Evolution Modes of Chili Pepper Industry Clusters under the Perspective of Social Network—An Example from Xinfu District, Xinzhou, Shanxi Province

Abstract

This study evaluates the progression and influencing factors of the chili pepper industry cluster in Xinzhou City, Shanxi Province from 2006 to 2020 from a social network standpoint, using both theoretical and empirical methods as well as incorporating field survey data. The findings reveal the following facts: (1) the chili pepper industry cluster underwent a steady evolution in the social network over the course of 15 years, evidenced by an increase in the network clustering coefficient from 0.157 to 0.470. The network scale expanded from 9 to 76 entities; thus it basically achieved maturity; (2) the development modes of the chili pepper industry cluster in Xinfu District can be summarized as follows: an “embryonic stage” (2006–2010), an “initial stage” (2011–2015), and a “developmental stage” (2016–2020), which are marked by a broker-centered industry mode during the embryonic stage, a cooperatives-centered industry mode during the initial stage, and a chili pepper association- and leading enterprise-centered industry mode during the developmental stage; (3) the policies, fund, market, labor, and external capital have a significant impact on the development of the chili industry cluster in the Xinfu District. During the embryonic stage, the primary influencing factors are fund (0.326) and market (0.309). During the initial stage, the primary influencing factors are market (0.162) and external capital (0.135). During the developmental stage, the primary influencing factors are policy (0.232) and market (0.232), with technology (−0.102) serving as a limiting factor. It is crucial to take into account natural resource endowment and industry mode features, foster technological advancement, and spur social capital involvement in developing chili pepper industry clusters. The government must create a supportive external environment for the chili pepper industry cluster’s growth to establish a solid foundation for the high-quality advancement of the agricultural industry cluster. The insights derived from this study can serve as a reference and source of inspiration for the growth of other vegetable industry clusters in China.

1. Introduction

Agricultural industry clusters play a crucial role in catalyzing industrial growth and invigorating rural regions, and are significant components of the country’s strategy to address imbalanced development in the agricultural sector [1]. The regional and specialized nature of these clusters enable a division of labor and collaborative efforts, a highly efficient innovation system, and a comprehensive industrial ecosystem, which traditional agriculture does not possess, thereby maximizing regional advantages and fostering agricultural industry innovation [2]. The vegetable industry, characterized by its high scale efficiency, minimal restrictions on greenhouse agriculture, and substantial technical contribution rate, is particularly well-suited to promote the benefits of agricultural industrial clustering. Vegetables, a staple food source for a significant portion of the world’s population [3], serve as an important economic pillar in the development of agriculture in many countries [4]. China is the world’s leading vegetable producer and consumer [5], and the vegetable industry generally boasts higher added value compared to major grain industries, making a significant contribution to comprehensive poverty alleviation and rural revitalization efforts in the country. Chili peppers, ranked as the third-largest vegetable crop globally after legumes and tomatoes [6], and the most extensively cultivated vegetable in China, serve as an instructive example for the study of industrial cluster formation and evolution. This study holds great value as a reference for the development of China’s vegetable industrial clusters.

The concept of social networks was initially propounded by the German sociologist Georg Simmel [7] and, following an extended period of theoretical development, has become increasingly crystallized: social relationships formed by the interaction between members and social networks, which are focused on the relationships and interaction of members [8]. The research methods employed in the study of social networks place a greater emphasis on examining relationships and interactions between individuals [9], rather than the conventional sampling survey techniques which are commonly utilized in sociology. These methods are widely employed in a variety of fields, including sociology, communication, and psychology [10,11,12], with many scholars focusing on the examination of individual relationships within networks [13,14], while only a limited number of them apply social network analysis to the study of agricultural industry clusters. The concept of industry clusters was first introduced by Porter [15], and the focus of his research was initially on the industrial sector [16,17]. However, with the advancement of agricultural modernization, agricultural industry clusters have emerged as a quintessential representation of this phenomenon in China and have become crucial in deepening our understanding of agricultural modernization [18]. Currently, most researchers concentrate on the competitiveness and impacts of agricultural industry clusters [19,20,21], while the underlying mechanisms of their formation are comparatively less explored [2]. This study combines social network analysis with actual survey data to investigate the network relationships between industrial entities in the Xinfu District chili pepper industry cluster and to characterize the pattern of formation and evolution of the studied chili pepper industry cluster between 2006 and 2020. The industrial entities encompass companies, cooperatives, and industry associations within the cluster, and the study draws upon network relationship and difference analyses of different stages of the industry cluster [22].

