Research on the Impact and Mechanism of Internet Use on the Poverty Vulnerability of Farmers in China

Abstract

Poverty vulnerability is an important indicator to achieve sustainable development of farmers out of poverty in China. Based on the data of China Family Panel Studies (CFPS) 2018, we research the impact and mechanism of Internet use on the poverty vulnerability of farmers in the context of the Internet plus strategy. The study found that under the poverty line of US $1.9 and US $3.1, vulnerable farmers accounted for 9.48% and 33.88% of the total sample, respectively. The use of the Internet can significantly reduce the poverty vulnerability of farmers. After using instrumental variables to overcome endogenous problems, and using the PSM method for robustness testing, the research conclusions are still valid. The mechanism shows that the use of the Internet reduces the level of poverty vulnerability by increasing income levels, enhancing the ability of farmers to obtain information, and promoting non-agricultural employment. Therefore, in the process of establishing and improving the long-term mechanism for poverty governance, it is necessary to fully promote the integration of the Internet and poverty vulnerabilities to further realize the long-term effectiveness and stability of poverty governance. Therefore, in the process of establishing and improving the long-term mechanism of poverty governance, we should first promote the full coverage of Internet infrastructure. Second, the government should improve the digital literacy of farmers. Third, the goal of the Internet plus strategy in the process of poverty control should be precise.

1. Introduction

Rural poverty is one of the great obstacles to the construction of a modern country. Eliminating poverty, realizing common prosperity of all people, and making all people share the fruits of reform and development are the important goals of building a well-off society in an all-round way. Since the reform and opening, China’s economy has achieved rapid and stable growth, the scale of government finance has been continuously expanded, and the government’s financial expenditure on people’s livelihood has continued to increase, providing a solid material foundation for poverty alleviation and intellectual support in poor areas. According to Poverty Alleviation: China’s Experience and Contribution released by China in 2021, under the guidance of the poverty alleviation strategy, 5.51 million people in rural Areas were living in poverty by the end of 2019, and the poverty incidence rate dropped to 0.6% from 1.7% in the previous year. By the end of 2020, all people in absolute poverty had been lifted out of poverty. In order to realize the transformation from absolute poverty management to relative poverty management, China faces two major challenges. The first is how to consolidate and improve the results of poverty alleviation in rural poor areas, especially in areas of deep poverty and continuous poverty. The second question is how to ensure the sustainable development of poor areas and avoid returning to poverty. However, China still faces severe challenges in its fight in poverty alleviation currently. The traditional standards for measuring poverty ignored the consideration of future risks. The poor are affected by their insufficient development capacity and the poverty environment. When facing a major negative impact of uncertainty in the future, they will have a higher risk to fall into poverty again. All of these will lead the problem of poverty vulnerability and in turn become an obstacle to the sustainable development of poverty alleviation.

In exploring the establishment of a long-term mechanism for the sustainable development of the poor population, we should fully realize that information occlusion will lead to cultural backwardness and narrow cognition, which fundamentally restrict the elimination of poverty. With the advent of the digital era, the Internet helps to break the barrier of blocked information and thus plays an important role in poverty alleviation for rural households. Since the implementation of China’s broadband strategy in 2013, rural Internet construction has made great progress. According to the 47th Statistical Report on China’s Internet Development, by December 2020, the number of rural Internet users in China reached 309 million, and the rural Internet penetration rate was 55.9%. Under the guidance of Internet plus strategic thinking, new forms of Internet plus business are emerging in rural areas, exerting an important influence on traditional agricultural production, life and cognition, and bringing important opportunities for increasing farmers’ income and eliminating poverty. In this context, by studying the impact and mechanism of the Internet on farmers’ vulnerability to poverty, we can not only improve the governance strategy of Internet plus targeted poverty alleviation, but also have important significance to establish a long-term mechanism for farmers’ sustainable development.

This article uses the Chinese family tracking data section in 2018, with expected poverty vulnerability measure method to measure the degree of poverty vulnerability. Furthermore, FGLS (Feasible Generalized Least Square) and Probit estimation models were used to empirically test the relationship between Internet use and farmers’ vulnerability to poverty, and the heterogeneity of the impact was analyzed in detail. On the basis of the above, we try to find ways to reduce the possibility of poverty among farmers in the future.

2. Literature Review and Influence Mechanism

With the advancement of digital rural construction, under the background of Internet plus targeted poverty alleviation, some scholars have conducted preliminary studies on whether Internet use has a poverty reduction effect. Li believes that the use of the Internet plays a significant role in reducing the multidimensional poverty of rural households [1]. The promotion and popularization of rural Internet informatization can effectively solve the capacity poverty of rural households, thus improving the quality of poverty alleviation. Hu and Zhang found that the popularization and use of Internet information technology could significantly increase the non-agricultural income of farmers in poor areas and reduce the incidence of poverty [2,3]. Zuo believes that the impact of the Internet on the relative poverty of low-income groups is significantly higher than that of middle- and high-income groups [4]. In a study of 580 wheat growers in four districts of Khyber Pakhtunkhwa Province in Pakistan, Khan, N. found that the larger the Mobile phone and Internet technology usage, the more substantial and positive the impact on agricultural income [5]. On the contrary, based on the data of 478 rural farmers from two regions in Ghana, Siaw found that Internet use could affect the accumulation of social capital of farmers, but if the Internet is used as a fantasy tool to eliminate poverty, it cannot play a real poverty reduction effect [6]. By investigating the demand of households for information and communication technology (ICT) access, Karakara and Osabuohien emphasizes the different threshold effects of households in Burkina Faso and Ghana in accessing ICT, and usage the features of the households should be taken into consideration when developing ICT access policies [7]. In addition, Abdulqadir and Asongu examined Internet access in 42 sub-Saharan African countries over the period 2008–2018 and found that Internet access has an asymmetric effect on economic growth in sub-Saharan Africa [8]. Jeffrey believes that the Internet cannot play a positive and effective role in poverty reduction due to the differences in the thinking and cognitive patterns of poor farmers [9]. Shao believes that in the early stage of the Development of the Internet, the farmers who first made use of the Internet thinking to start a business can achieve a significant increase in income, but in the later stage, due to the blind investment of farmers and the monopoly of the Internet industry, many farmers do not get dividends from the Internet, so the poverty reduction effect is very small [10]. At the same time, the development of Internet finance has further widened the income gap between urban and rural residents, thus exacerbating poverty [11].

