Use of Bankruptcy Prediction Models as an Auditing Tool

One of the best-known models for predicting corporate financial distress is the Altman’s Z-Score model (Altman, 1968, 1983, 1993). This model is constructed using linear relationship of financial ratios weighted by predefined coefficients. Initially the model was developed to assess the bankruptcy of publicly held manufacturers, but later it was re-estimated for private manufacturing, non-manufacturing and service companies. Therefore, the model is three folds: Z-model (original model); Z’-model (private firm model); Z’’-model (nonmanufacturing firm model). Eventually a firm is categorized based on its Z-score as non-bankrupt, bankrupt or gray area in each of the aforesaid three Z-score models.


From about 1985 onwards, the Z-scores gained wide acceptance by auditors, management accountants, courts, and database systems used for loan evaluation (Eidleman). Yet, the application of the model in Sri Lankan context as a prediction of future bankruptcy by the auditors is rare.



One such study was carried out by Lalith and Tanweer in the article of Altman’s Z-score models of predicting corporate distress: evidence from the emerging Sri Lankan stock. In this study the researchers purpose was to provide an out-of-sample test of the Z-score model and its variants by applying them to a sample of firms in the emerging Sri Lankan stock market.

In this regard, They examined the firms listed in the Colombo Stock Exchange during the period 1986 to 1997 to identify a sample of distressed firms and a paired control sample of non-distressed firms. They included firms, which have been delisted due to liquidation, litigation or losses as well as those with a negative net worth as distressed firms. Thereafter they selected a sample of non-distressed firms by matching the distressed sample firms’ size and industry. Finally couples of thirteen firms having similar size and within the similar industry were selected. The Z-score was calculated for both financially distressed firms and non-distressed firms.

Then examining the accuracy in classifying the distressed firms as distressed and non-distressed firms as non-distressed. The results show that these models have a remarkable degree of accuracy in predicting distress using financial ratios computed from financial statements in the year prior to distress. The overall success rate of 81%. Although this study was not constructed on hypothesis testing it gives an preliminary understanding of the application of the Altman’s Z-score models of predicting corporate distress. Nevertheless in the global arena, this concept is immensely used.


Antony and Yi (2008) used the Z-score model as a determination of risk factor by considering the agency theory when forming the judgement. The research teases out the serious agency problem in auditors’ going concern reporting and shows they have greater inappropriate reporting than the more traditional agent being the directors. The auditor is not merely considered as a monitoring mechanism but have agency features. The salient feature in Antony and Yi (2008) study is that it considered multilevel risks related to going concern opinion. In this study, they established five-level risk class which was derived from Australian Auditing Standard pronouncements to examine the appropriateness of auditors’ going concern reporting relating specifically to the likelihood of firm failure. The research focuses on all Australian listed companies within the building industry within the period 1989 to 2007. The study was based on the Altman’s Z-score bankruptcy prediction model which was practiced by many researchers. The five-level risk class was matched with the Altman Z-score bankruptcy prediction model. Companies with lower Z-scores (higher possibility of bankruptcy) are matched with a higher risk level of auditor modifications. The auditor’s risk class which were the dependant variable, were compared to Altman’s independent measure to determine the level of under or over reporting. Finally, this study established a relationship with the variables and concluded the auditor agent issue.


Lili in her study, indentified three criteria for the stress of a company. Firstly the auditor view. This criterion is most widely used in bankruptcy prediction and auditor going concern studies. Secondly the Altman’s Z-score to consider the level at which the company is in. i.e. is it in the bankruptcy, non-bankruptcy or in the group of gray area as mentioned in the previous paragraph. Thirdly , the Zmijewski Probability. Based upon this criterion, a company is considered as stressed if it has a probability of bankruptcy, calculated from Zmijewski’s (1984) model. Finally, the stock return considered as the fourth criterion. The dependent variable in the prediction models is each firm’s bankruptcy status. The independent variables are a set of bankruptcy predictors. The study revealed that there is a positive relationship between firm size and the probability of bankruptcy for stressed firms. Which is against the findings of Daniel et al in 2005 which is mentioned in the above paragraph. Among the four criteria tested, the auditor view criterion and the Altman Z-score criterion are less effective than the Zmijewski probability criterion and the stock return criterion.

When reviewing the previous literature, it was observed that many have attempted to create relationships between the financial indicators, order of auditor receiving evidence, size of the client, auditor being a Big 4, switching auditors, management accounting methods, experience and the framing of hypothesis, auditors attitudes toward the clients’ going concern, consensus of the auditors and disclosures and the auditors issuance of going concern qualification in the audit report. Apart from that some researchers found a relationship between the companies going bankruptcy and the auditor’s indication of going concern validity in the audit report. All the aforementioned dependant variables could be broadly categorize into two aspects. i.e. issuance of going concern opinion on the grounds of the indications of subject matter (in this case individual corporate) and auditors’ characteristics. Further, many have tried to validate the so-called phenomena of wealth at risk, self fulfilling prophecy and recency effect when receiving various types of evidence in the course of the audit. Among the others, Altman’s Z-score bankruptcy prediction model is considered to be a descriptive model to assess the future viability of an organization.
Despite the fact that auditors’ going concern opinion is not a guarantee for neither the future viability of the firm nor a prediction of future bankruptcy, as the code of ethics and auditing standards required the auditor should be skeptical when drawing conclusions. Further, they have a moral obligation to disclose the early warning signal of a financially distressed company to its interested parties.

References

Antony, Y. and Yi, W., 2008. Multi-risk level examination of going concern modifications. Managerial Auditing Journal, 25, pp.756-791.

Lalith P. S. and Tanweer H., Altman’s Z-Score Models of Predicting Corporate Distress: Evidence from the Emerging Sri Lankan Stock Market, Social Sciences Research Network [Online], Available at: http://papers.ssrn.com/sol3/Delivery.cfm/SSRN_ID1395229_code327186.pdf?abstractid=1395229&mirid=4, [Accessed 20th November 2010]

Lili, S., The effects of client size and stress criteria on bankruptcy prediction models: an empirical analysis [Online], Available at: http://aaahq.org/audit/midyear/04midyear/papers/bpt-stres%20size.doc, [Accessed 06th December 2010]