International Journal of Banking and Finance

Volume 10 | Issue 1 Article 3

8-14-2013

Modeling credit risk: An application of the rough

set methodology
Reyes Samaniego Medina
Pablo de Olavide University, Spain

Maria Jose Vazquez Cueto

Seville University, Spain

Follow this and additional works at: [Link]

Recommended Citation
Medina, Reyes Samaniego and Vazquez Cueto, Maria Jose (2013) "Modeling credit risk: An application of the rough set
methodology," International Journal of Banking and Finance: Vol. 10: Iss. 1, Article 3.
Available at: [Link]

This Journal Article is brought to you by the Faculty of Business at ePublications@bond. It has been accepted for inclusion in International Journal of
Banking and Finance by an authorized administrator of ePublications@bond. For more information, please contact Bond University's Repository
Coordinator.
 Medina and Vazquez Cueto: Modeling credit risk

The International Journal of Banking and Finance, Volume 10 (Number 1), 2013: Pages 34-56

MODELING CREDIT RISK: AN APPLICATION OF THE ROUGH SET

METHODOLOGY

Reyes Samaniego Medina and Maria Jose Vazquez Cueto

Pablo de Olavide University, Spain and Seville University, Spain
___________________________________________________________

Abstract
The Basel Accords encourages credit entities to implement their own models for measuring
financial risk. In this paper, we focus on the use of internal ratings-based (IRB) models for the
assessment of credit risk and, specifically, on one component that models the probability of
default (PD). The traditional methods used for modeling credit risk, such as discriminant analysis
and logit and probit models, start with several statistical restrictions. The rough set methodology
avoids these limitations and as such is an alternative to the classic statistical methods. We apply
the rough set methodology to a database of 106 companies that are applicants for credit. We
obtain ratios that can best discriminate between financially sound and bankrupt companies, along
with a series of decision rules that will help detect operations that are potentially in default.
Finally, we compare the results obtained using the rough set methodology to those obtained
using classic discriminant analysis and logit models. We conclude that the rough set
methodology presents better risk classification results.

Keywords: Rating, Credit risk, Basel Accords, Rough sets

JEL Classification: G21, G32

1. Introduction

The Basel Accords opened the way for and encouraged credit entities to implement their own
models for measuring financial risks. In this paper, we focus on the use of internal ratings-based
(IRB) models for the assessment of credit risk and, specifically, on one of the approaches to
model the probability of default (PD).

Published by ePublications@bond, 2013 1

 International Journal of Banking and Finance, Vol. 10, Iss. 1 [2013], Art. 3

The traditional methods used for modeling credit risk, such as discriminant analysis and
logit and probit models, start with several statistical restrictions. The rough set methodology
avoids these limitations and is presented as an alternative to the classic statistical methods.

The objective of our study is to apply the rough set methodology to a database composed of
106 companies that are debtors of the same financial entity to obtain the ratios that best
discriminate between healthy and bankrupt companies. A second objective is to find a series of
decision rules that will help detect potentially failing credit operations as a first step in modeling
the probability of default. Finally, we compare the results obtained using the rough set
methodology to those obtained using classic discriminant analysis and logit models. We
conclude that the rough set methodology presents good risk classification results.

This paper is structured as follows. Section 2 reviews the most significant empirical studies.
Section 3 introduces the theory of rough sets. In section 4, we continue with a description of the
sample of companies used for the empirical study. The empirical application is described in
section 5 wherein we first use the rough set methodology to determine the variables that may
explain the default, and then compare the results obtained using this methodology to those
obtained using classic discriminant analysis and logit models. Finally, in section 6, we draw a
series of conclusions.

2. Literature Review

The models for predicting business failure and estimating the probability of default (PD)
required by the Basel Accords have been the subject of several studies, conducted not only by
academics but also by the financial sector itself. All of the theoretical effort has been focused on
the modeling of the stochastic process associated with insolvency and on determining the
variables that must be included in these models. Among these traditional models, we can
distinguish between univariate and multivariate models. Univariate models examine the behavior
of each variable separately to explain any insolvency. One of the classic studies using this
method was conducted by Beaver (1966), who found a number of financial ratios that could
discriminate between healthy and bankrupt companies during the 5-year period prior to the
occurrence of the actual default. Other notable studies were those conducted by Courtis (1978)
and Altman (1993).

Unlike the univariate models, the multivariate models combine the information provided by
a set of variables. The study that pioneered this method was performed by Altman in 1968, in

35

[Link] 2
 Medina and Vazquez Cueto: Modeling credit risk

which he proposed a discriminant analysis that combined the information provided by 25

financial ratios. A wide variety of studies have been based on discriminant analysis, including
those by Dambolena (1980) and Laitinen (1991). In Spain, Cabedo et al. (2004) presented an
adaptation of the discriminant model to calculate the probability of default in companies. The
model was applied to a portfolio of hypothetical borrowers from the same sector to calculate the
regulatory capital according to the foundational IRB method of the Basel Accord. The
importance of this technique is demonstrated in the bibliographic review by Dimitras (1996).
After analyzing 158 articles on the prediction of business insolvency during the period from
1932 to 1994, Dimitras concluded that discriminant analysis is the model that is most frequently
used to resolve this type of problem.

Other authors have opted for logit and probit analysis. Ohlson (1980) was the first to apply
this type of technique to predict company insolvency. Wilson (1997) developed the Credit
Portfolio View model for McKinsey, establishing a discrete process with multiple periods. With
this methodology, the probability of default is obtained through logit functions of indices of
macroeconomic variables that, in some ways, represent the functioning of the economy
(Zmijewski, 1984). Dimitras (1996) found that the logit model is the second most frequently
used model for resolving the problem of company bankruptcy.

