Does industrial policy suppress corporate tax avoidance? —— a study on the perspective of provincial industrial policy

ABSTRACT Will industrial policy affect corporate tax avoidance? Based on China’s listed corporations which are in the current stage of industrial transformation and upgrading, this paper investigates how the provincial industrial policy impact corporate tax avoidance. The results show that provincial industrial policy can suppress corporate’s incentive to avoid tax and this relationship is stronger in non-SOE group. Further study finds that the release of provincial policy enhances the prospect of influenced industries, leading the improvement of firm performance, helping firms obtain more government subsidies and bank loans and finally suppress corporate tax avoidance. On the other hand, industrial policy is a crucial channel through which listed firms can build political connections with local government. To maintain the relationship with local government, the affected firms are willing to suppress corporate tax avoidance.


Introduction
Institutional economics holds that institutional arrangement is crucial to economic growth. Government policy, as an important institutional arrangement, has an important impact on the economic operation and development of the whole society. At the macro level, government policies can influence the long-term economic growth rate of a country (King & Rebelo, 1990). In microscopic perspective, government policies have direct interventions and influence on the daily business operations of enterprises (S. Chen et al., 2010). In the specific emerging and transiting phase of China's economic development, it is the priority of the government's current work to accelerate the adjustment of industrial structure. Due to its great advantages in guiding and adjusting industries, industrial policy has been a key measure employed by governments of different levels to optimise industrial structure and promote the industrial transforming and upgrading since the end of the 1970s. Up to now, China has become a country that carries out more industrial policy. Industrial policies can be divided as the ones issued by central government and the ones issued by local government. Most of the existing researches on industrial policy start from industrial policy issued by central government and discuss its impact on macroeconomic development and behaviours of microscopic firms. But over the past three decades, local government has been playing a key role in regional economic growth (Wu, 2004). After the decentralisation reform of central government and local government in China, local government has a considerable voice and acts like a baton in the process of local economic development. In this case, local governments compete in the process of pursuing regional economic interests, and their policies have a profound impact on the external environment of firms. Hence, it is of great practical significance to explore the effects of industrial policy on microscopic firms from the perspective of local governments. Existing studies mainly discuss effects of industrial policy on firm decisions from the perspectives of enterprise financing (S. Chen et al., 2010;Zhu et al., 2015), innovation efficiency (Li & Zheng, 2016), return on investment (Han & Hong, 2014), etc., but few studies touch upon that of industrial policies issued by local governments on tax avoidance of enterprises which help to evaluate the effect of industrial policy (macroscopic policy issued by government) more comprehensively and profoundly.
Theoretically, tax is one of the important ways in which industrial policies affect enterprises. In the institutional background of fiscal decentralisation in China, although local governments cannot adjust tax types and tax rates, they have great discretion in tax collection, management and enforcement. Hence, local government can give recessive tax preference to preferred industry. In line with industrial policies, local governments will also provide more financial subsidies, credit loan resources and related industry access permits for enterprises in the industrial catalogue (Eaton & Grossman, 1986;Kollmann et al., 2012;Musacchio et al., 2015;Neary & Leahy, 2000). Such industrial policy support has led to higher cash inflows (or lower cash outgoes) for affected companies, easing financial pressure and thus reducing tax avoidance motivation of enterprises. Additionally, industrial policy provides the channels and platforms for enterprises to establish and maintain mutual benefit and mutual trust with the local government. While benefiting from industrial policy of local government, enterprises investing on preferred industries will take tax position and enforcement of the local government into account by lowering tax avoidance motivation. In summary, industrial policies issued by local government may curb tax avoidance of enterprises in the jurisdiction. Previous literatures mainly discuss the factors affecting tax avoidance of enterprises from the basic features of enterprises, the peculiarity of management, the governance and other dimensions (Beck et al., 2014;D. Chen et al., 2016;Chyz, 2013;Dyreng et al., 2010;Edwards et al., 2016;Law & Mills, 2015;Rego, 2003), but few articles investigate the tax avoidance motivation of enterprises from the perspective of macroscopic government policy.