In the current research conclusions, whether the Internet can really play a role in poverty reduction remains to be further discussed. In addition, the poverty indicators in the research literature are measured by post-facto measurement paradigm, that is, the current poverty situation is reflected by income or consumption level, and the dynamic development of poverty in the future cannot be predicted. In fact, even if the current household income or consumption level of poor farmers is higher than the national poverty line, they may return to poverty again in the face of major fluctuations of the household income or consumption in the future [12]. Therefore, it is difficult to effectively link the impact of Internet use on poverty status with the sustainable development of poverty alleviation, nor can it dynamically predict the vulnerability of farmers’ poverty caused by consumption or income fluctuations. Since 1995, when the World Food Program first explained the theoretical connotation and extension of poverty vulnerability, the World Bank also defined the concept of poverty vulnerability for the first time in 2001 and included the probability of an individual falling into poverty again in the poverty research. The concept of vulnerability to poverty has also been interpreted differently in the academic world. Koomson believes that vulnerability to poverty is only one dimension of poverty [13]; Gallardo believes that poverty vulnerability refers to the ability of a family or individual to resist or bear external uncertain risks, as well as the family’s perception of external risks [14]. In terms of specific studies, most of them focus on the impact of national policies and government behaviors on the vulnerability of farmers to poverty, such as government transfer payment [15], integration of agriculture and tourism [16], subsistence allowance system [17], transfer of agricultural land [18], educational fiscal expenditure [19], relocation of poverty alleviation from other places [20]. The results show that the effective implementation of national policies reduces the incidence of poverty vulnerability of farmers. Some scholars also analyzed the impact on vulnerability to poverty from micro-perspectives such as household living pattern, non-agricultural employment, household registration, age difference, household debt and social interaction [21]. These research conclusions basically agreed that the higher the human capital, social capital or economic capital stock of peasant households, the lower the possibility of poverty in the future.

As mentioned above, the issue of poverty vulnerability has received continuous attention from scholars, but the existing research results on the poverty reduction effect of the Internet do not consider the issue of poverty vulnerability. Therefore, by measuring the poverty vulnerability of farmers, this paper puts Internet use and poverty vulnerability into the same analytical framework to test the impact of Internet use on the sustainable development of poverty alleviation of farmers. Through what pathway mechanism does Internet use affect poverty vulnerability? Utilizing a literature review, we summarize the following three theoretical paths.

First, Internet use has an impact on poverty vulnerability by affecting farmers’ agricultural income. In the digital information age, Internet emerging media plays a pivotal role in information interaction between the government and the public, which can timely deliver economic information of government departments, significantly reduce transaction costs and reduce transaction links [22,23]. Through the use of the Internet, farmers can quickly and conveniently obtain relevant agricultural information, which has a positive impact on agricultural technology to improve the sales of agricultural products and agricultural disaster prevention, thus reducing the cost of information acquisition and promoting the increase in income [24]. Internet information technology has promoted the large-scale digitalization and automation of agricultural production through the technological innovation of traditional agricultural production methods, improved production efficiency and realized the increase in farmers’ income. Internet use also helps to improve the decision-making efficiency of farmers [25]. In the rural Internet information construction is not perfect, the cost of farmers to obtain information is high, which increases the cost of agricultural production. On the contrary, the use of the Internet helps to establish effective channels of information exchange, thus maximizing farmers’ income. Therefore, the change of agricultural production thinking and mode, can help increase in income and improve the ability of farmers to resist the return to poverty.

Second, Internet use affects vulnerability to poverty by affecting non-agricultural employment. Pabilonia believes that Internet use promotes the improvement of human capital [26] and can promote the transfer of rural surplus labor force, thus obtaining more employment opportunities, which is an important reason for the increase in income level. On the one hand, currently there is a large amount of surplus labor in rural areas, and the allocation of labor resources is inefficient. To realize the effective flow of rural labor and obtain wage income through non-agricultural employment is an important way to increase farmers’ income [27]. Through the use of the Internet, farmers can timely and accurately collect a large number of non-agricultural employment information, promoting the effective allocation of rural surplus labor resources. Ma Junlong’s research shows that Internet use by rural residents can increase employment rate and non-agricultural employment income, and this impact effect is stronger in remote and poor areas [28]. Compared with rural residents who do not use the Internet, rural residents who are skilled in using the Internet have higher employment rates and less time out of work. On the other hand, Internet use can promote farmers to start their own businesses and improve their profits. The information-based poverty alleviation strategy based on Internet technology has promoted the development of e-commerce in rural areas, effectively improved the enthusiasm of rural households to start their own businesses, which can increase the current income of farmers and reduce the possibility of falling into poverty in the future. In the process of entrepreneurship, farmers can also effectively expand the family social capital network and promote information sharing, thus reducing their chances of falling into poverty in the future. In addition, the development of Internet finance can break the restrictions of regional conditions, realize the precise placement of funds, effectively solve the financing problems of farmers in entrepreneurship, improve the risk resistance ability of entrepreneurial projects, and reduce the possibility of falling into poverty due to entrepreneurial failure [29,30].

Third, Internet use affects farmers’ vulnerability to poverty by influencing income gap. Bound’s technology bias theory believes that highly skilled workers can quickly adapt to new Internet technologies, and Internet use has a greater impact on the wages of highly skilled employees than ordinary employees, thus resulting in the income inequality of workers [31]. The imbalanced allocation of Internet resources in different regions further widens the digital divide between the rich and the poor, and further aggravates the income gap. The information dividend brought by the Internet can only be enjoyed by the rich group with a higher socio-economic status, while the poor class with a lower socio-economic status is excluded, thus further widening the gap between the rich and the poor in society. Xu Zhuqing believes that the emergence of secondary digital divide causes the difference in information identification, utilization, and appreciation ability, which further widens the income gap among Internet users [32,33]. Using the data of Spanish counties, Whitacre found that the impact of Internet use on the income of the middle class was significantly greater than that of the lower class [34]. On the contrary, Song X believes that Internet information technology can significantly improve the level of regional economic development and play an effective buffer role in the imbalance between urban and rural economic development. The effective integration of Internet information technology and inclusive finance provides more development opportunities for low-income groups in rural areas and plays a positive role in alleviating the income gap between urban and rural residents [35]. The income gap further increases the unbalanced allocation of education and medical resources between regions, hinders the improvement of farmers’ livelihood capital, and then leads to poverty of farmers and raises the possibility of poverty in the future.

