Digital Infrastructure and Agricultural Global Value Chain Participation: Impacts on Export Value-Added

Abstract

[Objective] Digital infrastructure, with its fundamental and public good characteristics, can have a significant impact on export trade. This paper aims to analyze the impact and mechanism of digital infrastructure construction on the added value of agricultural exports by combining theory and empirical analysis. [Methodology] Based on the construction of the theoretical framework and the panel data of 61 economies from 2007 to 2021, the fixed effect model was used to explore the impact of the level of digital infrastructure on the added value of agricultural trade exports and the moderating effect of participation in the global agricultural value chain. [Results] (1) The construction of digital infrastructure is conducive to increasing the added value of agricultural exports. Specifically, a 1% increase in the level of digital infrastructure will promote a 0.159% increase in the added value of agricultural exports. (2) The construction of digital infrastructure affects the added value of agricultural exports through three mechanisms: enhancing labor productivity, optimizing the business environment, and promoting technological innovation. (3) Digital infrastructure has a more significant effect on enhancing the added value of agricultural exports in developed economies and those with higher levels of digital infrastructure. (4) Participation in the global value chain of agriculture has a moderating effect on the impact of digital infrastructure on the added value of agricultural exports.

1. Introduction

The world today is undergoing major changes unseen in a century, and the process of global economic and trade development is fraught with various uncertainties. The superimposition of factors such as trade protectionism, the wave of anti-globalization, extreme weather, and regional conflicts has led the global economy and trade into a period of deep adjustment [1].

Against this backdrop, the international agricultural product market has become increasingly volatile, and the challenges faced by global food security and agricultural product trade are also growing day by day [2]. Agricultural product trade is an important area of international economic and trade cooperation and also a key focus for ensuring food security [3].

As agriculture increasingly participates in the global value chain, international trade of agricultural products has also ushered in new development space [4]. This is conducive to optimizing resource allocation and reducing production costs [5], and also creates new development space for enhancing agricultural product trade. Against the background of the global agricultural value chain, agricultural products usually involve the division of labor and cooperation among multiple countries. At this time, the traditional indicators of total trade value have difficulty accurately reflecting the true level of trade benefits. Therefore, trade added value has become the focus of current research [6,7]. Amid the current increasingly complex international economic and trade situation, how to respond to challenges and increase the added value of agricultural trade has become a common concern of the world today.

With the accelerated evolution of industrial transformation, information technologies represented by 5G, the Internet of Things, big data, and artificial intelligence have permeated all aspects of global economic development [8,9,10,11], and the digital economy has brought new opportunities to agricultural product trade [12,13]. The construction of digital infrastructure determines the development direction of the digital economy [14]. Digital infrastructure features digitalization, integration, and inclusiveness, which is conducive to alleviating information asymmetry, optimizing resource allocation, and reducing trade risks, thereby promoting more efficient, convenient, and secure international trade in agricultural products [15,16,17,18].

Based on this, against the backdrop of an increasingly complex international economic and trade situation, this paper focuses on the impact effect and action path of digital infrastructure construction on the added value of agricultural trade exports. The innovation and contribution of this article are mainly reflected in the following aspects: First, it examines the impact effect of digital infrastructure on the benefits of agricultural trade from the perspective of export value added. Second, it explores the channels through which digital infrastructure affects the added value of agricultural exports from three aspects: labor productivity, the business environment, and technological innovation. Third, it conducts heterogeneity analysis based on the economic development level and digital infrastructure level of the economy. Fourth, it explores whether the impact of digital infrastructure on the added value of agricultural exports will be cross-influenced by participation in the global agricultural value chain.

The structure of this article is as follows. Firstly, the article analyzes the research background and existing literature. On this basis, it proposes a theoretical framework and research hypotheses, and explains the research area, research data, and research model. Secondly, the article empirically examines the relationship between digital infrastructure and the added value of agricultural exports, and analyzes the moderating effect of participation in the global value chain of agriculture. Finally, the article is summarized and analyzed by combining theory with practice.

2. Literature Review and Theoretical Framework

2.1. Literature Review

2.1.1. Agricultural Trade Value-Added

Against the background of economic globalization, traditional methods for statistics of total trade volume have difficultly accurately reflecting the true level of trade benefits. Therefore, scholars have turned to “trade value-added” for statistics, which mainly focuses on the domestic value added contained in export products [19,20]. Currently, the most widely used value-added accounting methods include the WWZ method and the nine-point method [21,22], both of which are based on input-output tables to disaggregate the total trade in order to obtain the real trade gains of individual countries (regions). With the continuous evolution of the accounting framework of agricultural trade value-added, scholars have also conducted a preliminary investigation into its influencing factors, and the research shows that factors such as land resources, foreign investment, institutional quality, total factor productivity, trade policies, and barriers may have a certain impact on the added value of agricultural trade [23,24,25].

2.1.2. Digital Infrastructure

Digital infrastructure has not yet formed a unified definition, and the early literature mainly focuses on the research of “information infrastructure” in a narrow sense [26], such as satellite, microwave, etc. In recent years, it has gradually transformed into “digital infrastructure” in a relatively broad sense [27], including information infrastructure and converged infrastructure formed by the digital transformation of traditional infrastructure. The measurement aspect of digital infrastructure mainly includes the single-indicator method and the comprehensive indicator method. The former refers to the selection of a single indicator from Internet development and broadband infrastructure as a proxy indicator for digital infrastructure, such as Internet penetration, broadband construction, and mobile base stations [28,29]. The latter usually selects multiple sub-indicators related to the degree of Internet construction and use for comprehensive calculation, mainly including indicators such as the Internet penetration rate, broadband subscriptions, telephone subscriptions, and international broadband rate [30]. In terms of the impact of digital infrastructure on trade value-added, most scholars believe that the construction of digital infrastructure has a positive influence on trade value-added. The channels of influence include reducing trade costs, promoting technological innovation, improving resource utilization efficiency, and expanding market boundaries [31].

