Impact of participating in global value chain on the carbon dioxide emissions of 1 China's equipment manufacturing industry 2

6 As the pillar industry in China's post-industrial era, the equipment manufacturing industry has played 7 an important role of providing technical equipment for downstream industries, which also brought 8 about a substantial increase in CO 2 emissions. Therefore, in order to find ways to reduce the carbon 9 dioxide emissions of the equipment manufacturing industry, this paper based on the global value chain 10 production length decomposition model, improved the CO 2 emission effect model and the STIRPAT 11 model to study the different impact of the GVC production length on the CO 2 emissions of China's 12 equipment manufacturing industry under different GVC participation modes. The study found that 13 extending GVC production length can effectively reduce CO 2 emissions, and the CO 2 reduction effect 14 of the simple GVC production length is the most significant. Besides, with the extension of the GVC 15 production length, the CO 2 emissions of high-tech industries have decreased, while the CO 2 emissions 16 of medium-technology industries have increased. In addition, the improvements of policy regulations, 17 factor structure and foreign investment will also reduce CO 2 emissions, but the expansion of production 18 scale and R&D investment will increase CO 2 emissions.


23
In the process of China's industrialization, China's equipment manufacturing industry assumes an 24 important function of providing equipment and technical means for downstream industries (Liu and 25 Zhu 2019), participating in the Global Value Chain (GVC) has greatly improved the level of production 26 technology and production efficiency, which has brought huge economic profits and technical 27 experience returns. But at the same time, because China's equipment manufacturing industry has a low 28 degree of participation in GVC and is a resource-consuming industry (Wang et al. 2021), it has long 29 become a major CO 2 emitter in China (Guy et al. 2020). Since 2000, its CO 2 emissions have accounted 30 for about 9% of China's total CO 2 emissions, and is about 10 times the level of CO 2 emissions in 31 industrialized countries (such as the United Kingdom, Germany, etc.), which is much higher than the 32 average level of CO 2 emissions of the global equipment manufacturing industry. Therefore, in the 33 context of China's accelerated implementation of the "1+X" planning system of "Made in China 2025", 34 the equipment manufacturing industry must take the "green" path of independent innovation and 35 sustainable development to achieve the transformation of high-quality and low-CO 2 production. Hereto, 36 what impact will the deepening of participation in GVC have on the CO 2 emissions of China's 37 equipment manufacturing industry? Will the different modes of participating in GVC have different 38 impact? What other factors also affect the CO 2 emission of China's equipment manufacturing industry? 39 The solution of the above problems will provide strong theoretical support for China's equipment 40 manufacturing industry to find a path of low-CO 2 development from the perspective of GVC. 41 To answer the above questions, the remainder of this paper is structured as follows: The "Literature 42 review" section briefly reviews the current literature. The and GVC production length. The research conclusions mainly include the following three points. 57 First of all, participating in GVC will increase the CO 2 emissions of equipment manufacturing 58 industry. The reason is that the low GVC participating degree of China's manufacturing industry is 59 harmful for energy-saving and CO 2 -reduction. The expansion of low-end production activities has 60 promoted the increase of CO 2 emissions. Based on GVC position index, although China's 61 manufacturing industry has improved trade competitiveness and basically shows an upward trend in 62 GVC (Wei and Zhang 2020), it is difficult for China's equipment manufacturing industry to escape the 63 development dilemma brought by the "low-end lock-in" of GVC (Chen and Wang 2015), which have 64 aggravated the pollution problem caused by CO 2 emissions (Sun and Du 2020), the phenomenon is 65 particularly obvious in capital and technology-intensive industries (Wang 2014). Moreover, in terms of 66 the GVC participation index, China's equipment manufacturing industry has a very high degree of 67 "backward participation" in the GVC (Pan 2019), which requires more energy and resource input (Zhao 68 et al. 2020), and makes CO 2 emissions increase (Chang et al. 2020). Meanwhile on the basis of the 69 GVC production length, after the extension of the GVC production length and the expansion of the 70 scale of processing trade, the CO 2 emissions generated by the manufacturing industry will also increase 71 before the widespread application of cleaner production technologies (Zhao and Yang 2020). The low 72 participating degree of China's manufacturing industry reflects the characteristics of weak technology 73 (Li and Yuan 2016), poor factor structure (Lu et al. 2018), and strong mass production demand (Kang 74 2018; Edger 2020), which are also important factors to promote the increase of CO 2 emissions (Xie et 75 al. 2018). 