Fernández (2005), in an attempt to combine univariate and multivariate analysis, conducted

an empirical study in which he used a prior univariate analysis to select those ratios with greater
discriminant power within each of the categories of ratios established from among the 23 ratios
initially considered.1 Subsequently, he performed a logit and probit multivariate analysis to
obtain scores for each company; these scores enabled a rating system to be established and
default probabilities to be assigned. Trucharte et al. (2002) developed a system for rating
borrowers by estimating a logistic regression model that utilizes economic and financial
information. The scores obtained are used to establish homogeneous categories in which the
various borrowers are classified or rated and the probability of default that can be assigned to
each category.

More recently, Altman and Sabato (2007) developed a distress predictor model specifically
for the SME sector and analyzed its effectiveness against that of a generic corporate model. They

1
These categories were liquidity, leverage, activity, debt cover and productivity.
36

Published by ePublications@bond, 2013 3

 International Journal of Banking and Finance, Vol. 10, Iss. 1 [2013], Art. 3

used a logit regression model technique on panel data from more than 2,000 US firms over the
period from 1994 to 2002.

In parallel with these studies, other methods have been explored to overcome the restrictive
hypotheses that models of statistical inference impose on the variables. These hypotheses usually
do not conform to reality and distort the results obtained. For these reasons, Eisenbeis (1977),
Ohlson (1980) and Zavgren (1983) questioned the validity of the traditional models. In
particular, techniques originating from the field of artificial intelligence have begun to be used;
programs have been produced that are capable of generating knowledge from empirical data and
then utilizing that knowledge to make inferences based on new data. Within this approach, we
can distinguish techniques that seek knowledge by identifying patterns in the data. Among these
are various classes of neural networks and other techniques that infer decision rules from the
base data. The rough set methodology belongs to this last group of techniques. Authors such as
Dimitras et al. (1998) and Daubie et al. (2002) have applied this technique to the classification of
commercial loans. Other authors, such as Ahn et al. (2000), have combined the rough set
methodology with neural networks to predict company failure. In Spain, various studies can
similarly be found that apply the rough set methodology for the prognosis of company
insolvency. Segovia et al. (2003) applied this technique to the prediction of insolvency in
insurance companies, and Rodríguez et al. (2005) utilized it for the same purpose in a sample of
small- and medium-sized enterprises (SMEs).

3. Theoretical Framework of the Rough Set Methodology

Rough set theory, first proposed by Z. Pawlak in 1982, is considered as an appropriate tool for
handling cases in which there is considerable vagueness and imprecision. More specifically, the
method is effective at working with problems of multidimensional classification (Pawlak et al.,
1994). The basic idea rests on the indiscernibility relation that describes elements that are
indistinguishable from each other. Its principal objective is to find basic decision rules that
enable the acquisition of new knowledge. Its key concepts are discernibility, approximation,
reducts and decision rules.

The point of departure for the method is the existence of an information/decision table in
which each element is characterized by a set of variables (attributes) and a decision variable that
classifies the element into one of two or more categories. Indiscernibility exists when two
elements are characterized by the same properties for all variables, but the categories in which
they are classified do not coincide. This is the basis of the rough set theory. In such a case, for
37

[Link] 4
 Medina and Vazquez Cueto: Modeling credit risk

each class of decision or category X and for each subset B that contains variables, two sets are
constructed; these sets are called the set of the lower approximation of the decision class and the
set of the upper approximation of the decision class, respectively. The set of the lower
approximation of decision class X with respect to the variables B, which is called BX, is given
by the group of all elements that, being characterized by B, belong to class X with complete
certainty. The set of the upper approximation of decision class X, which is called BX , is given
by the group of elements that, based on the information B that we possess, may belong to class
X, but about which we cannot be sure. The elements that are different between the two sets form
the "doubtful" elements, which are those elements that, using only the information contained in
B, are not known with complete certainty to belong to class X. When these different elements
exist (i.e., when the difference is not zero), it is said that class X is a rough set with respect to the
subset of variables B. This set can be characterized numerically by the quotient between the
cardinal of the set of lower approximation and the cardinal of the set of upper approximation.
This quotient is known as the "accuracy of approximation". If various decision classes exist, the
sum of the cardinals of the lower approximations divided by the total of all elements is known as
the "quality of approximation of the classification, by means of set B", and this is the percentage
of the elements that have been correctly classified.

Another important aspect of this technique is the reduction of the initial table of data, which
eliminates the redundant information. This process is performed using the reducts. A reduct is a
minimum set of variables that conserve the same capacity for classifying the elements as the full
table of information. A reduct is thus an essential part of the information and constitutes the most
concise way of differentiating between the decision classes.2

The final stage of the rough set analysis is the creation of decision rules, rules that allow us
to determine whether a given element belongs to particular decision classes. These rules
represent knowledge and are generated by combining the reducts with the values of the data
analyzed. A decision rule is a logical statement of the following type: "IF particular conditions
are met, THEN the element belongs to a particular decision class". These rules allow us to
classify new elements easily..3

2
The reducts were obtained based on the equivalence classes that defined the indiscernibility
relation on the set of observations.
3
For more detail on the formal mathematical aspects of the methodology, Komorowski et al.
(1999) may be consulted.
38

Published by ePublications@bond, 2013 5

 International Journal of Banking and Finance, Vol. 10, Iss. 1 [2013], Art. 3

4. Data and Variables

A. Selection of the Sample.

We adopted the following approach for both selecting the sample and choosing the independent
variables. Following Altman (1968), we have paired a number of healthy and bankrupt
companies of similar size and sector, thus taking a sample in which the bankrupt companies
represent 50% of the total. When selecting the sample, the data under consideration should be
obtained for the same period of time for healthy and bankrupt companies alike. However, the
companies in bankruptcy or suspension of payments tend to delay the presentation of their
accounting data in the time period prior to their declarations of insolvency. To overcome this
inconvenience, we have collected the accounting data from the last full year prior to the
bankruptcy from the most recent data available.