2.1.3. Agricultural GVC Participation

Agricultural global value chains (GVCs) refer to the division of labor in agricultural activities and product processing by multiple countries, including value addition and profit distribution throughout the entire process of production, marketing, and recycling. Measures of the division of labor status of agricultural GVCs mainly include indicators such as the vertical specialization index, upstream degree index, downstream degree index, production length index, and GVC location index [32,33,34]. Some scholars believe that participating in the global value chain division of labor can promote the increase of trade value-added. The influencing channels include industrial correlation, market reconstruction, technology spillover, forced improvement of human capital levels, promotion of infrastructure construction, enhancement of productivity, and reduction of trade costs [35].

2.1.4. Research Gap

Through this literature review, we found that although there are many studies discussing digital infrastructure and trade value-added accordingly, there are several shortcomings: First, most of the existing studies focus on manufacturing and service industries, and studies on the agricultural sector are yet to be supplemented. Second, the existing literature is unlikely to consider whether GVCs play a moderating role in the impact of digital infrastructure on agricultural trade value-added.

2.2. Theoretical Framework and Research Hypotheses

2.2.1. The Direct Impact of Digital Infrastructure on the Added Value of Agricultural Exports

Agricultural product export trade involves cross-border transactions, and there is an increase in costs at every stage, from providing cross-border products and matching customer demands to cross-border logistics transportation and the settlement of payment for goods. The construction of digital infrastructure is conducive to the intelligent management of agricultural product trade, reducing trade costs [36] and thereby enhancing the profits of agricultural trade. First, digital technology and modern agricultural biotechnology can enhance the quality and efficiency of planting, upgrade production methods, and reduce production costs [37], thereby increasing the added value of agricultural exports. Second, the development of digital infrastructure is conducive to breaking information barriers and promoting the full flow of information, thus improving the efficiency of information search, effectively reducing search costs [38], and helping to enhance the added value of agricultural exports. Third, the development of digital infrastructure can reduce communication costs for both trading parties. Trade entities can utilize the instant communication services and translation functions of Internet platforms to reduce communication costs and enhance communication efficiency [39], thereby promoting an increase in the added value of agricultural exports. Fourth, agricultural products are perishable and difficult to store, which leads to high transportation and storage costs during the trade process. The construction of digital infrastructure helps optimize the logistics system and effectively reduce transportation costs [40], thereby facilitating the increase in the added value of agricultural exports.

Therefore, referring to Figure 1, this paper proposes the following research hypothesis (H₁).

H₁. 

The construction of digital infrastructure helps to increase the added value of agricultural exports.

2.2.2. The Indirect Impact of Digital Infrastructure on the Added Value of Agricultural Exports

(1) Labor Productivity. The construction of digital infrastructure can significantly improve labor productivity and further promote the added value of agricultural exports. Firstly, digital infrastructure has the characteristic of inclusiveness, which helps to bridge the regional digital divide, improve the abundance of regional factors [41], optimize the allocation of production factors, and thereby enhance labor productivity. Secondly, the development of digital infrastructure can enhance the efficiency of labor information search and improve the matching level of the labor market [42], thereby driving an increase in labor productivity. Finally, with the construction and improvement of digital infrastructure, some labor forces with highly repetitive jobs are more likely to be replaced. This drives low-skilled workers to enhance their own skills, thereby contributing to the overall improvement of labor skills and further increasing labor productivity [43].

Labor productivity improvement is key to the growth of agricultural export value-added. On one hand, the construction of digital infrastructure provides support for the development of agricultural modernization, and with the help of technologies such as the Internet of Things, labor productivity can be improved [44], which, in turn, helps to promote an increase in the number of agricultural exports. On the other hand, as labor productivity increases, enterprises become more capable and inclined to produce high-quality products, thereby enhancing their competitive edge in the market. At the same time, they also have the motivation to seek higher value-added embedded positions in the global value chain [45] and, thus, improve their profitability.

Therefore, referring to Figure 1, this paper proposes the following research hypothesis (H₂).

H₂. 

The construction of digital infrastructure is conducive to enhancing labor productivity, thereby promoting the added value of agricultural exports.

(2) Business Environment. Digital infrastructure is universal and helps enhance market transparency and government regulatory efficiency, thereby optimizing the business environment. First, the construction of digital infrastructure has promoted the comprehensive opening and sharing of data, which is conducive to maintaining a fair market environment. For instance, big data technology can solve the problem of information asymmetry existing between traditional banks and enterprises, which is conducive to reducing financing costs and thereby alleviating the financing constraints and credit discrimination faced by small and micro enterprises [46]. Second, the construction of digital infrastructure is conducive to enhancing the efficiency of government services and supervision. The application of digital technology is conducive to achieving the precise supply of government public services and the scientific formulation of decision-making plans [47]. Regulatory authorities can also utilize artificial intelligence regulatory systems to achieve precise supervision [48], thereby optimizing the business environment.

The optimization of the business environment is an important guarantee for the growth of agricultural trade value-added. First, the optimization of the financial service environment is conducive to attracting capital inflows and reducing financing constraints, providing financial guarantees for agricultural enterprises [49] and thereby promoting the increase in the added value of agricultural exports. Second, a favorable business environment represents an effective market mechanism and a fair competitive environment, which is conducive to curbing enterprises’ rent-seeking motives [50] and creating a favorable external environment for the increase in the added value of agricultural exports. Third, the optimization of the government service environment is conducive to building a good relationship between the government and business. The improvement of the regulatory environment can better maintain market order [51], reduce the risks brought by policy uncertainties, and thereby increase the added value of agricultural exports.

Therefore, referring to Figure 1, this paper proposes the following research hypothesis (H₃).

H₃. 

The construction of digital infrastructure can enhance the added value of agricultural exports by optimizing the business environment mechanism.

(3) Technological Innovation. First, the construction of digital infrastructure can alleviate the problem of information asymmetry, reduce the cost of information search, promote technology spillover, and enhance innovation capabilities [52]. Second, the construction of digital infrastructure will give rise to new financing methods, which is conducive to providing financial support for enterprises and improving the efficiency of capital allocation [53]. Third, innovative talents are the key to innovation activities. The development of digital infrastructure provides favorable conditions for the aggregation of high-quality human capital, and at the same time, offers more convenient new models and platforms for talent exchange and learning and training, thereby promoting technological innovation [54].

