A study on the impact of digital trade on the position of services in global value chains

The drivers of China’s changing position in GVCs can be categorized into three dimensions: technological position, economic position and the structure of export value-added, with technological position being the dominant factor driving China’s rising position in GVCs [23]. At present, internationally, developed countries with front-end high technology still occupy the high value-added link of products and thus reside in the upstream high-end position of the global value chain, which further proves the dominant position of the influence of technological factors on the position of the global value chain. On the one hand, technological advancement can optimize the ratio of production factors within enterprises and enhance the value-added capacity of the factors, thus achieving the upgrading of GVC status.On the other hand, technological progress can also improve the production process, increase the technological content of the final product, improve the overall product quality, and thus enhance the GVC position. Different from the traditional trade era, the arrival of the Internet era has made digital technology integrate all kinds of innovative elements and become an important driving force for technological progress. Among them, information and communication technology can be put into production operations as a new type of production factor, but also through the formation of complementary advantages with other production factors or production sectors, resulting in technological spillover effects to drive enterprises and even the country’s position in the global value chain. Similarly, AI technology can further enhance production efficiency, help promote the transformation of enterprises from labor-intensive to technology-intensive, and thus promote the status of the global value chain from the middle and lower reaches to the upper reaches.

Factor endowments can be divided into traditional physical capital, human capital, and new endowments, such as independent innovation capacity. Physical capital includes machinery and equipment, plants, transportation and other infrastructures, and along with the improvement of its quality, it gradually increases the domestic added value of commodities in the export trade, which then affects a country’s position in the global value chain, and thus physical capital once occupied a dominant position in the industrial economy. However, with the advent of the knowledge economy, the quantity and benefits of human capital have far exceeded those of physical capital, and thus human capital has occupied a new dominant position. On the one hand, the cultivation of high-quality human resources can help optimize the national industrial structure, drive the added value of export products, and then affect the global value chain position. On the other hand, the “learning by doing” effect of human capital is also a strong driving force to promote a country’s position in the global value chain, and the positive externalities it brings will enhance the R…D innovation and production efficiency of countries. At the same time, the cultivation of high-quality talents will gradually enhance the independent innovation ability of enterprises, which will in turn lead to the improvement of their physical capital.Therefore, the three factor endowments are interconnected and together contribute to the enhancement of GVCs’ status. However, for developed countries, the contribution of financial credit, institutional environment and the level of innovativeness to GVC status is more prominent, while for developing countries such as China, the accumulation of physical and human capital is more conducive to the optimization and upgrading of GVC status.

OFDI mainly affects the upgrading of GVCs through three channels: market internalization, marginal industrial transfers, and reverse technology spillovers. Market internalization is the process by which home-country firms set up subsidiaries abroad, thereby lowering product transaction costs and improving production efficiency, which in turn leads to increased profits and contributes to the upgrading of the GVC position. Marginal industrial transfer means that if an industry in a country is in a relatively inferior state, the country will choose to transfer the industry to other countries for production, and the released production factors will be invested in high value-added industries, thus promoting the optimization of the overall industrial chain and coherence, so that its advantages in the global value chain will gradually become apparent, and further promote the country’s position in the global value chain. The reverse technology spillover effect refers to the fact that through greenfield investment or cross-border mergers and acquisitions, the country comes into close contact with the high-end production technology of developed countries, and then learns and imitates it and applies it to its own production and operation, so as to improve the quality of its own products and production efficiency and achieve the upgrading of the value chain.

The sample chosen for this empirical study is data from 50 countries and regions during the period 2005-2020. This paper sets up the following basic econometric model: (1) GVCposnt=α0+α1Pernt+α2Pernt2+θConnt+εnt

Subscripts n and t indicate country (or region) and year, respectively. GVC_pos_nt denotes the position in the global value chain, with larger values indicating that the economy is at the higher end of the chain. Per_nt refers to digital trade, Cont_nt contains all control variables, and ε_nt is a random disturbance term.

In order to better analyze how digital trade affects GVC status, the following recursive equation is constructed on the basis of equation (1): (2) TECnt=β0+β1Pernt+β2Pernt2+δConnt+εnt (3) GVCposnt=γ0+γ1Pernt+γ2Pernt2+φTECnt+δConnt+εnt

where γ₁,γ₂ represents the direct effect of digital trade affecting the GVC position of services, and β₁×φ,β₂×φ represents the mediating effect transmitted through STI and transaction costs. TEC_nt is the mediating variable, representing STI and transaction costs.

