Study on Causes of Differences in Tax Burden of Value-Added Tax from the Perspective of Industrial Linkage

1 Introduction

As a key component of the real economy, the manufacturing industry is the “ballast stone” and “stabilizer” of the national economy (Zheng and Zhang, 2022). The high-quality development of the manufacturing industry largely depends on the rational use of economic regulation and control policies, especially fiscal and tax policies. How to use proactive fiscal and tax policies to better serve the transformation and upgrading of the manufacturing industry and stable economic growth is not only the top priority to improve the resilience and integrity of the industrial chain, but also a key link to accelerate the improvement of the modern fiscal and tax system and improve the national governance capacity. During the accelerated construction of a new development pattern, the linkage among various industries in the domestic circulation will become increasingly close. In order to avoid industry hollowing and over-service, and improve the balance and coordination of industrial development, it is necessary to incorporate fiscal and tax policies of the manufacturing industry into the overall optimization of the industrial system and the economic system, and grasp the laws and characteristics of the input-output flow among industries. For example, in the fiscal and tax policy system, the reform of the VAT system is the focus of policy adjustment, which reflects the typical characteristics of the combination of micro-policies to stimulate the market vitality and structural policies to smooth the economic circulation.

Since the official imposition of VAT in China in 1994, the difference in the actual tax burden of VAT in industries has remained controversial (Wang et al., 2022). In fact, the core of the VAT system reform is to keep VAT revenue neutral. For example, the “VAT transformation” expands the scope of input tax deduction and further promotes enterprises to increase fixed asset investment (Ma et al., 2019); the “r eplacement of business tax with VAT” optimizes the VAT deduction chain (Li and Yan, 2018) and helps improve the specialized division of labor among industries (Fan and Peng, 2017). However, the “VAT transformation” and “replacement of business tax with VAT” focus on reducing the actual tax burden of VAT in the industry on the whole, and do not make a full explanation of the tax burden difference, causes and structural characteristics of the industry. In order to create a fairer tax environment, scholars proposed to simplify and consolidate the VAT tax rates tiers (Fang et al., 2022), so as to reduce the differences in tax deduction among enterprises in the industry. In addition, implementing structural tax reduction policies can reduce the distortion of social welfare levels (Ni, 2021).

As for the difference in the VAT tax burden in China’s manufacturing industry, most domestic scholars explained the causes from the perspective of the VAT system (Chen, 2013; Ni et al., 2020). The specific reasons can be classified into three categories as follows. 1. A non-single tax rate system is used, such as multiple tax rates; 2. The VAT deduction scope is limited. For example, under the production-oriented VAT system, enterprises that are highly dependent on fixed assets for production activities can obtain small amount of input tax deduction; 3. The VAT deduction is incomplete. Before the “replacement of business tax with VAT”, the VAT deduction chain between the manufacturing and service industries was discontinuous, which led to high VAT paid by manufacturing enterprises relying heavily on intermediate inputs from the service industry. In addition, the intervention of local governments is also an important factor affecting the actual tax burden of enterprise VAT. For example, the fluctuation in the VAT share ratio leads to changes in the tax collection intensity of local governments (Chen, 2017).

Some literatures study the difference of the actual tax burden from the perspective of enterprise heterogeneity. For example, Saez (2002) conducted a study on early commodity tax and believed that the consumer preference heterogeneity or corporate product heterogeneity led to different commodity tax rates among enterprises. Yin et al. (2021) discussed the scale of excess VAT burdens for Chinese manufacturing enterprises from the perspective of enterprise productivity heterogeneity and product heterogeneity, and regarded the distribution of comprehensive heterogeneity as an important factor leading to different excess VAT burdens of various industries. Compared with the above studies, this paper highlights the influence of the input-output correlation among manufacturing industries on the actual VAT tax burden of different industries from the perspective of industrial linkage, and further considers the moderating effect brought by the production network characteristics of various industries, so as to further explain the causes of the difference in the actual VAT tax burden of manufacturing enterprises in China.

