Abstract
The productivity of the agricultural sector depends on natural resources, and farmers are vulnerable to substantial weather risks globally. For many years, different insurance products have been promoted in China to manage the negative impact and risk associated with unpredictable natural events. Generally, different insurance products correspond to the various premium levels, and farmers are unaware of the insurance policies and premium calculations. This study focuses on farmers’ preferences toward the insurance of corn crop, one of China’s largest grain crop productions. This study uses Heckman’s two-stage model to focus on the farmers’ willingness to pay for multiple weather-based insurances, who suffered severe disasters. The contingent valuation method was used to collect the data about farmers’ willingness to pay and preferences from 252 households across three cities of Inner Mongolia, China. The result of empirical estimation indicates that farmers with less experience and high income tend to choose “60% of full-cost insurance product.” Farmers with a lack of specialization and lower diversified planting tend to choose “full-cost insurance product.” In contrast, farmers with higher education prefer “output value insurance product.” The empirical result indicates that the promotion of various agricultural insurance could be a valuable strategy to improve protection levels and manage the risk adjustment input in the agricultural sector. Therefore, this study provides insights into risk reduction design that suggests adopting and encouraging different agricultural insurance products.
1. Introduction
The agricultural sector is the primary source of food production for developing economies [1]. China’s agricultural industry has changed dramatically since the late 1970s. It grew at about 5 percent annually in the past three decades. While significant growth has occurred in almost all cropping sectors, the production of some crops has grown more rapidly [2]. Hence, crop structure has changed, diversifying from staple grains into higher-valued crops [3]. China has a vast territory and complex landscape. Natural disasters frequently occur, spread widely, and can cause sudden losses resulting in high volatile returns, especially in case of heavy drought, which seriously hinders the sustainable development of China’s agriculture and the stability of rural society. Extremely low temperatures and heavy drought during peak winter in southern China cause massive losses. Since the agricultural sector is highly dependent on the climate changes such as temperature, precipitation, and terrain, its development has faced various difficulties and challenges. Climate hazards increase the possibility of weather disasters for agriculture. China’s agricultural sector is prone to various losses due to weather, technological, and market uncertainties, and such losses arising from these risks have to be borne mainly by farmers. However, landholdings are so small that farming activities alone cannot adjust to uncertainties, especially the risks from extreme weather events and variability in climate, which are a great challenge and the most conspicuous trend in production for small farm householders. Risk management from climatic stresses is vital in promoting growth and protecting the agricultural sector [4]. Indeed, agriculture insurance has received considerable attention from the government, industry, academia, and the public. It is a valuable risk management tool that can prevent farmers’ loss and stabilize their income [5].
Agricultural insurance is an effective tool to promote agricultural development, increase farmers’ resilience, and help farmers survive and protect their assets. Crop insurance is considered a contingency risk management and an effective confining strategy to mitigate unexpected losses due to extreme weather events [6, 7]. To minimize the adverse impacts of risk in farming due to climate change and improve the disaster risk resistance of agriculture, the Chinese government officially launched agricultural insurance with the financial subsidy in 2007 [8]. The scale of China’s agricultural insurance has repeatedly hit new heights. The degree of risk protection that measures the core functions of agricultural insurance and the level of agricultural insurance protection (agricultural insurance amount/agricultural output value) have increased by six times, from 3.59% (2008) to 23.21% (2018) in the past 11 years. It has become an effective way to diversify and transfer agricultural risks [9], but generally the protection level is still low. China’s agricultural insurance protection level is 1/5 of the United States, 1/3 of Canada, and 1/3 of Japan [10]. China is far behind the developed countries. Therefore, four ministries and China’s commission jointly issued the “Guiding Opinions on Accelerating the High-Quality Development of Agricultural Insurance” (Goa) in 2019 to improve agricultural insurance’s ability and serve agriculture, countryside, and farmers. Expansion of the different types of agricultural insurance to meet the diverse needs of farmers is the core goal of “Goa.” Therefore, the development of different types of high-quality insurance products has become the focus of insurance industry.