The services GVC position index (GVC_pos) for 50 economies is used as an explanation variable, which is calculated to depict the services GVC position of the 50 countries in the sample. The data used are from the ADBMRIO database from 2005 to 2020.

Digital trade (Per) is chosen as the core explanatory variable, and the data are taken from the digital economy section of the UNCTAD database, which measures a country’s digital trade using the share of ICT goods trade exports in total trade exports. This statistical method is still insufficient, but it is still the most feasible accounting method among the single indicators reached by the current technology, and has been widely used in the ministries of commerce of various countries and regions in the world, which has a certain degree of rationality and feasibility. In addition, the corresponding control variables and intermediary variables have been selected, and all the variables selected for measurement and data sources are shown in Table 1.

The results of the impact of digital trade on the GVC status of the service sector are shown in Table 3. The robustness of the model is verified by gradually introducing control variables, and it is found that the core explanatory variables are consistently significant, with positive primary term coefficients and negative quadratic term coefficients, and the degree of fit gradually improves, showing that the model fit continues to improve. When other variables are not considered, the primary term coefficient of digital trade is significantly positive at the ɑ = 0.1 level, and the secondary term coefficient is significantly negative at the ɑ = 0.05 level, indicating that at the initial stage of the participation of digital services in international trade, the improvement of digital trade can contribute to the improvement of the GVC status of the service industry. With the deepening of digitization, the increase of digital trade on the GVC status of the service industry reaches a critical value, and then the GVC status of the service industry will decline with the increase of the level of digital trade, showing an inverted “U”-shaped trend.

After adding the control variables such as labor market size, foreign direct investment, GDP per capita, labor productivity, human capital, etc., the coefficient of the primary term of digital trade is still significantly positive at the level of ɑ=0.05, while the coefficient of the secondary term is still significantly negative at the level of ɑ=0.05, which indicates that the inverted “U”-shaped nonlinear relationship still exists.

For the coefficients of the control variables, the coefficient of the labor market size is significantly negative at the ɑ=0.01 level, indicating that the labor market size reduces the GVC status of the service industry, and labor-intensive countries are more likely to engage in the work of the intermediate links with low value added by virtue of their own demographic dividend, which leads to the lower GVC status of the service industry in the country. The regression coefficients are not significant after adding FDI and GDP per capita. Labor productivity is significantly negative at the ɑ = 0.1 level, indicating that an increase in labor productivity enables the labor force to perform more low-value work such as assembly and processing, thus reducing the GVC status of the service sector. The GVC status of the service sector is not significantly affected by adding the control variable of human capital.

Because of the consideration of the heterogeneous impact of digital trade on countries at different levels, this paper divides the 50 countries into three groups, namely, lower-middle-income countries, middle- and upper-middle-income countries, and high-income countries, and regresses each group of countries on this basis. Table 4 displays the regression results for different income levels.

For countries with different income levels, the impact of digital trade on the GVC status of services is different. The coefficient of the primary term of digital trade in low- and middle-income countries is positive at the significant level of ɑ=0.05, indicating that digital trade and service industry GVC status in low- and middle-income countries have a linear positive relationship, the reason is that low- and middle-income countries have a low level of digital trade, there is a large space for improvement, and the advancement of digital technology can improve their service industry GVC status, but due to the weakness of the digital foundation, the positive impact is smaller. The primary and secondary coefficients of the digital trade level of high- and middle-income countries are both significant at the level of ɑ=0.01, with the primary coefficient significantly positive and the secondary coefficient significantly negative, indicating that the inverted “U”-shaped relationship between the digital trade and the GVC status of the service industry in high- and middle-income countries, and that high- and middle-income countries already have a certain level of digital trade. The middle and high-income countries have a certain digital foundation, with the improvement of the level of digital trade will improve the GVC status of the service industry, but when the growth of digital trade to a certain extent, there will be “big data to kill maturity”, “data hegemony”, “digital monopoly” and other phenomena, hindering the service industry GVC status. However, when digital trade grows to a certain level, the phenomena of “big data killing”, “data hegemony” and “digital monopoly” will appear, hindering the improvement of the GVC status of the service industry. The impact of digital trade on the service industry GVC status of high-income countries is not significant, probably because high-income countries occupy the two ends of the “smile curve” due to the maturity of technology, and their own service industry GVC status is relatively high, so the impact of the increase in the level of digital trade is weaker on them.

