Wouldn’t it be great if you could just enter some simple financial data into a mathematical formula and the answer told us whether or not a company was heading for insolvency? Well that question has been considered by financial academics and practitioners for over 50 years and the outcome is that there is such a formula…..sort of.
Perhaps the most famous formula devised was that by Edward Altman, one time Assistant Professor of Finance at New York University, who came up with his Z-score. This looks fairly straightforward in that it is, in its original form, a sum of five variables, each multiplied by a specific coefficient, the variable being taken directly from a company’s financial statements. Released in 1968 this claimed a 72% success rate for predicting insolvency 2 years before the event. This is a pretty powerful result and led to fairly widespread acceptance by financial industry practitioners, but there were some quite decent caveats that went with those results. Primarily its use was only for publicly listed manufacturing companies with assets of more than US$1 million which narrowed the potential applicability somewhat. It should be noted though that the formula was adjusted over subsequent years to cover private and non-manufacturing companies as well with a claimed even higher predictability success rate.
Altman’s research also spawned interest in coming up with more powerful predictability formulae and a raft of analyses by numerous academics was created over the subsequent years utilising the statistical technique known as multivariate discriminant analysis with equal or better claims to predictive success.
So how come we haven’t heard much of these then? With this rate of success surely determining which businesses are heading into insolvency must be pretty simple these days?
Well, whilst these types of formulae are still used as part of credit assessments for large businesses where there is a depth of financial information available, there are known limitations that place restrictions on applicability. Firstly is that full financial statements must be obtained as in-depth financial information is required for any analysis – and these are sometimes not available. Secondly, although accuracy is statistically significant it is not fool-proof and it is psychologically difficult to rely on just one indicator which is not infallible, especially when there could be other ratios giving a contrasting view.
However, where financial statements are available the results can be surprising in their accuracy. It is very interesting to note that Carillion’s Z-score as at year ended 30 June 2016, the last full year prior to its insolvency and prior to the downgrades that were to come in 2017, was in the “Distress” zone. Interesting in two ways – firstly that the Z-score was accurate and secondly that this went unnoticed by the investment community at large (although a few short sellers came out quite well). Quite what this says about the latter probably warrants further consideration.
[Carillion is the listed giant UK construction company that went bust earlier this year].
And although Enron’s Z-score was not in the “Distress” zone prior to its bankruptcy filings the deterioration in its Z-score was the equivalent of a reduction in its bond rating of a BBB to B in the two years prior to filing – something missed by the ratings agencies.
So what about here in New Zealand?
In New Zealand the corporate landscape is quite different from the US where Altman’s formula, and most of the subsequent ones as well, was devised. The number of listed entities here is small and the majority of NZ business are small to medium enterprises where there is a known loss of predictive ability of the formulae and the amount of publicly available financial information is low.
In fact there was an interesting study conducted by Karen Van Peursem and Yi Chiann Chan of the Centre of Accounting, Governance and Taxation Research, Victoria University of Wellington in 2012 which looked at the use of these types of formulae to assess insolvency risk from publicly available information specifically in New Zealand. Their study looked at failed New Zealand listed corporations in the period 2001-2010 and the existence, if any, of prior media warnings of impending financial doom which could be used to show that a deteriorating financial performance had been observed prior to failure, and the study was summarised thus:
“We conclude that while using such patterns would have had merit, it would have been difficult to assert confidently as to the future viability of individual organizations……that is, corporate failure forecasts could have been of benefit to users, as long as such assessments had been properly qualified as subjects of ‘concern’ rather than ‘ likelihoods of failure’ “.
Patterns here relate to the changes in the various financial ratio numbers over time.
The study thus showed that although there was no conclusive evidence of financial failure predictability, there was some merit in observing changes to ratios over time as an indicator of potential financial problems.
This is disappointing. So despite the early promise it does not look as though corporate failure can be easily predicted in New Zealand, certainly on an individual basis. Does this, therefore, mean that credit assessments on the whole are somewhat less than useful?
Well certainly not as long as you understand the limits of what such an assessment can give you. On an individual basis it may be difficult to predict financial failure but there is great benefit in assessing the overall risk profile of your debtors as a whole. Managing your debtors ledger according to risk profile can give great benefits. If you can identify those of your customers that are of highest risk you can take much more appropriate and timely action. How helpful would it be to know that a customer who is currently 14 days late in paying your latest invoice is also your customer with the highest credit risk? Or that you have no security over your top 5 riskiest customers? Or that 40% of your total credit exposure is with one of your highest risk customers?
As the Victoria University study and the Enron example above also show it’s not necessarily the absolute risk score that is important but that changes in the risk score over time can also be indicative of failure.
This type of comparative analysis can be quite easily conducted from available information sources in New Zealand but it is rather a specialist skill. Sources used include information available at Companies Office, credit bureaux and even from your customer directly. This third party information can be added to business-specific information such as terms of trade, available security and credit exposure to give a surprisingly accurate complete risk profile. The caveat to this being that the person putting this all together has the requisite skill and knowledge.
So, in conclusion, the evidence does show that monitoring the financial health of companies over time can give rise to some level of failure predictability. It therefore seems sensible to conduct a credit risk analysis on your debtors ledger from time to time. Just make sure you engage an expert who can help put this together.