How Digital Skills Affect Rural Labor Employment Choices? Evidence from Rural China

Abstract

Expanding employment channels for rural households is a crucial means of enhancing the income of rural residents and enhancing the quality of rural employment. This study examines the impact of digital skills on rural laborers’ employment choices and explores the underlying mechanisms by using data from the China Family Panel Studies (CFPS) spanning 2014–2018. By employing various models, including the Probit, IV, mediated effects, and propensity-score-matching methods, the study reveals that digital skills have a significant impact on rural laborers’ employment choices. Specifically, digital skills increase rural labor’s employment opportunities in nonfarm and employed employment while reducing the proportion of informal employment. Additionally, the analysis indicates that the main channels through which digital skills influence rural labor’s employment choices are human and social capital. A heterogeneity analysis further reveals that work-study and social-entertainment skills have a more significant effect on rural laborers’ nonfarm and employed employment opportunities while inhibiting informal employment. Hence, to enhance the quality of future rural employment, the government must encourage rural workers to enhance their digital literacy and digital application skills while improving digital infrastructure.

1. Introduction

Employment is the biggest livelihood factor and the most basic support for economic development. Widening the employment channels of the rural labor force is an important way to increase income sources and improve living standards [1]. However, in many low- and middle-income countries, labor force retention in rural areas is prominent, and frictional unemployment and structural conflicts have increased [2]. The employment problem of unemployed and low-skilled people in rural areas has gradually become prominent owing to their lack of job skills. To improve the employment situation of rural residents, the Chinese government has taken a series of feasible measures. For example, it has promulgated employment-related policies and plans such as the Guidance on Stabilizing and Expanding Employment in the Development of the Digital Economy and the Outline of the Digital Countryside Development Strategy, which provide a sound policy environment for the stable development of employment. Although these measures have achieved certain results, the employment situation of China’s rural labor force is still unsatisfactory [1]. Therefore, actively transforming the traditional employment model and building a more diversified employment choice model has become the long-term driving force for sustainable employment for China’s rural labor force.

With the improvement of digital infrastructure and the popularization of intelligent devices, digital skills play an increasingly significant role in workers’ access to information. The improvement of digital infrastructure has greatly changed the lifestyle and communication behavior of rural residents and has a certain impact on their individual employment choices [3,4,5,6,7,8]. According to data released by the China Internet Network Information Technology Center, the Internet penetration rate in rural China reached 58.8% in 2022, and 160 million rural households have broadband access. The length of long-distance optical cable lines is also growing. Digitally skilled workers tend to have higher education levels, digital literacy, and job-searching ability [9,10]. In the US labor market data, Atasoy [2] found that the acquisition of digital skills can increase the employment rate by about 1.8 percentage points, with a greater impact in rural and remote areas. At the same time, Manuela [11] analyzed the employment situation in Albania’s rural areas and found that digital skills are conducive to the nonagricultural employment of rural labor, bring efficient and low-cost learning, and help improve the human capital level of labor. Digital skills can also promote entrepreneurial opportunities for rural laborers and improve the quality of life of rural residents [12,13,14]. Job searches based on digital skills have positive effects on the employment of rural laborers and can alleviate employment difficulties and frictional unemployment [15,16]. Agrawal et al. [17] claimed that with the help of digital technology, difficulties in obtaining employment information can be alleviated to a certain extent. Digital technology provides opportunities for social and economic development and helps to alleviate employment difficulties [18,19]. Adesugba et al. [20] found that more and more individual workers in Nigeria would search for employment information through the Internet. Digital skills can improve the imbalance between supply and demand in the labor market and narrow the digital divide [21,22,23]. Therefore, improving rural digital infrastructure and improving rural residents’ digital literacy and digital application ability have become key measures to promote rural labor employment.

The positive impact of digital skills in rural labor force employment has been confirmed by several studies [2,4,9,11,16]. However, the specific role of digital skills on the employment choices of the rural labor force remains unclear and therefore needs to be explored in more depth. The potential contributions of this paper are as follows: (1) This paper systematically analyzes the impact of digital skills on the employment choices of rural laborers in China and provides new perspectives for promoting the diversified employment of rural laborers. (2) To deal with possible endogeneity problems, this paper adopts PSM, variable replacement, and the IV-Probit two-step method for robustness testing, which provides a reference research method for similar studies. (3) This paper also discusses the potential mechanism of social capital and human capital that digital skills affect in rural labor’s employment choices.