Based on the above theoretical and practical background, this paper attempts to discuss the effects of industrial policy on tax avoidance of microscopic enterprises and its mechanism. The concrete research questions are as followed: First, does macro industrial policy of local government affect tax avoidance of microscopic enterprises? Secondly, are there any differences in the effects of industrial policies issued by local government on tax avoidance of enterprises with different property rights natures? Third, what is the mechanism by which industrial policies issued by local government affect tax avoidance of enterprises? Based on that, industrial policies of 31 provinces, cities and autonomous regions in China from 2008 to 2015 are manually collected, and A-share listed companies in China are taken as samples to conduct an empirical test. The results show that industrial policies issued by local government significantly weakens the tax avoidance motivation of enterprises in the jurisdiction and restrains their tax avoidance behaviour, and this effect is stronger in private-owned enterprises. Further mechanism test shows that the reduction of financing constraints (e.g. improvement of business performance, increase of government subsidies and bank credit/loan) and political connection are the ways that industrial policies affect tax avoidance of enterprises.
Possible contributions of this paper are as followed: First, different from previous studies of industrial policy in which the industries encouraged in the Five-year Plans documents of central and local governments were mostly used to measure industrial policy (S. Chen et al., 2010;Li & Zheng, 2016;Song & Wang, 2013), this paper measures industrial policies based on the actual industrial policies issued by each province, enriching the measurement indicators of industrial policies, more directly testing the economic consequences of the implementation of industrial policy. Second, it directly tests the effects of industrial policies issued by local government on enterprise decisions and preferences from the perspective of tax avoidance of enterprises, and diversifies researches on the correlation of existing macroscopic policies and behaviours of microscopic enterprises (Li & Li, 2014;Zhu et al., 2015). Furthermore, research on the mechanism of industrial policies issued by local governments affecting tax avoidance of enterprises helps us to understand the mechanism of transmission from macroscopic government policy to behaviours of microscopic enterprises. Third, the exploration of tax avoidance motivation of enterprises from the external perspective of industrial policy extends the related research on tax avoidance of enterprises (Brown & Drake, 2014;Cai & Rao, 2015;D. Chen et al., 2016;Hoopes et al., 2012), deepening our understanding of tax avoidance of enterprises

Theoretical analysis and hypotheses development
The government plays a very important role in the process of economic development (Easterly & Levine, 1997;Knack & Keefer, 1995). The development of a country relies on the organic integration of government power and market power, and industrial policy is an important tool to achieve this integration. In China, industrial policy has been an important part of the Five-Year Plans since 1953, and has always acted as a shadow of government intervention in the market economy (D. Chen et al., 2018). Besides Five-Year Plans, the central government, central departments such as the National Development and Reform Commission, the Ministry of Science and Technology, the Ministry of Industry and Information Technology, the Ministry of Finance and other ministries and commissions, and local governments as well as government departments at all levels, have issued a large number of industrial policies, making China a country with a large number of industrial policies. Over the years, industrial policies have played an important role in promoting China's industrial structure adjustment and industrial upgrading. However, with the occurrence of a new round of revolution of science and technology and industrial transformation as well as the deepening of globalisation, the background and conditions of China's industrial development in the current transition period have changed greatly, which makes new demands on industrial policy. China, being in a transition period, is characterised by the imperfect system and law and many uncertain factors. Hence, government intervention is a key external institutional factor that cannot be ignored (Xin et al., 2007). Local government is the key external environment factor of enterprises. After the decentralisation reform of central government and local government in China's government governance structure, local government has a considerable voice in the development of local economy and plays the role of a baton to the regional economy. After the decentralisation reform, the local government is committed to promoting the growth of regional economy, and industrial policy is an important means of its economic intervention. When local governments introduce and implement industrial policies, they often hope to achieve the purpose of accelerating industrial development and increasing fiscal revenue (Han et al., 2017). Under the new normal condition of economy, promoting the transformation and upgrading of industrial structure and optimising industrial structure are important part of government work, which propels Chinese governments at all levels to issue increasing industrial policies. In this paper, the industrial policies issued by the governments of 31 provinces, cities and autonomous regions in China from 2008 to 2015 are collected and sorted out. The results show that the number of industrial policies issued by local governments showed an overall upward trend (see Figure 1). Since 2008, industrial policies issued by provincial governments have been growing steadily. There were 124 industrial policies in 2009, a year with the largest number. This may be due to the outbreak of the financial crisis in 2008. In order to cope with the financial crisis, optimise the industrial structure and promote industrial upgrading, local governments issued industrial policies one after another.
