Can Agricultural Support and Protection Subsidy Policies Promote High-Quality Development of Grain Industry? A Case Study of China

Abstract

The issue of grain quality has attracted increasing attention with the gradual growth and stabilization of grain output. We take the successive agricultural support and protection subsidy policies introduced in 2015 as a quasi-natural experiment and use a multi-period double-difference model to analyze a panel of data from 298 prefecture-level cities in China from 2007 to 2020. Our findings confirm that implementing agricultural support protection subsidy policies has had initial success regarding grain output growth and quality improvement at the point of contradiction. This success is also due to its scale and technology effects, which can ensure that grain output and quality have achieved growth. The level of agricultural machinery also plays a crucial positive role in the impact of the subsidy policy on food production and food security. Further heterogeneity analysis shows that the impacts of agricultural support and protection subsidy policies on food quality and yield security differ significantly across different geographic regions, food crop production, and pollutant type levels. Therefore, the positive impact of subsidy policies on the food industry should be better promoted to enhance the sustainability and competitiveness of agriculture.

1. Introduction

Implementing an agricultural support and protection subsidy policy is of great practical significance for the high-quality development of the food industry [1]. China’s “four subsidies” policy included comprehensive income and production subsidies. Among them, the comprehensive income subsidies consisted of direct grain and comprehensive agricultural subsidies. The State’s comprehensive income subsidy policy aims to increase farmers’ incentive to produce and boost food production more directly. However, some debate in the academic community continues as to whether this policy has been effective in increasing food production and contributing to crop income [2,3]. Production-specific subsidies include two aspects: price support and a production material subsidy policy. Later, with the development of the food industry and the need for economic and social development, China combined the crop seed subsidies, direct subsidies for food farmers, and comprehensive agricultural subsidies into the original “four subsidies” policy and introduced a series of policies and measures to promote them. This development shows that China’s agricultural support subsidy policy has gradually been extended from pilot cities to the whole country. China’s food subsidy policy system is also changing gradually with the times. The direct or indirect support the government provides through the subsidy policy helps raise farmers’ income levels. Such economic incentives may encourage farmers to engage in food production more actively and promote the development of agriculture [4].

Moreover, subsidy policies can stabilize the price of agricultural products and improve the efficiency of food production, enhancing the agricultural industry’s technological level and competitiveness [5]. However, agricultural support and protection subsidy policies may also have negative effects. For example, excessive subsidies may lead to market price distortion, affecting the normal operation of the market, which is not conducive to the effective allocation of resources [6,7]. Therefore, when implementing the agricultural support and protection subsidy policy, the interests of all parties must be balanced, and the policy design must be optimized to ensure its positive impact on the food industry is maximized. This paper conducts an in-depth discussion of the impact of the agricultural support protection subsidy policy on grain quality and yield and takes the relevant policies introduced successively in 2015 as a quasi-natural experiment to verify the impact of this policy on achieving the high-quality development of the grain industry and the specific mechanisms of action.

Compared with previous research, the possible marginal contributions of this paper are mainly in the following aspects. First, from the research perspective, this paper focuses on analyzing the impact of agricultural support and the protection subsidy policy on the high-quality development of the food industry. It focuses on analyzing the significance of subsidy policy implementation on agricultural development. Second, regarding research content, the high-quality development of the food industry is subdivided into the aspects of food production and quality, and the impact of subsidy policy on food production and quality is analyzed. Through the above analysis, we can verify whether the agricultural support and protection subsidy policy achieves the “win-win” development of food production and quality. We also analyze the subsidy policy’s impact on food production quality and safety and consider the role of the important factor of agricultural mechanization production. This analysis is conducive to improving the efficiency of grain production and promoting the application of mechanized production in agricultural production.

This study bridges the gap of existing research based on the consideration of the real situation; for a long time, China’s related food problems have been mainly focused on the production aspect. In recent years, with the gradual alleviation of the food production problem, some scholars have begun to pay attention to the environment and quality of food production [8]; relatively speaking, the research on food production and quality is still relatively small. This paper tries to incorporate both food production and quality into an analytical framework to analyze the implementation of agricultural support and protection policies that can effectively promote the high-quality development of the food industry. At the same time, unlike the provincial data used in similar studies, the prefecture-level city data selected in this study is also more detailed, which makes the study convincing.

Section 2 of this paper is the literature review and research hypotheses. Section 3 is the modeling and data description, Section 4 is the analysis of empirical results, and Section 5 is the conclusion of the study.

2. Literature Review and Research Hypotheses

Many countries today have generally adopted food subsidy policies to protect and support food production to promote farmers’ income and guarantee national food security, and China is no exception [9,10]. Therefore, since the gradual implementation of the “four subsidies”-based agricultural subsidy policies in 2004, the impact of agricultural subsidies on food production has become the focus of attention for many scholars. After implementing the “four subsidies,” many scholars have explored the policy effects of subsidy policies on food security and the efficiency of funds from different perspectives around the “four subsidies” or direct subsidy policies in agriculture. The main research focuses on the effects of the “four subsidies” on increasing farmers’ incentives to grow food and promote and apply new and improved technologies [11,12]. Some studies have analyzed the impact of subsidy policies on the enhancement of comprehensive grain production capacity, market competitiveness, and the price of agricultural capital and labor [13,14], as well as on the willingness of farmers to transfer their farmland [15]. In addition, numerous studies have been conducted on the efficiency of food subsidy policies [16,17].

However, some scholars have argued that the effect of subsidy policies in stimulating farmers to increase grain production by expanding the area under cultivation is only apparent in the initial period of policy implementation and then gradually diminishes or even disappears [18,19]. This finding is strongly supported by the results of China’s financial support policies in the past 20 years. Therefore, we can find that any subsidy policy has adaptability and limitations under certain times and spaces.

Based on this reality, in 2016, China launched the comprehensive reform of three subsidies, merged the three subsidies into the agricultural support protection subsidy, and began to subsidize the main body of the grain moderate-scale operation to achieve the original intention of agricultural subsidies in promoting grain production and guaranteeing national food security. Some scholars have examined the impact of agricultural support protection subsidies on grain cultivation by large-scale farmers using the latest National Rural Fixed Observation Point Survey 2016 and 2017 data. The results of the study show that the agricultural support protection subsidy policy has a significant impact on expanding the grain sowing area of large-scale farmers [20]. The effect is manifested in the expansion of food production by promoting the transfer of more land by large-scale farmers, which has no significant effect on the planting structure of farmers. Some scholars use the triple-difference method to find that the subsidies for large-scale operations have a two-sided effect, increasing grain yield, labor productivity, grain income, and grain profits but causing large-scale operators to use more fertilizers and pesticides due to the subsidies easing their credit constraints and increasing their incentives to invest in factors [21]. Contrary to this view, some studies have argued that scaling up helps to reduce fertilizer and pesticide application [22,23,24].

The effects of the agricultural support and protection subsidy policy show that the policy positively incentivizes the physical capital input behavior of large rice farmers. When viewed through the lens of differences in resource endowments, subsidies for moderate-scale operations are more likely to incentivize young farmers to invest in agricultural production [25]. In addition, higher levels of education can lead to a greater positive effect of agricultural support and protection subsidies. From the perspective of differences in the scale of operation, the positive incentive effect of the moderate-scale operation subsidy is strongest for medium-sized, large rice farmers [26].

Further dissection of the agricultural support and protection subsidy policy reveals that the objectives of the policy have two main aspects. The first is to promote large-scale grain cultivation and operation and increase grain production. The second is to strengthen the protection of arable land fertility and ecological environment, reduce the use of pesticides and chemical fertilizers, achieve the green and healthy development of food, and improve food quality [27,28]. Indeed, food production and quality have always been difficult to trade-off. Based on national conditions and their planting interests, farmers usually choose to increase the amount of chemical fertilizer and sacrifice grain quality to improve yield. Thus, the question of whether the agricultural support protection subsidy policy can achieve a “win-win” of grain yield and quality is the concern of this paper. Whether the implementation of an agricultural support protection subsidy policy can reduce environmental pollution in food production, improve the quality of food production, ensure food quality and safety, improve food production, and ensure the safety of food production is the question that needs to be explored.

Accordingly, the hypotheses proposed in this paper are as follows:

H1. 

The implementation of an agricultural support protection subsidy policy can reduce environmental pollution in the process of food production, improve the quality of food production, and guarantee the quality and safety of food in the production chain.

H2. 

Implementing agricultural support and protection subsidy policy can help increase food production and guarantee food production safety.

H3. 

Scientific and technological progress (agricultural machinery level) is an important mechanism for the impact of agricultural insurance financial subsidy policy on food production and security.

3. Modeling and Data Description

3.1. Double-Difference Modeling

Program assessment, implementation assessment, and effect assessment are three important links in the policy evaluation process. First, program assessment is the process of comparing and evaluating different policy options. Through this process, the best implementation program can be identified to provide a basis for the smooth implementation of the policy. Second, implementation assessment is key to ensuring the policy is implemented. By assessing whether the policy and process implementation meet the design requirements, problems can be identified, and adjustments can be made promptly to ensure the effective implementation of the policy. Finally, effect assessment is the evaluation of the degree of impact of the policy. Quantitatively analyzing the impact of the policy can provide decision-making references for policymakers; it can also improve and perfect the policy.

