More Rational, More Attractive: Industrial Structure Rationalization and Migrant Workers’ Employment Choices in China

Abstract

Industrial structure rationalization could affect not only the employment structure but also the micro-employment choices of the labor force. Using the national individual-level survey data, we examine how regional industrial structure rationalization influences the employment status and location choices of migrant workers respectively using a probit model. The results show that industrial structure rationalization can significantly increase the probability that migrant workers choose regular and cross-provincial employment and reduce the probability of intra-provincial labor mobility, self-employment, and temporary employment. Higher employee welfare and income are the main mechanisms underlying the effect on migrant workers’ employment choices. Furthermore, the effects of industrial structure rationalization on the employment status and location choices of migrant workers differ according to gender, family size, and industry. These findings deepen the understanding of the relationship between industrial structure rationalization and labor migration and offer references for governments to promote regional sustainable development.

1. Introduction

Since the reform and Opening-up of China, industry has generated significant economic dividends that have led to rapid economic growth. Driven by policies to achieve sustainable development, China’s economy is gradually transitioning from high growth to high-quality growth. The reallocation of resources, the optimization of production processes, and the change in dominant industries have pushed the industrial structure from being dominated by energy-intensive and emission-intensive manufacturing industries to being dominated by innovation-oriented and knowledge-intensive industries [1], and the industrial structure tends to be rationalized. In this process, along with the change in industrial structure, the labor market will also make dynamic adjustments [2,3]. The upgrading of the industrial structure as a result of technical progress can lead to both job creation and loss [3,4]. As industries adopt information technology and automation, they shift from being labor-intensive to capital-intensive [1]. These changes in industrial structure enhance the difficulty of finding jobs in technologically lagging sectors and increase the unemployment rate of workers with low levels of education [4,5]. At the same time, the new industries formed by the rationalization of industrial structure also create new employment opportunities and upward mobility for the labor force, especially the high-skilled labor force [3,5]. In addition, the expansion of Internet-based services provides more employment options for laborers seeking to diversify their career choices [6,7].

Compared to the local workers, the migrant workers have relatively homogeneous skills and greater employment instability and, thus, are more sensitive to the labor market changes brought about by the industrial structure rationalization [8,9,10]. China’s household registration system reform, which began in the 1980s, has resulted in many laborers migrating from the rural areas to the cities. (In the 1950s, in order to ensure the implementation of the strategy of prioritizing the development of heavy industry, China adopted a household registration system (hukou system) that separated urban and rural populations. The hukou system was closely related to employment, social welfare, and the supply of consumer goods, and it effectively hindered labor mobility between urban and rural areas, regions, and industrial sectors. After the reform and opening up, the Chinese government has gradually reformed the hukou system, relaxing the restrictions on population migration and allowing rural laborers to work in urban areas and obtain local urban hukou registration. The number of rural–urban migrant workers increased from 67,136,000 in 1987 to 292,510,000 in 2021 (data from China National Bureau of Statistics). By 2021, migrant workers accounted for 39.18% of the total labor force, and labor mobility between rural and urban areas and between regions became very common. Moreover, due to the labor productivity gap between agriculture and other industries and regional income differences, rural workers will mostly choose to go out to work across regions and industries [11,12]. The large-scale migrant workers have greatly contributed to the economic growth of the inflow areas and become an important driving force for regional development [13,14]. Therefore, the effect of industrial structure rationalization on the employment decision and employment tendency of migrant workers is crucial for the sustainable development of the regional economy.

We use a probit model to estimate the effect of industrial structure rationalization on the employment choices of migrant workers. In general, workers have two primary employment options: job status and location. We use self-employment (SE), regular employment (RE), and temporary employment (TE) to define different employment status choices, respectively. We also define employment location choices based on the location of employment and home location as cross-provincial employment (CPE), cross-city employment within a province (CCEP), and cross-county employment within a city (CCEC). We use the inverse of the Thiel index to measure the level of industrial structure rationalization. To address possible endogeneity issues in the baseline regressions, we use an IV-probit model to re-estimate them, and we also conduct a series of robustness checks.

Controlling individual and household factors, we find that industrial structure rationalization improves the possibility that migrant workers choose regular employment and reduces the possibility that they would choose another employment status. Moreover, a more reasonable industrial structure can encourage the interprovincial migration of migrant workers. It shows that industrial structure rationalization can alter the job preferences of migrant workers, prompting them to select employment statuses with less income risk and more distant working locations. This effect varies by industry, gender, and family size. A mechanism analysis shows that the channels of the impact of industrial structure rationalization on the employment status and location choices of migrant employees would be different. First, industrial structure rationalization can improve the level of social welfare [15,16]. For migrant workers with lower risk tolerance, regular employment can provide a higher level of social welfare for migrant workers and guarantee the living standard of migrant workers in special circumstances, such as illness, unemployment, and work injury. Secondly, industrial structure rationalization can promote regional economic growth [13,17] and attract more distant migrant workers to local employment with higher income levels.

The main innovations of this study can be summarized as follows. First, existing research mainly focuses on the effect of industrial restructuring on employment structure [3,4]; however, it pays little attention to the micro effects of industrial structure rationalization on the employment decisions of the labor force, particularly the employment choices of migrant workers. The change in the employment preference of migrant workers, especially the change in the preference of employment location, may affect the regional development pattern, as they constitute a large-scale labor group. The continuous inflow of migrant workers can increase the supply of production factors and consumption demand and improve the potential for regional economic development. Our study thus focuses on the employment decisions of migrant workers and evaluates the individual effects of industrial structure rationalization on employment status and employment location choices. Our study enriches research on industrial structure adjustment and employment and provides supporting evidence for studies related to labor mobility and sustainable regional development.

Second, several studies have investigated the individual and family factors affecting career decisions, such as age, education, family size, and social networks [18,19,20,21]. Family inheritance (i.e., sons pick their fathers’ career status) and cultural traditions (i.e., certain locations have particular business traditions) can influence the employment decisions of migrant workers to some extent [22]. Some studies also examine the effect of employment support policies on migrant workers’ employment decisions [23]. However, it does not appear that any studies have considered changes in industrial structure as a factor for analyzing employment decisions from an individual-difference viewpoint so far. This study evaluates the impact of changes in industrial structure on individual employment choices and explores the specific channels of influence of rationalization of industrial structure on different employment decisions, thus filling the gap in research on the influence of industrial factors on the employment choices of workers.