76 Secondly, participating in GVC will decrease the CO 2 emissions of equipment manufacturing 77 expanded. First, the measurement methods for the degree of participating of GVC mainly stay in the 121 two aspects of GVC position index and GVC participation index, it is impossible to comprehensively 122 and scientifically describe the participating degree of GVC (Yuan and Qi et al. 2019). Secondly, it is 123 still not clear enough of the CO 2 emission effect model of GVC, and the internal links between CO 2 124 emissions and related indicators need to be clarified. Lastly, few studies have explored the specific 125 impact of participating in GVC on CO 2 emissions from the perspective of sub-sector of equipment 126 manufacturing industry. 127 Based on this, our study contributes to the previous literature in the following three aspects: (1) 128 According to the global input-output table, the GVC production length is decomposed from the 129 perspectives of the destination and the source of the value-added. And we distinguish the different 130 concepts of the GVC production length, GVC simple production length, the GVC production length 131 returned to the exporting country and the pure foreign GVC production length, which will help to 132 comprehensively describe the situation of the equipment manufacturing industry in GVC and analyze 133 the different results caused by different participating modes. (2) We update the analysis of 134 environmental pollution and supply models, establish a CO 2 emission effect model of the GVC, obtain 135 relevant economic indicators affecting CO 2 emissions. (3) We apply the CO 2 emission effect model of 136 GVC to a specific industry level, and deeply study the relationship between the GVC participation 137 degree and the CO 2 emissions of the equipment manufacturing industry, which will help to find the 138 effective measures to achieve CO 2 emission reduction targets while deeply participating in the GVC. 139 (1) There are only two industries in the world, namely industry 1 and industry 2, of which industry 1 146 is a high-CO 2 industry and industry 2 is other industries. Then the world only produces two products, 147 that is, industry 1 produces product X and industry 2 produces product Y. In addition, the production 148 process of the two types of products obeys the principle of constant return to scale. 149 (2) Product X is a high-CO 2 product, that is, the production of product X will discharge a large 150 amount of pollutants. Y is a low-CO 2 product, that is, the production of Y product does not emit any 151 pollutants. 152 (3) Pollutants only consider CO 2 emissions and ignore other environmental effects. 153 (4) The production only need two factors, they are labor (L) and capital (K). 154 Suppose the production function of potential output in the economy is: 157 In Equation (1), F is the production function, S is the total output of the industry, K is the capital 158 input, and L is the labor input. Since the production of products will be affected by internal and external 159 elements, the actual output will be lower than the potential output. Therefore, assume that the internal 160 element that affects production is only the factor input ratio (G), that is, the ratio of capital input (K) to 161 labor input (L); the external element that affects production is only the government's regulation of 162 reducing CO 2 emissions, in this case, r represents the rate of decrease of output. Since the input of 163 reducing CO 2 emissions will inhibit the increase of CO 2 emissions, the level of CO 2 emissions (CO 2 164 emissions per unit of output) can be expressed as decreasing function of r; the reciprocal form of production technology level (T) represents the 166 inhibitory effect of technological improvements on CO 2 emissions; and Then the actual production function of product X is: 168 (2) The CO 2 emissions during the production of product is: 169 Since industry 1 participates in the international division of production when producing X, the effect 170 of participating in GVC on production is ) (V  . At this time, the actual production function of product 171 X is: 172 Incorporating formula 3 into formula 4, the relationship between actual production level of product X 173 and CO 2 emissions can be obtained: 174 Since CO 2 emissions will cause negative externalization to the society, corresponding opportunity 175 costs must be paid, so the tax rate for CO 2 emissions is set to γ. According to the principle of 176 minimizing the cost of enterprises, under normal circumstances, enterprises will choose the optimal 177 arrangement of potential output and CO 2 emission levels to achieve the lowest production cost of 178 product. So we can construct the following function: 179 is the unit production cost of the potential output of product X, and is the production cost of capital and labor respectively. 