While the multi-period double-difference model is a powerful tool for causal inference, it has some limitations. First, the model requires that the treatment effect be stable, i.e., that the effect of the treatment on the outcome is consistent across time points. Second, it relies on the parallel trend assumption, i.e., in the absence of treatment, the trends of outcomes in the treatment and control groups should be parallel. In addition, the multiperiod double-difference model may be affected by time-invariant omitted variables, which may lead to biased estimates if they affect both treatment assignment and outcomes. Finally, the model may have difficulty dealing with dynamic treatment effects, i.e., treatment effects that vary over time.

In this study, the multi-period double-difference (multi-period DID) method is used to study the effect of the agricultural support protection subsidy policy, mainly because of the time difference in the implementation of the agricultural support protection subsidy policy in different provinces, which is piloted by some provinces first, and then gradually covered all the provinces in the country. Therefore, this method can estimate the policy effect more accurately. In this study, the multi-period DID is combined with the policy implementation time characteristics. Traditional DID assumes that all individuals in the treatment group begin to experience policy shocks at the same time. Multi-period DID is used in cases where individuals in the treatment group do not receive the treatment at the same time. The model constructed for multi-period DID is as follows:

$${\mathrm{Y}}_{\mathrm{i}\mathrm{t}}=\mathsf{\alpha }+{\mathsf{\mu }}_{\mathfrak{i}}+{\mathsf{\lambda }}_{\mathrm{t}}+{\mathsf{\theta }\mathrm{t}\mathrm{r}\mathrm{e}\mathrm{a}\mathrm{t}}_{\mathrm{i}}\times {\mathrm{p}\mathrm{o}\mathrm{s}\mathrm{t}}_{\mathrm{t}}+\mathsf{\beta }{\mathrm{X}}_{\mathrm{i}\mathrm{t}}+{\mathsf{ϵ}}_{\mathrm{i}\mathrm{t}}$$

(1)

In the equation, $\mathsf{\alpha }$ is a constant term, ${\mathsf{\mu }}_{\mathfrak{i}}$ is an individual fixed-effects variable, ${\mathsf{\lambda }}_{\mathrm{t}}$ is a time fixed-effects variable, ${\mathsf{\theta }\mathrm{t}\mathrm{r}\mathrm{e}\mathrm{a}\mathrm{t}}_{\mathrm{i}}\times {\mathrm{p}\mathrm{o}\mathrm{s}\mathrm{t}}_{\mathrm{t}}$ denotes the core dummy variable, and ${\mathrm{X}}_{\mathrm{i}\mathrm{t}}$ is the control variable in the model. In the model, ${\mathrm{p}\mathrm{o}\mathrm{s}\mathrm{t}}_{\mathrm{t}}$ is used to represent the time points in the treatment group that changed according to the individual i. The average treatment effect due to the policy can be obtained from the model constructed by the multi-period DID. Based on the model constructed by the multi-period DID, the average treatment effect brought by the policy can be obtained:

$$\begin{gathered} \mathsf{\theta }=\left\{\mathrm{E}\right.[{\mathcal{Y}}_{\mathrm{i}\mathrm{t}}\left|{\mathrm{t}\mathrm{r}\mathrm{e}\mathrm{a}\mathrm{t}}_{\mathrm{i}}\right.=1,{\mathrm{p}\mathrm{o}\mathrm{s}\mathrm{t}}_{\mathrm{t}}=1]-\mathrm{E}[{\mathcal{Y}}_{\mathrm{i}\mathrm{t}}\left|{\mathrm{t}\mathrm{r}\mathrm{e}\mathrm{a}\mathrm{t}}_{\mathrm{i}}\right.=1,{\mathrm{p}\mathrm{o}\mathrm{s}\mathrm{t}}_{\mathrm{t}}=0]\} \\ -\left\{\mathrm{E}\right.[{\mathcal{Y}}_{\mathrm{i}\mathrm{t}}\left|{\mathrm{t}\mathrm{r}\mathrm{e}\mathrm{a}\mathrm{t}}_{\mathrm{i}}\right.=0,{\mathrm{p}\mathrm{o}\mathrm{s}\mathrm{t}}_{\mathrm{t}}=1]-\mathrm{E}[{\mathcal{Y}}_{\mathrm{i}\mathrm{t}}\left|{\mathrm{t}\mathrm{r}\mathrm{e}\mathrm{a}\mathrm{t}}_{\mathrm{i}}\right.=0,{\mathrm{p}\mathrm{o}\mathrm{s}\mathrm{t}}_{\mathrm{t}}=0]\} \\ =\left({\mathrm{Y}}_1-{\mathrm{Y}}_0\right)-\left({\mathrm{C}}_1-{\mathrm{C}}_0\right)=(\mathsf{\theta }+{\mathsf{\lambda }}_{\mathrm{t}})-{\mathsf{\lambda }}_{\mathrm{t}} \\ {=(\mathrm{Y}}_1-{\mathrm{C}}_1)-({\mathrm{Y}}_0-{\mathrm{C}}_0)=(\mathsf{\theta }+{\mathsf{\mu }}_{\mathrm{i}})-{\mathsf{\mu }}_{\mathrm{i}} \end{gathered}$$

(2)

In the process of model utilization, the interaction term is usually replaced by $D_{it}$, which represents the dummy variable for individual i’s treatment in time t. Then, the constructed model can be written as follows:

$${\mathcal{Y}}_{\mathrm{i}\mathrm{t}}=\mathsf{\alpha }+{\mathsf{\mu }}_{\mathfrak{i}}+{\mathsf{\lambda }}_{\mathrm{t}}+{\mathsf{\theta }\mathrm{D}}_{\mathrm{i}\mathrm{t}}+\mathsf{\beta }{\mathrm{X}}_{\mathrm{i}\mathrm{t}}+{\mathsf{ϵ}}_{\mathrm{i}\mathrm{t}}$$

(3)

Similarly, the average treatment effect of the model can be expressed as follows:

$$\begin{gathered} \mathsf{\theta }=\left\{\mathrm{E}\right.[{\mathcal{Y}}_1\left|{\mathrm{D}}_{\mathrm{i}\mathrm{t}}\right.=1]-\mathrm{E}[{\mathcal{Y}}_1\left|{\mathrm{D}}_{\mathrm{i}\mathrm{t}}\right.=0]\} \\ -\left\{\mathrm{E}\right.[{\mathcal{Y}}_0\left|{\mathrm{D}}_{\mathrm{i}\mathrm{t}}\right.=1]-\mathrm{E}[{\mathcal{Y}}_0\left|{\mathrm{D}}_{\mathrm{i}\mathrm{t}}\right.=0]\} \\ {=(\mathrm{Y}}_{\mathrm{a}\mathrm{f}\mathrm{t}\mathrm{e}\mathrm{r}}-{\mathrm{Y}}_{\mathrm{b}\mathrm{e}\mathrm{f}\mathrm{o}\mathrm{r}\mathrm{e}})-({\mathrm{C}}_{\mathrm{a}\mathrm{f}\mathrm{t}\mathrm{e}\mathrm{r}}-{\mathrm{C}}_{\mathrm{b}\mathrm{e}\mathrm{f}\mathrm{o}\mathrm{r}\mathrm{e}})=(\mathsf{\theta }+{\mathsf{\lambda }}_{\mathrm{t}})-{\mathsf{\lambda }}_{\mathrm{t}} \\ =\left({\mathrm{Y}}_{\mathrm{a}\mathrm{f}\mathrm{t}\mathrm{e}\mathrm{r}}-{\mathrm{C}}_{\mathrm{b}\mathrm{e}\mathrm{f}\mathrm{o}\mathrm{r}\mathrm{e}}\right)-\left({\mathrm{Y}}_{\mathrm{b}\mathrm{e}\mathrm{f}\mathrm{o}\mathrm{r}\mathrm{e}}-{\mathrm{C}}_{\mathrm{b}\mathrm{e}\mathrm{f}\mathrm{o}\mathrm{r}\mathrm{e}}\right)=(\mathsf{\theta }+{\mathsf{\mu }}_{\mathrm{i}})-{\mathsf{\mu }}_{\mathrm{i}} \end{gathered}$$

(4)

In the above multi-period DID model, ${\mathcal{Y}}_{\mathrm{i}\mathrm{t}}$ denotes food security, which is mainly analyzed through quantitative and qualitative security dimensions. In terms of quantitative security, “grain output (10,000 tonnes)” is used as a measure. In terms of qualitative security, “pesticide application (10,000 tonnes), fertilizer in terms of quality security, “pesticide application (10,000 tonnes), and chemical fertilizer use (10,000 tonnes)” are used. The core explanatory variable in the model is whether the agricultural support protection subsidy policy is implemented, represented by ${\mathrm{D}}_{\mathrm{i}\mathrm{t}}$ in the model and takes the value of 1 if the agricultural support protection subsidy policy has been implemented in region i at time t; otherwise, it takes the value of 0. ${\mathrm{X}}_{\mathrm{i}\mathrm{t}}$ in the model represents the other control variables, including commonly used cropland area, per capita net income of rural residents, rural electricity consumption, total power of agricultural machinery, effective irrigated area, regional primary industry GDP (gross domestic product), and the number of agricultural lands in region i. The model also includes the number of agricultural lands in Region II, which is represented by ${\mathrm{X}}_{\mathrm{i}\mathrm{t}}$ in the model. Irrigated area, regional primary industry gross product, and the number of employees in regional primary industry. Time-fixed and city-fixed effects are also controlled in the model.