3. Research Design

3.1. Data Sources

The research data in this paper are from the Data of China Family Tracking Survey (CFPS) in 2018. These data are a large-scale, continuous, and interdisciplinary data survey project organized by the China Center for Social Sciences Surveys at Peking University. It aims to comprehensively reflect the dynamic changes in China’s economic, social, and demographic education and health by tracking and collecting data at the household and community levels. The China Family Tracking Survey has completed five nationwide data tracking surveys since 2008. In this paper, the cross-sectional survey data of 2018 were selected. According to the research needs, the number of valid samples was 7477 after variable screening and processing of missing values and outliers.

3.2. Variable Definition and Statistical Analysis

3.2.1. Dependent Variable: Vulnerability of Farmers to Poverty

The dependent variable of this paper is the vulnerability of farmers to poverty, that is, the possibility of farmers falling into poverty at a given time in the future. Gaiha and Imai (2008) further summarized three methods to measure vulnerability to poverty, namely, expected vulnerability measurement (VEP) vulnerability measurement of low expected utility (VEU) and risk exposure vulnerability measurement (VER). Both expected vulnerability and low expected utility vulnerability are ex ante measures, which can measure the incidence of future poverty of individuals or families. Risk exposure vulnerability is an ex-post measurement method, which is mainly used to assess the ability of individuals or families to resist unknown major risks. According to Chaudhuri’s theoretical interpretation of expected poverty vulnerability, the incidence of individual or family poverty vulnerability is mainly related to the family’s future economic welfare level, and family characteristic variables determine the expected welfare value and welfare fluctuation level, so this measurement method has good applicability to cross-section data [36]. In view of this, this paper adopts the expected poverty vulnerability method (VEP) to measure the poverty vulnerability of farmers and calculates the probability value of poverty occurrence in t+1 period through the family characteristics of farmer i in t period. According to the definition of poverty vulnerability, the poverty vulnerability of farmer i at time t can be expressed as:

$${\mathrm{Vep}}_{\mathrm{i},\mathrm{t}}=\mathrm{Prob}\left(\ln {\mathrm{con}}_{\mathrm{i},\mathrm{t}+1}<\ln \mathrm{poor}\right)$$

(1)

In Equation (1), Vep_i,t represents the poverty vulnerability of the ith farmer household in period t, ln con _i,t+1 represents the consumption level of the ith farmer household in period t + 1, and in poor represents the poverty standard. The specific calculation method is as follows:

Firstly, the per capita consumption equation was constructed to regression the logarithm of per capita consumption expenditure using Ordinary Least Squares (OLS). Then, OLS was used to square the residual of Equation (2) and then regression was performed to obtain Equation (3). According to the VEP theory, the consumption fluctuation term e_i is determined by household characteristic variables and follows the influence relationship of Equation (3).

$$\ln {\mathrm{con}}_{\mathrm{i},\mathrm{t}} ={\mathsf{\beta }}_0{\mathrm{X}}_{\mathrm{i},\mathrm{t}}+{\mathrm{e}}_{\mathrm{i}}$$

(2)

$${\mathrm{e}}_{\mathrm{i}}^2= {\mathsf{\theta }}_0{\mathrm{X}}_{\mathrm{i},\mathrm{t}}+{\mathsf{\epsilon }}_{\mathrm{i}}$$

(3)

In Equation (2), ln con_i,t represents the logarithm of the current per capita consumption expenditure of the ith peasant household. X_i,t represents a series of characteristic variables of peasant households that affect consumer expenditure, including gender, age, education level, marital status, health level, starting a business, government subsidies, family events, non-agricultural employment, social insurance, household per capita income, household borrowing, regional fixed effect, etc.

Secondly, in order to reduce the regression bias caused by heteroscedasticity, weighted regression was performed on Equations (2) and (3) by using FGLS (three-stage generalized least square method) to obtain the expected value E of logarithmic per capita consumption of farmers in the future, i.e.,

$$\mathrm{E}=\left[\ln {\mathrm{con}}_{\mathrm{i},\mathrm{t}}|{\mathrm{X}}_{\mathrm{i}}\right]={\mathrm{X}}_{\mathrm{i}}{\mathsf{\beta }}_{0\mathrm{FGLS}}$$

(4)

$$\mathrm{V}=\left[\ln {\mathrm{con}}_{\mathrm{i},\mathrm{t}}|{\mathrm{X}}_{\mathrm{i}}\right]={\mathrm{e}}_{\mathrm{i}}^2={\mathrm{X}}_{\mathrm{i}}{\mathsf{\theta }}_{0\mathrm{FGLS}}.$$

(5)

Finally, assuming that the logarithmic expectation of per capita consumption expenditure follows the positive distribution, when the logarithmic expectation of per capita consumption expenditure is X_iβ_0FGLS, and the logarithmic variance of per capita consumption expenditure is X_iθ_0FGLS, the poverty vulnerability of ith peasant household in period t can be expressed as:

$${\mathrm{Vep}}_{\mathrm{i},\mathrm{t}} =\mathrm{Prob}\left(\ln {\mathrm{con}}_{\mathrm{i},\mathrm{t}+1}<\ln \mathrm{poor}\right) =\left(\frac{\ln \mathrm{poor}-{\mathrm{X}}_{\mathrm{i}}{\mathsf{\beta }}_{0\mathrm{FGLS}}}{\sqrt{{\mathrm{X}}_{\mathrm{i}}{\mathsf{\theta }}_{0\mathrm{FGLS}}}}\right)$$

(6)

Based on different research purposes, scholars also choose different poverty vulnerability thresholds. Novignon and Mussa set the thresholds of poverty vulnerability as 0.5 and 0.75, that is, a peasant will have a 50% and 75% probability of falling into poverty in t+1 period [37]. Ward believes that if the sum of the probability of farmers falling into poverty in at least one period in the next two periods is 0.5, the single-period poverty vulnerability threshold should be set at 0.29 [38]. Based on the latest poverty standards of $1.9 and $3.1 per capita set by the World Bank, 0.29 was selected as the poverty vulnerability threshold, and the incidence of poverty in the future greater than 0.29 was defined as poor and vulnerable households, with a value of 1. Otherwise, non-poor vulnerable peasant households are assigned a value of 0.

Table 1. Distribution of poverty vulnerability of rural households in China and subregions.

