As the critical component of the terrestrial ecosystem, the forest ecosystem occupies a vital position in the living communities of mountains, rivers, forests, fields, lakes, and grasslands, providing water purification, soil conservation, carbon sequestration, raw material production and supply for the environment, and Important services such as sight-seeing.^3–8 China is an essential country in the world regarding forestry and its forest ecosystem functions.⁹ In 2021, the forest area of China was 220.44 million hectares, accounting for 22.96% of China's land area. Natural forest is the essence of forest resources with a more complex structure, richer biodiversity, and a more ecologically functional ecosystem. The Ninth National Forest Resources Inventory Data in China shows that the natural forests cover an area of 140 million hectares and account for 63.64% of the forest area, providing a wealth of ecological products and services for society. The ecosystem service it produces support sustainable human livelihoods and promotes Socio-economic Development.¹⁰

The natural forest protection project started in 1998 as a pilot project in state-owned forest areas in 12 provinces (autonomous regions and municipalities) and then was fully launched in 17 provinces (autonomous regions and municipalities) in 2000. The primary protection measure is to stop commercial logging of natural forests in the upper reaches of the Yangtze River and the upper and middle reaches of the Yellow River. In 2011, the area where natural forest logging was stopped was expanded to include key collective forest areas in the south, such as Guangxi, Jiangxi, and Fujian. The goal is to protect and restore natural forest resources and gradually improve the ecological functions of natural forests. By 2020, the state had invested 508.3 billion yuan in the wild forest protection project. The logging ban on natural forests reduced the supply of annual timber production by approximately 34 million cubic meters and helped to restore 1.296 million hectares of natural forests in China. The total carbon storage in forests is 9.186 billion tons, of which more than 80% of the contribution comes from natural forests. The ban has promoted increased forest area and forest stock in China.

Conservation and restoration of natural forests in the southern collective forest area has continuously improved the function of forest ecosystem service and generated ecological benefits but significantly impacted farmers’ income who depend on the forests. However, based on the financial compensation policy, the subsidy standard for natural forest conservation is low at 150–300 yuan/ha. The government needs to use transfer payments to make the compensation. Therefore, exploring the compensation standards for the logging ban is critical to the well-being of farmers in forest areas and protecting natural forest resources.

Financial compensation (FC) is similar to ecosystem service payment (ESP) and can serve as an efficient mechanism to convert the external non-market value of ecosystem services into economic incentives for internal actors to provide such services.^11,12 It regulates stakeholders’ interests through financial instruments and includes many incentive-based approaches to environmental management. It is an innovative approach that combines ecological protection with promoting sustainable socio-economic development.^13–16 Still, the FC is a more likely government-induced mechanism than a voluntary exchange. The Slope Conservation Program (SLCP) and the Natural Forest Protection Program are extensive ecosystem protection programs in China. FC for the SLCP involved 120 million households; including 32 million rural families.¹⁷ Current financial compensation, mainly indirect monetary subsidies, is largely dependent on third-party measurement or simple by the government, imposing economic and political challenges. It is not just a matter of economic efficiency but also fairness. For example, wetland habitats’ mitigation and restoration plans are wide, and it is difficult to track payments.¹⁸ The FC for biodiversity is difficult to measure, and compensation for this field also has significant limitations.¹⁹ Therefore, most programs rely mainly on adjusting economic means, supplemented by non-material indirect compensation.

Most FC calculation methods are based on cost analysis (e.g. opportunity cost) and benefits research using econometric methods. The protection of water resources and the compensation for soil erosion were explained based on the component–benefit analysis method, explored the marginal opportunity cost and market value method, and calculated the cost–benefit of ecological functions from both sides.²⁰ The applicability of ecological criteria of the winter wheat-fallow system in Hebei Province was evaluated based on the theory of opportunity cost approach and the logistic model.²¹ Moreover, the opportunity cost method was used to explore the relationship between the compensation standards for terraced fields in southwest China and farmers’ willingness to accept (WTA).²² Game theory was used to analyze the interaction between the upstream and downstream governments and the central government in the South-to-North Water Transfer Project. After implementing ecological compensation, the upstream government obtained 78% of the environmental benefits, and the high compensation fee reduced the downstream government's WTA.²³ Visible above that each of the above methods has its specific adaptability.

The opportunity cost method only considers the actual economic loss. It ignores farmers’ willingness, leading to low enthusiasm of farmers, and it isn’t easy to reflect on the changing process of a strategy. The contingent valuation method (CVM) is based on the respondents’ wishes and conforms to the principle of cost-effectiveness and maximizing benefits.²⁴ It can solve problems that are difficult to quantify in natural resources and environmental values, more feasible in operation.^24,25 Many scholars have used CVM to study the influencing factors of monetary compensation. For example, Peng et al.²⁶ combine CVM and Bayesian network models to analyze the willingness to pay for water reserves from the residents’ perspective influencing factors. Based on farmer survey data, Xiong and Kong²⁷ used CVM to study wetland farmers’ willingness and level of compensation. Results show that 87.8% of farmers are willing to accept payment, with an average price of US$858.81/household/year. Panwanitdumrong and Chen²⁸ used the binary selection method in CVM combined with the logistic model to evaluate the willingness to pay 1655 respondents in Thailand for coastal attractions without marine debris. The estimated average was 92.94 baht per person per visit. Sejati et al.²⁹ calculated the total economic value of lowland rice cultivation using CVM and analyzed the factors affecting land conversion from agricultural to non-agricultural uses. In addition, some scholars have used CVM to evaluate the economic value brought by the protection of intangible cultural heritage or applied the conditional value assessment method to estimate the compensation value related to ecological protection policy (WTA/WTP).^30–32 Therefore, as a narrative preference evaluation method,³³ CVM is suitable for calculating compensation or projects that are difficult to quantify in monetary terms. It can help promote the implementation of environmental protection projects and improve policies.

