Abstract: There is an interesting relationship between corporate R&D investment intensity and innovation efficiency of A-share listed companies in recent years. The R&D investment intensity is increasing year by year, but the innovation efficiency is not improving or even getting worse, which makes people wonder what is driving enterprises to continuously increase innovation investment? According to the existing research conclusions, product demand, innovation cost and the innovation behavior of other enterprises are the three factors that directly influence the decision-making of enterprise innovation. Obviously, due to the mismatch between innovation input and output, the driving force of innovation cannot be fully explained from the perspective of product demand and innovation cost. To some extent, this may have something to do with the behavior of other companies. Different from the theories of industrial cluster and social network, this paper explores the influence of other enterprises′ innovation behaviors on enterprises′ innovation behaviors from the perspective of peer effect.Based on the data of A-share listed companies which belongs to the industrial enterprises from the year 2006 to 2014 in China, this paper used the classical linear in means model and empirically tested the influence of peer effect on corporate R&D investment behavior. The results show that, firstly, the peer effect is an important external force to stimulate enterprises′ R&D investment. Enterprises will pay close attention to the innovation behavior of peer firms and take them as the reference point and driving force for themselves′ strategic decision of innovation. When the peer firms spend more on R&D in the previous year, the enterprises will increase themselves′ R&D investment in the coming year. Just because of the pull and catch-up effect between peer firms, the corporate R&D investment intensity is still increasing steadily despite the lack of profitability. Secondly, the more intense of the industry competition, the more susceptible that the enterprises′ innovation investment to the peer effect. Enterprises will pay more attention to the innovation strategies of competitors in highly competitive market. Fierce competition will prompt enterprises to respond more actively to the innovative behaviors of competitors. The more intense of the market competition, the higher the possibility of enterprise innovation strategy free riding. Thirdly, compared with enterprises with higher market position, the R&D investment of enterprises with lower market position are more susceptible to the peer effect. They are true followers. The innovation behaviors of the peer firms are valuable signals, which help the followers reduce the uncertainty of innovation investment. What′s more, imitating their peer firms especially the industry leaders are effective way for followers to catch up and surpass their competitors.In order to test the robustness of the above results, a series of supplementary tests were carried out. First, the linear mean model may have reflection problem. Specially, peer firms′ behavior can influence the firm′s decision, meanwhile, the firm′s behavior can also affect its peer firms′ decision. This simultaneous movements like a man and his reflection in a mirror. It′s hard to tell whether this mirror image cause the person′s movements or just reflect them. This simultaneity implies that the linear in means model may have endogenous problem. To solve this problem of endogeneity bias, this paper adopted the instrumental variable approach and choose peer firms′ idiosyncratic equity return shocks as an instrument for peer firms′ R&D investment intensity. The peer firms′ idiosyncratic equity return shocks is related to the peer firms′ R&D intensity but have nothing to do with the firm i′s R&D intensity, which means this instrumental variable has good representativeness. After testing, we know that this instrumental variable is not a weak instrumental variable. The empirical results after considering endogeneity are consistent with those before. Second, in addition to mutual imitation, the consistency of corporate innovation behavior may also be influenced by common factors. For example, industrial policies may make enterprises produce consistent expectations, which leads to the convergence of enterprises′ innovation behaviors. Although the correlation factors such as industry and year are controlled in the classical linear in means model, they may not be enough to explain all the common factors and then lead to biased conclusions. In order to improve the robustness of the research conclusions, samples which are greatly affected by industrial policies during the sample period will be removed and the regression test will be conducted again. As the national industrial policy is mainly inclined to strategic emerging industries during the sample period, we believe that strategic emerging industries are greatly affected by national industrial policy factors and 367 enterprises in strategic emerging industries are excluded from our sample. The results based on the remaining samples are consistent with the above conclusions. Third, consistent with most existing literatures, we used the new model which exclude contextual effect factors and re-examined the samples. The findings are not change. To sum up, after endogeneity test and a series of robustness tests, the above conclusions are still hold.This study contributes to better understanding of the external driving force of corporate innovation investment and provides some helpful ideas for stimulating corporate innovation. First, in order to stimulate enterprise innovation, the government should focus on fostering the industry competition atmosphere. Industry competition can promote innovation competition among the peer firms, and then improve the innovation level of the whole society. Second, the enterprises with lower market position are more susceptible to the peer effect means that most listed companies in China are still used to adopting the passive innovation strategy. Therefore, in order to lead more enterprises to investment in innovation with high efficiency and high quality, the government can set up some innovation models in various industries and widely publicize and greatly reward these enterprises. Through the demonstration and driving role of these benchmark enterprises, the innovation level of the whole society will be improved gradually. Last but not the least, in addition to seeing the positive side of peer effect, the government should be alert to the problem of inefficient investment under the influence of the peer effect at same time.

Key words: corporate innovation, peer effect, R&D intensity