Poverty Line	Sample Size	Per Capita $1.9 Costs/Day			Per Capita $3.1 Costs/Day
Subregions		Poverty Incidence	Mean of Vulnerability	Vulnerable Peasant Households Ratio	Poverty Incidence	Mean of Vulnerability	Vulnerable Peasant Households Ratio
Nationwide	7477	0.180	0.078	9.48%	0.395	0.339	33.88%
East	2823	0.128	0.049	5.79%	0.318	0.248	24.87%
Middle	2087	0.169	0.071	8.67%	0.391	0.319	31.96%
West	2567	0.234	0.117	14.07%	0.466	0.453	45.34%

reports the results of poverty vulnerability of Chinese peasant households. According to the international poverty line of US $1.9, the average vulnerability of the whole sample was 0.078, which account for 9.48%, and vulnerable farmers in the West, Middle and East of China accounted for 14.07%, 8.67% and 5.79%, respectively. If the national poverty standard of $3.1 is adopted, the mean value of vulnerability in the whole sample is 0.339, which account for 33.88%, and vulnerable farmers account for 45.34% 31.96% and 24.87%, respectively.

3.2.2. Independent Variable: Internet Use

The core independent variable of this paper is Internet use, that is, whether farmers use the Internet. If the interviewed farmers answered that they use computers to browse the Internet, they were defined as using the Internet and assigned a value of 1. Otherwise, if the Internet is not used, the value is 0. In the whole sample, farmers who use the Internet account for 15.34%, and those who do not use the Internet account for 84.66% Under the poverty standard of $1.9 and $3.1, the proportion of households that use the Internet is 3.27% and 15.06%, respectively, and the proportion of households that do not use the Internet is 11.11% and 38.37%.

3.2.3. Control Variable

In this paper, variables at three levels of household head’s individual characteristics, family characteristics and regional characteristics were selected as the control variables. Individual characteristic variables included age, sex, education level, marital status, and health level of head of household. Family characteristic variables include household per capita income, household major events, non-farm employment, social insurance, human connection. See

Table 2. Descriptive statistical analysis of variables.

Variables	Variable Definition and Description	Mean Value	Standard Deviation
Household consumption Expenditure per capita	Logarithm of per capita consumption expenditure of interviewed farmers in 2015	9.177	0.854
Household incomes per capita	Logarithm of per capita income of rural households interviewed in 2015	9.131	0.961
Internet using	Using Internet = 1, not Using Internet = 0	0.153	0.360
The householder age	Continuous variable	45.867	17.639
The household head gender	Male head of household = 1, female head of household = 0	0.506	0.500
Degree of Education	Illiterate = 0, elementary = 1, junior high = 2, high school and above = 3	1.147	1.041
The marriage status of the household head	Married = 1, unmarried = 0	0.759	0.428
Health Level	Very unhealthy = 1, relatively unhealthy = 2, fair = 3, relatively healthy = 4, very healthy = 5	3.047	1.281
Family Events	Major event = 1, no major event = 0	0.166	0.372
Non-agricultural Employment	Non-farm work = 1, non-farm work = 0	0.337	0.473
Social Insurance	Participation in social insurance = 1, Otherwise = 0	0.574	0.494
Human Interaction	Logarithm of expenditure of gift and money	7.745	1.081

for the definition and description of each variable.

3.3. Econometric Model Construction

In order to comprehensively investigate the impact of Internet use on farmers’ vulnerability to poverty, this paper constructed the following Probit Model.

$$\mathrm{pr}\left({\mathrm{Vep}}_{\mathrm{i}}=1\right)=\mathsf{\Phi }\left({\mathsf{\alpha }\mathrm{Internet}}_{\mathrm{i}}+{\mathsf{\beta }\mathrm{X}}_{\mathrm{i}}+{\mathsf{\mu }}_{\mathrm{i}}\right)$$

(7)

where, Vep_i represents the poverty vulnerability of farmers, Vep_i = 0 represents non-vulnerable farmers, Vep_i = 1 represents vulnerable farmers, Internet_i represents household Internet use, X_i is a series of individual household and regional characteristic variables affecting household poverty vulnerability, α β is the parameter to be estimated, β vector form, μ_i stands for the random disturbance term.

4. Empirical Analysis and Testing

4.1. The Impact of Internet Use on Farmers’ Vulnerability to Poverty and Regional Differences

Table 3. Baseline regression results of the impact of Internet use on farmers’ vulnerability to poverty.

Variable	$1.90 Poverty Line				$3.10 Poverty Line
	Whole Sample	West	Middle Region	East Region	Whole Sample	West Region	Middle Region	East Region
Internet Use	−0.338 ***	−0.514 **	−0.190	−0.424 **	−0.392 ***	−0.708 ***	−0.159	−0.391 ***
	(0.109)	(0.216)	(0.182)	(0.184)	(0.0619)	(0.124)	(0.106)	(0.100)
Age	−0.0317 ***	−0.00531	−0.0645 **	−0.0412 *	−0.0287 ***	−0.0135	−0.0281	−0.0430 ***
	(0.0123)	(0.0194)	(0.0252)	(0.0220)	(0.00934)	(0.0150)	(0.0189)	(0.0158)
Age Square	0.0003 ***	0.00005	0.0008 ***	0.0004	0.0003 ***	0.0001	0.0004 *	0.0004 **
	(0.0001)	(0.0002)	(0.0003)	(0.0002)	(0.0001)	(0.0002)	(0.0002)	(0.0002)
Sex	0.0402	0.0868	−0.0128	0.0180	0.0187	0.0233	−0.00958	0.0338
	(0.0549)	(0.0860)	(0.105)	(0.101)	(0.0391)	(0.0644)	(0.0733)	(0.0677)
Education Level	−0.0691 **	−0.146 ***	0.0398	−0.0550	−0.110 ***	−0.129 ***	−0.0592	−0.117 ***
	(0.0295)	(0.0453)	(0.0582)	(0.0561)	(0.0207)	(0.0333)	(0.0397)	(0.0365)
Marriage Status	0.455 ***	0.508 ***	0.572 ***	0.295 *	0.520 ***	0.464 ***	0.524 ***	0.548 ***
	(0.0915)	(0.135)	(0.208)	(0.163)	(0.0628)	(0.0957)	(0.132)	(0.110)
Health Level	−0.0558 **	−0.0627 *	0.00149	−0.0762 *	−0.101 ***	−0.0693 **	−0.108 ***	−0.123 ***
	(0.0225)	(0.0355)	(0.0422)	(0.0421)	(0.0164)	(0.0270)	(0.0305)	(0.0284)
Household Incomes per capita	−0.121 ***	−0.0837 **	−0.145 ***	−0.157 ***	−0.0918 ***	−0.0826 ***	−0.0983 ***	−0.111 ***
	(0.0264)	(0.0395)	(0.0542)	(0.0493)	(0.0191)	(0.0306)	(0.0372)	(0.0330)
Family Events	−0.857 ***	−1.003 ***	−0.615 ***	−0.950 ***	−0.768 ***	−0.865 ***	−0.742 ***	−0.674 ***
	(0.102)	(0.156)	(0.173)	(0.231)	(0.0565)	(0.0884)	(0.100)	(0.109)
Non-agricultural Employment	−0.686 ***	−0.590 ***	−0.701 ***	−0.828 ***	−0.584 ***	−0.605 ***	−0.570 ***	−0.570 ***
	(0.0691)	(0.110)	(0.133)	(0.126)	(0.0442)	(0.0738)	(0.0817)	(0.0761)
Social Insurance	−0.205 ***	−0.305 ***	−0.181	−0.108	−0.176 ***	−0.256 ***	−0.173 **	−0.107
	(0.0575)	(0.0908)	(0.115)	(0.102)	(0.0414)	(0.0703)	(0.0796)	(0.0687)
Human Interaction	−0.847 ***	−1.004 ***	−0.909 ***	−0.667 ***	−0.940 ***	−0.934 ***	−0.982 ***	−0.945 ***
	(0.0326)	(0.0545)	(0.0689)	(0.0527)	(0.0244)	(0.0407)	(0.0491)	(0.0403)
Constant Term	6.314 ***	6.852 ***	6.761 ***	5.148 ***	8.422 ***	8.195 ***	8.268 ***	8.224 ***
	(0.350)	(0.552)	(0.705)	(0.618)	(0.267)	(0.436)	(0.512)	(0.458)
Sample Size	7018	2398	1955	2665	7018	2398	1955	2665
Pseudo R2	0.347	0.370	0.341	0.305	0.351	0.329	0.323	0.360

