Dotcom Failure and Predictive Power of Financial Ratios

Abstract

This paper is a continued effort to alarm investors whether their investments are in good hands in case of more vulnerable companies, the dotcoms. I perform relatively simple multivariate discriminant analyses on the bankruptcy of dotcoms and provide evidence that the bankruptcy of dotcoms can be predicted using accounting-based financial ratios. The stepwise discriminant analysis shows extremely high cross-validation correct classification rate (96.3%). More striking results are from two arbitrarily chosen models: the stepwise five-variable model and the two-variable model. The former uses five variables from the stepwise discriminant procedure that show a significant mean difference between non-bankrupt and bankrupt dotcoms, while the latter uses two variables from the Altman’s model that displays a significant mean difference between the two groups of dotcoms. Both models report the overall correct classification rates of 82% and 84% respectively that are reasonably high when compared to those in prior studies ranging from 60% to 95%. Most striking is that the two-variable model is based on the Altman’s model that was originally developed only for manufacturing firms. Financial distress and failure have been an aching issue in Korean banking industry since 2011 and the present study suggests that the parsimonious models of bankruptcy forecasting may be a key to prevent investors and savers from entrusting their lifetime savings with failing technology firms and financial institutions.