181 By constructing a Lagrangian function, we can obtain the derivation of CO 2 emission C and output 182 Among them, θ is the Lagrangian multiplier. Then divide the two formulas in Equation (7) to 184 obtain the cost minimization conditions for the production of X products by the enterprise. 185 Under perfectly competitive market conditions, the result of market competition is in line with 186 Pareto optima. Then the net profit of the production of X product must be zero, so the profit function of 187 X product is set as , where X P is the relative price of X product relative to Y 188 product, and the price of Y product is defined as 1, we can obtain: 189 Then, the CO 2 emission level is: 191 The CO 2 emission function in Equation (3) can be rewritten as: 192 Equation (12) is the decomposition model of the CO 2 emission effect of product X participating in 193 the GVC. After taking the logarithm of both sides, we obtain: 194 is a constant term. As shown in Equation (13), the sign of production scale 195 (S) is positive, which means that as the production scale expands, CO 2 emissions will increase; the sign 196 of technical level (T), factor structure (G), and policy regulations (r) is negative, which means that CO 2 197 emissions will be reduced due to the improvement of technology, factor structure and policy 198 regulations; and it is expected that the increase of the GVC participating level will also have a negative 199 effect on CO 2 emissions. Divide the world into three parts: country A, country B and other countries (R). Each country has two 208 industrial sectors: Industry 1 and Industry 2. Then, the world input-output table will be reflected in 209 Matrix Z represents the intermediate inputs produced in one country and used in another country; 213 vector Y represents the final product produced in one country and used in another country; vector X 214 represents the total output of one country; vector Va represents one Country's direct value added. 215 Suppose the input coefficient matrix is X represents the diagonal matrix of X, at this time, 216 . And the total output X can be expressed as: 217 Y represents the sum of final products used in a country from other countries, D A represents the 218 domestic input coefficient, D Y represents the total domestic final products consumed by each country, 219 F A represents the import input coefficient, F Y represents the sum of final products exported, and E 220 represents total exports. According to the Leontief inverse matrix (B), we can rewrite Equation (14): 221 Among them, represents the domestic Leontief inverse matrix. Based on this, the 222 relationship between the value-added and the final product in Table 1 is: 223 It can be seen that the initial input (value-added) of an industry can only be absorbed by the final 224 product of the same industry. Therefore, the equation for the production process involved in the 225 value-added can be summarized as follows: 226 Vˆ matrix represents the sum of value-added in all production stages, each element of which 227 represents the value-added from an industry in one country, and the value-added is directly or indirectly 228 used by an industry in another country to produce final products. 229 Take the production length of each stage as the weight and add it up to get the total output of a 230 specific industrial department, we obtain: 231 Therefore, the average production length of the value-added in the final product is: 232 The average production length based on the forward industry linkage is: 233 Equation (20) measures the amount of supplementary value-added per unit of industry once, in 234 which the value-added of each industry can be seen as a whole. At this time, the longer the forward 235 production length is, the more downstream production stages the value-added participates in as a 236 substitute, and the higher its upstream production position is. 237 The production length based on the backward industry linkage is: 238 Equation (21) measures the total value-added input of final product in a specific industry. At this 239 time, the longer the backward production length is, the more upstream production stages of a particular 240 final product has, the lower the downstream production position of the product is. 241 According to the decomposition framework of value-added and final products proposed by Zhi et al. 242 (2017b), the production activities of a country can be broken down into 5 parts according to the 243 different situation of cross-border production activities: 244 In Equation (22), (1)The first part is the domestic value-added, which refers to the part of 245 domestically produced goods that are ultimately consumed domestically, represented by V_D.