In addition, using CVM is necessary to clarify from the perspective of WTP or the standpoint of WTA. WTP measures the willingness of respondents to pay for a good or service. WTA estimates the monetary compensation respondents are giving up a good or service. Therefore, the relevant measure depends on whether the respondent is the receiver of the ecosystem services (using WTP) or the provider (using WTA).^34–37 From the perspective of the ecological system, favorable ecological measures/policies may lead to a decrease in WTA.

Negative ecological impacts may lead to an increase in its value.³⁸ Given the increasing importance of sustainability issues, many scholars have used the WTP and WTA to estimate residents’ willingness or the benefits that projects can bring.^39–41 For example, Yin et al.⁴² conducted a survey study in Beijing, China, to estimate welfare losses from PM2.5. Truong et al. assessed the willingness of residents to pay for wetland conservation by conditional estimates. After further examining the factors affecting WTP, conservation measures were suggested worth taking.^39,43 Abebe et al.,⁴⁴ Mueller et al.,⁴⁵ Li et al.,⁴⁶ Aydoğdu et al.,⁴⁷ and other scholars have studied the protection of ecosystems such as watersheds and cultivated land from the perspective of WTP. The results of the effectiveness test of the method indicated that the current practice of using WTP to assess all values substantially might be widely biased.⁴⁸ Therefore, the importance of environmental items is argued hard to be quantified through regular market transactions from the perspective of WTA. Wang et al.⁴⁹ discussed the WTA of the interviewees in the area where the Grain for Green Project (GFGP) was implemented and further calculated the upper and lower values of the WTA. Al-Assaf et al.⁵⁰ measured the willingness of residents to accept compensation for sustainable ecosystem management in local communities. Wang et al.,⁵¹ based on the theory of expansion plan behavior (TPB), studied the willingness of Chinese rural households to accept degradable agricultural plastic film compensation and the proportion of willingness to compensate. As seen from the above, WTP is applicable to the improvement of welfare, and WTA is to the loss of welfare.

The FC policy can generally balance the contradiction between decision-makers and vested interests, but the specific amount of FC needs to be further discussed. Influencing factors or a single compensation amount calculation ignores the differences in external conditions, loses the pertinence of farmers with different characteristics, and hinders the promotion of environmental protection projects to a certain extent. The rationality of the FC standard should relate to the incentive effect of the policy on farmers as well as the efficiency. Therefore, in policy formulation, it is necessary to consider the amount of compensation suitable for both the payer and the recipient and its differences in the face of heterogeneous ecological balance for farmers. The CVM is just in line with the standard calculation of compensation items or protection items based on the wishes of the indebted. It is conducive to achieving a “win-win” state for both parties.

This article starts from the perspective of farmers’ WTA in the southern collective forest area and the context of implementing the ban policies. We will first examine the factors affecting farmers’ WTA, then the factors affecting the desired value of farmers’ WTA. The CVM is applied to calculate the compensation standard under the ban and compare whether the fund supply-demand between the government and the farmers is balanced. The standard calculation becomes the key to the study of this article. While the policy encourages farmers’ utility, it maximizes financial benefits and explores a new mechanism of “compensation (subsidies) based on benefits” for natural forests.

Data collection

Jiangxi Province is located in the south of China and the middle-lower reaches of the Yangtze River. The hilly and mountainous areas dominate the topography. The forest resources show a good trend and belong to the southern collective forest area. The data of China's ninth forest resources inventory shows that the forest coverage rate reaches 61.16%, ranking second in China, with a total forest area of 10.21 million hectares, of which the natural forest area accounts for about 63.89% of Jiangxi Province's forest land area. In 2016, Jiangxi was included in the first batch of the natural forest “comprehensive protection” provinces in addition to the natural forest protection project area, requiring the complete cessation of commercial logging of natural forests. The central government compensated farmers for the cessation of logging. A total of 4.05 billion yuan of natural forest protection subsidies will be issued from 2016 to 2020. The province is at the forefront of implementing the mandate to ban logging, which is typical and exemplary.

To understand the current status of FC in terms of the logging ban among farmers, survey data were collected during the summer of 2019 by a team from the School of Economics and Management of Jiangxi Agricultural University. Stratified random sampling was used to conduct a household questionnaire survey on the farmers in the study area. The specific method is: the first step is to select the study area. The sample of the study comes from 12 counties in Jiangxi Province. The forest coverage of each county is above 70%. They are distributed in southern Jiangxi (Dayu County, Chongyi County, Quannan County, Longnan County, Ganzhou City), Central Jiangxi (Yichun City, Tonggu County, and Jing'an County, Suichuan County, and Wan'an County of Ji'an City), and Northern Jiangxi (Xiushui County and Wuning County of Jiujiang City, Wuyuan County, and the Dexing City of Shangrao City). They are geographically representative. Five townships were selected and stratified based on the forest area as the standard. Then one sample administrative village was randomly selected from each township based on topography and regional distribution.

Farmers were sampled in the second step. The main target is to include ordinary farmers, excluding forestry management organizations. Ten farmers are randomly chosen from each village. Since Jiangxi Province is also a significant labor exporting province, the rural household in the following serial number of the roster will supply if the sampled farmers work outside. If the farmer was absent, then the respondent was discarded. The questionnaires mainly include individual characteristics of the householder (age, education level, etc.), family characteristics, forest land resources and management characteristics, evaluation of the ban policies, WTA, etc. One-to-one and face-to-face interviews were adopted, and the interviewee is the head of the household or the primary decision-maker of the family. At the end of the survey, the total number surveyed was 589. Because this study focused on the policy of compensation for the rural households who were asked to ban natural forest logging, 502 samples were obtained after deleting the farmers without natural forests. Finally, 486 valid sampled farmers are retained, removing some missing data and outliers.

