Anda di halaman 1dari 9


Interest in insolvency prediction has long been confined to academics, with most of the
published material restricted to business and accounting journals specializing in esoteric
and complicated subjects. A possible reason why insolvency prediction models have not
gained greater use in the business community is because it has been difficult to calculate
the results. With the wide spread use of personal computers, the utilization of an
insolvency prediction model is now practical and available to all. Now may be the time
when prediction models come into their own!

Four software programmes are reviewed here using five different prediction models. All
of the models reviewed here, but one, were developed using the statistical technique,
step-wise multiple discriminate analysis. This statistical technique gives weights to
financial ratios used to best differentiate or discriminate between failed and successful
companies. For example, 22 financial ratios were tested in developing the Altman Model
(1968). 66 companies were used - 33 failed and 33 successful. The first result was a
formula with 22 functions. The function that contributed the least to discriminating
between the failed and successful companies was dropped and the statistical software was
run again. This was repeated over and over each time dropping the ratio which least
contributed to discriminating between the failed and successful companies. In the case of
the Altman model, five functions remained.

The software we have reviewed here are easy to operate and give quick read outs. We
have not evaluated the models compared with each other because it is impossible to say,
in this kind of review, that one model is better or more accurate than another. One of the
great problems in developing and testing prediction models is that it is very difficult to
gather data on matched sets of failed and successful companies.

Some Words of Caution! All developers of prediction models warn that the technique
should be considered as just another tool of the analyst and that it is not intended to
replace experienced and informed personal evaluation. Perhaps the best use of any of
these models is as a "filter" to identify companies requiring further review or to establish
a trend for a company over a number of years. If, for example, the trend for a company
over a number of years is downward then that company has problems, which if caught in
time, could be corrected to allow the company to survive.

If bankers can identify companies in danger of failure sufficiently far in advance, then
corrective action can be taken. The banker can:

1. Decline to accept the company as a customer.

2. Encourage the company to identify its problems and take steps to rectify those
3. Encourage the principals of the company to inject more capital into the business.
4. Encourage the company to seek other financing.
BEAVER MODEL (1967 & 1968)

William Beaver (1967), one of the earliest researchers, was the first to use statistical
technique of univariate analysis to predict corporate failure on the basis of mean
values of thirty different financial ratios which were later reduced to fourteen
ratios. Beaver used a dichotomous classification test on two samples of firms (79
bankrupt firms and 79 non failed firms of approximately the same size and
industry group) based on the calculated ratios to discriminate between failed and
non-failed firms for as long as five years prior to failure. Beaver discovered that
there was significant difference between the financial ratios of failed firms and
non failed firms. Not only were they lower but deteriorated as failure approaches.
He then concluded that of the variables analyzed cash flow/total debt was the
single most important factor to consider in predicting failure. Beaver’s model was
able to accurately classify 78% and 87% of the sample of firms five years and one
year before failure respectively.

ALTMAN MODEL (U.S. - 1968)

Edward I. Altman (1968) is the dean of insolvency predictors. He was the first person to
successfully use step-wise multiple discriminate analysis to develop a prediction model
with a high degree of accuracy. Using the sample of 66 companies, 33 failed and 33
successful, Altman's model achieved an accuracy rate of 95.0%. Altman's model takes the
following form -:

The discriminant function of the form Z = v1 x1 + v2 x2 + v3 x3 + v4 x4 + v5 x5

Where v1,……………… v5 are the discriminant coefficients

x1, ………………..x5 are the independent variables

Z = 1.2x1 + 1.4x2 + 3.3x3 + 0.6x4 + 0.999x5

Z < 2.675; then the firm is classified as "failed"


X1=Working Capital/Total Assets

X2=Retained Earnings/Total Assets
X3=Earnings Before Interest and Taxes/Total Assets
X4=Market Value of Equity/Book Value of Total Liabilities
X5=Sales/Total Assets

This model was developed in 1978 at S.F.U. by Gordon L.V. Springate; following
procedures developed by Altman in the U.S. Springate used step-wise multiple
discriminate analysis to select four out of 19 popular financial ratios that best
distinguished between sound business and those that actually failed. The Springate model
takes the following form -:

Z = 1.03x1 + 3.07x2 + 0.66x3 + 0.4x4

Z < 0.862; then the firm is classified as "failed"

X1 = Working Capital/Total Assets
X2 = Net Profit before Interest and Taxes/Total Assets
X3 = Net Profit before Taxes/Current Liabilities
X4 = Sales/Total Assets

This model achieved an accuracy rate of 92.5% using the 40 companies tested by
Springate. Botheras (1979) tested the Springate Model on 50 companies with an average
asset size of $2.5 million and found an 88.0% accuracy rate. Sands (1980) tested the
Springate Model on 24 companies with an average asset size of $63.4 million and found
an accuracy rate of 83.3%.

FULMER MODEL (U.S. - 1984)

Fulmer (1984) used step-wise multiple discriminate analysis to evaluate 40 financial

ratios applied to a sample of 60 companies -30 failed and 30 successful. The average
asset size of these firms was $455,000. The model takes the following form -:

H= 5.528 (V1) + 0.212 (V2) + 0.073 (V3) + 1.270 (V4) - 0.120 (V5) + 2.335 (V6) + 0.575
(V7) + 1.083 (V8) + 0.894 (V9) - 6.075

H < 0; then the firm is classified as "failed"

V1 = Retained Earning/Total Assets
V2 = Sales/Total Assets
V3 = EBT/Equity
V4 = Cash Flow/Total Debt
V5 = Debt/Total Assets
V6 = Current Liabilities/Total Assets
V7 = Log Tangible Total Assets
V8 = Working Capital/Total Debt
V9 = Log EBIT/Interest

Fulmer reported a 98% accuracy rate in classifying the test companies one year prior to
failure and an average 81% accuracy rate more than one year prior to bankruptcy.


This is the only business failure prediction method outlined here that was not developed
using multiple discriminate analysis. This system was developed by William Blasztk in
1984. The essence of the system is that the financial ratios for the company to be
evaluated are calculated, weighted and then compared with ratios for average companies
in that same industry as given by Dunn & Bradstreet. One of this method's strengths is
that it does compare the company being evaluated with companies in the same industry.


This model is recommended by the Ordre des compatables agrees des Quebec (Quebec
CA's) and according to its developer is used by over 1,000 CA's in Quebec.

This model was developed under the direction of Jean Legault of the University of
Quebec at Montreal, using step-wise multiple discriminate analysis. Thirty financial
ratios were analyzed in a sample of 173 Quebec manufacturing businesses having annual
sales ranging between $1-20 million.

The model takes the following form -:

CA-Score = 4.5913 (*shareholders' investments (1)/total assets (1))

+ 4.5080 (earnings before taxes and extraordinary items
+ financial expenses (1)/total assets (1)) +
0.3936 (sales (2)/total assets (2)) - 2.7616

CA-Score < - 0.3; then the firm is classified as "failed"

1) Figures from previous period

2) Figures from two previous periods

* Shareholders' investments are calculated by adding to shareholders' equity the net
debt owing to directors.

This model, as reported in Bilanas (1987), has an average reliability rate of 83% and is
restricted to evaluating manufacturing companies.

1. Altman, Edward I., "Financial Ratios, Discriminant Analysis and the Prediction of
Corporate Bankruptcy". Journal of Finance, (September 1968): pp. 589-609.
2. Botheras, Donald A., "Use of a Business Failure Prediction Model for Evaluating
Potential and Existing Credit Risk". Unpublished M.B.A. Research Project,
Simon Fraser University, March, 1979.
3. "C.A. - Score, A Warning System for Small Business Failures", Bilanas (June
1987): pp. 29-31.
4. Fulmer, John G. Jr., Moon, James E., Gavin, Thomas A., Erwin, Michael J., "A
Bankruptcy Classification Model For Small Firms". Journal of Commercial Bank
Lending (July 1984): pp. 25-37.
5. Sands, Earl Gordon, "Business Failure Prediction and the Efficient Market
Hypothesis". Unpublished M.B.A. Research Project, Simon Fraser University,
November 1980.
6. Sands, Earl G., Gordon L.V. Springate, and Turgut Var, "Predicting Business
Failures". CGA Magazine (May 1983): pp. 24-27.
7. Springate, Gordon L.V., "Predicting the Possibility of Failure in a Canadian
Firm". Unpublished M.B.A. Research Project, Simon Fraser University, January

NOTE: Some non-Netscape browsers read these formulas incorrectly. If you have
doubts please refer to the source code.