Technological innovation is the internal driving force for increasing the added value of agricultural exports. On one hand, the wide application of Internet technology, Internet of Things technology, and big data statistics technology is conducive to promoting the digital transformation and upgrade of agriculture [55], achieving digitalization and intelligence in all links of agricultural product production, sales, storage, and transportation, and reducing the cost of transaction links. On the other hand, the promotion and application of blockchain technology, biotechnology, and mechanical equipment technology in the planting, processing, and storage of agricultural products can help improve the quality and safety standards of agricultural products [56] and increase their market competitiveness and added value.

Therefore, referring to Figure 1, this paper proposes the following research hypothesis (H₄).

H₄. 

Digital infrastructure development can positively affect the added value of agricultural exports through technological innovation effects.

2.2.3. The Moderating Role of Participation in Agricultural Global Value Chains

The concept of GVCs describes that different countries specialize in certain phases of a same production process, forming international production networks [57]. The embedding of the global value chain is divided into forward industrial correlation and backward industrial correlation. Forward embedding indicates that the industrial sector mainly participates in trade activities by providing intermediate goods, while backward embedding indicates that the industrial sector mainly participates in trade activities by processing imported intermediate goods. This paper selects the forward participation and backward participation in the global agricultural value chain as moderating variables.

(1) Forward Participation in Agricultural GVCs. Before agriculture moves forward to participate in the division of labor along the value chain, digital infrastructure can maximize trade benefits through channels such as enhancing labor productivity, optimizing the business environment, and promoting technological innovation. As forward participation in the global agricultural value chain continues to increase, it may diminish the influence of digital infrastructure on the added value of agricultural exports. First, the increase in forward participation in the global agricultural value chain may heighten competitive pressure. Enterprises may pay more attention to short-term profits and allocate more resources to their participation in the global value chain [58], while using fewer resources for the construction and application of digital infrastructure. This substitution relationship in resource allocation may lead to the weakening of the influence of digital infrastructure [59]. Second, the increase in forward participation in the global agricultural value chain may raise the complexity of the production process and coordination costs, thereby weakening the promoting effect of digital infrastructure.

Therefore, referring to Figure 1, this paper proposes the following research hypothesis (H₅).

H₅. 

Forward participation in agricultural GVCs negatively moderates the relationship between digital infrastructure and agricultural export value-added.

(2) Backward Participation in Agricultural GVCs. At present, the level of digital infrastructure construction shows regional differences worldwide. Some economies are confronted with problems such as technological backwardness and insufficient supply of digital talents [60], which have increased the cost of using data elements and digital technologies in the agricultural product trade process. These countries (regions) are generally at the lower end of the value chain. Most of them participate in trade activities by processing imported intermediate goods, that is, backward participation in the global agricultural value chain. In the process of backward participation in the global value chain, enterprises can, on one hand, enhance their technological innovation capabilities by learning and imitating the advanced production technologies of importers. On the other hand, during the process of importing intermediate goods, they may also receive technical training and guidance from exporters, thereby enhancing the level of human capital [61]. Therefore, backward participation in the global value chain can overcome the bottlenecks of technology and human capital through the learning-by-doing effect, thereby enhancing productivity and further promoting increases in the added value of agricultural exports. From this, it can be inferred that the higher the backward participation in the global agricultural value chain, the stronger the influence of digital infrastructure on the added value of agricultural exports.

Therefore, referring to Figure 1, this paper proposes the following research hypothesis (H₆).

H₆. 

Backward participation in agricultural GVCs positively moderates the relationship between digital infrastructure and agricultural export value-added.

3. Methodology and Data Source

3.1. Study Area and Data Source

This article conducts research using data from 61 economies from 2007 to 2021. The research areas include 30 European countries (regions), 25 Asian countries (regions), 3 North American countries (regions), 2 Oceania countries (regions), and 1 South American country (region). The samples are extensive and representative (

Table 1. Study area.

Region	Numbers of Economies	Names of Economies
Europe	30	Austria, Belgium, Bulgaria, Switzerland, Czech Republic, Germany, Denmark, Spain, Estonia, Finland, France, United Kingdom, Greece, Croatia, Hungary, Ireland, Italy, Lithuania, Luxembourg, Latvia, Malta, Netherlands, Norway, Poland, Portugal, Romania, Russia, Slovak Republic, Slovenia, Sweden
Asia	25	China, Cyprus, Indonesia, India, Japan, Republic of Korea, Turkey, Bangladesh, Malaysia, Philippines, Thailand, Vietnam, Kazakhstan, Mongolia, Sri Lanka, Pakistan, Laos, Brunei, Bhutan, Kyrgyzstan, Cambodia, Maldives, Nepal, Singapore, Hong Kong, China
North America	3	Canada, Mexico, United States
Oceania	2	Australia, Fiji
South America	1	Brazil

).

The data are mainly from the Asian Development Bank (https://mrio.adbx.online, accessed on 3 January 2025), UIBE GVC Indicators database (http://rigvc.uibe.edu.cn/english/D_E/database_database/index.htm, accessed on 3 January 2025), World Bank statistics (http://www.worldbank.org, accessed on 3 January 2025), International Telecommunication Union (ITU) (https://www.itu.int, accessed on 3 January 2025), World Trade Organization (WTO) (https://www.wto.org, accessed on 3 January 2025), Heritage Foundation (https://www.heritage-foundation.org, accessed on 3 January 2025), International Labor Organization (ILO) (https://www.ilo.org, accessed on 3 January 2025), and other international databases.

3.2. Variable Selection

3.2.1. Dependent Variable

The dependent variable of this study is the added value of agricultural exports (DVA). This paper refers to existing studies to measure the added value of a country’s (region’s) agricultural sector exports [21]. Specifically, total sectoral exports are divided into 16 small, 8 medium, and 4 large value-added components.