The impact mechanism is further analyzed to clarify how the agricultural support and protection subsidy policy achieves a “win-win” situation in terms of food production and quality. Agricultural specialization has been gradually applied to food production and management, and promoting agricultural machinery power is an important manifestation of agricultural specialization [29]. Implementing an agricultural support and protection subsidy policy in China is conducive to developing agricultural scale, intensification, and mechanization, which is conducive to accelerating the transformation of agricultural development. The development of agricultural scale, intensification, and mechanization will guarantee increasing food production and quality. Many agricultural demands for inputs will inevitably accompany agricultural support and protection subsidy policies. With the increase in inputs, the demand for agricultural science and technology will be promoted, and progress in agricultural technology will also be developed. Combined with the endogenous growth theory, endogenous technological progress is the power source of economic growth, and the progress of the level of agricultural technology is the power source of guaranteeing food security, especially driving the growth of food production (Wang et al., 2019) [30]. The previous analysis showed that the use of agricultural machinery improves the efficiency of food production, promotes the growth of food production, and improves the efficiency of pesticide and fertilizer use, leading to the use of fewer pesticides and fertilizers that can achieve better results, thus improving food quality. Therefore, the mechanism variable considered here is agricultural mechanized production, and the selected index is “the total power of agricultural machinery.” The mediation effect analysis method is used to explore the mediation effect of agricultural mechanized production in agricultural support and protection subsidy policy affecting food security. Finally, the model results are analyzed to further improve the research on the effect of agricultural support protection subsidy policy.

The mediator variable considers the level of local scientific and technological development and is measured by the total power of agricultural machinery. “Total power of agricultural machinery” is chosen as an indicator of agricultural mechanized production to explore whether the agricultural support protection subsidy policy can promote food security by enhancing the development of agricultural mechanization. In this paper, we refer to the methodology of previous researchers [31,32] and construct the mechanism analysis model as follows:

$${\mathrm{production}}_{\mathrm{it}}{=\mathsf{\alpha }}_0+{\mathsf{\alpha }}_1{\mathrm{G}}_{\mathrm{i}} \ast {\mathrm{D}}_{\mathrm{t}}+{\mathsf{\alpha }}_2{\mathrm{c}\mathrm{o}\mathrm{n}\mathrm{t}\mathrm{r}\mathrm{o}\mathrm{l}}_{\mathrm{i}\mathrm{t}}+{\mathsf{\mu }}_{\mathrm{i}}+{\mathsf{\lambda }}_{\mathrm{t}}+{\mathsf{\epsilon }}_{\mathrm{i}\mathrm{t}}$$

(5)

$${{\mathrm{m}\mathrm{a}\mathrm{c}\mathrm{h}\mathrm{i}\mathrm{n}\mathrm{e}\mathrm{r}\mathrm{y}}_{\mathrm{i}\mathrm{t}}=\mathsf{\beta }}_0+{\mathsf{\beta }}_1{\mathrm{G}}_{\mathrm{i}}\ast {\mathrm{D}}_{\mathrm{t}}+{\mathsf{\beta }}_2{\mathrm{c}\mathrm{o}\mathrm{n}\mathrm{t}\mathrm{r}\mathrm{o}\mathrm{l}}_{\mathrm{i}\mathrm{t}}+{\mathsf{\mu }}_{\mathrm{i}}+{\mathsf{\lambda }}_{\mathrm{t}}+{\mathsf{\epsilon }}_{\mathrm{i}\mathrm{t}}$$

(6)

$${\mathrm{p}\mathrm{r}\mathrm{o}\mathrm{d}\mathrm{u}\mathrm{c}\mathrm{t}\mathrm{i}\mathrm{o}\mathrm{n}}_{\mathrm{i}\mathrm{t}}{=\mathsf{\gamma }}_0+{\mathsf{\gamma }}_1{\mathrm{G}}_{\mathrm{i}}\ast {\mathrm{D}}_{\mathrm{t}}+{\mathsf{\gamma }}_2{\mathrm{m}\mathrm{a}\mathrm{c}\mathrm{h}\mathrm{i}\mathrm{n}\mathrm{e}\mathrm{r}\mathrm{y}}_{\mathrm{i}\mathrm{t}}+{\mathsf{\gamma }}_3{\mathrm{c}\mathrm{o}\mathrm{n}\mathrm{t}\mathrm{r}\mathrm{o}\mathrm{l}}_{\mathrm{i}\mathrm{t}}+{\mathsf{\mu }}_{\mathrm{i}}+{\mathsf{\lambda }}_{\mathrm{t}}+{\mathsf{\epsilon }}_{\mathrm{i}\mathrm{t}}$$

(7)

In the above model, ${\mathrm{m}\mathrm{a}\mathrm{c}\mathrm{h}\mathrm{i}\mathrm{n}\mathrm{e}\mathrm{r}\mathrm{y}}_{\mathrm{i}\mathrm{t}}$ is the total power of agricultural machinery, i.e., one of the mediating variables in this study, and the other variables are consistent with the meaning of the variables in the above double-difference model. Equation (5) represents the impact of the support and protection subsidy policy on food production, Equation (6) represents the impact of the support and protection subsidy policy on the total power of agricultural machinery, i.e., the impact of the policy on the mediator variable, and Equation (7) represents the impact of the support and protection subsidy policy and the mediator variable on the food production together. Fertilizer use and pesticide application are used as explanatory variables to test whether the use of total power of farm machinery will reduce the use of fertilizer and pesticides. The constructed model is as follows:

$${{\mathrm{f}\mathrm{e}\mathrm{r}\mathrm{t}\mathrm{i}\mathrm{l}\mathrm{i}\mathrm{z}\mathrm{e}\mathrm{r}}_{\mathrm{i}\mathrm{t}}\left({\mathrm{p}\mathrm{e}\mathrm{s}\mathrm{t}\mathrm{i}\mathrm{c}\mathrm{i}\mathrm{d}\mathrm{e}}_{\mathrm{i}\mathrm{t}}\right)=\mathsf{\alpha }}_0+{\mathsf{\alpha }}_1{\mathrm{G}}_{\mathrm{i}}\ast {\mathrm{D}}_{\mathrm{t}}+{\mathsf{\alpha }}_2{\mathrm{c}\mathrm{o}\mathrm{n}\mathrm{t}\mathrm{r}\mathrm{o}\mathrm{l}}_{\mathrm{i}\mathrm{t}}+{\mathsf{\mu }}_{\mathrm{i}}+{\mathsf{\lambda }}_{\mathrm{t}}+{\mathsf{\epsilon }}_{\mathrm{i}\mathrm{t}}$$

(8)

$${{\mathrm{f}\mathrm{e}\mathrm{r}\mathrm{t}\mathrm{i}\mathrm{l}\mathrm{i}\mathrm{z}\mathrm{e}\mathrm{r}}_{\mathrm{i}\mathrm{t}}\left({\mathrm{p}\mathrm{e}\mathrm{s}\mathrm{t}\mathrm{i}\mathrm{c}\mathrm{i}\mathrm{d}\mathrm{e}}_{\mathrm{i}\mathrm{t}}\right)=\mathsf{\gamma }}_0+{\mathsf{\gamma }}_1{\mathrm{G}}_{\mathrm{i}}\ast {\mathrm{D}}_{\mathrm{t}}+{\mathsf{\gamma }}_2\mathrm{m}\mathrm{a}\mathrm{c}\mathrm{h}\mathrm{i}\mathrm{n}\mathrm{e}\mathrm{r}\mathrm{y}+{\mathsf{\gamma }}_3{\mathrm{c}\mathrm{o}\mathrm{n}\mathrm{t}\mathrm{r}\mathrm{o}\mathrm{l}}_{\mathrm{i}\mathrm{t}}+{\mathsf{\mu }}_{\mathrm{i}}+{\mathsf{\lambda }}_{\mathrm{t}}+{\mathsf{\epsilon }}_{\mathrm{i}\mathrm{t}}$$

(9)

The baseline regression results of the double-difference model verified that the coefficients of the core explanatory variables in Equations (5) and (8) are significant and greater than 0, indicating that the food support and protection policies help to increase food production and reduce pesticide and fertilizer use. Moreover, whether ${\mathsf{\beta }}_1$ and ${\mathsf{\gamma }}_2$ in Equations (6), (7) and (9) are significant or not needs to be verified. If they are both significant, then we will see whether ${\mathsf{\gamma }}_1$ is significant; if either of ${\mathsf{\beta }}_1$ and ${\mathsf{\gamma }}_2$ is not significant, it is necessary to implement the Sobel test.

3.2. Variable Selection and Data Description

3.2.1. Explained Variables

At the level of food production security, the explanatory variable is the amount of food production. Grain production is one of the most important aspects of China’s food security, representing the production capacity of food, which is an important indicator of national food security. It reflects the food output capacity that can be successfully achieved domestically within a certain period, certain technical conditions, and inputs of production factors, specifically reflected in the total annual food production. At the food quality and safety level, the explanatory variables are pesticide and chemical fertilizer use. Using pesticides and chemical fertilizers pollutes the environment and harms human health. Agricultural fertilizer application is the amount of fertilizer used in agricultural production during the year, including nitrogen, phosphorus, potash, and compound fertilizer. Pesticide use refers to using chemicals and biological drugs that regulate plant growth, which can affect grain quality. Therefore, this paper considers pesticide and fertilizer usage as explanatory variables to carry out the analysis from the perspective of food quality and safety.