The failure of a business firm is an event which can produce substantial losses to
creditors and stockholders. Therefore, a model which predicts potential business failures
as early as possible would serve to reduce such losers by providing ample warning to all
interested parties. This was the sufficient motivation for Beaver (1967, 1968) and Altman
(1968) to develop models for predicting business failure based on the financial reports of
firms. Thus, attempts to develop bankruptcy prediction models began seriously sometime
in the late 1960's and continue through today. At least three distinct types of models have
been used to predict bankruptcy:
1) Statistical models (primarily, multiple discriminate analyses [MDA]), and
conditional logit regression analyses,

2) gambler's ruin-mathematical/statistical models, and

3) Artificial neural network models.

Most of the publicly available information regarding prediction models is based on

research published by university professors. Commercial banks, public accounting firms
and other institutional entities (bond ratings agencies, for example) appear to be the
primary beneficiaries of this research, since they can use the information to minimize
their exposure to potential client failures.

While continuing research has been ongoing for almost thirty years, it is interesting to
note that no unified well-specified theory of how and why corporations fail has yet
been developed. The available statistical models derive merely from the statistical
optimization of a set of ratios. As stated by Wilcox(1), the "lack of conceptual framework
results in the limited amount of available data on bankrupt firms being statistically 'used
up' by the search before a useful generalization emerges."

How useful are these models? Almost universally, the decision criterion used to evaluate
the usefulness of the models has been how well they classify a company as bankrupt or
non-bankrupt compared to the company's actual status known after-the-fact (that is ex
post). Most of the studies consider a type I error as the classification of a failed company
as healthy, and consider a type II error as the classification of a healthy company as
failed. In general type I errors are considered more costly to most users than type II
errors. The usefulness of fail/nonfail prediction models is suggested by Ohlson(2):

...real world problems concern themselves with choices which have a richer set of
possible outcomes. No decision problem I can think of has a payoff space which is
partitioned naturally into the binary status bankruptcy versus non-bankruptcy...I have
also refrained from making inferences regarding the relative usefulness of alternative
models, ratios and predictive systems... Most of the analysis should simply be viewed as
descriptive statistics - which may, to some extent, include estimated prediction error-
rates - and no "theories" of bankruptcy or usefulness of financial ratios are tested.

Subject to the qualifications expressed above, bankruptcy prediction models continue to

be used to predict failure.

The early history of researchers' attempts to classify and predict business failure (and
bankruptcy) is well documented in Edward Altman's seminal 1983 book, Corporate
Financial Distress(3). There appears to be no consensus on what constitutes business
failure. However, most businesses are considered to have failed once they have entered
formal bankruptcy proceedings.

Failure prediction models use different approaches to predict failure. Beaver(4) (1967),
one of the earliest researchers, used a univariate analysis of a number of financial ratios
to discriminate between failed and non-failed firms for as long as five years prior to
failure. Of the variables Beaver analyzed, he concluded that cash flow/total debt was the
single most important factor to consider in predicting failure. Beaver's work set the stage
for the multivariate analyses developed by Altman and others.

Possibly the most famous failure prediction model is Altman's Z-Score Model. Based on
multiple discriminate analysis (MDA), the model predicts a company's financial health
based on a linear discriminant function of the form:


X1=working capital/total assets
X2=retained earnings/total assets
X3=earnings before interest and taxes/total assets
X4=market value of equity/book value of total liabilities
X5=sales/total assets

The Z-Score model (developed in 1968) was based on a sample composed of 66

manufacturing companies with 33 firms in each of two matched-pair groups. The
bankruptcy group consisted of companies that filed a bankruptcy petition under Chapter
X of the United States bankruptcy act from 1946 through 1965. Based on the sample, all
firms having a Z-Score greater than 2.99 clearly fell into the non-bankruptcy sector,
while those firms having a Z-Score below 1.81 were bankrupt.
However it was discovered that the predictive ability of the model declined rapidly as the
number of years prior to failure increased.