Different insurance products with various protection levels indicate that farmers must pay different insurance premiums. A plethora review of literature pointed out that as protection level increases, the subsidy rate of fiscal funds should reduce [11]. Consequently, the farmers’ self-paid insurance premiums will increase. At present, China’s “bulk crop insurance scheme” only covers direct material costs. However, with an increase in price regarding input production, the protection level can no longer cover the direct material costs. In statistics issued by the “Compilation of National Agricultural Product Costs and Benefits,” the protection level in 2019 was 20 RMB/hm2 for corn insurance in Inner Mongolia Province, which is not high enough to cover the direct material cost (22 RMB/hm2). Thus, this leads to the following question: are farmers willing to pay different premiums related to the various agricultural insurance? When a premium is too high, farmers cannot pay for it, so agricultural insurance cannot survive in the market. If many farmers undertake WTP premiums, the demand will increase, and the insurance’s law of large numbers is satisfied, which will benefit the implementation of every agricultural insurance. On the other hand, there are indeed concerns that the blind implementation of new agricultural insurance products will not be effective enough, which can only get half the results for the twice effort. Therefore, to design and implement various types of agricultural insurances, it is indispensable to identify farmers’ WTP and analyze the influencing factors for different insurance products. Consequently, these steps will fulfill farmers’ needs, which will lead to the development of a stable and sustainable agricultural industry.
A plethora of literature on the WTP for different types of products reveals that most scholars adopted the protection ratio to increase the protection level [12]. Second, the key factors affecting the WTP for high-protection agricultural insurance are education level, risk aversion, disaster damage, planting scale, production specialization, income level, and risk aversion measures. Predecessors have extensive research content and rich results, but there is still research gap. In the context of the diversification and differentiation of China’s agricultural production entities, most scholars use the guaranteed ratio to increase the protection level, which cannot meet the risk protection needs of different production entities. According to the survey data from the Inner Mongolia Autonomous Region, China, the actual cost of corn in the eastern, central, and western regions of Inner Mongolia is 25.27 RMB/hm2, 30.27 RMB/hm2, and 33.33 RMB/hm2, respectively, with the maximum difference being 8.07 RMB/hm2. Suppose the protection level is designed according to different protection ratios, assuming that the insurance amount is 26.67 RMB/hm2, the protection ratios are 70% and 90%. In this case, the actual protection level is 18.67 RMB/hm2 and 24 RMB/hm2, respectively, and the difference between the two protection levels is 5.33 RMB/hm2. Compared with the actual cost, the difference of 8.07 RMB/hm2 cannot cover differentiated costs. Apart from that, it is challenging to meet the risk protection needs of farmers in different regions. Therefore, using different protection ratios to design the protection levels will reduce demand.
The primary purpose of this research is to help the local government, industry, and insurance companies decide on investing in new agricultural insurance products by providing consumers’ willingness to pay for agricultural insurance in Inner Mongolia, China. Therefore, this study tried to fill the gap by determining the impact factors that affect farmers’ WTP under different agricultural insurance products. Additionally, this study provides suggestions to improve the factors’ impact on farmers’ WTP and implement other agricultural insurance products. At the same time, to meet the risk protection needs of various entities, three different agricultural insurance products, 60% of total cost product, total cost product, and output product, were designed as a pilot project. During the study, corn growers are taken as the research example to obtain data on the actual willingness of farmers to pay. The contingent valuation method (CVM) was adopted to design the premium bid value. The Heckman two-stage model was employed to solve the problem of sample selection bias and analyze farmers’ WTP for corn insurance premiums and its influencing factors for different products.
The remaining study is arranged as follows: Section 2 discusses the detailed description of theoretical analysis, and this consistent part consists of the impact of the protection level, risk attitude of farmers, insurance awareness, operational status, and risk aversion on farmers’ wiliness to pay. Section 3 explores the data from the survey variable selection and explains the methodology used in the study. Section 4 provides the result of empirical analysis and discussion. Section 5 provides insight into the conclusion, discussion, and suggestion.
2. Theoretical Analysis
2.1. The Impact of Protection Level on Farmers’ WTP
In the early stages of agricultural insurance development in various countries globally, the protection level was relatively low, affecting farmers’ WTP [13–15]. China is also facing the same problem of a low protection level of agricultural insurance. Previous studies have pointed out that increasing protection levels can effectively increase farmers’ WTP [16–18]. A review of prior literature found that increasing protection level includes directly increasing the insurance amount and increasing the protection ratio, while the insurance amount remains unchanged. Most scholars suggest increasing the protection ratio to increase protection level [19–21]. While the optimal insurance theory assumes that farmers are risk averters and expect utility maximization, when a protection level is equal to their risk losses, farmers can maximize utility under actuarial fairs. It indicates that when the protection level is close to the size of risk loss, it is more beneficial to increase farmers’ WTP. Furthermore, only one protection level cannot meet the needs of various agricultural producers due to their priorities and differences. Therefore, increasing multiple protection levels can increase their WTP [16, 17].