3.2.2. Core Explanatory Variable

This paper takes the implementation of agricultural expenditure protection subsidy policy as the core explanatory variable. Due to the DID model adopted in this study, which refers to Callaway et al. (2021) for a multi-period DID analysis approach [33], the agricultural support protection subsidy policy is an important policy shock, and this variable is a dummy variable that takes the value of 1 if the policy is implemented, and 0 otherwise, it is denoted by $D_{it}$, which is the key variable of interest, where the coefficient of the variable is reflected by the estimation of the double-difference. The coefficient can portray the changes in local food production and fertilizer use after implementing the agricultural support protection subsidy policy, thus reflecting the effect of the policy.

3.2.3. Mechanism Variable

The mechanism variable is the level of rural scientific and technological development, expressed in terms of the total power of agricultural machinery. Specifically, it is a comprehensive indicator of the power of various agricultural machinery calculated according to power. In recent years, the digital, networked, and intelligent transformation of the agricultural industry has accelerated. Smart agriculture has begun to take effect, and the level of intelligence has gradually improved. China’s smart technology agricultural infrastructure is perfect, and the popularization of Chinese agricultural science and technology has gradually increased. The level of agricultural science and technology has an inextricable relationship with food development, and thus, this paper takes it as a mechanism variable to explore the mechanism played by the level of rural technology development.

3.2.4. Control Variables

Rural arable land situation refers to land used for agricultural production, including paddy fields, dry land, garden land, etc., and land used to grow crops, such as rice, wheat, corn, etc., to ensure the production and supply of food, which is expressed by the area of common arable land. It also includes the situation of farmers’ income, which is expressed as the per capita net income of rural residents, irrigation and electricity consumption, i.e., effective irrigated area and actual rural electricity consumption, and economic development and employment, i.e., agricultural GDP and the number of people involved in agricultural production.

$D_{it}$ in the model is the key variable of interest. The coefficient of this variable corresponds to the estimated quantity in double-differencing. The coefficient can portray the changes in local food production and fertilizer use after implementing the agricultural support and protection subsidy policy as a response to the effect of the agricultural support and protection subsidy policy. In this study, all the variables involved are defined, as shown in

Table 1. Specific definitions of relevant variables.

Variable Definitions	Metrics	Variable Symbols	Variable Unit
grain production	Annual grain production in prefecture-level cities	${production}_{it}$	tonnes
Agricultural support protection subsidy policy	Whether to implement an agricultural support protection subsidy policy	$D_{it}$	Dummy variable that takes the value of 1 if the policy is implemented and 0 otherwise
Situation of rural arable land	Area of commonly cultivated land	${land}_{it}$	Thousand hectares
Situation of farmers’ incomes	Per capita net income of rural residents	${income}_{it}$	Yuan
Rural electricity consumption	Rural electricity consumption	${electricity}_{it}$	Billion kilowatt-hours (kWh)
Level of rural scientific and technological development	Total power of agricultural machinery	${machinery}_{it}$	kilowatt (unit of electric power)
Irrigation in rural areas	Effective irrigated area	${irrigation}_{it}$	Thousand hectares
Economic development	Regional primary sector GDP	${primarygdp}_{it}$	Billions
Employment situation	Number of people employed in the primary sector	${primarypop}_{it}$	Ten thousand people
Pesticide usage	Pesticide application rate	${pesticide}_{it}$	tonnes
Chemical fertilizer use	Annual fertilizer use in prefecture-level cities	${fertilizer}_{it}$	tonnes
Specifics of grain production (rice, wheat, and corn)	Rice production	${prodg}_{it}$	tonnes
	Wheat production	${proxm}_{it}$	tonnes
	Corn production	${proym}_{it}$	tonnes
Time fixed effect	Year	${year}_t$	Year dummy variable, controlling for year effects
Area fixed effect	Area	${city}_i$	Regional dummy variables, controlling for regional effects

.

3.3. Data Sources

The data in this paper come from the official websites of provincial and municipal statistical and agricultural bureaus, specifically the China Rural Statistical Yearbook, China Statistical Yearbook, China Grain Yearbook, and EPS (Express Professional Superior) Three Rural Databases, etc. The period is from 2007 to 2020, and the data includes prefecture-level cities in 31 provinces except Hong Kong, Macao, and Taiwan. Indicators collected include the area of commonly cultivated land, grain output, sown area of grain crops, per capita net income of rural residents, rural electricity consumption, the total power of agricultural machinery, effective irrigated area, the gross domestic product of the primary industry, number of employees in the primary industry of the region, amount of pesticide application, fertilizer usage, and so on. Prefectural cities with huge missing value data and the varieties of grain cultivation that do not involve the three major staple grains were removed. Finally, 298 prefecture-level city data research samples were identified. The descriptive statistical analysis of relevant variables is shown in

Table 2. Results of descriptive statistics.

Variables	Observation	Mean	Standard Deviation	Minimum	Median	Maximum
${production}_{it}$	4172	200.72	207.11	0.21	285.22	1735.02
$D_{it}$	4172	0.38	0.48	0	0.41	1
${fertilizer}_{it}$	4172	18.01	16.78	0.04	17.35	121.88
${pesticide}_{it}$	4172	4.43	4.01	0.05	6.58	19.88
${land}_{it}$	4172	340.63	341.29	2.80	409.88	6448.21
${income}_{it}$	4172	10,341.13	6124.82	1236.03	10,944.61	20,2000.05
${electricity}_{it}$	4172	26.41	70.12	0.01	33.92	1101.22
${machinery}_{it}$	4172	305.22	282.95	1.68	347.81	2040.45
${irrigation}_{it}$	4172	184.49	156.69	0.10	255.69	955.47
${primarygdp}_{it}$	4172	571.43	821.22	3.25	688.09	20,986.81
${primarypop}_{it}$	4172	0.81	2.35	0.02	2.31	38.36

.

The descriptive statistics table shows differences in grain production in different regions; the maximum value is 17,350,200 tonnes, the minimum value is 0.21, the average value is 2,007,200 tonnes, and the standard deviation is 207.11. There are also differences in the amount of fertilizers used in the grain production. Table 2 shows that the maximum value reached 1,218,800 tonnes, the minimum value was 0.04 million tonnes, the average value was 180,100 tonnes, and the standard deviation was 1,801,000 tonnes. Differences in food production and food development in different regions, both in scale and structure, can also be observed.

4. Analysis of Empirical Results

4.1. Analysis of the Effect of Subsidy Policy

The results of the empirical study of the impact of agricultural support protection subsidy policy on food production are shown in

Table 3. Impact of agricultural support and protection subsidy policies on production.

	Model (1)	Model (2)
Variables	$\mathit{P}\mathit{r}\mathit{o}\mathit{d}\mathit{u}\mathit{c}\mathit{t}\mathit{i}\mathit{o}\mathit{n}$	$\mathit{P}\mathit{r}\mathit{o}\mathit{d}\mathit{u}\mathit{c}\mathit{t}\mathit{i}\mathit{o}\mathit{n}$
$D_{it}$	11.301 ***	9.651 **
	(1.251)	(4.247)
${land}_{it}$		0.016 ***
		(0.006)
${income}_{it}$		0.004
		(0.012)
electricityit		0.006 ***
		(0.001)
${irrigation}_{it}$		0.299 ***
		(0.021)
primarygdpit		0.032 ***
		(0.002)
${primarypop}_{it}$		0.145 ***
		(0.018)
Constant	197.048 ***	41.968 ***
	(10.770)	(10.762)
Year fixed	Yes	Yes
City fixed	Yes	Yes
Observation	4172	4172
${adj. R}^2$	0.166	0.157

Note: Values in parentheses are standard errors, *** and ** indicate significance at the 1% and 5% significance levels, respectively.

, and the explanatory variable is total food production. The empirical results in column (1) are those without control variables, and those in column (2) are the results with control variables.

Table 4. Inhibitory effect of agricultural support and protection subsidy policy.

	Model (1)	Model (2)	Model (3)	Model (4)
Variables	$\mathit{F}\mathit{e}\mathit{r}\mathit{t}\mathit{i}\mathit{l}\mathit{i}\mathit{z}\mathit{e}\mathit{r}$	$\mathit{F}\mathit{e}\mathit{r}\mathit{t}\mathit{i}\mathit{l}\mathit{i}\mathit{z}\mathit{e}\mathit{r}$	$\mathit{P}\mathit{e}\mathit{s}\mathit{t}\mathit{i}\mathit{c}\mathit{i}\mathit{d}\mathit{e}$	$\mathit{P}\mathit{e}\mathit{s}\mathit{t}\mathit{i}\mathit{c}\mathit{i}\mathit{d}\mathit{e}$
$D_{it}$	−0.262 ***	−0.203 *	−0.770 *	−0.208 *
	(0.101)	(0.112)	(0.465)	(0.125)
${land}_{it}$		0.016 ***		1.904
		(0.005)		(4.876)
${income}_{it}$		0.021 ***		−0.087
		(0.002)		(0.104)
${electricity}_{it}$		0.002		0.001
		(0.004)		(0.005)
${irrigation}_{it}$		0.008 ***		−5.772
		(0.002)		(18.261)
${primarygdp}_{it}$		−0.016 ***		−0.082
		(0.002)		(0.151)
${primarypop}_{it}$		0.021 ***		−1.231
		(0.002)		(7.521)
Constant	17.968 ***	5.057 ***	70.235 **	10.588 **
	(4.492)	(0.878)	(30.746)	(5.207)
Year fixed	Yes	Yes	Yes	Yes
City fixed	Yes	Yes	Yes	Yes
Observation	4172	4172	4172	4172
${adj. R}^2$	0.028	0.115	0.201	0.126

Note: Values in parentheses are standard errors, ***, **, and * indicate significance at the 1%, 5%, and 10% significance levels, respectively.

shows the coefficient of the core explanatory variable $D_{it}$ is significant regardless of whether the control variable is added. The coefficient of the core variable is significant at the 1% level when no control variable is added. The coefficient of the core variable is significant at the 5% level after the control variable is added. Moreover, the coefficient is always positive, indicating that the agricultural support and protection subsidy policy has significantly increased food production, and the share of the increase is 4.81%. Thus, hypothesis H1 is verified. That is, implementing an agricultural support protection subsidy policy helps promote the quantity of food production.

