Page 10 - IDEA Study 8 2017 Direct subsidies and R&D output in firms

(“crowd-in”) or substitute (“crowd-out”) private innovation inputs, such as R&D investments, and outputs, such as patents. In this sense output additionality is one of the most expected outcomes of direct public R&D interventions, which is why patents counts are considered in this respect. The existing empirical evidence on the effectiveness of R&D subsidy programmes is mixed and suggests that effect of public R&D programmes depends on many factors, including but not limited to the firm’s and R&D project’s characteristics, the firm’s past behaviour, the industry, the programme selection criteria, the subsidy amount and subsidisation rate, and the time since the subsidy under consideration (Zúñiga-Vicente et al., 2014). Moreover, the programme’s average success may be driven by a few successful projects, while there could be many cases without any sizable effect (Cunningham et al., 2012). Empirical findings for R&D subsidy programmes differ across countries, industries and programmes. Czarnitzki & Licht (2006) found a large degree of additionality in public R&D grants regarding patent applications in Germany. Hsu et al. (2009) found output additionality effects for R&D programmes in Taiwan, although they point out that patenting behaviour depends on evaluation criteria and differs across industry sectors. Cerulli and Poti (2012) found output additionality effects of publicly funded R&D expenditures on patent applications made by Italian firms, which are more oriented towards patenting and have lower decline of fixed capital. Other firms tend to exhibit crowding-out effects. Fornahl et al. (2011) found little evidence that R&D subsidies affect German biotech firms’ patent performance. The effect was somewhat larger for joint projects with two or more partners. Based on data for Italian firms, Catozzella & Vivarelli (2011) found that innovation subsidies had a negative effect on innovation productivity, measured as innovative sales over innovative expenditure. Based on Spanish data, Gelabert et al. (2009) found some evidence that subsidy programmes had crowding-out effects on appropriability (patents, models, trademarks, etc.) for firms with high levels of IP protection usage. Hence, there are estimates of both positive (“crowding-in”) as well as negative (“crowding-out”) effects and whether public R&D subsidies deliver the desired positive output additionality effects is an essentially empirical question. 8