Available online at www.sciencedirect.

com

Procedia Economics and Finance 3 (2012) 132 137

Emerging Markets Queries in Finance and Business

A statistical model of financial risk bankruptcy applied for

Romanian manufacturing industry
Monica Violeta Achima,*, Codruta Marea, Sorin Nicolae Borleab
a

Faculty of Economics and Business Administration, Babes-Bolyai University, Cluj-Napoca, Romania

b
Faculty of Economics, Vasile Goldis Western University, Arad, Romania

Abstract
The risk of bankruptcy was and is the subject of many research studies aiming to identify the time of bankruptcy, competing
factors to achieve this state, the main financial criteria which best expresses this orientation, the bankruptcy, etc. Although
there are models that have proved their viability over time, we note that the developed models are successfully applied in
space and time in which there were created. In this article we tried to develop a statistical model for prediction of
bankruptcy risk, available for the companies which operate in the Romanian space. Thus, we applied a multidimensional
analysis technique, namely Principal Component Analysis on two groups of companies in the manufacturing sector, during
the period 2000-2011, also including the global financial crisis impact. The financial crisis is the main factor that influenced
g
impact on the company's financial condition. The two groups of companies are selected from among the listed companies,
which are operating in the manufacturing area of activity, respectively one group consists in the listed companies and
another group consists in firms that will be unlisted in the next years. The statistical model can be successfully used to
predict the risk of bankruptcy for companies which activate in manufacturing area of activity in Romania.

2012 Published
Elsevier
Ltd. access
Selection
peer-review
under responsibility of the
underand
CC BY-NC-ND
license.
2012 TheAuthors.
Published byby
Elsevier
Ltd. Open
Selection and
peer review
under responsibility
of Emerging
Markets
in Finance and Business local organization.
Markets
Queries
in Finance and
Business
localQueries
organization
Keywords: bancruptcy risk; Principal Component Analysis, model; prediction;

Corresponding author. Tel: +40.264 41 86 52/3/4/5 , Fax: +40.264 41 25 70

E-mail address: monica.achim@econ.ubbcluj.ro, monicaachim@yahoo.com

2212-6716 2012 The Authors. Published by Elsevier Ltd. Open access under CC BY-NC-ND license.
Selection and peer review under responsibility of Emerging Markets Queries in Finance and Business local organization.
doi:10.1016/S2212-5671(12)00131-1

Emerging

Monica Violeta Achim et al. / Procedia Economics and Finance 3 (2012) 132 137

1. Introduction
Unprecedented global economic development culminating in the current economic crisis demonstrates that
the results of research studies on the prediction of bankruptcy risk are insufficient. Moreover, the financial
crisis increases the risks that affect the proper functioning of a company. This increase can be translated by an
increase in uncertainty about its ability to continue working, Robu & Mironiuc, 2012.
In this context, a review of predictive models of bankruptcy risk is imperative. A score function is
influenced by characteristics of the country or region for which it was created, by the economic and financial
development level of the country concerned, by the industry in which companies are operating, by the
accounting system used, by the influence of taxation, by the predominantly type of financing etc.
This paper aims to select the optimum combination of predictable financial indicators for the bankrupt
companies. In the future research we will continue our study with an application of multiple discriminant
analysis in order to achieve a classification of companies into "bankruptcy" and "non-bancruptcy" based on a
score function.
2. Literature review
The earlier work of Beaver, 1966, indicated that the financial ratios can predict the likelihood of
bankruptcy. His univariate study evidenced that the financial ratios of bankrupt firms generally differ from
those of non bankrupt firms and pointed out the importance of cash flow-to-debt ratio. The work began by
Beaver was continued by Altman, 1968, who introduced the multivariate discriminant technique for predicting
. Altman, 1968 found that his five ratios outperformed Beaver's cash flow to total debt
ratio. These ratios are: X1=working capital/total assets; X2=total retained earnings/total assets; X3=earnings
before interest and taxes/total assets; X4=market value of equity / book value of total debt; X5=sales/total
assets.
Both Beaver and Altman are considered pioneers of bankruptcy risk modelling based on financial criteria
aggregates by multiple discriminant analysis technique.
By using financial ratios, the accuracy of the prediction of bankruptcy of a company exceeds 90%,Chen
and Shimerda, 1981. Although, it should be noted that some researchers, that is to say, Morris, 1998, argue that
since the bankruptcy was due to unforeseeable events, therefore it can not be predicted.
imposed that financial ratios of a specific business to be best interpreted as a group, Walton et al. 2003, rather
than making judgments on individual ratios because the interpretation of one ratio may be altered by other
ratios of the same business.
Among the most popular financial ratios used often by researchers were:
Profitability ratio represented by return on assets Beaver, 1966; Deakin, 1972; Libby, 1975; Ohlson, 1980;
Lennox, 1999; Abdullah, 2008; Zulkarnain, 2001; Lykke et. al 2004; Siminica, 2005.
Leverage ratio represented by total liabilities to total assets Beaver, 1966; Deakin, 1972; Ohlson, 1980;
Zmijewski, 1984 ; Zavgren et Dugan, 1989; Mohamed 2001; Anghel, 2002; Lykke et. al . 2004; Abdullah;
2008.
Cash flow ratio, represented by cash to total assets or cash to current liabilities (Lennox 1999; Zavgren et
Dugan, 1989; Low et al., 2001 and Zulkarnain, 2001; Ivoniciu ,1998; Bailesteanu 1998, Anghel, 2002;
Size of activity Ohlson, 1980; Lennox, 1999; Shumway, 2001; Lykke et. al., 2004.
Investigating a vast literature over 170 studies focused on prediction failure problems Bellovary, 2007
found, among those 752 factors which are utilized in the individual studies, the first ten financial ratio, as
follows: Net income / Total assets; Current ratio; Working capital/Total assets, Retained earnings / Total assets;
Earnings before interest and taxes EBIT / Total assets; Sales / Total assets; Quick ratio; Total debt / Total assets
; Current assets / Total assets; Net income / Net worth.

