The persistence of innovation: a panel data investigation on manufacturing firms

This paper represents a contribution to empirical debate on the persistence of innovation in the firm, by exploiting an innovative panel database that, for the first time, links three waves of the Italian Community Innovation Survey with an administrative data source providing economic and financial information for firms in the Italian manufacturing sector, 1996-2003. By using both a dynamic logistic model and a Granger causality approach, we show that in order to innovate successfully it is much more important to have an adequate flow of profits during an appropriate time span rather than high profits only during one period before innovation. Our causality tests prove the existence of a dynamic interaction between innovation and profitability: successful innovation can, in the short run, generate the profitability conditions that can then enhance the financial resources needed to reinvest in new technological opportunities, thus causing the firm to persist in its innovative behaviour. We have also shown that another important source of persistence is represented by past innovative experience. A firm with consolidated innovative behaviour would have a higher probability of future successful innovation with respect to a firm that occasionally (or accidentally) innovates. Persistence in innovation enables a firm to take advantage of substantial technological and organizational learning effects, which improve with time.