Print Reference This Reddit This

How Financial Ratios Can Predict Corporate Failures Finance Essay

Financial crisis is the concrete embodiment of corporate failures. To study the causes of a listed company in financial distress and establish perfect, effective and operational early warning model not only has theoretical significance, but also has practical significance. With the scientific early warning model, a listed company can in time prevent and defuse financial crisis and improve the scientific nature of the crisis early-warning management; lenders (banks) will avoid high-risk loans; investors access to financial risk warning; government regulatory agencies can be with more effective and more scientific manner for market regulation, maintaining the stable operation of the market.

2 Literature Review

Beaver (1966) first proposed corporate financial early-warning analysis model, followed by many scholars in this field of research, research methods are also continually refined and improved. Among them is the landmark of Z-Score Discriminate model proposed by Beaver (1966) and Altman (1968, 1977), and the logic / probability of regression model proposed by Martin (1977) and Ohlson (1980). Zavgren (1985) and other scholars made further deepening. Aziz, Emanuel, and Lawson (1988, 1989) based on the cash flow model proposed company's value comes from the operators, creditors, shareholders and the sum of the Government's cash flow present value. According to the analysis of the paired data between bankrupt company and non-bankrupt company, they found that in the five years before bankruptcy, the two types of the company's means of cash flow from operations and cash payments have a significant difference in average income. According to statistics, the most frequently used method for analysis of the financial distress is multiple linear discriminate analysis and logistic regression method. In recent years, many countries have begun to experiment with new methods of financial distress prediction and have made some preliminary results, such as the use of various neural network models, in an attempt to overcome the shortcomings of previous methods. Meanwhile, some scholars have tried to explore the basis of economic theory of the enterprise's financial crisis, from the non-equilibrium theory, option pricing models and contract theory to analyze and forecast the financial crisis, and have achieved some results.

Throughout the status of research and application at home and abroad, classification methods to a multiple linear discriminate analysis (MDA) as the mainstream, this is the most effective method in the academic circles and the industry. However, this method has its own drawbacks, such as the more stringent assumptions that the MDA requests variables showed normal distribution and equal covariance matrix and linear independence, which are inconsistent with many practice, has been the controversial issues in this field of study on the quantitative analysis. In view of this, this article would try to ST, as defined in the British company is a sign of financial distress, using the financial data of listed companies to the mainstream method of MDA in the early warning of financial distress to make a comparative study, on the one hand, further test the empirical results of these two methods in the United Kingdom in the application; the other hand, verify the existing accounting system and accounting standards, the financial statements whether there is predictive value of information, financial crisis whether there is a trace could be found, so as to provide an effective early warning tool of financial analysis for investors, financial institutions and market regulators.

3 MDA Early Warning Model of Financial Distress

3.1 Introduction of MDA

The basic idea of multivariate linear discriminate analysis (MDA) in the s prediction research of financial distress is: According to the known observation of two different totalities, that is, a company with financial distress and a company without financial distress, and a number of random variable (financial indicators) reflect the differences in observed characteristics of the objects, based on historical data, using statistical methods filter out the index system with certain significant differences, to fit an optimal linear model for the classification of new things. The general form of discriminate function is:

  (1)

where: is discriminate score, to reflect the characteristic variables of the study object, such as financial ratios; is discriminate coefficient for each variable.

3.2 Ideas and methods of empirical research

In this paper, ST listed companies have been defined as an enterprise standard of corporate failures, which is a sign of financial distress for a listed company. This article first carried out the division of the ST types and stages of the company, and then analyzed the correlation model of the financial early-warning. Again it analyzed the all ST listed companies on the selected anomalies due to the financial situation, by primary 150 samples access to corporate financial distress, excluding the non-normal and data on the default company and, ultimately, obtained a sample of 83 companies in financial distress. At the same time, according to the matching principle of industry and size, it also selected 83 normal businesses as paired samples. On this basis, we have selected cover the length (short) period of solvency, operational efficiency, profitability, risk level, capacity development, capital structure, a total of six categories of 52 financial indicators, through the principal component analysis model for screening the forecast variables, and ultimately to obtain 16 principal component variables for financial distress prediction. This paper presents a new perspective as a quantitative model-based in order to carry out a qualitative analysis of financial distress prediction.

It should be noted that, despite the academic circles have done continuous research on the financial early-warning model, but so far people are still unable to accurately determine the financial variables included in the early-warning model. Therefore, at the choice of predictor variables, one can only determine the discrimination ratio as a standard to carry out a lot of "search work", in a number of different models, variable combinations, sample composition and estimation techniques determine the ability to find the best combination of forecasts model.

3.3 Selection of Analysis Variables of listed company in financial distress

3.3.1 Sample design and data processing

In this paper, the listed company by the plates of industry category in London Stock Exchange as a study object, the company has been special treatment (ST) because of the financial situation as a sign for an enterprise into financial distress, select all the ST companies from1998-2003 as the sample of companies financial distress.