Note: The brackets are standard deviations, and ***, **, * in turn means significant at the statistical level of 1%, 5%, and 10%.

reports the regression parameters of Probit Model of the impact of Internet use on farmers’ vulnerability to poverty. The whole samples regression results of Equations (1) and (5) show that no matter by $1.9 or $3.1 poverty line, Internet use are significantly 1% negative impact on the poverty vulnerability, showing that using the Internet can reduce the incidence of poverty vulnerability, the Internet can promote the sustainable development of future farmers out of poverty. The marginal effects of Equations (1) and (5) are −0.0377 and −0.0921, respectively, indicating that at the standard of $1.9 and $3.1, using the Internet can reduce the incidence of poverty vulnerability of farmers by 3.77% and 9.21%, respectively, which means that with the increase in poverty standards, Internet use has a greater impact on poverty vulnerability of farmers. At the same time, there are significant regional differences in the impact of Internet use on farmers’ vulnerability to poverty. The regression results of Equations (2)–(4) and (6)–(8) shows that, with the setting of $1.9 and $3.1 poverty line, the Internet use is of significant negative impact to poverty vulnerability in the east and west. The result shows that in the western and eastern regions, the Internet use can significantly promote the sustainable development of rural households in the future poverty alleviation. By comparing the marginal effect, it is found that the impact of Internet use on farmers’ vulnerability to poverty in western China is significantly greater than that in eastern China. However, the impact of Internet use on poverty and vulnerability of rural households in central China is not significant, and the promotion and popularization of Internet in rural areas in central China does not play a positive risk prevention and control role for farmers falling into poverty in the future. Whether this result is related to the number of samples (compared with the eastern and western, the central sample is the least) or affected by other unknown factors remains to be further explored.

In terms of the influence effect of control variables, under the poverty standard of $1.9 and in terms of individual characteristic variables of farmers, the impact of age on poverty vulnerability of farmers presents an inverted U-shape, that is, with the increase in the age of farmers households’ heads, the probability of poverty vulnerability of farmers households increases first and then decreases. The higher the level of education of household heads, the lower the incidence of poverty in the future. This is because the higher the level of education, the better the employment options and stable wage income, thus increasing the ability of households to withstand future risks, thus reducing the likelihood of poverty in the future. Marriage has a positive impact on poverty vulnerability of peasant households, indicating that married peasant households are more likely to fall into poverty in the future than unmarried peasant households. Health status has a negative impact on the vulnerability of farmers to poverty, indicating that the higher the health level, the lower the likelihood of falling into poverty in the future. Secondly, in terms of the characteristic variables of households, the per capita income of households with non-agricultural employment, social insurance and human relations have a significant negative impact on the vulnerability of households to poverty, indicating that the increase in household income, households with non-agricultural employment, and households with non-agricultural employment, have major events, participate in social insurance and promotion of human contact can reduce the probability of farmers falling into poverty in the future and promote the sustainable development of farmers getting rid of poverty in the future. In the case of using the poverty standard of $3.1, the impact of household characteristics and household characteristics variables on poverty vulnerability is the same.

4.2. Endogeneity Problem Handling Based on Instrumental Variable Method

Endogeneity problems are caused by data measurement errors and omission of variables and mutual causality between independent variables and dependent variables in empirical analysis models, thus leading to deviation of model regression results [39]. Therefore, this paper uses instrumental variable method to deal with the potential endogeneity of Internet use and selects farmer information source channels as instrumental variable. If farmer obtains information from the Internet, the value is assigned to 1; otherwise, the value is assigned to 0. On the one hand, the vulnerability of farmers to poverty is a prediction of the future poverty of the family, and the information source channel is the main way to reflect the current information source of farmers. Through the analysis of variables at two different time nodes, the influence of reverse causality can be effectively avoided. On the other hand, the information source channels are the heterogeneity characteristics of individual farmers, which is not directly related to the risk of the family falling into poverty in the future, so it satisfies the externality.

Table 4. Regression results of instrumental variable method for the impact of Internet use on farmers’ vulnerability to poverty.

