R&D subsidy output additionality: Evidence from programmes interaction and learning effects

Transcription

1 R&D subsidy output additionality: Evidence from programmes interaction and learning effects OLEG SIDORKIN AND MARTIN SRHOLEC CERGE-EI, PRAGUE (Draft: ) Extended abstract Introduction The relevance of public support for business R&D is a highly-debated topic (Cunningham et al. 2012). As government takes an active part in financing business R&D activity, there is a growing need to understand how effective these subsidies are and whether they achieve their goals. Policy makers and taxpayers worry that public funds are used to support R&D projects, which firms would carry out anyway, hence that the additionality effect is low (David et al and Klette et al. 2000). The additionality effects, in the sense of input, behavioural and output additionality, have been studied extensively in the empirical literature (García-Quevedo 2004, Cerulli 2010, Zúñiga-Vicente et al. 2014). Yet the vast majority of existing studies do not look beyond the main and direct effects of a single subsidy programme. Admittedly, there remains a number of important but under researched issues to be addressed by empirical research on this topic. Usually there are multiple funding schemes in place at the same time and similar policy instruments are implemented one after another. However, their interactions, both cross-sectional and intertemporal, have been seldom examined in the existing literature (Crespi et al. 2011, Cunningham et al. 2013). How much are output additionality effects of one programme affected by the same firms participating in other programmes? Do synergic or rather substitution effects prevail? Are there intertemporal learning effects at the firm-level? Are the interaction effects long-run or short-lived? Does the outcome depend on differences in the programme design, providers and policy coordination? Addressing these concerns and thoroughly evaluating R&D subsidy programmes is important for policy credibility, policy learning and maximizing the desired social impacts. The aim of this paper is to help with closing this research gap. Using extensive micro dataset that merges administrative data on R&D subsidy programmes participation with patenting records, balance sheets and income statements data and business register, we examine output additionality, synergy and persistence effects of eight main programmes for direct public support of business R&D in the Czech 1

2 Republic. More specifically, we consider effects of the IMPULS, TANDEM, TIP, PROGRESS, Projektova konsorcia and Trvala Prosperita administered by the Ministry of Industry and Trade (MIT), EUREKA administered by the Ministry of Education, Youth and Sports (MEYS) and ALFA administered by the newly established Technology Agency of the Czech Republic (TA CR); the combined duration of which spans over the period As the R&D subsidy programmes have different duration and partly overlap, it is possible to explore whether they complement or substitute each other for firms participating in multiple programmes. In addition, we focus on persistence effects of public R&D funding and explore the long-lasting effect of the programmes on innovation output. Research design As noted by Cunningham et al. (2013), an important synergy and persistence effects of R&D policy mix come from the fact that firms might receive multiple treatments. On one side, firms learn how to get public support for their activities (public R&D funding as well as other forms of support, which are not related to innovation directly). The persistence of firms participation in R&D subsidy and tax credit programmes is found, for example, by Busom et al. (2017). On the other side, the multiple treatments might reinforce or weaken the effects of public interventions. Neicu et al. (2016) found stronger behavioral additionality effects of R&D tax credits in the presence of R&D subsidies. In case of small and medium enterprises Radas at al. (2015) did not find any synergy between subsidies and tax credits in terms of R&D orientation or innovation output, although these results might be driven by relatively small size of firms, which received multiple treatments. Although effect persistence of public R&D interventions is of high importance, especially for policy makers, formal public evaluations as well as academic scholars rarely discuss the resulting long-term effects and focus mainly on immediate one-off effects (Cunningham et al., 2012). Another important question is whether the longterm effects of multiple public R&D instruments are stronger and have a longer duration than single-shot interventions. For example, according to Arqué-Castells and Mohnen (2015), the positive effect of R&D subsidies may take up to seven years. However, we are not aware of any studies which would study long-term additionality effects of firms, participating in multiple R&D incentive programs. Here we see a gap, which we could possibly address. The goal of this paper is to study the persistence of output additionality along with interaction effects of R&D subsidy programmes. In this paper, we focus on testing the following hypotheses: 1) Positive output additionality effects rapidy increase (short-term effects) and then fade away several years after the subsidy expires (long-term effects are limited). 2) Output additionality effects of individual subsidy programmes are not independent of each other. Firms participating in multiple subsidy programmes at the same time, especially with diverse designs and aims, entail stronger positive effects with longer duration. 2

