Summary of Doctoral thesis: Factors affecting the credit rating of commercial banks - Research in developed and emerging economies

1

ABSTRACT
Credit rating agencies such as Fitch, Standard & Poor’s and Moody’s do not mention the difference in
impact of affecting factors on credit ratings of commercial banks between developed markets and emerging
markets. However, some researchers have pointed out there is difference in affecting of finacial ratios to credit
rating of commercial banks between developed markets and emerging markets.
Thesis’s objective is to investigate the difference in impacting of systematical factors such as national
risk, banking system risk in country where the banks locate and specific factors of banks as ownership structure,
bank size and finacial ratios on bank credit ratings between developed markets and emerging markets.
First, we use One way anova analysis and choosing indepent variables for ordered logit model method to
indentify factor affecting to bank credit ratings in developed markets and emerging markets.
The results of the thesis indicate that systematical factors have a stronger impact on bank’s credit ratings
in emerging markets than developed markets. Meanwhile, financial ratios have less impact on bank’s credit
ratings in emerging than developed markets. Moreover, the thesis shows the existence of the difference in
affecting of ownership structure to bank’s credit rating between developed markets and emerging markets.
Basing on the empirical results, we have some policy implications for central banks of emerging markets
to raise the bank’s credit ratings in their countries. We also imply some methods for commercial banks in
emerging markets to enchance their credit ratings.


2

CHAPTER 1: INTRODUCTION
1.1 Study background
Investors and depositors have a great concern about bank’s credit rating. However, credit rating agencies
do not mention details the way and level of impacting of affecting factors on bank’s credit ratings. Besides, some
empirical studies have indicated there is difference in level of impacting of financial ratios on the credit ratings
of commercial banks in developed markets and emerging markets.
So it is essential to identify the difference in impacting of affecting factors on bank’s credit rating in
developed markets and emerging markets.

1.2 Research problem
Due to the study background above, we carry out the study to solve the detail research objectives
following:
Identify the difference in impacting of systematical factors such as national risk, banking sector risk and
bank specific features such as bank size, form of ownership and financial ratios on bank’s credit rating in
developed and emerging markets.
1.3 Research questions
Firstly, is there difference in impacting of systematical factors as national risk and banking sector risk on
bank’s credit rating between developed markets and emerging markets.
Secondly, is there difference in affecting of specific features as bank scale, form of ownership and
financial ratios to bank’s credit rating between developed markets and emerging markets.
1.4 Research objectives
(1): Analyze and compare the impact of systematical factors such as national risk and banking sector risk on
bank’s credit rating between developed markets and emerging markets.
(2): Analyze and compare the impact of specific features includes bank size, form of ownership and
financial ratios on bank’s credit rating between developed markets and emerging markets.
1.5 Scope of this study
The thesis focus on analyzing bank’s credit ratings and affecting factors to bank’s credit ratings at
developed markets and emerging markets in the period from 2010 to 2015.
1.6 Academic and empirical meaning of this study
1.6.1 Academic meaning
First, the thesis identifies the affecting factors to bank’s credit ratings at developed markets and
emerging markets
Second, this study indicates the difference in the impact of systematical factors and specific feature of
commercial banks to their credit ratings between developed markets and emerging markets.
1.6.2 Empirical meaning
First, identifying affecting factors and the impact level of these factors on bank’s credit ratings helps
banking governors in emerging markets define the credit risk of commercial banks. Moreover, the empirical
results of the thesis supply more reference foundation for banking governors in emerging markets to issue
regulations for ensuring the safety of commercial banks and enhancing the bank’s credit ratings in these

countries.


3
Second, defining the affecting factors on bank’s credit ratings helps commercial banks to choose suitable
solutions to raise their credit ratings.
1.7 Contribution of this thesis
The contribution of this study to the empirical literatures relating to bank’s credit ratings is that this
study clarifies the difference in the impacts of national risk, banking sector risk, bank size, form of ownership
and bank’s financial ratios to bank’s credit rating between developed markets and emerging markets.
1.8 Structure of this thesis
Chapter 1 “Introduction”.
Chapter 2 “Bank’s credit ratings in developed markets and emerging markets”.
Chapter 3 “Methodologies”.
Chapter 4 “Empirical results and analysis”.
Chapter 5 “Conclusions and policy implications”.