In general, there have been abundant studies on the causes of the distortion of VAT tax burden in the manufacturing industry, but most of them were carried out around the manufacturing industry as a whole. With the gradual formation of the domestic circulation and the deepening of specialized division of labor, it is increasingly important to analyze the tax burden issue from the perspective of industrial linkage. This paper holds that the VAT deduction mechanism will be of greater significance if the difference in the VAT tax burden of the manufacturing industry is discussed from the perspective of industrial linkage. According to the principle of VAT deduction, the VAT payable is the difference between the output tax and the input tax, in which the output tax and input tax are respectively the product of sales and purchases and the corresponding VAT rate, which contains the supply-demand relationship or input-output correlation among industries. Therefore, explaining the difference of the actual VAT tax burden in the manufacturing industry from the perspective of industrial linkage reflects the linkage between the input-output correlation and the actual VAT tax burden among industries, and provides a new way of thinking to solve the distortion of the VAT tax burden in the manufacturing industry.

Based on the existing studies, this paper first discusses the practical characteristics of the actual VAT tax burden in the manufacturing industry in China and the input-output correlation among industries, and on that basis, empirically tests the effect of industrial linkage on the actual tax burden of VAT, and considers the moderating effect of production network characteristics in the industry. It also conducts mechanism tests from the aspects of enterprise tax avoidance, tax shifting ability and substitution elasticity of intermediate inputs, and puts forward policy suggestions at the end to provide a realistic basis for optimizing VAT tax policies supporting the development of the real economy, and off er policy implications for building a modern VAT tax system and promoting the transformation and upgrading of the manufacturing industry.

2 Institutional Background and Cause Analysis

2.1 Characteristics and Causes of the Actual VAT Tax Burden in Manufacturing Industry

Based on the national tax survey data from 2007 to 2014, this paper calculates the actual VAT tax burden of manufacturing enterprises in China, and plots a diagram of the actual VAT tax burden of 15 manufacturing s ub-industries. Based on Figure 1, this paper will analyze the characteristics of actual VAT tax burden in the manufacturing industry in China from two aspects: tax burden in the industry and tax burden differences among the sub-industries.

Figure 1

Actual Tax Burden of VAT in Manufacturing Industries in China from 2007 to 2014

On the one hand, the overall VAT tax burden of the manufacturing industry was relatively high before 2010 for the following reasons: First, the reform of tax sharing system in 1994 led to insufficient local financial resources, and the reform of income tax sharing in 2002 further aggravated the local fiscal shortage (Liu, 2018), so local governments were motivated to strengthen the VAT tax supervision (Chen, 2017); Second, the production-based VAT implemented before 2009 resulted in the limited scope of enterprise input tax deduction; Third, tax collection and administration projects such as Golden Tax Project I and Golden Tax Project II strengthened the standardization of VAT invoice management, and the level of VAT tax burden was enhanced (Fan et al., 2018). Since 2009, with the introduction of policy measures such as “VAT transformation” and “replacement of business tax with VAT”, the tax burden on manufacturing enterprises has eased. Meanwhile, the increase in the share proportion of VAT divided to local governmets is conducive to promoting local governments to implement tax incentives for manufacturing enterprises to drive the development of regional industries (Liu and Mao, 2019).

On the other hand, the distribution of actual VAT tax burden in manufacturing sub-industries is relatively discrete, indicating that there are obvious differences in the actual tax burden borne by sub-industries. The causes of the difference in VAT tax burden among manufacturing sub-industries are complicated. For example, China’s VAT collection and management system faces the problem of differential VAT rates represented by multi-level tax rates and preferential tax rates, and the VAT deduction chain is short and incomplete in some sub-industries. Since 2011, the problem of tax differences among manufacturing sub-industries has eased. One reason is that a series of VAT reforms implemented in China since 2008 have achieved remarkable results. For example, the “VAT transformation” has transformed production-based VAT to consumption-based VAT, expanded the scope of enterprise input tax deduction, and the “replacement of business tax with VAT” has opened up the VAT deduction chain of manufacturing and service industries, and promoted tax fairness to a large extent. It should be emphasized that although the difference in VAT tax burden among different manufacturing sub-industries is shrinking in China, there are still tax distortions among the sub-industries, and the resulting imbalance has become a restricting factor for the high-quality development of China’s manufacturing industry and economy.

2.2 Characteristics and Causes of Industrial Linkage in Manufacturing Industry

With reference to the research of Fan and Peng (2017), this paper calculates the linkage between 15 manufacturing sub-industries in China and the industries that provide intermediate inputs for them from 2007 to 2014, and formulates the trend chart of the changes in industrial linkage.

