Improvement of Agricultural Supply Quality in China: Evidence from Jiangsu Province

Abstract

Promoting the reform of the agricultural supply side and its quality improvement are crucial for realizing agricultural modernization. This paper tests the varied trajectory of agricultural total factor productivity (ATFP) in Jiangsu Province over the past 21 years. The paper used the three-stage DEA empirical analysis method—Stochastic Frontier Approach (SFA), to address the uncertainty in the development of the agricultural industry. The paper introduces environmental variables such as urbanization level, import and export trade, financial support for agriculture, and transportation convenience. The research results show that: (1) the ATFP growth in Jiangsu Province presents a fluctuating trend; (2) The further sub-index research of ATFP in Jiangsu Province shows that the average rate of growth for agricultural technology efficiency (AEC) in Jiangsu is negative, indicating that the input cost of agricultural factors in Jiangsu increases and marginal efficiency decreases; (3) The empirical analysis of ATFP growth by region shows that there are still large differences in agricultural economic development level, the level of modern agricultural technology and ATFP in the southern, the central and the northern parts of Jiangsu. (4) Stage III: DEA empirical results showed that improving urbanization level, net export trade, and transportation convenience is conducive to improving agricultural production efficiency; financial support for agriculture is weakly conducive to improving agricultural production efficiency. On this basis, the paper puts forward countermeasures and suggestions to promote agricultural structural reform.

1. Introduction

The need for China to fully embark on a new modernisation process is today. Rural modernization is the foundation of national modernization and is vital to the overall modernization process. Green, high-quality agricultural products, and the improvement of supply quality are of great significance to the modernization of rural areas. China currently has 1/5 of the world’s population, with only 7% and 8% of global cultivated land and water, respectively [1]. As consumers’ need for healthy and safe agricultural products continues to rise, problems such as the oversupply of some agricultural products, shortage of high-quality agricultural products, resource degradation, and ecological and environmental pollution are becoming increasingly prominent. The reform of the agricultural supply system and its mechanism is a fundamental measure to improve the quality of agricultural supply. It is important to make up for the weak link in the mismatch between supply and demand of agricultural and rural development, and the supply structure must change with the change of the market demand structure. For the market oversupply of grain varieties, farming areas should be properly reduced while other varieties in short supply should be cultivated. Wang et al. [2] optimize the structure of agricultural products, achieve a high level of balance between the production and consumption of agriculture, improve the quality and efficiency of agriculture, establish a modern industrial agriculture system, promote rural flourishing, and realize the continual increase of peasants’ income.

Jiangsu Province in East China has consistently embraced advancing and strengthening agriculture through science and technology by maximizing its strengths as a scientific and educational powerhouse and encouraging the creative use of science and technology in agriculture. With rural revitalization and modernization, this goal seeks to ensure that the province experiences ongoing improvement in agricultural and rural development [3]. This modernization and revitalization have been spread across the province to ensure agricultural productivity. For example, there is a newly launched field precision production management system in Lanhuatang, a high-standard farmland in Nanjing, the capital city of Jiangsu, that can gather data from soil and seedling conditions and develop the best guidance plan for agricultural production based on the seedlings [4]. This system greatly reduces labor intensity, boosts productivity, and is cost-effective for agricultural inputs [5]. Jiangsu has pioneered the release mechanism for major agricultural science and technology needs in the nation and works to resolve the issue of scientific research and industrial coordination to give more agricultural production entities access to the benefits of scientific and technology innovation [6]. Again, five regional agriculture service centers have been established in Hai’an County since the year 2022 began, and these centers offer precise and focused services to unlock the effective management of agricultural technology services fully [5]. Every year, Jiangsu cultivates 50,000 scientific and technology demonstration entities and supports many eco-friendly, high-quality, and effective technologies [7]. Science and technology make up 70.9% of the province’s agricultural advancement, strengthening the province to increase its added value of agriculture, forestry, animal husbandry, and fishing from 340 billion yuan to 507.4 billion yuan [3]. Other information on the province is discussed in the literature review.

Based on the mechanism of improving agricultural supply quality, this paper’s objective is to empirically examine the changes in ATFP in Jiangsu Province and its sub-regions in the past 21 years by applying the three-stage DEA Model [8].

Taking inferences from the agricultural modernization situation in Jiangsu province, the paper puts forward countermeasures and suggestions to promote agricultural structural reform. Thus, relying on science and technological innovation, promoting the development of modern agricultural industry, implementing clean production, and integrating the development of primary, secondary, and tertiary industries. This paper provides a reference basis for the scientific decision-making for the Jiangsu provincial government and agricultural departments. It provides a valuable reference for the development of modern agriculture in other provinces of China and developing countries .

Towards the achievement of these objectives, the study’s contributions will mainly reflect in 2 aspects:

(1)

This article uses KS and T tests to observe the influence degree of different agricultural input and output indexes on agricultural production efficiency values. It selects the input-output indexes on this basis. The Malmquist productivity index of different regions is calculated with the tested variables as the input-output index. The output index increases rural residents’ per capita disposable income (POY) so that the input-output index is more objective and can truly reflect the influencing factors of agricultural TFP, filling the gap in the existing relevant research methods.

(2)

The structure of the DEA three-stage model. The paper uses the Stochastic Frontier Approach (SFA) in the second stage. It is a method of efficiency estimation using the stochastic frontier production function. The method is proposed independently by Aigner, Lovell & Schmidt (1977) and Meeusen & van den Broeck (1977), which is a parametric method [9]. Estimating the effect of external environmental variables on the efficiency values of each decision unit, the differential analysis was performed on the input variables and separated the external environmental influence and random error. This paper combines the advantages of the random envelope analysis (DEA-Malmquist) method and the stochastic frontier analysis (SFA) method to construct a three-stage DEA model. It empirically analyses the changing track of provincial, regional, and municipal ATFP in Jiangsu Province from 2000 to 2021. It makes up for the deficiencies in existing research on the empirical analysis of agricultural total factor productivity.

The main structure of this paper: First, this paper analyzes the background and significance of agricultural supply quality in China and the current situation of agricultural supply in Jiangsu. Second, this paper reviews the literature and studies the methods of improving the quality of agricultural supply and the measurement of agricultural total factor productivity. Third, a three-stage DEA model is constructed to empirically analyze the agricultural total factor productivity change in Jiangsu province and different regions from 2000 to 2021. Fourth, the results of the empirical analysis are discussed—fifth, conclusions and recommendations and provided. Based on the findings, it is imperative to maintain the rapid growth of ATFP, put forward suggestions on relying on agricultural technology and innovation, increase policy support, promote agricultural supply-side structural reform, and narrow the regional development gap.

2. Literature Review

2.1. Connotation of Improving the Quality of Agricultural Supply

Experts and scholars pointed out that China’s grain production, inventory, and import are increasing, and strengthening the agricultural supply-side reform is proposed [7]. In 2015, the grain output reached 62.1435 million tons, which led to difficulties in domestic grain supply and marketing [3]. The domestic grain imported by a large amount of high-quality foreign grain was in the dilemma of warehousing and foreign grain entering the market. In the late stage of industrialization, Jiangsu Province, with less than 4% of the country’s arable land and less than 0.06 hectare per capita, produced 5.6% of the country’s grain, 6.5% of vegetables, 3.6% of meat, and 6.5% of poultry eggs; created 6.1% of the country’s total output value of agriculture [6]. In 2022, the total output value of agriculture in Jiangsu Province reached 873.39 billion CNY, an increase of 3.9%, one of the higher rates in recent years . The total output value for the modern planting and breeding industry in Jiangsu Province reached 1 trillion yuan, and the total output value for the agricultural product processing industry reached 1.5 trillion yuan. In 2016, the total output value of agriculture in Jiangsu Province was 717.896 billion yuan, reaching 827.972 billion yuan in 2021, with an average annual rate of 3.07% [7].

The annual total grain output reached 37.69 million tons, a record high for five consecutive years, an increase of 2.08 million tons over 2015, ranking 8th in China (See Figure 1).

Moreover, ensuring quality agricultural products does not only guarantee basic food safety but also the internal requirement for developing green production. In establishing a national production quality and safety demonstration, Jiangsu province took the lead in establishing a five-level agricultural product quality and safety supervision system at the villages, townships, counties, and municipal and provincial levels. The proportion of green and superior foods reached 70.4%, and the total number of green food and organic produce reached 2981 Species types, with more than 5700 kinds of products and more than 2600 enterprises, ranking first in the total scale of the country. However, we should see that the problems on the supply side of agriculture still exist: High agricultural costs trigger government subsidies, which invert food prices and lead to a vicious circle of high inventories [3]. At the same time, consumers have a stronger demand for healthy and high-quality agricultural products [4]. The contradiction of agricultural product supply structure is prominent, and the continuous increment of farmers’ income is weak [10]. As a result, optimizing the supply structure and improving the quality of agricultural products is arduous.

Moreover, the increase in agricultural chemical inputs has led to the rising cost of agricultural production, the degradation of the agricultural ecological environment, and other problems, making the promotion of agricultural supply-side reform significant. Again, enhancing the quality of the agricultural supply system is to optimize the allocation of agricultural factors and accelerate agricultural modernization through the agricultural supply-side structural reform [6]. It is essential to realize the adjustment and change of agricultural production mode and promote cleaner production. It is also necessary to reform the “structural” issues, such as various business entities and types of agricultural operations. This includes the variety, quality, and quantity of related agricultural products, improving the system and mechanism of green development, and improving the supply of green high-end and superior produce [10].

2.2. Problems and Improving Factors on Agricultural Supply Quality

Scholars pointed out the five main problems restricting the supply side of China’s industry. The first is the supply side of products facing the structural imbalance of effective supply shortage and high inventory. China’s grain production, inventory, and import are increasing at the same time. Also, agricultural products are large in amount but poor in quality, with the majority of low-end products in quantity and few high-quality varieties. For example, the proportion of high-quality rice in Jiangsu province is still not high [11]. In 2022, the proportion of grain output in southern Jiangsu, central Jiangsu, and northern Jiangsu was 1:3:6 [6]. Wheat and rice accounted for 88%, but the proportion of good taste rice varieties accounted for only 40% of the total area of rice [12]. The over-application of chemical fertilizers and pesticides is unsuitable for developing green and modern agriculture, which does not meet the needs of consumers for superior, green, and safe agricultural products.