The formula is specified below:

$$\begin{array}{c}E=DVA\_FIN+DVA\_INT+\left(DVA\_INTrexI1+DVA\_INTrexF+DVA\_INTrexI2\right)+\\ \left(RDV\_FIN+RDV\_FIN2+RDV\_INT\right)+\left(MVA\_FIN+OVA\_FIN\right)+ \\ \left(MVA\_INT+OVA\_INT\right)+\left(MDC+ODC\right)+\left(DDC\_FIN+DDC\_INT\right)\\ =\left(DVA\_FIN+DVA\_INT+DVA\_INTrex\right)+RDV+ \\ \left(FVA\_FIN+FVA\_INT\right)+\left(FDC+DDC\right) \\ =DVA+RDV+FVA+PDC \end{array}$$

(1)

In Equation (1), DVA is the domestic value added absorbed abroad in total exports, RDV is the returned domestic value added, FVA is the foreign value added, and PDC is the double-counting component of total exports. The export value-added in this article is DVA. In the empirical analysis, the explained variable is the logarithm of DVA.

3.2.2. Core Explanatory Variable

This study selects digital infrastructure (FRA) as the core explanatory variable. This paper selects six indicators, fixed broadband subscription volume, fixed telephone subscription volume, mobile broadband subscription volume, mobile cellular subscription volume, Internet penetration rate, and international broadband speed, and uses the AHP-entropy method for combined weighting to obtain the digital infrastructure index. Referring to relevant literature [62], in the empirical analysis, this paper logarithmically processes the measured digital infrastructure index to reflect the level of digital infrastructure, specifically: $\ln FRA=\ln (FRA\ast 100)$.

3.2.3. Mediator Variable

According to the analysis part of the theoretical framework, this paper selects three mediating variables: labor productivity (LP), the business environment (DB), and technological innovation (NOV). For labor productivity, output per worker (constant 2015 GDP: USD) is used to measure it; for the business environment, the business environment index is used to measure it; and for the technological innovation effect, papers in scientific and technological journals are used to measure it.

3.2.4. Moderator Variable

This article uses the global value chain participation index to describe the degree of agricultural participation in the global value chain [35].

The forward GVC participation index is defined as:

$$GV{C}_{\mathrm{f}}=I{V}_{ir}/E_{ir}$$

(2)

The backward GVC participation index is defined as:

$$GV{C}_{\mathrm{b}}=F{V}_{ir}/E_{ir}$$

(3)

where i denotes the agriculture; r represents the country; $I{V}_{\mathrm{ir}}$ represents the country’s indirect agricultural value-added exports; $F{V}_{\mathrm{ir}}$ stands for foreign value added; and $E_{\mathrm{ir}}$ is the total export value of agricultural products.

3.2.5. Other Control Variables

In order to reduce the bias of neglected variables, based on the existing studies, the following variables are introduced as control variables in this paper: (1) land resources (LRs), measured using the share of cultivated land area in land area; (2) foreign investment (FDI), measured using net FDI inflows as a share of GDP; (3) regulatory quality (RQ), measured using the regulatory quality index; (4) economic freedom (EFI), measured by the global economic freedom index; and (5) whether to join the World Trade Organization (WTO). The descriptions of the main variables in this article are shown in

Table 2. Description of variables.

Variable Type	Variable	Variable Description	Expected Symbol	Obs	Mean	Std. Dev.	Min	Max
Dependent Variable	lnDVA	The logarithmic value of the added value of agricultural exports	/	915	7.125159	1.910924	0.1018066	11.29243
Core Explanatory Variable	lnFRA	The logarithmic processing value of digital infrastructure	+	915	5.686417	0.9411408	0.8137257	8.937631
Mediator Variable	LP	Output per worker (USD)	+	915	51,270.04	46,858.47	1565.69	257,976.9
	DB	Ease of doing business index	+	915	69.65242	10.51569	40.6806	88.70221
	NOV	Papers in scientific and technological journals	+	900	34,513.44	80,216.33	0.92	729,030.1
Moderator Variable	GVCf	Forward participation in agricultural Global Value Chains index	+	915	0.2163338	0.1620188	0.0087101	0.9193935
	GVCb	Backward participation in agricultural Global Value Chains index	+	915	0.2092498	0.1226278	0.0147087	0.5917655
Other Control Variables	LR	Cultivated land (percentage of land area)	+	915	19.16313	14.43028	0.7544339	62.475
	FDI	Net FDI inflows (percentage of GDP)	+	915	8.146677	27.98512	−117.3747	449.0828
	RQ	Regulatory quality index	+	915	0.6813351	0.8779764	−1.186762	2.252235
	EFI	Global economic freedom index	+	915	66.56918	9.366937	44.2	90.2
	WTO	Whether to join the World Trade Organization (Yes = 1, No = 0)	+	915	0.9628415	0.1892534	0	1

.

3.3. Baseline Model

This paper is based on the fixed effects model and conducts an empirical test using data from 61 economies from 2007 to 2021.

The baseline model is as follows:

$$\ln DV{A}_{it}={\alpha }_0+{\alpha }_1\ln FR{A}_{it}+{\alpha }_2{X}_{it}+{\delta }_i+{\lambda }_t+{\epsilon }_{it}$$

(4)

where subscripts i and t denote country (region) and year, respectively. $\ln DVA$ represents the logarithmic value of the added value of agricultural exports; $\ln FRA$ represents the logarithmic processing value of digital infrastructure; $X$ denotes control variables including land resources, foreign investment, regulatory quality, economic freedom, and the presence or absence of WTO accession; ${\delta }_i$ is an individual fixed effect; ${\lambda }_t$ is the time fixed effect; and ${\epsilon }_{it}$ is the random error term.

4. Results and Analysis

4.1. Analysis of Agricultural Export Value-Added and Digital Infrastructure Level

The changing trends of the added value of agricultural exports in various regions from 2007 to 2021 are shown in Figure 2. During the 15 years from 2007 to 2021, the added value of agricultural exports in various economies generally showed an upward trend. By region, Europe maintained a leading position in the added value of agricultural exports from 2007 to 2021. The other regions ranked from high to low in terms of the added value of agricultural exports were North America, Asia, South America, and Oceania.

