FIRMS GROWTH, SIZE AND INNOVATION AN INVESTIGATION INTO THE ITALIAN MANUFACTURING SECTOR

Abstract

This article aims at understanding the determinants of Italian small- and medium-sized enterprises’ (SMEs) turnover growth having in mind the fact that the Italian economic system relies substantially on small firms which have traditionally managed to stay competitive by adopting strategies such as the creation of well-integrated social and institutional clusters or specialising in the production of quality goods (the so called Made in Italy). However, the growing pressure coming from the Far East has rendered this production system vulnerable, challenging its international competitiveness. Building on a conceptual model, we found that, on average, young firms are more likely to experience positive growth; moreover, turnover growth is positively associated with firms’ size, process innovation, product innovation and organisational changes. In contrast, marketing innovation does not considerably affect Italian SMEs growth. When restricting our focus to a sub-sample of innovative firms, we found that those firms investing directly in innovating activities are almost 30% points more likely to experience positive growth, which is significantly affected also by workers and managers’ re-qualification. Finally, among innovative firms, process innovation and organisational changes are, by far, the most influential innovating strategies. The model was tested using a unique database which collects data for the year 2004, over a sample of 2600 SMEs.

Keywords:

• SMEs
• Growth
• Innovation
• Ordered probit

Keywords:

• JEL Classification
• L1
• O31
• C24

Acknowledgements

The authors are grateful to Cesare Imbriani, Gustavo Crespi and John Micklewright for their helpful comments. The usual disclaimers apply.

Notes

1 Note that throughout this article SMEs are defined as firms with less than 250 employees. Specifically, we shall define micro-firms as those with less than 10 employees; small firms as those with more than nine employees but less than 50; and medium firms as those with more than 49 employees but less than 250.

2 A detailed list of all used variables is presented in Section 3.2.

3 Note that a less than 5% increase or decrease in turnover is considered as an unaltered turnover.

4 In the survey, firms are classified into 10 sectors. We restrict our analysis to nine sectors dummies plus one latent sector (clothing), which serves as base.

5 Note that, given the available data, we select independent variables to have a one-to-two years time lag between the date at which firm innovative strategies are observed and the date at which its potential impact on turnover change is observed. Moreover, for the sake of clarity, we report simple descriptive statistics for the full set of independent variables in Table A1 in the appendix.

6 The ordered probit model, in its general form, is discussed in Wooldridge (2002) and Maddala (1983).

7 The variance–covariance matrix of the ordered probit model is corrected for heteroscedasticity using a ‘sandwich’ estimator (Huber, 1967).

8 Firms were classified as innovators or non-innovators on the basis of their answer to the following question: ‘Did your firm introduce any innovation in product, in processes, in organisation and in marketing during the last two years?’

9 The pseudo-residuals (McFadden R²) from the ordered probit model are used to inform on the adequacy of the estimated models. Machin and Stewart (1990) provide the computational details for the pseudo-residuals for the ordered probit model. For a detailed description of these indexes, see: http://www.gseis.ucla.edu/courses/ed231c/notes3/fit.html.

10 Note that, in the interpretation of the role of firms’ size, it might be argued that there is a possible problem of endogeneity. More precisely, one might think that firms experienced turnover growth decided to increase the scale of production by expanding the number of workers (in the short term) and the capital investments (in the long term). In this article, we support the view that this relation is not so obvious for two main reasons. First, the economic theory suggests that industries might be characterised by increasing, decreasing or constant returns to scale, hence it is possible to conclude that this problem is particularly pertinent for those industries characterised by increasing returns to scale. Secondly, since we are using categorical data, it is very likely that firms do not move from one size-class to another over the considered period. This last assumption is supported by the fact that half of the firms declared not to have recorded employment change over the period 2001–2002 (see Table A2 in the appendix).

11 Recently Onida (2004) pointed out how Italian private banks have typically adopted an asset-based lending strategy towards SMEs. However, under the pressure of the Basel 2 agreements, which aim to make the framework more risk-sensitive and representative of modern banks’ risk management practices, lending strategies are shifting towards a cash-flow based approach.

12 The different response rates were tested using Pearson χ², which resulted statistically significant.

13 Marginal effects indicate the percentage point change of probability if the value of the indicator variable changes from zero to one.

14 Note that models 1 and 2 are not strictly comparable as some variables are not observed on the whole sample. In order to allow comparisons between the whole sample and the innovative firms’ sample, we estimated model 1 on the innovative firms’ sample. The results of this further estimation are reported in the appendix (Table A3). It is worth noting that the core differences observed by comparing models 1 and 2, and hereafter discussed, are fully confirmed by comparing model 1 with the findings reported in the appendix.

15 This definition is grounded on Pavitt's taxonomy (Pavitt, 1984).

16 In our analysis, we also investigated the interaction between process innovation and organisational change (see Table A4 in the appendix). It is interesting to note that the interaction term is negatively signed and significant at conventional level of 5%, suggesting that the Italian firms are more likely to benefit from process innovation and organisational change separately.