The governments of most advanced countries offer some type of financial subsidy to encourage firm innovation and productivity. This paper analyses the effects of innovation subsidies using a unique Swedish database that contains firm level data for the period 1997-2011, specifically information on firm subsidies over a broad range of programmes. Applying causal treatment effect analysis based on matching and a diff-in-diff approach combined with a qualitative case study of Swedish innovation subsidy programmes, we test whether such subsidies have positive effects on firm performance. Our results indicate a lack of positive performance effects in the long run for the majority of firms, albeit there are positive short-run effects on human capital investments and also positive short-term productivity effects for the smallest firms. These findings are interpreted from a robust political economy perspective that reveals that the problems of acquiring correct information and designing appropriate incentives are so complex that the absence of significant positive long-run effects on firm performance for the majority of firms is not surprising.