Income tax expenditure is very important cash expenditure in business operations, which directly affects the cash flow of enterprises. According to China's current income tax rate of 25%, enterprises are supposed to contribute a quarter of their pre-tax profits to the tax department. In order to reduce cash expenditure, enterprises have the tax avoidance motivation. Indeed, tax avoidance can cut down tax expenditures, increase cash (Mills, 1998), diminish the demand for debt financing, and reduce corporate leverage (Graham & Tucker, 2006;Hasan et al., 2014). Studies have shown that tax avoidance can reduce cash outflow, increase cash flow and further value up enterprises (Desai & Dharmapala, 2009b). Despite possible benefits, tax avoidance of enterprises can be risky and costly. The cost of corporate tax avoidance mainly includes direct cost and indirect cost (Desai & Dharmapala, 2009a). Specifically, the direct costs mainly include the expenses paid to relevant organisations and personnel for tax planning, and the time spent on tax audit. Indirect costs mainly include litigation and other defence expenses if sued by tax authorities, back income tax, fines, and stricter supervision by tax authorities in the future as well as political and reputational costs (Hasan et al., 2014).
Industrial policy issued by local government may affect the tax avoidance motivation of enterprises from the following three aspects: First, while tax avoidance of enterprises can save the tax expenditure, increase cash (Mills, 1998), it is risky and costly as once the tax avoidance of enterprises is uncovered, the cost could be far higher than the benefits of tax avoidance. In China, one important way of the industrial policies issued local governments to support enterprises is to provide tax preference, including tax reductions and tax rebates. Upon issuance of industrial policies by local government, supported enterprises can get tax preference, saving tax expenditures without risks of tax avoidance. As a result, the eased tax burden of enterprises naturally weakens the motivation for tax avoidance. Second, the Cash Flow Effect Theory of tax avoidance of enterprises demonstrates that tax avoidance extent of enterprises is negatively correlated with debt financing, that is, tax avoidance increases post-tax cash flow of enterprises (Dyreng et al., 2008;Edwards et al., 2016;Law & Mills, 2015), thus reducing the external financing needs of enterprises. The Cash Flow Effect Theory of tax avoidance shows that tax avoidance of enterprises is an important way to alleviate financing constraints (Beck et al., 2014). The more financing constraints, the more likely it is for enterprises to save cash expenditure through tax avoidance. In China, listed companies are generally faced with financing constraints, and it is an important means for them to obtain funds by loaning from financial institutions. In the transition period, the development of financial institutions in China is dominated by banks (Allen et al., 2005), and the banking industry influences the economic development through the allocation of credit loans. However, banks do not allocate credit in isolation but under the influence of the institutions in which they operate (Fang, 2007), which means China's transiting economy is envolved in government intervention. To achieve economic growth, maintain macroeconomic stability and ensure employment, local governments regulate the economy by formulating various economic policies, including industrial policies. One of the important means for local governments to carry out industrial policies is to encourage or restrict the flow of credit funds among different industries by intervening in the financial market. Industrial policies encourage credit funds to favour industries supported by industrial policies, which reduces the external financing constraints of affected enterprises. According to the Cash Flow Effect Theory of tax avoidance, tax avoidance of enterprises is negatively correlated with debt financing (Liu et al., 2017). Enterprises affected by industrial policies can obtain credit support from financial institutions and their financing constraints are reduced. Hence, their tax avoidance motivation decreases. Third, at the present stage, the dominant role of government in economic resource allocation has not changed (S. Chen et al., 2010), and government intervention is still a key external institutional factor that cannot be ignored (Xin et al., 2007). In the absence and imperfection of formal institutions, enterprises are more inclined to seek political connection. Previous studies revealed that political connections can bring various resources and advantages to enterprises, such as preferential bank loans (Houston et al., 2014;Khwaja & Mian, 2005;Sapienza, 2004), priority in obtaining government orders (Agrawal & Knoeber, 2001;Tahoun, 2014), political protection and judicial intervention (Chaney et al., 2009;Correia, 2014;Gordon & Hafer, 2005), which can help enterprises grasp policy trend and external information and enhance their information advantages. Industrial policies are often a variety of encouragement policies carried out by governments to optimise the industrial structure and protect the emerging industries. Industrial policy provides the channel and platform for the enterprise to establish and maintain a connection featuring mutual benefit and mutual trust with local governments. On the one hand, enterprises supported by industrial policies benefit from financing, subsidies and tax, on the other hand, the entrance of enterprises into the industries supported by industrial policy can help the government achieve macroeconomic goals by meeting the industrial development needs of local governments, and then to help local governments to hit political goals and performance assessment indicators. Hence, industrial policy can strengthen the connection between enterprises and the government. In order to maintain this close connection, enterprises supported by industrial policies will take tax position and enforcement of local governments into account while benefiting from the industrial policies. And the tax avoidance motivation of enterprises will be weakened. Additionally, tax avoidance of enterprises can lead to a reputational cost. If a company is found to do tax avoidance activities by the regulators after being supported by industrial policies, its reputation will be seriously affected and its political connection with the government will be impaired. Hence, in order to protect their own reputation, maintain a good relationship with the government and reduce the political cost of tax avoidance, enterprises affected by industrial policies will have a weaker tax avoidance motivation. Therefore, research hypothesis 1 is proposed: Hypothesis 1: Industrial policies help to curb tax avoidance of affected enterprises.
