The paper studies the medium-term effect of being a persistent or an occasional innovator on firm economic return within a “counterfactual” setting. Compared to previous studies on this subject the novelty of this paper is above all methodological. We make use of a random coefficient model estimated within a treatment evaluation setting. Compared to previous more standard regression analyses this approach allows us to assess not only the punctual effect of a persistent/occasional innovation strategy on the “operating profit margin” (OPM), i.e. the so-called “average treatment effect” (ATE) or “average treatment effect on treated” (ATET), but the “entire distribution” of this causal effect over companies as a function of their observable characteristics. Relying on a “distribution” rather than on one single statistical moment (such as the “mean”, as usual) brings with it a great amount of further information about the “heterogeneous” response of firms to the innovative event.

We exploit a longitudinal dataset of Italian manufacturing firms with more than ten employees, where 451 companies are observed for 9 years, from 1998 to 2006. We obtained it by performing a merge of the last three waves of the Capitalia/Unicredit survey (8^th,9^th and 10^th survey).

When compared with a non-innovative strategy, our results show a strong better economic performance of the group of firms that continuously and directly implement their innovating capacity and output instead of resting on past acquired skills. Also occasional innovation produces good OPM differential outcomes and we estimate a difference with the persistent behavior of about only 1 percentage point lower. Nevertheless we have found that, at least at a descriptive level, persistent innovation allows for a dynamic advantage against occasional “one-time-only” innovative behaviors, an aspect that deserve further evidence and attention especially when looked through an evolutionary-Schumpeterian analytical lens.