The Xinfu District of Xinzhou City in Shanxi Province boasts a rich history of chili pepper cultivation which is characterized by its optimal natural conditions. Over the years, the scale of chili pepper cultivation has consistently exceeded 100,000 mu (approximately 16,474 acres), making it a significant producer of chili peppers in Shanxi Province and North China. Through extensive industry growth, the Xinfu District has formed a prototypical chili pepper industry cluster that covers the cultivation, processing, sales, and research and development aspects of the industry. This paper selects the chili pepper industry cluster of the Xinfu District as its research subject, and analyzes the defining features, modes, and determining factors of this industry cluster with the aim of providing a scientifically grounded and data-driven basis for its continued high-quality development. Furthermore, this study establishes an instructive reference for the examination of the characteristics, evolution, and motivating forces behind other vegetable-based industry clusters.

2. Materials and Methods

2.1. Study Area

The Xinfu District in Xinzhou City, Shanxi Province is positioned between from N38°13′ to 38°41′ and from E112°17′ to 112°58′ (see Figure 1). The district covers a total area of 1954 square kilometers and is distinguished by its topographical features, which are characterized by a hill-laden terrain with a westward slope that descends towards the east and that is encircled by mountains on three sides. It is endowed with a temperate continental monsoon climate and has an average temperature of over 15 °C from April to October, providing an optimal climate for the cultivation of chili peppers. The chili pepper industry in the Xinfu District comprises three primary categories: pigment chili, dried chili, and long slim chili, encompassing roughly 12 varieties, including locally grown L3 and Xinlu chili, which are renowned for their superior quality. The main chili pepper varieties generally exhibit attractive colors and agreeable spiciness, contributing to the rapid development and widespread distribution of the chili pepper industry, both domestically and internationally. The majority of sales are primarily within the domestic market, while international sales serve as a supplement. The chili pepper industry has formed a significant and influential chili pepper industry cluster.

2.2. Data Sources and Processing

The data utilized in this study were obtained through field research and telephonic interviews. Between April and June of 2021, the chili pepper industry within the specified region was subjected to examination through 58 in-person surveys and 18 telephonic interviews, yielding a total sample size of 76 respondents and a response rate of 97.5%. Of these participants, cooperatives constituted the most significant demographic, accounting for 88% of the sample, while companies and industry associations represented 10.5% and 1.5%, respectively. The predominance of cooperatives can be attributed to the predomination of cooperative development within the chili pepper industry cluster in Xinfu District, as opposed to a comparatively limited number of enterprises. All research participants, except industry associations, are referred to using their initials in this study, as documented in

Table 1. Abbreviations of the subjects of the survey.

1. Pepper Association	2. WMCX	3. CX	4. HTY	5. WHS	6. CS	7. KP	8. HYL
9. ZJ	10. CY	11. WM	12. HBNF	13. MS	14. SY	15. HuiF	16. HY
17. AM	18. QS	19. FTM	20. TDH	21. QEH	22. JS	23. JY	24. XY
25. TTX	26. XSY	27. JM	28. QR	29. WXN	30. LP	31. SD	32. LLXM
33. TD	34. CSL	35. XDF	36. XP	37. BLT	38. HYN	39. WL	40. YH
41. LT	42. PH	43. CYN	44. JYW	45. JJ	46. BY	47. JX	48. LQ
49. Hyun	50. HYT	51. YS	52. SX	53. HJ	54. HL	55. SZM	56. CSZZ
57. LY	58. GYH	59. SH	60. SL	61. EX	62. JW	63. XT	64. NXF
65. JYF	66. SC	67. AW	68. RF	69. QJX	70. DY	71. JFT	72. HF
73. GZ	74. GY	75. QX	76. YW

.

