Research on the Impact of Economic Policy Uncertainty on Corporate R&D Innovation: Evidence of China’s Strategic Emerging Industries

Based on the data of China’s A-share listed strategic emerging industry companies from 2011 to 2019, this paper discusses the impact of economic policy uncertainty on R&D innovation. The research shows that the uncertainty of economic policy inhibits the R&D innovation of companies in strategic emerging industries; In the analysis of heterogeneity, it is found that under different property rights and asset scales, the uncertainty of economic policy can inhibit the R&D innovation of companies in strategic emerging industries, especially for non-state-owned or small-scale companies. Based on the research conclusion, this paper suggests that the government should continue to build a fully competitive, fair, transparent and open business system, and strategic emerging industry companies should take measures to improve corporate governance quality to alleviate the impact of economic policy uncertainty.


Introduction
China's strategic emerging industry refers to major technological breakthrough and development demand as the foundation, to the overall and long-term development of economy and the society important leading role, growth potential is huge industry, fusion is the emerging science and technology and the depth of

Literature Review and Hypothesis
R&D is one of the most important investment activities of a company, especially for companies that are strategic emerging industries with innovation as their natural demand. According to the theories of Myers and Majluf [4], Jensen and Meckling [5], information asymmetry and agency conflict have great adverse effects on corporate investment decisions, for example, managers may have adverse selection and moral hazard. When the market doesn't know much about the company's operating conditions, the managers may have adverse selection, while the capital providers may take countermeasures by increasing the cost of capital use. The company may have financing constraints, and the company will give up relevant profitable investment projects due to lack of funds, resulting in insufficient investment of the company. When managers allocate enterprise resources with private interests, under incomplete contracts, managers will have moral hazard. By allocating projects that may not be suitable for shareholders to gain their own interests, such as blind expansion and building a "business empire", the result is over-investment. Different from conventional investment projects, R&D innovation projects have a series of outstanding characteristics, such as long-term, large investment, invisibility, high risk, strong confidentiality, opaque information disclosure, and difficult value evaluation. It can be seen that information asymmetry and agency problems are more prominent in R&D innovation decisions of companies. A large number of studies have found that the information asymmetry between the inside and outside of the company and the conflict between principal and agent are considered as the main distortion factors of the inefficiency of R&D innovation of the company [6].
It has an important influence on the R&D and innovation process of government companies, especially at present, China government clearly regards strategic emerging industries as the main thrust of high-quality economic development. According to the definition of Gulen and Ion [7], the uncertainty of economic policy means that the micro-unit of the market has no way to predict the scenario in which economic policy makers (governments) will be in our country and how to change existing economic policies (such as existing government subsidies, tax incentives, etc.). It can be seen that the uncertainty of economic policy is an unavoidable systemic risk of the company. Pastor and Veronesi [8] thinks that the uncertainty of the environment will make the company's investment behavior cautious, especially for high-risk innovative projects, and the company even deals with this uncertainty by reducing its investment. Based on the view of investment information, uncertainty increases the value of waiting for new information. If the uncertainty always exists, the company will take short-term and steady investment projects as the second-best choice. Because of the long cycle and high probability of failure of innovative projects, companies become more sensitive to innovative investment. At the same time, under the uncertainty of economic policy, it is more difficult to obtain information, so that managers can't effectively evaluate the value of R&D innovation, which eventually leads companies to be more cautious about R&D investment. Based on the theoretical framework of information asymmetry and agency problem, information asymmetry and agency conflict are the main distortion factors that cause the inefficiency of R&D innovation. Venky et al. [9] found that the uncertainty of macroeconomic policies reduces the quality of the company's information environment and aggravates the information asymmetry inside and outside the company. Moreover, the higher the external uncertainty is, the more beneficial it is for managers to cover up their self-interest behaviors, and it is easier for managers to cover up "bad news", and it is easier for them to blame their own lack of efforts in R&D and innovation on the uncertainty of economic policies. Wang and Liu [10] found that when the external macro uncertainty is large, the company insiders are more likely to have earnings management, which in turn makes the agency conflict of the company more serious and ultimately damages the company's R&D and innovation. From the perspective of financing constraints, strategic emerging industry companies are in the growth stage of their life cycle, and need a lot of funds to support their R&D and innovation. However, the company's business activities will be affected by the uncertainty of economic policies, which will lead to the rise of the company's risk of debt and the risk of project default. At the same time, the uncertainty of economic policies will affect investors' judgment of the future situation, reduce their willingness to invest abroad and increase the capital cost, which will lead to the lack of financial support for R&D and innovation projects with high risks. However, high-risk innovation projects often have a long-term development for the company. To sum up, this paper puts forward the research hypothesis: Under the background of strategic emerging industries, the uncertainty of economic policy inhibits the company's R&D innovation. CSMAR is one of the largest accounting and economy-related databases in China. CSMAR provides reliable data for scholars to study Chinese listed companies. In the end, 521 companies were selected as samples in this paper, which is close to 543 by Wang and Zhang [2]. The final sample number of this paper is 4172. The statistical analysis software used in this paper is Stata 17 version.

