Welcome to Seminar in Finance

“Financial Ratios as Predictors of Failure”

Published in (Journal of Accounting Research, PP.71-111. 1966 )

A Research ArticleBy

By William H. Beaver

Presentation By: Rajendra Lamsal, M. Phil. Roll. No. 05

• Background and Purpose• Methodology• Limitations• Data Analysis

– Profile Analysis– Dichotomous classification Test– Likelihood Ratios

• Summary and conclusion• Suggestion for Future Research

Presentation Outline

Background and Purpose

• Ratio Analysis involves the use of several ratios by a variety of users.

• However, little efforts has been directed towards the formal empirical verification of their usefulness.

• Purpose of this study is to provide an empirical verification of the predictive ability of accounting data.

Methodology(I) Paired- Sample design: For each failed firm in the sample, a Non-failed firm of the

same industry and assets size was selected.(II) Selection of Failed and Non-Failed Firm Samples were selected from Moody’s Industrial Manual. Total 79 firms were selected as failure firms under 38 industries. The failed firm were classified according to industry and assets

size and each firm was assigned a three digit number as SIC system.

The Non-failed firm were selected from News Front Magazine on which 12000 firms were grouped according to industry(using SIC) and size.

Methodology Cont. ….

(III)Collection of data: The financial statement data of the failed

firms were obtained from Moody's five years prior to failure.

The financial statements of the non failed firms were also obtained for the same fiscal years as those of their failed mates. The time period of study was from 1954 to 1964 .

(IV) Computation of Ratios (Computed total 30 ratios under the major 6 parts )

1. Cash flow ratios(4). CF/TD2. Net income ratios(4). NI/ TA3. Debt to total asset ratios(4). TD/TA4. Liquid asset to total asset ratios(4). WC/TA5. Liquid asset to current debt ratios(3). CA/CL6. Turnover ratios(11). No-Credit Interval

Methodology Cont. ….

Limitations

• Inferences drawn from this study apply only to firms that are members of the population.

• It cannot draw inferences regarding a single observation-only about pairs of observations.

• The findings will not provide any insight in to the potential predictive power of industry and assets size because the pairing for assets size is not perfect.

Data Analysis

Ratio Prediction

Cash flow to total debt Nonfailed > failed

Net income to total assets Nonfailed > failed

Total debt to total assets

Failed > Nonfailed

Working capital to total assets Nonfailed > failed

Current ratio Nonfailed > failed

No credit interval Nonfailed > failed

I.Profile AnalysisTable:2 Prediction of Mean value

I.Profile Analysis (Comparison of Mean Values)Cash flow toTotal Debt Net Income to Total Assets

1 2 3 4 5-0.25

-0.2

-0.15

-0.0999999999999998

-0.0499999999999998

2.22044604925031E-16

0.0500000000000003

0.1

0.15

Survived

Failed

Years Before Failure1 2 3 4 5-0.3

-0.2

-0.1

0

0.1

0.2

0.3

0.4

0.5

0.6Sur-vived

Failed

Years Before Failure

Data Analysis Cont…

Data Analysis Cont…

Total debt to total assets Working capital to total assets

1 2 3 4 50

0.1

0.2

0.3

0.4

0.5

0.6

0.7

0.8

0.9

Survived

Failed

Years Before Failure 1 2 3 4 50

0.05

0.1

0.15

0.2

0.25

0.3

0.35

0.4

0.45

0.5

Survived

Failed

Years Before Failure

Data Analysis Cont…

Current Ratio No credit interval

1 2 3 4 50

0.5

1

1.5

2

2.5

3

3.5

4

Survived

Failed

Years Before Failure1 2 3 4 5

-0.2

-0.15

-0.1

-0.05

0

0.05

0.1

0.15

0.2

Survived

Failed

Years Before Failure

Data Analysis Cont…

Total Assets

• The difference in means is evident for at least five years before failure, with the difference increasing as the year of failure approaches.

• The evidence suggests that there is a difference in the ratios of failed and non failed firm.

1 2 3 4 55

5.5

6

6.5

7

7.5

8

8.5

Survived

Years Before Failure

Failed

Data Analysis Cont…

II.Dichotomous Classification Test • It makes a dichotomous prediction (with the help of

an optimal cutoff point-to minimize the percent of incorrect prediction).

• If the firm's ratio is below the cutoff point, the firm is classified as failed.

• The process of finding the optimal cutoff point based upon trial and error.