4

CHAPTER 2: BANK CREDIT RATINGS IN DEVELOPED MARKETS AND
EMERGING MARKETS
2.1 The overview of bank’s credit ratings
2.1.1 Concepts of bank’s credit ratings
Bank’s credit ratings issued by credit rating agencies are ordinal measures that should not only reflect
the current positions of banks but also provide information about their future financial positions (Bellotti et al,
2011a).
2.1.2 Bank’s credit rating methodologies
2.1.2.1 The Uniform Financial Institutions Rating System - UFIRS
This system is adopted by the Federal Financial Institutions Examination Council in 1979. At first, this

system is applied in United State. After that this system is used by many countries due to the recommendation of
the Federal Reserve.
2.1.2.2 Bank’s credit rating methodologies of credit rating agencies
Fitch evaluates the bank’s credit rating through 2 phases:
 Phase 1: Accessing bank’s viability rating – VR bases on 5 factors: operating environments, bank size,
management capability, risk management and financial positions of commercial banks.
 Phase 2: Accessing bank’s final credit ratings by combining bank’s viability ratings and supports from
government and group.
The same as Fitch, Standard & Poor’s evaluates bank’s credit rating within 2 steps:
 Step 1: identify bank’s stand alone credit profile bases on 6 factors include: economic risk, industry risk of
country where banks locate; business position; capital and earnings; risk position; funding and liquidity.
 Step 2: evaluate bank’s final credit rating by combining bank’s stand alone credit profile and external
supports from government and group.
2.2 The affecting factors on bank’s credit rating
According to The Uniform Financial Institutions Rating System and bank’s credit rating methodologies
of credit rating agencies, we realize that bank’s credit ratings are affected by the following factors: economic
risk, industry risk of country where banks locate, external supports from government and group and some
specific features of banks.
2.2.1

The affecting of macro factors on bank’s credit ratings
Banking operations are very sensitive to macro factors’ varieties. Especially, changes in economic

policies or political systems have strong affects on bank’s credit ratings in these countries.
2.2.2 The affecting of government supports and group supports on bank’s credit ratings
Fitch (2104) supposes that government supports to belonging commercial banks help to improve these
banks’ credit ratings.
Moreover, supports of big and prestige groups have positive impacts on banks’ credit ratings. According
to Moody’s (1999), groups utilize their scale advantages, risk diversification capabilities and management
experience to help belonging commercial banks.

2.2.3 The affecting of specific features on banks’ credit ratings


5
According to Standard & Poor’s (2011a), specific features of commercial banks have impacts on bank’s
credit rating included: bank size and business position; asset quality; capital and earnings; funding and liquidity.
Credit rating agencies analyze these factors to identify bank’s stand alone credit profiles. After that, credit rating
agencies combine bank’s stand alone credit profiles, economic risks and supports from governments or groups to
determine bank’s credit rating.
2.3 Economic features and commercial bank characteristics in developed markets
2.3.1 Economic features in developed markets
First, developed countries have a high level of per capita GNP.
Second, developed countries have post-industrial economies.
Third, developed countries have a high standard of living.
2.3.2 Commercial bank characteristics in developed markets
First, banking system in developed countries has a high degree of competition.
Moreover, commercial banks in developed markets have a higher level of services diversification than
commercial banks in emerging markets.
Finally, the regulatory frameworks controlling the banking operations in developed countries are better
than emerging markets.
2.4 Economic features and commercial bank characteristics in emerging markets
2.4.1 Economic features in emerging markets
First, emerging markets are countries being in transition process from a closed and less developed
economy to an opened and developed economy.
Second, instability of financial system in emerging markets is an important feature discussed by a lot of
researchers.
Third, financial liberalization is taking place in emerging markets to over come the instability of
financial systems.
Final, GDP growth rates in emerging markets are usually higher than developed markets.
2.4.2 Bank characteristics in emerging markets