In the development of our model, we employed a database, provided by a Spanish savings

bank that contains information on companies that requested and obtained a loan from the bank.
These companies were divided into two groups: healthy and bankrupt. In particular, the sample
of bankrupt companies used for the analysis only included those companies with loans whose
unpaid debt, whether of interest or principal, amounted to a percentage of more than 10% of the
full risk accepted. The computation date was December 31, 2003.

The group of healthy companies, i.e., those that did not default in the time horizon
considered, was selected using the individual pairing technique, controlled by those
characteristics that could affect the relationships between financial ratios and failure. Each
company in the failed group was matched with a healthy company of the same industry and same
approximate size. In relation to the sector, the pairing was achieved at a level of four digits of the
C.N.A.E. (National Classification of Economic Activities) of 1993. The criterion adopted for
pairing by size was the total assets.

As a homogenizing factor for all of the companies, we controlled for the conditions that the
total amount of the customer's operations with the savings bank, i.e., their live risk, should
exceed 60,120 euros and that the companies should all be public limited companies (PLCs),
which facilitates access to their accounting statements.

39

[Link] 6
 Medina and Vazquez Cueto: Modeling credit risk

In total, the sample contained 106 Spanish companies, 53 failed and 53 healthy, with a very
diverse spread of economic activities.4

B. Selection of the Independent Variables for the Models

The independent variables chosen for the construction of the models were selected from the
financial statements, principally from the balance sheet and the profit and loss account, of the
companies that comprise the sample. These accounting statements were extracted from the SABI
database, developed by Informa, S.A., which includes more than 95% of the companies that
present their accounts in the Mercantile Register in Spain. Given that most of the companies that
went bankrupt presented their financial statements neither in the preceding year nor in the two
years prior to the date of default, we obtained the latest data available corresponding to the year
prior to the company bankruptcy, as explained above. Thus, the year t-1 corresponds to that of
the latest available accounts.

Table 1: Ratios considered in the analysis

This table summarizes the variables that are potentially explanatory for company bankruptcy.
These, in general, include ratios of liquidity, indebtedness, structure, rotation, generation of
resources and profitability.

Liquidity Ratios

R1=Current assets / current liabilities

R2=(Quick + available assets) / current

Degree to which the company's assets liabilities
that can be liquidated, in the short R3=Available assets / current liabilities
term, are sufficient to meet the
payments required for the short-term
debts contracted.
R4=(Quick + available assets - current
liabilities) / (Operating costs + Personnel costs
+ Variation provisions + Other operating
costs)

Indebtedness Ratios

4
We have excluded property development and property sales companies from the analysis, as
these companies have characteristics that are very peculiar and different from other companies.
In the assessment of the loan application made by this type of company, the decisive factor for
granting the loan is the viability of the specific project for which the loan is sought. This
information is not reflected in corresponding accounting statements.
40

Published by ePublications@bond, 2013 7

 International Journal of Banking and Finance, Vol. 10, Iss. 1 [2013], Art. 3

R5=Long-Term Debt / Net Worth

R6=Net Worth / Total Liabilities

R7=Long-Term Debt / (Long-Term Debt +

Relationship between the different Current Liabilities)
components of the liabilities, in the
short and long term, and the
company’s own funds and between R8=Financial Costs / (Long Term Debt +
the cost of the debt and the liabilities Current Liabilities)
or the profits and funds generated. R9=Financial Costs / (Gross Profits +
Provision for Amortization)

R10=Financing costs / Gross profits

R11=Long-Term Debt / Total Liabilities

Structural Ratios

R12=(Current Assets - Current Liabilities) /

Proportionality between the balance Total Assets.
sheet items of assets and liabilities
and in the composition of these items.

R13=Current Assets / Total Assets.

Rotation Ratio

Measure of the dynamism of the R14=(Current Assets – Current Liabilities) /

business activity in relation to the (Net Turnover + Other Income from
structure of the company. Operations)

Resource Generation Ratios

Relationship of the self-financing R15=(Net Profit/Loss for period +

capacity of the company to various Amortization Provision) / (Net Turnover +
accounting magnitudes. Other Income from Operations)

41

[Link] 8
 Medina and Vazquez Cueto: Modeling credit risk

R16=(Net Profit/Loss for period +

Amortization Provision) / Current Liabilities

R17=(Net Profit/Loss for period +

Amortization Provision) / (Long-Term Debt +
Current Liabilities)

R18=(Net Profit/Loss for period +

Amortization Provision) / Total Liabilities

R19=(Gross Profits + Amortization Provision)

/ Current Liabilities

Profitability Ratios

R20=(Operating Profit/Loss + Financial

Income + Profits from financial investments +
Exchange rate gains) / Total Assets

R21=Profit/Loss from ordinary activities /

Comparison of the profit obtained at Total Liabilities
various levels with the resources
invested R22=Pre-tax Profits / Net Worth
R23=Pre-tax Profits / Total Liabilities

R24=Profit/Loss for the period / Net Worth

R25=Gross Profits / Total Assets

Source: Trujillo et al., (2004)

The accounting information derived from the sample selected was subjected to a meticulous
study with the aim of detecting and resolving possible anomalies or significant incidents that
could distort the final analysis. Those atypical companies with clear and insuperable anomalies
in their accounts were excluded from the sample. For example, those companies that presented
profits despite being in a situation of default were eliminated.