3 3) Firms learn from participation in R&D subsidy programmes, i.e. a past participation in public subsidy programmes increases the output additionality effects of future programme participation, thus there is a stronger long-term effect on repeatedly treated firms. Methodology and data Non-random assignment of R&D subsidies, makes it problematic to draw conclusions about subsidy effects only by comparing data on firms who received and did not receive subsidies (Zúñiga-Vicente et al., 2014). Firms, which received subsidies are different from firms which did not, therefore it is pertinent to separate effect of subsidy selection (selection bias) from the actual effect of subsidies (average treatment effect on the treated). For this purpose, we use non-parametric propensity score matching techniques (Angrist and Pischke, 2009) and as a robustness check conditional difference-in-differences model (Blundell and Costa Dias, 2000). The analysis is based on a large firm-level dataset constructed by merging micro data from several sources. First, the public support history for each firm relies on administrative data from the Research and Development and Innovation Information System of the Czech Republic ( which provides complete information about public tenders, subsidy providers, programmes, individual projects, subsidy receivers as well as their outputs in the field of R&D policy. Second, data on IP applications is derived from European Patent Office s PATSTAT database (Spring 2016), which is the largest international database of 90 million intellectual property documents. Finally, financial and structural data about the firms is extracted from Bureau Van Dijk s Amadeus databased, which provides balance sheets and income statements data of Czech firms with 25 and more employees. Expected results So far, we found evidence in support of the thesis that output additionality (crowd-in effects) increases within first three years after receiving subsidy. Business enterprises participating in the IMPULS, TIP and ALFA programmes registered 5-12% increase in the probability to apply for IP protection within one to three years after receiving the subsidy. Nevertheless, the output additionality effects hold only for Czech IP protection only. In contrast, the effects on international IP application probability were weak and statistically insignificant. We are going to expand the analysis to the other relevant subsidy programmes cited above and extend the time span to at least four and five years find out whether the additionality effects weakens over time. We also confirmed that participation in public R&D subsidy programmes in the past increases the probability of receiving an R&D subsidy in the future. Moreover, the effect fades away with lags, that is the most recent experience with public funding is more important and there is a persistence among subsidy receivers. Here our goal is to add lagged effects of participation in different types of public support programmes and investigate the interactions thereof with regards to the output additionality effects. Currently we work on estimating full synergy and 3

4 persistence effects of the R&D subsidy programmes interactions in other to test the main the hypotheses. References Angrist, J. D., & Pischke, J.-S. (2009). Mostly harmless econometrics an empiricist s companion. Princeton: Princeton University Press. Arqué-Castells, P. & Mohnen, P. (2015), Sunk Costs, Extensive R&D Subsidies and Permanent Inducement Effects. The Journal of Industrial Economics, 63: DOI: /joie Blundell, R. and Costa Dias, M. (2000). Evaluation methods for nonexperimental data. Fiscal Studies, 21, Busom, I., Corchuelo, B. & Martínez-Ros, E. (2017) Participation inertia in R&D tax incentive and subsidy programs. Small Business Economics, 48(1): DOI: /s Cerulli, G. (2010) Modelling and Measuring the Effect of Public Subsidies on Business R&D: A Critical Review of Econometric Literature. Economic Record, 86, Crespi, G., Maffioli, A., Mohnen, P. and Vázquez, G. (2011) Evaluating the Impact of Science, Technology and Innovation Programs: a Methodological Toolkit. Impact=Evaluation Guidelines, Technical Notes, No. IDB-TN-333, Inter-American Development Bank. Cunningham, P., Gök, A., & Larédo, P. (2012). The Impact of Direct Support to R&D and Innovation in Firms. NESTA Working Paper No. 13/03. Cunningham, P., Edler, J., Flanagan, K., Laredo, P. (2013). Innovation policy mix and instrument interaction: a review. NESTA working paper No. 13/20. David, P. A., Hall, B., Toole, A. A. (2000) Is public R&D a complement or substitute for private R&D? A review of the econometric evidence. Research Policy, 29, García-Quevedo, J. (2004) Do Public Subsidies Complement Business R&D? A Meta-analysis of the Econometric Evidence. Kyklos, 57, Klette, T. J., Møen, J., Griliches, Z. (2000) Do subsidies to commercial R&D reduce market failures? Microeconometric evaluation studies. Research Policy, 29, Neicu, D., Teirlinck, P. & Kelchtermans, S. (2016) Dipping in the policy mix: Do R&D subsidies foster behavioural additionality effects of R&D tax credits? Economics of Innovation and New Technology, 25(3): Radas, S., Anić, I.-D., Tafro, A., & Wagner, V. (2015) The effects of public support schemes on small and medium enterprises. Technovation, 38:

5 Zúñiga-Vicente, J. A., Alonso-Borrego, C., Forcadell1, F. J. and Galán, J. I. (2014) Assessing the Effect of Public Subsidies on Firm R&D Investment: A Survey. Journal of Economic Surveys, 28,