It is believed due to the characteristics of public goods that private firms are undertaking R&D investments funded by themselves less than the optimal social level of R&D. Based on the market failure argument, governments of many countries are practicing fiscal policies to alleviate the market failure problem and to further private firms' R&D investments. Korean government has also been practicing both R&D subsidy programs and tax credits on R&D. The purpose of this report is, specially focusing on tax policy towards R&D, to measure the effectiveness of each policy tool on the R&D investments funded by private firms and to deduce some policy implications. 

Using a large number of firm-level panel data over 2002～2004, we estimate both user costs of R&D investments and how responsive the R&D investments funded by private firms are to changes in user costs and/or government subsidies. It is also shown under a number of assumptions and proxies for those unobserved details related to individual firms' tax and financial informations that the estimated user costs have approximation errors at most 10%.  

As we can expect as a priori, the empirical results show the increase in tax credits on R&D leads to the decrease in the user costs, which in turn increases the R&D funded by private firms. It is also found that R&D funded by firms responds more elastically to the changes in their user costs the lower the economic depreciation rate for current expenditures. In addition, tax-price elasticity for small & medium firms is more elastic than large firms. Even if the estimated coefficient on lagged dependent variable is insignificant for some cases, the long-run tax-price elasticity is quite larger for small & medium firms than large firms.  

on the other hand, we cannot find any statistically significant evidence for the hypothesis that the government subsidy programs increase R&D investments funded by private firms. Specially, the crowding-out effect of government subsidy is prominent for small & medium firms even if the magnitude itself is very small. Since the magnitude of crowing-out effect is not large, it could be said that the purpose of government subsidy programs has been accomplished as intended if government subsidy programs aimed to increase total R&D level. However, we have to draw the opposite conclusion if they were initially designed to increase R&D funded by individual firms.