R&D Innovation of the Company
Input and output are two levels of R&D innovation. R&D investment is generally measured by the ratio of R&D investment to operating income or the ratio of R&D investment to total assets, and R&D innovation output is generally reflected by the number of patent applications and other related indicators. Because the country generally adopts the reward policy of giving high subsidies to patent applications of strategic emerging industries, there is a "Great Leap Forward" phenomenon in which companies apply for patents for subsidies, but ignore the utility or quality of patents. Moreover, because strategic emerging industries are technological innovation-oriented companies, compared with the total assets, the market-oriented operating income can better reflect the future value of the company. Based on this, this paper measures the company's R&D innovation by dividing the company's R&D expenditure by its operating income (expressed as RD). As the explained variable of this paper's modeling, the larger RD is, the better R&D innovation of the company is.

Economic Policy Uncertainty
The measurement of economic policy uncertainty is not easy. Because economic policies are promulgated and implemented by the government, many literatures use the stability of government officials as a substitute variable for economic policy uncertainty, and because economic policies are often transmitted through financial instruments, some literatures also use the stability of local fiscal revenue and expenditure as a substitute variable for economic policy uncertainty.
Since Baker et al. [11] used the method of text retrieval, filtering and analysis of South China Morning Post to construct a statistical index reflecting the uncertainty of China's economic policy, it has been widely used by domestic scholars.
Yang et al. [12] think that compared with the perspective of government offi- wider range of texts, is more comparable, and more comprehensively reflects the fluctuation of economic policies. Based on this, using the practice of Yang et al. [12] for reference, the uncertainty of economic policy is measured by Baker index. Because the Baker index provides monthly data on the website of Policy Uncertainty, and in order to eliminate seasonal fluctuation, and other variables in this paper are annual data, this paper uses the geometric average method to convert monthly data into annual data, and makes natural logarithm processing, so as to maximize the coordination of data time period and data magnitude, as shown in the following model (1).
The monthly data of economic Policy Uncertainty provided on the website of Policy Uncertainty is represented by. Economic policy uncertainty is expressed by EPU. The larger EPU is, the more unstable economic policy is month EPU .

Control Variables
In order to improve the explanatory power of the research model and learn from the practices of Yang et al. [12] and Gu et al. [13], this paper uses company size, investment opportunity, leverage level, operating cash flow ratio, asset structure, property right nature and ownership structure as control variables. The variables used in this paper are described in Table 1.

Model Design
In order to avoid possible problems of reverse causality and endogeneity, this paper draws on the practice of Gu et al. [13]. In this paper, explanatory variables and control variables are delayed by one period, and the test models are controlled by year and industry, as shown in Model (3). If economic policy uncertainty inhibits firm R&D innovation, EPU coefficient is significantly negative.
In the above test model (2), i is the company of different samples, t is the current period, and CV is the general name of all control variables.   In order to preliminarily judge the correlation degree between variables, this paper makes Pearson correlation analysis. As shown in Table 3, the correlation coefficient between R&D innovation (RD) and economic policy uncertainty (EPU) is significantly negative at the confidence level of 1%. The correlation conclusion supports the research hypothesis, which preliminarily shows that economic policy uncertainty inhibits R&D innovation of strategic emerging industry companies. Table 3 shows that the correlation coefficient between two variables is not large (both less than 0.5), which indicates that the problem of multicollinearity is in the test model of this paper.

Mean Value Test
Because the R&D innovation variables of strategic emerging industry companies belong to continuous data, obey or approach normal distribution, and meet the conditions of double independent sample T test, this paper takes the median of economic policy uncertainty (EPU) as the cut point, divides all samples into two groups, and takes the R&D innovation (RD) of the company as the test variable for double independent sample mean T test. As shown in Table 4, the confidence level of R&D innovation of companies with high and low economic policy uncertainty is significantly different by 1%, which shows that when the economic policy is highly uncertain, the R&D innovation of companies is lower, which supports the research hypothesis.   Note: The upper right marks ** and * indicate that the confidence level is 1% and 5% respectively.