First, asset and loan growth rates of commercial banks in emerging markets are at high level.
Second, Suarez (2001) indicates that equity capital of commercial banks in emerging markets do not
present the financial capability of commercial banks as in developed markets.
Third, earning capability of commercial banks, presented by net profit/average total assets ratio, in
emerging markets is higher than commercial banks’ in developed markets.
Final, quality of financial information issued by commercial in emerging markets is not reliable (Vives,
2006). In these countries, issuing financial information of commercial banks has a lot of problems due to slow
intuitional reforms.
2.5 The impact of information asymmetry on bank’s credit rating in emerging markets
2.5.1 Concept of information asymmetry
Information asymmetry occurs when one party of a financial transaction have more sufficient
information than the other party. And this may lead to moral hazard or adverse selection.
2.5.2 The reasons cause the impaction of information asymmetry on bank’s credit rating in emerging markets


6
Information asymmetry between credit rating agencies and credit rated banks always occurs in credit
rating process. The reason make information asymmetry have a strong impact on bank’s credit rating in
emerging market may due to the essence of bank’s credit ratings and the quality of bank’s financial information
in these countries.
2.5.3 The impact of information asymmetry on bank’s credit ratings in emerging markets
Most of bank’s credit ratings in emerging markets are unsolicited rating. These credit ratings base almost
on public information of rated banks. So that the credit rating agencies can not access the credibility and
accuracy of the information especially when publishing information frameworks and accounting standards are
not strict in emerging markets. In this case, the credit rating agencies focus on evaluating the bank’s operation
environment and skip accessing the specific financial ratios of rated banks. So that the information asymmetry
causes the the difference in the impact of economic risks and specific features of commercial banks on bank’s
credit rating between developed markets and emerging markets.
2.6 Literature reviews
Empirical researches relating bank’s credit rating can be divided into 2 strands:

The first one presented by studies that search and try to identify the reliability of rating assignments.
The second strand is focused on empirical researches that investigate the prediction models for bank’s
credit ratings.
2.6.1

Reliability of rating assignments
Researches of Poon and Firth (2005), Poon et al (2009), Shen et al (2012).

2.6.2 Prediction models for bank’s credit ratings
2.6.2.1 The studies applied statistical techniques
Poon et al (1999), Matousek and Stewart (2009), Caporale et al (2012).
2.6.2.2 The studies applied artificial intelligent
Boyacioglu et al (2009), Ioannidis et al (2010), Bellotti et al (2011a, 2011b), Chen (2012).
2.7 Research gaps and thesis’ analyzing framework
2.7.1 Research gaps
We notify that the previous studies have not mentioned the differences in impact of some factors
include: economic risk, industry risk and bank’s ownership form on bank’s credit ratings between developed
markets and emerging markets. Moreover, the number of financial ratios used in these studies is limited.
2.7.2

Thesis analyzing framework
We have 2 groups of factor in our research model. The first one presents systematical factors such as

economic risk and industry risk of the countries where banks locate. The other group presents the specific
features of commercial banks including bank’s ownership form, bank size and financial ratios. We apply One
way – ANOVA analyzing on bank’s financial ratios and indepenent variables selection method for ordered logit
model to indentify the factors impacting bank’s credit ratings in developed markets and emerging markets.The
next step, we access the model’s reliability and test the model’s assumptions. Finnaly,we analyse the impact of
affecting factors and the difference in impact of these factor on bank’s credit ratings between developed markets
and emerging markets.