2.3. Methods

2.3.1. Main Network Features

By using the UNCINET software, the network relationships between entities in the industrial cluster were constructed. The primary indicators include network density, clustering coefficient, average path length, and network centrality, which reflect the development characteristics of the Xinfu chili pepper industrial cluster’s social network. The specific formulas are as follows:

①

Network density:

Network density refers to the ratio of the number of actual edges in the social network of the industrial cluster to the number of maximum possible edges. The network density of the cluster shows the development level of the cluster as a whole, and also reflects the closeness of the connection between entities. The calculation formula is as follows:

$$\frac{m}{n\left(n-1\right)\raisebox{1ex}{$1$}\!\left/ \!\raisebox{-1ex}{$2$}\right.}=\frac{2m}{n\left(n-1\right)}$$

(1)

where n is the total number of individual entities in the chili pepper industrial cluster and m refers to the actual number of relationships contained in the chili pepper industrial cluster. n(n−1)/2 is the maximum number of relationships that can theoretically be formed. The higher the network density value, the closer the connection between entities.

②

Clustering coefficient

The clustering coefficient is mainly used to describe the degree to which the entities in the social network are connected into groups. The formula for the clustering coefficient C, defined from the perspective of transitivity according to Barrat and Weight, is shown as follows [23]:

$$C=\frac{3N\mathsf{\Delta }}{N_3}$$

(2)

where N₃ represents the total number of potential triangles in the social network model, and 3NΔ represents the number of groups that connected three nodes in the social network model.

③

Average path length

The average path length represents the average length of the path between any two entities in the social network. The average path length could reflect the strength of the relationship between entities in the social network. The smaller the average path length, the greater the connectivity of the nodes in the network. In this paper, an undirected and unweighted network is selected. Thus, L is the average distance between any of the two nodes:

$$L=\frac{1}{\frac{1}{2}n\left(n+1\right)}\sum d_{ij} i>j$$

(3)

In the formula, n represents the number of chili pepper industry entities in the social network, and dij represents the shortest path length from i to j.

④

Intermediary centrality

Intermediary centrality represents the degree to which entities of the social network are in the center of the network, and the calculation formula is as follows:

$$B{C}_{A{B}_i}={\mathsf{\Sigma }}_j^n{\mathsf{\Sigma }}_k^n{b}_{\dot{J}k}\left(i\right)\mathit{,\hspace{1em}}j\ne k\ne i\wedge j<k$$

(4)

where bjk (i) refers to i’s ability to control j and k, and the probability that i is on the shortest path between j and k.

2.3.2. Quadratic Assignment Problem (QAP) Analysis

QAP is a sophisticated method of social network analysis that compares the grid values of corresponding elements across multiple matrices to derive relationship coefficients. Based on relevant QAP studies [24,25,26] and an understanding of the actual background of the chili pepper industry cluster in Xinfu District, six key indicators, including technology, market, labor, capital, policy, and external capital, were selected to identify the impact factors for the evolution of the cluster social network. This paper employs QAP regression analysis to evaluate the regression relationship between multiple matrices of the chili pepper industrial cluster in Xinfu District and a single matrix. The specific equation used in the analysis is represented as follows:

$$Y={\theta }_n+{\theta }_1{x}^{\left(1\right)}+{\theta }_2{x}^{\left(2\right)}+\cdots {\theta }_n{x}^{\left(n\right)}+a$$

(5)

where Y and X(i) represent n-order relationship matrices, and θ₀, θ₁, and θ₂ are regression coefficients.