Twenty-five ratios were selected by choosing a broad set of variables that are potentially
explanatory for company bankruptcy based on the frequency and efficacy with which the ratios
have been used in other predictive models of company insolvency or in the analysis of banking
risks.

42

Published by ePublications@bond, 2013 9

 International Journal of Banking and Finance, Vol. 10, Iss. 1 [2013], Art. 3

The variables used include ratios of liquidity, indebtedness, structure, rotation, generation
of resources and profitability. The specific ratios considered in the analysis are given in Table 1.

5. Empirical Application

A. Application of the Rough Set Mrthodology

For the empirical application, the values of the 25 economic/financial ratios shown in Table 1
were calculated for each of the 53 bankrupt companies for the financial year before entering into
default, and a similar procedure was adopted for each matched healthy company. This process
produced a table of information containing 106x25 data items. An additional column that
indicates whether the company in question is in a situation of bankruptcy or health is included in
the table of information. Thus, we have assigned the value 0 to bankrupt companies and the
value 1 to the matched healthy companies. We thus obtain an information-decision table with
106x26 data items.

From these data, we determine which ratio or ratios of the 25 variables serve to explain a
company’s state of default as the first step in calculating the probabilities of default.

First, Napierian logarithms of the values of the ratios were computed to avoid problems with
the normality of the variables when applying the discriminant analysis. Then, given the nature of
the variables considered, we proceeded to discretize the values. This is not an essential
requirement for the application of the technique, but it facilitates the interpretation of the results;
it is a more consistent way to identify bankrupt or healthy companies when the values of the
variables considered fall within the same range but do not coincide exactly. For this, we utilized
the codification given in Table 2.5

The next step is to determine the accuracy provided by the explanatory variables using the
ROSE software. The quality of the approximation is 1.6

5
We discretized the variables by grouping them into four ranges based on the number of
observations belonging to each range. For this, we utilized the ROSE software, provided by the
Institute of Computing Science of the Poznan University of Technology, and we thank the
Institute for making this software available to us.
6
The quality of the approximation is expressed by the ratio between the number of companies
classified correctly and the total number of companies that comprise the sample.
43

[Link] 10
 Medina and Vazquez Cueto: Modeling credit risk

Table 2: Codification ranges of the variables

Napierian logarithms of the values of the ratios are first computed. Then, the values are
discretized according to the ranges identified in this table.

CODIFIED VALUE

Variables
0 1 2 3

(-inf, (0.00434058, (0.00437793, (0.00467188,

R1 +inf)
0.00434058) 0.00437793) 0.00467188)
(-inf, (0.00131066, (0.00215777, (0.00653424,
R2 +inf)
0.00131066) 0.00215777) 0.00653424)
(-inf, 3.01154e- (3.01154e-005, (4.19928e-005, (0.000818944,
R3 +inf)
005) 4.19928e-005) 0.000818944)
(-inf, - (-0.00126886, - (-0.000469693, - (-0.000412292,
R4 +inf)
0.00126886) 0.000469693) 0.000412292)
(-inf, - (-0.000651859, (0.000583821, (0.00176257,
R5 +inf)
0.000651859) 0.000583821) 0.00176257)
(-inf, 1.38138e- (1.38138e-005, (0.00072438, (0.00077608,
R6 +inf)
005) 0.00072438) 0.00077608)
(-inf, 1.46802e- (1.46802e-005, (6.811e-005, (0.000260608,
R7 +inf)
005) 6.811e-005) 0.000260608)
(-inf, 4.45802e- (4.45802e-005, (6.1332e-005, (0.000253627,
R8 +inf)
005) 6.1332e-005) 0.000253627)
(-inf, - (-0.00032564, (0.00129406, (0.00412803,
R9 +inf)
0.00032564) 0.00129406) 0.00412803)
(-inf, - (-0.00032564, (0.00170486, (0.00355543,
R10 +inf)
0.00032564) 0.00170486) 0.00355543)
(-inf, (0.00308829, (0.00355161, (0.00430769,
R11 +inf)
0.00308829) 0.00355161) 0.00430769)
(-inf, - (-0.00169722, (3.00724e-005, (0.000163319,
R12 +inf)
0.00169722) 3.00724e-005) 0.000163319)
(-inf, (0.00131974, (0.00255195, (0.00265304,
R13 +inf)
0.00131974) 0.00255195) 0.00265304)
(-inf, - (-0.000937993, (0.000271131, (0.00183654,
R14 +inf)
0.000937993) 0.000271131) 0.00183654)
R15 (-inf, 9.4422e- (9.44223e-006, (7.94461e-005, (0.000744896,
44

Published by ePublications@bond, 2013 11

 International Journal of Banking and Finance, Vol. 10, Iss. 1 [2013], Art. 3

006) 7.94461e-005) 0.000744896) +inf)

(-inf, 3.03552e- (3.03552e-005, (0.000232274, (0.000290301,
R16 +inf)
005) 0.000232274) 0.000290301)
(-inf, 6.49981e- (6.49981e-006, (0.000152804, (0.000620383,
R17 +inf)
006) 0.000152804) 0.000620383)
(-inf, 6.4581e- (6.45818e-006, (1.90195e-005, (0.000153841,
R18 +inf)
006) 1.90195e-005) 0.000153841)
(-inf, 5.4507e- (5.45079e-005, (5.56532e-005, (0.000473715,
R19 +inf)
005) 5.56532e-005) 0.000473715)
(-inf, - (-0.000115164, (5.14362e-005, (0.00121164,
R20 +inf)
0.000115164) 5.14362e-005) 0.00121164)
(-inf, - (-0.000148011, (4.68013e-005, (0.00120903,
R21 +inf)
0.000148011) 4.68013e-005) 0.00120903)
(-inf, 1.16457e- (1.16457e-005, (0.000203684, (0.00417805,
R22 +inf)
005) 0.000203684) 0.00417805)
(-inf, 7.0378e- (7.0378e-007, (1.22549e-005, (0.000117764,
R23 +inf)
007) 1.22549e-005) 0.000117764)
(-inf, 1.10402e- (1.10402e-005, (0.000426896, (0.00416212,
R24 +inf)
005) 0.000426896) 0.00416212)
(-inf, 1.2036e- (1.2036e-005, (2.53061e-005, (0.000161758,
R25 +inf)
005) 2.53061e-005) 0.000161758)
Inf. = Infinite.