For a long time, the state-owned enterprises take the dominant position in our national economy. In state-owned enterprises, SASAC (State-owned Assets Supervision and Administration Commission), as the major shareholder that actually makes decisions, is an integral part of the government department, which determines the natural close connection between state-owned enterprises and the government. Currently, the stateowned banking sector dominates the financial system of China (Allen et al., 2005). Due to the natural connection with the state-owned banking sector, state-owned enterprises have significant advantages in external financing and are subject to fewer financing constraints (Brandt & Li, 2003). Meanwhile, due to the natural connection with the government, state-owned enterprises have innate advantages in policy support and financial subsidies (Luo & Yang, 2011). Additionally, state-owned enterprises are less willing to reduce tax avoidance in order to maintain good relations with the government. In summary, industrial policy has little impact on tax avoidance motivation and tax avoidance of state-owned enterprises. In the current situation that the government is still in control of the financial system, the problem of 'ownership credit discrimination' is particularly prominent (Allen et al., 2005;Brandt & Li, 2003;Fang, 2007), and private enterprises are faced with more serious external financing constraints. After the issuance of industrial policies by local governments, the support of industrial policies widened the financing channels of affected private enterprises and eased the financing constraints of supported private enterprises. Hence, after the financing constraints are eased, the affected private enterprises will have lower tax avoidance motivation. Due to the different political backgrounds of the major shareholders, compared with state-owned enterprises, private enterprises pay more attention to the construction of political connections (He et al., 2013). The issuance of industrial policies provides a good opportunity for the affected enterprises to establish political connections with the government. In order to obtain preferential policies and maintain a continuous and sound relationship with the government, the enterprises affected by industrial policies are more willing to reduce tax avoidance. Accordingly, research hypothesis 2 is proposed: Hypothesis 2: Compared with that in state-owned enterprises, the inhibition effect of industrial policy on tax avoidance is stronger in private enterprises.

Sample selection
In this paper, the A-share listed companies in Shanghai and Shenzhen from 2008 to 2015 are taken as the initial samples, and selected according to the following criteria: (1) Exclude financial companies; (2) Exclude companies with negative pre-tax profits, companies with actual tax rate greater than 1 or less than 0, and companies with income tax expenses less than 0; (3) Exclude the samples with missing data; (4) All continuous variables are winsorised at the 1st and 99th percentiles. This yields a final sample of 13,363 firm-year observations.

Measurement of industrial policies
In the existing researches, the industrial policies issued by central government such as the Five-year Plans or the catalogue of key industries published by the National Development and Reform Commission are mostly used to measure the variable of industrial policy. In this study, it is believe that industrial policies by central government are suitable for studies of macroscopic effects of industrial policies, while industrial policies issued by local governments are suitable for studies of effects of industrial policy on microscopic enterprises. This can be attributed to several reasons. First, industrial policy is an important means for the local government to intervene in the local economy and adjust the industrial structure in the jurisdiction. Compared with industrial policies issued by central government, industrial policies issued by local governments have a more direct and profound impact on the enterprises within their jurisdiction. Second, industrial policies issued by local governments are to guide the industrial structure according to the advantages of the region. It is featured by making the best use of its situation, more in line with the local conditions. Third, in policies issued by central government, supported industries are listed typically without detailed means of support while in those issued by the local government, concrete means of support is often explicitly put forward, which helps us to further study the channel and mechanism of the effects of industrial policy to the financial behaviours of microscopic enterprises.
