The Employment Effect Of Chinese Enterprises Embedded In The Global Environmental Value Chain

: The employment effect of enterprises embedded in global value chains has important 8 theoretical value, but existing research has ignored the impact of cross-border pollution transfer on 9 employment under the division of labor system within the value chain. This study constructs a 10 global environmental value chain (GEVC) analysis framework to combine economic and 11 environmental issues and establishes a theoretical model to discuss the impact of the degree of 12 enterprise embeddedness in the GEVC on employment. Using 2000 – 2006 data from the China 13 Industry Business Performance and China Customs databases, the study finds that an increase in 14 the degree of enterprise embeddedness has a significant inhibitory effect on employment, 15 especially for female laborers, lower-skilled laborers, state-owned enterprises, private enterprises, 16 and enterprises in the eastern region. The research also shows that the cost increase effect 17 enhances the negative effect of increased GEVC embeddedness on employment, while the 18 innovation promotion effect and the foreign direct investment effect serve to mitigate the negative 19 effect. The results provide a reference for developing countries seeking to effectively protect 20 people's livelihood and employment while achieving a leap in the division of labor along the green 21 value chain.


26
For more than 40 years since the reform and opening-up of its economy, China has actively 27 integrated into the division of labor system of the global value chain; owing to its low cost of labor 28 and resource endowments, the country has gr adually become the world's second largest economy 29 and a global manufacturing center. The division of labor system of the global value chain is 30 characterized by value added, which has transformed the division of labor of the traditional 31 supply of labor. This study combines the concepts of ecological footprint and ecological 90 compensation with the most cutting-edge, firm-level global value chain embeddedness 91 measurement method to determine the economic benefits and environmental costs of enterprises 92 integrating into the global value chain, that is, the degree of enterprises' embeddedness in the 93 GEVC. It studies the relationship between GEVC embedding and employment and uses a variety 94 of methods to control for endogeneity. This approach is conducive to a more accurate assessment 95 of the employment effects of enterprises' embedding in the value chain, and provides a reliable 96 reference for developing countries to maintain the stability of the labor market. 97 Second, this study also compares and analyzes the differences in the employment impact of 98 GEVC embedding and traditional export trade, the impact of GEVC embedding on laborers of 99 different genders and with different types of skills, and the embedding of enterprises under 100 different forms of ownership and in different regions. The differences in the impact of the GEVC 101 on the level of employment in these heterogeneous cases provides theoretical support for the 102 precise implementation of the policy of "stabilizing growth and securing employment" in 103 developing countries. 104 Third, this paper examines the impact mechanism of GEVC embedding on employment by 105 analyzing three channels: the cost increase effect, the innovation promotion effect, and the foreign 106 direct investment effect. This helps reveal the in-depth mechanism of the impact of enterprise 107 embeddedness on employment and provides possible path support for improving the relationship 108 between GEVC embedding and employment. 109 The rest of this paper is structured as follows. Section 2 presents the theoretical model. Section 110 3 introduces the index measurement and model setting. Section 4 describes the empirical analysis. 111 Section 5 presents the analysis of the influence mechanism. Section 6 provides concluding 112 remarks and presents implications for policy. 113 114 2. Theoretical model 115 Based on the theoretical analysis framework of Feenstra and Hanson (1996) and Antweiler et al. 116 (2001), this paper builds a labor demand model under the GEVC. In the model, we assume that 117 there are only two countries in an open economy-one is a developed country and the other is a 118 developing country. The developing country uses two production factors, K and L, to produce two 119 products, X and Y, where the marginal returns of K and L are r and w, respectively. The production 120 process of product X can be decomposed into a continuous intermediate product Z, while product Y is independently produced by the developing country for domestic sales and 122 export. Assuming that the developing country mainly produces intermediate product Z, which is 123 pollution-intensive in the production process of product X, and Y is the benchmark unit of pricing, 124 the relative price of intermediate product Z in the domestic market is P. 125 It is assumed that pollutant emissions M (undesirable output) are generated by intermediate 126 product Z during the production process, and product Y is a clean product with no undesirable 127 outputs from its production process. Given the existence of government environmental regulatory 128 policies, the potential outputs are set at a θ ratio of undesirable outputs. In the case that the 129 production function is a Cobb-Douglas function with constant returns to scale, the desirable and 130 undesirable outputs of intermediate product Z are, respectively: 131 It is assumed that the influence function of developing countries' embedding in the GEVC on 138 We assume that the government levies a tax, t, on unit pollution emissions, and after considering 144 the level of clean technology, the government's carbon tax T can be expressed as: 145 It is assumed that is the unit production cost when the potential output is F. 147 Therefore, the optimal production decision for the enterprise embedded in the GEVC is to select 148 the potential output and the optimal pollution emission level so that the unit production cost of Z is 149 The first-order condition of cost minimization can be obtained by constructing a Lagrange 153 function as follows: 154 By dividing Equation (9) and Equation (10), we obtain: 157 Assuming that the market is perfectly competitive, the profit of intermediate product Z is 0: 159 Substituting the first-order condition into Equation (12), the following equation can be solved: 161 Next, by substituting the above equation into Equation (4), we obtain: 163 where DVAR represents the domestic value-added rate of enterprise exports; where EFCi, EFEi, and EFPi represent the carbon, energy, and pollution footprints, respectively. The  GDP as the total amount of ecological compensation for that year. After the above adjustments, the 220 equation of the degree of embeddedness in the GEVC is as follows: 221 where eDVAR represents the export domestic value-added rate of enterprises considering 223 environmental costs, that is, the degree of enterprise embeddedness in the GEVC. 224 (2) Econometric model setting 225 In order to study the impact of GEVC embedding on employment, this study takes the scale 226 of employment as the explained variable and the degree of embeddedness of enterprises in the 227 GEVC as the core explanatory variable to establish the following econometric model: 228 where i refers to the enterprise, t describes the year, and lnlab represents the logarithm of the Therefore, the estimated coefficient of capital intensity is negative. In the actual measurement, the 248 average annual balance of the net fixed assets of the enterprise is used to represent the capital 249 factor, and the number of employees of the enterprise represents the labor factor. The data from China Customs is monthly data, so we first aggregate the monthly data to the 281 annual level, and then, we use the "two-step method" to merge the two databases, following Yu 282 (2015). After that, we filter and sort the merged data to obtain the sample for this study. The 283 statistical description information is shown in Table 1. 284 Insert Table 1 here 285