Variable	$1.90 Poverty Line				$3.10 Poverty Line
	Whole Sample	West Region	Middle Region	East Region	Whole Sample	West Region	Middle Region	East Region
Internet Use	−1.065 ***	−1.287 **	−0.247	−1.169 **	−1.537 ***	−2.165 ***	−0.819	−1.262 ***
	(0.307)	(0.647)	(0.637)	(0.514)	(0.179)	(0.366)	(0.393)	(0.306)
Other Variables	control	control	control	control	control	control	control	control
Constant Term	6.285 ***	6.968 ***	6.803 ***	5.532 ***	8.173 ***	7.971 ***	8.523 ***	8.439 ***
	(0.336)	(0.543)	(0.830)	(0.612)	(0.267)	(0.488)	(0.502)	(0.448)
F value in the first phase	269.64	80.06	77.76	108.09	269.64	80.06	77.76	108.09
p value in Wald inspection	0.029	0.033	0.092	0.015	0.000	0.000	0.010	0.008
t value of instrumental variable	21.44	11.3	11.01	13.52	21.44	11.03	11.01	13.52
Sample Size	7018	2398	1955	2665	7018	2398	1955	2665

Note: The brackets are standard deviations, and ***, ** in turn means significant at the statistical level of 1%, 5%, and 10%.

reports the results of regression using IV Probit. The regression results of Equations (1) and (5) show that under the poverty standard of $1.9 and $3.1, Internet use has a significant negative impact on poverty vulnerability of farmers, that is, Internet use can significantly reduce the occurrence probability of poverty vulnerability of farmers. In the regression results by region, Internet use can significantly reduce the probability of the occurrence of poverty vulnerability of rural households in the western and eastern regions, which is completely consistent with the Probit regression results above, indicating that the research conclusions of this paper have certain robustness.

4.3. Robustness Test Based on Propensity Score Matching (PSM)

In order to further verify the robustness of the above empirical conclusions, this paper chose the Propensity Score Matching (PSM) method for robustness test. The core of the propensity score matching method is to make the observation data as close as possible to the random trial data through the matching re-sampling method, so as to realize the robust inference of the causal relationship between the core independent variable and the dependent variable. According to the design idea of PSM method, it is assumed that Vep_i is the result variable of poverty vulnerability of farmers, and farmers using the Internet are set as the treatment group, ${\mathrm{Vep}}_{\mathrm{i}}^1$ is the result of poverty vulnerability of farmers using the Internet, and farmers not using the Internet are set as the control group, ${\mathrm{Vep}}_{\mathrm{i}}^0$ is the result of poverty vulnerability of farmers who do not use the Internet. The average processing effect of Internet use on farmers’ vulnerability to poverty (ATT) is estimated by using K-nearest Neighbor matching radius matching kernel matching local linear regression matching and Mahalanakis matching methods. The average treatment effect difference (ATT) in

Table 5. ATT difference results of PSM method robustness test.

Variable	$1.90 Poverty Line				$3.10 Poverty Line
	Sample	ATT Difference	Standard Deviation	t Value	Sample	ATT Difference	Standard Deviation	t Value
The Neighbor Mat (k = 4)	before matching	−0.086	0.009	−9.66	before matching	−0.233	0.014	−16.28
	after matching	−0.224	0.009	−2.45	after matching	−0.086	0.019	−4.57
Radius Matching	before matching	−0.086	0.009	−9.66	before matching	−0.233	0.014	−16.28
	after matching	−0.250	0.010	−2.47	after matching	−0.087	0.018	−4.83
Kernel Matching	before matching	−0.086	0.009	−9.66	before matching	−0.233	0.014	−16.28
	after matching	−0.026	0.010	−2.70	after matching	−0.090	0.017	−5.22
Local Linear Regression Matching	before matching	−0.086	0.009	−9.66	before matching	−0.233	0.014	−16.28
	after matching	−0.256	0.111	−2.30	after matching	−0.090	0.023	−3.85
Markov Match	before matching	−0.086	0.009	−9.66	before matching	−0.233	0.014	−16.28
	after matching	−0.177	0.006	−2.82	after matching	−0.098	0.015	−6.73

shows, whether prior to or after the match, Internet use negatively affected farmers’ vulnerability to poverty at the significance level of 1%. This shows that under the poverty standard of $1.9 and $3.1, using the Internet can indeed reduce the probability of farmers falling into poverty risk in the future and it is of a positive impact on promoting the sustainable development of farmers out of poverty. The regression results of propensity score matching method were completely consistent with the previous Probit and IV Probit regression results, which further confirmed the robustness of the research conclusions in this paper.

4.4. Heterogeneity Analysis of the Impact of Internet Use on Farmers’ Vulnerability to Poverty

The above empirical analysis results show that Internet use plays a positive role in promoting the sustainable development of farmers’ poverty alleviation. Considering the difference in the impact of gender income level and type of Internet use on the empirical results, this paper further analyzed the heterogeneity of the impact of Internet use on poverty vulnerability of farmers.

Table 6. Results of heterogeneity analysis.

$1.9 Poverty Standard	Gender Difference		Differences in Income Levels			Different Types of Internet Use
	Female	Male	Low Income	Middle Income	High Income	Internet Work	Social Activities on the Internet	Internet Commerce	Time Spent on the Internet
Vulnerability of Peasant Household Poverty	−0.352 ***	−0.475 ***	−0.583 ***	−0.386 ***	−0.571 *	−0.0423 **	−0.0684 ***	−0.135 ***	−0.0105 ***
	(0.109)	(0.103)	(0.125)	(0.136)	(0.338)	(0.0175)	(0.0186)	(0.0263)	(0.00379)
Other Variables	Control	Control	Control	Control	Control	Control	Control	Control	Control
Constant Term	−0.123	−0.439 *	−0.573 **	−0.136	−0.366	−1.359 ***	−1.024 **	−1.062 ***	−1.254 ***
	(0.241)	(0.247)	(0.254)	(0.372)	(0.786)	(0.404)	(0.415)	(0.411)	(0.408)
Sample Size	3715	3882	3345	3007	666	3405	3405	3405	3394
Pseudo R2	0.050	0.069	0.054	0.078	0.117	0.037	0.042	0.053	0.038
$3.1 Poverty Standard	Gender Difference		Differences in Income Levels			Different Types of Internet Use
	Female	Male	Low Income	Middle Income	High Income	Internet Work	Social Activities on the Internet	Internet Commerce	Time Spent on the Internet
Vulnerability of Peasant Household Poverty	−0.514 ***	−0.540 ***	−0.465 ***	−0.517 ***	−0.989 ***	−0.064 ***	−0.047 ***	−0.187 ***	−0.006 **
	(0.081)	(0.068)	(0.076	(0.077)	(0.217)	(0.012)	(0.014)	(0.018)	(0.0023)
Other Variables	Control	Control	Control	Control	Control	Control	Control	Control	Control
Constant Term	1.015 ***	0.639 ***	0.545 ***	1.045 ***	0.461	−0.0994	0.0504	0.389	−0.107
	(0.221)	(0.217)	(0.200)	(0.268)	(0.561)	(0.324)	(0.331)	(0.333)	(0.328)
Sample Size	3354	3664	3345	3007	666	2995	2995	2995	2988
Pseudo R2	0.061	0.077	0.060	0.064	0.127	0.065	0.059	0.091	0.057