Methodology

Data collection

The CVM is used to investigate how much the WTA of the farmers participates in natural forest conservation. WTA reflects the opportunity costs of the farmers for the change: farmers who own the natural forests give up their production and lifestyles. The research group used the open-question format to ask the interviewee for the minimum compensation. What is the minimum amount of FC wanted every year? Therefore, the questionnaire includes four features: personal characteristics of household heads, household characteristics of farmers, the cognition of the ban and ecological protection, and the WTA.

This article adopts the double-bounded CVM. First, the farmer was asked whether he or she was willing to participate in the ban. If the answer is “yes,” the farmer would be further asked whether he accepts an initial compensation level ( $W_{1}$ ); if the farmer is unwilling to get it, the compensation will steadily increase to various levels W_i until it is acceptable. The WTA will show at ( $W_{1}$ , $W_{i}$ ). There is a pre-determined level W_max. If farmers still tend to raise the level of compensation beyond W_max, the WTP will be in the range ( $W_{\max}$ , $+ \infty$ ). For farmers willing to accept the initial payment, the amount of compensation is gradually reduced until the farmer is unwilling to accept it. If the farmers’ compensation amount is 0, ask for specific reasons. Therefore, the double-bounded is the half-open and half-closed interval data or the right-censored data will be obtained, which is the farmers’ compensation level interval.

Calculation method

In the double-bounded CVM, The WTA uses the interval median for that farmer's bid interval. In the research process, some farmers expressed their willingness to stop logging natural forests even without subsidies. The WTA for these farmers is determined to be 0 considering the genuine desire. In the CVM assessment method, this study uses a non-parametric approach to measure the value of farmers’ willingness to be reimbursed, which is calculated as follows:

E (W T A) = \sum_{i = 1}^{n} P_{i} \times W_{i}

(1)

Among them,

W_{i}

represents the ith bid amount selected by the sampled farmer, P_i represents the probability that the farmer chooses

W_{i}

and n is the amount of money available for selection. The farmers’ WTA and the minimum estimating value of the farmers’ WTA can be further obtained according to formula (1):

E (W T A)_{\min} = E (W T A)_{\max} \times Compensation rate

(2)

Models

The farmers’ participation in FC for the ban can be divided into two stages: (1) whether farmers are willing to accept payment; (2) why acceptance of the compensation level. Due to the non-random selection of samples, the above operation will cause sample selection bias when analyzing WTA. Heckman’s two-stage model, which was accepted to avoid bias in sample selection,^41,52,53 contains the selection equation and the result equation. The selection equation has at least one explanatory variable that is not included in the result equation. The two-stage factors are not identical variables.^54,55

The first stage of the selection equation analyzes the influencing factors of WTA by the Probit model. If “farmers are willing to accept FC from the government for natural forest logging ban.” $Y_{i}$ = 1, $Y_{i}$ = 0 indicates that they are unwilling to accept compensation, and unpredictable latent variables determine the observable. Therefore, the expression is as follows⁵⁴:

Y^{*} = ω_{i} γ + μ_{i}

(3)

In formula (3), ω is the independent variable that affects whether the farmer is willing to accept compensation. ϒ denotes the parameter to be estimated. i is the farmer surveyed for the ith position. The independent error term µ follows a normal distribution.

In the second stage of the result equation, the inverse Mills ratio calculated in the first stage as a correction variable, the OLS model is used to explore further the factors influencing farmers’ willingness to accept compensation. The expression of the equation is:

E (Z_{i} | Y = 1, O b s e r v a b l e) = x_{i} β + ρ σ_{ε} λ (- ω_{i} γ)

(4)

In formula (4),

x_{i}

is the row vector of the explanatory variable,

β

is the column vector of the parameter to be estimated,

ρ

is the correlation coefficient of the error terms of the two equations,

σ_{ε}

is the covariance of the error terms of the two equations, and

λ

is the inverse Mills ratio.

Following Heckman's two-stage regression model, this paper further uses the quantile regression (QR) model to analyze the key factors that affect the amount of farmers’ willingness to accept compensation. Suppose the quantile regression model is:

Y_{q} (x_{i}) = x_{i}^{'} β_{q}

(5)

In formula (5),

Y_{q} (x)

is the linear function of x,

β_{q}

is the “

q

quantile regression coefficient,” and i is the

i

th farmer household under investigation.

{\hat{β}}_{q} = min_{β_{q}} \sum_{i : Z_{i} \geq {x^{'}}_{i} β_{q}}^{n} q | \underset{i}{Z} - \underset{i}{x^{'}} \underset{q}{β} | + \sum_{i : Z_{i} < {x^{'}}_{i} β_{q}}^{n} (1 - q) | \underset{i}{Z} - \underset{i}{x^{'}} \underset{q}{β} |

(6)

In formula (6),

Z_{i}

is the dependent variable, that is, the amount of compensation for farmers’ willingness to participate in the ban, and

x_{i}

is the independent variable vector.

Variable selection

The dependent variable is the WTA of farmers for the ban. Referring to Xiong and Kong (2017), the independent variables have the following categories:

Individual characteristics of the householder, including male, age, education level, and employment type. The employment type of the householder is set in a fixed order according to the degree of separation from agriculture of the head of household.

Family characteristic variables, including the proportion of household labor force, annual per capita non-forest income, whether to engage in forestry-related business activities, and the degree of fragmentation of woodland. We take the logarithm of the original data of the annual per capita non-forest income index.

Natural forest logging ban and ecological protection-related variables, including whether to participate in forestry training, the village collective's willingness to accept payment, evaluation of the implementation of ecological compensation policies, and evaluation of the natural forest logging ban policy.

The specific settings and descriptive statistics of the variables are shown in Table 1.

Table 1.

Description and assignment of variables in the model.