The main purpose of this paper is to explore the impact of digital skills on rural labor force employment choices, including the importance, role, and impact mechanisms of digital skills in rural labor force employment choices. Specifically, this paper aims (1) to explore the role of digital skills in offering employment choices for rural laborers and the impact of digital skills on enhancing employment opportunities for rural laborers; (2) to elucidate the mechanisms of the role of digital skills in influencing the employment choices of rural laborers; and (3) to analyze the facilitative effects of digital skill uses on rural labor force employment choices and the differences in the effects of employment choices on different groups of the rural labor force, including the effects of different genders and regions. Through these studies, we can gain a deeper understanding of the role of digital skills on the employment decisions of rural laborers and provide more-powerful policy suggestions for promoting the employment of rural laborers.

2. Materials and Methods

2.1. Theoretical Analysis and Hypotheses

2.1.1. Analysis of Digital Skills and Employment Choices for Rural Laborers

The relationship between digital skills and employment has been receiving attention from the academic community. The acquisition of digital skills can broaden the access of rural laborers to external information, improve their efficiency in obtaining employment information, and facilitate the reasonable matching of rural laborers with employment positions, thus increasing the probability of employment [24,25]. In addition, the acquisition of digital skills can reduce the cost of information acquisition and alleviate technical and knowledge-based economic poverty in rural areas, which has a positive impact on the development of the digital economy in rural areas. The mastery of digital skills can optimize the matching of rural labor with labor market information and match high-quality employment resources to cities and regions with low information levels, thus alleviating the current uneven distribution of employment resources [26]. The first hypothesis is formulated as follows:

Hypothesis 1. 

Digital skills can facilitate the employment choices of rural labor.

2.1.2. Analysis of Social Capital, Digital Skills, and Employment Choices for Rural Laborers

Studies on the relationship between individual social capital and employment have shown that the individuals’ possession of rich social capital is one of the key factors affecting their employment [27]. The acquisition of digital skills can strengthen the social network communication of rural laborers, broaden the scope and depth of social relationships, and contribute to the accumulation of social capital [23]. This social capital can help rural workers to obtain employment information and can provide employment opportunities and resources, thus improving their employment quality and employment probability [4,20,22]. Empirical studies have shown that in rural South Africa, having an extensive social network can significantly increase the probability of employment among rural residents [21]. Digital skills, as a bridge for information exchange, can not only expand an individual’s social network but also help maintain the long-term value and stability of social capital [28,29]. Specifically, the acquisition of digital skills can help farmers gain access to more employment information and opportunities by expanding private social networks and engaging in online communication, thus increasing their employment options and development opportunities in the nonfarm sector [27]. On the basis of the above findings, the following hypothesis is proposed:

Hypothesis 2. 

Social capital mediates the effect of digital skills on rural labor’s off-farm employment, employed employment, and informal employment choices.

2.1.3. Analysis of Human Capital, Digital Skills, and Employment Choices for Rural Laborers

With the rapid development of digital technology, digital networks have become a new learning platform that overcomes the limitations of time and geographical learning resources and provides more-convenient conditions for achieving learning resource sharing and online communication [16,17]. The mastery of digital skills can enable rural laborers to access learning resources such as vocational skills training at a lower cost without leaving home, thus improving their skills and cognitive abilities [18]. Vocational skills training can enhance the human capital level of rural workers, and having a higher level of human capital will increase their labor skill level and productivity, thus making it easier to find employment opportunities [19]. In addition, the level of human capital supported by digital skills plays an important role in the employment decisions of the rural labor force [30]. On the basis of the above findings, the following hypothesis is proposed:

Hypothesis 3. 

Human capital mediates the effect of digital skills on rural labor’s off-farm employment, employed employment, and informal employment choices.

2.2. Methods

2.2.1. Data Sources

This paper is based on data from the China Family Panel Studies (CFPS) from 2014 to 2018. The sample covers 25 provinces, municipalities, and autonomous regions, which is widely representative. Given that the study population is the rural labor force, aged 19 to 65, we keep only the data of rural households. After removing some samples with apparently abnormal data or missing main variables, a final sample of 5598 valid samples was obtained.