Next, we construct the reducts. Because there are correlations between the explanatory
variables, the program produces many reducts. Specifically, the program produced 18,241
reducts with and between 6 to 12 variables. Importantly, there are no core elements; there is no
variable that is essential for the classification or is more relevant than any other. The frequency
of appearance of each variable is shown in Table 3.

Table 3: Frequency of appearance of each variable in the reducts

The program produced 18,241 reducts with and between 6 to 12 variables

Variable R1 R2 R3 R4 R5 R6 R7

Frequency 23.78% 30.55% 42.46% 23.26% 26.31% 29.58% 24.89%

Variable R8 R9 R10 R11 R12 R13 R14

45

[Link] 12
 Medina and Vazquez Cueto: Modeling credit risk

Frequency 44.01% 32.34% 48.32% 33.71% 23.26% 38.31% 28.20%

Variable R15 R16 R17 R18 R19 R20 R21

Frequency 31.96% 31.96% 24.75% 25,60% 24.33% 28.39% 30.72%

Variable R22 R23 R24 R25

Frequency 39.72% 37.99% 34.67% 29.52%

We selected 3 of all possible reducts. The selection criteria are the following: first, the
reduct should contain the smallest possible number of variables; second, the variables should
present a high frequency of appearance in the different reducts; and finally, they should be
formed by the smallest number of ratios for each category considered. The ratios belonging to
each of the reducts are shown in Table 4.

Table 4: Selected reducts

The ratios belonging to each of the reducts

Reducts Variables
1 {R3,R10,R13,R17,R22,R25}
2 {R3,R10,R13,R14,R17,R20,R25}
3 {R3,R11,R13,R14,R15,R23}.
The first 2 reducts have been chosen because they include the ratios R3, R10 and R13, the
quotients that have high percentages of appearance and contain a ratio in each of the categories
studied. The third reduct was chosen because it presents ratios from all of the categories. We
limited our search to those reducts formed from a maximum of 7 ratios (one more than the
number of categories considered).

With each of the 3 reducts, decision rules were generated using the lem2 procedure;7 these
are shown in Table 5. The following points can be noted regarding the items in the table. We can
observe that the number of rules varies from one reduct to another; we find that 24, 28 and 22
rules are required to classify 100% of the observations correctly. The first reduct is the one that

7
Chan et al., (1994).
46

Published by ePublications@bond, 2013 13

 International Journal of Banking and Finance, Vol. 10, Iss. 1 [2013], Art. 3

Table 5: Selected reducts and corresponding decision rules

In all sub-tables below the decision rules are generated using the lem2 procedure.

Table 5.1: Reduct 1: Ratios R3, R10, R13, R17, R22, and R25

Rules of Classification Correct Classification

(D1 = 0): Healthy
1. (R3 = 2) & (R13 = 3) & (R17 = 1) & (R25 = 3) 9.43%
2. (R3 = 0) & (R10 = 3) 26.42%
3. (R10 = 2) & (R17 = 2) & (R22 = 1) & (R25 = 3) 3.77%
4. (R17 = 0) 33.96%
5. (R10 = 3) & (R17 = 2) & (R22 = 2) 5.66%
6. (R3 = 2) & (R13 = 1) & (R25 = 2) 11.32%
7. (R3 = 0) & (R17 = 1) 13.21%
8. (R10 = 1) & (R13 = 3) & (R22 = 3) 3.77%

9. (R3 = 0) & (R17 = 3) 5.66%

10. (R3 = 2) & (R13 = 0) & (R25 = 2) 5.66%

11. (R22 = 0) 37.74%

(D1 = 1): Bankrupt
12. (R10 = 1) & (R22 = 2) & (R25 = 3) 49.06%
13. (R10 = 2) & (R17 = 2) & (R22 = 2) & (R25 = 3) 16.98%
14. (R3 = 2) & (R17 = 3) & (R22 = 2) 9.43%
15. (R13 = 2) 9.43%
16. (R3 = 3) & (R17 = 3) 24.53%
17. (R3 = 2) & (R13 = 3) & (R25 = 2) 5.66%
18. (R25 = 1) 3.77%
19. (R3 = 2) & (R13 = 1) & (R25 = 3) 15.09%
20. (R3 = 0) & (R17 = 2) & (R25 = 2) 1.89%
21. (R3 = 2) & (R17 = 2) & (R22 = 3) 3.77%
22. (R3 = 1) & (R25 = 3) 7.55%
23. (R3 = 2) & (R10 = 3) & (R13 = 3) & (R17 = 2) & (R22 = 1) 1.89%
24. (R3 = 3) & (R25 = 2) 3.77%
47

[Link] 14
 Medina and Vazquez Cueto: Modeling credit risk

requires the fewest rules to identify the bankrupt companies. We can see also that the first reduct
is the one that has the rules with the greatest power of classification. Thus, the rules 11 [(R22 =
0) => (D1 = 0)] and 12 [(R10 = 1) & (R22 = 2) & (R25 = 3) => (D1 = 1)] of the first reduct
classify, respectively, 37.74% of the bankrupt companies and 49.06% of the healthy ones. No
other individual rule of the second or third reducts reaches such high percentages.