(2) Part 246 2 represents the Ricardo trade part, that is, the final products exporting to foreign countries are directly 247 consumed, which is only cross-border once, denoted by V_RT. (3)The part representing cross-border 248 production activities is divided into simple cross-border production activities and complex cross-border 249 production activities. Simple cross-border production activity refers to the part of the intermediate 250 product produced in one country and directly used by the importing country for production and 251 consumption, the production activity is only cross-border once, denoted as V_GVC_S. Complex 252 cross-border production activity refers to the part of intermediate goods produced in one country that 253 are used by the importing country for production and exported to a third country, denoted as V_GVC_C. 254 (4) The complex cross-border production activities involve two categories according to whether they 255 return to the exporting country. Among them, 4a is the part returned to the exporting country and 256 absorbed by the exporting country, denoted by V_GVC_D. 4b is the part that is indirectly absorbed by 257 the importing country and exported to other trading partner countries after being processed, denoted by 258

V_GVC_F. 259
According to this, the GVC production length in Equation 19 is divided into five parts: 260 Similarly, as shown in Equations 24 and 25, the forward production length (PLv) and the backward 261 production length (PLy) can also be divided into five parts. Among them, the part that participates in 262 GVC activities include the GVC production length (PLv_GVC, PLy_GVC), the simple GVC production 263 length (PLv_GVC_S, PLy_GVC_S), the GVC production length returning to the exporting country 264 (PLv_GVC_D, PLy_GVC_D) and the pure foreign GVC production length (PLv_GVC_F, 265 PLy_GVC_F). The decomposition model of production length is shown in Fig. 1. 266  In Equation (29), the explained variable C represents the CO 2 emissions of each sub-industry of 291 China's equipment manufacturing industry, which is a substitute for the environmental pressure in the 292 IPAT model; V represents the core explanatory variables related to the GVC, including forward and 293 backward GVC production length (PLv_GVC, PLy_GVC) and its decomposed parts; Policy represents 294 the policy regulation, which is expressed by the amount of industrial pollution control investment based 295 on the method of Peng and Li (2013),and the weight is the ratio of the total investment in fixed assets 296 of equipment manufacturing industry to China's total investment in fixed assets, then the industrial 297 pollution control investment is calculated according to China's total pollution control investment; Scale 298 represents the production scale, instead of the population size in the original IPAT model, it is measured 299 by per capita output value; G_factor is the factor structure, replacing the per capita assets in the 300 original IPAT model, expressed by the ratio of the total fixed assets of industrial enterprises above 301  represents a constant term; 1  to 7  represent the coefficient of each 307 variables, ε represents a random disturbance term, i represents an industry, and t represents time. The results of the forward and backward GVC production length of China's equipment manufacturing 322 industry is shown in Fig. 2 and Fig. 3. China's equipment manufacturing industry has become deeply 323 participating in GVC, and the change trend of forward and backward GVC production length is similar, 324 but the backward GVC production length is always longer than the forward GVC production length. In 325 2001, the production length of the forward and backward GVC production length increased rapidly, 326 which thanks to the tremendous progress that China made after joining the WTO in 2000, the 327 convenience of participating in the international division of production has been promoted, and the 328 technology spillovers from developed countries has increased. After a difficult growth process, the 329 GVC participating level of China's equipment manufacturing industry achieved a major leap again in 330 2009. Comparing the GVC production length under different participating modes, the largest increase 331 part is the GVC production length returning to the exporting country, followed by the pure foreign 332 GVC production length. The variation of the GVC production length and the simple GVC production 333 length is almost the same. It shows that the impetus provided by participating in GVC is far greater 334 than that of China's independent research and development. 335    that the fixed effect model is better than the random effect model. Based on this, the empirical 376 estimation results are as follows. 377 The first three columns in Table 3  The fixed effect model is more accurate in estimating the indicators in this article. After continuing to 383 add annual dummy variables in column (3), the impact of policy regulation and foreign investment on 384 CO 2 emissions has become significant, and the coefficient of determination R 2 has increased to 0.853, 385 indicating a 9.6% change in CO 2 emissions of the equipment manufacturing industry can be explained 386 by missing variables that change with time but not with industry. 387 388 Note: The robust standard errors in parentheses, ***, **, and * indicate significant at the level of 1%, 5%, and 10% respectively.