Since the Commission is based on the announced results from the annual reports of listed companies two years ago to determine whether there is abnormal financial situation and decide whether want them for special treatment, so the first two years using the annual reports of listed companies to forecast whether it would be obvious ST will exaggerate the predictive power of the model. Therefore, this option ST listed companies are to predict the first three years to determine whether it will ultimately fall into financial distress, that is, if a listed company in 2003 was special treatment, we use the data in 2000 to predict the annual report.

In order to eliminate the influence of the factors of different years, industry and the scale of asset on financial distress prediction, we have based on the principle according to the ratio of 1:1 selects the healthy financial listed companies as paired samples:

(1) consistent study period, such as financial distress enterprises adopt the data in 2000, then the normal financial company also use the data in 2000. (2) Paired samples are the same similar with the types of industry in financial distress. (3) Paired samples have the similar size of the total assets of the enterprises in financial distress. (4) The exclusion of ST companies with pure B shares and companies with serious accounting fraud. (5) The exclusion of companies missing or unreasonable data and the companies within two years is ST. (6) The exclusion of ST companies by other abnormal conditions.

Accordingly, this study identified 300 samples, while excluding the non-normal ST companies and missing data companies. Finally the typical samples with completed data are a total of 166. Define the combination of 0 for the company in financial distress, the combination of 1 for the company in healthy finance, and with estimated 126 samples (including 63 companies in financial distress, 63 companies in healthy finance ), predictive samples are 40 (including 20 companies in financial distress, 20 companies in healthy finance). Selected covering a long (short) period of solvency, operational efficiency, profitability, risk level, capacity development, capital structure, a total of six categories are 52 financial indicators, deleting data, the default variable to be 26 initial variables, and then through principal component analysis method to determine the model predictor variables.

3.3.2 Variable test and screening

The 16 principal components get by the above methods as new variables use stepwise discriminate analysis method for screening independent variables.

1, test with equal group mean value

To compare the means of 16 principal components of financial indicators in different combinations are equal, we conducted a test of equal means. The results are shown in Table 1.

Table 1 Groups of equal mean test results

Variable

Group 0

Group 1

Wilks'λ

F

df1

df2

Sig.

Mean

SD

Mean

SD

Z1

-1.2607

2.31791

.8269

1.26904

.765

28.220

1

92

.000

Z2

.1485

1.22360

-.2111

1.59304

.984

1.526

1

92

.220

Z3

.4273

.90003

-.2455

1.73900

.941

5.742

1

92

.019

Z4

.0632

.99773

-.3933

1.34948

.963

3.530

1

92

.063

Z5

-.2244

.92329

-.1062

1.39349

.997

.240

1

92

.025

Z6

.1005

1.34329

-.2409

.92412

.979

2.006

1

92

.060

Z7

-.1205

1.11573

.4026

1.13050

.948

5.082

1

92

.027

Z8

.1248

1.31975

.0226

.75649

.998

.205

1

92

.052

Z9

-.0056

.85203

-.0663

.88944

.999

.114

1

92

.036

Z10

-.0809

.83065

.1170

1.24683

.991

.837

1

92

.063

Z11

.0675

.63701

.0271

1.06312

.999

.051

1

92

.121

Z12

.0365

.88595

.0473

.94020

1.000

.003

1

92

.054

Z13

-.0514

.47404

.1764

1.15547

.983

1.633

1

92

.004

Z14

-.0185

.47379

.1951

1.24191

.986

1.269

1

92

.263

Z15

-.0170

.74669

-.1185

.76928

.995

.420

1

92

.018

Z16

.0542

.87501

.0002

.72362

.999

.104

1

92

.074

It can be seen from Table 1, at the probability level of significance of 10%, Z1, Z3 ... and other 13 financial indicators in the two groups of samples there were significant differences on the mean.

2, variable screening

Use stepwise discriminate analysis for variable screening. The model criteria are generally based on pre-designated F-value and probability level. Only when the calculated F value for a variable is greater than the specified value, the variable can enter the final discriminate function. This article will define the corresponding probability of significant F value as 0.1.

Table 2 Variables inspection table

Tolerance

F to Remove

Wilks' Lambda

Z8

1.000

18.080

Z5

.994

18.035

.907

Z4

.994

8.207

.814

Z3

.993

14.937

.796

Z2

.955

10.653

.736

The process of variable screening is shown in Table 2, the last predictive variables to enter the model are Z8, Z5, Z4, Z3, Z2. And the λ value of Z2 (Lambda = 0.736) is the smallest, indicating Z2 in the enterprise's financial distress prediction has a more important role, followed by Z3 and Z4.

3, multicollinearity test

In order to avoid multicollinearity, this paper adopts tolerance (TOL) and variance inflation factor (VIF) for multicollinearity tests for selected five variables. The results are shown in Table 3:

Table 3 Multicollinearity test

Z2

Z3

Z4

Z5

Z8

TOL

0.924

0.968

0.924

0.869

0.96

VIF

1.082

1.033

1.082

1.151

1.042

In general, when the TOL is less than 0.1, or VIF greater than 10, it believes that there is multicollinearity. It can be drawn from Table 6, these selected variables do not exist multicollinearity.