According to Figure 2, the industrial linkage of China’s manufacturing industry shows an obvious upward trend overall, indicating that the input-output correlation between manufacturing sub-industries is becoming closer, and the advantages of vertical specialized division of labor are gradually emerging. In addition, the industrial linkage of the manufacturing industry has obvious hierarchical characteristics in China. The industries with a linkage degree higher than 25% are mostly technology intensive, and the industries with a linkage degree lower than 25% are mostly labor intensive or resource intensive. The reason is that under the vertical specialized division of labor system, the intermediate goods required by the manufacturing sub-industries are more dependent on external purchases rather than production in the industry. The value chain of technology-intensive industries is long, and there are many kinds and large quantities of outsourced intermediate goods, hence the backward linkage between such industries and other industries are closer. In contrast, the value chain of production activities in labor-intensive industries is short, and the characteristics of specialized division of labor are not obvious, thus the industrial linkage is low.

Figure 2

Linkage of Manufacturing Sub-Industries in China from 2007 to 2014

4 Analysis of Benchmark Regression Results

5 Study on Moderating Effect of Production Network Characteristics

Characteristics of the industrial production network are used to measure the position of an industry in the production network, including upstream degree, centrality degree and block characteristics. Specifically, production network indicators such as upstream degree, centrality degree, and block type are descriptions of industry characteristics from different dimensions. Upstream degree is used to measure the distance between an industry and the final demand end. The higher the upstream degree of an industry is, the further it is from the final demand end. Centrality degree is mainly used to indicate whether the industry is at the core of the production network or at the edge. Blocks are classified according to the ratio between the transfer from industries in the block to industries out of the block and imports by industries in the block from those outside. This paper calculates production network indicators of each manufacturing sub-industry to examine whether the position of the industry in the production network has a moderating effect on the relationship between the industrial linkage and the actual VAT tax burden.

This paper includes interaction terms of industrial linkage and production network characteristics on the basis of Model (3) to build the following econometric model:

Wherein, Net_jt refers to production network characteristics of industry j. This paper uses upstream degree (Upstr), centrality (Centr) and block characteristics (Block) to represent production network characteristics, μ_j represents industry effect ^[1], and δ_t represents time effect. The analysis on the effect of production network characteristics in the manufacturing industry by introducing interaction terms can not only test the statistical significance of production network characteristic variables in a statistical sense, but also show whether production network characteristics can moderate the influence of industrial linkage on the actual VAT tax burden in an economic sense.

This study takes upstream degree, centrality and block characteristics as moderating variables and estimates the results of Model (4). Table 3 shows the results, wherein, column (1) shows the estimates with upstream degree (Upstr) as a production network characteristic variable (Net). Column (2) shows the estimates with centrality (Centr) as a production network characteristic variable. Column (3) shows the estimates with block characteristics (Block) as a production network characteristic variable.

The results in Table 3 show that the estimated coefficient of the upstream degree characteristic of the manufacturing industry is significantly negative. It indicates that the closer an industry is to the upstream of the industrial chain, the lower the actual VAT tax burden it bears, because on the one hand, the actual VAT burden of upstream industries in the industrial chain is greatly affected by policy factors, and on the other hand, upstream industries are mostly engaged in basic production activities such as raw material processing, and the government usually provides more tax incentives to such industries. Therefore, the upstream degree of the industry has a significant negative influence on its actual VAT tax burden. However, the interaction coefficient between industrial linkage and upstream degree is positive, but not significant. This indicates that the characteristic of upstream degree has no significant moderating effect on the relationship between industrial linkage and the actual VAT tax burden.

The interaction coefficient of industrial linkage and centrality is significantly positive. This indicates that the actual VAT tax burden of industries close to the center of the production network is less affected by the correlation of intermediate goods, compared to industries located at the edge of the production network. A possible explanation for this is that manufacturing enterprises with a higher degree of centrality have obvious information advantages in production activities, and can preferentially choose to cooperate with more outstanding enterprises, thus forming more stable input tax deduction for intermediate goods. In addition, because of the superior position, enterprises in industries with a higher degree of centrality bear lower communication and transaction costs with enterprises in other industries, which further enhances the stability of intermediate goods transactions between this industry and other industries. Therefore, the linkage of industries with a high degree of centrality has a relatively limited negative influence on the actual VAT tax burden.