2.2. Risk Attitudes Affecting Farmers’ WTP
According to the utility function theory’s hypothesis, the utility function is strictly quasi-concave, which means that the farmers are risk-averse, and they have demand for insurance currently. Both theory and empirical evidence have proved that when risk aversion is higher, the farmers are more willing to pay for agricultural insurance [22, 23]. Therefore, for different agricultural insurance products, other conditions are the same. Assuming that farmers’ expected risk loss is lower than the current protection level, as the protection level increases, the protection capability of insurance will increase. Besides, a stronger WTP means a higher risk aversion. Farmers with high-risk aversion tend to choose a high-protection product to spread risk as much as possible.
2.3. Effect of Insurance Awareness on Farmers’ WTP
Insurance awareness is people’s knowledge, opinions, thoughts, theories, and mentality about insurance. Insurance awareness and insurance demand are positively correlated; the higher the insurance awareness, the higher the demand [24]. Many scholars used education levels as a regressor to observe the insurance awareness of farmers. The findings show that the higher the education level, the stronger the insurance awareness of farmers who would use insurance to spread risk [15, 25]. Besides, the high-protection insurance product will disperse risk as much as possible to meet their risk protection needs. In other words, the higher the protection, the stronger the impact of the education level on farmers’ WTP.
2.4. Impact of Operation Status on Farmers’ WTP
Predecessors used planting area, income level, degree of production specialization, degree of disaster loss, and damage as factors to measure farmers’ production operation status and to analyze its impact on farmers’ WTP [19, 26, 27]. Then, for the high-protection product, the effect of operation status on farmers’ WTP is as follows: firstly, the farmers with high-income levels are more capable of paying a premium. Therefore, they have the stronger WTP for the high-protection product, which means a higher premium. Secondly, the high degree of specialization indicates that the farmers have a large scale of land areas, a high degree of mechanization, robust disaster prevention, loss reduction measures, and proper field management measures. Therefore, if the total investment of farmers increases, the expected risk loss would be significantly higher. Choosing a high-protection product can improve its utility level when expecting consequential risk loss [28]. In this case, higher specialization indicates a higher WTP for the high-protection product. Thirdly, according to the expected utility function theory, it is believed that disaster loss affects farmers’ insurance purchasing decision-making behavior, and agricultural insurance is a predisaster preventive measure [29]. Thus, the current disaster loss degree does not affect the recent insurance purchase decision, and the historical disaster loss degree will affect recent insurance purchasing decisions. Some studies found that the largest disaster in the past five years significantly affected farmers’ WTP [20, 30–32]. Farmers with the higher degree of historical disaster damage would have more significant expected risk loss, and they choose high-protection products, which can increase the utility level. Consequently, their WTP for agricultural insurance with the high-protection product would be more incredible.
2.5. Effect of Risk Aversion Measures on Farmers’ WTP
From the perspective of insurance demand theory, risk aversion measures substitute agricultural insurance products, and the demand for both is negatively correlated. Risk aversion measures are divided into preprevention and posttreatment. Preprevention measures include diversification of production, decentralizing arable land, nonagricultural income, and purchasing insurance products. At the same time, posttreatment involves the input of factors that can control agricultural production risks (such as pesticides and herbicides) and the use of formal or informal means of credit [15, 25, 33, 34]. Increasing the protection degree would effectively enhance covering losses and increase income. So, the alternative of risk aversion measures is relatively weak, indicating that farmers lack measures to cope with risks, so they are willing to pay agricultural insurance [35].
3. Materials and Methods
3.1. Data Sources and Descriptive Statistics
3.1.1. Data Sources
The Inner Mongolian region is the biggest and the main demonstrative grassland regional area. It is the 3rd largest Chinese subdivision, consisting of approximately 463000 square meters and 12% of China’s total land area. This region (Figure 1) was selected for the study because this is the central area of corn production in China, also known as the northern spring-sown corn area in China. Inner Mongolia ranked 5th largest of corn’s sown and production in China. The central corn-producing regions are in the eastern region of Inner Mongolia. The dataset used in this study was collected from the field survey from July to September 2017. Three cities in the east part of Inner Mongolia, including Hulunbuir, Xing’an League, and Tongliao, were selected, while two counties were randomly selected in each city, and a total of 252 valid questionnaires were obtained. The distribution of observations is shown in Table 1.