Additionally, it negatively impacts living beings’ survivability [13]. It is necessary to adjust the agricultural production structure and introduce a wide range of nutritious and resilient crops in line with the principle of comparative advantage. This will give full play to industrial benefits to diversify the agri-food system. It will also be driven by growing market demand for high-quality products. Since 2020, the COVID-19 pandemic has aggravated the imbalance between agricultural products’ imports and exports. China’s imports of agricultural products from the USA in 2020 reached 162.74 billion yuan, an increase of 66.9% from 2020, among which soybeans, pork, and cotton increased by 56.3%, 223.8%, and 121.7%, respectively [14].

Jiangsu province has a high grain inventory. In 2020, the price of corn in Jiangsu exceeded 3.0 yuan/kg, the price of wheat exceeded 2.6 yuan/kg, and the price of rice exceeded 3.2 yuan/kg, which is the highest value in recent years, and the international grain price difference of 0.6 yuan per kilogram . In addition to the export quota restrictions and the price “ceiling”, some grain depots in Jiangsu province are at full capacity. The second is the agricultural “structural distortion” phenomenon and production “environmental negative effect” mainly caused by the agricultural supply-side problem. Experts have studied and constructed an unbalanced indicator system of agricultural supply structure, including 20 indicators such as effective supply, factor allocation, green development, and farmers’ income increase [14]. The supply side of produce shows three problems, according to [15]. The shortage of agricultural arable land and water resources, and the non-point source pollution of the agricultural ecological environment; agricultural products on the supply side lack international competitiveness in quality and price; innovation in agricultural technology are insufficient. Third is the total balance of agricultural products that have changed from the binary equilibrium of production and demand to the ternary balance of production, demand, and import. The agricultural infrastructure is not perfect. The supply is insufficient, the per capita cultivated land resource is less than 0.1 hectares, and the scale of farming production is limited [16]. Again, the supply of essential modern factors plays a prominent role in farm development, such as finance, insurance, intelligent equipment, etc. In developing modern agriculture, informatization, digitalization, intelligence, and other aspects have become “new weaknesses”.

Urbanization is a global trend. As far as China is concerned, rapid urbanization and opening in 1978 have resulted in rural-to-urban migration, loss of arable land, environmental degradation, and new dietary demands on food production [17]. The fourth problem is that China’s agricultural product market policy choices are subject to more and more external constraints, which has put the agricultural product pricing policy implemented for many years into a dilemma. The tariff policy is strengthened by the commitment to enter the WTO, and the tariff level does not match the gap in China’s basic competitiveness [18]. The fifth and final problem is that agricultural science and technology progress lags. For a long time, technological innovation has not contributed much to China’s agricultural growth [5]. The extensive growth relying solely on labor and capital led to an imbalance between the supply and need of agricultural goods [3]. The development of green agriculture relies on technological innovation, while the level of agricultural machinery and equipment, the ability to resist natural disasters, the strength of new technology promotion, and the degree of scientific and technology transformation in China are still weak, and scientific research cannot be combined with actual products [16]. At the same time, Chinese farmers’ understanding of agricultural science and technology is not deep enough and has not been extensively applied in rural areas [19].

2.3. Research on the Method of TFP

Wen (1993) [20], Ji (2016) [21], and Kuosmanen, T et al. (2009) [22] measured the national agricultural total factor productivity by the DEA-Malmquist index method, production function method and the window DEA-Malmquist index method, respectively. They believe that the growth of China’s agricultural total factor productivity is mainly due to the progress of cutting-edge agricultural technology, while the agricultural technology efficiency is deteriorating. Wang et al. [23] calculated the agricultural total factor productivity of various provinces in China by using SBM directional distance function and the Luenberger productivity index. They found that technological progress and scale change are the main reasons for improving agricultural total factor productivity in China. Experts [24] and scholars [9,25] mostly use four methods to measure agricultural total factor productivity: the production functional relationship method, the growth adjust accounting index method, the stochastic frontier approach (SFA), and the DEA-Malmquist index method. Characteristics of the above methods are shown in Section 2.3.1, Section 2.3.2, Section 2.3.3. Empirical analysis shows that if China’s agricultural economy achieves a higher growth level, it needs to change from an investment-dependent growth type to a technology-dependent Kuznets growth type [4,24]. It needs to improve factor productivity, reduce the factor input costs, moderate scale operation, reduce the input of agricultural chemical products, and improve the utilization rate and the competitiveness of the grain market [24]. Outside agricultural supply quality, various studies have used the DEA method in assessing total factor productivity in other fields.

Table 1. Summary of the empirical literature on DEA approaches in supply quality.

Authors	Objective	Data and Method Used
Wen [20]	Factor Productivity Change in China’s Farming Sector	DEA-Malmquist index Data: Related to agricultural economic development in 1952–1989
Ji et al. [21]	Agricultural total factor productivity growth	DEA-Malmquist index Data: Related to agricultural economic development in 2000–2014
Wang et al. [23]	Agricultural Efficiency and Total Factor Productivity Growth in Various	SBM, Luenberger productivity index Data China’s agricultural economic data from 1978 to 2011
Shabanpour et al. [26]	Future planning for benchmarking and ranking sustainable suppliers	Deterministic, robust double frontiers DEA
Jauhar and Pant [27]	Evaluating sustainable suppliers	Deterministic, DEA with DE and MODE
Li et al. [28]	Agriculture Total-Factor Energy Efficiency	DEA-Malmquist index Data: Related to energy development in 1993–2014
Tavassoli and Farzipoor [29]	Proposed new stochastic super-efficiency	Stochastic and deterministic
Mohammad Tavassoli [30]	Evaluating and ranking sustainable suppliers in unified formwork	Deterministic, stochastic, and fuzzy-DEA
Li et al. [31]	Performance of Metabolism in China	Deterministic, DEA-Malmquist Data: the emergy evaluation indicators system with 23 indicators in 2009–2015
Zheng [32]	Energy efficiency evaluation	Deterministic, DEA-SBM-Malmquist index Data: emergy development in 2000–2019
Molinos-Senante and Maziotis [33]	Benchmarking the efficiency of water and sewerage companies	DEA-StochasticData: water resource use in 2000–2019
Ding et al. [34]	Assessing industrial circular economy performance and its dynamic evolution	Deterministic, An extended DEA-Malmquist index Data: China’s Yangtze River Delta region over 2012–2017
Chen et al. [35]	A three-stage SBM-DEA model with non-point source pollution and CO2 emissions	DEA combined with the Slack-Based Measure (SBM) Data: influences of environmental factors and random errors and explore the real AGTFP of 30 provinces in China from 2000 to 2017.
Pokharel and Featherstone [36]	Examining the productivity growth of agricultural Cooperatives	Biennial Malmquist Index (BMI)
Khoshroo et al. [37]	A new double frontier-based Malmquist productivity index	Deterministic, DEA-Malmquist index
Wanke et at. [38]	an approach based on generalized auto-regressive moving averages	Stochastic DEA-ratio
Pourmahmoud and Bagheri [39]	Evaluating healthcare systems during the COVID-19 Pandemic	BCC-Malmquist-DEA
This paper	Proposed BCC-DEA-Malmquist Index in Agricultural supply quality	Stage III: DEA.: Data: Related to agricultural economic development in 2000–2021

reviews some of the relevant literature that used similar methods.

2.3.1. Parameteranalys-Production Function Method

(1) Characteristic. The Experts apply the C-D production function model from Kumar et al. and Los et al. The source of economic growth is decomposed into three parts: technology progress, technology catch-up (technology efficiency), and input factor accumulation. As technology changes and its efficiency change, capital and labor are introduced into the agricultural production model to construct agricultural production function:

$$\ln Y_{it} = {\beta }_0+{\beta }_{1}T{P}_{it}+{\beta }_{2}E{C}_{it}+{\beta }_3\ln L_{it}+{\beta }_4\ln A_{it}+{\beta }_5\ln M_{it}+{\beta }_6\ln F_{it}+v_i+{\epsilon }_{it}$$

(2) Application result analysis. The model starts from a production perspective, focuses on the impact of agricultural technology progress and efficiency change on the agricultural economy in different regions, and is compared with other input elements. Taking Jiangsu Province as an example, the difference in total factor productivity in southern Jiangsu central Jiangsu is compared with other input elements. Taking Jiangsu Province as an example, the difference in total factor productivity in southern Jiangsu, central Jiangsu, and northern Jiangsu has been compared and analyzed [21].

2.3.2. Stochastic Frontier Approach (SFA)

(1) Agriculture is an industry with noise. Consistent with the production nature of the agricultural industry, the Stochastic front approach can use different frontiers for different samples. So, it has an advantage in measuring the absolute efficiency of total agricultural factor production. It can distinguish between the statistical and management error terms and make assumptions about the model. The accuracy of the assessment of agricultural total factor production efficiency will not be affected by random error. It distinguishes between productivity, random noise, and error terms. Therefore, the estimation results of the SFA are more accurate. However, the SFA is only applicable for single output input circumstances. There are defects in the case of more input and more output.

(2) Application result analysis. Lv & Meng [9] use the three-stage DEA model to calculate the agricultural production efficiency in Jiangsu Province. Combined with the cluster analysis method, the production efficiency of all the cities in Jiangsu Province is divided into three types. It can be seen that external environmental factors and random factors have certain effects on agricultural production efficiency influence. The improvement of urbanization level, net export, and the convenience of traffic are the positive factors to agricultural production efficiency, while financial support in agriculture is the negative factor to agricultural production efficiency. Governments of each region in Jiangsu Province should combine farming with the characteristics of its efficiency, focusing on improving the management level and expanding the scale of agricultural production to ameliorate the agricultural production efficiency.