Note: The brackets are standard deviations, and ***, **, * in turn means significant at the statistical level of 1%, 5%, and 10%.

reports the heterogeneity of the impact of Internet use on peasant household vulnerability to poverty. Gender disaggregated regression shows that Internet use has a significantly smaller impact on the vulnerability of female heads of peasant household to poverty than male heads of peasant household under the poverty standard of $1.9. It is concluded by calculating the marginal effect, Internet use reduced the risk of future poverty for female and male heads of peasant household by 5.91% and 7.17%, respectively. The difference held true at the $3.10 poverty level. Internet use reduced the risk of future poverty for female and male heads of peasant households by 17.93% and 17.97%, respectively. Therefore, in the process of Internet popularization and promotion in rural areas, more attention should be paid to rural women’s Internet use, the scope and skills of rural women’s Internet use should be improved, and the Internet thinking should be actively cultivated to reduce the risk of female farmers falling into poverty in the future with the support of Internet information technology.

The regression results of income level group show that under the poverty standard of $1.9, the impact of Internet use on poverty vulnerability of low-income farmers is significantly greater than that of high- and middle-income farmers. Internet use reduced the poverty vulnerability probability of high-, middle- and low-income peasant household by 7.64%, 4.91% and 10.69%, respectively. At the same time, under the poverty standard of $3.1, the impact of Internet use on the vulnerability of high-income households to poverty is significantly greater than that of low-income and middle-income households. Internet use reduced the poverty vulnerability probability of high-, middle- and low-income peasant household by 30.77%, 16.79% and 16.64%, respectively. Under different poverty criteria, there are significant differences in the impact of Internet use on poverty vulnerability of farmers with different income levels. Therefore, it is necessary to strengthen the promotion of the Internet for middle- and low-income farmers, organize Internet skills training, improve the Internet plus thinking of middle and low income farmers, and make the Internet become a starting point for farmers to increase production and income.

The grouped regression results of the types of Internet use showed that, under the poverty criteria of $1.9 and $3.1, peasant households engaged in Internet work, engaged in social activities on the Internet, engaged in Internet commerce and increased time spent on the Internet all had a significant negative impact on peasant households’ vulnerability to poverty. Therefore, the active development of rural Internet office, the promotion of Internet social function in rural areas, the active expansion of rural Internet business activities, and the scientific and reasonable arrangement of farmers’ online time play an important role in reducing the risk of poverty in the future and improving the sustainable development of farmers out of poverty.

5. The Mechanism of the Impact of Internet Use on Farmers; Vulnerability to Poverty

Rural Internet construction has played a positive role in promoting rural revitalization and poverty alleviation. How then does Internet use have an impact on farmers, realization of poverty alleviation, and sustainable development? This paper proposes the following three impact paths: First, Internet use can affect household income of farmers, improve their living standards and quality of life, and enhance the ability of farmers to withstand major unknown risks, to reduce the possibility of falling into poverty in the future and achieve sustainable development of poverty alleviation. Second, Internet use can affect the way and channel of farmers’ information acquisition, improve farmers’ information acquisition ability, reduce information asymmetry, and then reduce the incidence of poverty vulnerability [40]. Third, Internet use can affect farmers’ off-farm employment, promote farmers’ entrepreneurship and Internet commerce, improve farmers’ risk awareness, broaden employment paths, and thus reduce the risk of falling into poverty in the future. Therefore, this paper adds the interaction terms of household income, information access and off-farm employment to further verify whether Internet use affects household poverty vulnerability through the above paths.

5.1. Examining the Mechanism of Vulnerability to Income Poverty of Rural Households Using Internet

Income is one of the important factors affecting poverty. To improve the income level of farmers, enhance their ability to generate income, combine poverty alleviation with intellectual support, and realize the sustained and stable growth of farmers’ income is an important way to change farmers’ production thinking and enrich the diversity of income in the new era. In the context of the introduction of Internet into villages, the promotion of Internet plus thinking can significantly improve the income level of middle-aged and elderly farmers with low education levels and narrow the income gap among peasant households. Zhou’s research shows that the use of the Internet can increase the added value of agricultural products, reduce information barriers, and realize the smooth production and sales channels of agricultural and sideline products, so as to increase farmers’ income [41]. Based on this, this paper adds the interaction item between Internet use and peasant household income, and the regression results are shown in

Table 7. Interaction mechanism analysis.

Variables	$1.90 Poverty Line			$3.10 Poverty Line
	(1)	(2)	(3)	(4)	(5)	(6)
Internet Use	−3.067 **	−0.835 ***	−0.436 ***	−0.195 ***	−0.268 ***	−0.637 ***
	(1.346)	(0.277)	(0.101)	(0.0669)	(0.0532)	(0.0880)
Peasant Household Income	−1.917 ***			−1.759 ***
	(0.0603)			(0.0446)
Internet Use × Peasant Household Income	−0.330 **			−0.051 **
	(0.157)			(0.022)
Information Access		−0.0861 ***			−0.0976 ***
		(0.0189)			(0.0141)
Internet Use × Information Access		−0.122 *			−0.0849 **
		(0.0668)			(0.0425)
Non-agricultural Employment			−0.528 ***			−0.483 ***
			(0.0578)			(0.0403)
Internet Use × Non-agricultural Employment			−0.201 ***			−0.252 **
			(0.0257)			(0.104)
Other Variables	Control	Control	Control	Control	Control	Control
Constant Term	15.40 ***	0.141	−0.172	16.07 ***	1.312 ***	1.106 ***
	(0.510)	(0.192)	(0.173)	(0.441)	(0.169)	(0.157)
Sample Size	8089	7597	7597	7018	7018	7018
Pseudo R2	0.548	0.063	0.079	0.434	0.076	0.086

Note: The brackets are standard deviations, and ***, **, * in turn means significant at the statistical level of 1%, 5%, and 10%.