The top ten economies in terms of the average level of digital infrastructure from 2007 to 2021 were: China, the United States, India, Japan, Hong Kong, China, Brazil, Russia, Germany, Indonesia, and France (Figure 3). Among them, three economies belong to Europe, five economies belong to Asia, one economy belongs to North America, and one economy belongs to South America. During the 15 years from 2007 to 2021, the digital infrastructure levels of various economies generally showed an upward trend.

4.2. Baseline Model Estimates

The results of columns (1) to (6) in

Table 3. Estimation results of the baseline model.

Variables	(1)	(2)	(3)	(4)	(5)	(6)
$\ln FRA$	0.1298 *** (0.0361) [0.0590, 0.2006]	0.1206 *** (0.0364) [0.0492, 0.1920]	0.1162 *** (0.0364) [0.0448, 0.1876]	0.1136 *** (0.0366) [0.0418, 0.1855]	0.1127 *** (0.0366) [0.0407, 0.1846]	0.1069 *** (0.0370) [0.0342, 0.1796]
LR		0.0262 * (0.0143) [−0.0018, 0.0542]	0.0298 ** (0.0144) [0.0016, 0.0580]	0.0278 * (0.0147) [−0.0011, 0.0566]	0.0279 * (0.0147) [−0.0009, 0.0567]	0.0281 * (0.0147) [−0.0007, 0.0569]
FDI			0.0010 * (0.0005) [0.0000, 0.0020]	0.0010 * (0.0005) [−0.0000, 0.0019]	0.0010 * (0.0005) [−0.0000, 0.0019]	0.0010 * (0.0005) [−0.0000, 0.0019]
RQ				0.0558 (0.0832) [−0.1076, 0.2192]	0.0380 (0.0876) [−0.1339, 0.2098]	0.0340 (0.0876) [−0.1380, 0.2060]
EFI					0.0039 (0.0059) [−0.0076, 0.0154]	0.0031 (0.0059) [−0.0085, 0.0147]
WTO						0.1144 (0.1080) [−0.0977, 0.3265]
Fixed Years	YES	YES	YES	YES	YES	YES
Fixed Individual	YES	YES	YES	YES	YES	YES
N	915	915	915	915	915	915
R2	0.3358	0.3385	0.3415	0.3419	0.3422	0.3431

Note: *, **, and *** denote 10%, 5%, and 1% significance levels, respectively; standard errors are in parentheses; and the upper and lower limits of the 95% confidence interval are shown in square brackets.

show that as control variables are gradually added, the coefficients of digital infrastructure are all significantly positive. This indicates that digital infrastructure can significantly boost the domestic added value of agricultural exports, thus verifying H₁. According to the results in column (6), and combined with the sample mean, a 1% increase in the level of digital infrastructure will promote an increase of USD 5.046 in the added value of agricultural exports, accounting for 0.1069% of its mean. Economically, although this improvement is relatively small, in the context of actual policies, it may be of great significance. For many developing economies, agriculture is an important export sector. Any measure that can increase the added value of agricultural exports may be of great significance to economic development.

The results of the control variables show that the coefficient of land resources is significantly positive, indicating that the improvement of land resources is conducive to the enhancement of the added value of agricultural exports. Agricultural products have a strong dependence on land resources [63], and abundant land resources are conducive to promoting the large-scale operation of agriculture, improving the productivity and unit output of agricultural products and, thus, promoting the enhancement of agricultural export value-added. Foreign investment can provide sufficient financial support for the agricultural sector, which is conducive to promoting the differentiation and diversification of agricultural products [64] as a way to promote the enhancement of the added value of agricultural exports. The quality of regulation has a positive impact on the added value of agricultural exports. The improvement of the quality of government regulation can inhibit speculative behaviors and maintain a fair competitive market environment [65], which, in turn, enhances the added value of agricultural exports. The improvement of the level of economic freedom is conducive to the formation of an open trade environment for agricultural products [66], thus increasing the added value of agricultural exports. The economy’s accession to the WTO is conducive to reducing trade barriers and broadening the international market and, thus, can promote the increase in the added value of agricultural exports.

4.3. Endogeneity and Robustness Tests

4.3.1. Endogeneity Test

The treatment of model endogeneity in this paper focuses on two main aspects.

(1) Although the main factors affecting added value within the exporting country have been considered in the previous section, there may still be important variables left out. Considering that per capita income level, human capital, and the level of the rule of law also have an impact on trade value-added, three control variables are added on the basis of the baseline model here: per capita GDP (thousands of US dollars per person), human capital (total labor force, in logarithmic terms), and the level of the rule of law. The data are sourced from the World Bank. The results are presented in column (1) of

Table 4. Endogeneity and robustness tests.

Variables	(1)	(2)	(3)	(4)
$\ln FRA$	0.1520 *** (0.0406) [0.0723, 0.2316]	0.1218 *** (0.0388) [0.0458, 0.1977]	2.6932 *** 0.2096 [2.2823, 3.1040]	0.1138 *** (0.0387) [0.0378, 0.1898]
Anderson LM		792.887 (0.0000)
Cragg–Donald Wald F		11,000.000 [16.38]
Control Variables	YES	YES	YES	YES
Fixed Years	YES	YES	YES	YES
Fixed Individual	YES	YES	YES	YES
N	915	915	915	915
R2	0.3521	0.3475	0.3744	0.3508

Note: *** denotes 1% significance level; standard errors are in parentheses; and the upper and lower limits of the 95% confidence interval are shown in square brackets. In the relevant tests of instrumental variables, the values in parentheses are the p-values of the corresponding statistics, and the values in square parentheses are the critical values of the Cragg–Donald Wald F-statistic at the 10% significance level.

and show that the regression coefficient for digital infrastructure remains significantly positive.