Variable name	Variable meaning and assignment	Mean	SD	Med.
(1) Dependent variable
WTA	0 = no; 1 = yes	0.71	0.45	1
WTA level	The actual amount (Yuan)	48.22	44.32	30
(2) Individual characteristics of a householder
Male	0 = female; 1 = male	0.94	0.23	1
Age	Actual age	55.94	10.58	55
Education level	1 = primary school and below; 2 = junior high school; 3 = high school (secondary school); 4 = college or undergraduate and above	1.90	0.83	2
Employment type ^a	1 = forest farmer; 2 = part-time agricultural job; 3 = non-agricultural employment; 4 = unemployed (child support, government subsidy)	–	–	–
(3) Characteristics of rural households
The proportion of household labor force	Labor force/family population (%)	0.59	0.27	0.6
Annual per capita non-forest income	Actual amount/population (Yuan)	25881.29	98203.78	13,000
Whether to engage in forestry-related business activities	0 = no; 1 = yes	0.61	0.49	1
Fragmentation degree of woodland	Woodland area (hectare)/number of family woodland plots	1.48	1.83	0.88
(4) Natural forest logging suspension and ecological protection awareness
Whether to participate in forestry training	0 = no; 1 = yes	0.29	0.46	0
Village collective willingness to accept payment	Except for the sample household head, other farmers’ willingness to accept payment in the village	0.72	0.30	0.8
Evaluation of the implementation of ecological compensation policies	1 = unclear; 2 = dissatisfied; 3 = fair; 4 = satisfied	2.55	1.33	3
Evaluation of the policy of stopping logging in natural forests	1 = unsatisfied; 2 = fair; 3 = satisfied	2.16	0.94	3

Note^a: Forest farmers accounted for 51.6%, part-time agricultural jobs accounted for 16%, non-agricultural employment accounted for 27.3%, and unemployed (child support, government subsidy) accounted for 5.1%; the unemployed farmers here are older people who are unable to work. WTA: willingness to accept.

Empirical results

Measurement results of farmers’ compensation

The median of farmers’ WTA is chosen to substitute the level of WTA in the interval. According to formula (1), the upper limit of the average compensation value for farmers’ WTA can be calculated as 723.3 yuan/hectare in 486 cases. The 350 households surveyed are willing to accept payment from the government for the logging ban, accounting for 72.02% of the total surveyed households, nearly 30% of which were unwilling to receive the compensation. Perhaps the condition of forest land resources is too good, and the economic income from the forest land far exceeds the financial compensation, so they are not willing to accept payment. Therefore, according to Kristrom's Spike formula,⁵⁶ the lower limit of the average value of farmers’ WTA can be calculated: as 723.3*72.02% = 520.95 yuan/ha. In summary, farmers expect to receive the government's compensation standard for the ban is 520.95–723.3 yuan/hectare, as shown in Tables 2 and 3.

Table 2.

Heckman's two-stage model estimation results.

Variable name	Model (1)The first stage: Probit		Model (2)The second stage: OLS
Variable name	Coe.	S.E.	Coe.	S.E.
(1) Individual characteristics of householder
Male	−0.243	0.324	16.464*	9.851
Age	0.005	0.008	−0.716***	0.247
Education level	0.198*	0.102	0.429	3.11
Employment type	−0.132*	0.078	−3.825	2.673
(2) Characteristics of rural households
The proportion of household labor force	0.029	0.298	−8.506	9.828
Degree of woodland fragmentation	0.112*	0.058	−2.152*	1.282
Annual per capita non-forest income	−0.029	0.074	−0.089	2.225
Whether to engage in forestry-related business activities	0.044	0.152	7.192	4.951
(3) Natural forest logging suspension and ecological protection awareness
Whether to participate in forestry training	−0.084	0.179	-	-
Village collective willing to accept payment	2.675***	0.266	-	-
Evaluation of the policy of stopping logging in natural forests	0.094	0.078	−6.882***	2.513
Evaluation of the implementation of ecological compensation policies	0.044	0.058	1.522	1.871
Constant	−1.589	0.976	89.81***	28.654
λ	13.901*	8.283
Rho/sigma	0.325		42.770
Wald chi²	25.27
Prob > chi²			0.005
N	486		350

Note: *, **, ***are significant levels at 10%, 5%, and 1%, respectively; the values are partial regression coefficients, which are the results after rounding.

Table 3.

Comparison of compensation amount of farmers’ willingness to participate in the ban.

Calculation method	All farmers	Farmers willing to accept
The contingent valuation method (CVM)	520.95 yuan/ha	723.3 yuan/ha
Parameter estimation (Heckman selection model)		517.95 yuan/ha

Influencing factors of farmers’ willingness and level of compensation

Influenced by individual characteristic variables, this study applied the CVM to measure the WTA of the farmer. The test results showed that the minimum value of VIF (variance inflation factor) is only 1.05, and the maximum is 1.29. The mean value is 1.15, and the variance expansion factor and tolerance of each variable are within a reasonable range, indicating that the model has no multicollinearity problem. When the average value of VIF is ≥ two, and the maximum value of VIF exceeds 10 or 1/VIF is < 0.1, there is a problem of collinearity between variables.

The MLE estimation is carried out first using the Heckman two-stage selection model. The model results show that the Wald chi² statistic is significant, and the likelihood ratio test of rho is not 0 and p = 0. The corresponding p < 0.01 indicates that the Heckman selection model is appropriate and can solve the particular bias problem in the survey samples. The regression results obtained using the Heckman two-stage estimation method are shown in Table 2, where the inverse Mills ratio value passes the statistical test at the 10% confidence level.