2.2.2. Descriptions of Variables

The variables used in this paper are shown in

Table 1. Definitions of variables and descriptive statistical results.

Variable	Description or Definition	All		Mastered Digital Skills		Failed to Master Digital Skills
		Mean	S.D.	Mean	S.D.	Mean	S.D.
Nonagricultural employment	1: Yes; 0: No	0.228	0.420	0.282	0.450	0.169	0.375
Employed work	1: Yes; 0: No	0.277	0.447	0.357	0.479	0.187	0.390
Informal employment	1: Yes; 0: No	0.944	0.229	0.917	0.274	0.973	0.161
Digital Skills	1: Yes; 0: No	0.524	0.499	-	-	-	-
Work-based learning skills	1: Yes; 0: No	0.803	0.398	0.624	0.485	-	-
Social-entertainment skills	1: Yes; 0: No	0.913	0.282	0.833	0.373	-	-
Gender	1: Male; 0: Female	0.569	0.495	0.564	0.496	0.574	0.496
Age	Continuous variable	48.941	9.281	45.870	9.313	52.288	8.123
Nation	1: Han; 0: Other	0.009	0.092	0.014	0.117	0.003	0.051
Marital status	1: Married; 0: Others	0.883	0.321	0.895	0.307	0.871	0.335
Political status	1: Chinese communists; 0: Others	0.083	0.277	0.091	0.287	0.075	0.264
Education level	1: Never went to school; 2: Primary school; 3: Junior high school; 4: Senior middle school; 5: Junior college; 6: Undergraduate; 7: Master and above	2.357	1.087	2.575	1.095	2.116	1.025
Health condition	Self-assessment indicators	5.391	1.241	5.410	1.182	5.272	1.294
Brains condition	Self-assessment indicators	5.272	1.297	5.480	1.184	5.043	1.376
Family size	Continuous variable	3.981	1.847	3.982	1.728	3.980	1.971
Farm income	Continuous variable	9.106	1.028	9.152	1.107	9.056	0.932
Social Capital	1: Attend nonacademic education; 0: No	0.056	0.229	0.090	0.287	0.018	0.132
Human Capital	1: Get help from others; 0: No	0.964	0.186	0.948	0.223	0.982	0.133
Observations		5598		2935		2663

Note: Calculated on the basis of CFPS data from 2014 to 2018.

.

(1)

Dependent Variable

The dependent variable studied in this paper is the employment status of rural labor. Referring to the research method of Song and He [31], we classify them into three categories, according to the employment sector: whether they are in nonfarm employment, employed employment, or informal employment.

(2)

Independent Variable

The core independent variable in this paper is digital skills, i.e., rural workers take a value of 1 in the current period relative to their latest mastery of digital skills in the previous period and 0 otherwise. On this basis, this paper draws on Mou et al.’s study [8] to further subdivide digital skills into two dummy variables, namely work-based learning skills and social-entertainment skills, to explore the possible differential impact of different types of digital skills.

(3)

Control Variables

This paper controls for exogenous factors, such as gender, age, ethnicity, and family size, that affect the rural labor force employed in the study of He and Song [32]. In addition, this paper controls for some endogenous factors affecting employment, including marital status, political affiliation, health status, intelligence level, years of education, and net household income.

(4)

Mediator Variable

Human capital is one of the important factors affecting the entry of rural labor into the labor market [33]. Most scholars use workers’ education level and average years of education as proxy variables of human capital [34]. However, given the dynamic nature of human capital and the subjective initiative of workers, this paper uses the indexes proposed by Qi and Chu [29] to measure human capital, including participation in nonacademic education and whether people obtain help from others to find jobs.

(5)

Statistical Description of the Variables

As can be seen from Table 1, the proportion of nonagricultural employment for rural labor with digital skills is 28.2%, higher than that of the group without digital skills (16.9%). In addition, the proportion of rural workers with digital skills in employed employment was 35.7%, which was significantly higher than that of those without digital skills (18.7%). However, the proportion of rural workers with digital skills in informal employment was 91.7%, lower than the 97.3% of those without digital skills. These descriptive statistics suggest that mastering digital skills may influence the employment choices of rural workers. In addition, there are some differences between the characteristics of the two groups, those with and those without digital skills. For example, slightly fewer men than women have digital skills; moreover, the average age of those with digital skills is about 45, lower than the average age of those without digital skills (52). It should be noted that samples from those younger than 19 years old and those older than 65 years old are excluded from this study because most of these samples are not in the range of job market choices [33].