• The financial ratio analysis was used for at least five years before failure

Data Analysis Cont…

Ratio Years Before Failure

1 2 3 4 5

Cash flow to total debt .13(.10)

.21(.18)

.23(.21)

.24.(24)

.22(.22)

Net income to total assets .13(.12)

.20(.15)

.23(.22)

.29(.28)

.28(.25)

Total debt to total assets .19(.19)

.25(.24)

.34(.28)

.27(.24)

.28(.27)

Working capital to total assets .24(.20)

.34(.30)

.33(.33)

.45(.35)

.41.(35)

Current ratio .20(.24)

.32(.27)

.36(.31)

.38(.32)

.45(.31)

No- credit interval .23(.23)

.38(.31)

.43(.30)

.38(.35)

.37.(30)

Total assets .38(.38)

.42(.42)

.45(.42)

.49(.41)

.47(.38)

Table:3 Result of dichotomous analysis

Data Analysis Cont…

Results from table:3 (First test results are in parenthesis &

second test results are out of parenthesis): Ability to predict failure is strongest in the cash-

flow to total-debt ratio.Second test provides narrower error percentages

between 1st and 5th years before failure than ‘ random-prediction model’.

All ratios do not predict equally well.The net-income to total assets ratio predicts

second best.

Data Analysis Cont…

Results from table:4 Table 4 shows the percentage error form the

paired analysis to the unpaired analysis. The reduction in predictive power is not

overwhelming (devastating). It means that there is small industry effect. The predictive ability of the ratios is

understated to the extent of the industry effect, which if taken into account, improves the predictive power of ratios.

Data Analysis Cont…

Ratio

Correlation coefficient

Proportion of variance explained

Sub-sample Sub-sample

A B A B

Cash flow/Total debt .12 .20 .0144 .0400

Net income/Total assets .22 .18 .0484 .0324

T. Debt /T. Assets -.09 -.06 .0081 .0036

Working cap./T. Assets -.15 -.01 .0225 .0001

Current ratio -.04 .02 .0016 .0004

No-Credit interval -.02 .15 .0004 .0225

Table:5 Correlation with total assets

Data Analysis Cont…

No evidence of strong correlation has been observed, and the results conform to the hypothesis that the ratios are uncorrelated with assets size.

Data Analysis Cont…

Type I and Type II errors are probabilities of error conditional upon the actual status of the firm.

17 out of 79 failed firms in the first year before failure were misclassified, so the Type I error is 22%(17/79), and 4 out of 79 nonfailed firms misclassified, so the Type II error is 5%(4/79).

The total error is(17+4) out of 158 is app13% (21/158). The evidence shows that cash-flow to total-debt ratio has

a striking ability to classify both failed and nonfailed firms more accurately than the random prediction model, and

Type I error increases as the time period before increases but Type II error remains remarkably stable over the five year period. So ratios can predict nonfailed firms more accurately than the failed firms.

Contingency Tables (Table: 6)

Data Analysis Cont…

Analysis of Likelihood Ratios

Histograms are used to assess the likelihood ratios form the financial ratios.

The fig.2 shows that the distribution of nonfailed firms is surprisingly stable in each years before failure.

The distribution of ratios of failed firms shift farther to the left as failure approaches, and the gap between the failed and nonfailed firms become greater.

The smaller the overlap of the two distributions, the lower the percentage error in the classification test,

Analysis of Liklihood Ratios

Data Analysis Cont…

To arrive at the conditional prob., possible events are first viewed as being dichotomous- failed or nonfailed based on certain factors.

After the financial ratio is observed, then the likelihood of failure and nonfailure is formed.

It is the prob. that the observed ratio that would appear if the firm were failed- P(R/F), and the observed ratio that would appear if the firm were nonfailed.

Joint prob. is the product of the prior probabilities times the likelihood estimates.

Analysis of Likelihood Ratios Cont.

Conclusions

• Author found that the most successful predictors is cash flow to total debt ratio

• Second important predictor is net income to total asset ratio,

The predictive power of the liquid asset ratios is much weaker.

• The ratios do not predict failed and non-failed firms with the same degree of success.

Suggestions for Future Research

This study is univariate analysis – a multivariate ratios analysis should be pursued to predict failure of the firms.

This analysis can be enhanced further by comparison with other predictors of failures such as market rate of return.

Comments

• Pair sampling design is strong.• Limitations are clearly stated.• Some explanations are very hard to understand.

Thanking You!