The following sections are organized as follows: Section 2 constructs the theoretical framework; Section 3 presents the model used in the research and introduces the data sources; Section 4 presents the empirical findings; Section 5 investigates the mechanisms and potential heterogeneity; and Section 6 concludes the paper and puts forward useful policy suggestions.

2. Conceptual Framework

2.1. Industrial Structure Rationalization and Employment

Industrial structure rationalization is a dynamic adjustment process in which the inter-industry and intra-industry factor layout within an economy tends to be rationalized according to the existing demand structure and technology level [24,25]. Industrial structure rationalization needs to optimize the factors of production of different industrial sectors in the national economy based on the current market demand, including the rational adjustment of factors of production, such as raw materials, labor, and capital inputs; and the rationalization of the scale of production capacity of primary products, intermediate products, and final products, as well as the structure of revenue distribution [24,26]. In particular, industrial structure rationalization also requires the optimization of the technological structure, which implies the upgrading of production processes, production equipment technology, and the improvement of the technological innovation capacity of enterprises. With the rationalization of industrial structure, advanced technology can be popularized in the industry, and it makes the optimal allocation of resources and simplification of the production process to improve the productivity level. Furthermore, industrial structure rationalization will promote the emergence of new industries and the expansion of existing industries, forming a diversified economic pattern. This diversification reduces dependence on specific industries and mitigates the risks associated with economic fluctuations and external shocks [27].

The impact of industrial structure rationalization on employment is comprehensive. First, industrial structure rationalization leads to changes in the employment structure, which may lead to unemployment of low-skilled laborers, while providing more employment opportunities for high-skilled laborers [2,3,4,5]. However, the emergence of new industries may also absorb those in the unemployed labor force and encourage them to change their skills [6,7,28]. Second, due to productivity growth, technological innovation, and demand for skilled labor, a more rational industrial structure can provide more stable employment opportunities and higher wages [2,8,29]. Stable employment arrangements will attract workers seeking stable and long-term career prospects, promoting regional social stability and sustainable economic development. In addition, higher income levels brought by a more rational industrial structure will attract external labor to flow into the local area due to the widening of the regional income gap. Based on the above analysis, we propose the first hypothesis:

H1: 

Industrial structure rationalization encourages migrant workers to select regular employment and interprovincial migration.

2.2. Industrial Structure Rationalization and Employment Choices of Migrant Workers

Compared to the local workers, industrial structure rationalization can have a greater impact on the employment status and employment location choices of the migrant workers [8,9,10]. First, industrial structure rationalization is often accompanied by improvements in working conditions, access to social benefits, and overall welfare within the industry [15,16]. First, industrial structure rationalization is often accompanied by improvements in working conditions, access to social benefits, and overall welfare within the industry [15,16]. Instead, temporary and self-employed workers are responsible for their social insurance premiums. In China, a majority of migrant workers from rural areas have no assets in the cities where they live, and their wages are spent on household needs, making them less risk-tolerant than local workers in multidimensional and persistent poverty [30,31,32]. In this situation, more stable and comprehensive social benefits can attract migrant workers to choose regular employment rather than self-employment and temporary employment through the increase in social welfare.

Industrial structure rationalization can create new jobs and raise wages [8,29,33]. The primary objective of rural migrant workers who choose to work in cities is to diversify their income sources and lessen the risks connected with agricultural revenue caused by bad weather or natural disasters [34]. The prospect of rising wages and improved living standards motivates migrant workers to continue to migrate to regions with more rationalized industries, making regions with more rationalized industrial structures a more viable option for workers seeking to improve their livelihoods [8]. Therefore, the industrial structure rationalization would prompt migrant workers to choose to migrate across provinces over longer distances rather than within provinces through higher incomes. Based on the above analysis, we propose the second hypothesis:

H2: 

Industrial structure modification encourages migrant workers to select regular employment and interprovincial migration by enhancing social welfare and income.

3. Data and Empirical Strategy

3.1. Data and Variables

Migrant workers’ data are from the China Migrants Dynamic Survey (CMDS) published by the National Health Commission of China, which includes personal and household characteristics, employment position, mobility, and residential intentions. (The data source of CMDS is a public database, which can be found at the following link: “https://www.chinaldrk.org.cn/wjw/#/data/classify/population (accessed on 20 March 2024)”). We selected migrant workers with rural household registration who have lived in the inflow area for more than a month and are aged 15 or older in 31 provinces. Excluding samples that migrated for reasons such as marriage, education and training, military service, or old age, our study sample consists of 633,045 persons who were questioned between 2014 and 2018. The China Statistical Yearbook and China Regional Economic Statistical Yearbook, both issued by the National Bureau of Statistics of China, provide the provincial-level industrial structure statistics.

The dependent variable is migrant workers’ employment choices, including their employment status and location. Initially, employment status is classified as self-employment, regular employment, and temporary employment. Self-employment refers to business ownership or choosing to start a business. Regular employment entails signing an employment contract with an employer and having rights and benefits protected by the Labor Law, typically for a lengthy period with a high degree of stability. The employer must provide social insurance for the regularly employed workers. Temporary employment entails an employment relationship with an employer that is often short-term and less stable than regular employment. Employers do not provide temporary workers with social insurance. These three employment statuses are assigned dummy variables.

Second, we categorize migrant workers’ employment location choices as cross-provincial employment, cross-city employment within a province, and cross-county employment within a city, meaning that a migrant worker chooses to work in a different province from one’s home province, in a different city within the same province as one’s hometown, and in a different county within the same city as one’s hometown, respectively. We also set dummy variables for the selection of employment locations.

Industrial structure rationalization is the independent variable, which is measured by the index of industrial structure rationalization. The rationalization of industrial structure is the inter-industry convergence quality, which indicates the degree of cooperation between industries and the effectiveness of factor resource use [35]. Given that production elements may be freely moved, a higher degree of coupling between input and output structures indicates a more mature stage of industrial development and a higher level of productivity. The Theil Index is a measurement of the gap between industrial development and economic equilibrium; it is therefore an appropriate indication of the degree of industrial structure rationalization. Based on the relationship between industrial structure and economic development and Gan et al. (2011) [36], we use the three-dimensional industry classification method to classify industries. The level of industrial structure rationalization in each province is measured as follows:

$$TL={\displaystyle \sum }_{j=1}^n(\frac{Y_j}{Y})ln(\frac{Y_j}{L_j}/\frac{Y}{L})$$

(1)

where j denotes the industry category and is divided into three industries: the primary industry, which is represented by agriculture and animal husbandry; the secondary industry, which is represented by manufacturing; and the tertiary industry, which is represented by the service industry. Moreover, j is assigned the values 1, 2, and 3, respectively; n denotes the number of industrial sectors; and Y_j and L_j denote the output value and the number of employees in industry j, respectively. When TL equals 0, the industrial structure is the most reasonable.

