Into Thin Air: Using a Quantile Regression Approach to Explore the Relationship Between R&D and Innovation

Applying quantile regression to 760 Finnish firms, we show that the relationship between R&D and firm performance is less straight forward than so far assumed. OLS regression analysis fails to capture the effect of R&D expenditure at different locations on the performance distribution. We reveal that R&D matters, especially on the medium quantiles, while regressing against the upper quantiles of the economic gains from innovation distribution exhibit decreasing returns scale in R&D. Our results confirm that Gaussian statistics fail to capture the most interesting part of the distribution – namely the extreme observations located in the tails.