关键词: 创新投入, 营改增, 政策评价, 双重差分法

Abstract: Does the "Business Tax Replaced with VAT" policy enhance the innovative investment of the firms? Extant research provides mixed results. It has been widely believed in the literature that there is indeed a correlation between the "Business Tax Replaced with VAT" policy and corporate innovation. However, from the research theme, most of the existing research focuses on the effect of "Business Tax Replaced with VAT" policy on business performance, and few literature pay attention to its relationship with corporate innovation behavior. In the results of the research, the previous research did not solve the endogenous problem, and it is still inconclusive whether the "Business Tax Replaced with VAT" policy has suppressed or promoted the level of innovation investment.In view of this, this paper intends to extend the scope of research to the relationship between the "business reform and reform" policy and the innovation behavior of all service enterprises, and empirically test the impact of the "business reform" policy on the level of innovation investment.Starting from January 1, 2012, the"Business Tax Replaced with VAT" policy owns the characteristics of pilot experiments conducted in batches and batches, and is in line with the conditions for policy analysis using the dual difference method. The data from the A-share listed service companies of the Shanghai and Shenzhen Stock Exchanges from 2008 to 2014 were used as empirical samples, excluding ST, ST* and missing important information. All data is from CSMAR. The company's asset size, operating capacity, growth capacity, age of listing, nature of ownership and whether it belongs to high-tech enterprises are the control variables of this paper. They are divided into treated group and control group. This paper adopts whether or not it belongs to the target enterprises of the"Business Tax Replaced with VAT" policy in other pilot cities except Shanghai in 2013, and assigns treated = 1 to all Beijing Municipalities that were included in the "Business Tax Replaced with VAT" policy target in 2013. Pilot enterprises in Jiangsu Province, Anhui Province, Fujian Province, Guangdong Province, Tianjin City, Zhejiang Province and Hubei Province formed a treatment group. The August 1st, 2013 is used as the segmentation node of the experimental period and the non-experimental period, that is, the "Business Tax Replaced with VAT" policy is changed from the pilot of some provinces and cities to the time of promotion throughout the country. Time = 0 is in the non-experimental period, and time = 1 is in the experimental period. The control variables in this paper,including asset size, operating capacity, growth ability, age of listing, nature of ownership and whether it belongs to high-tech enterprises, are also the propensity matching covariates used for PSM-DID, which makes the above-mentioned treated group sample individuals and control groups The basis for the one-to-one exact match of the sample individuals. The covariate balance of the participation tendency matching scores was examined, and the results showed that the coefficients of enterprise scale, high-tech and enterprise listing years in the covariate are -0.131、6.41、2.206、-0.053 and significant at the level of 0.05, indicating that the covariation vector can be used for the weight calculation of the propensity score matching.In the correlation test, the correlation coefficient between the main control variables used in this paper and the dependent variable named enterprise innovation is tested. The main control variables selected in this paper and the enterprise logarithm with the natural logarithm of R&D investment as one year behind are the enterprise innovations. The correlation between them is high andit can play a controlling role. In the descriptive statistics of each variable, the standard deviation of R&D input was 3.721, indicating a large difference between individual samples. There is a big difference between the business growth rate, the maximum and minimum extremes between the ROA and the company's listing period, which indicates that it is necessary to use the PSM method for control. Regression analysis was performed using DID and PSM-DID respectively. The results showed that the coefficient of the intersection term was 0.077, which was not significant at the significance level of 0.05. In other words, the "Business Tax Replaced with VAT" policy has no significant impact on the R&D investment of related companies. After relaxing the same variance hypothesis, the coefficient of the intersection term is still 0.077, which is not significant at the significance level of 0.05. In the regression results of these models, whether the coefficient of the experimental period (Time) is 1.376, the coefficient of enterprise nature (SOE) is 1.196, and the coefficient of high-tech (TECH) is 3.679. They are significant at the level of significance of 0.01, indicating that the level of innovation behavior of enterprises in the experimental data is higher than that of non-experimental enterprises, and the level of innovation behavior of state-owned enterprises and high-tech enterprises is also higher than that of private enterprises and non-high-tech enterprises. The coefficient of enterprise size (Lnsize) is 0.41 and is significantly positive, indicating that the enterprise innovation investment of large enterprises is higher than that of small and micro enterprises. The company's growth ability (GROWTH) and enterprise listing (IPOAGE) factors are -0.001 and -0.062 respectively, which are consistently significant, indicating that companies with higher operating profit growth rates and companies with longer listings have relatively lower levels of innovation input. The time effect is further fixed, and the coefficient of the intersection term is 0.267, which is still not significant at the significance level of 0.05, indicating that the coefficient results of the two insignificant positive are stable in different methods and different hypotheses. Therefore, the implementation of the "Business Tax Replaced with VAT" policy has a positive and insignificant impact on corporate innovation behavior, and this conclusion is relatively credible.The paper further believes that the possible reason for the above results is that the "Business Tax Replaced with VAT" policy itself has an overall positive impact on corporate innovation. However, due to the uncertainty in the initial stage of the policy, the company’s return on investment in innovation is temporarily suppressed. This has caused related companies to hold a wait-and-see attitude during this period. This conclusion reveals that it is necessary to introduce a series of relevant incentive policies to reduce information asymmetry in the initial stage of policy implementation. The government should reasonably and appropriately regulate the expectations of enterprises for future innovation returns, so that the positive effects of policies can be implied more smoothly.

Key words: innovation investment, business tax replaced with VAT, policy evaluation, difference-in-difference