389
With the acceleration of globalization, the impact of participating in GVC on CO 2 emissions cannot 390 be ignored. Since production length variables change with time and individual differences, columns (4) 391 to (7) in Table 3 is based on column (3) and added variables reflecting the production length of China's 392 equipment manufacturing industry to test the specific impact of GVC production length on CO 2 393 emissions. Except for the forward production length, the impact of the backward production length, the 394 forward GVC production length and the backward GVC production length on the CO 2 emissions of the 395 equipment manufacturing industry have all passed the test at a significant level of 1%, and the 396 coefficients of the three are all negative, which means that for every 1% increase in the backward 397 production length, the forward GVC production length and the backward GVC production length, the 398 level of CO 2 emissions will drop by about 5.3%, 3.7%, and 5.5%, respectively. The CO 2 emission 399 reduction effect of the GVC is obvious, which further supports the derivation in the previous model. 400 Column (6) shows that the extension of the forward GVC production length can effectively reduce CO 2 401 emission, indicating that the equipment manufacturing industry has achieved technological 402 improvement through imitation, learning and secondary innovation in the process of moving upstream 403 to the international division of production. But this reduction is still not enough to drive the overall 404 transformation of the CO 2 emissions of intermediate products in the equipment manufacturing industry, 405 making the CO 2 emission reduction effect from the perspective of forward production length not 406 obvious, which also shows that the production technology of industrial intermediate products of 407 China's domestic equipment manufacturing industry is still not environmentally friendly, and the 408 improvement of clean technology in the domestic equipment manufacturing industry has stuck in a 409 "bottleneck period". It can be seen from columns (5) and (7) that the CO 2 emission reduction effect of 410 the backward production length and the backward GVC production length is similar, both are greater 411 than the result of the forward GVC production length. It proves that while the participation of China's 412 equipment manufacturing industry in GVC activities has increased, the clean level of final product 413 production technology has been greatly improved, and it has gradually moved from a low-tech, 414 high-CO 2 production stage to a high-tech, low-CO 2 production stage. 415 According to the data in each column in Table 4, whether it is the forward GVC production length or 416 the backward GVC production length, the extension of the simple GVC production length has the best 417 reduction effect on CO 2 emissions, both are significantly negative at the 1% level. However, the length 418 of complex GVC production length has little effect on CO 2 emissions, and only the backward pure 419 foreign GVC production length has passed the test and has a positive effect on CO 2 emissions. This  Note: The robust standard errors in parentheses, ***, **, and * indicate significant at the level of 1%, 5%, and 10% respectively.

The impact of the relevant economic indicators on CO2 emissions 441
Observing the data in Tables 3 and Table 4, we can see that policy regulations, factor structure, and 442 foreign investment have a negative impact on CO 2 emissions, and scale effects and R&D investment 443 will promote the increase of CO 2 emissions. In addition to R&D investment, the effects of other 444 indicators on CO 2 emissions are in line with expected results. The specific analysis is as follows: (1) 445 The coefficient of policy regulation is maintained at around -0.4, and the promotion of CO 2 emission 446 reduction is not obvious. This is related to China's industrialization development stage during 447 2000-2011, and environmental regulation did not take effect until it is over.
(2) The production scale 448 has an increasing effect on CO 2 emissions, because in the process of joining the international division 449 of production, China's equipment manufacturing industry has undertaken the transfer of high-CO 2 450 emission industries from developed countries, and production is mainly based on high-energy and 451 high-polluting activities. The expansion of production scale will lead to an increase in CO 2 emissions, 452 which is consistent with the reality. (3) The factor structure has a restraining effect on CO 2 emissions. 453 The factor structure of China's equipment manufacturing industry is changing from labor-intensive to 454 capital-intensive, and it is still in the process of moving towards technology-intensive. The prospects 455 for reducing CO 2 emissions through the adjustment of the factor structure are great. (4) There is a 456 significant negative correlation between foreign investment and CO 2 emissions, indicating that the 457 clean technology learned from the investing country can inhibit CO 2 emissions with the spillover effect 458 of FDI technology. (5) The effect of R&D investment on CO 2 emissions is positive and insignificant, 459 which is consistent with the results of Wang et al. (2015). The reason is that, on the one hand, because 460 the current Chinese enterprises cannot effectively allocate R&D resources, the actual investment in 461 clean technology is much lower than expected; on the other hand, it is because the current level of CO 2 462 emission reduction technology of the equipment manufacturing industry is extremely low. 463 464

Robustness check 465
For the purpose of further examine the robustness of the empirical results, this paper removes 5% of 466 the extreme values from both ends, and performs regression test on the sub-samples to eliminate the 467 influence of non-randomness on the regression results (   Note: The robust standard errors in parentheses, ***, **, and * indicate significant at the level of 1%, 5%, and 10% respectively.  Table 6  481   and Table 7. 482 483 Note: The robust standard errors in parentheses, ***, **, and * indicate significant at the level of 1%, 5%, and 10% respectively.