3.1.2. Descriptive Statistics
The survey results show that farmers between the ages of 41 and 50 most frequently occur in the sample size, with an average of 47 years. The average corn planting area per household is 4.5 hm2, while 49.2% of the household sown corn under 3.33 hm2, and 23% are planted over 6.67 hm2. The phenomenon of land transfer is significantly common, and the transfer area of dry land is more significant than irrigated land. The majority of the farmers’ education level is elementary. Additionally, there are more specialized farmers than part-time farmers. On average, these households have approximately four family members. About 70% of households have only 2 or fewer laborers work as labourers. The average household corn planting income accounts for 50% of the total agricultural income. The observations are consistent with the rural situation in Inner Mongolia, which is a clear indication of the valid and representative selection of observation (Table 2).
3.2. Variable Design and Descriptive Statistics
To measure the dependent variables, firstly, the questionnaire excluded the farmers who had never been contacted by the insurance company or did not understand agricultural insurance. Resultantly, they never purchased or did not purchase the agricultural insurance within the last three years. Secondly, this study did not include the farmers who purchased agricultural insurance but did not understand the contract’s terms and conditions. Thirdly, several products and premium bid values were fixed under the conditions that the existing premium subsidies and insurance rates remain unchanged. ① According to the statistical data of “Compilation of National Agricultural Product Costs and Benefits,” there are three products based on corn’s direct material, land, and labour cost in Inner Mongolia: first, 60% of the total cost product with the value of 43 RMB/hm2, including “direct material cost + land cost or labour cost”; second, total cost product with the value of 61 RMB/hm2, including “full production cost = direct material cost + land cost + labour cost”; and third, output product with its value of 69 yuan/hm2. ② The premium bid value was fixed using the contingent valuation method (CVM) in the double-boundary dichotomous method. The intermediate bid value was calculated based on the existing premium, which increased and decreased by 60%, respectively, under each product where five bid values were set (Table 3). On this basis, the following questions were asked from the interviewees: to judge the willingness of farmers to pay, “With this product, are they willing to buy the insurance? If they are willing to buy, how much are they willing to spend? (The two-way inquiry method was used to obtain the WTP level). If they are unwilling to buy, moved to the next product, and loop until the end of three products.” During the study, to eliminate farmers’ precautionary psychology, selecting the premium bid value is mentioned in the middle part of the questionnaire to obtain the actual level of farmers’ WTP.
This study selected three aspects and 11 independent variables to measure explanatory variables. The measurement and descriptive statistics of specific variables are shown in Table 4.
4. Empirical Analysis
4.1. The Heckman Two-Stage Model
There might be biasedness if the OLS method is used to estimate the sample selection [36]. Therefore, this study used the Heckman two-stage model to avoid such bias. Based on the expected utility function theory, referring to the insurance demand model of Rothschild [37], the following model is constructed based on the Heckman two-stage model. When farmers purchase insurance, their income is unobservable; in other words, the net income is a latent variable, and thus, is used as a proxy variable to measure:where refers that the farmer has purchased one type of corn insurance product, ; refers that the farmer has not purchased any products, . refers to an explanatory variable group, including , , , and ; indicates a variable of individual characteristics; represents a variable of production and operation status; is a variable of risk aversion measures; shows the degree of risk attitude; means the parameter to be estimated; and indicates the random disturbance item.
4.1.1. The First-Stage Model
In the first stage, the probit model is used, assuming that follows the standard normal distribution, and the model is constructed with “whether to purchase this type of agricultural insurance” as the dependent variable.
Among them, is the probability of farmers to buy insurance, and are the probability density function of the standard normal distribution and the corresponding cumulative distribution function, respectively.
4.1.2. The Second-Stage Model
The second-stage model took “the highest level of farmers’ WTP agricultural insurance premiums under the product” as the dependent variable and used the inverse Mills ratio , which is obtained in the first stage. Then, the calculation formula of is calculated using the following equation:
It is necessary to select identification variables to identify the equation. According to previous research, risk attitudes can be used as effective identification variables for insurance decision-making. Therefore, risk attitude is chosen as the identification variable in this study and only put in the first-stage model. The modified variable is introduced into the second-stage model to correct the selective bias and to build the second-stage model.
Among them, is the natural logarithm concerning the WTP of the i farmer; is the parameter to be estimated, and is the error term. Besides, OLS is employed to estimate the above equation. If is significant, it is indicated that the sample is featured with a selection bias problem, and it is suitable to use the Heckman two-stage model.
The empirical results of the Heckman two-stage model for three products show that the is significant at the statistical levels of 10%, 1%, and 10%, respectively, indicating that the Heckman two-stage model can correct the sample selection bias problem.