To ensure that the Theil Index positively reflects the level of industrial structure rationalization, we use the logarithm of the inverse of the Theil Index as an indicator of industrial structure rationalization (ISR). The larger the ISR is, the more reasonable the industrial structure adjusts.

$$ISR=ln(\frac{1}{TL})$$

(2)

Control variables include individual and household characteristics that influence the employment decisions of migrant workers, such as gender, age, education, marital status, and family size.

Table 1. Descriptive statistics.

Variables	Observations	Mean	SD	Maximum	Minimum
Employment status
SE (=1, self-employment)	633,045	0.3816	0.4855	1	0
RE (=1, regular employment)	633,045	0.5313	0.4990	1	0
TE (=1, temporary employment)	633,045	0.0850	0.2789	1	0
Employment location
CPE (=1, cross-provincial employment)	633,045	0.5151	0.4998	1	0
CCEP (=1, cross-city employment within a province)	633,045	0.3097	0.4623	1	0
CCEC (=1, cross-county employment within a city)	633,045	0.1751	0.3801	1	0
Industrial structure rationalization
ISR	31	0.1007	0.1002	0.5227	0.0201
Individual and family characteristics
Gender (=1, male)	633,045	0.5896	0.4919	1	0
Age	633,045	35.401	9.7407	60	15
Education	633,045	9.6393	2.8724	19	0
Marriage (=1, yes)	633,045	0.8186	0.3854	1	0
Family size	633,045	2.8600	1.2330	10	1

shows the descriptive statistics of the main variables. The average age of the sample is 35.4 years old and completed compulsory education in China. Most of them were married and under pressure to raise children and support their parents, which represented the actual composition of migrant workers in China.

3.2. Empirical Strategy

To estimate the impact of industrial structure rationalization on migrant workers’ employment choices, we set the following probit model:

$$Probit\left(E{S}_{ipt}=1\right)={\beta }_0+{\beta }_{1}IS{R}_{pt}+{\beta }_2{X}_i+area+{\lambda }_t+{\epsilon }_{ipt}$$

(3)

$$Probit\left(E{L}_{ipt}=1\right)={\beta }_0+{\beta }_{1}IS{R}_{pt}+{\beta }_2{X}_i+area+{\lambda }_t+{\epsilon }_{ipt}$$

(4)

where β₁ denotes the key coefficient of interest, i represents the individual, p represents the province, and t represents the year. ES_ipt denotes the choice of the employment status of migrant worker i in province p and year t. If migrant workers choose SE, RE, or TE, the corresponding dummy variable ES_ipt = 1; otherwise, ES_ipt = 0. EL_ipt denotes the choice of the employment location of migrant worker i in province p and year t, which is a dummy variable for CPE, CCEP, or CCEC. We set the dummy variables as we do for employment status. ISR_pt denotes the degree of Industrial structure rationalization in province p in year t, and X_i represents the set of control variables. We also control area and ${\lambda }_t$ for region and year-fixed effects.

4. Results

4.1. Main Results

Table 2. Regression results: industrial structure rationalization and employment status of migrant workers.

Variables	SE			RE			TE
	(1)	(2)	(3)	(4)	(5)	(6)	(7)	(8)	(9)
ISR	−0.3027 *** (0.002)	−0.2219 *** (0.003)	−0.0758 *** (0.001)	0.3377 *** (0.002)	0.2376 *** (0.003)	0.0824 *** (0.001)	−0.1109 *** (0.003)	−0.0626 *** (0.004)	−0.0095 *** (0.001)
Gender		0.0458 *** (0.004)	0.0156 *** (0.001)		−0.0963 ** (0.003)	−0.0334 ** (0.001)		0.1265 *** (0.005)	0.0193 *** (0.001)
Age		0.0432 *** (0.001)	0.0147 *** (0.000)		−0.0528 *** (0.001)	−0.0183 *** (0.000)		0.0309 *** (0.002)	0.0047 *** (0.000)
Age squared		−0.0004 *** (0.000)	−0.0002 *** (0.000)		0.0003 *** (0.000)	0.0002 *** (0.000)		−0.0003 *** (0.000)	−0.0001 *** (0.000)
Marriage		0.4151 *** (0.007)	0.1417 *** (0.002)		−0.4313 *** (0.006)	−0.1495 *** (0.002)		0.2107 ** (0.010)	0.0321 *** (0.001)
Education		−0.0371 *** (0.001)	−0.0127 *** (0.000)		0.0428 *** (0.001)	0.0148 *** (0.000)		−0.0145 *** (0.001)	−0.0022 *** (0.000)
Family size		0.1581 *** (0.002)	0.0540 *** (0.001)		−0.1728 *** (0.002)	−0.0599 *** (0.001)		0.0398 *** (0.002)	0.0061 *** (0.000)
Constant	−1.1112 *** (0.006)	−2.3312 *** (0.026)		0.9753 *** (0.005)	2.3670 *** (0.025)		−1.6945 *** (0.009)	−2.4097 *** (0.050)
Area fixed effects		√	√		√	√		√	√
Year fixed effects		√	√		√	√		√	√
Observations	633,045	633,045	633,045	633,045	633,045	633,045	633,045	633,045	633,045
R-squared	0.0246	0.0982	0.0982	0.0308	0.1196	0.1196	0.0042	0.0242	0.0242

Notes: The dependent variables in Columns (1)–(3), Columns (4)–(6), and Columns (7)–(9) are self-employment, regular employment, and temporary employment, respectively. The results in Columns (1), (2), (4), (5), (7), and (8) are the estimated coefficients, and Columns (3), (6), and (9) indicate the estimated marginal effects. The standard errors are shown within parentheses. “√” represents the inclusion of fixed effects in the model. ** p < 0.05, and *** p < 0.01.

shows the estimated results of the impact on employment status. In Columns (1)–(3), higher levels of industrial structure rationalization reduce the possibility of migrant workers’ self-employment. The effect of industrial structure rationalization on the self-employment of migrant workers is depicted in Column (3). The rational industrial structure rationalization decreases the probability of self-employment by 7.58% for migrant workers.