4 Prediction of MDA model

According to Fisher criteria, to make the maximum discriminate scores to distinguish two different totalities, the finally obtained linear discriminate function must determine the efficiency of the largest group of linear discriminate variable Z.

According to Fisher linear discriminate model, we can calculate the means of the two types and, "discriminate point" is determined according to the principle of symmetry category: Z’= (+)/ 2

We will compare every company's Z score and discriminate point Z’, if Z> Z' then it can be regards as the healthy financial company, and vice versa ruled the company in financial distress. By the application SPSS11.5 software and run the MDA analysis model, the analysis results obtained are in Table 4 below:

By the discriminate coefficient in Table 4, we can get a linear discriminate function:

Z = 0.121 + 0.474×Z2 - 0.258×Z3 - 0.348×Z4 - 0.247×Z5 + 0.404×Z8

Table 4 Standardized discriminate function coefficient table

Function

1

Z2

.474

Z3

-.258

Z4

-.348

Z5

-.247

Z8

.404

(Constant)

.121

From the coefficient of the group center of gravity in Table 8, it can be seen that the means of the two combined Z scores are -0.797 and 0.866, respectively. Therefore, according to the principle of symmetry classification, discriminate point Z’=(+)/ 2 = 0.0345, When taking every company's financial indicators into the discriminate function, the obtained Z-score is greater than 0.0345, then it can be regards as the healthy financial company, the contrary is the company in financial distress. Discriminate results are as follows (see Table 5):

Table 5 The results of Multiple discriminate analysis

Original value

Estimation sample

Prediction sample

Predicted value

Total

Integrated accuracy rate %

Predicted value

Total

Integrated accuracy rate %

0

1

Counting

0

53

10

63

84.85

17

3

20

82.5

1

9

54

63

4

16

20

Percentage

0

84

16

100.0

85

15

100.0

1

14.3

85.7

100.0

20

80

100.0

The error rate of the miscarriage of justice for the company in financial distress (0) to the company in healthy finance (1) is expressed by type of I e error rate of the miscarriage of justice for the company in financial distress (0) to the company in healthy finance (1) is expressed by type of II, it can be seen that with error rate for estimated samples with multiple discriminate analysis is 16%, the error rate by type of II is 14.3%, the integrated accuracy rate is 84.85%; and the error rate by type of I for predicted samples is 15%, the error rate by type of II is 20%, the integrated accuracy rate is 82.5%.

5 Recommendations

The quantitative early-warning model provides an effective early warning signal, but to make more accurate measurements for the operation of listed companies’ failures, it also needs for some qualitative detailed analysis of the situation.

a, changes in ownership in listed companies: The transfer of shares of listed companies, especially the changes of large shareholders and controlling shareholders, generally begin to occur in the operations of the company to start or has worsened, a major shareholder is to look for a strong partner to bail the company and its own out circumstances. Such as asset restructuring of listed companies are mostly based on transfer of shares as the beginning.

b, self-discipline and stability of the situation of the board and senior management: ① self-discipline. Concern about whether the directors and executives suspected of corruption, smuggling and other economic crimes, because these acts not only harm the company's image, but also often bring enormous economic losses to listed companies, ② stability of personnel. Board is the company's business strategy-maker, and senior executives is an execution for the established strategy, if the frequent changes in boards and senior executives, will definitely lead to the instability and the implementation of bias for the development of strategic and operational guidelines, thus affecting the company's normal operations . Therefore, it must also attach importance to the company with frequent changes of personnel.

6 Conclusion

MDA technology for early warning of financial distress of listed companies have a high predictive ability, which further demonstrated that in the existing accounting system and accounting standards, the financial statements can provide a wealth of useful information for prediction of financial distress, the financial crisis has disappeared to be found.

A listed company in financial crisis or corporate failures is a concrete manifestation of survival of the fittest under the conditions of market economy. In this context, the financial crisis highlights the growing importance of advance warning. Financial indicators contain useful information for the prediction of financial distress, using the first three years of financial indicators, the MDA model can effectively predict the probability of financial crisis for ST companies; however, with the comparison, by the use of the previous four years and the first five years financial indicators, the long-term early warning capability of the MDA model is not good enough. To this end, it needs the combination of the quantitative and qualitative indicators for monitoring early-warning model for the real-time monitoring of company changes, to establish a financial distress prediction method for listed companies with quantitative model-based, supplemented by qualitative analysis, to ensure to provide a reliable basis for decision making on the businesses, governments, creditors, prevention and risk mitigation.

Need help with your literature review?

Our qualified researchers are here to help. Click on the button below to find out more:

Literature Review Service

Related Content

In addition to the example literature review above we also have a range of free study materials to help you with your own dissertation: