Analysing capital structure of spanish chemical companies using self-organizing maps
Abstract
Purpose
This paper aims to analyses the capital structure of the Spanish chemical industry during the period between 1999 and 2013, with a twofold objective. First, to determine whether the assumptions of pecking order theory are fulfilled throughout the study's timeframe. Second, by using data covering the years before the crisis and the worst years thereof, this study shows how the crisis has affected the capital structure of the companies included in this sample.
Design/methodology/approach
A particular kind of unsupervised neural network, self-organizing maps, is applied. This methodology allows to cluster firms avoiding the need to establish relationships between the different variables involved in the problem beforehand.
Findings
Companies are clustered into groups with different degrees of accomplishment of the pecking order theory. The hypothesis about risk is the one that experience a greater variation in the period before and after the crisis. Moreover, companies' capital structure has been lightly disrupted by the crisis.
Originality/value
The originality of this paper lies in applying an unprecedented methodology to the problem of capital structure. Therefore, the capital structure problem can be approached without setting any function relationship previously.
Keywords
Citation
Camara-Turull, X., Fernández-Izquierdo, M.Á. and Sorrosal-Forradellas, M.T. (2017), "Analysing capital structure of spanish chemical companies using self-organizing maps", Kybernetes, Vol. 46 No. 06, pp. 947-965. https://doi.org/10.1108/K-05-2016-0112
Publisher
:Emerald Publishing Limited
Copyright © 2017, Emerald Publishing Limited