Benchmark Regression
The data in this paper belong to panel data. After Hausman test, the model (2) adopts the fixed effect regression analysis of control year and industry. In view of the fact that the sample data has the unbalanced mixed characteristics of cross-section and time series, in order to avoid the heteroscedasticity and correlation of time series and cross-section data in unbalanced mixed data, this paper uses Petersen's method to process the annual and company two-dimensional

Heterogeneity Analysis
Based on the research of Yang et al. [12], companies with government background have more "relationship" resources to alleviate the impact of economic policy uncertainty on R&D innovation. This paper divides the data into nonstate-owned and state-owned groups according to the nature of property rights for regression analysis. Based on the research of Li and Tang [14], the R&D innovation of companies with different asset SIZEs is different. This paper takes the median of company size (the natural logarithm of total assets at the end of the period) as the cut point, and divides the data into two groups: large asset size and small asset size, and makes regression analysis.  Columns (2) and (3) of Table 5 are heterogeneity analysis of different property rights, and columns (4) and (5) are heterogeneity analysis of different asset sizes. The EPU coefficient of data is significantly negative at 5% confidence level in state-owned property rights and large asset scale groups, and significantly negative at 1% level in non-state-owned property rights and small asset scale groups, which indicates that economic policy uncertainty inhibits R&D innovation of companies in different property rights and asset scales, and economic policy uncertainty inhibits R&D innovation of non-state-owned companies and companies with smaller assets more obviously. The reason may be that state-owned companies and companies with large assets are relatively more politically connected, have more "relationship" resources, and have more rent-seeking advantages for government policy information. Therefore, these companies are more resistant to economic policy uncertainty.

The Method of Replacing Variables
In order to verify the reliability of the above empirical conclusions, this paper uses the robustness test method of substitution variables to make regression analysis on the test model again. The regression results of robustness test are shown in Table 6. Compared with the above regression results, there is no qualitative change in the regression results of robustness test, which indicates that the above research conclusions in this paper have good reliability. The specific inspection is as follows. First, because the R&D innovation of companies in strategic emerging industries will be mostly manifested in the number of patent applications and the number of authorizations, Yang et al. [12] used the logarithm of the total number of patent applications plus 1 (expressed by APP) and the total number of authorizations plus 1 (expressed by GRA) as the explained variables. The results are shown in (1) and (2) of Table 6.  Table 6.

Instrumental Variable Method
Considering that the model may have endogenous problems and reverse causality problems, this paper uses the method of instrumental variables to test it. Referring to the method of Gu et al. [13], we use the economic policy uncertainty indicators of the United States as instrumental variables respectively, and make two-stage regression. In the first stage of regression, the uncertain indicators of American economic policy are significantly positively correlated with those of China's economic policy. Table 7 shows the results of the second stage regression, and it can be found that EPU estimation coefficients are significantly negative, indicating that the regression results show that the benchmark regression is robust.
In addition, because the company's R&D innovation may be influenced by the early performance of the company's R&D investment in the later period, in order to avoid this influence, this paper refers to the practice of Fang et al. [16],

Countermeasures and Suggestions
The research conclusion of this paper has important reference value for the practice of government and company.
Economic policy is crucial to the sustainable development of strategic emerging industries, and the uncertainty of economic policy has a negative impact on R&D innovation of strategic emerging industries. However, economic policy will change with the complicated and changeable domestic and international economic situation. As the economist Zhou Qiren pointed out, facing the unpredictable uncertainty, government management departments need a set of systems, and the government needs to further improve the market business quality and break the rent-seeking phenomenon of political association privileges.
Create a fully competitive, fair, transparent and open business system, for example, improve the property rights protection system and contract system, give full play to the market's role in resource allocation, and reduce the interference of government actions on the company's innovation activities. The top-level system design needs to maintain the continuity of basic policies and stabilize the company's innovation expectations. At the same time, the government management department further supervises the quality of accounting information of the company and improves the protection system for small and medium investors.
From the company's point of view, strategic emerging industries are the key nurturing and supporting projects of the government, and the government has issued many economic policies to support them. Therefore, strategic emerging industries are greatly affected by the uncertainty of economic policies, which is the systematic risk that inhibits R&D and innovation of strategic emerging in-

Further Research Directions
According to Zhou and Jiang [17], Economic policy uncertainty is the inevitable systemic risk of the company. In the face of unpredictable economic policies, the company does not just passively accept them. On the micro level, the company can actively implement relevant mechanisms to mitigate the impact brought by the uncertainties of economic policies. For example, the company can alleviate information asymmetry and agency conflict through sound accounting policies, good internal control and other mechanisms, and the company can also undertake more social responsibilities. However, this paper does not discuss how firms' micro policies respond to economic policy uncertainties. Therefore, the subsequent research of this paper can focus on how the internal mechanism of the company can mitigate the impact of economic policy uncertainty on the company's R&D innovation.