4.2. Analysis on the Impact of Three High-Protection Products on Farmers’ Participation Rate
Among the three agricultural insurance products, the participation rate of farmers is declining for the higher-protection product. In particular, the participation rate of 60% full-cost product was the highest (82%); the participation rates of full-cost and output value products were 76% and 71%, respectively (Table 5). Compared with China’s crop insurance participation rate in 2016 (85.15%), the participation rate of three agricultural insurance products is low but higher than the lower limit (50%) of the insurance company’s underwriting base. It is concluded that farmers demand different high-protection insurance products, but the participation rate is not positively correlated.
4.3. Comparative Analysis on Influencing Factors of Farmers’ WTP under Three Products
Age, years of farming, risk attitude, planting area, and net household income significantly affect farmers’ WTP: age, planting area, risk aversion, and household net income increase the farmers’ WTP level. Additionally, the farmers with more years of farming experience are less willing to pay for all kinds of insurance (Table 6). (1) By comparing two covered cost products, estimated results indicate that the degree of production specialization and the level of diversified planting are significantly affecting the WTP for the total cost product. While the impact on 60% of the total cost product was not significant, which indicates that the lower specialization, a single field management measure, has the fewer disaster prevention and loss reduction measures. Besides, farmers with a lower diversified planting level and a single risk aversion measure were willing to use total cost product. (2) Comparing total cost product and output product, estimation results revealed that the education level of farmers has a significant impact on the WTP for output product, but the impact on the WTP for cost product is not significant. Farmers with high education levels were willing to use high insurance premiums to protect their output value, which revealed that the higher the education level of farmers, the stronger the awareness of insurance, the greater the acceptance of the higher-protection product, and the stronger the WTP.
4.4. Results of Farmers’ WTP Level under Three Products
This study adopts three methods to calculate the farmers’ maximum premium WTP level (Table 7). With the continuous improvement of calculation methods, the WTP level decreased, indicating the sample mean, and the sample expected value may overestimate the farmers’ WTP level. The results of Heckman’s simulation are as follows: farmers’ WTP level of 60% of cost product, total cost product, and output product is 0.63 RMB/hm2, 0.79 RMB/hm2, and 0.88 RMB/hm2, respectively.
5. Results, Discussion, and Recommendations
This study concluded that farmers have different demands for three different agricultural insurance products based on the empirical analysis. Farmers with low specialization and diversified planting levels have a strong WTP for the total cost product. By comparing cost product and output product, the education level of farmers significantly affects the WTP for the output product. Therefore, companies and governments should provide different products for farmers with their characteristics. Evidence from the survey highlighted that the farmers with their features have different preferences for various products. Different insurance products have their target groups. To increase the needs of farmers: 60% of total cost product is positioned at farmers with low scale and risk averse. Farmers with low production specialization and diversified planting levels are considered as a target group of the total cost product. Farmers with a large scale of operation and high school education level and above are considered as a target group of the output insurance product.
Agricultural insurance serves agriculture, the countryside, and farmers. Besides the increasing demand for insurance products, other essential factors determine whether insurance implementation is exemplary or not. Agricultural modernization development and insurance rating are crucial prerequisites for implementing high-protection agricultural insurance products. The signs of agricultural modernization development are the efficiency of production factors, improved quality, transformed operation methods, increased agricultural output, and rural economic development [38]. Farmers with modern agricultural techniques, high education level, a strong knowledge of insurance, and high income can pay premiums for high-protection agricultural insurance products. Focus on these factors would increase the demand for high-protection agricultural insurance products.
High-protection insurance should implement gradually. From the perspective of farmers’ needs, studies have shown that farmers demand the full-cost product, while, from the standpoint of financial supporting capabilities, the high-protection product will inevitably increase the total amount of premium subsidies. Therefore, finance supporting capacity is crucial in ensuring high-protection agricultural insurance implementation. With the financial support, acceptance, and recognition of local governments, the insurance industry can effectively implement this insurance product. Therefore, it is concluded that the first total cost insurance product should be carried out as a pilot project and gradually shift to the output insurance product.
This study also offers some limitations. Since this study tries to attempt farmers’ different characteristics for different insurance products, this does not imply the impact of government policy and corporate sector services on farmers’ WTP in those years. There was almost certainly room for improvement for considering other characteristics of the farmer that are not being considered due to the limited data availability. Further studies are encouraged for policy management.
Data Availability
The data presented in this study are available on request from the corresponding author.
Conflicts of Interest
The authors declare that they have no conflicts of interest.
Acknowledgments
This research work was supported by the National Natural Science Foundation of China (grant no. 72173069), the National Natural Science Foundation of China (grant no. 71863028), the National Natural Science Foundation of China (grant no. 72163026), and the Inner Mongolia Science and Technology Project (grant no. 201802108).