A reexamination of the firm innovation process: sensitivity to sample and estimation methods

In this article, we re-examine the innovation process through the CDM model. Compared to the existing literature, this study offers several contributions. First, it relies on an unusually large dataset of 509,033 firms from nine European countries – Bulgaria, the Czech Republic, Estonia, Spain, Hungary, Lithuania, Portugal, Romania, and Slovakia – over the period 1998–2016. This extensive dataset allows us to explore cross-country heterogeneity, as well as potential temporal trends across multiple survey waves. Second, the paper provides a systematic and detailed review of the vast CDM literature, offering a structured synthesis of prior findings and highlighting the main areas where results diverge across studies. Third, methodologically, we compare three alternative estimation strategies, which enables us to evaluate the robustness of our findings and to identify potential sources of heterogeneity in estimated relationships. Across all specifications, we find that R&D investment has a positive effect on the share of new products in sales, which subsequently enhances firm performance. Promoting innovation can have a substantial impact on performance. However, the magnitudes of these effects vary depending on the country, the estimation method, and the treatment of potential biases. In some countries, innovation generates stronger positive spillover effects on firm performance, while others are more effective at transforming R&D into innovation but face challenges in converting this innovation into productivity gains. This implies, on the one hand, that public policies must be context-specific, and on the other hand, that the choice of estimation method and the treatment of potential biases can significantly affect the robustness and validity of the results.