As can be seen from Table 3, the interaction coefficient between industrial linkage and block characteristics is significantly -0.093, indicating that compared with the industries in the spillover block, the correlation of intermediate goods in industries of the beneficiary block has a more significant negative effect on the actual VAT tax burden. Most industries in the spillover block are basic industries of the national economy. The intermediate products of the industry in this block have obvious spillover effect. The deduction of the inflow of intermediate goods is not an important factor affecting its VAT tax burden. Therefore, for the industries in the spillover block, the linkage with other intermediate goods industries cannot significantly offset their actual VAT tax burden. An industry in the beneficiary block can receive intermediate goods from other industries within the block as well as those from industries in the overflow block (Li et al., 2014). The sources of intermediate inputs required for its production are more extensive. Hence, industrial linkage has a more significant influence on the actual VAT tax burden of industries in the beneficiary block.

6 Intermediate Mechanism Test

According to the study of Jiang (2022), there may be some errors in the analysis of the intermediate mechanism by using the traditional mediation model. Therefore, this paper mainly observes the influential effect of industrial linkage of the manufacturing industry on mediators to judge the action mechanism of industrial linkage on the actual VAT tax burden of manufacturing enterprises. The models built are shown in Equations (5) and (6) as below:

Wherein, M_ijt refers to the mediator. In this paper, variables such as corporate tax avoidance (Taxavoid), tax shifting ability (Taxshift) and substitution elasticity of intermediate inputs (Es) are selected successively for testing. Specifically, this paper determines whether an enterprise evades tax (Taxavoid) according to whether it pays additional VAT in the current year. Additionally, the integrative bargaining capability of an enterprise relative to its suppliers and customers is an important indicator reflecting its tax shifting ability (Anderson et al., 2001). In view of this, this paper takes integrative bargaining capability as the proxy variable of tax shifting, and uses the net trade credit of an enterprise to measure its integrative bargaining capability (Wei and Zhu, 2019). The greater the net trade credit, the higher the integrative bargaining capability. Meanwhile, with reference to the research of Atalay (2017) and Ni (2021), this paper estimates the substitution elasticity of intermediate inputs in China’s manufacturing sub-industries, and tests it as a mediator.

It should be noted that the endogeneity of mediators is an important factor that leads to the judgment bias of traditional mediation models (Jiang, 2022). Based on this consideration, this paper focuses on the relationship between mediators and the actual VAT tax burden, so as to avoid endogenous problems caused by reverse causality. Specifically, the difference of actual VAT tax burden in manufacturing sub-industries may give rise to the tax avoidance motivation of enterprises, but it is not the factor that leads to the tax avoidance behavior of enterprises. Similarly, an enterprise’s bargaining capability determines its capability of tax shifting which directly affects its actual VAT tax burden. The substitution elasticity of intermediate inputs is used to measure the difficulty of mutual substitution between different intermediate inputs, which depends on the nature of intermediate inputs. Therefore, all the mediators used in this paper are exogenous. Table 4 shows the results of the intermediate mechanism test.

It can be seen from columns (1) and (2) of Table 4 that the influence coefficient of industrial linkage on corporate tax avoidance is significantly positive. Compared with the industries with a lower degree of linkage between intermediate inputs, the industries with a higher degree of linkage conduct transactions with other intermediate goods industries more frequently, and theoretically have more opportunities to evade tax. From the perspective of location, industries with a higher degree of linkage between intermediate inputs are mainly characterized by the production of final demand goods. The VAT input deduction rules are only applicable to different industries in the industrial chain. Even if the final demander obtains sales invoices of enterprises in downstream industries, they cannot be used for input tax deduction, and the final demander may lack the motivation to require enterprises to truthfully issue VAT invoices (Naritomi, 2019). On the contrary, the authenticity of VAT invoices issued by upstream industries determines the input tax deduction of downstream industries. There are checks and balances and mutual supervision between the upstream and downstream industries. It is unfeasible for upstream industries to evade tax artificially (Tian and Fan, 2016).

The influence coefficient of the linkage between the industry of the enterprise and other intermediate goods industries on the tax shifting ability is significantly positive. The dependence of the manufacturing industry on other industries is a key factor affecting its tax shifting ability (Ahern, 2012). Structural supply-demand imbalance exists in some manufacturing sub-industries in China (Dong et al., 2015), which is reflected in the low rate of capacity utilization of some basic industries, and the need to rely on external demand to absorb internal excess capacity. At this point, the supplier of intermediate goods is more dependent on the demander of intermediate goods, thus it is easier for the latter to lower the procurement costs and realize tax shifting. Namely, enterprises in the industry closely linked to the intermediate goods industry are more able to carry out tax shifting. It should be noted that when enterprises shift the tax burden onto suppliers of intermediate goods by lowering the price of intermediate goods, it may lead to a decrease in input tax, and further result in an illusion of increased nominal VAT tax burden, but in fact, the actual tax burden of enterprises does not increase.