2.3.3. Data Envelopment Analysis (DEA)

(1) Characteristic. DEA is a well-known family of mathematical programming tools for assessing the relative efficiency of a set of comparable processing units (a.k.a. decision-making units, DMU). One of the strong points of DEA is its non-parametric character, which means that only the observed input consumption values and output production amounts are needed to assess the relative efficiencies of the DMU properly. The way to do this is by extrapolating, from the observed sample of inputs and outputs, a set of possible operating points, assuming some technology. The most common technologies are constant return to scale (CRS) and variable return to scale (VRS). Both consider linear combinations of the inputs and outputs of the existing DMU [25].

(2) Application result analysis. Agricultural TFP continues to grow, but the factor contribution varies. From the measurement results of the geometric average and the cumulative index, agricultural TFP across the country and in the eastern, central, and western regions showed varying degrees of growth. It makes up for the TFP loss caused by decreased technology efficiency. The regional differences and interprovincial differentiation of agricultural TFP growth and its composition changes were obvious. The eastern region grew the fastest, the main force driving the national agricultural TFP growth. The central and western regions lack technological progress, and TFP growth is weak. As a traditional big agricultural province and a major grain-producing area, the central region reduces the technology efficiency, while the delay of technology progress affects the growth of agricultural TFP. In addition, the interprovincial differentiation of agricultural TFP growth is also relatively obvious. TFP growth in most eastern provinces showed a strong trend. TFP growth in the central and western provinces is mostly lower than the national average and even at the lowest level [31].

2.3.4. The SBM Directionality Distance Function

(1) Characteristic. With an input or output non-zero relaxation (Slack), the current literature using DEA is mainly based on the radial direction, the angle of the (oriented) traditional method. On the one hand, with the presence of relaxation variables, the radial DEA will overestimate the efficiency of the evaluation object, making the calculation result inaccurate. On the other hand, the angle of DEA needs to choose whether to calculate the efficiency value based on input or output orientation. Both the input and the output aspects cannot be considered simultaneously, leading to the distortion of the efficiency value. The non-radial, non-angular SBM (Slack-based Measure, SBM) directional distance function can overcome the above defects [32].

(2) Application result analysis. To adapt to SBM and directional distance function, a new productivity measurement method—the Luenberger productivity index has been used to measure agricultural total factor productivity in each province of China.

The results show that the level of agricultural inefficiency in the eastern regions was significantly lower than that in the central and western regions, unproductive output; inefficient draft animal input and inefficient sown area are the main sources of agricultural inefficiency in China; unproductive output, inefficient draft animal input, and inefficient sown area are the main sources of agricultural inefficiency in China; from 1995 to 2008, the total factor productivity growth rate of China’s agriculture was 5.58%, mainly reflected in the technological progress and scale change, the eastern region of the agricultural total factor productivity is the highest; the eastern region has the highest agricultural total factor productivity. The improvement of the education level of agricultural practitioners promotes the improvement of agricultural efficiency and total factor productivity in China; the popularization of mechanization is conducive to the growth of agricultural total factor productivity in China [23].

2.4. Limitations of Existing Studies

2.4.1. External Environmental Impact and Uncertainty

From the existing research, there are still some problems worth further discussion: The production environment of agricultural production is often very complex, and the production cycle is very long, which will inevitably be affected by external factors and other uncertain factors. Agricultural total factor productivity is inevitably affected by these factors when measuring the input-output efficiency of agricultural factors. However, most of the literature studies do not consider external environmental problems, and the relevant conclusions are more or less mixed with the interference of external environmental factors, so the research results are different. How will external factors affect agricultural total factor productivity and its influence degree? How should we eliminate the influence of external environmental factors and investigate the growth of agricultural total factor productivity? These questions are worth studying.

2.4.2. Study of the Measurement Methods of Agricultural Total Factor Productivity Growth

The stochastic envelope analysis (DEA-Malmquist) method and the stochastic frontier approach (SFA) are the most widely used. However, the efficiency growth value measured by the former will not be affected by external factors. Although the latter considers external factors, it is limited to eliminating the influence of random interference, and there are some defects in the measurement technology.

The three-stage DEA-Malmquist method effectively combines the advantages of the above two methods and can fully consider external environmental factors. At present, it has enriched successful experience in other related fields, but few literature has noticed the application of this method in the measure of agricultural total factor productivity growth.

2.4.3. Major Research Contributions in This Paper

To address the uncertainty in the development of the agricultural industry, the paper used the three-stage DEA empirical analysis method-Stochastic Frontier Approach (SFA). The paper introduces the environmental variables, such as urbanization level, import and export trade, financial support for agriculture, and transportation convenience, in the second-stage SFA model, estimating the effect of external environmental variables on the efficiency values of each decision unit, the differential analysis was performed on the input variables and separated the external environmental influence and random error, it fills the gap in the existing relevant empirical research.

The empirical results of SFA indicate that the improvement of urbanization level, net export trade and transportation convenience is conducive to improving agricultural production efficiency; financial support for agriculture is not conducive to improving agricultural production efficiency. All regions in Jiangsu province should combine the local economic level and resource endowment and improve the management level or adjust the scale of agricultural production to improve agricultural production efficiency [40].

2.5. Hypotheses

To achieve this study’s objective, some basic theoretical hypotheses have been formulated based on the mechanism for improving the quality of agricultural supply. The fundamental way to improve the quality of agricultural supply is to promote the optimization of industrial structure and the upgrading of product quality through supply-side structural reform. The total factor productivity in agriculture (AFP) plays a significant role.

Hypothesis 1: 

Agricultural total factor productivity in Jiangsu province shows a rising trend from 2000–2021 [41].

Hypothesis 2: 

Agricultural production has noise and uncertainty, which was affected by environmental factors and statistical noise. The SFA model separates the influencing factors and proposes the path and measures for upgrading the agricultural supply structure in Jiangsu Province.

Hypothesis 3: 

Improving agricultural total factor productivity has an obvious positive effect on optimising agricultural structure and economic growth in Jiangsu Province [41].

The mechanism of the main path is to enhance effective institutional supply. It adjusts the mismatch between the supply and demand of agricultural products and meets consumers to seek green, diversified, personalized products. The production of green and high-quality agricultural products, such as rice, wheat, soybeans, vegetables, and fruits, has increased [42] (See Figure 2 for its mechanism).

3. Materials and Methods

3.1. Research Variables and Data Sources

3.1.1. Input and Output Variables

This paper uses the data of agricultural input and output indicators in Jiangsu Province from 2000 to 2021 to calculate the ATFP of different regions. The current literature has great controversy on this issue, and its conclusions are not uniform and comparable. To improve the effectiveness of this kind of study, this paper first takes all the variables mentioned in the existing literature into consideration. It uses the KS test and T-test methods to eliminate the factors that have less impact gradually. The agricultural output indicators adopted include gross agricultural production (AGDP), grain output (AF), and annual disposable income of rural residents (POY). The agricultural input indicators adopted include total sown area (ACUL), grain sown area (AFCUL), agricultural effective irrigation area (IRR), number of total agricultural employees (LAB), the total power of agricultural machinery (MACH), the amount of agricultural chemical fertilizer input (FER), and rural electricity consumption (ELEC) (see

Table 2. The agricultural input and output indicators.

Variable Name	Description	Variable Type
Gross agricultural production (AGDP)	The output of agriculture, forestry, animal husbandry, and fishery products and their by-products within one year is multiplied by the price of their respective unit products.	Output Indicators
Grain output (AF)	The planting area of rice, wheat, corn, soybean, and sorghum is multiplied by yield per unit area.	Output Indicators
Annual disposable income of rural residents (POY)	In one year, the sum of the wage income, net operating income, net property income, and net transfer income of individual rural residents.	Output Indicators
Total sown area (ACUL)	The total area of food crops, such as grains, legumes, and potatoes sown throughout the year.	Input Indicators
Agricultural effective irrigation area (IRR)	For the sum of irrigated land areas in paddy fields and dry land that can be irrigated normally.	Input Indicators
Number of total agricultural employees (LAB)	The labor force that the whole society, directly participates in the production activities of agriculture, forestry, animal husbandry, and fishery.	Input Indicators
The total power of agricultural machinery (MACH)	Including tillage machinery, drainage and irrigation machinery, harvesting machinery, agricultural transportation machinery, plant protection machinery, animal husbandry machinery, forestry machinery, fishery machinery, and other agricultural machinery, internal combustion engine by engine horsepower into tile (special) calculation, motor by power into watt calculation.	Input Indicators
The amount of agricultural chemical fertilizer input (FER)	The total amount of nitrogen, phosphorus, potassium fertilizer, and compound fertilizer used per year; the application amount should be calculated according to the discounted purity amount.	Input Indicators
Rural electricity consumption (ELEC)	The annual total electricity consumption of rural production and living in the current year after deducting the electricity consumption of state-owned industry, transportation, and infrastructure units in rural areas.	Input Indicators

).

The data of this study are mainly from the Statistical Yearbook of NBS and Jiangsu Province from 2001 to 2022 (http://www.stats.gov.cn, accessed on 18 June 2023), the Statistical Yearbook of Rural Economic Development [43], and the official data of the FAO [44].

3.1.2. Environmental Factors

These external factors have a certain impact on agriculture Production efficiency and will benefit it. The specific analysis is as follows:

(1)

The level of urbanization. It is expressed as a proportion of the urban population to the total population. The increased level of urbanization implies an increase in the opportunity cost of labor, a tight supply of factor resources, and the need for agricultural production to move towards intensification, which is conducive to an increase in agricultural production efficiency.

(2)

Financial expenditure. This is a calculation of the amount of expenditure (in billion yuan) on agriculture, forestry, and water affairs in the budget expenditure of each region. Financial support for agriculture can increase farmers’ motivation to cultivate land and reduce fallow and abandoned land.

(3)

Import and export trade. It represents the total amount of goods entering and leaving the country in each region (unit: billion yuan). Import and export trade has broadened the distribution channels for agricultural products, with broader sales markets, enhancing the market value of agricultural products and enabling higher utilization of agricultural production factors.

(4)

Transportation convenience. The development of transport infrastructure facilitates the movement of talent and reduces the “time distance” between businesses, which helps to generate positive externalities and improve the output of production factors. The density of the road network (total miles of graded roads/total area of the region) is chosen to measure accessibility (unit: km/km²).