286
(1) Baseline regression analysis 287 The baseline regression results are shown in Table 2. As shown in Column (1) of Table 2,   Moreover, enterprises with a higher degree of GEVC embeddedness have better and cleaner 300 production technology, as well as a higher level of automation and intelligence, which may also 301 reduce employment demand. 302

303
Insert Table 2 here 304 305 For the comparison with traditional trade, we examine the impact of export intensity (ex) on the 306 scale of employment in Columns (3) and (4) of Table 2. The export intensity index is expressed by 307 the ratio of export delivery value to industrial sales output value. The results show that an increase 308 in export intensity can significantly promote the expansion of the scale of employment, mainly 309 because with increased exports, enterprises must employ more workers to meet the production 310 demand, which brings a positive effect to the labor market. In terms of control variables, the 311 coefficient of total factor productivity is significantly negative, indicating that the negative 312 indicating that the more established and larger the enterprise, the greater the demand for labor, 318 which is also consistent with expectations. 319 (2) Potential endogeneity analysis 320 In order to avoid the interference of endogeneity with respect to the estimation results, this 321 analysis uses two instrumental variables, the lagged term of GEVC embeddedness and the average 322 Insert Table 3  In the initial stage of the development of the division of labor system within the value chain, 346 more employment opportunities are created for women in developing countries. In particular, the 347 processing and manufacturing industries focused on the export of electronic products attract a 348 large number of female workers (Osterreich, 2020). When the degree of embeddedness of 349 enterprises in the GEVC increases, the impact on workers may be heterogeneous based on gender. 350 Therefore, this study further examines the impact of GEVC embeddedness on female and male 351 employees, as shown in Columns (1) and (2) of Table 4. The results show that with the 352 improvement in the degree of GEVC embeddedness, enterprises' demand for female and male 353 employees decreases, and the impact on women is greater than that on men. This may be because 354 women are generally less educated than men in China, and the cost of maternity, labor protection, 355 and special care for female employees increases the financial burden on the enterprise. Therefore, 356 when the degree of GEVC embeddedness of enterprises increases, the female members of the 357 labor force will be more adversely affected. higher-skilled employees are less exposed to unemployment risk. 371

372
Insert Table 4   impact of enterprises' level of GEVC embeddedness on employment. Therefore, this study divides 393 the total sample into two regional groupings, the eastern region and the combined central and 394 western regions, to analyze the impact of GEVC embeddedness on employment. The results in 395 Columns (4) and (5) of Table 5 show that an increase in enterprises' degree of embeddedness in 396 the GEVC has a significant negative impact on employment in the eastern region, while the 397 impact on employment in the central and western regions is not significant. This may be because 398 the level of environmental regulation and labor costs in the eastern region are relatively high. 399 When the degree of embeddedness in the GEVC increases, investments in pollution control further 400 squeeze the labor input factors and reduce the demand for labor. 401

Analysis of the influencing mechanism 405
The baseline regression results show that the increase in the degree of embeddedness of 406 enterprises in the GEVC has a significant negative impact on the scale of employment, while the 407 heterogeneity test results show that the impacts vary widely according to employee gender and 408 skill level as well as enterprise ownership type and regional location. Together, these results 409 indicate that the effect on employment will be influenced through three channels: the cost increase 410 effect, the innovation promotion effect, and the foreign direct investment effect. Therefore, we 411 establish the following model to test the influence mechanism: 412  Foreign direct investment is expressed by the share of foreign capital in the paid-in capital of the 421 enterprise. Table 6 reports the test results of the influence mechanism. 422 423 Insert Table 6 here 424 425 According to the results in Column (1) of Table 6 Based on the perspective of environmental cost, this paper explores the relationship between 508 GVC embeddedness and labor employment and its influence mechanism at the micro level.