2.2.3. Model Selection

(1)

Baseline Regression Model

The dependent variable in this paper is a binary dummy variable, and traditional regression models often have problems with predicted values exceeding the range of the dummy variable when dealing with this type of variable. Additionally, the Probit model is widely used to deal with such problems so that the regression results are in the normal range. Therefore, following Yin et al.’s (2019) study [35], we choose the Probit model to analyze the impact of digital skills on the employment choices of rural laborers. The baseline regression model is set as follows:

$$Pr(emplo{y}_{it}=1)=\varnothing \left(\alpha +\beta \times digitalskil{l}_{it}+\sum {\theta }_j{X}_j\right)$$

(1)

The explanatory variable employs three employment choice variables, taking a value of 1 for nonfarm employment, employed employment, and informal employment and 0 for agricultural jobs, self-employment, and formal employment. The core explanatory variable Digitalskill is a dummy variable, i.e., rural workers take 1 if they have newly acquired digital skills in the current period relative to the previous period; otherwise, they take 0. In addition, Digitalskill is further subdivided into two dummy variables, namely work-study skills and social-entertainment skills, to discuss the possible differentiated effects of different digital skill types. Here, i denotes the rural labor force, and t represents the year. X_j is a control variable, including the year and individual, their household, and regional characteristics that may affect rural labor force employment. β indicates the marginal effect of digital skill mastery on rural labor force employment choices.

While digital skills influence the employment choices of rural laborers, the employment choices of rural laborers themselves may influence whether individuals possess digital skills [36]. Those with digital skills may have higher education, social status, and other personal characteristics that affect their employment. To effectively solve the possible endogeneity problem between digital skills and rural labor employment, this paper uses the propensity-score-matching (PSM) method proposed by Rosenbaum and Rubin [37], which can verify the relationship between digital skills and rural labor employment.

(2)

Mediating Effect Model

In the transmission mechanism analysis, digital skills may influence the employment choices of rural labor through the human capital channel and the social capital channel. In this paper, a standardized mediating effects model is used for further empirical testing to analyze the indirect effect of the explanatory variable (X) on the explanatory variable (Y) through the mediating variable (M), by using the following equations:

$$Y=cX+e_1$$

(2)

$$M=aX+e_2$$

(3)

$$Y=c^{\prime }X+bM+e_3$$

(4)

The mediating effect model was constructed with labor force employment choice (employment) as the explanatory variable Y, human capital and social capital as the mediating variable M, and digital skills as the explanatory variable X.

3. Results

3.1. Baseline Regression Results

The results in

Table 2. Impact of digital skills on different types of employment choices for the rural labor force.

Variables	Nonagricultural Employment		Employed Work		Informal Employment
	(1)	(2)	(3)	(4)	(5)	(6)
Digitalskill	0.155 *** (0.014)	0.095 *** (0.015)	0.089 *** (0.014)	0.046 *** (0.014)	−0.050 *** (0.008)	−0.022 *** (0.007)
Age	1.209 *** (0.193)	0.821 *** (0.199)	1.125 *** (0.183)	0.797 *** (0.189)	−0.434 *** (0.088)	−0.149 (0.097)
Square of age	−0.813 *** (0.188)	−0.440 ** (0.194)	−0.816 *** (0.178)	−0.506 *** (0.184)	0.303 *** (0.089)	0.036 (0.098)
Gender	0.060 *** (0.012)	0.035 *** (0.012)	0.029 *** (0.011)	0.009 (0.011)	−0.024 *** (0.006)	−0.011 * (0.006)
Nation	0.225 *** (0.061)	0.168 *** (0.064)	0.222 *** (0.056)	0.174 *** (0.059)	−0.064 *** (0.023)	−0.039 * (0.023)
Family size	0.208 *** (0.030)	−0.144 *** (0.031)	0.195 *** (0.029)	0.126 *** (0.010)	−0.059 *** (0.017)	−0.028 (0.017)
Marital status		−0.042 ** (0.019)		−0.053 *** (0.017)		0.019 ** (0.009)
Political status		0.056 *** (0.020)		0.056 *** (0.019)		−0.015 * (0.009)
Education level		0.050 *** (0.005)		0.029 *** (0.005)		−0.025 *** (0.003)
Health condition		0.182 *** (0.034)		0.167 *** (0.032)		−0.025 (0.016)
Brains condition		−0.136 *** (0.034)		−0.139 *** (0.033)		0.045 *** (0.018)
Farm income		−0.316 *** (0.035)		−0.247 *** (0.033)		0.083 *** (0.013)
Year_Fe	Yes		Yes		Yes
Province_Fe	Yes		Yes		Yes
Pseudo R2	0.102	0.143	0.071	0.109	0.135	0.199
N	5598	5598	5598	5598	5598	5598