Columns (4)–(6) show a significantly positive effect of industrial structure rationalization on regular employment. In Column (6), industrial structure rationalization increases the possibility that migrant workers select regular employment by 8.24%. Columns (7)–(9) demonstrate the negative impact on temporary employment choices. The probability of a migrant worker opting for temporary employment reduces by 0.95%. These results show that industrial structure rationalization can significantly promote the possibility of migrant workers choosing regular employment and inhibit their choice of self-employment and temporary employment. In other words, industrial structure rationalization can change the employment status choice of the migrant workers, making them more favorable to the long-term stable employment status.

Table 3. Regression results: Industrial structure rationalization and employment locations of migrant workers.

Variables	CPE			CCEP			CCEC
	(1)	(2)	(3)	(4)	(5)	(6)	(7)	(8)	(9)
ISR	0.7113 *** (0.002)	1.1802 *** (0.004)	0.3730 *** (0.001)	−0.4254 *** (0.002)	−0.7058 *** (0.003)	−0.2268 *** (0.001)	−0.5681 *** (0.002)	−0.9320 *** (0.004)	−0.2116 *** (0.001)
Gender		0.0749 *** (0.004)	0.0237 *** (0.001)		−0.0349 *** (0.004)	−0.0112 *** (0.001)		−0.0517 *** (0.004)	−0.0117 *** (0.001)
Age		−0.0046 *** (0.001)	−0.0014 *** (0.000)		0.0023 * (0.001)	0.0007 * (0.000)		0.0037 ** (0.002)	0.0008 ** (0.000)
Age squared		0.0000 ** (0.000)	0.0001 ** (0.000)		−0.0000 (0.000)	−0.0000 (0.000)		−0.0000 (0.000)	−0.0000 (0.000)
Marriage		0.0686 *** (0.006)	0.0217 *** (0.002)		−0.0564 *** (0.006)	−0.0181 *** (0.002)		−0.1910 *** (0.007)	−0.0043 *** (0.002)
Education		−0.0460 *** (0.001)	−0.0145 *** (0.000)		0.0318 *** (0.001)	0.0102 *** (0.000)		0.0206 *** (0.001)	0.0047 *** (0.000)
Family size		−0.0170 *** (0.002)	−0.0054 *** (0.001)		0.0151 *** (0.002)	0.0049 *** (0.001)		−0.0007 (0.002)	−0.0001 (0.000)
Constant	1.9436 *** (0.006)	3.9897 *** (0.026)		−1.6502 *** (0.006)	−3.1556 *** (0.026)		−2.5227 *** (0.007)	−2.5774 *** (0.040)
Area fixed effects		√	√		√	√		√	√
Year fixed effects		√	√		√	√		√	√
Observations	633,045	633,045	633,045	633,045	633,045	633,045	633,045	633,045	633,045
R-squared	0.1163	0.1930	0.1930	0.0438	0.0812	0.0812	0.0714	0.1170	0.1170

Notes: The dependent variables in Columns (1)–(3), Columns (4)–(6), and Columns (7)–(9) are cross-provincial employment, cross-city employment within a province, and cross-county employment within a city, respectively. The results in Columns (1), (2), (4), (5), (7), and (8) are the estimated coefficients, and Columns (3), (6), and (9) indicate the estimated marginal effects. The standard errors are shown within parentheses. “√” represents the inclusion of fixed effects in the model. * p < 0.1, ** p < 0.05, and *** p < 0.01.

shows the effects of industrial structure rationalization on the employment locations of migrant workers. The estimated coefficients in Columns (1)–(3) are significantly positive, demonstrating that industrial structure rationalization can increase the possibility of CPE choices by 37.3%. Significantly negative coefficients in Columns (4)–(9) suggest that a higher level of industrial structure rationalization discourages migrant employees from selecting CCEP or CCEC. The likelihood of migrant workers selecting CCEP and CCEC declined by 22.68 and 21.10 percentage points, respectively. The findings have statistical significance. These results demonstrate that industrial structure rationalization can promote inter-provincial migration of migrant workers and suppress intra-provincial and intra-city migration. We use the level of industrial structure rationalization at the location of the workplace: regions with a more rational industrial structure can attract migrant workers from farther distances to come to local employment. This is consistent with the findings of existing studies [37,38].

All of these findings indicate that industrial structure rationalization can promote the choice of regular employment for migrant workers and attract them to migrate across provinces to those with more rational industrial structures, and Hypothesis H1 is verified.

4.2. Robustness Checks

The main findings demonstrated that industry structure adjustment could affect the choices of migrant workers, but industrial structure changes may also be influenced by labor migration patterns. Referring to Ngai and Pissarides (2007) and De Marco (2017) [38,39], we use regional technological progress as an instrumental variable (IV) to avoid the impact of reverse causality. Technological progress can effectively promote industrial structure rationalization with little influence on the employment options of migrant workers. Current research indicates that technological progress, especially as represented by AI and robots, affects unemployment rates and labor market demand, but technological progress has little effect on the employment status and location decisions of a worker who already has a job and can be used as an effective instrumental variable [2,3,4,5]. We measure regional technical advancement by taking the logarithm of patents granted.

Table 4. Technological progress, industrial structure rationalization, and employment choices.

Variables	Employment Status			Employment Location
	(1)	(2)	(3)	(4)	(5)	(6)
ISR	−0.5310 *** (0.009)	0.7012 *** (0.009)	−0.4073 *** (0.011)	0.8049 *** (0.009)	−0.5156 *** (0.009)	−0.4510 *** (0.012)
Controls	√	√	√	√	√	√
Area fixed effects	√	√	√	√	√	√
Year fixed effects	√	√	√	√	√	√
Observations	633,045	633,045	633,045	633,045	633,045	633,045
R-squared	0.5810	0.5810	0.5810	0.5810	0.5810	0.5810
First stage
Technological progress	0.1699 *** (0.001)
Wald	1495.14 (0.0000)	3364.85 (0.0000)	1124.57 (0.0000)	2341.00 (0.0000)	565.87 (0.0000)	2089.09 (0.0000)
Weak IV	3804.50 (0.0000)	6669.15 (0.0000)	1332.42 (0.0000)	6940.16 (0.0000)	3150.46 (0.0000)	1404.49 (0.0000)

Notes: The dependent variables in Columns (1), (2), and (3) are self-employment, regular employment, and temporary employment, respectively. The dependent variables in Columns (4), (5), and (6) are cross-provincial, cross-city, and cross-county employment within the province, respectively. “√” represents the inclusion of fixed effects in the model. *** p < 0.01.

shows the results of the IV-probit model. The technical advancement has a considerable positive effect on industrial structure rationalization in the first-stage regression and passes the weak IV test. At the second stage of regression, the impact of industrial structure rationalization on the employment decisions of migrant workers remains significant. The estimated coefficients of the IV-probit model are significantly higher than the main results for both employment status and location. So, the main results are reliable.

