Innovation and Productivity in Irish Firms

Abstract

This paper uses an endogenous switching technique which allows us to utilize micro-econometric data to construct counterfactual scenarios of the innovation-productivity relationship in Irish firms. A firm's innovation effort, capital intensity, firm size, location and its operating environment are key variables in explaining a firms' propensity to innovate. However, the importance of these factors differs across innovation types. We find mixed results on the effect of innovation on the productivity of innovators across innovation types. The results indicate in the counterfactual analysis that all types of innovation have a positive effect on the productivity levels of non-innovating firms.

Résumé

La présente communication utilise une technique de communication endogène, qui nous permet d'utiliser des données micro-économiques grâce auxquelles nous sommes en mesure de constituer des scénarios imaginaires du rapport innovation—productivité dans des entreprises en Irlande. Les initiatives, l'intensité en capital, la taille de l'entreprise, sa situation géographique, ainsi que son environnement d'exploitation, sont des variables clés pour expliquer la tendance à l'innovation d'une entreprise. Toutefois, l'importance de ces facteurs varie selon les types d'innovation. Nous établissons des résultats mitigés concernant l'effet de l'innovation sur la productivité d'innovateurs dans différents types d'innovations. Les résultats indiquent, dans l'analyse imaginaire, que tous les types d'innovation ont un effet positif sur les niveaux de productivité d'entreprises non innovantes.

Resumen

este estudio usa una técnica de sustitución endógena, que nos permite utilizar datos microeconométricos con los que podemos construir escenarios hipotéticos de la relación entre innovación y productividad en las firmas irlandesas. El trabajo de innovación de la firma, la intensidad del capital, el tamaño de la firma, su ubicación y entorno de explotación son variables clave para explicar la propensión a innovar de la firma. Sin embargo, la importancia de estos factores difiere entre los tipos de innovación. Encontramos una mezcla de resultados sobre el efecto de la innovación en la productividad de los innovadores entre los tipos de innovación. Los resultados indican en el análisis hipotético que todos los tipos de innovación tienen un efecto positivo sobre los niveles de productividad de las firmas no innovadoras.

摘要

本文运用了内源性切换分析方法。通过该方法，我们能够用微观计量经济学数据构建爱尔兰公司的创新与生产力之间关系的反事实情景。一个企业的创新努力、资本密集度、规模、地理位置及其经营环境是确定企业创新倾向的关键要素。然而，这些因素对于不同类型的创新来说，重要性不尽相同。我们发现，创新努力对于创新者生产力的影响随不同的创新类型而异。反事实分析结果表明，各种类型的创新都能对非创新性企业的生产力水平产生积极的影响。

Keywords:

• Innovation
• productivity
• endogenous

JEL Classification:

• 031
• 032

The authors would like to thank Maria Abreu who made valuable comments on an earlier version of this paper at a RSAI-British and Irish Section Annual Conference.

The authors would like to thank Maria Abreu who made valuable comments on an earlier version of this paper at a RSAI-British and Irish Section Annual Conference.

Notes

1. All of these studies refer to product innovation in manufacturing firms.

2. The BEEP's series have been principally employed to examine aspects of the business environment in developing countries and to examine the impact of criminal activity and corruption on business development in developing countries. In terms of innovation studies, Roper (2009) has a working paper examining innovation policy in the Balkan countries. Aghion et al. (2002) examined competition, innovation and economic growth in transition countries. Ş eker (2012) conducted a study on importing, exporting and innovation in developing countries. A detailed analysis of the data collection method may be obtained in Synovate (2005).

3. On the theoretical grounds outlined in the literature section between the differences between product and process innovation we argue that the innovation process for new products and services and upgraded products and services is likely to be similar (firms try to differentiate their product/service from other products and services in the market place) and hence we have combined these measures. The correlation between the measures for new and improved products and services is 42%.

4. In order to account with issues of specification and degrees of freedom there are less independent variables specified in the regime equations to those in the selection equation.

5. In this case, other innovation efforts by the firm may be spending on R&D and/or introducing a process innovation and/or introducing an organizational innovation.

6. In this case, other innovation efforts by the firm may be spending on R&D and/or introducing a product innovation and/or introducing an organizational innovation.

7. In this case, other innovation efforts by the firm may be spending on R&D and/or introducing a product innovation and/or introducing a process innovation.

8. We decided not to include the results from the innovation effort models and the factors affecting productivity levels (other than the effect of innovation) in the main results section as we are specifically interested on the drivers of product, process and organizational innovations and their effects and counterfactual effects on the productivity of innovating and non-innovating firms in this study. The estimations for the innovation effort step indicate that the processes explaining R&D activity, process innovations, and organizational innovations are dependent. Hence, we use predicted values for innovation effort in the selection equation of the endogenous switching model. Furthermore, the innovation effort of firms is not found to have a significant effect on the productivity levels of innovating or non-innovating firms. Results from these steps are available from the authors on request.

9. These variables were excluded from the final regression estimates. The result estimations that include interaction variables measuring the regional and inter-firm characteristics are available from the authors on request.

10. A likelihood ratio (LR) test of independent equations is inherently estimated with the Movestay command in Stata. The LR test indicates that we can reject the null hypothesis (at the 1% level) that the equations measuring process innovation and productivity are independent and we cannot reject the null hypothesis (prob > χ² = 0.000) that the equations measuring product/service innovation and productivity and organizational innovation and productivity are independent (prob > χ² = 0.340 and prob > χ² = 0.7752 respectively).