The above 2 rules tell us, first, that if the value of the Napierian logarithm of the profitability
ratio R22 (Pre-tax profits / Net Worth) is less than 1.16457e-005, then the company must be
classified as bankrupt. Second, if the value of the Napierian logarithm of the ratio of
indebtedness R10 (Financial Costs /Gross Profits) is between 0.00032564 and 0.00170486, that
of the profitability ratio R22 (Pre-tax profits / Net Worth) is between 0.000203684 and
0.00417805 and that of the ratio R25 (Gross Profits / Total Assets) is greater than 0.000161758,
the company must be classified as healthy.

Table 5.2: Reduct 2: Ratios R3, R10, R13, R14, R17, R20, and R25

Rules of Classification Correct Classification

(D1 = 0): Healthy
1. (R13 = 3) & (R14 = 1) & (R17 = 1) & (R25 = 3) 11.32%
2. (R10 = 3) & (R25 = 2) 22.64%
3. (R10 = 3) & (R14 = 2) 9.43%
4. (R10 = 0) 16.98%
5. (R3 = 2) & (R14 = 0) & (R17 = 2) 5.66%
6. (R10 = 2) & (R14 = 3) 1.89%
7. (R10 = 3) & (R13 = 0) 15.09%
8. (R3 = 0) & (R13 = 3) 26.42%
9. (R3 = 3) & (R13 = 0) & (R17 = 2) 1.89%
10. (R20 = 0) 32.08%
11. (R3 = 0) & (R17 = 3) 5.66%
12. (R14 = 1) & (R20 = 1) & (R25 = 2) 9.43%
13. (R13 = 0) & (R14 = 0) 16.98%
14. (R10 = 2) & (R13 = 3) & (R20 = 1) 3.77%

48

Published by ePublications@bond, 2013 15

 International Journal of Banking and Finance, Vol. 10, Iss. 1 [2013], Art. 3

15. (R20 = 3) 1.89%

16. (R10 = 3) & (R20 = 2) 1.89%
(D1 = 1): Bankrupt
17. (R10 = 1) & (R17 = 3) & (R20 = 2) 28.30%
18. (R10 = 2) & (R14 = 1) & (R17 = 2) & (R20 = 2) 15.09%
19. (R3 = 2) & (R10 = 1) & (R20 = 2) 20.01%
20. (R3 = 2) & (R13 = 1) & (R25 = 3) 15.09%
21. (R13 = 2) 9.43%
22. (R10 = 1) & (R14 = 2) 32.08%
23. (R3 = 1) & (R25 = 3) 7.55%
24. (R3 = 0) & (R13 = 1) & (R17 = 2) 3.77%
25. (R3 = 2) & (R13 = 3) & (R25 = 2) 5.66%
26. (R3 = 3) & (R13 = 1) 15.09%
27. (R14 = 2) & (R17 = 3) 26.42%

28. (R3 = 2) & (R10 = 3) & (R13 = 3) & (R14 = 1) & (R17 = 1.89%
2) & (R20 = 1)

Table 5.3: Reduct 3: Ratios R3, R11, R13, R14, R15, and R23

Rules of Classification Correct Classification

(D1 = 0): Healthy
1. (R3 = 0) & (R11 = 2) 26.42%
2. (R11 = 3) 26.42%

3. (R15 = 1) & (R23 = 3) 3.77%

4. (R3 = 2) & (R13 = 1) & (R14 = 0) 3.77%
5. (R23 = 0) 47.17%
6. (R13 = 0) & (R23 = 1) 3.77%
7. (R11 = 2) & (R13 = 3) & (R15 = 2) & (R23 = 2) 5.66%
8. (R15 = 3) & (R23 = 2) 7.55%
9. (R14 = 0) & (R15 = 3) 5.66%

49

[Link] 16
 Medina and Vazquez Cueto: Modeling credit risk

10. (R3 = 2) & (R11 = 0) & (R14 = 3) 7.55%

11. (R13 = 1) & (R23 = 1) 1.89%
12. (R11 = 0) & (R13 = 1) & (R14 = 1) & (R23 = 2) 1.89%
(D1 = 1): Bankrupt
13. (R13 = 1) & (R15 = 2) & (R23 = 3) 20.75%
14. (R3 = 2) & (R15 = 2) & (R23 = 3) 30.19%
15. (R3 = 2) & (R11 = 2) & (R15 = 1) & (R23 = 2) 7.55%
16. (R11 = 0) & (R14 = 2) 30.19%
17. (R3 = 3) & (R15 = 2) 30.19%
18. (R11 = 1) & (R23 = 2) 11.32%
19. (R13 = 2) 9.43%
17. (R3 = 3) & (R15 = 2) 30.19%
18. (R11 = 1) & (R23 = 2) 11.32%
19. (R13 = 2) 9.43%
20. (R3 = 1) & (R14 = 3) 1.89%
21. (R3 = 1) & (R15 = 2) 5.66%
22. (R3 = 2) & (R11 = 2) & (R13 = 3) & (R23 = 1) 1.89%

To verity these classification results, we performed a cross validation with 10 passes. We

present the results for each reduct in Table 6.

The first and third reducts are clearly more robust than the second, with respect to
percentages of correct classification. The third reduct, in addition to meeting the previously
imposed requirements, is the one that presents fewest type I and II errors (13.42% and 11.18%,
respectively, for the validation sample).