488
The first four columns of Table 6 show the estimated results of the impact of forward GVC 489 production length of the high-tech industry on CO 2 emissions, which is negative at a significant level of 490 5%; The first four columns in Table 7 indicate that the CO 2 emissions of high-tech industries are 491 subsequently reduced as the backward GVC production length is extended. The overall effect of the 492 backward GVC production length on CO 2 emissions is better than the forward GVC production length. 493 The main reason for this phenomenon is that the production of high-tech industries in China's 494 equipment manufacturing industry is mainly to provide high-level intermediate products to other 495 countries. In this process, the level of production increases with deeply participating in GVC activities, 496 and thus makes CO 2 pollution in the production process continues to decrease. Due to the shortage of 497 labor resources and the increase of basic production costs in China, the simple processing and 498 production part of China's equipment manufacturing high-tech industry has begun to move to other 499 developing countries, resulting in a stronger CO 2 emission reduction effect caused by extending the 500 length of backward GVC production length. 501 The last four columns in Table 6 show that the CO 2 emissions of the medium-tech industry are 502 affected by the extension of the forward GVC production length, that is, for every 1% increase of the 503 forward GVC production length, CO 2 emissions will increase by about 1%. This is because the 504 medium-technology industry in China's equipment manufacturing industry is still dominated by 505 labor-intensive production. In the process of participating in the GVC, it has not completely separated 506 from the low value-added and high CO 2 -emission production stage, and the CO 2 emissions level of 507 production is relatively high. The average growth rate of CO 2 emissions from the transportation 508 equipment manufacturing industry in 15 years was 8.36%, the CO 2 emissions increased by 458.57 509 million tons, and the average annual growth rate of the fabricated metal products industry reached 510 6.99%. The coefficients of the backward GVC production length related indicators in the last four 511 columns of Table 7 are all positive, indicating that the backward GVC production length also has a 512 driving effect on the increase of CO 2 emissions in the medium-tech industry. The reason is that the 513 medium-tech industry in China's equipment manufacturing industry has a low position in the 514 international division of production. This is because the fabricated metal product industry and 515 transportation equipment manufacturing industry have higher requirements for precision parts, and the 516 core technology manufacturing capabilities of China's equipment manufacturing industry are still weak, 517 the change from basic core components, basic core technology and basic core materials to high-tech, 518 high-end products and high-end components is very slow. China's medium-tech equipment 519 manufacturing industry mainly provides final products to other countries in the form of OEM (Original 520 Equipment Manufacturer). Therefore, the GVC participation mode based on backward linkage will 521 generate more CO 2 . 522 It is worth noting that the regression results of the high-tech industries are in the same direction as 523 the overall regression results, and will reduce CO 2 emission, but the results of the medium-tech 524 industries are opposite to the overall results. This may be because high-tech industries deepen the 525 participation in the GVC by improving the level of research and development, while the increased 526 participation in the GVC of the medium-tech manufacturing industry is at the expense of producing 527 more resource-intensive products. China is committed to reducing resource-intensive production and 528 encouraging high-tech R&D production activities, so that the CO 2 emission reduction effect of 529 high-tech industries is stronger than the CO 2 promotion effect of medium-tech industries, which in the 530 end will reduce the CO 2 emissions of the whole equipment manufacturing industry. 531 532

533
Based on the decomposition framework of the GVC production length and the model of the impact of 534 participating in GVC on CO 2 emissions, this paper derives the core indicators and constructs a 535 STIRPAT model of the impact of participating in GVC on CO 2 emissions, clarifies the specific impact 536 of different GVC participating modes on the CO 2 emissions of the equipment manufacturing industry, 537 and analyzes the industry heterogeneity of this impact. The main conclusions of the study are as 538 follows: 539 First of all, the extension of the forward production length of the GVC can effectively reduce CO 2 540 emissions. The extension of the forward simple GVC production length has the best effect, and the 541 forward complex GVC production length has no effect on CO 2 emissions; the CO 2 emission reduction 542 effect of the backward GVC production length and the backward simple GVC production length is 543 significant, which is better than the result of the forward GVC production length, and the extension of 544 the pure foreign GVC production length also has a slight CO 2 emission reduction effect. It shows that 545 the improvement of cleaner production technology in China's equipment manufacturing industry at the 546 current stage mainly stays at the simple production stage of GVC, only reducing CO 2 emissions in the 547 processing and assembly links. Hence, the cleaner production technology of complex GVC production 548 activities needs to be improved urgently. 