We also replace the independent variable as a robustness test. We use the industrial structure deviation to indicate the level of industrial structure rationalization, calculated as follows:

$$ISD={\displaystyle \sum }_{j=1}^n\left(\frac{Y_j}{Y}\right)\left|\frac{Y_j/L_j}{Y/L}-1\right|$$

(5)

where a lower ISD represents a more rational industrial structure. For convenience, we use the inverse of the ISD to test and report the estimation results in

Table 5. Deviation in industrial structure and employment choices of migrant workers.

Variables	Employment Status			Employment Location
	(1)	(2)	(3)	(4)	(5)	(6)
ISD	−0.2043 *** (0.003)	0.2218 *** (0.003)	−0.0685 *** (0.004)	1.1524 *** (0.004)	−0.6628 *** (0.003)	−0.9544 *** (0.005)
Controls	√	√	√	√	√	√
Area fixed effects	√	√	√	√	√	√
Year fixed effects	√	√	√	√	√	√
Observations	633,045	633,045	633,045	633,045	633,045	633,045
R-squared	0.0976	0.1192	0.0244	0.1824	0.0751	0.1151

Notes: The dependent variables in Columns (1), (2), and (3) are self-employment, regular employment, and temporary employment, respectively. The dependent variables in Columns (4), (5), and (6) are cross-provincial, cross-city, and cross-county employment within the province, respectively. The standard errors are shown within parentheses. “√” represents the inclusion of fixed effects in the model. *** p < 0.01.

. After replacing the independent variables, the coefficients remain the same and indicate that the influence of industrial structure deviation on the employment decisions of migrant workers is similar to that shown by the main results. The rational adjustment of the industrial structure increases the probability that migrant workers will select regular employment and CPE, which supports the primary results.

In addition, because large cities are more appealing to workers in China, the results may be biased by migrant workers’ preferences for major centers [40]. Thus, we exclude the samples who work in regional center cities and then re-estimate the models. As shown in

Table 6. Industrial structure rationalization and employment choices of migrant workers, excluding regional central cities.

Variables	Employment Status			Employment Location
	(1)	(2)	(3)	(4)	(5)	(6)
ISR	−0.3610 *** (0.008)	0.3972 *** (0.008)	−0.0858 *** (0.011)	1.0702 *** (0.009)	−0.4258 *** (0.009)	−1.0674 *** (0.010)
Controls	√	√	√	√	√	√
Area fixed effects	√	√	√	√	√	√
Year fixed effects	√	√	√	√	√	√
Observations	187,727	187,727	187,727	187,727	187,727	187,727
R-squared	0.0998	0.1292	0.0344	0.1125	0.0191	0.1342

Notes: The dependent variables in Columns (1), (2), and (3) are self-employment, regular employment, and temporary employment, respectively. The dependent variables in Columns (4), (5), and (6) are cross-provincial, cross-city, and cross-county employment within the province, respectively. The standard errors are shown within parentheses. “√” represents the inclusion of fixed effects in the model. *** p < 0.01.

, the findings are similar to the main results. This suggests that the influence of industrial structure rationalization on the employment decisions of migrant workers is not a consequence of big-city preferences and that the main findings are efficient and robust.

5. Mechanisms and Heterogeneity Analysis

5.1. Social Welfare and Employment Status

Industrial structure rationalization increases the level of social welfare, of which social insurance is the most important component [15,16]. In China, social insurance covers unemployment, pension, work injury, medical, and maternity insurance, which can guarantee a basic quality of living and reduce the income risks of workers in the event of incapacity. For regularly employed workers, enterprises and society afford most of the social security expenditures, and individuals need to pay fewer social security expenditures to receive comprehensive and stable social security [30]. However, temporary and self-employed laborers typically do not receive employee benefits from their employers and need to pay for their social insurance, which is often prohibitively expensive for rural migrant workers. Therefore, the increased social benefits brought about by the industrial structure rationalization may become the main channel for migrant workers to choose regular employment. We use the number of social insurances the migrant workers owned as a proxy variable for social welfare, including pension insurance, unemployment insurance, work injury insurance, medical insurance, and maternity insurance.

The results are shown in

Table 7. Industrial structure rationalization, social welfare, and employment status.

Variables	Social Welfare	SE	RE	TE
ISR	0.0942 *** (0.002)	−0.2233 *** (0.004)	0.2516 *** (0.004)	−0.0810 *** (0.005)
Social welfare		−0.2710 *** (0.002)	0.3069 *** (0.002)	−0.1397 *** (0.003)
Controls	√	√	√	√
Area fixed effects	√	√	√	√
Year fixed effects	√	√	√	√
Observations	523,349	523,349	523,349	523,349
R-squared	0.2015	0.1226	0.1606	0.0594

Notes: Standard errors are within parentheses. “√” represents the inclusion of fixed effects in the model. *** p < 0.01.

. Column (1) indicates that industrial structure rationalization improves migrant workers’ social welfare by 9.42%. The estimated coefficients in Columns (2), (3), and (4) indicate that the influence of industrial structure rationalization on self-employment, regular employment, and temporary employment, respectively, decreases when social welfare is taken into consideration. This result suggests that the rational industrial structure could increase social welfare received by migrant employees, thereby encouraging them to choose regular employment over self-employment or temporary work. In conclusion, a higher level of social welfare owing to industrial structure rationalization is a significant factor in the employment status choices of migrant workers.

