Abstract: With the increasingly fierce market competition and the rapid development of science and technology, R&D has been becoming an indispensable value-added activity. Considering long duration, high uncertainty, and mismatch between cost and income period, the accounting standard allows managers to capitalize development expenditures and record them on the balance sheet. According to statistics in our sample, 19.07% of the firm-year observations capitalize over 80%development expenditures, and the highest proportion of capitalization is 97.4% from 2007 to 2017. Then what is the impact of this phenomenon on firm innovation? Does high proportion of development expenditures capitalization enhance or hinder firm innovation? What is the mechanism? These questions remain unexplored in the previous literature.Theoretically, managers face earnings announcement pressure as the contractual system arrangement and capital market operation. In order to meet the earnings target, in addition to improving financial performance, managers usually conduct earnings management, such as cutting R&D investment. However, this will result in insufficient investment in R&D and is not conducive to the long-term development of companies. In 2007, the new accounting standards stipulate that firms can capitalize development expenditures that meet certain conditions. Under the capitalized accounting treatment, since development expenditures are included in the balance sheet, they do not affect the current earnings, mitigating managers′ tendency to cut R&D investment to meet an earnings target, increasing the enthusiasm for R&D investment. This can alleviate the problem of insufficient R&D investment, and thus promoting firm innovation. However, under high proportion of R&D capitalization, even if the initial R&D project is difficult to form intangible assets or the efficiency is lower than expected, managers might still choose to allocate funds to that project. This is because the abandon of a R&D project requires a one-time conversion of capitalized development expenditures into the income sheet, resulting in a great negative impact on current earnings. The higher the proportion of development expenditures capitalization, the greater the pressure to terminate the earnings announcements generated by the R&D project, and the greater the possibility of managers inefficient R&D investment. This will cause inefficient internal resource allocation and make firms lose the opportunity to invest in new R&D projects, which might hinder firm innovation. Based on the above analysis, although the impact of R&D capitalization on R&D investment is consistent in different situations, the effect of R&D investment outputs, firm innovation, might be different. Therefore, we predict that when the proportion of development expenditures capitalization is lower than the critical value, capitalization can alleviate the lack of R&D investment by reducing the earnings announcement pressure on the managers, and thus promote firm innovation. On the contrast, when the proportion of development expenditures capitalization is higher than the critical value, capitalization will encourage managers to over-invested inefficient R&D projects, and thus hinder firm innovation. In other words, the degree of development expenditures capitalization has an inverted U-shaped relationship with firm innovation.Using the data of Chinese A-share listed firms between 2007 and 2017, this paper examines the impact of R&D capitalization on firm innovation. Through quadratic function model regression, we find that there is an inverted U-shape relationship between the capital expenditures rate and firm innovation. Our results hold with several robustness tests, including replacement of firm innovation measurement methods, change of sample period and change of initial sample．The mediation effect test shows that when the capital expenditures rate is lower than the critical value, R&D capitalization improves firm innovation by inhibiting the manager′s incentive to cut the R&D investments in response to earnings considerations. In contrast, when the capital expenditures rate is higher than the critical value, R&D capitalization hinders firm innovation because managers are more likely to over-invested non-optimal R&D project to avoid a negative report consequence in the form of R&D asset impairment. Further examining the moderation effect of the earnings announcements pressure, we find that the inverted U-shape relationship between the development expenditures capitalization rate and firm innovation is more pronounced for firms with higher earnings announcement pressure.This paper contributes to the existing literature in two ways. First, we contribute to the stream of literature that examines the determinants of firm innovation. Previous studies mainly focus on the impact of equity incentives, corporate finance and personal characteristics of executives on firm innovation, but there is little literature on the impact of accounting policies on firm innovation. Accounting policies not only reflect business operations, but also have economic consequences that affect management decisions. Development expenditures accounting treatment is the most relevant policy choice for company innovation activities. Based on the perspective of capitalization of development expenditures, this paper explores the differences between firm innovation outputs in different development expenditures capitalization rate. Second, this paper enriches the research on the economic consequences of R&D accounting treatment choice. Previous studies investigate the economic consequences of research and development expenditure accounting choice from the perspective of information disclosure behavior or R&D investment decision-making. Different from their research, this paper examines the impact of development expenditures capitalization on firm innovation and its mechanism of action, and thus provides new evidence that development expenditures accounting policy can affect final outputs of firms′ value-added activities.This study not only contributes to literature on firm innovation and R&D accounting treatment choice, but also has significant implications for regulators and investors. Our findings show that when there is a large proportion of development expenditures capitalization, managers facing the pressure of earnings announcements are more likely to allocate funds to the continuing inefficient R&D project. Excessive capitalization of development expenditures will not only reduce the efficiency of internal resource allocation, but also convey the illusion of value-added activities to the capital market, which might mislead investors in decision-making. This has a strong inspiration for regulators to strengthen the regulation of accounting treatment choice for firm development expenditures. At the same time, it can remind investors that they cannot only focus on the promotion of R&D expenditure capitalization on firm innovation. It is necessary to understand the inhibition effect of excessive development expenditures capitalization on firm innovation, and to be alert to firms that have the trap of development expenditures capitalization.

Key words: development expenditure accounting policy, inefficient R&D investment, earnings announcement pressure