Note: *** p < 0.01, ** p < 0.05, * p < 0.1. Robust standard errors in parentheses. Regression coefficients obtained from the Probit model have been transformed into marginal effects.

report the marginal effects of digital skills on the nonfarm employment, employed employment, and informal employment of the rural labor force by introducing exogenous and endogenous control variables. The results in columns (1) and (2) show that after the exogenous and endogenous variables were controlled for, the probability of the nonfarm employment of rural laborers with digital skills is 9.5 percentage points higher than that of laborers without digital skills, and digital skills have a significant positive effect on the nonfarm employment of rural laborers. The results in columns (3) and (4) show the effect of digital skills on the employed employment of rural laborers; overall, the marginal effect of digital skills is significantly positive; and digital skills make rural laborers 4.6% more likely to be employed. The results in columns (5) and (6) show that digital skills reduce the probability that the rural labor force will be in informal employment by 2.2%.

3.2. Robustness Test Results

3.2.1. Propensity-Score-Matching Estimation

The study in this paper focuses on the influence of digital skills on the employment choice of rural laborers, taking into account that the employment choice itself may influence the self-selection behavior of whether individuals acquire digital skills. For this reason, this paper adopts the propensity-score-matching method for processing and chooses three methods: proximity matching, radius matching, and kernel matching. The results of the balance test showed that the proportion of deviations between groups of the matched sample variables were all less than 10%, and there were no significant differences between the treatment and control groups, which satisfied the balance hypothesis. In

Table 3. Results of propensity score matching.

Variables	Nonagricultural Employment			Employed Work			Informal Employment
	NNM	RM	KBM	NNM	RM	KBM	NNM	RM	KBM
ATT	0.068	0.079	0.083	0.042	0.045	0.048	−0.014	−0.016	−0.015
ATU	0.060	0.060	0.062	0.033	0.035	0.036	−0.010	−0.009	−0.010
ATE	0.064	0.070	0.073	0.038	0.040	0.042	−0.012	−0.010	−0.012

Data in parentheses are t-values. Note: The radius in K = 4 radius matching is set to 0.01, and the width of kernel-based matching is the default.

, the ATT values in the three matching results ranged from 0.0678 to 0.0790 for nonfarm employment and from 0.0424 to 0.0448 for employed employment and passed the 1% significance test.

3.2.2. Regression Estimation with Variable Substitution

To test the robustness of the results of the baseline model, this paper uses the importance of Internet information to replace digital skills as an explanatory variable [38]. The higher the importance of an individual’s use of the Internet as an information channel, the greater the probability of acquiring digital skills and the frequency of using the corresponding skills, and there is a strong correlation between the two. According to the results in

Table 4. Results of variable substitution regression.

Variables	Nonagricultural Employment	Employed Work	Informal Employment
Information Channels	0.050 *** (0.009)	0.021 ** (0.009)	−0.014 *** (0.004)
Control variables	Yes	Yes	Yes
Year_Fe	Yes	Yes	Yes
Province_Fe	Yes	Yes	Yes
R2	0.141	0.108	0.200
N	5598	5598	5598

Note: Standard errors are in parentheses; ***, ** denote 1%, 5% significance levels, respectively.

, after replacing digital skills with the importance of Internet information as an explanatory variable, the probability of both nonfarm employment and employed employment significantly increases and is positive at the 1% and 5% statistical levels, respectively, controlling for exogenous and endogenous variables. Meanwhile, informal employment is significantly negative at the 1% statistical level.

