Are there non-linearities between SME growth and its determinants? A quantile approach

This article uses quantile regressions to study the relationship between the growth of Portuguese small and medium enterprises (SMEs) and its determinants. To control a sample selection bias, we use a two-step estimation. In the first stage, we verified that past growth, size, age, R&D intensity, and cash flow are important determinants for the survival of Portuguese SMEs. Then, after controlling for possible bias of data, we now find that relationships between determinants and growth in Portuguese SMEs present significant non-linearities over growth distribution: (i) Gibrat's Law cannot be rejected in the context of reductions in the size of Portuguese SMEs, but is rejected when company size increases; (ii) debt and R&D intensity are restrictive determinants of Portuguese SME growth when there is reduction in size, being determinants promoting growth in the case of increased size; (iii) age is a restrictive factor of Portuguese SME growth only when there is an increase in size; and (iv) cash flow, although a determinant promoting Portuguese SME growth regardless of the level of growth, takes on greater importance when there is a reduction in size. Copyright 2010 The Author 2010. Published by Oxford University Press on behalf of Associazione ICC. All rights reserved., Oxford University Press.