In this paper, industrial policies of 31 provinces, cities and autonomous regions in China from 2008 to 2015 are manually collected. Specifically, log into the government portal websites of all districts, search the local government documents one by one and Then, read the specific contents of the support documents of the industrial policy one by one and make clear of the specific industries supported by the documents accordingly, and match these industries one by one with the industry classification stipulated in the Guidelines for the Industry Classification of Listed Companies. Finally, if the industrial policy support document does not specify the supported industry, it is excluded.

Measurement of tax avoidance extent
This paper adopts three methods to measure tax avoidance extent of enterprises. First, the difference between nominal income tax rate and effective income tax rate (TA1) is used to measure the tax avoidance extent of enterprises (H. Chen et al., 2015;Liu & Ye, 2013). The higher value shows the greater tax avoidance extent of enterprises. In China, there are many preferential policies for corporate income tax, and the same effective tax rate may not accurately capture the tax avoidance extent of enterprises (D. Chen et al., 2016). Since various tax preference can be reflected by different nominal tax rates applicable to enterprises, we adopt the difference between nominal tax rate and effective tax rate to measure tax avoidance of enterprises. Secondly, the difference between nominal income tax rate and cash ETR (TA2) is used to measure tax avoidance extent of enterprises (Armstrong et al., 2015;Salihu et al., 2015). The higher value means more tax aggressive. Finally, based on the method of Desai and Dharmapala (2006), we use the book-tax difference (TA3) to measure tax aggressiveness. Since data such as income tax expense of subsidiaries are not available, the average nominal tax rate of the merged company cannot be accurately calculated. In addition, owing to the importance of data such as income tax of parent company expense in calculating tax avoidance indicators, we refer to the general method in existing research (Cai & Rao, 2015;Liu & Zhao, 2019;Wang & Xu, 2016). The nominal tax rate of the parent company in WIND database is used to calculate the tax avoidance extent of the listed company.

Control of variables
To control other factors influencing tax avoidance of enterprises, the following variables are controlled: Enterprise scale, enterprises of different scales receive different attention, and their willingness and the tax avoidance extent are different; The asset-liability ratio, the tax shield effect of debt will affect the tax avoidance extent; ROA, book-to-market ratio and profitability of enterprises will affect tax avoidance motivation, so we take ROA, bookto-market ratio and profitability as controlled variables; The enterprise's asset attributes will also affect tax avoidance motivation and tax avoidance extent, so we take the fixed assets ratio, intangible assets ratio, inventory ratio and monetary fund ratio as the controlled variables; Enterprises' capability for development will have an impact on the tax avoidance extent, so this paper takes the growth rate of total assets and the ratio of investment income as the controlled variables; There may be differences between stateowned enterprises and private enterprises in tax avoidance motivation and tax avoidance extent, so enterprise ownership is taken as the controlled variable; Existing studies have found that ownership concentration and dividend distribution also affect tax avoidance extent of enterprises, so we take them as controlled variables; In addition, in order to control the impact of the nominal tax rate and the difference between the nominal tax rate of parent and subsidiary companies on tax avoidance of enterprises, we take the nominal tax rate and the difference between the maximum and minimum value of all applicable tax rates of the enterprise group as the controlled variables.
In this paper, the nominal tax rate of the company is from the WIND database, and the rest of the financial data are from the CSMAR database.

Model specifications
To estimate the relationship between industrial policies and tax avoidance of enterprises, we use model (1): where i indexes firms, t indexes time. Herein, the explained variables TA i,t refers to the tax avoidance extent of company i in year t. IP i,t stands for whether company i is affected by industrial policy in year t. Since most of the industrial policies do not specify the validity period when they are issued, we set the validity period of the industrial policies that do not specify the validity period as 5 years. Upon issuance of the industrial policy, the value of the variable IP of the supported companies in the current and subsequent 5 years is 1; otherwise, the value of IP is 0. Table 1 lists definitions and evaluations of main variables in this study.
To solve the endogeneity problem, we use model (2) to study dynamics effects of industrial policy on tax avoidance. Herein, before-1, before 0, After 1 and after 2 respectively represent the previous year, the current year, the second year and the third year when the company is first affected by the industrial policy. If a company is affected by industrial policy for the first time in a given year, the value of the previous year before-1 is 1, the value of the current year before 0 is 1, the value of the second year after 1 is 1, and the value of the third year after 2 is 1. The rest of the variables are defined as mentioned above. Table 2 shows the descriptive statistics of main variables in this study. As observed, averages of TA1, TA2 and TA3 are 0.016, 0.024 and −0.003, respectively. Overall, difference of nominal income tax rate and effective income tax rate, difference of nominal income tax rate and cash income tax rate and BTD of listed companies in China are 1.6%, 2.4%,  −0.3%, respectively. The average IP is 0.264, indicating that 26.4% of listed companies in China were affected by industrial policies issued by local governments during the sampling period. Therefore, industrial policies issued by local government affect listed companies in the jurisdiction in various aspects.

