Firm persistence in technological innovation: the relevance of organizational innovation

Abstract

Organizational innovation favors technological innovation, but does it also influence persistence in technological innovation? This paper empirically investigates the pattern of technological innovation persistence and tests the potential impact of organizational innovation using firm-level data from three waves of French Community Innovation Surveys. The evidence indicates a positive effect of organizational innovation on persistence in technological innovation, according to the various measures of organizational innovation. Moreover, this impact is more significant for complex innovators, i.e. those who innovate in both products and processes. The results highlight the complexity of managing organizational practices with regard to the technological innovation of firms. They also add to understanding of the drivers of innovation persistence through the focus on an often forgotten dimension of innovation in a broader sense.

Keywords:

• organizational innovation
• technological innovation
• persistence

JEL Classification:

• O31
• C23
• C25
• L20

Acknowledgements

The authors thank the three anonymous reviewers of this paper, the participants of the special EINT workshop in Torino in March 2013 as well as those of the various seminars where the paper was presented (GATE and ESDES in Lyon, IREGE in Annecy, BETA in Strasbourg, CEPS in Luxembourg).

Notes

1. The concepts of state dependency and persistence are important in the analysis of outcomes over time. ‘Persistence’ describes whether a particular condition, innovation in our case, is brief or long-lasting, while ‘state dependency’ indicates whether the chance of experiencing a condition depends on having experienced the same condition in the past.

2. The CIS databank has been made available to two of the authors under the mandatory condition of censorship of any individual information.

3. NACE is the ‘statistical classification of economic activities in the European Community’ used uniformly by all member states. We classified manufacturing industries according to their global technological intensity with NACE Revision 1.1 for the t–2 and t−1 periods, whereas t was covered by NACE Revision 2, according to the Eurostat classification (http://epp.eurostat.ec.europa.eu/cache/ITY_SDDS/Annexes/hrst_st_esms_an9.pdfSDDS/Annexes/hrst_st_esms_an9.pdf).

4. Note that compared to the CIS2008 (2006–2008) population (Table 6), our sample shows a slight bias in favor of product innovators because the percentage of innovative firms is slightly superior in our sample (63% in our sample compared to 39% in CIS2008) to that of the CIS2008 population. This overestimation of innovative firms is quite logical and is due to the merging of three different waves (Evangelista 2000).

5. A methodological change between the CIS2006 and CIS2008 reintegrated ‘knowledge management’ back into ‘new business practices for organizing procedures’ for CIS2008.

6. The total amount of in-house R&D is given directly in CIS. The total amount of external R&D is a variable that we constructed from an average of three inputs: (1) the amount dedicated to the purchase of external R&D; (2) the acquisition of machinery, equipment, and software dedicated to R&D; and (3) the acquisition of external knowledge.

7. In addition, the individual average of firm size (Sizemean) is positive and significant for pure product innovators and negative and significant for pure process innovators, which indicates substantial correlations between these variables and unobserved individual heterogeneity.

8. In the first step, we also estimated simple models, assuming the absence of individual effects and exogenous initial conditions. The persistence parameters were positive and highly significant for all innovator profiles. However, in these unrealistic conditions, overestimation of the dependent variable is likely; the significance of the persistence parameters, therefore, does not mean that true persistence exists. Results are available on request.

9. CIS2006 reports four organizational practices: Business practices; Knowledge management; Work organization; and External relations. For harmonization with CIS2008, where only three organizational practices are reported, we decide to group CIS2006 Business practices and Knowledge management so that we obtain only one binary variable that we call ‘business practices’.

10. We interpret DivOrg as a measure of the diversity of organizational innovation. It should depict the diversity of new practices implemented by the firm.