The independent and dependent variables analyzed through QAP regression are presented in matrix form, as illustrated in the equation below:

$$Y=\left(\begin{array}{ccc}0& \cdots & y_{1,n}\\ ⋮& \ddots & ⋮\\ y_{n,1}& \cdots & 0\end{array}\right)$$

$$X^i=\left(\begin{array}{ccc}0& \cdots & X_{1,n}^i\\ ⋮& \ddots & ⋮\\ X_{n,1}^i& \cdots & 0\end{array}\right) i=1,\cdots ,n$$

(6)

3. Results

3.1. Analysis of the Evolution Modes of Chili Pepper Industry Cluster

This investigation employs a blended approach of field surveys and data analysis to comprehensively elucidate the seminal occurrences in the progression of the chili pepper industry cluster in Xinfu District. By conducting structured interviews, it was determined that several key events had a significant impact on the development of the cluster, including the predominance of brokers in 2006, the shift from the broker form to the cooperative form around 2010, the arrival of a leading chili pepper enterprise from Hunan Province in 2015, the establishment of the “China (Xinzhou) Chili pepper Industry Park” in 2017, and the formation of the Xinfu Chili Association in 2019. The study selected three time nodes, in 2010, 2015, and 2020, to demarcate the development of the Xinfu chili pepper industry cluster into three epochs, i.e., from 2006 to 2010, from 2010 to 2015, and from 2015 to 2020. The methodology of social network analysis is then employed to characterize the network relationships of the chili pepper industry cluster in each period, resulting in cluster network structure diagrams for 2010, 2015, and 2020, as well as the clustering coefficient, network density, and average path length values of the cluster networks. With the support of field survey information and the results of the analysis of the social network, this study seeks to synthesize the salient features and developmental trends of each stage. The field survey information and data model analysis corroborate one another, thus establishing three historical stages in the progression of the chili pepper industry cluster in Xinfu District, namely the “embryonic stage”, the “initial stage”, and the “developmental stage”.

3.1.1. Analysis of Social Network Characteristics

As evidenced by

Table 2. Data on the main characteristics of social networks.

Index	2010 Numeric Value	2015 Numeric Value	2020 Numeric Value
Number of subject	9	30	76
Agglomeration coefficient	0.157	0.298	0.470
Network density	0.236	0.236	0.064
Average path length	2.063	2.010	2.739

, the number of cluster entities, which stood at 9 in 2010, has consistently grown, reaching 76 in 2020, exhibiting a substantial expansion of the cluster network. Over the course of a decade, the number of cluster entities has increased by 10.7 times, indicating a positive trajectory in the development of the industry’s foundation. The clustering coefficient, which stood at 0.157 in 2010, has steadily risen to 0.470 in 2020, with an upward fluctuation of 0.313, demonstrating an increasingly stronger clustering trend among the entities within the network and a gradual establishment of connections and clustering cores. The network density of the cluster remained at 0.236 in both 2010 and 2015. However, it dropped to 0.064 in 2020, indicating a decrease in the closeness of connections among entities, leading to a more dispersed overall network. The average path length of the cluster experienced significant changes, with a low level of 2.010 in 2015 and a rise to 2.739 in 2020, reflecting a transformation from strong to weak average relationship intensity among entities from 2010 to 2020.

3.1.2. Embryonic Stage (2006–2010)

As depicted in Figure 2, the operation mode of the chili pepper industry cluster is relatively straightforward in this stage. During this period, the chili pepper industry cluster was in its infancy stage, and the operating entities were primarily chili brokers. The primary mode of operation involved brokers purchasing chili peppers from farmers and exporting them to remote markets to capture the price differential. The development of the industry at this stage was contingent upon two factors. On the one hand, the inherent natural resources and advantageous growing conditions ensured the yield and quality of chili peppers. On the other hand, brokers leveraged their informational asymmetry to establish a small-scale market monopoly through their personal connections to markets of other provinces. By 2010, there were only nine related entities that had established preliminary cooperative relationships, mostly with a single-faceted cooperative path and connections to two–three other entities. According to the calculations carried out in this study, the top three entities in terms of centrality were QS (31.25), JLL (15.18), and JCHF (12.50), with a substantial disparity in centrality. During this period, the network had yet to form a significant center due to the small scale of the chili pepper industry and that interdependent relationships could easily form among entities.