3.2. Selection of Research Variables

In the non-parametric DEA analysis, the selection of input-output indicators is the premise and key for the analysis. Different input-output indicators will bring about differences in the relative efficiency of decision units. Charnes et al. [45] pointed out that the number of decision units in the non-parametric DEA analysis was at least twice the input-output index. If more input-output indicators are used, the number of relatively effective decision-making units will be relatively large, reducing the degree of differentiation [46]. To avoid this problem in DEA analysis, this paper draws on multiple tests proposed by Banker to measure the difference in efficiency values between models [47], including Kolmogorov-Smirnov (KS) and T-tests. Based on these methods, they can test whether the distribution of two populations is significantly different. The input-output index can be selected objectively. Taking the set of two sets of efficiency estimates as an example, the null hypothesis of the KS test is no significant difference in the distribution of the set of two sets of efficiency estimates value. The known set A is the set of efficiency estimates value with special variables, denoted as $\stackrel{}{{\widehat{E}}_a}$, the set B is the set of efficiency estimates value with no special variables, written as ${\widehat{E}}_b$, the cumulative distribution function of two sets of efficiency values is denoted as $S_a$(E) and $S_b$(E) respectively, and the test statistic is: $MaxE=\{S_a\left(E\right)-S_a\left(E\right)\}$.

$$\mathrm{r}\left[D>{\left(\frac{n^a{n}^b}{n^a+n^b}\right)}^{\frac{1}{2}}z\right]=e^{-2{\mathrm{z}}^2}(\mathrm{z}>0)$$

According to the KS test and T-test, indicators set are screened. Firstly, we set the benchmark model, and under the condition of variable returns to scale (VRS), all the above input and output indicators are included to get the numerical efficiency collection. Secondly, we remove the assumptions and variables in the benchmark model step by step to get another numerical collection. Finally, we compare the numerical collection according to the significant difference between the distribution and the mean value to determine whether the variables remain in the benchmark.

3.3. Uncertainty in the DEA-Malmquist Model

From the research method of measuring ATFP, parametric and non-parametric methods measure agricultural production technology progress and production efficiencies. Since the agricultural system is a multi-input and multi-output system, the real data is both certain and affected by uncertainty. It is difficult to determine the importance of the inputs when the parametric approach is not objective. This can affect the reliability and robustness of the results. This uncertainty issue makes measures to eradicate data uncertainties very important. When estimating parameters, it is necessary to assume the functional form and the distribution of error terms, and it cannot effectively deal with the case of multiple outputs. At the same time, the obtained TFP results are highly dependent on the input factor and output elasticity. If the capital elasticity coefficient is set artificially, the subjective factors of its parameter value have a greater impact. If regression analysis is used to determine the coefficient, the parameter value will rely heavily on the sample data, and it is difficult to avoid the impact of data errors [48]. The non-parametric method does not need to set the form of the production function and does not require the data of input factor price. In agricultural production, on the one hand, it is hard to obtain data on factor prices; on the other hand, the prices of some factors will be distorted due to government intervention. Therefore, this paper uses the three-stage DEA Model to measure agricultural TFP and decomposes the index into technology progress and efficiency change. A method proposed by Fried et al. can better evaluate the efficiency of DMU (Decision Making Unit, decision unit). It can effectively eliminate the interference of the external environment and random error in efficiency measurement, and its construction and application include three stages [49].

3.4. Stage I: Traditional DEA model (BCC Model)

Utilizing the DEA-Malmquist index method, efficiency is analyzed and computed. Since all conventional DEA models use section data from a single year as their analytic sample, they can only reflect the relative levels of efficiency of various economies over the same time. They cannot examine how these levels have changed over time. The panel data, which cannot only measure the total factor efficiency into technology efficiency, technology progress, and scale efficiency, can be analyzed by the DEA-Malmquist index model to clarify the reasons for the total factor productivity change of the research subject. [50] For the estimation of production frontiers, such as in agriculture, the DEA nonparametric technique is used in operations research and economics. It is used to experimentally assess the decision-making units’ (DMUs’) production efficiency. The method is used for benchmarking in operations management, where a set of indicators is chosen to benchmark the performance of manufacturing and service operations, despite having a strong connection to production theory in economics [50]. In benchmarking, the effective DMUs may instead result in a “best-practice frontier” rather than a “production frontier”. The DEA model is the method to evaluate the relative efficiency of the same unit or multi-input and multi-output economic system [51]. The Malmquist productivity efficiency index is used to examine the dynamical production efficiency between the multi-input and multi-output variables across a period and test the change of TFP. Its model is:

$$M_0(x_t,y_t,x_{t+1},y_{t+1})=\left[\frac{D_0^{t+1}(x_{t+1},y_{t+1})}{D_0^{t+1}(x_t,y_t)}\times \frac{D_0^t(x_{t+1},y_{t+1})}{D_0^t(x_t,y_t)}\right]$$

(1)

$D_0^t\left(x_t,y_t\right)D_0^t\left(x_t,y_t\right)$ and $D_0^{t+1}\left(x_t,y_t\right)D_0^{t+1}\left(x_t,y_t\right)$ are the input distance functions obtained by comparing the production point with the frontier technology at the same time, that is, (t and t + 1) $D_0^t\left(x_t,y_t\right)D_0^t\left(x_t,y_t\right)$, and $D_0^{t+1}\left(x_t,y_t\right)D_0^{t+1}\left(x_t,y_t\right)$ are the input distance functions gained by comparing the production point with the frontier technology during the mixing period, which can be further decomposed into the product of the technology change index EC and the technology progress index TC, that is, $TFP=EC\times TC$, among of which:

$$EC=\frac{D^{t+1}(x^{t+1},y^{t+1})}{D^t(x^t,y^t)}$$

(2)

$$TC=\sqrt{\frac{D^{t+1}(x^{t+1},y^{t+1})\times D^t(x^t,y^t)}{D^{t+1}(x^{t+1},y^{t+1})\times D^{t+1}(x^t,y^t)}}$$

(3)

Under the condition of allowing the variable scale revenue, the technology variable efficiency ($EC$) variable index can be further decomposed into the product of pure technology efficiency change index ($PTE$) and scale efficiency change index ($SE$), that is, $EC=PTE\times SE$, among of which:

$$PTE=\frac{{{\displaystyle D}}^{t+1}({{\displaystyle x}}^{t+1},{{\displaystyle y}}^{t+1}|VRS)}{{{\displaystyle D}}^t({{\displaystyle x}}^t,{{\displaystyle y}}^t|VRS)}$$

(4)

$$SE=\frac{{{\displaystyle D}}^{t+1}({{\displaystyle x}}^{t+1},{{\displaystyle y}}^{t+1}|CRS)}{{{\displaystyle D}}^{t+1}({{\displaystyle x}}^{t+1},{{\displaystyle y}}^{t+1}|VRS)}\times \frac{{{\displaystyle D}}^t({{\displaystyle x}}^t,{{\displaystyle y}}^t|VRS)}{{{\displaystyle D}}^t({{\displaystyle x}}^t,{{\displaystyle y}}^t|CRS)}$$

(5)

It can be concluded from the above the Malmquist index can be eventually decomposed into:

$$M(x^{t+1},y^{t+1},x^t,y^t)=EC\times TC=PTE\times SE\times TC$$

(6)

If TFP > 1, productivity shows a growth trend; on the contrary, if TFP < 1, it means that productivity is declining.

3.4.1. Technology Efficiency Change Index (EC)

Including the pure technology efficiency variable index (PTE) and the scale efficiency variable index (SE). TE is a measure of the comprehensive efficiency of the decision-making unit. PTE is a measure of the resource allocation efficiency of the decision-making unit. SE measures the scale efficiency of resources invested by decision-making units. With the return of scale unchanged, the catching-up degree of the production possibility boundary of each DMU from period t to t + 1 is measured. EC > 1 indicates improvement in technology efficiency If EC < 1 means the technology efficiency decreases.

3.4.2. Index of Change in Technology Progress (TC) and the Scale Efficiency Index (SE)

TC indicates the degree of innovation of production technology changes. If TC > 1, it means that the production technology efficiency has increased. If TC < 1, production technology has declined. SE indicates the degree to which DMU approaches the optimal scale in the long run. If SE > 1, DMU is close to the optimal scale in the long term; SE < 1 indicates deviation from the long-term optimal scale.

3.5. Stage II: Construction of the SFA Model

The main task of the second stage of model construction is to estimate the impact of external environmental variables on each decision unit’s efficiency values, perform a difference-in-difference analysis of the input variables, and strip out the two exogenous factors of external environmental effects and random errors.

Fried et al. (2002) believed that the magnitude of the role of external environmental effects, random errors, and internal management efficiency factors could be measured separately [52] by constructing an SFA model to derive the redundancy of DMU inputs due to internal management inefficiencies only.