An analysis of Table 3 and Table 4 indicates that the agricultural support and protection subsidy policy appears to have achieved a “win-win” situation regarding food production growth and quality improvement, which is not in line with traditional perceptions. In fact, after 2004, a large part of China’s food growth momentum came from using pesticides and chemical fertilizers, which indicates that the growth of food production was at the expense of food quality. How, then, can agricultural support and protection subsidy policy achieve a “win-win”? A possible answer is that the agricultural support protection subsidy policy puts forward more stringent requirements on the quality of arable land, and some farmers take the initiative to reduce the amount of pesticides and chemical fertilizers to obtain higher subsidies. However, such decisions may result in a decline in food production. However, from the perspective of short-term interests, it is possible to actively reduce the use of pesticides and fertilizers in exchange for subsidies. The agricultural support and protection subsidy policy has released the financial pressure and accelerated the development of agricultural mechanization. The popularity of mechanization has improved the efficiency of food production and brought into play the scale effect of the policy, which has contributed to the growth of food production. At the same time, mechanization has improved the efficiency of pesticides and fertilizers, prompting lesser use of pesticides and fertilizers to achieve better results, thereby improving grain quality.

The improvement of China’s productivity level and production technology in recent years has ensured that food quantity security continues to be met; at the same time, China’s food quality also has higher standards and requirements. However, with the implementation of the green revolution in agriculture and the adoption of “rough” production methods, food pollution has increased, and food quality and safety have a higher standard and requirement for “rough” production methods. The research shows the factors affecting the quality and safety of food include the following aspects. First, more pesticides and chemical fertilizers are left in the food, which leads to harmful substances. Second, a large amount of industrial waste pollution is gathered in the food, including harmful substances caused by farmland, atmosphere, and water. Third, the quality of food is lowered due to the mildew and breakage caused by the improper treatment in the food purchasing and storage process. Fourth, the chemical additives added to the food during the processing process reduce the quality of food. Chemical additives are added during processing. While ensuring the safety of food quantity, the issue of food quality and safety should not be ignored. The quality and safety of food discussed in this paper refer to the quality and safety problems caused by the excessive use of pesticides and chemical fertilizers in food production and the harmful substances left in the land and food.

An empirical study of agricultural support and protection subsidy policy on food quality and safety adopts the use of pesticides and chemical fertilizers as an indicator of food quality and safety. The use of and pollution by pesticides and chemical fertilizers are common in agricultural production and food cultivation. In grain cultivation, the more pesticides and chemical fertilizers are used, the more serious water, air, and soil pollution will be caused, resulting in poorer arable land strength and grain quality. At the same time, the nitrogen, phosphorus, and other elements present in chemical fertilizers can also lead to eutrophication of rivers, lakes, and other waters, posing a threat to food quality and safety development. When pesticides and chemical fertilizers are used less, it means that the degree of green health of food is better, i.e., the quality and safety of food are more guaranteed. Hence, the research objective is to determine whether the implementation of an agricultural support and protection subsidy policy can effectively constrain the use of pesticides and chemical fertilizers to guarantee the quality and safety of food. Therefore, this model chooses local fertilizer use as the explanatory variable in the multi-period DID. The obtained empirical results are tabulated in Table 4.

Table 4 reports the effects of agricultural support and protection subsidy policies on pesticide and fertilizer usage, with columns (1)–(2) presenting the results of the empirical modeling of the dampening effect of agricultural support and protection subsidy policies on fertilizer usage. Columns (3)–(4) are the results of the empirical model of the agricultural support protection subsidy policy on pesticide usage. From the regression results, as expected from the above hypotheses, implementing an agricultural support and protection subsidy policy effectively suppresses pesticide and fertilizer usage. Columns (1) and (3) represent the regression results without control variables, while columns (2) and (4) represent the regression results with control variables.

In terms of the impact of agricultural support and protection subsidy policy on fertilizer use, the regression results in column (1) show the coefficient of the core explanatory variable, D_it, which is significant at the 1% level when no control variables are added. The regression results in column (2) show that the core explanatory variable D_it coefficient is significant at the 10% level when control variables are added. Implementing the agricultural support and protection subsidy policy reduces fertilizer use by 1.13%. Its weight is obtained by dividing the coefficient of the core explanatory variable in the double-difference model by the corresponding mean value. In terms of the effect of agricultural support and protection subsidy policy on the application of pesticides, the regression results in column (3) are significant at the 10% level for the core explanatory variables without adding the control variables. After adding the control variables, the regression results in column (4) are obtained. The core explanatory variables are still significant at the 10% level, and the implementation of an agricultural protection policy promotes the reduction of pesticide use by 4.69%.

4.2. Robustness Tests

4.2.1. Parallel Trend Test

In constructing the double-difference model, it is necessary to satisfy that before the policy implementation stage, the trend of variable changes in the treatment and control groups in the experiment over time remains consistent, and the parallel trend assumption can be used for testing. When the agricultural support protection subsidy policy meets the parallel trend test, the interference from the time trend to this study can be excluded. Thus, it can be proved that the impact of the agricultural support protection subsidy policy on food production and fertilizer use is entirely due to the shock of the policy itself.

4.2.2. Replacing the Research Sample

To further study the robustness of the model results, the top ten provinces ranked in grain output in 2020, namely Heilongjiang Province, Henan Province, Shandong Province, Anhui Province, Jilin Province, Inner Mongolia Autonomous Region, Hebei Province, Jiangsu Province, Sichuan Province, and Hunan Province, are selected based on the sample dataset. These provinces are used as the research sample to explore the impact of agricultural support and protection subsidy policies on grain yield and quality security. The results obtained from the model are shown in

Table 5. Empirical results of multi-period double-difference in grain-producing provinces.

	Model (1)	Model (2)	Model (3)
Variables	$\mathit{F}\mathit{e}\mathit{r}\mathit{t}\mathit{i}\mathit{l}\mathit{i}\mathit{z}\mathit{e}\mathit{r}$	$\mathit{F}\mathit{e}\mathit{r}\mathit{t}\mathit{i}\mathit{l}\mathit{i}\mathit{z}\mathit{e}\mathit{r}$	$\mathit{F}\mathit{e}\mathit{r}\mathit{t}\mathit{i}\mathit{l}\mathit{i}\mathit{z}\mathit{e}\mathit{r}$
$D_{it}$	−0.378 ***	6.472 ***	−0.243 ***
	(0.115)	(1.961)	(0.066)
${land}_{it}$	0.011 ***	0.161 ***	0.922
	(0.001)	(0.021)	(1.324)
${income}_{it}$	0.001	0.021 ***	0.093
	(0.001)	(0.007)	(0.117)
${electricity}_{it}$	0.002	0.001	0.006
	(0.003)	(0.001)	(0.013)
${irrigation}_{it}$	0.009 ***	0.228 ***	0.782
	(0.002)	(0.033)	(0.811)
${primarygdp}_{it}$	−0.012	−0.032 ***	−0.028
	(0.016)	(0.004)	(0.293)
${primarypop}_{it}$	0.026 ***	0.329 ***	0.081
	(0.004)	(0.021)	(0.128)
Constant	2.869 **	26.105	30.121 ***
	(1.207)	(23.518)	(6.025)
Year-fixed	Yes	Yes	Yes
City-fixed	Yes	Yes	Yes
Observation	1762	1762	1762
${adj. R}^2$	0.269	0.216	0.329

Note: Values in parentheses are standard errors, *** and ** indicate significance at the 1% and 5% significance levels, respectively.

.

In large grain-producing provinces, an important source of economic growth includes grain production and supply; therefore, compared with non-main grain-producing provinces, the governments of large grain-producing provinces focus extra attention to the issue of grain production and will actively respond, follow up, and implement the national grain policy. By choosing data from large food-producing provinces, it is possible to test whether the agricultural support and protection subsidy policy affects food security in regions that emphasize food production and further test the robustness of the baseline regression results.

Model (1) in Table 5 indicates the effect of agricultural support and protection subsidy policy on fertilizer use, and model (2) indicates the effect of agricultural support and protection subsidy policy on grain yield. Model (3) indicates the impact of agricultural support protection subsidy policy on pesticide application. Through screening, the sample individuals obtained are 1762. The coefficients of the core explanatory variables in models (1), (2), and (3) are all significant at the 1% level. The agricultural support protection subsidy policy still has a significant inhibitory effect on fertilizers and pesticides and an equally significant enhancement effect on grain output in provinces with higher grain output, and the model passes the robustness test.

4.2.3. Sampling with Bootstrap

To further validate the robustness of the model, the analysis is carried out based on the Bootstrap method, which is used to estimate the sampling distribution by taking multiple samples. Bootstrap is an important estimation statistic in nonparametric statistics. Several samples are drawn from the original sample using a repeated sampling technique in this testing process. Then, based on the drawn samples, the statistic is estimated, and the above steps are repeated N times to obtain N statistics. Finally, the indicators of the statistics are estimated. The data were sampled 600 times, and the mean values of the statistics of 600 samples in the corresponding years were obtained to analyze the robustness of the model. The results of the model obtained are shown in

Table 6. Empirical results of robustness analysis based on Bootstrap sampling.

