Are young SMEs' survival determinants different? Empirical evidence using panel data

Abstract

Based on two subsamples of Small and Medium-sized Enterprises (SMEs), 495 young SMEs and 1350 old SMEs, using probit regressions, we investigate whether the survival determinants of young SMEs are different from those of old SMEs. The multiple empirical evidence obtained allows us to conclude that the survival determinants of young SMEs are considerably different from those of the old SMEs: (1) determinants related to scale effects, financial situation and macroeconomic situation are of greater relative importance in explaining young SMEs' survival; and (2) a determinant related to technological intensity is of greater relative importance in explaining old SMEs' survival. The results let us provide different guidelines for economic policy in general, and industrial policy in particular, in support of young and old SMEs.

Keywords:

• old SMEs
• panel data
• survival
• young SMEs

JEL Classification:

• C33
• G32
• L16
• L21

Acknowledgement

The authors gratefully acknowledge partial financial support from FCT, program POCTI.

Notes

¹ Studies about survival (Agrawal and Audretsch, 2001; Cabral and Mata, 2003; Esteve-Pérez and Mañez-Castillejo, 2008; Siriopoulos and Lalountas, 2008; Holmes et al., 2010; Vaona, 2010; Giovannetti et al., 2011) have neglected specific study of the difference between the survival determinants of young and old firms.

² Monetary variables are deflated according to inflation in Portugal. We construct a price index from 1999 to 2006; 2006 is taken as the base year.

³ We decide to introduce separately, and together, the variables measuring companies' intrinsic characteristics and macroeconomic variables in the regressions, since we can expect significant correlation between debt and interest rates. By doing so, we can check whether the possible collinearity between debt and interest rate means significant alteration to sign, magnitude and statistical significance of estimated parameters measuring relationships between debt and survival and between interest rates and survival.

⁴ Just as Blanco-Mazagatos et al. (2007), we consider the following dummy sector variables: (1) agriculture, (2) forestry and fishing, (3) construction, (4) manufacturing industry, (5) wholesale and retail trade and (6) services.

⁵ We consider dummy annual variables so as to check for possible macroeconomic impacts not measured by the interest rates and GNP on the survival of young and old SMEs.

⁶ As can be observed, whatever the regression estimated, the parameters measuring relationships between determinants and the survival of young and old SMEs are quite similar regarding the sign, magnitude and statistical significance of estimated parameters. We find that the problem of possible collinearity between explanatory variables will not be very relevant in this study, particularly concerning the possible collinearity between debt and interest rate.