133

134

Monica Violeta Achim et al. / Procedia Economics and Finance 3 (2012) 132 137

3. Methodology
3.1. Sample and data
The sample consists in 53 companies that activate in the Romanian industrial sector, more specifically
manufacturing. These are all companies that are listed or were listed at the Bucharest Stock Exchange.
According to the methodology, companies are divided into two categories:
A part of them, more precisely a number of 35 companies have a good financial situation we will call them
his assessment is done by keeping the company in the category
the entire analyzed period 2000-2011.
Another part consists in a number of 18 companies, that have financial difficulties
This assessment is based on the financial deterioration during the period of analysis, to the
extent that these companies reach the point of unlisted over the analyzed period 2000-2011.
Data was collected from the financial statements of the companies for three years in range among the period
2000-2011, as it is provided by the site of Bucharest Stock Exchange, www.bvb.ro.
3.2. Selection of variables
Based on the results of specialized studies but also on the source of information we have accessed, we
selected as financial variables representative for the company's financial condition, the following financial
ratios:
Table 1. Financial variables taken into account in our study

Financial variables
Return on equity (ROE)
Return on assets (ROA)
Financial expenses ratio 1
Financial expenses ratio 2
Capital engaged ratio
Financial leverage ratio
Receivables turnover
Assets turnover
Working capital ratio
Currents assets turnover
Flexibility ratio
Debts turnover
Unpaid debts ratio
Current liquidity
General solvency
Solvency on long term
Source: own prelucrations

Formula
Profitability ratios
Return
Operating profit/ Total assets
Financial expenses / Operating result
Financial expenses/Sales
Leverage ratios
Capital engaged / T
Total debts /
Activity ratios
Sales /Receivables
Sales /Total assets/
Own equity/ Fixed assets
Sales/Current assets
Working capital/Total assets
Sales/Total debts
Unpaid debts / Total debts
Liquidity and solvency ratios
Current assets/Currents debts
Total assets/Total debts
Long term debts /(Total assets-Currents debts)

The variables for the analysis represent the average values for the period considered.

135

Monica Violeta Achim et al. / Procedia Economics and Finance 3 (2012) 132 137

3.3. Methods of work

For the present research we have employed the principal component analysis. This is a multidimensional
analysis technique used to partition a large number of variables into a smaller number of groups called
principal components Pearson, 1901. While between the initial variables there are relationships in the form of
correlations, between the principal components obtained, correlations are nil, i.e. the principal components are
independent among them. The technique is very much used as it allows for dimension reduction without the
loss of information. However, only components with an eigenvalue greater than 1 are preserved.
Each principal component obtained is a combination of several initial variables:
Ci = ai1*X1 + ai2*X2

ik*Xk,

(1)

