Measuring non-R&D drivers of innovation : the case of SMEs in lagging regions

In order to better capture non-R&D based processes related to Learning by Doing, Using and Interacting (DUI) as a basis for policy advice, this paper empirically identifies DUI mode drivers of SME innovation. For the first time, a large set of conceptually derived indicators is used in a self-conducted survey. Using lasso regression as a data-driven selection technique capable of handling such a large number of potential predictors, we find that DUI learning involves a wide range of elements beyond interaction with external actors. Moreover, our results suggest that the relevance of DUI learning for predicting SME innovation depends on both the region and the type of innovation output. SME innovation in lagging regions is strongly related to the DUI mode, which is particularly pronounced in the case of intra-firm learning processes. These results suggest that R&D capacity is not the only main driver of SME innovation, especially in lagging regions, and therefore provide an indication of how firms can compensate for unfavourable conditions in their regional innovation environment. This in turn implies going beyond innovation policy in the narrow sense to a more holistic approach that may include links with other policy areas.