Conventional accounting ratios and their relative power in the prediction of corporate failure : a local investigation

Abstract

The prediction of corporate failure is an effective tool which helps a broad range of persons and institutions such as banks, business entities, employees, creditors, shareholders and investors amongst others. Many forecasting models have been built and tested in terms of their accuracy to discriminate between failing and non-failing companies. Accounting ratios as a basis to discriminate between the two categories feature in many models used in research on this topic. Conventional accounting ratios have the advantage of being easy to calculate and widely available from public financial statements without the need of inside-firm knowledge. This dissertation, inspired by past empirical research, adopts a univariate approach in assessing whether the available ratios individually discriminate between failing and non-failing films and the extent of their predictive power. The objective is achieved by taking a sample of corporate failures and matching these with non-failed firms, the derivation of the accounting ratios of the two classes are compared for any significant difference which signals prediction, using the elaborate technique of regression analysis. The dummy variable model applied shows that in Malta only two accounting ratios, out of those which could be computed from the available data, disc1iminate significantly between failed and non-failed firms. The linear probability model shows that local corporate failure is mostly sensitive to the current ratio indicating that it is the best ratio to predict of corporate collapse.