. Regression results of Equation (1) in Table 7 show that, under the poverty standard of $1.9, the interaction term between Internet use and farmers’ income is negative at the significance level of 5%, indicating that Internet use can reduce farmers’ chances of falling into poverty in the future by influencing farmers’ income path. Meanwhile, the regression results of Equation (4) show that under the poverty standard of $3.1, the interaction term is also significantly negative at the statistical level of 5%, which further proves the establishment of the influence effect of this mechanism.

5.2. An Empirical Study on the Interaction Mechanism among Internet Use, Information Acquisition and Peasant Household Vulnerability to Poverty

In the Era of Internet plus, it is of great significance to obtain convenient information, reduce information asymmetry between farmers and government, farmers and other market subjects, and improve the ability of accurate information identification to achieve sustainable poverty alleviation for farmers. The Internet as the information link breaks the traditional way of communication between farmers and the outside world. The openness of communication in the information age helps farmers timely obtain the policy interpretation of rural revitalization and understand the connotation of the era of rural revitalization. At the same time, the Internet information exchange model can promote the accumulation of farmers’ social capital and play a good role in promoting the expansion of communication channels, improving the ability to obtain information, and achieving lasting poverty alleviation. Regression results of Equations (2) and (5) in Table 7 show that under the $1.9 and $3.1 poverty standards, the interaction between Internet using and the access to information under the significance level of 10% and 5%, respectively, is negative, the Internet use can reduce the vulnerability of peasant household poverty and realize the sustainable development of future farmers out of poverty by influencing the paths of farmers’ access to information.

5.3. A Mechanistic Test of Internet Use, Non-Agricultural Employment and the Vulnerability of Rural Households to Poverty

The incidence of vulnerability to poverty is higher when the income of the primary industry producers is relatively lower than that of those working in the secondary and tertiary industries. With the promotion and use of the Internet in rural areas, farmers can timely understand non-agricultural employment information through the Internet, promote the transfer of rural surplus labor force, improve the non-agricultural employment rate of rural residents, and then increase the income of farmers [42]. At the same time, the Internet can also promote the occurrence of entrepreneurial behavior of farmers, promote the transformation of traditional production and operation mode, improve the efficiency of non-agricultural employment, and create a good material foundation for the realization of lasting poverty alleviation. Regression results of Equations (3) and (6) in Table 7 show at the poverty level of $1.9 and $3.1, the interaction between Internet use and off-farm employment is negative at the significance level of 1% and 5%, respectively, indicating that Internet use can reduce the incidence of poverty vulnerability of farmers by affecting off-farm employment.

6. Conclusions

Improving the ability of farmers to deal with major risks in the future and realizing sustainable development of poverty alleviation are the core connotation of consolidating the achievements of targeted poverty alleviation and realizing comprehensive and lasting poverty alleviation [43]. The popularization of Internet in rural areas provides a new solution for sustainable development of poverty alleviation for farmers. Based on the cross-sectional data of China Household Tracking Survey in 2018, this paper adopts the international poverty standard line of $1.90 and $3.100 to calculate the occurrence probability of poverty vulnerability of farmers through FGLS method, and empirically tests the relationship between Internet use and poverty vulnerability of farmers. The findings from the study are as follows: First, Internet use significantly reduces the incidence of poverty vulnerability of farmers, and Internet use can promote poverty alleviation and sustainable development of farmers in the future, especially when the poverty line of $3.1 is adopted, Internet use has a more significant effect on reducing poverty vulnerability of farmers. Second, there is significant heterogeneity in the impact of Internet use on peasant households’ vulnerability to poverty. There are significant differences in the impact of different gender, income level and types of Internet use on peasant households’ vulnerability to poverty. Third, the path mechanism shows that improving farmers’ income level, enhancing farmers’ ability to obtain information, and promoting farmers’ realization of off-farm employment is an important mechanism by which Internet use affects farmers’ vulnerability to poverty.

In particular, in order to ensure the robustness of the conclusions of the study, as mentioned above, while carrying out the above research, (1) we use instrumental variable method to deal with possible endogeneity problems, and get almost consistent regression results with Probit; (2) the PSM method is used to estimate the average treatment effect of poverty vulnerability through k-nearest neighbor matching, radius matching and other methods. The regression results further confirm the previous conclusions. In addition, on the premise of maintaining the consistency of dependent variables, independent variables and control variables with this study, the Probit regression analysis of CFPS2016 data was carried out. Although maybe affected by factors such as sample size and Internet penetration rate (in 2016, China’s rural Internet penetration rate was only 33.1%, which increased to 38.4% in 2018), the significance was slightly lower than that of this study, it still supports the findings of this study to a certain extent. Therefore, the research conclusion of this paper is valid.

The research conclusion of this paper also has the following policy recommendations. First, actively promote the full coverage of peasant households’ Internet infrastructure and promote the comprehensive popularization of the Internet in rural areas is a must [44]. As an important new tool for poverty alleviation and sustainable development, the Internet should cultivate farmers’ scientific thinking and skills on the Internet, correctly guide farmers to use the Internet, and realize the effective integration of traditional thinking and Internet plus thinking. Secondly, the government should strengthen the training of Internet use skills for farmers, rural women, and rural low-income groups in the rural areas of central and western China, enrich and improve farmers’ access to information channels, speed up the construction of professional information platforms, eliminate the poverty of information access, and achieve sustainable development of poverty alleviation. Finally, an Internet plus long-term poverty alleviation mechanism should be established to provide lasting internal impetus for poverty alleviation and sustainable development. Further recommendations include the breaking of the traditional thinking of poverty alleviation, giving full play to the Internet plus entrepreneurship, Internet plus agricultural education, Internet plus financial, Internet plus medical, and other fields. The results are: (i) the achievement of sustainable poverty alleviation of rural households; (ii) the empowerment of the role of the Internet in the allocation of information resources in poverty-stricken areas; and (iii) the promotion of the concentration of all kinds of resources to poverty-stricken and vulnerable rural households. Previous studies on poverty in underdeveloped countries, such as sub-Saharan Africa, Pakistan, have indicated that Internet use can increase farmers’ income and eliminate poverty vulnerability. This paper has similar research conclusions; therefore, the corresponding countermeasures and suggestions are also applicable to the underdeveloped countries.

It is worth noting that the data used in this paper are relevant data from CFPS in 2018, so the relevant countermeasures and suggestions still need to be viewed dialectically. Meanwhile, due to the uneven regional distribution of the samples, the impact of the use and popularization of the Internet in the central region on poverty vulnerability is not obvious, and the follow-up studies will continue to pay attention to and discuss this reason in depth.