According to the analytical framework of Nadvi and Barrientos [
6], the effects of industrial clusters on poverty reduction include direct and indirect effects. The direct effect is that farmers can participate in the cluster, which includes three aspects. The first is farmers can receive job opportunities. McCormick [
7] analyzes six cases of industrial clusters in Africa and finds that although industrial clusters only provide low-paid job opportunities, these job opportunities can help families survive and escape from poverty. Through investigating the evolution of industrial clusters of children’s clothing in Zhili County, Zhejiang Province, China, Fleisher et al. [
8] argue that these industrial clusters have employed more than 200,000 workers by 2005, accounting for approximately one-third of the employment of this industry in China. According to statistics on industrial clusters in Wenzhou City, Zhejiang Province, Sonobe et al. [
9] draw a conclusion that the average number of employees in Wenzhou increases from 47 in 1990 to 339 in 2000. The second is farmers’ income can be increased. Many studies conclude that industrial clusters can raise workers’ wages. Sonobe et al. [
9] emphasize industrial clusters have played an important role in increasing labor productivity. Visser [
10] examines Peru’s garment, he finds that the wage of firms in the cluster is higher than that of firms not in the cluster in the same industry. Fowler and Kleit [
11] consider that the number of industrial clusters has a significant effect on reducing poverty. The third is to enhance the ability. The feasible competence theory proposed by Amartya Sen indicates that poverty deprives the basic ability or power, not only low incomes [
12]. Because technology and information may spill in a region, the enterprises in the industrial clusters can further promote the spread and application of new knowledge, technologies and ideas among enterprises. Through “learning by doing” in the cluster, the capabilities of skills and management of entrepreneurs have been improved [
9]. In addition, industrial clusters also allow the poor to have more rights, especially for women, which is also an indication of improving their ability. Batra and Aneja [
13] explore the role of UNIDO’s textile cluster in Chanderi in improving women’s textile capacity, increasing their self-esteem and involving women more in industrial clusters; the indirect effect is mainly reflected in the externalities of the industrial clusters. First, agglomeration economies and joint actions. Agglomeration economies can reduce costs and allow small businesses to enter the market, which can increase their chances of obtaining jobs and improve their abilities to obtain more income. Cluster joint actions can further strengthen the capacity of local firms and reduce the vulnerability to external shocks [
6]. Second, in terms of promoting local economic development: Promoting the development of industrial clusters is widely regarded as a valuable goal of regional economic development [
11]. Industrial clusters, through providing high-quality products and services, can improve infrastructure, which further attracts suppliers, customers and public investors to invest in the infrastructure, which may benefit the whole business development in this region [
14].
There are few empirical articles on the role of industrial clusters in poverty alleviation. Nadvi and Barrientos [
6] believe that the role of industrial clusters in poverty alleviation is a subject that remains to be developed. They analyze this subject from three perspectives: The first is cluster characteristics. Some types of clusters may directly affect poverty, such as rural clusters, informal urban clusters, small and medium enterprises clusters, micro-enterprise or family workshop clusters, labor-intensive industrial clusters and employees consisting of women, migrant workers or non-skilled workers. The second is the processing of clusters. The scale effect of economic agglomeration can reduce transaction costs and improve employee capabilities. The third is cluster dynamics. Through joint operations, the capabilities of local companies have been improved and the risks caused by external shocks have been reduced. Fowler and Kleit [
11] are the first and probably the only ones to systematically verify the relationship between industrial clusters and poverty. Based on cross-sectional data at the county level in the United States, they used the spatial error model, under the control of the population, local economic factors, local geographical and spatial factors, concluding that the number and depth of industrial clusters in small and medium-sized cities across the country have a positive and significant effect on the reduction of poverty, while in big cities, the number of industrial clusters does not significantly reduce poverty. Long and Zhang [
15], although not directly studying the relationship between industrial clusters and poverty, focus on the research of the performance of industrial clusters, which is less involved in previous studies. Through the employ of China’s industrial census in 1995 and the economic census of enterprise-level micro-data in 2004, they emphasize that firms have better and full factor growth in some regions where China’s industrial clusters have developed well.
Industrial clusters may affect poverty in a variety of ways. The affection channels are mainly through farmers participate in industrial clusters, to obtain more job opportunities and non-agricultural income; through the externalities of industrial clusters, local farmers can obtain more convenient infrastructure and obtain access to the dripping effect from regional economic development. Based on this, Fowler and Kleit [
11] have added that due to population migration, the population base in the poverty regions will be reduced or increased, which in turn may increase or decrease the incidence of poverty. They provide a good theoretical stance and empirical reference for the study of the role of industrial clusters in poverty. However, industrial clusters in the United States are the agglomeration of large companies, while in China are the agglomeration of SMEs. The differences are relatively large. More importantly, the economic society environment, resource endowment and poverty conditions in the United States and China are also quite different. What is the role of industrial clusters with Chinese characteristics in poverty alleviation? Therefore, from the household-level, this paper will examine the impact of the number of industrial clusters, the type of industrial agglomerations and agglomeration levels on farmers’ income and rural poverty reduction.