(2) To enhance the validity of the estimation, referring to relevant studies [67,68,69], the instrumental variable selected in this paper is the lag period of digital infrastructure, and the two-stage least squares method (2SLS) is used to estimate the model. The digital infrastructure level of the previous period will significantly affect the current period’s digital infrastructure level through development accumulation and path dependence, meeting the correlation requirement. The digital infrastructure lagging behind by one period is also mainly affected by the factors of the lag period, meeting the requirements of exogeneity. The use of digital infrastructure lagging by one period can effectively handle the endogeneity problems caused by reverse causality and accurately assess the causal effects. The results are shown in column (3) of Table 4. The sign, size, and significance of the digital infrastructure remain robust.

4.3.2. Robustness Checks

This paper’s test for model robustness focuses on the following two aspects.

(1) Replace the fixed-effect model with the logit model. According to the mean of the explained variable, replace it with a binary variable. Specifically, assign the data greater than the mean to 1, the data less than the mean to 0, and the remaining variables to be consistent with the benchmark regression. The results are shown in column (3) of Table 4, and the regression coefficient of digital infrastructure remains robust.

(2) Winsorization. All continuous variables were truncated by 1% to control the bias caused by extreme values on the estimation results. After sample processing, the statistical distribution of the main variables was relatively stable. The results are shown in column (4) of Table 4, and there is no significant difference from the baseline model.

4.4. Heterogeneity Tests

To analyze the specific performance of digital infrastructure in different economic development levels and economies with different levels of digital infrastructure, this paper conducts the following tests. In terms of heterogeneity in economic development levels, referring to the World Bank’s classification standard for per capita national income levels, high-income economies are classified as developed economies, and middle- and low-income economies are classified as developing economies. In terms of heterogeneity in digital infrastructure levels, 61 economies were classified into high digital infrastructure level economies and low digital infrastructure level economies based on the average digital infrastructure level of each economy from 2007 to 2021 (the top 30 were high digital infrastructure level economies, and the bottom 31 were low-level economies).

Table 5. Heterogeneity tests.

Variables	(1)	(2)	(3)	(4)
$\ln FRA$	0.4289 *** (0.0831) [0.2657, 0.5922]	0.1551 * (0.0788) [0.0001, 0.3102]	0.3358 *** (0.0500) [0.2376, 0.4340]	0.0307 (0.0560) [−0.0793, 0.1408]
Control Variables	YES	YES	YES	YES
Fixed Years	YES	YES	YES	YES
Fixed Individual	YES	YES	YES	YES
N	585	330	450	465
R2	0.4457	0.2818	0.5609	0.2713

Note: *, and *** denote 10%, and 1% significance levels, respectively; standard errors are in parentheses; and the upper and lower limits of the 95% confidence interval are shown in square brackets.

shows the test results for developed economies in item (1) and for developing economies in item (2). The results show that compared to developing economies, digital infrastructure has a greater impact on developed economies. This might be because economies with a higher level of economic development can provide sufficient technical, talent, and financial support for the construction of digital infrastructure [70]. Compared to developed economies, developing economies have problems such as limited human capital supply scale, relatively backward technological levels, and imperfect market mechanisms, which have led to the potential value of digital infrastructure not being fully realized.

Table 5 shows the test results of economies with high digital infrastructure levels in item (3) and those with low digital infrastructure levels in item (4). The results show that digital infrastructure has a significant positive impact on the added value of agricultural exports in high-level economies, while the impact on low-level economies is not significant. This might be because economies with low levels of digital infrastructure are relatively backward in terms of network coverage, data center construction, cloud computing capabilities, etc., which makes agricultural enterprises face many difficulties when using digital technologies for production and sales [70].

4.5. Mediation Effects Test

Referring to relevant research, this paper examines whether the construction of digital infrastructure can influence the added value of agricultural exports through mechanisms such as enhancing labor productivity, optimizing the business environment, and promoting technological innovation [71,72,73].

The model is as follows:

$$\ln L{P}_{it}={\beta }_0+{\beta }_1\ln FR{A}_{it}+{\beta }_2{X}_{it}+{\delta }_i+{\lambda }_t+{\epsilon }_{it}$$

(5)

$$\ln D{B}_{it}={\theta }_0+{\theta }_1\ln FR{A}_{it}+{\theta }_2{X}_{it}+{\delta }_i+{\lambda }_t+{\epsilon }_{it}$$

(6)

$$\ln NO{V}_{it}={\gamma }_0+{\gamma }_1\ln FR{A}_{it}+{\gamma }_2{X}_{it}+{\delta }_i+{\lambda }_t+{\epsilon }_{it}$$

(7)

where subscripts i and t denote country (region) and year, respectively. $\ln LP$ denotes the logarithmic value of labor productivity; $\ln DB$ denotes the logarithmic value of the Ease of Doing Business index; $\ln NOV$ denotes the logarithm of the total number of scientific and technical papers; and all other variable settings are consistent with the baseline model. In this case, the sample for the study of technological innovation mechanisms was deleted from the original with one economy with significant data gaps.

Column (1) of

Table 6. Mediation effects test.

Variables	(1)	(2)	(3)
$\ln FRA$	0.1285 *** (0.0076) [0.1136, 0.1434]	0.0207 *** (0.0060) [0.0088, 0.0325]	0.3944 *** (0.0274) [0.3407, 0.4481]
KHB (Reduced)	0.4830 *** (0.0497) [0.3856, 0.5803]	0.4484 *** (0.0356) [0.3786, 0.5182]	0.7785 *** (0.0627) [0.6556, 0.9014]
KHB (Full)	0.3545 *** (0.0349) [0.2862, 0.4228]	0.3366 *** (0.0307) [0.2764, 0.3968]	0.2050 *** (0.0391) [0.1284, 0.2816]
KHB (Diff)	0.1285 ** (0.0639) [0.0031, 0.2538]	0.1118 *** (0.0361) [0.0411, 0.1826]	0.5735 *** (0.0859) [0.4052, 0.7418]
Control Variables	YES	YES	YES
Fixed Years	YES	YES	YES
Fixed Individual	YES	YES	YES
N	915	915	900
R2	0.6375	0.4955	0.6738

Note: **, and *** denote 5%, and 1% significance levels, respectively; standard errors are in parentheses; and the upper and lower limits of the 95% confidence interval are shown in square brackets.

presents the mechanism test results of labor productivity. The results show that the coefficient of digital infrastructure is significantly positive, indicating that the construction of digital infrastructure can effectively promote the growth of labor productivity. Combined with the KHB test, it can be concluded that the construction of digital infrastructure can promote the profits of agricultural export trade by enhancing labor productivity [43]. Therefore, H₂ was verified.