From Table 2, in the model's first stage, the Educational level in characteristics of rural households passed the statistical test at 10%, which was significantly positive, suggesting higher education of a householder has a more vital willingness to accept payment when other conditions remain unchanged. The possible reason is that education affects their ecological cognition. Compared with farmers with low education, farmers with higher education have a more vital environmental protection awareness and a higher degree of response to national policies. The employment of the householder has a significant adverse effect on the willingness to be paid. The lower the householder occupations are related to agriculture and forestry, the weaker their willingness to accept payment. It shows that the level of dependence on the type of employment of farmers in agriculture and forestry determines their desire to be compensated. With the agricultural economy under serious threat, the gradual loss of ecosystem services that support sustainable livelihoods is stimulating rural–urban migration. Farmers’ generally low-income level in developing countries has led to a gradual separation.^57,58 For the above two reasons, farmers are increasingly looking for non-agricultural employment in their job choices. Obviously, the less they rely on agriculture and forestry for a living, the less willing they are to accept payment.

Among the cognitive variables of natural forest stoppage and ecological protection, the village's collective willingness to compensate is positively significant at the 1% confidence level in the first stage of the selection regression. It shows that the external environment affects the individual's willingness to accept payment. The higher proportion of the village collective's willingness to take payment will drive the individual's willingness to get a bonus. There may be a mutual influence between relatives, friends, and neighbors, that is, the “neighborhood effect.” Apart from monetary compensation to farmers, the government has strengthened environmental protection and increased farmers’ awareness of ecological conservation. It is conducive to implementing the project and encourages more farmers to participate in the logging ban voluntarily. To maximize the efficiency of financial subsidies while promoting ecological benefits.

In the second stage of the model, the age of the householder in the personal characteristics variable passed the statistical test at the 1% confidence level and was significantly negative. The householder age significantly impacts the WTA for the ban. When other conditions remain unchanged, the older the head of household, the lower the compensation required for the ban, and the amount is significantly lower than for young householders. There are possible explanations for this. Older the householder will reduce the time spent on farming due to health conditions and other reasons. And the research shows that their economic income mainly comes from the alimony given by their children and government subsidies. In summary, it can be concluded that the old householder would decrease the expectations of forestry business benefits. And the way of the elder survival has changed from farming to child support and government subsidies. Therefore, the government can replace the monetary compensation by improving the rural pension system under the ban.

The gender of the householder passed the statistical test at the 10% confidence level. It has a notable positive effect on the compensation of farming households, indicating that the expectation of willingness to be paid is higher for the male householder than for females. This result may be that the householders surveyed are primarily male. Among the variables related to the ban and ecological cognition, the evaluation of the natural forest ban policies passes a statistical test at the 1% confidence level in the regression, showing a noticeable negative impact on the value of WTA. That indicates that under the circumstance that other conditions remain unchanged, the higher the degree of support of farmers to the logging ban, the better the evaluation, and the lower the demand for compensation.

It is noteworthy that the forestland fractionalization in the household characteristics passed the statistical test at the 10% confidence level in both the first and second stages of the model regressions but in the opposite direction. It indicates that farmers with highly fragmented forestland want to be compensated but do not ask for higher compensation. The reason for the opposite trend of their relationship in the first stage may be that, despite the high degree of woodland fragmentation, farmers own the property rights of woodlands and have the right to benefit from the forests. When the woodlands are stopped logging, the farmers’ economic benefits from the woods are damaged. Therefore, the farmers prefer to receive specific financial compensation. In addition, as forest land fragmentation increases, the labor and transportation costs of farmers’ forestland management become higher, which increases the input cost of production factors and decreases farmers’ willingness to cultivate their forestland. They prefer to choose compensation from government funds. At the same time, highly fragmented forestland will lead to low production efficiency, thus affecting the return on inputs and outputs. Therefore have lower expectations of the compensation fund standard.

Farmers’ willingness to accept payment: Parameter estimation

Using stata16 statistical software, according to formula (4) and the expectation calculation method in the Heckman selection model, the WTA to participate in the logging ban is adjusted after removing the sample selection bias to 517.95 yuan/ha. See Table 3 for details.

QR analysis

To further analyze various factors on the value of the wishes of the farmers repaid in full the ban policies, Table 4 reports the main findings of OLS and quantile regression.

Table 4.

The estimation results of the quantile regression model.

Variable name	Model 3 OLS	Model 4 QR_10	Model 5 QR_50	Model 6 QR_90
(1) Individual characteristics of household head
Male	18.373*** (6.190)	−2.851 (3.335)	3.32 (3.606)	63.296*** (17.506)
Age	−0.722*** (0.310)	−0.061 (0.082)	−0.215** (0.108)	−1.968*** (0.681)
Education level	−0.38 (3.10)	1.268 (0.909)	0.384 (1.240)	−12.846 (10.056)
Employment type	−3.70 (2.270)	−1.677** (0.826)	−1.749 (1.139)	−10.169 (7.658)
(2) Characteristics of rural households
The proportion of household labor force	−8.43 (9.990)	1.185 (2.913)	−6.778 (5.116)	−7.184 (27.680)
Fragmentation degree of woodland	−2.703*** (0.890)	−0.009 (0.021)	−0.013 (0.036)	−0.443*** (0.132)
Annual per capita non-forest income	0.360 (2.960)	−0.009 (0.454)	0.118 (1.310)	3.477 (7.156)
Whether to engage in forestry-related business activities	6.682 (4.880)	0.615 (1.389)	1.355 (2.673)	17.392 (19.689)
(3) Natural forest logging suspension and ecological protection awareness
Evaluation of the policy of stopping logging in natural forests	−7.401*** (2.80)	−0.139 (0.801)	−3.557** (1.674)	−14.211 (9.859)
Evaluation of the implementation of ecological compensation policies	0.780 (1.940)	−0.018 (0.408)	−1.373 (1.141)	5.961 (6.903)
Constant	94.372** (38.390)	22.555*** (7.269)	58.022*** (14.068)	188.462* (102.739)
N	350	350	350	350
R ²	0.08	F	3.75

Note: *, **, and *** indicate significance levels of 10%, 5%, and 1%, respectively; the values are partial regression coefficients, and the brackets are the robust standard errors.