549 Secondly, for high-tech industries, the extension of the forward and backward GVC production 550 length will reduce CO 2 emissions; while the extension of the forward and backward GVC production 551 lengths of the medium-tech industry will increase CO 2 emissions. This shows that the level of cleaner 552 production in China's high-tech industries is increasing with the deepening of the participating degree 553 of the GVC; however, the production of the medium-tech industries still relies on basic advantages 554 such as abundant labor resources, and has been locked in the low-end link of the GVC. In addition, the 555 high-tech industry has developed more vigorously, driving the overall CO 2 emission reduction trend of 556 the equipment manufacturing industry to improve. Therefore, China should implement an industry 557 differentiation policy, improve the overall competitiveness of high-tech industries, and promote the 558 realization of qualitative changes in low-tech industries (Chenyao et al. 2020). 559 Thirdly, policy regulations, factor structure and foreign investment can effectively reduce CO 2 560 emissions, but the expansion of production scale and R&D investment will increase CO 2 emissions. It 561 shows that, in recent years, China's improvement in environmental regulations, the adjustment of factor 562 structure and the introduction of foreign capital have brought positive CO 2 emission reduction effects, 563 but the problem of high-CO 2 activities in the export intermediates production and inefficient use of 564 R&D funds still exists. 565 Based on the above conclusions, the following policy implications are proposed: 566 Firstly, continue to deepen the degree of participating in GVC and move out of low-end production 567 activities. In the context of participating in GVC, the extension of the GVC production length will 568 bring great potential for CO 2 reduction worldwide, especially in manufacturing sector (Rilong et al. 569 2020). It provides strong evidence for China's unswerving participation in the international division of 570 labor and adherence to opening up. Therefore, China's equipment manufacturing industry should 571 actively respond to the "Belt and Road" initiative, cooperate with countries along the "Belt and Road" 572 in production activities, and undertake more high-value-added, low-CO 2 production activities from 573 developed countries. Transfer low-end production activities to other developing countries where 574 resources and labor are cheaper. After that, China's equipment manufacturing industry can further 575 extend the GVC production length, and be involved in the high-end production link of the GVC totally 576 (Chenyao et al. 2020). 577 Secondly, maintain the advantages of intermediate production in simple GVC activities and improve 578 the clean production technology level of complex GVC activities. At present, China's equipment 579 manufacturing industry has made great progress in simple GVC production and has reached the 580 requirements of cleaner production, but it still needs to improve the CO 2 emission reduction effect of 581 complex GVC activities. Accordingly, on the one hand, it is necessary to optimize the import quality of 582 intermediate products through learning the manufacturing technology and processing technology of 583 high-tech intermediate products, improve the level of intermediate products exported in complex GVC 584 activities, and extend the GVC production length returning to the exporting countries. On the other 585 hand, to encourage equipment manufacturing enterprises to "go global" means enterprises need to 586 conduct in-depth cooperation with multinational companies in R&D, design, brand building, etc., seek 587 new path of participating in GVC with technological innovation to get out of the dilemma of "low-end 588 lock-in" and extend the length of pure foreign GVC production length. 589 Thirdly, keep the clean production advantages of high-tech industries, accelerate the transformation 590 and upgrading of medium-tech industries, and enable the equipment manufacturing industry to achieve 591 the CO 2 emissions reduction of the entire industry. For high-tech industries, while vigilant against the 592 implementation of restrictions by countries with high-income, we should strive to achieve more 593 advanced technological breakthroughs, seize the strategic position of high-end production links, and 594 steadily move to the top of the GVC. The medium-tech industry needs to expand the production scale 595 of high value-added intermediate products through the extensive introduction of advanced low-CO 2 596 production technologies, reduce dependence on the export of pollution-intensive intermediate products, 597 and gradually transform from the high-CO 2 GVC participation channels to high-tech channels. By this 598 way, the entire equipment manufacturing industry will achieve CO 2 emission reductions eventually. 599 Last but not least, continue to strengthen environmental control and foreign investment, improve the 600 factor structure and R&D expenditure utilization. The factor structure can be transformed to technology-intensive by increasing the skilled labor and 608 R&D personnel. (4) The government need to strictly supervise the destination of R&D expenditures, 609 allocate R&D expenditures reasonably, guide enterprises to use R&D expenditures efficiently, and 610 promote enterprises to increase investment in independent innovation, so that the enterprises of 611 equipment manufacturing industry could climb to the higher level of participating in GVC through its 612 own capabilities. 613 614