	Model (1)	Model (2)	Model (3)
Variables	$\mathit{F}\mathit{e}\mathit{r}\mathit{t}\mathit{i}\mathit{l}\mathit{i}\mathit{z}\mathit{e}\mathit{r}$	$\mathit{P}\mathit{e}\mathit{s}\mathit{t}\mathit{i}\mathit{c}\mathit{i}\mathit{d}\mathit{e}$	$\mathit{P}\mathit{r}\mathit{o}\mathit{d}\mathit{u}\mathit{c}\mathit{t}\mathit{i}\mathit{o}\mathit{n}$
$D_{it}$	−0.319 ***	−0.524 **	4.661 **
	(0.084)	(0.227)	(1.805)
${land}_{it}$	0.035 **	0.366	0.279 ***
	(0.014)	(0.297)	(0.094)
${income}_{it}$	0.058 **	−0.107	0.051
	(0.026)	(0.095)	(0.098)
${electricity}_{it}$	0.001	0.038 ***	0.064 ***
	(0.006)	(0.011)	(0.018)
${irrigation}_{it}$	0.019 ***	−0.392 *	0.115 *
	(0.003)	(0.219)	(0.063)
${primarygdp}_{it}$	−0.048 ***	−0.088	0.081 ***
	(0.017)	(0.174)	(0.026)
${primarypop}_{it}$	0.015 ***	−0.603 *	0.392 **
	(0.001)	(0.281)	(0.177)
Constant	8.109 ***	7.922 **	10.553 ***
	(1.455)	(2.331)	(2.618)
Year-fixed	Yes	Yes	Yes
City-fixed	Yes	Yes	Yes
Observation	1322	1322	1322
${adj. R}^2$	0.495	0.381	0405

Note: Values in parentheses are standard errors, ***, **, and * indicate significance at the 1%, 5%, and 10% significance levels, respectively.

.

The empirical results in Table 6 show that the coefficients of the core explanatory variables in models (1) to (3) are significant. Moreover, the conclusions obtained from the robustness test are consistent with those obtained from the benchmark regression model and indicate that the model results remain robust after adopting the Bootstrap repeated sampling method, i.e., the effect of the agricultural support protection subsidy policy remains significant.

4.2.4. Propensity Score Matching Test

The propensity score matching (PSM) method was used for robustness testing. The robustness test using this method helps mitigate the effect of confounding variables and other variables on the model’s validity. The basic idea of the PSM method embodies finding samples in the control group that can be matched with the treatment group and conducting regression analysis based on the matched samples. The empirical results obtained using this method are shown in

Table 7. Robustness results for propensity score matching methods.

	Model (1)	Model (2)	Model (3)
Variables	$\mathit{F}\mathit{e}\mathit{r}\mathit{t}\mathit{i}\mathit{l}\mathit{i}\mathit{z}\mathit{e}\mathit{r}$	$\mathit{P}\mathit{e}\mathit{s}\mathit{t}\mathit{i}\mathit{c}\mathit{i}\mathit{d}\mathit{e}$	$\mathit{P}\mathit{r}\mathit{o}\mathit{d}\mathit{u}\mathit{c}\mathit{t}\mathit{i}\mathit{o}\mathit{n}$
$D_{it}$	−0.312 ***	−0.183 ***	3.861 ***
	(0.089)	(0.055)	(0.839)
${land}_{it}$	−0.001	0.925	−0.011 **
	(0.005)	(1.577)	(0.005)
${income}_{it}$	0.001	0.054	0.013 **
	(0.001)	(0.121)	(0.005)
${electricity}_{it}$	−0.002	0.003	0.001
	(0.003)	(0.005)	(0.002)
${irrigation}_{it}$	0.007 *	−0.219	0.145 ***
	(0.004)	(0.652)	(0.041)
${primarygdp}_{it}$	−0.012 ***	−0.021	0.006 ***
	(0.002)	(0.043)	(0.002)
${primarypop}_{it}$	0.028 ***	−0.353	0.112 ***
	(0.003)	(0.472)	(0.041)
Constant	9.521 ***	10.212 ***	117.198 ***
	(1.725)	(2.332)	(20.336)
Year fixed	Yes	Yes	Yes
City fixed	Yes	Yes	Yes
Observation	1208	1208	1208
${adj. R}^2$	0.047	0.102	0.062

Note: Values in parentheses are standard errors, ***, **, and * indicate significance at the 1%, 5%, and 10% significance levels, respectively.

.

Empirical research often uses a combination of PSM and double-differencing to re-estimate the outcome of shocks to policy. In this study, the neighborhood matching approach is adopted for the treatment. From the empirical results in Table 3, Table 4, Table 5, Table 6 and Table 7, model (1) represents the analysis of the PSM test for fertilizer use, model (2) represents the analysis of the PSM test for pesticide application, and model (3) represents the analysis of PSM test for grain yield variable. Models (1) and (2) were analyzed from the perspective of grain quality and safety, and model (3) was analyzed from the perspective of grain yield safety. After regression using PSM-DID, the core explanatory variables were significant at a 1% level, indicating that the model passed the robustness test.

4.3. Heterogeneity Analysis

4.3.1. Heterogeneity Analysis Based on Geographical Conditions

Chinese provinces are divided into eastern, central, and western regions to analyze further the impact of agricultural support and protection subsidy policies on grain yield and quality under different regional levels. Regarding the division of the eastern, central, and western provinces, referring to the National Bureau of Statistics website and other relevant information (without considering Hong Kong, Macao, and Taiwan), the collected data were grouped and regressed. The results of the heterogeneity analysis of different regions obtained are shown in

Table 8. Effectiveness of different regional dimensions in contributing to grain production.

	Model (1)	Model (2)	Model (3)	Model (4)	Model (5)	Model (6)
Variables	Eastern Region		Central Region		Western Region
$D_{it}$	7.657 ***	7.574 ***	23.423 ***	8.648 ***	4.722 ***	3.833 ***
	(2.308)	(2.515)	(2.039)	(2.986)	(0.951)	(1.406)
${land}_{it}$		0.069 **		0.014 ***		0.051 ***
		(0.028)		(0.005)		(0.007)
${income}_{it}$		0.002		0.003		0.001
		(0.002)		(0.005)		(0.002)
${electricity}_{it}$		0.001		0.005 ***		0.002
		(0.002)		(0.001)		(0.035)
${irrigation}_{it}$		0.297 ***		0.103 *		0.002
		(0.033)		(0.053)		(0.026)
${primarygdp}_{it}$		−0.006 ***		−0.007 ***		−0.011 ***
		(0.001)		(0.001)		(0.002)
${primarypop}_{it}$		0.424 ***		0.088 **		0.161 ***
		(0.031)		(0.038)		(0.033)
Constant	224.554 ***	4.297	212.538 ***	90.842 ***	129.529 ***	69.785 ***
	(21.421)	(21.895)	(13.263)	(18.365)	(10.581)	(9.510)
Year fixed	Yes	Yes	Yes	Yes	Yes	Yes
City fixed	Yes	Yes	Yes	Yes	Yes	Yes
Observation	1946	1946	1148	1148	1078	1078
${adj. R}^2$	0.102	0.171	0.110	0.281	0.025	0.172

Note: Values in parentheses are standard errors, ***, **, and * indicate significance at the 1%, 5%, and 10% significance levels, respectively.

.

According to the empirical regression results of different regions in Table 8, the agricultural support protection subsidy policy plays a significant role in promoting food production in the eastern, central, or western regions, and the coefficients of the core explanatory variables in Models (1) to (6) are all significant at the 1% level. Models (1) and (2) indicate that the eastern region is used as the research sample to explore the impact of agricultural support protection subsidy policy on food production. Model (1) does not include control variables, and model (2) includes relevant control variables. Models (3) and (4) indicate that the central region is used as the research sample to explore the impact of agricultural support protection subsidy policy on food production. Model (3) did not include control variables, and model (4) added control variables. Models (5) and (6) indicate that the western region is used as the research sample to explore the impact of agricultural support protection subsidy policy on food production, where model (5) does not add control variables and model (6) adds control variables. The model coefficients of core explanatory variables indicate that agricultural support protection subsidy policy has a significant role in promoting the growth of food production in the East, middle and West. Among them, implementing the agricultural support protection subsidy policy promotes food production by 1.32% in the eastern region, 2.03% in the central region, and 1.14% in the western region. In comparison, the best results regarding food production growth of the agricultural support and protection subsidy policy were achieved in the central region from the perspective of different regions. A possible reason is that in addition to Shanxi Province in the central region, Henan, Hubei, Hunan, Jiangxi, Shanxi, and Inner Mongolia are all major grain-producing areas, indicating that the effect of agricultural support protection subsidy policy in promoting grain production increase is more obvious in major grain-producing areas.

Based on the division of different regions and considering the impact of agricultural support protection subsidy policy on food security, the explanatory variable in the model is the amount of chemical fertilizer used. The empirical results of the model obtained are shown in

Table 9. Effect of fertilizer suppression at different regional levels.