where X denotes the initial variables, i the number of the component, k the number of initial variables that
form the component and aik the coefficients of the variables.
For the principal components analysis to be applied, correlations have to be present between the initial
variables. Moreover, there are two tests that show if running the analysis for the sample used makes sense. The
first one is the Kaiser-Meyer-Olkin Measure of Sampling Adequacy which should be greater than 0.5 - the
y
which tests that the variables are unrelated, based on the correlation matrix. Evidently, for the principal
component analysis to be useful, the null hypothesis must be rejected, i.e. the significance level must have
values lower than 0.05.
At least half of the variance of the initial variables should be explained by the solutions obtained. If the
opposite occurs, the original variables with such communality values must be removed and the analysis run
again. Among the validation methods there is the identification of complex structures. These appear when one
original variable has loading greater than 0.4 in more than one component. When this happens, that specific
variable must be excluded from the analysis. The last method used for validation is the Cronb
coefficient, Cronbach, 1951. This is computed for each of the components obtained, and is measuring the
legitimacy of getting together the objects of analysis. The minimum acceptable level is 0.6. Below this level,
the internal consistency is poor or unacceptable. Above this level, the consistency is questionable, between 0.6
and 0.7, acceptable, 0.7 to 0.8, good, 0.8 0.9, and excellent, above 0.9, George & Mallery, 2003; Kline, 1999;
Cortina, 1993.
4. Results
The initial list of variables used was presented above, in the methodological part.
When running the analysis, the KMO test was situated at the limit value of 0.49999, aproximatley 0.5, while
the Bartlett test had a significance level of 0.000 the best fit possible. This was an indicator that the analysis
had some errors. At a closer look in the communalities table we identified three variables for which the
variance accounted for by the factors where smaller than 0.5. Consequently, they were excluded and the
analysis rerun. These variables are: Financial expenses ratio 1, Unpaid debts ratio, General solvency .The
results of the KMO test did not improve in the second stage of the analysis either. When checking again for
errors, they were discovered in the Rotated Component Matrix. The phenomenon of complex structure
appeared for two variables: Capital engaged ratio and Debts turnover. They both had correlations greater than
0.4 in more than one component. Consequently, they were removed and the analysis run again.
statistically significant, meaning that the analysis makes sense for our sample of data but another variable had
to be removed due to the communality value: Receivables turnover.

136

Monica Violeta Achim et al. / Procedia Economics and Finance 3 (2012) 132 137

Finally, after all the validation procedures were passed, the principal component analysis identified four
principal components, explaining more than 85% of the total variance. The composition of the 4 groups is:
Component 1 is made up of: Return on assets, Working capital ratio, Flexibility ratio, Current liquidity.
Component 2 consists in the following variables: Return on equity, Financial leverage ratio.
Component 3: Total assets turnover, Current assets turnover
Component 4: Financial expenses ratio 2, Solvency on long term
The standardized equations of the four components are:
C1 = 0.009 * Return on equity + 0.371 * Return on assets + 0.093 * Financial expenses ratio 2 +
+ 0.005 * Financial leverage ratio + 0.03 * Total assets turnover + 0.286 * Working capital ratio
- 0.018 *Current assets turnover + 0.336 * Flexibility ratio + 0.202 * Current liquidity +
+ 0.101* Solvency on long term
C2 = -0.484 * Return on equity - 0.066 * Return on assets - 0.135* Financial expenses ratio 2 +
+ 0.490 * Financial leverage ratio + 0.083 * Total assets turnover - 0.020 * Working capital ratio
- 0.009 * Current assets turnover + 0.056 * Flexibility ratio + 0.006* Current liquidity +
+ 0.075* Solvency on long term
C3 = -0.016 * Return on equity + 0.218 * Return on assets - 0.109 * Financial expenses ratio 2 +
+ 0.050 * Financial leverage ratio + 0.459 * Total assets turnover + 0.011 * Working capital ratio +
+ 0.478 * Current assets turnover - 0.03 * Flexibility ratio - 0.184* Current liquidity +
+ 0.121 * Solvency on long term
C4 = 0.054 * Return on equity + 0.288 * Return on assets + 0.574 * Financial expenses ratio 2
- 0.054* Financial leverage ratio - 0.198 * Total assets turnover - 0.057 * Working capital ratio +
+ 0.107 * Current assets turnover + 0.086 * Flexibility ratio - 0. 079 * Current liquidity+
+ 0.516 * Solvency on long term
As the smallest value obtained was for the first component, 0.660, the reliability analysis is satisfied.
5. Conclusions
The goal of the present analysis was to see if there is possible to group the most important variables found in
the economic literature in respect to the goal of finding possible combinations of financial indicators that could
be used to assess the bankruptcy probability of a company. For this, we have used a sample of 53 Romanian
manufacturing companies, both with good financial situation and with precarious financial situations.
Running the principal component analysis, with all its validation methods, we have succeeded in grouping
the variables under discussion for the present sample in such a way as the information represented by them to
be relevant. The total information in the sample was divided between the four components and 6 independent
variables. However, the four principal components obtained account for more than 85% of the total variance.
Additionally, the validation tests applied have demonstrated that the sets of variables can easily be substituted
by these components in further analyses.
However, we are aware of the drawbacks of the present research, among which the most important is the
volume of the sample. The minimum admitted for such an analysis is 50, while our sample consists in 53
companies. So the results, even though validated, are accepted with caution, mainly due to the above presented
aspect. Future developments of this research include the extension of the sample by adding more companies
and running simulations on random samples generated with the same properties as the initial one.
Moreover, cautiousness is
than 0.7 in all cases. The component for which the coefficient has the value of 0.66 is the first one. Following
the interpretation of this value, the internal consistency of the variable exists, but is questionable.
Future research will involve, beside the above presented issues, using the components obtained in order to