Column (2) of Table 6 presents the mechanism test results of the business environment. The results show that the coefficient of digital infrastructure is significantly positive, indicating that the improvement of digital infrastructure levels has promoted the optimization of the business environment. Based on the KHB test, it can be found that the development of digital infrastructure can alleviate financing constraints, enhance market transparency, improve the efficiency of government services and supervision, and thereby optimize the international business environment. The optimization of the business environment is conducive to providing sufficient financial support and a fair and effective market mechanism for agricultural export trade enterprises, thereby increasing the added value of agricultural exports [46]. Therefore, H₃ was verified.

Column (3) of Table 6 presents the mechanism test results of the technological innovation effect. The results show that the coefficient of digital infrastructure is significantly positive, indicating that digital infrastructure has promoted the improvement of the innovation level. Based on the KHB test, it can be known that the construction and improvement of digital infrastructure is conducive to promoting the circulation of innovative elements such as information, capital, and talent, thereby promoting technological innovation, and technological innovation is the internal driving force for increasing the added value of agricultural exports [52]. Therefore, H₄ was verified.

4.6. Moderating Effects Test

To examine the moderating role of participation in the global value chain of agriculture, this paper introduces agricultural GVC participation and its interaction with digital infrastructure for moderating effect analysis using the following econometric model:

$$\ln DV{A}_{it}={\alpha }_0+{\alpha }_1\ln FR{A}_{it}+{\alpha }_2\ln FR{A}_{it}\times GVC{P}_{it}+{\alpha }_{3}GVC{P}_{it}+{\alpha }_4{X}_{it}+{\delta }_i+{\lambda }_t+{\epsilon }_{it}$$

(8)

where $GVCP$ denotes agricultural GVC participation, and all other variable settings are consistent with the baseline model. The results are shown in

Table 7. Moderating effects test.

Variables	(1)	(2)
$\ln FRA$	0.1425 *** (0.0360) [0.0718, 0.2132]	0.2090 *** (0.0501) [0.1106, 0.3073]
$GV{C}_{\mathrm{f}}$	2.8264 *** (0.1684) [2.4958, 3.1570]
$\ln FRA\ast GV{C}_f$	−0.6529 *** (0.1412) [−0.9300, −0.3757]
$GV{C}_b$		0.9975 *** (0.3813) [0.2492, 1.7459]
$\ln FRA\ast GV{C}_b$		0.6751 ** (0.2555) [0.1735, 1.1767]
Control Variables	YES	YES
Fixed Years	YES	YES
Fixed Individual	YES	YES
N	915	915
R2	0.5371	0.3514

Note: **, and *** denote 5%, and 1% significance levels, respectively; standard errors are in parentheses; and the upper and lower limits of the 95% confidence interval are shown in square brackets.

.

Column (1) of Table 7 indicates that the interaction between forward participation in the global agricultural value chain and digital infrastructure is significantly negative, opposite to the coefficient direction of the core explanatory variable, indicating a clear substitution relationship between digital infrastructure construction and forward participation in the global agricultural value chain. When agriculture has not yet engaged in forward participation in the division of labor in the value chain, the construction of digital infrastructure provides a material foundation for broadening trade channels and reducing trade costs [74], and business entities can make use of the Internet, the Internet of Things, big data, and other technological means to achieve the maximization of trade benefits. However, as forward participation in the global agricultural value chain increases, enterprises may allocate more resources to their involvement in the global value chain, while using fewer resources for the construction and application of digital infrastructure, thereby weakening the influence of digital infrastructure on the added value of agricultural exports [58]. Therefore, forward participation in the global agricultural value chain has a negative moderating effect on the impact of digital infrastructure on the added value of agricultural exports. Therefore, H₅ was verified.

Column (2) of Table 7 indicates that the interaction between backward participation in the global agricultural value chain and digital infrastructure is significantly positive, in the same direction as the coefficient of the core explanatory variable, suggesting that backward participation in the global agricultural value chain positively moderates the promoting effect of digital infrastructure. When a country (region) participates backwards in the global value chain, it can enhance its technological innovation capacity and human capital level by learning and imitating advanced production technologies [61], which is conducive to mobilizing the enhancement effect of digital infrastructure on agricultural trade benefits. Therefore, the higher the backward participation in the global agricultural value chain, the stronger the influence of digital infrastructure on the domestic added value of agricultural exports. Therefore, H₆ was verified.

5. Discussion

This article analyzes the role of digital infrastructure construction and participation in the global value chain of agriculture on the added value of agricultural exports. The main findings are as follows: First, the construction of digital infrastructure has a positive effect on the increase of the added value of agricultural exports, and its impact on developed economies and high levels of digital infrastructure is obvious. Liu’s [13] research indicates that the digital economy has significantly increased agricultural product exports by reducing trade costs and intensifying market competition. Zhou’s [17] study pointed out that the development of broadband infrastructure can significantly promote the growth of export trade in Chinese cities. Broadband infrastructure affects export trade through direct channels that enhance information efficiency, thereby reducing logistics costs, improving trade efficiency, and lowering trade barriers. Tang [70] conducted a study using panel data from 30 regions in China from 2006 to 2017 and found that the positive impact of digital infrastructure shows significant heterogeneity. In regions with high levels of economic development, research and development, and traditional infrastructure development, the positive impact of digital infrastructure is more pronounced. In conclusion, relatively few studies have examined the relationship between digital infrastructure and agricultural product trade from the perspective of added value. Based on the elaboration of the mechanism by which digital infrastructure promotes the added value of agricultural exports, this article once again confirms the viewpoint that digital infrastructure drives the added value of agricultural exports through empirical analysis.