Based on the above analysis, this article categorizes the 2006–2010 development period as the “embryonic stage” of the Xinfu District chili pepper industry cluster, and establishes the industrial model (see Figure 3) of this period. Characteristics of the embryonic stage include a relatively small market scale, simple market structure, low level of industrial maturity, short industrial chain, and limited connections among the entities. During this stage, the industry is in its nascent stage and is influenced by multiple factors which may pose difficulties in industrial transition or cause instability in the industrial chain. To facilitate the growth of the industry, it is necessary to continue to leverage the advantages of natural resources while also striving to build the local market, enhance the industrial chain, and establish a more mature local industrial network.

3.1.3. Initial Stage (2011–2015)

As depicted in Figure 4, cooperatives have begun to play a crucial role in the operation of the chili pepper industry cluster, yet brokers remain active within the market domain. The dominant business mode entails the cooperatives acquiring chili peppers from farmers and brokers, with some of the chili pepper raw materials undergoing initial processing by the cooperatives, which then become advanced raw materials for sale to external markets, while the other portion of chili peppers is sold directly to external markets in the raw form. From 2011 to 2015, the cluster network underwent a breakthrough development, and a relatively complex network structure took shape, with high levels of internal interconnectivity and low dispersal within the network, indicating close collaboration and effective communication among the entities, and with a high proportion of acquaintances. The top three entities during this period were XCD (11.22), TD (8.78), and ZH (8.37), with relatively small disparities among them and no significant network center, leading to balanced development among the various operators and high participation of the emerging actors.

Based on the analysis above, this paper categorizes the period from 2011 to 2015 as the “initial stage” of the chili pepper industry cluster in Xinfu District (see Figure 5). This stage was characterized by vigorous growth and dynamism, with the formation of a nascent local market and the gradual refinement of the supply chain. The cluster’s ability to spur the development of local farmers also increased, along with a rapid growth in the number of participating entities. The evolution of an agricultural industry cluster into its initial stage typically indicates a region’s long-term potential for developing the industry, along with a foundational level of processing capabilities. However, the structure of the industry cluster was not yet fully established during this stage, and the overall level of processing was often low. Information exchange was scarce, making it difficult to form a mature industrial system, and the cluster’s overall competitiveness could not be guaranteed.

3.1.4. Developmental Stage (2016–2020)

During this period (see Figure 6), the “China (Xinzhou) Chili Pepper Industry Park” was established, and as a result, several leading enterprises emerged within the cluster. At the same time, an industry association appeared within the cluster, connecting small and micro cooperatives, as well as farmers and leading enterprises in the park, providing an information exchange platform for the entire industry and enhancing the efficiency of resource and information flow. The industrial mode of this period matured, with farmers, brokers, and cooperatives obtaining chili seedlings for planting and processing through leading enterprises. The chili pepper association integrated market information and fed back the information to farmers, brokers, and cooperative societies. Finally, leading enterprises began to process chili within the cluster, with the chili entering the market as high-grade raw materials after preliminary processing, as well as entering the market in the form of chili sauce, capsaicin, and chili essence after intensive processing. In this stage, the cluster had achieved resource integration and formed an industrial chain of “seedling cultivation—planting—preliminary processing—intensive processing,” significantly increasing the added value of the product. As can be seen from Figure 6, the scale of the network was large and the number of entities was substantial in this period, with obvious network centers and a mature network structure. The top three entities, with the highest centrality in the network, were pepper association (37.51), TTX (13.76), and LT (11.40), with pepper association having extremely high centrality in the network and a tremendous impact on the development of the entire network. TTX and LT were both leading enterprises, which have a significant impact on the cluster’s development. Furthermore, data from Figure 6 and Table 2 also indicate that the network was relatively dispersed during this period due to the rapid development of the cluster and a large number of emerging entities. Although the emerging entities were connected to traditional ones, their participation in the cluster network was relatively low, and the connection and collaboration among the entities in the cluster were relatively lagged compared to the expansion speed of the cluster scale.