5.2. Income and Employment Location

Industrial structure rationalization can contribute to urban economic growth by raising corporate profits and employee wages [35,40,41]. The income effect of industrial structure rationalization may be a significant factor in the employment location choices of migrant workers and then increase regions’ attractiveness to migrant workers. Intra-provincial migrant workers typically choose to work in provinces with a higher level of earnings. As a result, the increase in income level brought about by the industrial structure rationalization may become the main channel for migrant workers to choose cross-provincial employment. We use the migrant worker’s monthly income as a proxy variable for the income level to explore potential mechanisms to understand how industrial structural adjustment affects migrant worker’s choice of employment location.

In the first column of

Table 8. Industrial structure rationalization, income, and employment location.

Variables	Income	CPE	CCEP	CCEC
ISR	0.1170 *** (0.001)	1.1668 *** (0.003)	−0.7000 *** (0.003)	−0.9179 *** (0.004)
Income		0.1927 *** (0.003)	−0.0705 *** (0.003)	−0.1720 *** (0.004)
Controls	√	√	√	√
Area fixed effects	√	√	√	√
Year fixed effects	√	√	√	√
Observations	631,181	631,181	631,181	631,181
R-squared	0.1485	0.1979	0.0823	0.1209

Note: Standard errors are within parentheses. “√” represents the inclusion of fixed effects in the model. *** p < 0.01.

, the results indicate that industrial structure rationalization can lead to an approximate 0.12 increase in migrant workers’ monthly income. When both the income and industrial structure are involved in the model, the estimated coefficients in Columns (2), (3), and (4) are less than the main results, demonstrating that industrial structure rationalization can change migrant workers’ employment location choices by raising their incomes. Based on the above analysis, the industrial structure rationalization can promote migrant workers’ choice of long-term employment and cross-provincial employment by increasing social welfare and income levels. Therefore, hypothesis H2 is also confirmed.

5.3. Heterogeneity by Industry, Gender, and Family Size

Industrial structure rationalization has considerably different effects on the employment choices of migrant workers in different industries. In Appendix A 

Table A1. Heterogeneity of employment choices according to industry.

Variables	Employment Status			Employment Location
	SE	RE	TE	CPE	CCEP	CCEC
Agriculture
ISR	−0.1886 *** (0.022)	0.1887 *** (0.024)	0.0545 * (0.032)	1.2657 *** (0.030)	−0.7106 *** (0.026)	−1.0488 *** (0.037)
Observations	19,243	19,243	19,243	19,243	19,243	19,243
R-squared	0.0606	0.1033	0.1080	0.2070	0.1414	0.1026
Mining
ISR	0.3637 *** (0.066)	−0.1944 *** (0.066)	−0.2272 (0.159)	0.6556 *** (0.046)	−0.4093 *** (0.047)	−0.4343 *** (0.050)
Observations	5714	5714	5714	5714	5714	5714
R-squared	0.0339	0.0809	0.1993	0.0843	0.0402	0.0423
Manufacturing
ISR	−0.3027 *** (0.013)	0.2936 *** (0.012)	−0.0739 *** (0.018)	1.4159 *** (0.013)	−0.8597 *** (0.012)	−1.2816 *** (0.019)
Observations	61,649	61,649	61,649	61,649	61,649	61,649
R-squared	0.1499	0.1395	0.0252	0.2367	0.1231	0.1799
Energy Supply
ISR	0.0560 (0.054)	−0.0344 (0.050)	−0.0308 (0.069)	1.0578 *** (0.047)	−0.6945 *** (0.047)	−0.6589 *** (0.049)
Observations	2682	2682	2682	2682	2682	2682
R-squared	0.0964	0.0940	0.0602	0.1651	0.0732	0.0961
Building
ISR	−0.0951 *** (0.011)	0.0896 *** (0.010)	−0.0011 (0.014)	1.2648 *** (0.013)	−0.8112 *** (0.012)	−0.9049 *** (0.015)
Observations	53,054	53,054	53,054	53,054	53,054	53,054
R-squared	0.0223	0.0884	0.2069	0.1991	0.0918	0.1100
Retail
ISR	−0.1096 *** (0.007)	0.1177 *** (0.008)	0.0309 *** (0.009)	0.8058 *** (0.006)	−0.4411 *** (0.006)	−0.7427 *** (0.008)
Observations	148,788	148,788	148,788	148,788	148,788	148,788
R-squared	0.1186	0.1784	0.0855	0.1273	0.0545	0.0776
Transportation
ISR	−0.1712 *** (0.014)	0.1806 *** (0.014)	−0.0631 *** (0.022)	1.4299 *** (0.018)	−0.7465 *** (0.015)	−1.0356 *** (0.019)
Observations	23,914	23,914	23,914	23,914	23,914	23,914
R-squared	0.0612	0.0661	0.0475	0.2797	0.1001	0.1566
Accommodation and Food Service
ISR	−0.1304 *** (0.009)	0.1350 *** (0.009)	−0.0162 (0.012)	1.1418 *** (0.010)	−0.6759 *** (0.009)	−0.8066 *** (0.011)
Observations	90,802	90,802	90,802	90,802	90,802	90,802
R-squared	0.1494	0.1740	0.0474	0.1674	0.0658	0.0859
Telecom and Network Services
ISR	−0.2157 *** (0.027)	0.1875 *** (0.024)	−0.0326 (0.033)	1.3636 *** (0.026)	−0.8863 *** (0.023)	−0.8629 *** (0.027)
Observations	9941	9941	9941	9941	9941	9941
R-squared	0.1420	0.1587	0.0878	0.2988	0.1547	0.1442
Finance
ISR	−0.0705 (0.050)	0.0477 (0.047)	0.0446 (0.065)	1.4163 *** (0.039)	−0.7401 *** (0.032)	−0.8932 *** (0.038)
Observations	4020	4020	4020	4020	4020	4020
R-squared	0.0881	0.1042	0.1681	0.3122	0.1200	0.1456
Real Estate
ISR	−0.0261 (0.027)	0.2021 *** (0.026)	−0.2474 *** (0.030)	1.2910 *** (0.028)	−0.7528 *** (0.024)	−0.9758 *** (0.035)
Observations	10,133	10,133	10,133	10,133	10,133	10,133
R-squared	0.1278	0.2490	0.1710	0.2238	0.0968	0.1151
Renting
ISR	−0.1627 *** (0.039)	0.2043 *** (0.037)	−0.1169 ** (0.047)	1.1054 *** (0.038)	−0.6602 *** (0.037)	−0.8402 *** (0.044)
Observations	3867	3867	3867	3867	3867	3867
R-squared	0.1274	0.1590	0.0747	0.2065	0.0987	0.1204
Consulting Services
ISR	−0.1390 *** (0.053)	0.1111 ** (0.049)	−0.0047 (0.070)	1.5141 *** (0.057)	−0.8816 *** (0.046)	−1.1151 *** (0.065)
Observations	2515	2515	2515	2515	2515	2515
R-squared	0.0897	0.0870	0.0400	0.3356	0.1674	0.1943
Environmental Protection
ISR	−0.0653 (0.067)	0.0750 (0.055)	−0.0444 (0.070)	1.3312 *** (0.056)	−0.8252 *** (0.050)	−0.9010 *** (0.061)
Observations	1989	1989	1989	1989	1989	1989
R-squared	0.0649	0.0628	0.0623	0.2832	0.1289	0.1371
Residential Services
ISR	−0.1725 *** (0.008)	0.1964 *** (0.008)	−0.0560 *** (0.010)	1.1114 *** (0.008)	−0.6214 *** (0.008)	−0.8413 *** (0.010)
Observations	91,809	91,809	91,809	91,809	91,809	91,809
R-squared	0.0884	0.1029	0.0235	0.2061	0.0714	0.1075
Education
ISR	−0.1920 *** (0.039)	0.1646 *** (0.035)	−0.0354 (0.050)	1.3125 *** (0.034)	−0.6252 *** (0.027)	−0.8795 *** (0.033)
Observations	6697	6697	6697	6697	6697	6697
R-squared	0.1139	0.0982	0.0462	0.2860	0.0925	0.1344
Public Health
ISR	−0.2655 *** (0.058)	0.2645 *** (0.049)	−0.1066 * (0.060)	1.3370 *** (0.050)	−0.7355 *** (0.040)	−0.7925 *** (0.054)
Observations	2913	2913	2913	2913	2913	2913
R-squared	0.1503	0.1304	0.1684	0.2715	0.0900	0.1081
Social Work
ISR	−0.2113 ** (0.085)	0.5655 *** (0.074)	−0.5012 *** (0.070)	1.2864 *** (0.083)	−0.5511 *** (0.069)	−1.1689 *** (0.115)
Observations	1321	1321	1321	1321	1321	1321
R-squared	0.0639	0.1894	0.1461	0.2399	0.0969	0.1612
Entertainment
ISR	−0.2554 *** (0.038)	0.2164 *** (0.036)	−0.0081 (0.047)	1.2278 *** (0.036)	−0.6882 *** (0.034)	−0.9716 *** (0.044)
Observations	5179	5179	5179	5179	5179	5179
R-squared	0.1576	0.1702	0.0670	0.2386	0.1044	0.1434
Public Administration
ISR	−0.2300 *** (0.071)	0.2827 *** (0.066)	−0.3809 *** (0.130)	1.2175 *** (0.053)	−0.4815 *** (0.041)	−0.9201 *** (0.051)
Observations	2705	2705	2705	2705	2705	2705
R-squared	0.1164	0.0805	0.1184	0.2588	0.0848	0.1448
Controls	√	√	√	√	√	√
Area fixed effect	√	√	√	√	√	√
Year fixed effect	√	√	√	√	√	√