7

CHAPTER 3: METHODOLOGIES
3.1 Research model
3.1.1 Ordered logic model
The objectives of the thesis are identifying the different in impact of affecting factors on bank’s credit
rating between developed markets and emerging markets and interpreting the impact of these factors on bank’s
credit ratings. So that, we apply ordered logic model as our main analyzing model in this study. Because ordered
logic model is suitable for presenting the result of object classifying process into ordinary ratings (Greene,
2002). Otherwise, basing on the direction of variables’’ coefficients we can identify the impact directions of
corresponding factors on bank’s credit ratings. Meanwhile, we can not achieve these objectives with non linear
models such as neutral network or support vector machines. Moreover, we can use interacting variables in
ordered logic model to access the difference in impact of affecting factors on bank’s credit ratings between
developed markets and emerging markets.
The ordered logic model is presented as following:

y* is a dependent variable and unobserved. We can only observer
Y = 1 if y* ≤ 1
= 2 if 1 < y* ≤ µ1
= 3 if µ1 < y* ≤ µ2
…
= J if µj-1 < y*
µ1 , µ2 ,… µj-1 are thresholds calculated by the models.
β is coefficients presenting the impact of independent variables on the dependent variable.
Ɛ is a stochastic error term. Ɛ has a standard distribution, a 0 average value and variance equaling 1.
3.1.2 Definition and measurement of dependent variable
The dependent variable of the model is bank’s credit ratings issued by Fitch. The dependent variable is
given symbol yi and coded from 1 to 9.

3.1.3 Definition and measurement of independent variable
The independent variables of model are classified into 2 groups:


The first group presents the systematic factors affecting bank’s operation environment.



The other group have some subgroups presenting bank’s specific features such as form of ownership,
bank’ size and financial ratios.

3.2 The research data
The thesis’ research data is crossed-sectional data including bank’s credit ratings, bank’s financial ratios
and macro economic factors affecting bank’s operation environment.
The thesis’ research data is divided into 2 data sets. The first data set has 296 observations of bank’s
credit ratings and bank’s financial ratios in developed markets. The second data set has 282 observations of
bank’s credit ratings and bank’s financial ratios in emerging markets. The list of developed markets and


8
emerging markets is referent to World Economic Outlook 2014 (IMF, 2014). We apply systematical sample
selection method with 2 steps to select observations for these 2 data sets
The bank’s credit ratings are Fitch’s bank’s credit rating assignments from 2013 to 2015. The bank’s
financial ratios between 2010 and 2014 are provided by Bank scope.
3.3 Research hypothesis
Hypothesis 1 (H1): there are differences in impact of economic risk and industry risk of countries where
banks locate on bank’s credit ratings between developed markets and emerging markets.
Hypothesis 2 (H2):there are differences in impact of international financial group ownership on bank’s
credit ratings between developed markets and emerging markets.
Hyppothesis 3 (H3):there are differences in impact of government ownership on bank’s credit ratings

between developed markets and emerging markets.
Hyppothesis 4 (H4):there are differences in impact of bank’s total asset size on bank’s credit ratings
between developed markets and emerging markets.
Hyppothesis 5 (H5):there are differences in impact of financial ratios on bank’s credit ratings between
developed markets and emerging markets.
3.4 Analyzing data process of the thesis
The data analyzing process of the thesis to achieve the thesis’ objectives and answer the research
question of the thesis is described by the following diagram:
Diagram 3.1: Thesis’ analyzing data process

Apply One way Anova
analyzing on bank’s
financial ratios
Apply variable
selection process for
Ordered logic model

Identify the
affecting factors
on bank’s credit
rating

Access the
creditability and
assumptions of
the research
model

Merge 2 data
sets and add

interacting
variables
Access the
differences in impact
of affecting factors on
bank’s credit ratings

Source: Author’s inference from literature review
Step 1, to identify the differences in impact of affecting factors on credit ratings between developed
markets and emerging markets we must indicate the specific factors affecting on bank’s credit ratings. To
achieve this purpose, we separately apply one way anova analyzing process and variable selection process for
ordered legit model on the data sets of developed markets and emerging markets.
Step 2, we apply BIC ratios (Bayesian information criteria) to compare the crediability of model
inferring from the above proccess to random models. Besides that, we also test the model’s assumptions such as
muticollinearity, heteroskedasticity and missing essential variables.


