Public R&D support and firm performance: A multivariate dose-response analysis

•Output additionality for R&D-starters and R&D-experienced firms.•Multivariate dose-response of all public R&D instruments to business enterprise sector.•Analyse the intensive and the extensive margins, and the policy instrument mixture.•No effects of R&D support on labour productivity or return on assets.•R&D instruments should favour R&D-starters over R&D-incumbents. We analyse all the major sources of direct and indirect research and development (R&D) support to the business enterprise sector in a single country, Norway, for the period 2002–2013, treating the financial support for R&D from several instruments as a multivariate dose exposure. The output additionality of support to incumbent firms that regularly perform R&D (R&D-incumbents), which obtain about 65 per cent of all R&D support to business enterprises, is insignificant for any instrument or policy mixture. However, the estimated additionality of support to R&D-starters (firms without prior R&D activity), which obtain about 30 per cent of all R&D support, is generally positive. In this firm category, the main instruments for direct R&D support in Norway generate significantly less output and economic activity per NOK 1 million in support than do tax credits, despite the fact these instruments manage large project portfolios at considerable administrative costs. We do not identify positive effects of R&D support on labour productivity or the return on assets for any of the instruments. Our main policy implication is that R&D instruments for the business enterprise sector should be designed in favour of R&D-starters over R&D-incumbents, that is, shifting the focus from the intensive to the extensive margin.