Heterogeneity in speed of adjustment using finite mixture models

Abstract

Many empirical analyses of firms' speed of leverage adjustment (SOA) impose a strong constraint: an average SOA is estimated for all firms in a sample. We demonstrate the usefulness of finite mixture models (FMM) in corporate finance by analysing estimates of firms' SOA. Applying FMM to a sample of US firms during 1972–2017, we find five distinct types of firm behaviours, each with its own SOA. Moreover, the same explanatory variables can have quite differing effects across the groups. We also offer the applied researcher a battery of validation techniques that can be used in a FMM context. FMM should be a standard part of finance researchers’ tool-kits.