According to the regression results of the OLS model 3 in Table 4, the variables that have passed the statistical significance test are male, age, fragmentation degree of woodland, and evaluation of the policy of stopping logging in natural forests. In addition, in the quantile regression model, among the statistically significant variables, the effect of employment type on the expected magnitude of farmers’ willingness to participate in the ban showed a stable first and then a sharp increase. However, it was not significant at the high quantile, indicating that the type of employment of farmers has a more dramatic effect on farmers with lower compensation for the ban. The possible explanation is that there is a particular income gap between farmers who rely on forestry income and farmers who rely on non-agricultural employment. Perhaps the low compensation level can already make up the income gap between migrant workers and farmers, so the type of employment only impacts farmers with low compensation levels. From the results of Model 4, among the interviewed farmers whose WTA is inadequate, some and only the type of employment of farmers impact them. Model 5 showed that the age and evaluation of the ban policies affect the medium compensation amount, and both are negatively significant. From the point of view of the coefficients, under the same compensation level range, the evaluation of the logging ban policy has a more meaningful effect on the value of compensation intentions than age for the same degree of compensation levels. In addition, it can be seen from Model 6 that among the groups of farmers with a higher expected value of WTA, the order of the degree of influence from high to low is male, age, and degree of woodland fragmentation.

From the variable point of view, the age of the farmer households was negatively significant at the 50th and 90th quantiles and adopted the statistical test at the 1% sensitivity level at the 90th quantile. That shows that age impacts farmers with middle and high compensation levels, and the older the householder is, the lower the level of compensation. This may be explained by the fact that elderly farmers cannot continue their forestry activities and that their children provide sufficient financial support for their living expenses. Younger householders need to shoulder more family responsibilities and prefer to get more compensation to subsidize their living. Therefore, the age of householders has an impact on the medium and high levels of compensation. As the level of compensation increases, the degree of influence of age on it shows an upward trend. The degree of woodland fragmentation and the evaluation of the ban policy passed the statistical tests at the 1% and 5% significance levels at the 90th and 50th points, respectively, and both showed negative significance. That indicates that the scale of woodland is affected by farmers with a high level of compensation, while those affected by the evaluation of the ban policy are those with a moderate compensation level. The effect of forest land scale on farmers with high compensation levels may be caused by the fact that the input of production factors is a long-term and continuous process, and the low compensation level is challenging to make up for the sunk cost of the earlier input. The evaluation of the natural forest suspension policy impacts household heads with a medium compensation level, which may be explained by the insufficient understanding of household heads of the policy, lack of ecological cognition, and lack of recognition of the long-term benefits of environmental protection. Therefore, the government should consider the characteristics of farmers’ endowments and guide farmers’ understanding of the policy while promoting natural forest deforestation projects and establishing ecological compensation mechanisms.

Conclusion and discussion

Protecting natural forests is essential to restoring the natural forest system, protecting biological diversity, and increasing ecological carrying capacity. Based on 350 valid household information from 486 sampled rural households in the southern collective forest area, this study used the double-boundary dichotomy in the CVM to estimate WTA for the ban.

The results found that 72% of the households are receptive to receiving a certain amount of compensation as an incentive to participate in the ban policy. The WTA estimated is between 520 and 723 yuan/ha, much higher than the subsidy of 225 yuan/ha currently provided by the government for the ban.

The key factors influencing farmers’ willingness to participate in the logging ban are farmers’ education level, employment type, fragmentation of woodland, and village collective's willingness to logging ban. The factors that affect the WTA include gender, age, fragmentation of the woods, and evaluation of the logging ban policy. The older the age, the higher the degree of fragmentation of woodland, and the better the evaluation of the natural forest ban, the lower WTA farmers expect from the government. External factors will also have a tremendously influential effect on the level of compensation.

Based on the estimated WTA and the gap from the current payment, the raising funds should be broadened, and FC standards should be reasonably increased. It is necessary to mobilize the power of direct and indirect stakeholders of forest ecological benefits, implement “paid use of resources,” raise funds to the greatest extent, and improve the compensation standard for farmers. All whole society should be invited to participate in ecological governance and protection projects through donations, volunteer services, etc.

The government should strengthen the publicity of ecological protection and improves the ecological awareness of farmers. The level of ecological awareness directly determines the implementation of the ban. Therefore, efforts must be made to create environmental awareness, improve farmers’ awareness and engage in FC independently. Specifically, it can be carried out to strengthen and enrich the training content of farmers’ environmental education and training platforms.

A diversified and differentiated compensation mechanism should be established to combine financial and non-financial compensation. China's existing natural forest subsidy mechanism mainly comes from direct transfer payments from the central and regional governments, and the compensation method is relatively simple. Indirect transfer payment methods can be enriched by increasing non-agricultural employment opportunities, granting non-agricultural employment tax reductions and exemptions, and strengthening rural elderly medical insurance and old-age security. The payments between various regions should be balanced through the above methods to supplement the transfer of financial resources while ensuring ecological benefits.

Footnotes

Author statement

Wenmei Liao: Conceptualization, Writing-Original draft preparation, Writing-Reviewing, and Editing. Danyang Ye: Methodology, Software, Visualization, Resources. Ruolan Yuan: Data curation. Yaoqi Zhang: Writing-Reviewing and Editing and Supervision. Qian Deng: Resources, Software, Validation.

Declaration of conflicting interests

The author(s) declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.

Funding

The author(s) disclosed receipt of the following financial support for the research, authorship, and/or publication of this article: This work was supported by the National Natural Science Foundation of China (grant numbers 71873060, 71934003, and 72263017).

ORCID iD

Wenmei Liao

Author biographies

Wenmei Liao is a professor at the School of Economics and Management, Jiangxi Agricultural University, Nanchang, China. His research interests include agricultural and forestry economics and management, focusing on rural society and governance.

Danyang Ye obtained a master's degree in Rural Development from Jiangxi Agricultural University in 2022, and her main research direction is forestry economy.