Considering the possible heterogeneity of the impact of digital trade on the status of GVCs in different industries, the industries are divided into the primary industry (agriculture, forestry, animal husbandry and fishery), the secondary industry (mainly manufacturing industry), and the tertiary industry (mainly service industry) according to the standard of China’s national economic industry classification to carry out a further regression in groups, and the regression results are shown in Table 5.

For the primary, secondary and tertiary industries, the impact of digital trade development on the status of global value chains is positive, but the coefficients of the primary and secondary industries are still not significant even at the ɑ=0.1 level of significance, which indicates that there is indeed heterogeneity in the impact of digital trade on the status of global value chains in different industries. At the same time, it also shows that the impact of digital trade on the value chain of the service industry is positive and significant, but the impact on the primary industry and manufacturing industry is not significant, which may be due to the fact that although the digital trade is developing rapidly, its import and export volume is still relatively small compared with the primary and secondary industries, so the pulling effect on the primary and secondary industries is not obvious. In addition, there is a certain lag in the technology spillover effect of digital trade, and the integration of digital technology with primary industries and manufacturing industries is not close enough. Therefore, China, as a large manufacturing country, is still at the middle and low end of the value chain, and should grasp the opportunity of the development of the digital economy, take advantage of the dividends of the digital economy, accelerate the integration and development of digital technology and the primary and secondary industries, realize the transformation and upgrading of industries, and embed itself in the middle and high end of the global value chain.

In this section, technological innovation (RD) and production costs (ct) are added and empirically tested using a mediated effects model. The two indicators are explained next:

Technological innovation (RD): there are many academic measures of technological innovation, and most scholars use R…D capacity to measure it. Considering the research effect, this paper selects the number of resident patent applications to measure technological innovation, and the data comes from the WDI database.

Production cost (ct): digital trade accelerates the digitization of goods and services, simplifies the transaction process, reduces transaction costs, and can enhance the position of the global value chain. This paper uses the proportion of Internet users in each country who use the Internet (CT) as a proxy variable for transaction costs.

In the process of testing the impact mechanism, two mediating variables of technological innovation and production cost are added, and the test results are shown in Table 6. From the table, it can be seen that the impact of digital trade on technological innovation and trade costs passes at least the significance test of ɑ=0.05, while the impact of technological innovation and trade costs on the GVC status of the service industry also passes the significance test, and there is a mediating effect, indicating that digital trade can improve the competitive advantage of the service industry of each economy in the international market by reducing the cost of production and enhancing the ability of technological innovation, thus promote the improvement of the GVC status of the service industry.

In this paper, the original hypothesis was rejected through the Dobbins Wu Hausmann test, which concluded that there was some endogeneity, and the regression was conducted using two-stage least squares (2sls) using per capita Internet usage as an instrumental variable for the core explanatory variables. The results are shown in Table 7.

The findings show that the sign of the coefficients of digital trade did not change significantly, and Kleibergen-PaapWaldrkF is greater than the maximum instrumental variable critical value at the level of the statistic Stock-Yogo test ɑ=0.01, and the weak instrumental variable hypothesis is rejected. Meanwhile, the estimated coefficients of the core explanatory variables increase compared with the benchmark regression, indicating that after controlling for endogenous problems, the facilitating effect of digital trade on China’s embeddedness in the GVC position is more obvious. Therefore, China’s position in GVCs is enhanced significantly by the development of digital trade.

In order to prevent the influence of outliers on the results, the paper carries out the treatment of shrinking the tails before and after 1% for all continuous type variables, and the results are shown in Table 8. After performing the shrinking treatment, the direction and significance of continuous variables do not change significantly, the coefficients of the core explanatory variables are positive and remain significant at least at the level of ɑ=0.1, and the control variables do not change significantly, the coefficients of the labor market size are still negative and significant at the level of ɑ=0.01, which indicates that the conclusions do not change after the outliers are eliminated. Digital trade can enhance China’s global value chain embedded position, indicating that the model is robust.