Based on the analysis outlined above, this study categorizes the period from 2016 to 2020 as the “developmental stage” of the chili pepper industry cluster’s development in Xinfu District (see Figure 7). During this stage, the cluster boasts a relatively comprehensive industrial ecosystem, featuring a substantial local market size, a brisk increase in the number of entities, an ongoing refinement of the industrial chain, the advent of leading enterprises and trade associations, and systematic policy support. The transition to the “developmental stage” signifies that the cluster has attained an initial degree of general competitiveness and holds the potential to form a preeminent industry cluster on a national scale. Nevertheless, there are still persisting issues at this stage, such as an unfinished industrial chain, a yet-to-be-established regional industrial brand, relatively insufficient competitiveness among the leading enterprises, and a limited number of trade associations. Therefore, it is not appropriate to term this stage as the “culminated stage” or “mature stage,” but rather the “developmental stage” is more appropriate for the development characteristics of the industrial cluster during this period.

3.2. Interpretations of QAP Results

As indicated by

Table 3. Identification of influencing factors of pepper industrial cluster network in Xinfu District in 2010, 2015, and 2020.

Years	Index	Normalize Regression Coefficients	Significance Test
2010	Technology	0.212	0.078
	Policy	−0.094	0.477
	Labor	0.297	0.004
	Capital	0.326	0.003
	Market	0.309	0.006
	External capital	−0.179	0.120
2015	Technology	−0.010	0.411
	Policy	0.084	0.035
	Labor	0.122	0.004
	Fund	0.095	0.018
	Market	0.162	0.001
	External capital	0.135	0.002
2020	Technology	−0.102	0.008
	Policy	0.232	0.000
	Labor	0.102	0.014
	Fund	0.127	0.004
	Market	0.232	0.000
	External capital	0.132	0.003

, during the embryonic stage (2006–2010) of the cluster network relationship, only three variables—capital, market, and labor—were found to be valid through examination. The standardized regression coefficients of the three variables were 0.326, 0.309, and 0.297, respectively. This is due to the labor-intensive nature of the chili pepper industry, which is heavily reliant on labor in its early stages of development. The nascent cluster was unable to form strong competitiveness, attracting policy support and external capital, and its development could only rely on capital and market. The relatively small differences in the coefficients of capital, labor, and market during this period indicate that they have a similar impact on the formation of the cluster network.

During the initial stage (2010–2015), the chili pepper industry cluster in Xinfu District underwent a stage of rapid development. Factors such as policy, labor, capital, market, and external capital were evaluated, with standardized regression coefficients of 0.084, 0.122, 0.095, 0.162, and 0.135, respectively. Among these, market and external capital exhibited the highest standardized regression coefficients, due in part to the introduction of TTX, a leading chili pepper industry enterprise from Hunan, in 2012, and the subsequent merger and establishment of a subsidiary with the local leading cooperative XCD, resulting in the formation of a regional leader in the chili pepper industry. Although labor and capital remain key drivers, the standardized regression coefficients for these factors have significantly decreased compared to 2010, with the labor indicator dropping from 0.297 to 0.122 and the capital indicator declining from 0.326 to 0.095. This indicates that local enterprises are faced with increased opportunities and a certain degree of reduced reliance on existing factors.

