中文摘要
既有研究表明，研发税收抵免会增加研发支出。我们利用美国各州研发税收抵免的交错实施，考察其对创新本身的影响。具体而言，我们考察了十项常见的专利特征，这些特征在现有研发激励文献中很少受到关注或完全未获关注。我们的实证方法表明，研发税收抵免降低了研发的使用者成本并增加了研发支出，但我们没有发现总体证据表明，此类抵免增加了实施抵免政策州的专利申请量。税收抵免也没有提高以专利引用衡量的专利科学质量。另一方面，研发税收抵免提高了专利的新颖性，并且我们观察到实施抵免政策州的专利市场价值出现了幅度较大且显著的增长。我们的所有总体结果均由实际抵免力度较大的州和规模较大的企业所驱动，因为此类企业创造了绝大多数专利。这些结果对研发公共政策具有重要意义。
Abstract
Previous work suggests that research and development (R&D) tax credits increase R&D expenditure. We exploit the staggered adoption of state-level R&D tax credits in the United States to examine their effect on innovation itself. In particular, we consider ten commonly-studied patent characteristics that have received little or no attention in the extant literature on R&D incentives. Our empirical approach suggests that R&D tax credits reduce the user cost of R&D and increase R&D expenditure, but we find no aggregate evidence that such credits increase patenting in credit-adopting states. Nor do credits increase the scientific quality of patents, as captured by patent citations. On the other hand, R&D tax credits increase patent novelty and we see large and significant increases in the market value of patents in credit-adopting states. All of our aggregate results are driven by states with larger effective credits and by larger firms, because such firms produce the vast majority of patents. These results have important implications for R&D public policy.