According to columns (5) and (6) of Table 4, the estimated coefficient of the influence of industrial linkage on the substitution elasticity of intermediate inputs is significantly positive. This indicates that highly linked industries have stronger substitution elasticity for intermediate inputs required for production. The production linkage between manufacturing sub-industries is becoming closer, which means that the professional division of labor system is becoming increasingly perfect, the vertical production mode is gradually formed, and a large number of intermediate goods enterprises are gathered in each production link, resulting in that enterprises with strong industrial linkage can obtain intermediate inputs from more channels. Therefore, the higher the linkage of manufacturing sub-industries is, the greater the substitution elasticity of intermediate inputs will be. It can be seen that the substitution elasticity of intermediate inputs determines the flow direction of intermediate inputs and is an important factor affecting the types of intermediate inputs and the level of tax burden for manufacturing enterprises. Also, it shows that the greater the substitution elasticity of intermediate inputs, the more serious the distortion of VAT tax burden (Ni, 2021).

7 Conclusions and Policy Implications

Studying the differences of actual VAT tax burden among different sub-industries of the manufacturing industry in China and their causes from the perspective of industrial linkage is not only a practical basis for improving the fiscal and tax incentive policies related to the manufacturing industry, but also a basis for addressing the differences of actual VAT tax burden in the manufacturing industry. From the perspective of industrial linkage, this paper studies the differences of actual VAT tax burden among different sub-industries of the manufacturing industry in China and their causes, discusses the important role of manufacturing production network characteristics in explaining the differences of actual VAT tax burden, and draws the following main research conclusions:

Industrial linkage has a significant negative influence on the actual VAT tax burden of the manufacturing industry, that is, the closer the intermediate input connection between an industry and other industries, the lower the actual VAT tax burden it bears. The characteristics of production network have a significant moderating effect on the relationship between the linkage of manufacturing sub-industries and the actual VAT tax burden. Industries in the center of the production network realize a relatively stable deduction of input tax by optimizing the source of intermediate inputs, and further weaken the impact of industrial linkage changes on their actual VAT tax burden. Compared with the industries in the spillover block, industries in the beneficiary block achieve a greater degree of VAT deduction through a widely connected input-output network, further reducing the actual VAT tax burden of the industries in the block.

According to the mechanism test results, enterprises in industries of the downstream industry chain lack the supervision of VAT invoices from the final demand side, hence they have the motivation to reduce tax burden through tax avoidance. The demander of intermediate goods tends to shift its tax burden onto the supplier of intermediate goods, thus effectively easing the VAT tax burden. Enterprises with a higher industrial linkage are more able to carry out tax shifting. In addition, the substitution elasticity of intermediate inputs is an important factor affecting the types of intermediate inputs and the tax burden of manufacturing enterprises. The substitution elasticity of intermediate inputs required in the production process is greater for highly linked manufacturing sub-industries, so they can adjust the types of intermediate goods to reduce the actual tax burden.

According to this study, on the one hand, structural tax reduction policies should be implemented in a targeted manner to improve the accuracy and effectiveness of VAT related policies on the basis of fully considering the location characteristics of various industries in the production network. For example, for industries at the core of the production network, larger fiscal subsidies or tax incentives should be given to give full play to their basic or key supporting role for the overall manufacturing industry, while for industries at the edge of the production network, fiscal and tax support policies should be appropriately reduced to ensure the relative fairness of the actual tax burden of various industries in the production network to the maximum extent. On the other hand, efforts should be made to improve or protect the integrity of the VAT deduction chain and prevent the distortion of tax burden that may occur in the process of tax collection and administration, such as improving the VAT supervision mechanism among enterprises in the upstream and downstream industries, building external constraints of joint supervision by the government and society, mobilizing the internal circulation mechanism of the production network in the manufacturing industry, and optimizing the connection mode of the upstream, midstream and downstream of the manufacturing industry, forming the internal motivation of tax compliance and mutual checks and balances of various industries in the production network, and ensuring high-quality development of the manufacturing industry and stable economic growth with tax equity.