Second, the increase in the added value of agricultural exports through the construction of digital infrastructure is achieved by enhancing labor productivity, optimizing the business environment, and promoting technological innovation. Aleca [43] found that the use of digital infrastructure plays a significant role in enhancing labor productivity in EU countries. Verhoogen [45] pointed out that factories with higher labor productivity produce higher-quality goods and export a larger share than those with lower labor productivity. Li [46] pointed out that the construction of digital infrastructure can optimize the business environment. Cui’s [50] research, which included an analysis based on the gravitational equation model, found that a favorable business environment can promote trade openness, and countries with lower development levels benefit more from a continuously improving business environment. Hussain [52] pointed out that digital infrastructure plays a crucial role in enhancing innovation performance. Osei [54], in an analysis based on data from 28 African countries from 2011 to 2019, found a positive correlation between digital infrastructure and innovation. Cao’s [55] research indicates that compared to traditional agricultural supply chains, platform participation based on new technologies such as blockchain can increase the total surplus. Li’s [56] research indicates that technological progress has provided new methods for enhancing traceability systems. An agricultural product traceability system developed using blockchain technology can effectively improve the efficiency of the agricultural product trade system. These scholars respectively analyzed the relationships among labor productivity, the business environment, technological innovation, digital infrastructure, and trade but did not incorporate them into a unified framework for research. This study finds that digital infrastructure can boost the added value of agricultural exports by enhancing labor productivity, optimizing the business environment, and promoting technological innovation.

Third, forward participation in the global agricultural value chain will negatively moderate the influence of digital infrastructure construction on the added value of agricultural exports. Backward participation will play a positive moderating role in the impact of digital infrastructure construction on the added value of agricultural exports. Calatayud [75] examined the impact of manufacturing enterprises’ participation in the global value chain in Sub-Saharan Africa. The research indicates that participation in the global value chain has a positive effect on productivity. Wang’s [76] research indicates that uncertainty has suppressed companies’ exports, but participating in the global value chain helps to mitigate the negative impacts brought by uncertainty. Tian’s [61] research indicates that backward participation in the global value chain is conducive to the progress of developing countries, which can import advanced products and learn from them. Forward participation in the value chain is more beneficial for developed countries. At present, there are relatively few studies analyzing the moderating role of the global agricultural value chain in the impact of digital infrastructure on the added value of agricultural exports. The analysis in this article is conducive to providing a theoretical basis for promoting the international trade of agricultural products and integrating them into the global value chain system.

This article also has some limitations and shortcomings. First, regarding the selection of indicators for the level of digital infrastructure, this paper refers to some excellent journal articles and strives to choose indicators that are widely recognized and well-established by the majority of scholars as much as possible. However, given the data availability and complexity of digital infrastructure indicators, the selected indicators may have limitations. If other indicator data can be obtained in the future, the selection of digital infrastructure indicators can be improved. Second, when analyzing the mechanism by which digital infrastructure promotes the added value of agricultural exports, this paper only selected some mediating variables and did not analyze or verify other possible mediating variables. Further exploration of other channels of action is needed in future research. Third, this paper selects 61 economies as the research objects. The research conclusions and suggestions may not be fully applicable to other countries and regions around the world. If more data on economies can be obtained in the future, more comprehensive research can be conducted. Fourth, this paper does not take into account the possible impacts of significant differences in technology diffusion among different countries. If more data can be obtained in the future, a more comprehensive study will be performed.

6. Conclusions and Policy Implications

6.1. Conclusions

Based on the panel data of 61 economies from 2007 to 2021, this paper examines the impact of digital infrastructure construction and participation in the global agricultural value chain on the added value of agricultural exports and draws the following research conclusions. First, the construction of digital infrastructure is conducive to promoting an increase in the added value of agricultural exports. Second, an increase in the added value of agricultural exports through the construction of digital infrastructure is achieved by enhancing labor productivity, optimizing the business environment, and promoting technological innovation. Third, forward participation in the global agricultural value chain plays a negative moderating role in the impact of digital infrastructure construction on the added value of agricultural exports. Backward participation in the global agricultural value chain plays a positive regulatory role. Fourth, the impact of digital infrastructure construction on developed economies and those with higher levels of digital infrastructure construction is more pronounced.

6.2. Policy Implications

Therefore, this article puts forward the following countermeasures and suggestions: First, in the process of promoting international trade in agricultural products, the construction of digital infrastructure should be gradually improved. On one hand, we should continue to increase investment and support in the field of digital infrastructure, promote the overall construction level of digital infrastructure within regions, and narrow the digital divide. On the other hand, we should use the role of digital infrastructure as an information bridge to improve the level of agricultural informatization, independent innovation ability, and management decision-making ability of agricultural business entities, and give full play to its enhancement effect on the added value of agricultural trade. Second, we should strengthen international cooperation in agriculture and integrate more deeply into the global value chain system. This includes encouraging agricultural enterprises to participate in the international division of labor and expand overseas markets, making full use of digital information platforms and regional cooperation platforms to improve the level of embeddedness in the value chain and their ability to control it, expanding the space and potential for agricultural value chain cooperation, reducing the inhibiting effect of embeddedness in the global value chain on the enhancement of the benefits of agricultural trade, and strengthening the positive impact of embeddedness in the global value chain. Third, economies at different levels of development should adopt differentiated trade strategies. Economies with relatively low levels of economic development and digital infrastructure development can give priority to strengthening talent and cultivating education, promoting scientific and technological innovation, optimizing industrial structure, improving market mechanisms, and enhancing regional cooperation, etc., to gradually narrow the gap between them and developed economies. For instance, in the process of participating in the global value chain, developing economies can learn advanced technologies from other economies, enhance their control and autonomy over the value chain through the learning-by-doing effect, and break through the predicament of low-end lock-in. Economies with high levels of economic and digital infrastructure development should, on the premise of increasing financial input into digital infrastructure construction and establishing and improving regulatory mechanisms, further deepen international agricultural cooperation, actively play a leading role, and create a favorable international business environment to better promote the improvement of agricultural trade benefits.