Note: Standard errors are within parentheses. * p < 0.1, ** p < 0.05, and *** p < 0.01.

, we classify migrant workers into 20 industries. After separating the migrant workers in different industries, the industrial structure rationalization has a significant difference on the choice of employment status and employment location of the migrant workers. On the one hand, the impact of industrial structure rationalization on the choice of employment location of migrant workers is similar across industries. As in the main results, we find that industrial structure rationalization encourages migrant workers to choose CPE in different industries and reduces the probability of employment within the province. This may be attributed to the fact that the regional income gap brought about by industrial structure rationalization is prevalent across industries and will have a similar facilitating effect on the cross-province employment of migrant workers.

However, the effects of industrial structure rationalization on the employment status choices of migrant workers show industry heterogeneity. There are no statistically significant effects of industrial structure rationalization on the employment status choices in industries with a large proportion of state-owned businesses and high skill requirements for employees (e.g., finance, real estate, energy, and environmental protection). However industrial structure rationalization encourages migrant workers to select regular employment in the manufacturing and social service sectors (including social work and social organizations). In some cities, social service enterprises provide their employees with social insurance, such as pensions and medical care, which may be the primary reason why migrant workers in social service industries prefer regular employment. These industries can attract low-skilled migrant workers and provide them with stable employment opportunities and comprehensive social security, thus contributing to regional social stability and sustainable development.

Moreover, industrial structure change has a significantly positive impact on migrant workers’ selection of temporary agricultural employment. This implies that industrial structure rationalization may have unexpected positive effects on the employment and income of migrant workers in agriculture. Agricultural output has a distinct seasonal cycle relative to other industries. Temporary migrant workers might contribute to agricultural production during different seasons and in different places to gain more revenue.

Additionally, we examine heterogeneity in terms of gender and family size. Industrial structure rationalization has a greater impact on female migrant workers than on male migrant workers. And the impacts vary depending on the size of the family. The results are reported in Appendix A 

Table A2. Heterogeneity of employment status choices according to gender and family size.

Variables	Less than 3 Persons			3–5 Persons			More than 5 Persons
	SE	RE	TE	SE	RE	TE	SE	RE	TE
Male:
ISR	−0.2311 *** (0.008)	0.2536 *** (0.008)	−0.1146 *** (0.011)	−0.1886 *** (0.005)	0.2036 *** (0.005)	−0.0393 *** (0.006)	−0.1737 *** (0.032)	0.1748 *** (0.033)	−0.0180 (0.042)
Controls	√	√	√	√	√	√	√	√	√
Area fixed effects	√	√	√	√	√	√	√	√	√
Year fixed effects	√	√	√	√	√	√	√	√	√
Observations	131,846	131,846	131,846	234,473	234,473	234,473	7266	7266	7266
R-squared	0.0998	0.1235	0.0492	0.0395	0.0509	0.0084	0.0344	0.0479	0.0155
Female:
ISR	−0.2474 *** (0.010)	0.2607 *** (0.010)	−0.1033 *** (0.014)	−0.2326 *** (0.006)	0.2412 *** (0.006)	−0.0401 *** (0.009)	−0.3190 *** (0.043)	0.3261 *** (0.044)	0.0097 (0.061)
Controls	√	√	√	√	√	√	√	√	√
Area fixed effects	√	√	√	√	√	√	√	√	√
Year fixed effects	√	√	√	√	√	√	√	√	√
Observations	93,895	93,895	93,895	160,946	160,946	160,946	4619	4619	4619
R-squared	0.1571	0.1736	0.0506	0.0617	0.0731	0.0145	0.0846	0.1113	0.0221

Note: Standard errors are within parentheses. “√” represents the inclusion of fixed effects in the model. *** p < 0.01.

