Page 3 - IDEA Studie 07 2023 TACR

Study 7/2023
Are Subsidies to Business R&D Effective? Regression Discontinuity Evidence
from the TA CR ALFA Programme1
MAY 2023
MATĚJ BAJGAR, MARTIN SRHOLEC
Summary
• Governments subsidise business research and experimental development (R&D) to promote development of the economy, because externalities and information asymmetries inherent to the innovation process make private funding of these activities fall short of what is socially desirable. Nevertheless, how effective such subsidies are and whether they achieve their goals is an open question that needs to be studied empirically.
• This study leverages the state-of-the-art method of regression discontinuity (RD) that allows us to come very close to making causal inferences about the effects of subsidies, to find out whether the Technology Agency of the Czech Republic’s (TA CR) ALFA programme stimulated new business R&D inputs, outputs, and positive economic impacts that would not have happened otherwise.
• Our results show that the subsidies significantly stimulated R&D expenditures in small and medium size enterprises (SMEs), but not in large ones. In SMEs, the effect is strongly positive on both publicly and privately funded R&D, both during a subsidised project and afterwards, so there is evidence of persistent crowding in. For large firms, in contrast, the subsidy appears to have only changed the structure of R&D expenditure during the project – increasing funding from public sources at the expense of private ones and capital expenditures at the expense of current expenditures – and the effects largely fizzle out after the project expires.
1 This study represents the authors’ own views and not the official position of the Czech Academy of Sciences’ Economics Institute nor the Charles University Centre for Economic Research and Graduate Education (CERGE). An earlier draft of this study was presented to TA ČR on 24th May 2022. We would like to thank Dan Münich, Štěpán Jurajda and the staff of TA ČR, in particular Martin Bunček, for their invaluable comments. We also thank the Czech Statistical Office (CZSO) and TA ČR for providing access to their microdata. The authors have undertaken the econometric estimations in this study on the basis of a confidentiality agreement in place during Martin Srholec’s work for the CZSO, which was part of a collaboration on the OECD microBeRD project. The study was produced with support from the Czech Academy of Sciences as part of its Center for Research, Development and Innovation Analysis programme (RaDIAC). All remaining omissions and errors are our own.
1