	Model (1)	Model (2)	Model (3)	Model (4)	Model (5)	Model (6)
Variables	Eastern Region		Central Region		Western Region
$D_{it}$	−0.064	−0.417 ***	−0.342 *	−0.811 ***	−0.778 ***	−0.336
	(0.127)	(0.148)	(0.176)	(0.267)	(0.255)	(0.389)
${land}_{it}$		0.011 ***		0.008 *		−0.001
		(0.002)		(0.005)		(0.002)
${income}_{it}$		0.002		0.011 ***		0.001
		(0.022)		(0.001)		(0.001)
${electricity}_{it}$		0.001		0.021 ***		0.032 ***
		(0.031)		(0.001)		(0.002)
${irrigation}_{it}$		0.008 ***		−0.016 ***		−0.013 *
		(0.002)		(0.005)		(0.007)
${primarygdp}_{it}$		−0.031 ***		−0.012 ***		−0.038 ***
		(0.002)		(0.002)		(0.006)
${primarypop}_{it}$		0.028 ***		0.041 ***		0.021 ***
		(0.003)		(0.002)		(0.004)
Constant	19.144 ***	3.063 ***	19.688 ***	11.111 ***	13.944 ***	0.793
	(1.078)	(1.157)	(0.109)	(1.644)	(0.155)	(2.631)
Year-fixed	1946	1946	1148	1148	1078	1078
City-fixed	Yes	Yes	Yes	Yes	Yes	Yes
Observation	Yes	Yes	Yes	Yes	Yes	Yes
${adj. R}^2$	0.011	0.229	0.004	0.136	0.009	0.101

Note: Values in parentheses are standard errors, *** and * indicate significance at the 1% and 10% significance levels, respectively.

.

The empirical results in Table 9 show that models (1) and (2) for the eastern region, the impact of agricultural support protection subsidy policy on fertilizer use, in the absence of control variables, the coefficient of model (1) is not significant, after adding control variables, the coefficient of model (2) is significant at the 1% level, the reduction of fertilizer use is at 0.95%, indicating that in the eastern region. Thus, the implementation of an agricultural support protection subsidy policy effectively suppresses the use of fertilizer and promotes the improvement of arable land quality and grain quality, which can have a good effect on promoting food quality and safety, indicating that in the eastern region, the implementation of agricultural support and protection subsidy policy effectively suppresses the use of chemical fertilizer, promotes the improvement of arable land quality and grain quality, and can play a good role in promoting grain quality and safety. Models (3) and (4) represent the impact of agricultural support protection subsidy policy on fertilizer use in the central region. In model (3), without adding control variables, the coefficient is significant at the 10% level, and after adding relevant control variables, the coefficient of the core explanatory variables in the model is significant at the 1% level, and the use of fertilizer decreases by 1.51%, indicating that in the central region, the agricultural support protection subsidy policy can significantly reduce fertilizer use. The policy effect in the central region is significantly higher than that in the eastern region, indicating that the agricultural support and protection subsidy policy has a more prominent inhibitory effect on fertilizer use in the central region than in the eastern region. Models (5) and (6) represent the effect of the agricultural support protection subsidy policy on fertilizer use in the western region. Model (5) is the result of the model without adding control variables, and Model (6) is the result of the model with the addition of relevant control variables. The coefficients of the core explanatory variables in the model (6) are not significant, indicating that in the western region, the inhibitory effect of agricultural support and protection subsidy policy on fertilizer use is not significant, which also indicates that the effect of agricultural support and protection subsidy policy on suppressing fertilizer use is not prominent in the western region.

Based on the above different regional divisions, the heterogeneity analysis is carried out at the level of pesticide use to explore the impact of agricultural support protection subsidy policy on food quality and safety, and the empirical results of the obtained model are shown in

Table 10. Pesticide inhibition effects at different regional levels.

	Model (1)	Model (2)	Model (3)	Model (4)	Model (5)	Model (6)
Variables	Eastern Region		Central Region		Western Region
$D_{it}$	−0.112	−0.117 ***	−0.311 *	−0.398 ***	−0.219	−0.286
	(0.134)	(0.038)	(0.171)	(0.122)	(0.191)	(0.269)
${land}_{it}$		0.053		0.005		0.006
		(0.061)		(0.008)		(0.009)
${income}_{it}$		0.019		0.061		0.055
		(0.022)		(0.192)		(0.042)
${electricity}_{it}$		0.091 ***		0.033 ***		0.066
		(0.007)		(0.008)		(0.086)
${irrigation}_{it}$		0.003		−0.039		−0.023
		(0.002)		(0.072)		(0.039)
${primarygdp}_{it}$		−0.051		−0.012		−0.024
		(0.043)		(0.016)		(0.031)
${primarypop}_{it}$		0.061		0.038		0.095
		(0.052)		(0.072)		(0.071)
Constant	17.301 ***	5.152 ***	16.322 ***	10.439 ***	13.944 ***	3.662 ***
	(5.053)	(1.032)	(3.221)	(2.431)	(2.155)	(1.153)
Year-fixed	Yes	Yes	Yes	Yes	Yes	Yes
City-fixed	Yes	Yes	Yes	Yes	Yes	Yes
Observation	1946	1946	1148	1148	1078	1078
${adj. R}^2$	0.052	0.064	0.098	0.112	0.072	0.166

Note: Values in parentheses are standard errors, *** and * indicate significance at the 1% and 10% significance levels, respectively.

.

Table 10 shows the effect of food support and protection subsidy policy on pesticide application in food security, where models (1), (3), and (5) represent the absence of added control variables, and models (2), (4), and (6) represent the addition of control variables in the model. Models (1) and (2) represent the empirical results in the eastern region, models (3) and (4) represent the empirical results in the central region, and models (5) and (6) represent the empirical results in the western region. The empirical results show that in terms of the effect of suppressing pesticide application, in the eastern region, the policy implementation promotes the reduction of pesticide application by 1.77%, and in the central region, the policy implementation promotes the reduction of pesticide application by 2.91%. Similar to the case of fertilizers, the effect of the food support protection subsidy policy is more significant in the East and the center and not significant in the West.

The main reason for the weaker effect in the western region is that this region has many ethnic minorities, and the understanding and implementation of the policy and the acceptance of the policy are far less than that in the eastern and central regions. Moreover, compared to the eastern and central regions, the use of fertilizers in the western region is low, making it difficult for the policy to play a role. China’s main food production areas are concentrated in the central and eastern regions. From the geographical environment and economic development, the western region of food development is still behind the eastern and central regions. In recent years, the western region has emphasized the “development of seed industry, to create the western seed capital” and continues to promote the western region’s industrial restructuring and the development of characteristic advantageous industries in the new era to create a higher level of “heavenly grain silo” and other policy objectives. However, at the current stage, the policy effect of the food support protection subsidy policy is still reflected in the eastern and central regions.

4.3.2. Heterogeneity Analysis Based on Grain Types

In the heterogeneity analysis, the effects of agricultural support and protection subsidy policies on the yields of the three major staple grains, i.e., rice, wheat, and corn, are examined separately, and the results of the empirical analysis are shown in

Table 11. Effectiveness of agricultural support and protection subsidy policies in promoting yields of different grain crops.

	Model (1)	Model (2)	Model (3)	Model (4)	Model (5)	Model (6)
Variables	$\mathit{P}\mathit{r}\mathit{o}\mathit{d}\mathit{g}$	$\mathit{P}\mathit{r}\mathit{o}\mathit{d}\mathit{g}$	$\mathit{P}\mathit{r}\mathit{o}\mathit{x}\mathit{m}$	$\mathit{P}\mathit{r}\mathit{o}\mathit{x}\mathit{m}$	$\mathit{P}\mathit{r}\mathit{o}\mathit{y}\mathit{m}$	$\mathit{P}\mathit{r}\mathit{o}\mathit{y}\mathit{m}$
$D_{it}$	4.781 ***	1.389 **	2.125 ***	0.617 **	2.882 ***	0.837 **
	(0.529)	(0.666)	(0.235)	(0.296)	(0.319)	(0.401)
${land}_{it}$		0.007 ***		0.003 ***		0.004 ***
		(0.002)		(0.001)		(0.001)
${income}_{it}$		0.001		0.003		0.002
		(0.001)		(0.002)		(0.002)
${electricity}_{it}$		0.043 ***		0.041 ***		0.033 ***
		(0.002)		(0.002)		(0.001)
${irrigation}_{it}$		0.127 ***		0.056 ***		0.076 ***
		(0.008)		(0.004)		(0.005)
${primarygdp}_{it}$		−0.023 ***		−0.015 ***		−0.025 ***
		(0.002)		(0.002)		(0.02)
${primarypop}_{it}$		0.083 ***		0.061 ***		0.062 ***
		(0.021)		(0.006)		(0.011)
Constant	83.352 ***	17.753 ***	37.045 ***	7.8900 ***	50.247 ***	10.702 ***
	(0.326)	(4.552)	(0.145)	(2.0233)	(0.196)	(2.744)
Year fixed	Yes	Yes	Yes	Yes	Yes	Yes
City fixed	Yes	Yes	Yes	Yes	Yes	Yes
Observation	4172	4172	4172	4172	4172	4172
${adj. R}^2$	0.021	0.156	0.021	0.156	0.021	0.156

Note: Values in parentheses are standard errors, *** and ** indicate significance at the 1% and 5% significance levels, respectively.

.