Empirical test of effects of industrial policy on tax avoidance of enterprises
This paper tests the effects of industrial policies issued by local government on tax avoidance of listed companies in the jurisdiction. The basic regression results ate listed in Table 3. According to the regression results (1), (3) and (5) in Table 3, it can be seen that the regression coefficients of industrial policy IP are −0.004 and −0.009 and −0.002, respectively, and are significant at the significance level of at least 10%. This result indicates that after the issuance of industrial policies by local governments, tax avoidance extent of affected enterprises affected by industrial policies is significantly reduced compared with that before the issuance of industrial policies. In other words, the issuance of industrial policies by local governments significantly curbed tax avoidance of enterprises in the jurisdiction. This result supports hypothesis 1 of our study, indicating that the industrial policies of local governments may make it feasible for the affected listed companies to get direct benefits such as tax relief, and at the same time, the affected companies may get government fiscal subsidies. Additionally, enterprises affected by industrial policies are willing to reduce tax avoidance in order to maintain a friendly relationship with the government and reduce the reputational and political costs of tax avoidance activities. The regression results of dynamic effects are shown in Table 3 (2), (4) and (6). The results show that the coefficients of before-1 and before0 are not significant, while the coefficients of after 1 are significantly negative, and the coefficients of After 2 are negative but not significant. Results of dynamic effects reveals that industrial policies by local governments significantly reduced the tax avoidance extent of affected enterprises, and this effect is the most significant in the second year after the policy is issued. In order to further discuss the impact of corporate governance characteristics on correlation of industrial policy and tax avoidance of enterprises, we studied the impact of the nature of ultimate property rights on the correlation. Specifically, all the samples are divided into two groups, state-owned enterprise group and private enterprise group. Then, model (1) and model (2) are used to carry out regression for the two sample groups respectively. The regression results are shown in Table 4. In column 1, 5 and 9 of Table 4, when the enterprise is a state-owned one, the regression coefficient of industrial policy IP is negative but not statistically significant; In column 3, 7 and 11, when the enterprise is not a state-owned one, the regression coefficients of industrial policy IP are −0.006, −0.013 and −0.003 respectively, and are significant at the significance level of at least 10%. The results show that compared with state-owned enterprises, the reduction of tax avoidance extent of private enterprises is more significant after the issuance of industrial policies by local governments, and the inhibition effect of industrial policies on tax avoidance of enterprises is stronger in private enterprises. The dynamic effect test results in Table 4 also indicate that the inhibitory effect of industrial policy on tax avoidance of enterprises is stronger in private enterprises. Hence, hypothesis 2 is verified.

Robustness test
In the above regression, we define the industrial policy IP as 1 within 5 years after the issuance of the industrial policy when the enterprise is supported by the industrial policy of the local government. In order to ensure the robustness of the results of this paper, industrial policy is defined in other ways. In this part, we define industrial policy IP as 1 in the years when the enterprise is supported by the industrial policy of the local government upon the year of policy issuance and in the following 4 years; otherwise, the value is 0. With the newly defined industrial policy IP, models (1) and (2) are used to verify research    ***, **, and * indicate that the parameter estimate is significantly different from zero at the 1%, 5%, and 10% level.    hypothesis 1 and 2. The results are shown in Table 5. It can be seen from Table 5   results show that the restraining effect of local government industrial policies on tax avoidance of enterprises is the most significant in the second year after the issuance of the policies. In the state-owned enterprise group in column 3, 9 and 15 of Table 5, the regression coefficients of IP are not statistically significant, while in the private enterprise group in Column 5, 11 and 17, the coefficients of IP are −0.006, −0.013 and −0.003, respectively, and statistically significant. The dynamic effect regression results by grouping also show that effects of industrial policy on tax avoidance of enterprises is more significant in private enterprises. The regression results in Table 5 show that the results in this paper are robust.