Ruolan Yuan received her master's degree in Agricultural economics and management from Jiangxi Agricultural University in 2021. She is a PhD candidate in Agricultural and forestry economic management at Jiangxi Agricultural University. Her current research interests are agricultural and forestry economic theory and policy.

Yaoqi Zhang is a professor at the School of Forestry and Wildlife Sciences, Auburn University, AL, USA, and his research interests include Sustainable forest management and rural development and Forest land use change.

Qian Deng is a PhD candidate from the School of Economics and Management at Jiangxi Agricultural University. She was awarded the Master's Degree by the University of York (UK) and now works as a lecturer at the School of Foreign Languages, Jiangxi Agricultural University. Her research areas include agricultural and forestry economics and management, focusing on rural society and governance.

References

Milošević

Nikolić

Milošević

, et al. Managing resources based on influential indicators for sustainable economic development: a case study in Serbia. Sustainability 2022; 14: 4795.

Costanza

De Groot

Sutton

, et al. Changes in the global value of ecosystem services. Glob Environ Change 2014; 26: 152–158.

Augustynczik

ALD

Asbeck

Basile

, et al. Diversification of forest management regimes secures tree microhabitats and bird abundance under climate change. Sci Total Environ 2019; 650: 2717–2730.

Janke

Finlay

Hobbie

. Trees and streets as drivers of urban stormwater nutrient pollution. Environ Sci Technol 2017; 51: 9569–9579.

Kicklighter

Melillo

, et al. A contemporary carbon balance for the northeast region of the United States. Environ Sci Technol 2013; 47: 13230–13238.

Rodrigues

Jacobi

Figueira

JEC

. Wildfires and their impact on the water supply of a large neotropical metropolis: a simulation approach. Sci Total Environ 2019; 651: 1261–1271.

Schaubroeck

Alvarenga

Verheyen

, et al. Quantifying the environmental impact of an integrated human/industrial-natural system using life cycle assessment; a case study on a forest and wood processing chain. Environ Sci Technol 2013; 47: 13578–13586.

Chang

Zou

Yoshino

, et al. Ecological policy benefit valuation based on public feedback: forest ecosystem services in Wuyishan nature reserve, China. Sci Total Environ 2019; 673: 622–630.

Niu

Wang

Liu

, et al. Economical assessment of forest ecosystem services in China: characteristics and implications. Ecological Complexity 2012; 11: 1–11.

10.

Cao

Suo

Xia

, et al. Net value of forest ecosystem services in China. Ecol Eng 2020; 142: 105645.

11.

Zheng

Robinson

Liang

, et al. Benefits, costs, and livelihood implications of a regional payment for ecosystem service program. Proc Natl Acad Sci U S A 2013; 110: 16681–16686.

12.

Engel

Pagiola

Wunder

. Designing payments for environmental services in theory and practice: an overview of the issues. Ecol Econ 2008; 65: 663–674.

13.

Bennett

Gong

Scarpa

. Hungry birds and angry farmers: using choice experiments to assess “eco-compensation” for coastal wetlands protection in China. Ecol Econ 2018; 154: 71–87.

14.

Dasgupta

Hamilton

Pagiola

, et al. Environmental economics at the world bank. Rev Environ Econom Policy 2020; 2: 4–25.

15.

Wunder

Engel

Pagiola

. Taking stock: lessons learnt for the design of payment for environmental sciences programs. Ecol Econ 2008; 65: 834–852.

16.

Vincent

. Microeconomic analysis of innovative environmental programs in developing countries. Rev Environ Econ Policy 2020; 4: 221–233.

17.

Ouyang

Zheng

Xiao

, et al. Improvements in ecosystem services from investments in natural capital. Science 2016; 352: 1455–1459.

18.

Salzman

Bennett

Carroll

, et al. The global status and trends of payments for ecosystem services. Nat Sustain 2018; 1: 136–144.

19.

Maron

Gordon

Mackey

, et al. Conservation: stop misuse of biodiversity offsets. Nature 2015; 523: 401–403.

20.

Sun

Yang

, et al. Ecological compensation estimation of soil and water conservation based on cost-benefit analysis. Water Resour Manage 2013; 27: 2709–2727.

21.

Xie

Cheng

. Factors influencing farmer willingness to fallow winter wheat and ecological compensation standards in a groundwater funnel area in Hengshui, Hebei Province, China. Sustainability 2017; 9: 839.

22.

Liu

Yang

, et al. A dynamic eco-compensation standard for Hani Rice Terraces System in southwest China. Ecosyst Serv 2019; 36: 100897.

23.

Gao

Shen

, et al. An evolutionary game analysis of governments’ decision-making behaviors and factors influencing watershed ecological compensation in China. J Environ Manage 2019; 251: 109592.

24.

Arrow

. Valuing environmental preferences: theory and practice of the contingent valuation method in the US, EU, and developing countries. New York, USA: Oxford University Press on Demand, 2001.

25.

Whittington

Pagiola

. Using contingent valuation in the design of payments for environmental services mechanisms: a review and assessment. World Bank Res Obs 2012; 27: 261–287.

26.

Peng

Zhang

Yin

, et al. Study of impact factors of willingness to pay regarding water reserve of South-to-North Water Diversion Project in Beijing based on Bayesian network model. J Clean Prod 2018; 184: 569–578.

27.

Xiong

Kong

. The analysis of farmers’ willingness to accept and its influencing factors for ecological compensation of Poyang Lake wetland. Procedia Eng 2017; 174: 835–842.

28.

Panwanitdumrong

Chen

. Are tourists willing to pay for a marine litter-free coastal attraction to achieve tourism sustainability? Case study of Libong Island, Thailand. Sustainability 2022; 14: 4808.

29.

Sejati

Arifien

Maad

. Economic valuation of rice agricultural land in Bogor regency. J Phys Conf Ser 2020; 1517: 12024.