3.2.3. IV Estimation

To deal with the possible endogenous problems in the Probit model, this paper selects provincial Internet penetration rate as the instrumental variable of digital skills [39]. Studies have shown a correlation between digital skills and provincial Internet penetration. To further ensure the effectiveness of instrumental variables, a weak instrumental variable test was conducted in this study. The results showed that statistical value F in the first stage of

Table 5. Regression results of instrumental variables.

Iv	Nonagricultural Employment	Employed Work	Informal Employment
Internet penetration rate	0.002 ** (0.001)	0.002 ** (0.001)	0.002 ** (0.001)
Control variables	Yes	Yes	Yes
Year_Fe	Yes	Yes	Yes
Province_Fe	Yes	Yes	Yes
R2	0.476	0.476	0.476
N	5598	5598	5598
Phase I F-statistic	316.6	316.6	316.6
Phase II
Digital skills	7.528 ** (3.423)	7.077 ** (3.326)	−10.651 ** (5.022)
Control variables	Yes	Yes	Yes
Year_Fe	Yes	Yes	Yes
Province_Fe	Yes	Yes	Yes
Wald chi2	19.76	17.86	17.49
N	5598	5598	5598

Note: Standard errors are in parentheses; ** denote 5% significance levels, respectively.

far exceeded the critical value of 10%, indicating that instrumental variables were not weak. At the same time, the regression results of the first stage in Table 5 show that digital skills have a significant impact on the nonagricultural employment, employed employment, and informal employment of the rural labor force, with the provincial Internet penetration rate as the instrumental variable under the condition that the control variables remain unchanged, and the result significance level is 5%.

3.3. Mechanism Test Results

By looking at

Table 6. The mediating effect.

Variables	Human	Nonagricultural	Employed	Informal	Social	Nonagricultural	Employed	Informal
Digital skills	0.502 *** (0.086)	0.323 *** (0.052)	0.154 *** (0.054)	−0.246 *** (0.081)	−0.197 ** (0.093)	0.325 *** (0.053)	0.145 *** (0.056)	−0.246 *** (0.081)
Human		0.293 *** (0.081)	0.332 *** (0.084)	−0.533 *** (0.101)
Social						−2.218 *** (0.183)	−0.301 *** (0.156)	−0.533 *** (0.101)
Control variables	Yes	Yes	Yes	Yes		Yes	Yes	Yes
R2	0.190	0.145	0.102	0.210	0.094	0.189	0.086	0.210
N	5598	5598	5598	5598	5598	5598	5598	5598

Note: Standard errors are in parentheses; ***, ** denote 1%, 5% significance levels, respectively.

, it can be observed that digital skills have a significant effect on human capital and social capital at the 1% level. Specifically, the coefficient of human capital is significantly positive and has a significant positive effect on both the off-farm and employed employment of the rural labor force, but it has a significant negative effect on informal employment. In contrast, the coefficient of the social capital channel shows a significant negative effect.

3.4. Heterogeneity Test Results

The results of the baseline regression and robustness tests verify that digital skills have a significant promotion effect on the nonfarm employment and employed employment of rural laborers and a significant adverse impact on informal work. Despite controlling for the control variables at the individual, household, and regional levels, the individuals selected for the questionnaire survey still need to be completely homogeneous, and their promotion effects have significant variability across groups. Therefore, this paper discusses the heterogeneity of the influence of digital skills on employment choice from the perspectives of different digital skills, gender, age, and different regions.

3.4.1. Heterogeneity in the Use of Digital Skills

Table 7. Impact of different digital skills on the employment of rural labor force.

Variables	Nonagricultural Employment	Employed Work	Informal Employment
Work-study skills	0.069 *** (0.014)	0.029 ** (0.139)	−0.331 *** (0.103)
Control variables	Yes	Yes	Yes
R2	0.140	0.098	0.227
N	5598	5598	5598
Social-entertainment skills	0.057 *** (0.019)	0.010 * (0.018)	−0.143 * (0.082)
Control variables	Yes	Yes	Yes
R2	0.138	0.098	0.218
N	5598	5598	5598

Note: Standard errors are in parentheses; ***, **, * denote 1%, 5%, and 10% significance levels, respectively.

reports the estimated results of the employment impact of the rural labor force’s acquisition of different digital skills. The results show that work-study skills and social-entertainment skills significantly affect nonfarm employment at the 1% level, where work-study skills make the largest positive contribution to nonfarm employment. Work-based learning skills significantly affect employed employment at the 5% level, while social recreational skills significantly affect employed employment at the 10% level. It is noteworthy that work-based learning skills and social recreation skills, however, significantly reduce the probability of informal employment.