Further analysis
Further, we study the mechanism of industrial policies issued by local governments inhibiting tax avoidance of enterprises. The industries guided and supported by industrial policies are usually those with good prospects. The issuance of industrial policies greatly enhances the society's expectation on the prospects of the industry. If an enterprise enters these industries, its business performance is likely to improve. Hence, tax avoidance motivation of enterprises with superior performances decreased upon issuance of industrial policies by local governments. In order to verify this mechanism, we set the dummy variable of business performance change ROA_H, which is 1 when the company's ROA change is greater than the median, and 0 otherwise. Further, we add the cross multiplier of ROA_H and industrial policy IP into the regression model to test the effects of industrial policies issued by local governments on tax avoidance of enterprises in the jurisdiction. The results are shown in Table 6. As observed, the cross multiplier coefficients of ROA_H and IP are significantly negative, suggesting a lower tax avoidance motivation of companies with good business performance after the issuance of local government policies. Additionally, the results in Table 6 show that the improvement of business performance is an important mechanism for industrial policies to curb tax avoidance.
Government subsidy is a common means for the government to guide and support industrial development. Government subsidy enables enterprises to obtain the funds needed for operation and development and increases the cash inflow of enterprises. Hence, the willingness of enterprises to reduce cash outflow through tax avoidance may be weakened. In order to test this mechanism, we set the dummy variable SUBS_H of government subsidy change. When the change of government subsidy obtained by the company is greater than the median, the value is 1; otherwise, it is 0. Furthermore, we add SUBS_H and industrial policy IP into the regression model to test the effects of industrial policies issued by local governments on tax avoidance extent of enterprises in the jurisdiction. The results are shown in Table 7. As observed, except for the first measurement method, the coefficients of SUBS_H and industrial policy IP are all significantly negative, indicating that tax avoidance motivation of enterprises with more government subsidy became weaker upon issuance of industrial policies by local governments. Therefore, government subsidy is an important mechanism of industrial policy to curb tax avoidance of enterprises.
When enterprises cannot obtain the funds needed for operation and development from financial institutions, the motive to reduce cash outflow through tax avoidance becomes stronger and the tax avoidance extent of enterprises becomes more frequent. In order to ease the financing constraints of enterprises and guide enterprises to enter the industries supported by industrial policies, local governments often use credit support when issuing industrial policies, so that enterprises affected by industrial policies can have the access to bank loan. Therefore, tax avoidance motivation of enterprises with more support in bank credit/loan may be weaker. To verify this mechanism, we set the dummy variable LOAN_H of bank loan change, which is 1 when the change of corporate loan is greater than the median of the samples, and 0 otherwise. Herein, the loan includes longterm loan. Further, we add the cross multiplier of LOAN_H and industrial policy IP into the regression model to test the effects of industrial policies issued by local governments on tax avoidance of enterprises in the jurisdiction. The results are shown in Table 8. As observed, except for the first measurement method, the cross multiplier coefficients of LOAN_H and industrial policy IP are significantly negative, which generally indicates that tax avoidance motivation of enterprises with more credit loans decreased upon the issuance of industrial policies by local governments. Therefore, bank credit/loan is an important mechanism for industrial policies to curb tax avoidance of enterprises.
In China, which is in the transition stage, due to the strong uncertainty of the external environment, political connection is an important link connecting enterprises and the government. Therefore, establishing and maintaining political connection is an important way for enterprises to deal with the uncertainty of the external environment. In order to optimise the industrial structure, the government often issues various industrial encouragement policies in the period of protecting emerging industries. During the period of economic transition, there are many uncertain factors, and the industrial policy of the government has greatly changed the macro environment of companies. When companies are uncertain about the future of the industry, they may rely more on internal government information to make decisions. At this time, politically connected enterprises are more likely to obtain the government's internal information and make economic decisions based on it. In order to keep and maintain their relationship with the government so as to better make use of industrial policy, politically connected companies have a weaker motive for tax avoidance .In addition, tax avoidance of enterprises has reputational costs. If a company is found to have tax avoidance activities after being supported by industrial policies, its reputation will be seriously affected and its political connection with the government may be weakened. To sum up, enterprises supported by industrial policy have weaker tax avoidance motivation in order to maintain and strengthen political connection.