The second stage of the analysis will build on Fried’s ideas. To correspond to the input-based DEA model of the first stage, an input-oriented SFA model is developed, and the N input difference variables of M decision units are analyzed separately. The input difference variables expression is as follows:

$$s_{ij}=x_{ij}-x_{ij}^{*}\left(i=1,2,\dots ,M,j=1,2,\dots ,N\right)$$

(7)

In the formula: $s_{ij}\ge 0$, $x_{ij}$ is the actual value of the jth input of the decision unit $i$, and $x_{ij}^{*}$ is the ideal value of the $j\mathrm{th}$ input of the decision unit $i$. Using the $s_{ij}$ obtained in the first stage as the dependent variable and the external environmental explanatory factor $z_i$ as the independent variable, construct the SFA regression equation:

$$s_{ij}=f_j\left(z_i,\beta \right)+v_{ij}+u_{ij}\left(i=1,2,\dots ,M,j=1,2,\dots ,N\right)$$

(8)

In the equation: $f_j\left(z_i,\beta \right)$ indicates the impact of external environmental variables on the i input difference value $s_{ij}$; Generally, take $f_j\left(z_i,\beta \right)=z_i^T\beta$, $\beta$ indicates the regression coefficient of the external environment explanatory variable $z_i$, $v_{ij}$ represents a random statistical error, and assumes to follow a normal distribution; $u_{ij}$ represents an error due to technology inefficiency, and assumed to obey truncated normality distribution; $v_{ij}+u_{ij}$ is a composite random term. Assuming $v_{ij}+N\left(0,{\sigma }_{jv}^2\right)$ represents the random error of the $j\mathrm{th}$ input of decision unit $i$, or $u_{ij}-N^{+}\left({\mu }_j,{\sigma }_{j\mathrm{u}}^2\right)$ is the perturbation term for internal management inefficiency, and that the $u_{ij}\ge 0$ reflect managerial inefficiency. $v_{ij}$ and $u_{ij}$ is not relevant. As seen from the setting of the above random term, the influence of external factors on production efficiency was excluded. First, separate the random disturbance term from the compound error term. According to the conditions of managing inefficiency:

$$\widehat{E}\left[u|v+u\right]={\widehat{\mu }}_{*}+{\widehat{\sigma }}_{*}\frac{f\left(-{\widehat{\mu }}_{*}/{\widehat{\sigma }}_{*}\right)}{1-F\left(-{\widehat{\mu }}_{*}/{\widehat{\sigma }}_{*}\right)}$$

(9)

In the equation: ${\widehat{\mu }}_{*}=-{\widehat{\sigma }}_u\widehat{\epsilon }/{\widehat{\sigma }}^2$,${\widehat{\sigma }}_{*}={{\displaystyle \widehat{\sigma }}}_{{}_u}{\widehat{\sigma }}_v/{\widehat{\sigma }}^2$, and ${\widehat{\sigma }}^2={\widehat{\sigma }}_u^2+{\widehat{\sigma }}_v^2$; $f$ and $F$ represent the density function and cumulative distribution function of the standard normal distribution, respectively. An estimate of the random error can be obtained:

$$\widehat{E}\left[v_{ij}|v_{ij}+u_{ij}\right]=s_{ij}-z_i\widehat{\beta }-\widehat{E}[u_{ij}|v_{ij}+u_{ij}] \left(i=1,2,\dots ,M,j=1,2,\dots ,N\right)$$

(10)

After using the SFA equations for regression analysis, the impact of external environmental effects on production efficiency $z_i\widehat{\beta }$ and the impact of random errors $\widehat{E}\left[v_{ij}|v_{ij}+u_{ij}\right]$ can be obtained separately. Therefore, the interference of external factors on input can be adjusted to measure the actual production efficiency of each decision-making unit objectively. Based on the most effective DMU, the input number of other samples based on its input amount has been adjusted. The adjusted input number of elements is as follows:

$${x^{\prime }}_{ij}=x_{ij}+\left[{\max }_i\left\{z_i\widehat{\beta }\right\}-z_i\widehat{\beta }\right]+\left[{\max }_i\left\{{\widehat{v}}_{ij}\right\}-{\widehat{u}}_{ij}\right]\left(i=1,2,\dots ,M,j=1,2,\dots ,N\right)$$

(11)

The formula ${\max }_i\left\{z_i\widehat{\beta }\right\}-z_i\widehat{\beta }$ aims to eliminate the influence of external environmental effects, while the function ${\max }_i\left\{{\widehat{v}}_{ij}\right\}-{\stackrel{\wedge }{u}}_{ij}$ is to place all DMUs under the same external environmental conditions, facing the same random factors and production at the same level. Equation (10) indicates that the final difference in production efficiency is determined by internal management factors.

3.6. Stage III: Adjusted DEA Model

The adjusted DMU, the input amount ${x^{\prime }}_{ij}$, and the original output value $x_{ij}$ are evaluated again using the two DEA models in the first stage. The production efficiency value of each DMU is the efficiency value after removing external environment effects and random error [52].

4. Results and Discussion

4.1. Descriptive Statistics

The descriptive statistics of the variables used can be seen in

Table 3. Descriptive statistical analysis of the variables.

Variables	N	Mean	Maximum	Minimum	Std. Dev.
LAB	286	70.92885	233.14	15.03	0.4722
ACUL	286	587.4057	1472.02	129.8	3.54.04
MCH	286	310.2785	778.75	41.44	1.8127
FER	286	24.8415	70.3405	4.48	0.1790
ELEC	286	104.1644	656.31	1.9691	1.39
AGDP	286	371.1113	1311.61	39.32	2.6764
POY	286	12885.6	41487	2597	86.0807

NB: VRS is variable returns, to sale, AGDP is gross agricultural production, AF is grain output, POY is the annual disposable income of rural residents, ACUL is a total sown area, AFCUL is grain sown area, IRR is agricultural effective irrigation area, LAB is the number of total agricultural employees, MACH is the total power of agricultural machinery, and ELEC is and rural electricity consumption.

. It shows their mean, maximum, minimum, and standard deviation values. According to the descriptive results, the results show large differences between different areas. So, it is necessary to test for the stationarity of the panel data.

4.2. Unit Root Test Results

Following the larger differences between different areas within the data, it was necessary to conduct a unit root test to determine the stationarity of the variables. The LLC test was conducted to determine the unit-roots of the variables. According to

Table 4. LLC test results.

Variable	LLC
LAB	−11.3005 (0.0000)
ACUL	1.8518 (0.0000)
MACH	−1.6608 (0.0000)
FER	−1.8441 (0.0000)
ELEC	1.6247 (0.0000)
AGDP	−1.5997 (0.0000)
POY	−1.5139 (0.0000)

NB: VRS is variable returns, to sale, AGDP is gross agricultural production, AF is grain output, POY is the annual disposable income of rural residents, ACUL is a total sown area, AFCUL is grain sown area, IRR is agricultural effective irrigation area, LAB is the number of total agricultural employees, MACH is the total power of agricultural machinery, and ELEC is and rural electricity consumption.

, it can be seen that all variables passed the LLC test at the 1% significance level. This confirms that the variables are stationary and can be used for further analysis.

4.3. Results of the Variable Selection

Firstly, we removed the assumption of the variable return to scale, and a set of efficiency values is obtained under the assumption that the return of scale is constant. Comparing this set with the efficiency set of the benchmark model, the mean changed by 0.001. Both the KS and T tests considered that the hypothetical efficiency values were not significantly changed after removing VRS, so the assumption of CRS can be adopted in this paper’s analysis. Secondly, AGDP and AF have been separately removed. The result shows that AGDP rejects the null hypothesis at 5% and 1% significant levels. The efficiency values change significantly before and after removal, which should be retained in the benchmark model. However, AF reflects a single-grain yield, and the efficiency value did not change significantly after removal; thus, it should be removed. Further removing the POY, the efficiency mean changed by 0.0018, but the KS and T-test results are the same. The KS and T-test considered that the efficiency means the change significantly before and after the variable removal, rejected the null hypothesis at the 10% significant level and retained in the model. The POY reflects the agricultural production on the rural economy, and the influence degree of farmers’ life makes the model setting more reasonable. On the analysis of the agricultural input index, the (ACUL) is first retained. The test results are similar to the analysis results of rural residents’ per capita disposable income, and the mean value of efficiency changes significantly, so this index should be included in the model. Following the sequent removal of ACUL and (IRR), a significant change in the mean of efficiency could be retained. If ACUL, IRR and LAB, and both KS tests were removed, T-test rejected the null hypothesis at the significance level of 1%, which indicated that the number of LABs had a significant impact on the agricultural production efficiency of each city. Subsequently, when ACUL, IRR, and MACH were removed, the efficiency mean value changed significantly, and the effect of fertilizer use on the efficiency value was more obvious. The null hypothesis was rejected under both tests. Finally, when ACUL, IRR, and ELEC were removed, the null hypothesis was rejected by both the KS test and T-test. Thus, the ELEC index should also be retained in the benchmark model. The results of the variable selection are shown in

Table 5. Results of the KS test and T-test.

Removed indicator	Change of Mean Value	KS test	T-test	Conclusion
VRS	0.003	0.454	0.665	Remove Indicator
AGDP	0.021	1.732 **	3.856 ***	Reserve Indicator
AF	0.017	0.360	0.481	Remove Indicator
POY	0.008	1.617 **	3.006 ***	Reserve Indicator
ACUL	0.013	0.557	0.519	Remove Indicator
ACUL + AFCUL	0.026	0.618	1.562 *	Remove Indicator
ACUL + IRR	0.089	1.316 **	2.507 ***	Reserve Indicator
ACUL + IRR + LAB	0.058	2.162 ***	3.511 ***	Reserve Indicator
ACUL + IRR + MACH	0.003	1.925 ***	3.374 ***	Reserve Indicator
ACUL + IRR + FER	0.075	0.673	1.021 *	Remove Indicator
ACUL + IRR + ELEC	0.039	1.774 ***	3.382 ***	Reserve Indicator

Note: ***, **, and * show significant levels of 1%, 5% and 10%, respectively. NB: VRS is variable returns, to sale, AGDP is gross agricultural production, AF is grain output, POY is the annual disposable income of rural residents, ACUL is the total sown area, AFCUL is grain sown area, IRR is agricultural effective irrigation area, LAB is the number of total agricultural employees, MACH is the total power of agricultural machinery, and ELEC is and rural electricity consumption.

.

Finally, six input variables were selected: LAB (ten thousand people), IRR (one thousand hectares), ACUL (thousand hectares), MACH (ten thousand kilowatts), application amount of agricultural fertilizer FER (ten thousand tons), and rural electricity consumption ELEC (hundred million kilowatt-hours). Two output variable indicators were selected: AGDP (million yuan) and POY (yuan).

4.4. Results of the DEA-Malmquist, i.e., Stage I

This paper uses the 21-year time series data from 2000 to 2021 and the Deap 2.1 software to calculate the DEA-Malmquist index model on the panel data of agricultural input and output indicators of 13 cities in Jiangsu Province.