Altman subsequently developed a revised Z-Score model (with revised coefficients and
Z-Score cut-offs) which dropped variables X4 and X5 (above) and replaced them with a
new variable X4 = net worth (book value)/total liabilities. The X5 variable was allegedly
dropped to minimize potential industry effects related to asset turnover.

Around 1977, Altman developed jointly with a private financial firm (ZETA Services,
Inc.) a revised seven-variable ZETA model based on a combined sample of 113
manufacturers and retailers. The ZETA model is allegedly "far more accurate in
bankruptcy classification in years 2 through 5 with the initial year's accuracy about
equal." However, the coefficients of the model are not specified (without retaining ZETA
Services). The ZETA model is based on the following variables:

• return on assets
• stability of earnings
• debt service
• cumulative profitability
• liquidity/current ratio
• capitalization (five year average of total market value)
• size (total tangible assets)

Many additional bankruptcy prediction models have been developed since the work of
Beaver and Altman. Lev (1974)(5), Deakin (1972)(6), Ohlson (1980)(7), Taffler (1982)
(8), Platt & Platt (1980)(9), Gilbert, Menon, and Schwartz (1990)(10), and Koh and
Killough (1990)(11) amongst others have continued to refine the development of
multivariate statistical models. Almost all of these traditional models have been either
matched-pair multi-discriminate models (such as Altman's) or logit models (such as
Ohlson's). A 1997 study by Begley, Ming and Watts(12) concludes:

Given that Ohlson's original model is frequently used in academic research as an

indicator of financial distress, its strong performance in this study supports its use as a
preferred model.

Wilcox(13), Santomero(14), Vinso(15) and others have adapted a gambler's ruin

approach to bankruptcy prediction. Under this approach, bankruptcy is probable when a
company's net liquidation value (NLV) becomes negative. Net liquidation value is
defined as total asset liquidation value less total liabilities. From one period to the next, a
company's NLV is increased by cash inflows and decreased by cash outflows during the
period. Wilcox combined the cash inflows and outflows and defined them as "adjusted
cash flow." All other things being equal, the probability of a company's failure increases,
the smaller the company's beginning NLV, the smaller the company's adjusted (net) cash
flow, and the larger the variation of the company's adjusted cash flow over time. Wilcox
uses the gambler's ruin formula (Feller, 1968) to show that a company's risk of failure is
dependent on 1) the above factors plus 2) the size of the company's adjusted cash flow "at
risk" each period (i.e., the size of the company's bet).

Using an alleged more robust statistical technique, Vinso(16) extended Wilcox's

gambler's ruin model to develop a safety index. Based on input concerning the variability
of "expected contribution margin amounts," the index can be used to predict the point in
time when a company's ruin is most likely to occur (called first passage time).

The statistics used in gambler's ruin approaches are somewhat formidable (especially to
the average business reader). However, both Wilcox and Vinso richly describe some of
the factors which most affect business failure. For example, Wilcox states:

The (cash) inflow rate ... can be increased through higher average return on investment.
However, having a major impact here usually requires long-term changes in strategic
position. This is difficult to control over a short time period except by divestitures of
peripheral unprofitable businesses...The average outflow rate is controlled by managing
the average growth rate of corporate assets. Effective capital budgeting ... requires
resource allocation emphasizing those business units which have the highest future
The size of the bet is the least understood factor in financial risk. Yet management has
substantial control over it. Variability in liquidity flows governs the size of the bet. This
variability can be managed through dividend policy, through limiting earning variability
and investment variability, and through controlling the co-variation between profits and
investments...True earnings smoothing is attained by control of exposure to volatile
industries, diversification, and improved strategic position. (Emphasis added)

Vinso supports Wilcox's emphasis on cash flow processes and stresses the importance of
debt capacity:

Before deriving a mathematical model for determining the risk of ruin, it is necessary to
describe the process. (First), a firm has some pool of resources at time = 0 of some size
U0, which are available to prevent ruin (similar to Wilcox's beginning NAV). (Then),
earnings come to (the) firm from revenue(s)...less the costs incurred in producing (the