Yes, believe it or not, it is possible to reliably estimate the probability of a company going bankrupt! The z score was published in 1968 by renowned professor of the NYU Stern School of Business Edward I. Altman.
The z score has since been used to measure the financial distress of a company i.e. its financial health. The z score is linear combination of various ratios and weighted coefficients of firms that declared bankruptcy and matching them to a set that survived bankruptcy in the same industry.
The predictive accuracy of Altman’s z score is remarkable, attaining up to 90% accuracy for a one year time horizon and 72% for a two year horizon!
Nevertheless, this accuracy applies mainly to publically held manufacturing firms with assets larger than $1 million. Notably, since the Altman concept is essentially a balance-sheet driven model, it’s predictive powers are not generally recommended for use with financial institutions. This is attributable to the occasionally opaque accounting practices of financial firms and in particular the usage of off-balance sheet items.
These off-balance sheet or OBS items are actually assets or liabilities of an institution. So, what’s so opaque about them? Well, since these OBS’s are not recorded on the firms balance sheet, it makes them difficult to identify, quantify or track since they are usually only stated in the adjacent notes. This gives OBS the appearance of being potentially hidden liabilities, although the companies do not necessarily intend to mislead or deceive investors.
Take an example of a firm purchasing a machine using debt. If the company has already reached its required debt-to-asset ratio it cannot purchase the new machine without breaching its banks credit-line restriction. By creating a subsidiary or special purpose entity (SPE) that officially owns the machine, the firm then leases the machine from the subsidiary. With the SPE, the company can control and utilize the machine without violating the credit covenants. Pretty neat!
To be fair, most OBS are perfectly legal and good practice! However, there are certain cases that one might consider borderline e.g. Lehman Brothers during the financial crisis of in 2008. The transaction or ‘accounting gimmick’ was referred to as ‘Repo 105’. In a repo, the seller promises to re-purchase an asset at a later stage. This way the seller can raise cash quickly e.g. to pay down debt. However, recording the transaction as a genuine ‘sale’ is the part does not generally appear intuitive, since the risky asset is ‘temporally’ taken off the ‘published’ books. (A special case is deemed legal when the asset is valued at 105% or more cash is received). No wonder any model might find it challenging to dissect and contextualize these OBS’s!
Nevertheless, the Altman z score has been used for 50 years and is extremely valuable in the industry and endorsed worldwide by auditors. As with every model, all have their strengths and limitations. The key thing is to know where models perform best and where extra care is needed. After half a century of development the Altman z score has been enhanced and immensely refined and represent one of the best tools of predictive analysis in the industry.
To learn more about the Altman z score, meet the honourable Professor Altman himself in Geneva 11th April 2018 or in Zurich 12th April 2018. The event is hosted by CFA Switzerland.
To apply go to:
Event in Geneva: http://swiss.cfa/Lists/Events%20Calendar/DispForm.aspx?ID=2496
Event in Zurich: http://swiss.cfa/Lists/Events%20Calendar/DispForm.aspx?ID=2497