. The results indicate that, for male migrant workers, the impact of industrial structure rationalization on employment status choices diminishes as family size grows. The effect of industrial structure rationalization on female migrant workers’ employment is lowest in 3–5 member families. In these families, women must devote more time to childrearing and then have less time to work outside. In contrast, in large families with five or more members, grandparents can help to share parental responsibilities and care for their grandchildren. Also, larger families require more individuals to work to raise their incomes. Hence, we observe a stronger effect on the employment status decisions of female migrant workers with large families than with smaller families.

Similar to the results for employment status choices, we find gender and family size disparities in the effects of industrial structure rationalization on migrant workers’ selection of employment location. The results in Appendix A 

Table A3. Heterogeneity of employment location choices according to gender and family size.

Variables	Less than 3 Persons			3–5 Persons			More than 5 Persons
	CPE	CCEP	CCEC	CPE	CCEP	CCEC	CPE	CCEP	CCEC
Male:
ISR	1.0448 *** (0.008)	−0.7287 *** (0.008)	−0.7666 *** (0.009)	1.2089 *** (0.006)	−0.6870 *** (0.005)	−0.9850 *** (0.007)	0.9346 *** (0.036)	−0.4933 *** (0.035)	−0.8982 *** (0.047)
Controls	√	√	√	√	√	√	√	√	√
Area fixed effects	√	√	√	√	√	√	√	√	√
Year fixed effects	√	√	√	√	√	√	√	√	√
Observations	131,846	131,846	131,846	234,473	234,473	234,473	7266	7266	7266
R-squared	0.1831	0.0960	0.1073	0.1880	0.0718	0.1179	0.1281	0.0660	0.0920
Female:
ISR	1.1839 *** (0.010)	−0.7450 *** (0.009)	−0.8600 *** (0.011)	1.2288 *** (0.007)	−0.6890 *** (0.006)	−1.0107 *** (0.009)	0.9622 *** (0.048)	−0.5233 *** (0.046)	−0.9073 *** (0.061)
Controls	√	√	√	√	√	√	√	√	√
Area fixed effects	√	√	√	√	√	√	√	√	√
Year fixed effects	√	√	√	√	√	√	√	√	√
Observations	93,895	93,895	93,895	160,946	160,946	160,946	4619	4619	4619
R-squared	0.2312	0.1052	0.1279	0.2048	0.0774	0.1270	0.1437	0.0790	0.0966

Note: Standard errors are within parentheses. “√” represents the inclusion of fixed effects in the model. *** p < 0.01.

indicate that the rational industrial structure has a smaller impact on the employment location decisions of migrant workers from small families with less than three members and large families with more than five members than on those from families with three to five members. That may be due to the fact that smaller families spend less on consumption, and larger families have more sources of income and more parental support. In contrast, migrant workers with 3–5 family members face higher child-rearing costs than those with smaller families and receive less assistance for child-rearing than those with large families. Therefore, migrant workers with 3–5 family members prefer a higher wage by choosing CPE.

6. Conclusions and Discussion

Currently, the spread of new technologies represented by the Internet and artificial intelligence is changing the industrial structure and bringing about changes in the labor market. From the perspective of industrial restructuring, our study examines the micro impact of industrial structure rationalization on the employment choices of migrant workers. By raising employee welfare, industrial structure rationalization promotes migrant workers to select regular employment and reduces the possibility of self-employment and temporary employment. In addition, this study demonstrates that industrial structure rationalization attracts cross-provincial migrant workers and decreases migrant employees’ employment within a province by increasing their earnings. The heterogeneous analysis indicates industry-specific disparities in the impact of industrial structure rationalization on migrant workers’ employment choices. Also, we find that the effect of industrial structure on female migrant workers is greater than that of male migrant employees and varies depending on the size of the family.

Our study focuses on the employment choices of migrant workers and the underlying mechanisms, and it contributes to the literature by giving a realistic perspective to explain the ignored micro effects of industrial structure rationalization on the labor force. Industrial structure rationalization will inevitably bring changes in the labor market. Industrial structure rationalization is an inevitable process of regional industrial development and an important driving force for regional economic growth. Industrial structure rationalization will inevitably bring changes in the labor market. As a vulnerable group in the labor market, migrant workers’ employment preferences and behaviors are often ignored. The existing literature on industrial structure adjustment and the labor market pays little attention to its impact on the migrant workers. And the literature focusing on the employment of the migrant workers pays little attention to analyzing the change in industrial structure as a factor affecting the employment of migrant workers. In China, due to the large size of the migrant workers, their employment choices will affect the regional economic pattern to a large extent. Therefore, this paper explores the effect of industrial structure adjustment on the employment status and employment location of migrant workers from a micro perspective, enriching the research on industrial structure adjustment and the employment of migrant workers.

We also provide research evidence for the government to coordinate the relationship between industrial structure and employment structure to achieve sustainable regional economic development. Industrial structure rationalization is an inevitable process of regional industrial development and an important driving force for regional economic growth. Based on the importance of migrant workers’ job decisions to the regional economy, the government should issue social insurance policies for migrant workers in the process of industrial restructuring, expanding the provision of social welfare to ensure stable employment of migrant workers. At the same time, the government should pay attention to the income level and income stability of the migrant workers to ensure that the migrant workers can receive income in time. Also, the government can also provide additional subsidies for migrant workers experiencing difficulty due to unemployment, work-related injuries, and other reasons. Through a series of policies, the government can attract migrant workers to continue to flow into the region and optimize the structure of local human capital, thus promoting the sustainable development of the region.

In addition, enterprises should adjust their production mode in time to satisfy market demand in the process of rationalization of industrial structure. While expanding the scale of production, enterprises should provide the labor force with a more stable employment environment and higher income. For migrant workers, the rationalization of industrial structure brings new employment opportunities and raises higher requirements for personal skills. Migrant workers should increase the accumulation of human capital and improve their skills. Migrant workers can also seek occupational diversification in the expansion of the service industry in order to achieve sustainable personal development.