Firstly, it is obvious that the agricultural total sown area, agricultural total personnel number, fertilizer usage input elements of Jiangsu province, and the influence of agricultural production efficiency value can be found by selecting the model variables. DEA test results show that the agricultural total sown area, agricultural total personnel number, fertilizer usage input elements of Jiangsu province, and the influence of agricultural production efficiency value in the input-output variable have an important impact on the improvement of total factor productivity. This shows that in Jiangsu, agricultural production, land, labor, and production input still occupy a more important position. Secondly, technological progress in agricultural production is the main reason for the change in TFP. However, the contribution of technology efficiency is not too obvious. Besides, the difference in technology progress and efficiency in agricultural production in different regions of Jiangsu Province is large.

4.4.1. Analysis of the Empirical Results of the Agricultural Malmquist Index (ATFP) in Jiangsu Province

From 2000 to 2021, the annual average Malmquist index and its decomposed technology efficiency change index and progress change index of all cities in Jiangsu Province, and the technology efficiency change index, including the APTE and the scale efficiency index (ASE) calculation results are shown in

Table 6. Malmquist Index and Composition of Agriculture in Jiangsu Province from 2000 to 2021.

Years	Agricultural Technology Efficiency Change Index (AEC)	Agricultural Technology Progress Indexb (ATC)	Agricultural Pure Technology EFFICIENCY Change Index (APTE)	Agricultural Scale Efficiency Change Index (ASE)	Malmquist (ATFP)
2000–2001	1.008	0.991	0.998	1.01	0.998
2001–2002	1.013	1.051	1.01	1.003	1.065
2002–2003	0.984	1.008	0.984	1	0.992
2003–2004	0.995	1.111	1.003	0.992	1.106
2004–2005	0.989	1.059	0.997	0.992	1.048
2005–2006	1	1.059	0.989	1.011	1.058
2006–2007	0.997	1.091	0.995	1.002	1.088
2007–2008	1.005	1.127	1.004	1.001	1.133
2008–2009	1.005	1.106	1.003	1.001	1.111
2009–2010	1.009	1.087	1.009	1.001	1.098
2010–2011	0.991	1.161	0.999	0.992	1.15
2011–2012	0.995	1.09	0.997	0.998	1.085
2012–2013	1.006	1.086	1.005	1.001	1.093
2013–2014	0.99	1.051	0.993	0.997	1.042
2014–2015	0.985	1.082	0.992	0.993	1.065
2015–2016	0.982	1.077	0.996	0.987	1.058
2017–2018	1.002	1.021	0.997	0.997	1.071
2018–2019	1.005	1.055	1	1.002	1.068
2019–2020	0.994	1.093	1.002	0.999	1.053
2020–2021	0.989	1.13	0.996	1.001	1.085
Mean Value	0.997	1.077	0.998	0.999	1.074

.

4.4.2. Horizontal Analysis of the Change in Agricultural Total Factor Productivity in Jiangsu Province

During the 21 years from 2000 to 2021, the ATFP of Jiangsu Province increased in volatility, the higher level of growth, and the average annual growth rate was 7.4% (see Table 4). However, the average growth rate of agricultural technology efficiency is −0.2%, among which the average growth rate of the change index of pure technology efficiency is −0.2% and the average growth rate of the change index of scale efficiency is −0.01%, both of which show negative growth. From 2000 to 2021, the main driving force of the overall growth trend of agricultural total factor productivity in Jiangsu province is the progress of agricultural technology. Because agricultural pure technology efficiency and scale efficiency have negative growth, the growth of agricultural technology efficiency lags behind the progress of agricultural technology, which restricts the improvement of agricultural production efficiency. The empirical results are consistent with the production function analysis by Ji et al. [53]. The main reason is that the management system is not sound, and there are institutional constraints. Therefore, in the process of agricultural development in Jiangsu Province, we should pay attention to the research and development of agricultural science and technology and technology innovation and the improvement of scale operation and management efficiency.

4.4.3. Longitudinal Analysis of Agricultural Total Factor Productivity Change in Jiangsu Province

Concerning the longitudinal analysis of the empirical results, the productivity index of other years was greater than 1 except for 2000–2001 and 2002–2003, when the index was less than 1. From 2003 to 2008, ATFP, ATC, and AEC in Jiangsu showed an obvious rising trend. After 2004, China implemented development policies to support agriculture and increase agricultural investment. Such as reducing agricultural tax, direct subsidies to farmers, financial policies to support agriculture, and financial benefits to agriculture. As such, it adjusted and optimized agricultural structure and accelerated the promotion of improved varieties and agricultural mechanization. Affected by the international financial crisis that erupted in 2008, the total agricultural factor productivity index declined for two consecutive years [54].

From 2010 to 2011, ATFP grew rapidly again, with a growth rate of 15.0%, reaching a peak in the past 21 years. The main reason is that in 2010, China again promulgated a series of sustaining policies for the farmers and agriculture, such as increasing subsidies for improved varieties, raising the ceiling price of household appliances to rural areas, and expanding the field of financial support for agriculture. To a large extent, the burden of farmers has been reduced, the impetus for agricultural development has been greatly enhanced, and the ATFP has been promoted. From 2011 to now, the ATFP of Jiangsu Province has been in a steady growth trend [53].

From 2020–2021, agriculture’s total factor productivity growth in Jiangsu province reached 8.5% and grew faster. The main reason is that Jiangsu province has strengthened the construction of the income guarantee for grain farmers and the interest compensation system for major grain-producing areas. We formulated a list of policies to support grain production and improved policies for subsidizing the combined cultivation of rapeseed, soybean, and corn. We implemented policies to subsidize rice, arable land, subsidies for the purchase and application of agricultural machinery, and minimum grain purchase prices. In 2020, the green prevention and control coverage rate of crop diseases and pests in Jiangsu Province was 47%. Emphasis was placed on promoting efficient fertilization technologies such as water and fertilizer integration. The amount of fertilizer and pesticide used in the province decreased by 10.6% and 13.7%, respectively, compared with 2015. The total power of agricultural machinery in Jiangsu reached 51.94 million kilowatts, an increase of 7.6% over 2015, and the total level of agricultural mechanization reached 88%. The change index of technology progress greatly increased, reducing agricultural production costs, realizing scale operation, and improving marginal agricultural output.

In 2022, the contribution rate of agricultural technology progress in Jiangsu Province reached 71.8%, an increase of 6% over 2016, with an average annual rate of 7.7%. This is because, firstly, Jiangsu Province has vigorously promoted the application of new technologies and planting modes, such as the leaf age model, population quality, precise quantitative cultivation, and other new technologies, and is in a leading position in domestic agricultural popularization and application. Many high-yield and high-quality varieties have been cultivated, including 19 special varieties, with the yield ranking first in China. Secondly, 11 new varieties have been successfully cultivated in China, such as “Sujiang pig” and “Suqin green shell eggs”, ranking second in China. Thirdly, Jiangsu is a big network province, and its development speed and scale of “Internet Plus” are in the leading position in China. It has strengthened the construction of information technology, Internet of Things technology, and service platform and realized the traceability of agricultural product quality and the application of digital technology [55].

4.4.4. AEC Is Negative Growth

The average rate of the agricultural technology efficiency index was −0.3%, which shows that its input shows a gradual decline in marginal benefits. It shows that in the new century, the agricultural scientific research system in Jiangsu province is relatively successful in innovating agricultural scientific research technology. However, it is still insufficient to adapt to the application of agricultural frontier technology and new technology.

4.5. Analysis of ATFP in Each City of Jiangsu Province

4.5.1. The Changes in ATFP

Table 5 and Figure 3 show that the ATFP index of 13 cities in Jiangsu Province has been greater than 1 in the past 21 years, indicating that the agricultural productivity of all cities in Jiangsu Province has been increasing continuously. However, there are still differences in the growth of the ATFP index. Among them, Wuxi has the fastest growth rate of 13.5%, followed by Suzhou, Changzhou, Nanjing, Zhenjiang, and Xuzhou, with 11.9%, 11.1%, 11.0%, 10.0%, and 8.1%, respectively. The slowest growth rate was Huai’an, with a growth rate of only 0.1%, followed by Suqian, Yancheng, and Nantong, with growth rates of 3.4%, 4.2%, and 4.6%, respectively. At the same time, according to the ranking of per capita GDP of all cities in Jiangsu Province in 2021 (from high to low) includes Wuxi, Suzhou, Nanjing, Changzhou, Zhenjiang, Yangzhou, Nantong, Taizhou, Yancheng, Huai’an, Xuzhou, Lianyungang, and Suqian. The ranking of the per capita GDP of Jiangsu Province is similar to that of the ATFP index, indicating that per capita GDP has a certain role in promoting the growth of ATFP.

4.5.2. Changes in the Change Index of Agricultural Technology Progress

From 2000–2021, ATC’s changes in 13 cities of the province are greater than 1, of which Wuxi has the highest ATC value, with an average annual rate of 13.5%. Suzhou, Changzhou, and Nanjing also have an average growth rate of more than 10.0%. Except for Yancheng, the average yearly rate of the other nine cities also reached more than 5.0%. This shows that all cities in Jiangsu Province have actively focused on the R&D and promotion of new agricultural technologies in the past 21 years, and agricultural technology progress has been enhanced rapidly. Similarly, Figure 2 shows that the main contribution of the growth of the ATFP index still comes from agricultural technology progress. However, all the indicators of the top 6, Wuxi, Suzhou, Changzhou, Nanjing, Zhenjiang, and Xuzhou, show positive growth. It can be seen that, in addition to the variable index of technology progress, the variable index of technology efficiency plays a key role in agricultural productivity [56].

According to the change analysis of agricultural technology efficiency (AEC), the 13 cities in Jiangsu Province are divided into three categories. Firstly, cities with AEC > 1 include Xuzhou, Nanjing and Zhenjiang, indicating that the utilization efficiency of agricultural production technology in these three cities has increased rapidly in the past 21 years, and the change index of technology efficiency in Xuzhou has the highest growth rate of 1.2%. Secondly, cities with AEC = 1 include Wuxi, Changzhou, Suzhou, and Huai’an, indicating that the efficiency of agricultural technology utilization in these four cities has remained stagnant in the past 21 years and needs to be improved. Thirdly, cities with AEC < 1 include Lianyungang, Yancheng, Yangzhou, Taizhou, Nantong, and Suqian, among which Suqian has the lowest, with an average growth of −2.2%.