Monica Violeta Achim et al. / Procedia Economics and Finance 3 (2012) 132 137

construct a score function for the manufacturing sector in Romania that could be used to analyze the risk of
bankruptcy. An extension of this goal is to construct score functions for bankruptcy analysis for other sectors
and even for the entire Romanian economy.
Acknowledgements
This work was supported by the project "Post-Doctoral Studies in Economics: training program for elite
researchers - SPODE" co-funded from the European Social Fund through the Development of Human
Resources Operational Programme 2007-2013, contract no. POSDRU/89/1.5/S/61755.
References
Altman E. I. 1968. Financial ratios, discriminant analysis and the prediction of corporate bankruptcy, Journal of Finance, Vol. 23, No. 4,
pp. 589 610.
Altman E. I. 1980. Commercial bank lending: process, credit scoring and costs of error in lending Journal of Financial and Quantitative
Analysis, Vol. XV, No.1, pp. 35 51.
Aziz, M. A., and Humayon A. 2006. Predicting Corporate Bankruptcy: Where we Stand? Corporate Governance, 2006, v6(1), 18-33.
Bellovary et. al. 2007 A Review of Bankruptcy Prediction Studies: 1930 to Present, Journal of Financial Education, Volume 33, Winter
2007.
Chen, K. H., and Thomas A Shimerda. 1981. An Empirical Analysis of Useful Financial Ratios Financial Management, Spring 1981,
v10(1), 51-60.
Collins, R. A.1980. An Empirical Comparison of Bankruptcy Prediction Models , Financial Management, Summer 1980, v9 (2), 52-57.
Cronbach, L. J. 1951. Coefficient alpha and the internal structure of tests. Psychometrika, 16(3), 297-334
Dugan, M. T., and Christine V. Zavgren, 1989. How a Bankruptcy Model Could be Incorporated as an Analytical Procedure.
The CPA Journal, May 1989, v59(5), 64-65.
Cortina, J.M. 1993. What is coefficient alpha? An examination of theory and applications. Journal of Applied Psychology, 78, 98-104
George, D., & Mallery, P. 2003. SPSS for Windows step by step: A simple guide and reference. 11.0 update (4th ed.).
Kline, P. 1999.The handbook of psychological testing (2nd ed.). London: Routledge
Journal of Accounting Research 18(1): 109-131.
Pearson, K. 1901. On Lines and Planes of Closest Fit to Systems of Points in Space (PDF). Philosophical Magazine 2 (6): 559 572.
Siminica M, Circiumaru D., Murarita I. 2004. Consideratii privind analiza discriminanta Revista Tinerilor Economisti Anul II~ Nr.2,
Aprilie 2004, pp.81-86
Robu M-A., Mironiuc M., Robu I-B, Un model practic pentru testare a ipotezei de "going-con
cotate, in,
Revista Audit Financiar, no. 2/2012, pp. 13-25
Journal of Business 74(1): 101-124.
Journal of Banking and Finance 8(2): 199-227.
Walton P, Haller A., Raffournier B. 2003. International accounting, second edition, Ed. Thomson, London

137