Prediction of Greek company takeovers via multivariate analysis of financial ratios

Abstract

This case study evaluates the financial features of Greek firms that were taken over during the period 1983–1990. A sample of acquired firms and a sample of equivalent non-acquired firms are considered with the objective of distinguishing between them, based upon their differences in sixteen financial characteristics 1–3 years prior to each takeover. Linear and quadratic discriminant analysis and logit models are developed with factor analysis input. Prediction results are mixed, mainly due to the similar financial ratio profiles between acquired and non-acquired firms. Most models classify correctly a significant proportion of acquired or non-acquired firms, but not both (with one exception). The only model that provides significantly correct predictions for both acquired and non-acquired firms in either the calibration or holdout sample is a linear discriminant function with six financial ratios (two from each of the three years prior to takeover). The reasons for these modelling difficulties are discussed.

Keywords:

• Multivariate statistics
• factor analysis
• discriminant analysis
• logit
• regression
• financial ratios
• mergers
• acquisitions