Given the above results, we are thus able to confirm that the rough set methodology leads to
good results in the classification of healthy and bankrupt companies and indicates which
variables are the most relevant of those considered. Thus, by applying these rules to new credit
operations, a bank would be able to detect possible defaults.

50

Published by ePublications@bond, 2013 17

 International Journal of Banking and Finance, Vol. 10, Iss. 1 [2013], Art. 3

Table 6: Percentages of correct classification

This table includes the results of a cross validation with 10 passes.

Correct Correct Correct

Reduct classification: classification: classification:
bankrupt healthy total
R3,R10,R13,R17,R22,R25 83.98% 87.90% 86.00%
R3,R10,R13,R14,R17,R20,R25 66.67% 84.83% 76.27%
R3,R11,R13,R14,R15,R23 86.58% 88.82% 88.82%

B. Comparison with the Discriminant Analysis

As mentioned above, discriminant analysis was the first technique that was widely used. It
still remains the most frequently utilized technique for measuring company insolvency.
Therefore, we have compared the results of the rough set methodology with those produced by
the discriminant analysis.

We used the same set of data to perform this comparison8. First, we utilized Box's M-
statistic to test the equality of the variance-covariance matrices between the two groups. From
the results, we can accept this hypothesis at a 5% level of significance. The discriminant function
(obtained using the ascending stepwise method, with Snedecor's F criterion for entry between
0.06 and 0.09) and the statistics associated with the model are shown in Table 7.

Table 7: Discriminant functions and statistics of the models

This table reports information on the discriminant analysis. The discriminant function is
obtained using the ascending stepwise method, with Snedecor's F criterion for entry between
0.06 and 0.09.

Discriminant function Wilks'

P-Value
(discriminant canonical function with standardized coefficients) lambda

Z = 0.554 R4 + 0.769 R6 – 0.337 R9 0.691 <0.0000

Base on the avove table, we can deduce that, according to the discriminant analysis, the
liquidity and indebtedness ratios are the most relevant items for determining the possible

8
The Napierian logarithms of the data were used, and their values were typified.
51

[Link] 18
 Medina and Vazquez Cueto: Modeling credit risk

insolvency of a company. The liquidity ratio positively affects a company’s probability of being
classified as healthy. With respect to the ratios of indebtedness, their influence will depend on
their definition. The ratio R6 (Net Worth / Total liabilities) has a positive influence, while the
ratio R9 (Financial Costs / (Gross Profits + provision for amortization)) has a negative influence
on the probability of the company in question being considered healthy.

Table 8 gives the results of the classification with the discriminant function applied to the
original sample and validated using the cross procedure.

Table 8: Percentages of correct classification base on discriminant analysis

This table reports the results of the classification with the discriminant function. The original
sample is used. The results are validated using the cross procedure.

Correct Correct Correct

classification: classification: classification:
bankrupt healthy total
Original sample 71.7% 83.0% 77.4%
Validation sample 71.7% 83.0% 77.4%

From tables 6 and 8 together, it can be deduced that the rough set methodology is a more
useful tool than discriminant analysis for the classification of the defaulting companies in the
database considered for our empirical application. With the correct choice of reducts, only 6 of
the 25 variables considered in the analysis are required to correctly classify 100% of the original
observations. Furthermore, more validation samples are correctly classified, giving type I and II
errors of 13.42% and 11.18%, respectively, compared to the errors of 28.3% and 17%,
respectively, that are given by applying discriminant analysis and utilizing the forward method
for the twenty-five variables.

5.3. Comparison with the Logit Analysis

Because logit analysis is a widely used technique for categorizing two groups, we have
compared the results obtained using the rough set methodology with the results obtained using
the logit model. The practical benefits of the logit methodology are that it does not require the
restrictive assumptions of the discriminant analysis.

52

Published by ePublications@bond, 2013 19

 International Journal of Banking and Finance, Vol. 10, Iss. 1 [2013], Art. 3

In the logit model, we applied a statistical forward stepwise selection procedure on the
selected variables. The logit model function resulting from our sample of companies is shown in
Table 9. The Wald test for each of the predictors is statistically significant. Additionally, the
logit-likelihood test is statistically significant.

Table 9: Logit model results

This table includes the details of applying the logit model. A statistical forward stepwise
selection procedure is employed.

Coefficient (β) Standard error Wald gl Sig. Exp(β)

R4 1.271 .602 4.462 1 .035 3.566

R6 4.671 1.851 6.370 1 .012 106.765
R8 -9.568 5.268 3.299 1 .069 .000
R21 12.630 5.589 5.106 1 .024 305642.138
Constant .002 .492 .000 1 .996 1.002
Z = 1.271 R4 + 4.671R6 – 9.568R8 + 12.630R21 + 0.002

Table 10 shows the results of the designation of healthy and failed companies. Based on
tables 6 and 10, it can be deduced that the rough set methodology is more useful than the logit
model over the sampl of companies considered.

Table 10: The result of the classification between healthy and failed companies (logit model).

Forecast group in which the company falls

Total
Failed Healthy
Failed 81,6% 18,4% 100%

Healthy 26,5% 73,5% 100%

Overall correct percentage 77,6%

53

[Link] 20
 Medina and Vazquez Cueto: Modeling credit risk

6. Conclusion

In this study we present an alternative methodology to the classic discriminant analysis and logit
model for determining the variable(s) that serve to explain the failure of a company to meet its
debt repayments as the first step in determining the probability of default (PD).