30.

Vondolia

Kusi

King

, et al. Valuing intangible cultural heritage in developing countries. Sustainability 2022; 14: 4484.

31.

Mutandwa

Grala

Petrolia

Estimates of willingness to accept compensation to manage pine stands for ecosystem services. For Policy Econ 2019; 102: 75–85.

32.

Zhang

Wang

, et al. A scientometric review of emerging trends and new developments in agricultural ecological compensation. Environmental Science and Pollution Research 2018; 25: 16522–16532.

33.

Hanemann

. Discrete continuous models of consumer demand. Econometrica: J Econometric Soc 1984; 52(3): 541–561.

34.

Diener

O’Brien

Gafni

. Health care contingent valuation studies: a review and classification of the literature. Health Econ 1998; 7: 313–326.

35.

Henderson

. Consumer's surplus and the compensating variation. Rev Econ Stud 1941; 8: 117–121.

36.

Van Exel

NJA

Brouwer

WBF

van den Berg

, et al. With a little help from an anchor: discussion and evidence of anchoring effects in contingent valuation. J Socio Econ 2006; 35: 836–853.

37.

Rotteveel

Lambooij

Zuithoff

NPA

, et al. Valuing healthcare goods and services: a systematic review and meta-analysis on the WTA-WTP disparity. PharmacoEconomics 2020; 38: 443–458.

38.

Feng

Liang

, et al. Factors influencing willingness to accept in the paddy land-to-dry land program based on contingent value method. J Cleaner Prod 2018; 183: 392–402.

39.

Baffoe

Mizunoya

Yabar

. Determinants of rice farmers’ willingness to pay for conservation and sustainable use of Swampy Wetlands in Ghana’s Northern and Ashanti regions. Agriculture 2021; 11: 507.

40.

Venkatachalam

. The contingent valuation method: a review. Environ Impact Assess Rev 2003; 24: 89–124.

41.

Eskandari-Damaneh

Noroozi

Ghoochani

, et al. Evaluating rural participation in wetland management: a contingent valuation analysis of the set-aside policy in Iran. Sci Total Environ 2020; 747: 141127.

42.

Yin

Pizzol

Jacobsen

, et al. Contingent valuation of health and mood impacts of PM2.5 in Beijing, China. Sci Total Environ 2018; 630: 1269–1282.

43.

Truong

. Estimating residents’ willingness to pay for wetland conservation using contingent valuation: the case of Van Long Ramsar Protected Area, Vietnam. Biodiversitas: Journal of Biological Diversity 2021; 22: 4784–4793.

44.

Abebe

Dagnew

Zeleke

, et al. Willingness to pay for watershed management. Resources 2019; 8: 77.

45.

Mueller

Soder

Springer

. Valuing attributes of forest restoration in a semi-arid watershed. Landsc Urban Plann 2019; 184: 78–87.

46.

Shi

, et al. Effects of social interactions and information bias on the willingness to pay for transboundary basin ecosystem services. J Environ Manag 2021; 296: 113233.

47.

Aydoğdu

Sevinç

Cançelik

, et al. Determination of farmers’ willingness to pay for sustainable agricultural land use in the GAP-Harran Plain of Turkey. Land (Basel) 2020; 9: 261.

48.

Nguyen

Knetsch

Mahasuweerachai

. WTP or WTA: a means of determining the appropriate welfare measure of positive and negative changes when preferences are reference dependent. Environ Resour Econ 2021; 78: 615–633.

49.

Wang

Adamowski

Wang

, et al. Farmers’ willingness to accept compensation to maintain the benefits of urban forests. Forests 2019; 10: 691.

50.

Al-Assaf

Albalawneh

Majdalawi

, et al. Local communities’ willingness to accept compensation for sustainable ecosystem management in Wadi Araba, South of Jordan. Sustainability 2021; 13: 5190.

51.

Wang

Zhang

, et al. Environmental knowledge, risk attitude, and households’ willingness to accept compensation for the application of degradable agricultural mulch film: evidence from rural China. Sci Total Environ 2020; 744: 140616.

52.

Ren

Wimmer

, et al. Incentive mechanism for promoting farmers to plant green manure in China. J Cleaner Prod 2020; 267: 122197.

53.

McGurk

Hynes

Thorne

. Participation in agri-environmental schemes: a contingent valuation study of farmers in Ireland. J Environ Manag 2020; 262: 110243.

54.

Heckman

. Sample selection bias as a specification error. Econometrica: Journal of the Econometric Society 1979; 47(1): 153–161.

55.

Greene

. Sample selection bias as a specification error: a comment. Econometrica: Journal of the Econometric Society 1981; 49: 795–798.

56.

Kriström

. Spike models in contingent valuation: theory and illustrations. Arbetsrapport-Sveriges Lantbruksuniversitet, Institutionen Foer Skogsekonomi (Sweden) 1995; 79: 1013–1023.

57.

Lotfinasabasl

Khosroshahi

Saeedifar

, et al. Precipitation trend analysis and drought assessment of Jazmorian basin using trend analysis methods and optimized drought indices. Iranian Journal of Range and Desert Research 2019; 25.

58.

Chu

Zhan

Wang

, et al. Households’ willingness to accept improved ecosystem services and influencing factors: application of contingent valuation method in Bashang Plateau, Hebei Province, China. J Environ Manag 2020; 255: 109925.

Financial compensation for natural forest logging ban: Standard calculation based on willingness to accept

Abstract

Keywords

Introduction

Data collection

Methodology

Data collection

Calculation method

Models

Variable selection

Empirical results

Measurement results of farmers’ compensation

Influencing factors of farmers’ willingness and level of compensation

Farmers’ willingness to accept payment: Parameter estimation

QR analysis

Conclusion and discussion

Footnotes

Author statement

Declaration of conflicting interests

Funding

ORCID iD

Author biographies

References