Partial Least Squares Based Financial Distressed Classifying Model of Small Construction Firms
​
Amirah-Hazwani Abdul Rahim, Ida-Normaya M. Nasir, Abd-Razak Ahmad and Nurazlina Abdul Rashid
​
​
Abstract. The study on the classification of firms’ financial distress was made popular by Altman (1968). Up until today, banks use Altman’s ratio to rate credit credibility of potential borrowers. Since then many works replicate, improvise or use different statistical and non-statistical methods to improve the classification rate of financial distress. Most of these works dealt with information gathered from large companies as information on small companies are limited and not easily available. The aim of this research is to fill in the gap and extends the work done in Abd Razak and Wan Asma’ (2012) by looking at the predictive ability of information gathered from Malaysian small firms. It tries to determine the financial covariates that can classify the distressed firms from the healthy ones and to investigate whether a partial least squares discriminant analysis (PLS-DA) is a more efficient model than a logit model in classifying the distressed from the healthy ones. The result of Logistic Regression and PLS-DA are found to be close. PLS-DA has the advantage that is not affected by multicollinearity, because its components are orthogonal.