Basing our arguments on a sample of sound and bankrupt companies and on a set of 25
financial ratios that are potential explanatory factors for the defaults occurring in the sample, we
have shown that the rough set methodology can be a valid alternative to discriminant analysis
and to the logit model when there is a need to classify objects into two different classes. In
addition to obtaining acceptable percentages of correct classification using rough sets, there is no
need to assume any type of prior statistical behavior of the variables involved in the
classification, unlike discriminant analysis, which requires normality in the distributions and
equality in the variance-covariance matrices. Furthermore, the variables are included as they are
presented, with no need for any transformation. Among the more significant advantages of this
methodology is that it eliminates redundant information, and it expresses the dependencies
between the variables considered and the results of the classification through decision rules for
which the language is closer to that normally utilized by the experts.

___________________________________________________________

Authors Information: Reyes Samaniego Medina is the corresponding author and is with Pablo
de Olavide University, Department of Financial Economics and Accounting, Carretera de Utrera,
Km. 1 (41013) Seville, Spain. Tel.: +34 954 34 98 45. E-mail address: rsammed@[Link]. M.
José Vázquez Cueto, Departamento de Economía Aplicada III, Universidad de Sevilla (41018)
Sevilla, Spain (España).

54

Published by ePublications@bond, 2013 21

 International Journal of Banking and Finance, Vol. 10, Iss. 1 [2013], Art. 3

References

Altman, E.I. (1968). Financial ratios, discriminate analysis and the prediction of corporate
bankruptcy. Journal of [Link] 589-609.
Altman, E.I. (1993). Corporate financial distress and bankruptcy. John Wiley. New York.
Altman, E.I., Sabato, G. (2007). Modelling credit risk for smes: evidence from the U.S.
market. Abacus, Vol. 43, 3, 332-357
Ahn, B.S., Cho, S.S., Kim, C.Y. (2000).The integrated methodology of rough set theory and
artificial neural network for business failure prediction. Expert Systems with Applications,
18.
Basle Committee on Banking Supervision (2001).The internal ratings-based approach.
Consultive Document. Supporting Document to the New Basle Capital Accord, January.
Basel Committee On Banking Supervision (1999). Credit risk modelling: current practices
and applications. April.
Beaver, W. H. (1966). Financial ratios as predictors of failure. Journal of Accounting
Research 4, 71-127
Cabedo,J., Reverte, J.A., Tirado, J.M. (2004). Riesgo de crédito y recursos propios mínimos
en entidades financieras. Revista Europea de Dirección y Economía de la Empresa, 13, 2.
Chan. C., Grzymala-Busse, J. (1994).On the two local inductive algorithms: PRISM and
LEM2. Foundations of Computing and Decision Sciences, 19, 185–204.
Comité De Supervisión Bancaria De Basilea (2004). Convergencia internacional de medidas y
normas de capital. June.
Courtis, J.K. (1978). Modelling financial ratios: categories framework. Journal of Business,
Finance and Accounting, 5 (4).
Dambolena, I.G., Khoury, S.J. (1980). Ratio stability and corporate failure. The Journal of
Finance, 35.
Daubie, M.; Leveck, P.; Meskens, N. (2002). A comparison of the rough sets and recursive
partitioning induction approaches: an application to commercial loans. International
Transactions in Operational Research 9, 681-694.
Dimitras, A.I., Zanakis, S.H., Zopounidis, C. (1996). A survey of business failures with an
emphasis on prediction methods and industrial applications. European Journal of
Operational Research, 90.
Dimitras, A.I., Slowinski, R., Susmaga, R, Zopounidis, C. (1998). Business failure prediction
using rough sets. European Journal of Operational Research, 114.
Eisenbeis, R.A. (1977). Pitfalls in the application of discriminant analysis in business and
economics. Journal of Finance, 32, 875-900.
Fernandez, J.E. (2005). Corporate credit risk modeling: quantitative rating system and
probability of default estimation. SSRN: [Link]
Komorowski, Z. (1999). Rough sets: a tutoriain rough-fuzzy hybridization - a new trend in
decision makin. S.K. Pal and A. Skowron, Eds., 3-98, Springer-Verlag Singapore Pte
Ltd.

[Link] 22
 Medina and Vazquez Cueto: Modeling credit risk

Laitinen, E.K. (1991). Financial ratios and different failure processes. Journal of Business
Finance and Accounting,18/5.
Ohlson, J.A. (1980). Financial ratios and the probabilistic prediction of bankruptcy. Journal
of Accounting Research (Spring) 109-131.
Pawlak Z. (1982). Rough sets. Int. J. Computer and Information Sci., 11,341-356
Pawlak,Z.; Slowinski,R.(1994). Decision analysis using rough sets. International
Transactions in Operational Research 1, 107-114.
Rodriguez, M., Díaz, F. (2005). La teoría de los rough sets y la predicción del fracaso
empresarial. Diseño de un modelo para pymes. Revista de la Asociación Española de
Contabilidad y Administración de Empresas, 74.36-39
Segovia, M.J., Gil, J.A., Vilar, L., Heras, A.J. (2003). La metodología rough set frente al
análisis discriminante en la predicción de insolvencia en empresas aseguradoras. Anales
del Instituto de Actuarios Españoles, 9.
Trujillo, A., Martin, J.L. (2004). El rating y la fijación de precios en préstamos comerciales:
aplicación mediante un modelo logit. Revista Europea de Dirección y Economía de la
Empresa, 13. Nº. 2.
Trucharte, C., Marcelo, A. (2002). Un sistema de clasificación (rating) de acreditados.
Estabilidad Financiera, 2.
Wilson, T. (1997). Portfolio credit risk. Risk Magazine, Sept., 111-117.
Zavgren, C.V.(1983). The prediction of corporate failure: the state of the art. Journal of
Financial Literature 2, 1-37.
Zmijewski, M.E. (1984). Methodological issues related to the estimation of financial distress
prediction models. Studies on Current Econometric Issues in Accounting Research.

56

Published by ePublications@bond, 2013 23