According to the empirical results in Table 11, models (1) and (2) represent the impact of agricultural support and protection subsidy policy on rice yield, models (3) and (4) represent the impact of agricultural support and protection subsidy policy on wheat yield, and models (5) and (6) represent the impact of agricultural support and protection subsidy policy on corn yield. Among them, models (1), (3), and (5) represent the empirical results without adding control variables. Models (2), (4), and (6) represent the empirical analysis results after adding control variables. From the above analysis, it can be concluded that the agricultural support protection subsidy policy has a significant promotion effect on rice yield, wheat yield, and corn yield, and from the results of the empirical analysis, the implementation of the agricultural support protection subsidy policy promotes the increase in rice yield by 1.73%, the increase in wheat yield by 1.38%, and the increase in corn yield by 1.61%, in which the agricultural support protection subsidy policy promotes the rice yield. The effect of the increase is the largest, followed by corn yield and wheat yield. Rice is one of the most important food crops in China, accounting for about 30% of China’s total food production. Rice has also become the main food crop and is an advantageous leading industry in many cities. Hence, the sensitivity to the policy is also the strongest, and it is very intuitively embodied in the yield boost.

4.4. Further Analysis of Impact Mechanisms

The results of the mechanism test in

Table 12. Results of the mechanism analysis of the impact of agricultural support and protection subsidy policies on grain production.

	Model (1)	Model (2)	Model (3)	Model (4)
Variables	$\mathit{M}\mathit{a}\mathit{c}\mathit{h}\mathit{i}\mathit{n}\mathit{e}\mathit{r}\mathit{y}$	$\mathit{P}\mathit{r}\mathit{o}\mathit{d}\mathit{u}\mathit{c}\mathit{t}\mathit{i}\mathit{o}\mathit{n}$	$\mathit{F}\mathit{e}\mathit{r}\mathit{t}\mathit{i}\mathit{l}\mathit{i}\mathit{z}\mathit{e}\mathit{r}$	$\mathit{P}\mathit{e}\mathit{s}\mathit{t}\mathit{i}\mathit{c}\mathit{i}\mathit{d}\mathit{e}$
$G_i\ast D_t$	2.845 ***	5.132 ***	−0.373 ***	−0.391 ***
	(0.221)	(1.125)	(0.062)	(0.019)
${machinery}_{it}$		0.372 ***	−0.009 ***	−0.012 ***
		(0.021)	(0.002)	(0.003)
${land}_{it}$	1.749 ***	0.0213	0.072 ***	0.572 ***
	(0.216)	(0.023)	(0.014)	(0.083)
${income}_{it}$	0.013	0.002	0.082	0.064
	(0.041)	(0.019)	(0.078)	(0.097)
${electricity}_{it}$	0.863 ***	0.127 ***	−0.071 ***	−0.021 ***
	(0.151)	(0.014)	(0.008)	(0.003)
${irrigation}_{it}$	0.293 ***	0.168 ***	0.041 **	0.199 ***
	(0.164)	(0.021)	(0.013)	(0.024)
${primarygdp}_{it}$	0.363 ***	0.167 ***	−0.842 ***	−0.721 ***
	(0.014)	(0.011)	(0.012)	(0.028)
${primarypop}_{it}$	0.531 ***	0.903 ***	−0.623 **	−0.934 ***
	(0.072)	(0.066)	(0.049)	(0.022)
Constant	3.415 ***	7.957 **	3.655 ***	5.921 ***
	(0.536)	(3.524)	(0.123)	(0.793)
Year fixed	Yes	Yes	Yes	Yes
City fixed	Yes	Yes	Yes	Yes
Observation	4172	4172	4172	4172
${adj. R}^2$	0.241	0.143	0.812	0.593

Note: Values in parentheses are standard errors, *** and ** indicate significance at the 1% and 5% significance levels, respectively.

show that in column (1), the coefficients of the mediating variables are significant, suggesting that policy implementation enhances the development of agricultural machinery power. The regression coefficient in column (2) is significantly positive, reflecting that the food support protection subsidy policy and the total power of agricultural machinery jointly promote the growth of food production, and the agricultural support protection subsidy policy and the development of agricultural machinery simultaneously promote the growth of food production. Columns (3) and (4) indicate that the food support protection subsidy policy and the total power of agricultural machinery reduce the use of pesticides and fertilizers. With the development of agricultural machinery power, mechanical production can serve as a more effective alternative to human labor; compared to the previous use of fertilizers and pesticides to stimulate agricultural production, the development of mechanical power not only improves the efficiency of food production but also helps to reduce the frequency of pesticides and fertilizer use. The coefficients of the above variables are all significant at the 1% level, indicating that the total power of agricultural machinery plays a part in mediating the effect of food support and protection subsidy policies on food production and food quality. The combination of Sobel and Bootstrap tests found that the Sobel test is significant at the 1% level, and the direct and indirect effects are significant after carrying out the Bootstrap test. The above analysis also verified hypothesis H3.

The results of the mechanism test in Table 12 show that the coefficient of the mediator variable indicated in column (1) is significant, indicating that policy implementation promotes the development of agricultural machinery power. The regression coefficient in column (2) is significantly positive, indicating that the food support protection subsidy policy and the total power of agricultural machinery jointly promote the growth of food production. Columns (3) and (4) indicate that the food support protection subsidy policy and the total power of agricultural machinery together reduce the use of pesticides and chemical fertilizers. Mechanical production can replace human labor more when the power of agricultural machinery is more developed. Compared with the previous method of using chemical fertilizers and pesticides to stimulate agricultural production, the development of mechanical power not only improves the efficiency of food production but also helps to reduce the frequency of the use of pesticides and chemical fertilizers. The coefficients of the above variables are all significant at the 1% level, indicating that the total power of agricultural machinery plays a partial mediating effect, driving the impact of food support and protection subsidy policies on food production and quality. The combination of Sobel and Bootstrap tests indicated that the Sobel test was significant at the 1% level, and the direct and indirect effects were significant after carrying out the Bootstrap test. The above analysis also verified hypothesis H3.

5. Conclusions and Policy Recommendations

5.1. Research Conclusions

This paper analyzes the effects of the agricultural support protection subsidy policy by combining the multi-period DID. The study has the following relevant conclusions. The agricultural support protection subsidy policy, which was successively promoted after 2015, has obvious technical and scale effects and can increase food production to a certain extent. It effectively reduces the impact of food quality contaminated by pesticides and fertilizers in food production. The study’s results still hold after a series of robustness tests, including replacing the research sample, using the Bootstrap method, and performing a propensity score matching test. An analysis of the mechanism of how agricultural support and protection subsidy policies affect food security and quality shows that the level of agricultural machinery plays a crucial positive role in the impact of subsidy policy on food production and food security. Further heterogeneity analysis shows that the impacts of agricultural support protection subsidy policies on food quality and food yield security are also significantly differentiated across different geographic regions, food crop production, and pollutant type levels. Specifically, agricultural support and protection subsidy policies are most effective in promoting food production growth in the central region. At the same time, the promotion effect is relatively weak in the eastern and western regions. Regarding suppressing fertilizer use and pesticide application, the effect of the subsidy policy is not prominent in the western region, and it is more significant in the eastern and central regions. In addition, the effect of the subsidy policy in promoting the growth of rice production is the largest, followed by corn and wheat production.

5.2. Policy Recommendations

The results of this study contribute to the further exploration of the effects of the current agricultural support and protection subsidy policy and contribute to the further improvement of China’s grain subsidy system. In China’s plan to promote rural agricultural modernization, emphasis is placed on improving the compensation mechanism for the benefits of the main grain-producing areas. The main grain-producing areas are under the influence of the national grain policy, which inevitably affects grain quantity and quality security. This is highly consistent with the existing conclusions that agricultural support and protection subsidy policies are conducive to promoting arable land protection, as well as the findings of some scholars who found that agricultural support and protection subsidy policies have a positive impact on arable land protection as well as on the moderate-scale operation of food [34,35]. The agricultural support and protection subsidy policy further promotes grain production in main grain-producing areas, allowing main grain-producing areas to play their roles, fully thereby guaranteeing national food security.

Moreover, to achieve the high-quality development of the agricultural support and protection subsidy policy for the food industry, it is necessary to comprehensively consider several aspects, such as policy design, implementation mechanisms, and environmental sustainability. On the one hand, it is necessary to scientifically formulate the subsidy policy to ensure that the agricultural support and protection subsidy policy is framed following scientific principles and the actual situation. The government needs to understand the demand for agricultural production, market conditions, and farmers’ incomes through in-depth research and scientifically set the targets and ranges of subsidy policies. On the other hand, a differentiated approach to subsidies should be adopted. Different policies should be formulated according to the situation of different regions and agricultural producers. This approach will help bring the incentive effect of subsidies into play more accurately and lead to more efficient and sustainable agricultural production. Improving agricultural technology also appears to be crucial. Linking subsidy policies to technological progress and innovation encourages agricultural producers to adopt advanced technologies and management practices. The findings can help to improve agricultural production efficiency, reduce production costs, and promote high-quality agricultural development.

5.3. Research Shortcomings and Future Prospects

This paper discusses the role of the agricultural support and protection subsidy policy through the scale effect and technology effect, initially achieving the “win-win” of food production and quality, promoting the high-quality development of food. The results of this paper confirm that the implementation of the agricultural support protection subsidy policy has achieved “preliminary results” in the “contradiction” between food production growth and quality improvement, i.e., it is difficult to achieve the simultaneous growth of food production and quality, i.e., the implementation of the agricultural support protection policy has promoted the growth of food production and quality. The implementation of an agricultural support protection subsidy policy has a “preliminary effect,” that is, the implementation of an agricultural support protection policy has promoted the growth of food production and quality, and its driving force mainly comes from the scale effect and technology effect. The shortcoming is that it has not been possible to verify whether the scope for agricultural support subsidies to promote food quality has come at the expense of yields. In the following research, this paper will further explore whether the implementation of agricultural support and protection subsidy policy has taken away the space for food quality improvement at the expense of yield growth.