During the developmental phase (2016–2020), the significance of technology, policy, labor, fund, market, and external capital was verified through standardized coefficients, with values of −0.102, 0.232, 0.102, 0.127, 0.232, and 0.132, respectively. Among these factors, policy and market had the highest standardized regression coefficients of 0.232, making them the most influential driving factors of network structure development in this stage. This is attributable to the initiation of the “China (Xinzhou) chili pepper industry Park” in 2017 by the Xinfu District government, located in the Gaocheng township, which accommodates 13 enterprises. The establishment of the park facilitated the gathering of industrial elements and the flow of knowledge and technology, resulting in a marked impact on the development of the cluster’s entities and forming an initial industrial chain as well as a group of leading enterprises. Hence, the chili pepper industry cluster in Xinfu District is gradually maturing.

For 2020, the standardized regression coefficient of technological factors was −0.102, exerting a restrictive effect on the development of the chili pepper cluster. This was due to the comparatively low investment and output of technology by the entities within the chili pepper industry in Xinfu District, which primarily existed as technological islands and failed to optimize their efficiency. There was a severe disconnect between the chili pepper cluster and technological factors. As revealed by interviews, enterprises within the region lacked technological innovation capacity, which, coupled with the absence of high-end new products for a prolonged period, eroded the confidence of some businesses to continue investing, thus resulting in the intensification of homogeneous competition among primary agents. Additionally, the slow development of crop protection technology, severe pest infestations, declining soil fertility, and difficulties in the recovery of abandoned plastic film further compounded the issue. All of these problems led to a reduction in the chili pepper production capacity and a decrease in farmers’ planting intentions in core production areas such as Gaocheng township in 2020.

4. Discussion

This paper endeavors to characterize the development stages of the chili pepper industry cluster in Xinfu District by utilizing a social network perspective, combining on-site research information with network data analysis. The characterization of the three stages may exhibit a certain degree of subjectivity and tentative nature; however, it aligns with the results of the data model analysis, thus exemplifying the reciprocal validation of research facts and scientific analysis. The purpose of stage division in this study is to uncover the formation and evolution mechanisms of the agricultural industry cluster through examination of its historical development cycles, and to provide a reference for future research on various agricultural industry clusters. Furthermore, the paper objectively investigates the impact factors of each historical period of the agricultural industry cluster through QAP analysis: delving deeper into the scientific question of “why” beyond explaining “what” constitutes its development cycle. However, compared to existing studies, this study may still have limitations, such as a limited data sample size, insufficiently meticulous stage division, and a lack of comparative case studies. Further research will be conducted utilizing multiple clusters and larger data samples for a more in-depth analysis, thereby refining the results of this paper.

During the embryonic stage of the chili pepper industry cluster, a lack of policy support and external capital rendered it heavily dependent on labor and funding. The labor-intensive industry demands a vast amount of labor. Prior to 2010, China’s rural labor force was relatively ample, allowing chili cultivators to hire workers at a relatively low cost to perform tasks such as planting, harvesting, loading, and transportation. However, research has shown that the rural labor force in Xinfu District has become severely insufficient and is significantly aging as of 2020 [27]. The majority of the youth labor force has been attracted to cities, and the age of the rural labor force is mostly around 60 years old, which has been accompanied by an increase in hourly wages. The decrease in production efficiency and the rise in labor costs pose a potential threat to the development of the chili pepper industry. When the demographic dividend disappears, the sustainable development of the industry usually requires innovation and upgrading through technological means [28]. This issue may not only exist in the chili pepper industry cluster, but is also a challenge faced in the development of other agricultural industry clusters [29,30].

During the initial phase, the critical factors of the industrial cluster lie in the brokers and cooperatives. Brokers play a unique and crucial role in rural China, often serving as intermediaries for information exchange and possessing specialized knowledge and market resources [31]. After the accumulating of experience during the embryonic stage, a significant number of brokers catalyzed the development of the industry by establishing cooperatives during the initial stage. Cooperatives have a certain degree of resource integration capabilities and possess preliminary processing capacities for agricultural products [32]. Compared to the embryonic phase, the emergence of cooperatives has brought greater standardization to the development of the industry and has also been more effective in inspiring farmers to participate in the chili cultivation sector.