Table 7. Agricultural Malmquist Index and its Composition in Jiangsu Province (2000 to 2021).

Cities	AEC	ATC	APTE	ASE	ATFP	Rank
Nanjing	1.002	1.107	1	1.002	1.11	4
Wuxi	1	1.135	1	1	1.135	1
Xuzhou	1.012	1.068	1.009	1.003	1.081	6
Changzhou	1	1.111	1	1	1.111	3
Suzhou	1	1.119	1	1	1.119	2
Nantong	0.991	1.056	1	0.991	1.046	10
Lianyungang	0.999	1.06	1	0.999	1.059	8
Huai’an	1	1.001	1	1	1.001	13
Yancheng	0.994	1.047	1	0.994	1.042	11
Yangzhou	0.993	1.076	0.995	0.997	1.068	7
Zhenjiang	1.002	1.097	1	1.002	1.1	5
Taizhou	0.992	1.067	0.994	0.998	1.059	8
Suqian	0.978	1.058	0.982	0.996	1.034	12
Mean Value	0.997	1.077	0.998	0.999	1.073

and Figure 3 shows that the utilization efficiency of agricultural production technology in these six cities has not improved or decreased in the past 21 years, and we should attach great importance to the utilization efficiency of agricultural production technology [7].

4.6. Change Analysis of Regional ATFP in Jiangsu Province

According to the geographical distribution of the Huai River and the Yangtze River from north to south, Jiangsu Province is divided into three regions—the southern Jiangsu, the central Jiangsu, and the northern Jiangsu for empirical research. Southern Jiangsu refers to Wuxi, Zhenjiang, Suzhou, Nanjing, and Changzhou. Central Jiangsu refers to Taizhou, Nantong, and Yangzhou. Northern Jiangsu refers to Huai’an, Xuzhou, Lianyungang, Suqian, and Yancheng.

The average ATFP of southern, central, and northern Jiangsu showed obvious gradient differences, decreasing successively during the 21 years from 2000 to 2021. Among them, the average growth rate of ATFP in southern Jiangsu is 11.5%, and the average rate of the TC index is 11.4%, which is ahead of that in central and northern Jiangsu. (See

Table 8. Malmquist index and its composition of regional agriculture in Jiangsu from 2000 to 2021.

Regions	AEC	ATC	APTE	ASE	ATFP
Southern Jiangsu	1.001	1.114	1	1.001	1.115
Central Jiangsu	0.992	1.066	0.996	0.995	1.057
Northern Jiangsu	0.997	1.047	0.9986	0.998	1.043

and Figure 4). Although the per capita cultivated area in central and northern Jiangsu is relatively large, and the fishery is relatively developed, southern Jiangsu has a higher level of modernization, large investment in the promotion of new agricultural technologies, and a high degree of land intensification, so the agricultural production efficiency is far higher than that in central and northern Jiangsu. Therefore, the regional distribution of agricultural TFP growth in Jiangsu province and the economic development of each region are highly related. Thus, ATFP growth is higher in regions with good economic development, while ATFP growth is lower in regions with backward economic development (see

Table 9. Estimation results of the influence model of Agricultural Technology Progress and Efficiency Change on Agricultural Economic Growth in Jiangsu Province.

Variable	Southern Jiangsu		Central Jiangsu		Northern Jiangsu
	Fixed Effect	Stochastic Effect	Fixed Effect	Stochastic Effect	Fixed Effect	Stochastic Effect
lnLAB	−1.2902 ***	−1.8158 ***	−0.2185 **	−0.4458 ***	−0.6141 **	−1.0135 ***
	(0.1492)	(0.2158)	(0.0830)	(0.1251)	(0.1933)	(0.1429)
LnMACH	0.3229 **	1.1360 ***	2.0362 ***	1.8062 ***	0.6477 ***	0.6607 ***
	(0.1266)	(0.2295)	(0.2080)	(0.3435)	(0.1298)	(0.1165)
LnFER	−0.9215 ***	0.1692	−0.3961 ***	−0.2765	−0.0231	0.4869 **
	(0.1425)	(0.2360)	(0.1228)	(0.2096)	(0.3003)	(0.1966)
lnACUL	2.4983 ***	0.7888 ***	2.0502 ***	0.4982	3.2209 ***	0.9273 ***
	(0.2656)	(0.2473)	(0.3280)	(0.4033)	(0.5833)	(0.1073)
EC	0.2063 *	0.8545 *	0.3923 **	0.8248 *	0.3689 *	0.4169
	(0.2845)	(0.5996)	(0.3340)	(0.5732)	(0.2293)	((0.3049)
TE	−0.3458	−0.5830	0.2687 *	0.2842 *	0.2434 **	−0.2058 ***
	(0.2488)	(0.5419)	(0.2377)	(0.4107)	(0.2980)	(0.4017)
_cons	−3.7998 ***	0.6213	−17.3033 ***	−5.9361 ***	−16.8266 ***	−1.5479
	(1.1892)	(1.1806)	(2.0260)	(1.0710)	(4.3504)	(1.0585)
Hausman
Prob > chi2	0.0000	0.1101	0.0000
Prob > chi2	0.0000	0.0000	0.0000	0.0000	0.0000	0.0000

Note: *, **, *** Represents significant at 10%, 5%, 1% levels, respectively.

).

In addition, the average growth rate of the EC index is 0.1%, the only region in the province where all the indexes are positive. The average rate of ATFP and technology progress change index in central Jiangsu was 5.7% and 6.6%, respectively, while the AEC index was negative, with an average rate of 0.8%. The average growth rates of ATFP and ATC in north Jiangsu were 4.3% and 4.7%, respectively, and the AEC index also showed negative growth, with an average growth rate of −0.3%. In general, although the agricultural AEC in the three regions is different, the gap is not large, and they all tend to be close to 1, indicating that the main reason for the large difference in ATFP in each region is the difference in the level of agricultural technology progress. The gradient of development in southern, central, and northern Jiangsu is quite different. The advantages of resource endowment in the agricultural and industrial layout have not been brought into play. It cannot adapt to the new stage of the division of labor development of capital and technology elements and has not formed a distinctive industrial system [57]. These findings correspond to the studies of [29,33], who found similar results in their respective fields of study.

4.7. The Stage II: SFA Regression Analysis

In the second stage, using the great likelihood method, this paper uses Coelli’s FRONTIER Version 4.1 software to conduct a regression analysis of the input redundancies obtained in the first stage. The redundancy of LAB, ACUL, MACH, FER, and ELEC from 2000 to 2021 was used as the dependent variable, and the four external influences of urbanisation level, financial support to agriculture, import/export trade, and accessibility were used as the independent variables. The specific parameter estimates are shown in

Table 10. Results of the SFA regression.

	Labour Input Slack Variable	Land Input Slack Variable	Agricultural Machinery Power Input Slack Variable	Fertilizer Input Slack Variable	Power Consumption Input Slack Variable
constant term	128.215 ***	938.798 ***	736.431 ***	60.888 ***	543.348 ***
	(125.906)	(3.594)	(4.376)	(61.583)	(6.348)
urbanization level	−121.139 ***	−650.607 **	−579.015 ***	−61.388 ***	−398.195 ***
	(−119.382)	(−1.680)	(−2.729)	(−62.968)	(−9.768)
Financial support for agriculture	0.00004 ***	−0.0003 **	0.00004	0.00003 ***	0.00004 **
	(2.740)	(−1.617)	(−0.326)	(14.534)	(−0.394)
Import and export trade	−9.855 ***	29.705	−28.436	−7.537 ***	−12.339 *
	(−9.630)	(0.318)	(−0.423)	(−5.981)	(−2.031)
Convenient transportation	−16.130 ***	−161.731 **	−135.878	−15.778 ***	−35.490 ***
	(−14.718)	(−1.985)	(−1.498)	(−18.246)	(−9.226)
${\sigma }_{vi}^2$	640.568 ***	11,556.479 ***	13,318.702 ***	121.315 ***	3555.4907 ***
	(641.034)	(223.359)	(12,689.504)	(121.548)	(2456.893)
γ	1.000 ***	0.580 *	0.973 ***	1.000 ***	0.732 *
	(49,307.593)	(1.749)	(10.893)	(157,684.132)	(4988.599)
Log-likelihood	−53.798	−76.186	−72.533	−43.388	−61.047
LR test of the one-sided error	3.679	0.146	1.628	7.325	4.582

Note: Numbers in parentheses are the corresponding estimated t-statistic. *, **, *** Represents significant at the 10%, 5%, and 1% levels, respectively.

.

4.8. Stage III: DEA Empirical Results When Adjusted for Input Variables

After excluding the influence of environmental variables and random factors, the following conclusions can be drawn from

Table 11. Comparison of the adjusted cities’ AEC, APTE, and ASE and the values before the adjustment from 2000–2021.

Cities	Agricultural Technology Progress			Agricultural Pure Technology Efficiency			Agricultural Scale Efficiency
	AEC1	AEC3	Direction	APTE1	APTE3	Direction	ASE1	ASE3	Direction
Nanjing	1.002	0.965	↓	1.000	0.989	↓	1.002	0.975	↓
Wuxi	1.000	0.965	↓	1.000	1.000	—	1.000	0.965	↓
Xuzhou	1.012	1.018	↑	1.009	1.009	—	1.003	1.009	↑
Changzhou	1.000	0.994	↓	1.000	1.001	↑	1.000	0.893	↓
Suzhou	1.000	0.976	↓	1.000	1.000	—	1.000	0.976	↓
Nantong	0.991	0.998	↑	1.000	1.000	—	0.991	0.998	↑
Lianyungang	0.999	1.009	↑	1.000	1.006	↑	0.999	1.003	↑
Huaian	1.000	1.010	↑	1.000	1.008	↑	1.000	1.002	↑
Yancheng	0.994	1.003	↑	1.000	1.000	—	0.994	1.003	↑
Yangzhou	0.993	0.995	↑	0.995	0.996	↑	0.997	0.999	↑
Zhenjiang	1.002	1.001	↓	1.000	1.000	—	1.002	1.001	↓
Taizhou	0.992	0.999	↑	0.994	0.999	↑	0.998	1.000	↑
Suqian	0.978	0.999	↑	0.982	1.000	↑	0.996	0.999	↑
Mean value	0.997	0.999	↑	0.998	0.9998	↑	0.999	0.987	↓

.