3.4.2. Gender Heterogeneity

A regression analysis was performed by gender grouping, controlling for individual, family, region, and year variables.

Table 8. Impact of digital skills on the employment of different genders in the workforce.

Variables	Nonagricultural Employment		Employed Work		Informal Employment
	Men	Women	Men	Women	Men	Women
Digital skills	0.222 *** (0.061)	0.353 *** (0.070)	0.092 (0.061)	0.255 *** (0.072)	−0.295 *** (0.101)	−0.022 (0.119)
Control variables	Yes	Yes	Yes	Yes	Yes	Yes
R2	0.150	0.138	0.092	0.121	0.174	0.258
N	3185	2413	3185	2413	3185	2413

Note: Standard errors are in parentheses; *** denote 1% significance levels, respectively.

shows that at the significance level of 1%, digital skills have a significant impact on male and female nonfarm employment, and the probability of nonfarm employment of rural women is higher than that of men. However, when it comes to employed employment, digital skills have a significant effect only on women, not on men. For informal employment, the influence coefficient of digital skills on the male sample is −0.295, which is significant at the 1% significance level, but not on the female sample.

3.4.3. Regional Heterogeneity

In this paper, the overall samples were divided into eastern, central, and western samples according to their respective regions, and a regression analysis was conducted. According to the results in

Table 9. Impact of digital skills on rural labor force employment in different regions.

Variables	Nonagricultural Employment			Employed Work			Informal Employment
	East	Center	West	East	Center	West	East	Center	West
Digital skills	0.273 *** (0.077)	0.292 *** (0.087)	0.253 *** (0.078)	0.170 ** (0.079)	0.089 (0.088)	0.211 *** (0.080)	−0.331 *** (0.128)	−0.044 (0.150)	−0.126 (0.130)
Control variables	Yes	Yes	Yes	Yes	Yes	Yes	Yes	Yes	Yes
R2	0.152	0.132	0.129	0.106	0.083	0.108	0.231	0.216	0.181
N	1901	1589	2108	1901	1589	2108	1901	1589	2108

Note: Standard errors are in parentheses; ***, ** denote 1%, 5%, and significance levels, respectively.

, digital skills have a significant impact on the nonfarm employment of the rural labor force in the eastern, central, and western regions, at the 1% level of significance. In addition, digital skills also have significant effects on the employed employment of the labor force in the eastern and western regions, but not in the central region. There is also a negative effect of digital skills on the rural labor force in the eastern region at the 1% statistical level, while there is an inhibitory effect in the central and western regions.

4. Discussion

After using the Probit model regression, robustness test, and mediation effect test, this paper finds that digital skills have important effects on the employment choice of rural labor in China. Specifically, digital skills significantly improve the possibility of nonagricultural employment and employed employment for China’s rural labor force while also helping to promote the transition of the rural labor force from informal to formal employment. Further research finds that digital skills have significant effects on the nonagricultural employment and employed employment of China’s rural labor force through the improvement of human capital but have inhibitory effects on informal employment. In addition, this paper also discusses the mediating effect of social capital on the nonagricultural employment and employed employment of rural labor, but no significant results are found. In terms of heterogeneity, research shows that work-learning skills and social recreation skills have a significant impact on the nonagricultural employment and employed employment of rural labor, but their contribution to informal employment is not obvious. Digital skills also help increase rural women’s chances of achieving off-farm and employed employment. Finally, the study also found that there may be regional differences in the impact of digital skills on rural labor employment in different regions.