During the developmental stage, two key factors emerged in the chili pepper industry of Xinfu District: the trade association and leading enterprises. The entry of leading enterprises brought the standardization of technology and regulation of management models, as well as substantial demand to the local market, creating an impact on the local market while also infusing the market with robust vitality [33]. Ultimately, the entry of leading enterprises was the result of policy support. Therefore, policy always plays a critical role in industrial development. The trade association provides a platform for information exchange and enables the cluster to evolve from an initially loose state to an organized and efficient existence. The trade association effectively serves as the pivot of the network.

The results of the QAP analysis indicated that technological constraints play a limiting role in the progress of the cluster during its developmental stage. Currently, the acquisition of specialized human capital and technological advancements is predominantly achieved through the employment of professional and technical staff. However, the lack of educational establishments and research institutions in the area, along with the absence of established expert workstations, highlights the criticality of technological innovation in propelling the growth of agricultural industry clusters. There have been several successful precedents in China, such as the vegetable industry cluster of Shouguang, Shandong province [34], which serves as a good reference for Xinfu District through incorporating science and technology companies, universities, and research institutions in the construction and development of agricultural industry clusters. When these educational and research bodies are integrated into the social network model as entities, they can significantly influence the network as a whole. Therefore, it is crucial for Xinfu District’s chili pepper industry to prioritize the significance of human capital and technology in its pursuit of sustainable industrial growth. To achieve this, investment in talent and technological advancements should be increased while consolidating the existing foundation of the industry.

The object of this study is the chili pepper industry cluster, whose development modes hold significance for both vegetable industry clusters and even agricultural industry clusters as a point of reference. However, the development of other vegetable industry clusters may differ from that of the chili pepper industry cluster, as even industry clusters with similar foundations may exhibit new characteristics and transformations due to differences in their underlying conditions and developmental paths. Throughout the three stages of change in the industrial mode in this study, the center of the industry evolved from brokers to cooperatives and, ultimately, to leading enterprises and trade associations. The question remains: will modernization factors impact the evolution of the industry’s center in the future? Another crucial direction for future research is the maintenance of data continuity and the ongoing regulation and optimization of the development mode of the studied industry cluster.

5. Conclusions

This study utilizes the method of social network analysis and incorporates research interview data to characterize the evolution of the chili pepper industry cluster network in Xinfu District at various stages. It clarifies the factors that influenced the development of the cluster and establishes development modes for each stage of the industrial cluster. The conclusions are summarized as follows:

(1)

Between 2006 and 2020, the number of entities within the chili pepper industry cluster social network in Xinfu District increased from 9 to 76, and the overall network density declined from 0.236 to 0.064. The network structure became increasingly integrated and evolved a clear network center.

(2)

Based on social network analysis and confirmed by actual research information, this study divides the development history of the cluster into three stages: the embryonic stage (2006–2010), the initial stage (2011–2015), and the developmental stage (2016–2020). During the embryonic stage, an industry mode centered on brokers was formed, while, during the initial stage, a cooperative-centered mode appeared. Then, in the development stage, a mode centered on the chili association and leading enterprises was formed.

(3)

Labor exerted a positive effect on all three stages; however, its influence waned, declining from a coefficient of 0.297 during the embryonic stage to 0.122 and 0.102 in the initial and developmental stages, respectively, representing a decrease of 0.195. Meanwhile, the impact coefficients of capital and market declined from 0.326 and 0.309 in the embryonic stage to 0.127 and 0.232 in the initial stage. Policy support, however, demonstrated greater potency in the developmental stage, with its coefficient rising from 0.084 to 0.232. The impact of external capital remained relatively consistent across both stages, being 0.135 and 0.132. Technology, on the other hand, emerged as a hindrance in the developmental stage, with a negative impact coefficient of −0.102.