The empirical results show that the comprehensive technology efficiency value of Wuxi City and Suzhou City has changed from 1 to less than 1, which shows that the agricultural production in these two cities is no longer efficient after excluding the influence of environmental variables and random factors. This is mainly because the scale efficiency value decreases while the pure technology efficiency remains 1. Yancheng’s comprehensive technology efficiency value changed from less than 1 to 1, and the direct reason is that the scale efficiency value increased from 0.994 to 1.003, indicating that the adjusted agricultural production in Yancheng is relatively efficient. The comprehensive technology efficiency value increased in 8 cities and decreased in 5 cities. The TFP’s average value in Jiangsu province increased, which indicates that the comprehensive technology efficiency value of all cities in Jiangsu province is improving. The pure technology efficiency values increased in seven cities (Changzhou, Lianyungang, Huaian, Yangzhou, Taizhou, and Suqian). The other six cities (Wuxi, Xuzhou, Suzhou, Nantong, Yancheng, and Zhenjiang) had unchanged values, and one city (Nanjing) had reduced values. The average comprehensive technology efficiency of the province has increased.

The empirical results also indicate that the improvement of urbanization level, net export trade, and transportation convenience is conducive to improving agricultural production efficiency; financial support for agriculture is weakly conducive to improving agricultural production efficiency. All regions in Jiangsu province should combine the local economic level and resource endowment and improve the management level or adjust the scale of agricultural production to improve agricultural production efficiency.

5. Conclusions and Implications

5.1. Summary of Results

Jiangsu is an economically developed province. In the 21st century, Jiangsu has accelerated the pace of agricultural modernization, transformed the agricultural development mode, significantly improved the supply capacity of grain, optimized the agricultural and industrial structure, and sustained the growth of farmers’ income. All aspects have achieved obvious results and gradually moved towards a technology-intensive agricultural development model of capital, intelligent replacement of labor, and land investment [58]. Based on the findings, it is imperative to maintain the rapid growth of ATFP. The results indicate that from 2000–2021, the average annual rate of ATFP in Jiangsu Province was 7.4%, the largest. From 2018 to 2021, it showed steady growth, with an overall upward volatility trend. The level of comprehensive agricultural productivity in the province is developing well. Agricultural total factor productivity growth plays an important role in economic growth. Also, the improvement of the ATFP synchronized with the progress of agricultural science and technology must be kept. The indicators of agricultural productivity in Jiangsu Province are consistent with agricultural technology progress in both horizontal and vertical directions. This trend is mainly driven by advancing agricultural science research rather than improving agricultural technology efficiency [57]. In addition, narrowing the gap ATFP growth gap among regions. Table 4 shows that from 2000 to 2021, the average annual growth rate of agricultural total factor productivity in 13 cities in Jiangsu Province exceeded 1 and continued to show a positive growth trend, but there was an obvious gap between different regions. With the improvement of ATFP, the average growth rate of the ATFP index in southern Jiangsu is 11.5%, and the average rate of the ATC index is 11.4% in southern Jiangsu. The degree of agricultural intensification, modern agricultural development, and ATFP are higher than those in central Jiangsu, while the level of agricultural development in central Jiangsu is higher than that in northern Jiangsu. The overall level of development in central and northern Jiangsu is still relatively low, so Jiangsu Province has focused on shortening the regional development gap and giving more support to talents, technology, and policies in northern and central Jiangsu [7].

5.2. Policy Implications

Based on these results, the following suggestions are made to policymakers. These will guide in outlining effective policies to enhance agricultural supply quality in China.

5.2.1. Adhere to the Principle of Supply Matching Consumption and Achieve a High Level and Balanced Upgrading in the Supply of Agricultural Products

Policymakers must construct the industrial layout of agricultural products to adapt to consumption upgrading. Based on ensuring the balance of grain supply and demand and the self-sufficiency of grain rations, the principle of “storing grain in the land and technology” is implemented. Firstly, strengthen the comprehensive production capacity, steadily promote agricultural production, and achieve the provincial grain production target of more than 35 million tons. Maintain 2.4 million hectares of rice planting area while appropriately adjusting the scale of wheat fields. Consolidate and improve maize cultivation and improve the quality of maize cultivation. Determine key agroecological zones. Secondly, consolidate the advantage of characteristic agriculture-led industries [17]. Jiangsu Province has realized the development pattern of four major industries represented by high-quality rice, green vegetables, efficient aquaculture, and ecological animal husbandry. It is necessary to make targeted adjustments to the agricultural and industrial structure according to the new consumption needs. To improve the quality of special wheat, it is necessary to carry out “ japonica replace indica” and “fine refinement” projects. Promote the product structure optimization of livestock, poultry, and aquatic products. Vigorously develop high-quality edible rice and corn, promote the cultivation and promotion of new varieties, establish a batch of high-quality food processing bases, promote the food processing industry to reduce losses, improve production efficiency, and strengthen the post-production commercialization of vegetable basket products and characteristic agricultural products to meet the consumption needs of the people. Integrate the development of the agricultural processing industry with the development of production, leisure tourism, and innovative agriculture in the main production areas. Accelerate the development of the “Internet plus” intelligent development trend of the agricultural product circulation industry and cultivate network marketing and other new business forms. Also, they must cultivate well-known agricultural product brands and improve the market competitiveness of agricultural products. They should make full use of the market advantages of enterprises and the public welfare functions of the government to create several distinctive brands that are famous at home and abroad. Create a high-quality, high-value-added “JiangSu” brand and regional agricultural brand. Secondly, we should establish a green and safe brand image. On this basis, improve the traceability management mechanism of production and the traceability of agricultural products. Meanwhile, strengthen the monitoring and treatment of major animal epidemics, and continue to strengthen production safety and quality control. Continue to increase the construction level of “three products and one standard”, enhance the credibility of agricultural product brands, and enhance product demonstration [59].

5.2.2. Enhance the Efficiency of Supply and Allocation of Production Factors by Relying on Scientific and Technological Acceleration

Firstly, increase the output of arable land. Vigorously develop many seed production, breeding, and sales integration enterprises. Promote the fine allocation of germplasm resources of key food crops and build an industrial cultivation base of high-quality germplasm resources. Secondly, we should vigorously cultivate and develop new agricultural production entities. Give continuous and stable support to family farms and professional cooperatives. Vigorously develop new agricultural scale operation and service subjects represented by leading enterprises and promote various forms of agricultural scale production [17]. Again, popularize the application of new science and technology to enhance the efficiency of agricultural resource utilization. Improve the level of agricultural mechanization, develop the production technology of agricultural machinery and agronomy integration, reduce the input of production factors, and increase the output of agricultural products. Continuously improve the effectiveness of factor allocation. Solve the problem of insufficient quantity and low quality of agricultural labor force and increase the training of skilled rural labor force. Moreover, invest and develop a variety of modern production factors, combine new factors, and promote the improvement of agricultural total factor productivity and supply-side structural reform [58].

5.2.3. Build a Green and Clean Production System for Improving the Safety and Quality of Agricultural Products

Firstly, optimize the rotation planting system. Scientific crop rotation and interplanting should be carried out, and food safety should be taken as the premise to realize the rational distribution of crops and promote the development and technology popularization of new fertilizers. Actively explore the efficient use of organic fertilizers and trace elements. Realize the organic combination of unified defense and green control. Secondly, develop ecological and circular agriculture. According to the bearing capacity of the soil, scientifically plan the development direction of the livestock and poultry breeding industry, promote the coordinated development of the industry, and build a circular agriculture demonstration base [59,60]. Thirdly, improve agricultural products’ quality and safety level to demonstrate, guide and popularize new agricultural production techniques to standardize equipment, techniques, and products. Highlight local characteristics, and build high-quality agricultural product brands with local characteristics, with Jiangsu’s geographical regional characteristics and cultural heritage as the core. Fourthly, accelerate the development of new forms of creative agriculture. Promote the integration of rural production, living and ecological resources, and upgrade the industrial value chain of agriculture [61].

5.2.4. Accelerate the Comprehensive Rural Reform and Implement the Rural Revitalization Strategy

Systems for rural land property rights, basic agricultural support and protection, rural social governance, and establishing sound systems and mechanisms for agricultural modernization must be improved.

Firstly, improve the reform of the separating “three rights” of rural land. Implement the Opinions on Improving the Measures for the Separation of Rural Land Ownership and Contract Rights and Management Rights issued by China, promote the separation of ownership rights, contract rights, and management rights of rural land, and guide the transfer of land management rights to “enter the market for a transaction and standardize operation”. Carry out the pilot project of land transfer risk guarantee fund. Efforts should be made to develop multi-forms of agricultural scale operations such as joint farming and planting, alternative farming and planting, and farmland trusteeship [62,63]. Secondly, accelerate the legislation and reform of the rural collective property rights system. A new rural collective economic development model featuring clear property rights, stable income, reasonable distribution, and democratic management will be promoted throughout the province [64,65]. Thirdly, promote the innovation of rural financial service mechanisms. Promote rural land finance and expand the pilot project of mortgage loans for rural land management rights. Establish a provincial-level agricultural credit guarantee system to improve the efficiency of financing risk compensation funds for new agricultural business entities. We should encourage the development of rural cooperative finance in various fields and at a broad level [66].

5.3. Future Research Direction and Focus

Future research will mainly focus on the following aspects. Further, strengthen the research framework by combining modern economic growth theory with empirical analysis, for example, paying attention to the impact of the digital economy on the quality of agricultural supply [67]. More attention should be paid to the research of ATFP at the county level and conduct an empirical analysis on the impact elasticity of scientific and technology progress, integration of primary, secondary, and tertiary industries, carbon emissions, and other factors to promote high-quality development of the economy as a whole [68,69].