Digital skills have significant effects on the nonfarm and employed employment of rural laborers in China, which is consistent with the findings of Atasoy [2], Zhao and Xiang [4], Mao et al. [10], Manuela [11], Badea et al. [16], and Campos et al. [26]. In addition, our study supports the findings of Song and He [31] on the positive effect of digital skills on the formal employment of the rural labor force in China. Digital skills increase the probability of nonfarm and employed employment while decreasing the probability of informal employment. Our findings are similar to those of Kuhn and Mansour [13], Agrawal et al. [17], Adesugba et al. [20], and Liu and Diao [27], which show that digital skills contribute most to human-capital-mediated effects on nonfarm and employed employment. However, digital skills have a dampening effect on informal employment, which is where our study differs from previous studies. Our study shows that good digital literacy and occupational skills improve the ability of rural laborers to find and choose employment, giving them a competitive advantage in the labor market [10]. On the other hand, there is no significant mediating effect of social capital in the employment decisions of rural labor. This suggests that digital skills can increase the probability of employment by reducing the need for help from others, decreasing the level of dependence on others, and facilitating active job search opportunities [29]. In addition, the study found that work-study skills and social-entertainment skills contributed the most to nonfarm and employed employment among the rural labor force, which is consistent with the findings of Qi and Chu [29] and He and Song [32]. However, their contribution to informal employment is not significant. These results suggest that individuals with work-based learning skills are relatively scarce in resource-limited rural areas and play an irreplaceable role in promoting nonfarm and employed employment. At the same time, the positive impact of online learning, socializing, and entertainment on informal employment is offset by their negative impact, suggesting that the contribution of these skills to informal employment is insignificant. Our study further reveals that digital skills have a more significant impact on nonfarm and employed employment among rural women compared with that among men, which is consistent with the findings of Dettling [25] and Lin et al. [40]. This may be because rural male laborers need a stable and continuous income to support their families and thus prefer formal employment with stable employment relationships and social security. In contrast, rural female laborers are more likely to choose flexible employment activities to take care of their families. Finally, our study finds that the nonfarm employment effect of digital skills is significantly higher in eastern and central China than in western China, while the employed employment effect is higher in western China than in eastern and central China. This may be due to the more widespread digital infrastructure and technology adoption in the eastern and central regions than in the western region, such that digital skills generate more nonfarm employment opportunities for rural laborers in eastern and central China.

Digital skills play an active role in promoting diverse employment in the rural labor force and realizing sustainable employment. First of all, the mastery of digital skills can provide more job opportunities for rural laborers, expand employment channels, and improve their employment quality. The mastery of digital skills can also help rural workers better adapt to the needs of the labor market and improve their competitiveness, thus achieving sustainable employment. Second, the government should strengthen the construction of rural digital infrastructure, guide the rural labor force to use the Internet for job hunting and vocational training, stimulate their enthusiasm to learn digital skills, and improve the application of digital skills in the labor market. In this way, the rural workforce can better master digital skills, better adapt to the job market in the digital age, and achieve employment sustainability. Third, mastering digital skills can help narrow the regional digital development gap and promote coordinated regional development. The mastery of digital skills can improve the competitiveness of the rural labor force in the digital economy and promote the development of the digital economy in rural areas, thus narrowing the gap between urban and rural digital development and promoting the coordinated development of regions.

However, our study has some limitations. First, because of data limitations, we used mainly the latest information on digital skills acquired by the rural labor force relative to the previous period, without exploring the depth and quality of their digital skills. Therefore, it is necessary to further expand the definition of digital skills and find relevant data to explore the impact of digital skills on the employment decisions of the rural labor force. Second, we did not deeply investigate the path of digital skills’ impact on different age groups. Because rural laborers of different ages have different levels of digital skills, this may affect employment decisions in different ways. Therefore, further quantitative studies in this area are needed in the future.

5. Conclusions

With the development of digital technology, digital skills have continuously penetrated various fields of people’s lives and become an important force to improve people’s quality of life. This study aims to explore the impact of digital skills on the employment choices of the rural labor force in China and its mechanism of action. By analyzing the China Family Panel Studies data from 2014 to 2018, this study finds that digital skills have positive effects on rural laborers’ off-farm and employed employment choices, and the findings are verified in the tests of multiple empirical methods. Digital skills can significantly affect rural labor’s nonfarm and employed employment by increasing the level of human capital, while the effect on informal employment is not significant. In addition, this study also finds that work-study skills and social-entertainment skills can increase the probability that rural laborers will obtain nonfarm and employed employment, but they do not significantly contribute to informal employment. Digital skills also promote nonfarm employment among rural female laborers and also contribute to the transfer of rural male laborers from informal to formal employment. These findings provide useful references for the development of digital skills and employment policies for the rural labor force.