In order to test the effects of political connection on correlation of industrial policy and tax avoidance of enterprises, the cross multiplier PC*IP of political connection and industrial policy is added into the model. Herein, we mainly focus on the regression coefficient of the cross multiplier PC*IP. The regression results are shown in Table 9. In this paper, the specific measurement method of political connection is as followed: if the chairman or general manager of an enterprise worked in a government agency at present or in the past, the company has political connection, and the value of PC is 1; otherwise, the value of PC is 0. As can be seen from Table 9, the coefficient of the cross multiplier of PC and IP is significantly negative, indicating that industrial policies provide a channel and platform for enterprises to establish and maintain a relationship featuring mutual benefit and mutual trust with local governments. Investment in industries supported by industrial policies is a channel for enterprises to establish connection with the government. While benefiting from industrial policies, enterprises will also take into account the tax position and enforcement of local governments. Therefore, in order to keep and maintain this connection, tax avoidance motivation of enterprises will be weakened. Table 9 indicates that political connection is an important mechanism for industrial policies to curb corporate tax avoidance.

Conclusions
In the new normal state of economic development, an important goal of China's economic development is to optimise industrial structure and rationalise industrial layout. In order to achieve this goal, local governments have introduced industrial policies to promote the transformation and upgrading of industrial structure. The implementation effect of industrial policies is an important indicator to test the effectiveness of industrial policies. The effects of industrial policies issued by local government on tax avoidance motivation and tax avoidance extent of enterprises in the jurisdiction are studied in this paper. The study finds that industrial policies issued by local government significantly curbed tax avoidance of affected enterprises, especially those that are not state-owned enterprises. Furthermore, we investigated the mechanism of industrial policy affecting tax avoidance of enterprises. It is found that the reduction of financing constraints (business performance, government subsidy, bank credit/loan) and political connection are important mechanisms by which industrial policy influences tax avoidance of enterprises. Upon issuance of industrial policies by local governments, the society usually has high expectations on the prospects of industries guided and supported by industrial policies and enterprises respond to industrial policies issued by local governments by investing in these industries. And their business performance tends to improve. Meanwhile, affected enterprises can have government subsidy and advantage in bank credit/loan, which increases the cash inflow of enterprises. Hence, enterprises lower their intention to reduce cash outflow through tax avoidance so that tax avoidance of enterprises will be cut down. Additionally, industrial policy provides a channel and platform for enterprises to establish and maintain a connection featuring mutual benefits and mutual trust with local government. While benefiting from industrial policies, enterprises will also take into account the tax position and enforcement of local governments. Therefore, in order to keep and maintain this connection, tax avoidance motivation of enterprises will be weakened.
This study is of theoretical and practical significance. Theoretically, this study expands the research on the effects of government's macro policies on the behaviour of micro economic entities. Most of the existing researches on industrial policy start from the industrial policies issued by central government and discuss the effects of macroscopic policy on behaviours of microscopic enterprises. Different from previous studies, this paper, from the perspective of regional industrial policies, explores the effects of industrial policies issued by local governments on tax avoidance of enterprises within their jurisdiction, expanding the research scope of existing studies and enriching existing literature. In addition, the existing literature mostly analyzes the motivation and influencing factors of tax avoidance of enterprises from the perspective of the internal of enterprises. This study investigates the effects of external policy environment on tax avoidance of enterprises from the perspective of industrial policy, enriching the research on the influencing factors of tax avoidance of enterprises and complementing related literature. In the current economic transition period, many industrial policies are carried out in China as an important means for governments at all levels to accelerate the adjustment of industrial structure and optimise the industrial structure. However, the debate on the practical role of industrial policy has not stopped, and some scholars even put forward the theory that industrial policy is useless. This study provides empirical evidence to verify the practical role of industrial policy. It is found in this paper that the industrial policies issued by local governments greatly affect the microscopic behaviours of enterprises in their jurisdictions, and inhibit the motivation of tax avoidance and other illegal activities. This conclusion shows that the industrial policies issued by local governments in China are useful and conducive to the development of local economy. Hence, local governments should further give play to the positive role of industrial policies to guide and promote the transformation and upgrading of regional industrial structure, so as to further optimise the regional industrial structure. The study also shows that after the decentralisation of central and local governments, local governments play an important role in the process of regional economic development. In all districts, the transformation of government functions should be further deepened and the positive role of local governments in the process of economic development should be given full play to. Additionally, it is found that the external environment is an important factor affecting the tax avoidance motivation of enterprises, and the government should further improve the external governance environment to curb tax avoidance of enterprises.