9
Step 3, to achieve the first and the second research objectives, we merge the data set of commercial
banks in developed markets and the data set of commercial banks in emerging markets. We also add a dummy
variable named Emer. This proxy takes 1 value if banks locate in developed markets and 0 value in case banks
locate in emerging markets. After that, we apply interact proxies between Emer variable and each variable
presents systematical factors and bank’s specific factors. Finally, we regress again the Ordered logit model with
all variables indentified from previous analyzing steps and these interact proxies. In case the cofficients of the
interact variables are significant that mean there are difference in impact of corresponding variables on bank’s
credit ratings between developed markets and emerging markets.


10

CHAPTER 4: EMPIRICAL RESULTS AND ANALYSIS
4.1 One way Anova analyzing bank’s financial ratios
We separately apply one way Anova analyzing for each data set in the thesis.
Table 4.1: Average values of bank’s financial ratios classified by bank’s credit ratings
in emerging markets
Bank’s
credit
rating

Number
of Obs

Standard
deviation

Average

Min

Max

LnAss:Logarit bank’s total assets

Average

Standard
deviation

Min

Max

AssGrow:Average grow rate of bank’s assets over 3 years

B

70

8.1763

1.3779

5.4189

10.8731

0.2477

0.2719

-0.0512

1.2881

BB

64

8.9984

1.3439

6.3988

13.1649

0.1615

0.1277

-0.0920

0.5908

BBB

116

9.9381

1.6938

5.8197

13.4052

0.1601

0.1167

-0.1883

0.6149

A

32

11.1366

1.9862

6.3651

14.9549

0.1009

0.0645

-0.0326

0.2870

Total

282

9.4235

1.8310

5.4189

14.9549

0.1755

0.1729

-0.1883

1.2881

CreGrow:Average grow rate of loan over 3 years

LoanLoss_Ln:Overdue loan/total loan

B

70

0.2679

0.3014

-0.1466

1.4153

9.2476

14.2697

0.0600

89.9940

BB

64

0.1717

0.1271

-0.1175

0.6352

6.1456

7.2884

0.2160

37.2630

BBB

116

0.3186

1.5222

-0.1357

16.5153

4.8128

6.4004

0.0000

33.9010

A
Total

32
282

0.1160
0.2497

0.0727
0.9901

-0.0319
-0.1466

0.3488
16.5153

3.9422
6.1173

4.1475
9.1790

0.4840
0.0000

19.0650
89.9940

LoanLoss_Equ:Overdue loan/Equity

B

70

9.0775

13.1440

0.5800

LoanPro_Loan:Loan provision/Average total loan

86.0830

77.8620

172.4143

1.3510

985.2000

BB

64

5.9166

7.1224

0.4070

40.8250

33.3714

50.3338

3.2010

335.6260

BBB

116

5.1344

10.4250

0.0000

100.0000

30.5691

62.7480

0.0000

592.5900

A
Total

32
282

4.6897
6.2403

6.3130
10.2703

0.4030
0.0000

27.4330
100.0000

23.6996
42.1649

17.3031
99.7057

3.8960
0.0000

62.9140
985.2000

Equ_Ass:Equity/Total assets

Equ_Loan:Equity/Net loan

B

70

13.0284

7.5463

4.8960

54.4000

12.5905

9.9021

-35.8200

46.3100

BB

64

11.4510

4.6231

4.6550

29.9590

11.3953

4.4502

4.5650

27.6880

BBB

116

11.4910

7.0204

2.2510

45.6580

12.1437

10.6638

2.7870

87.1290

A

32

10.0933

6.3050

1.2650

36.4300

9.7345

7.1818

1.3450

43.6430

Total

282

11.7050

6.6405

1.2650

54.4000

11.8114

9.0263

-35.8200

87.1290

Equ_ShortCap:Equity/Short term capital

Equ_Debt:Equity/Total debts

B

70

27.2697

22.0246

7.4000

117.8660

19.8259

17.1355

5.6050

122.1600

BB

64

22.0198

10.9775

9.9440

71.3420

16.0915

10.6321

5.7420

76.8130

BBB

116

28.5163

67.7129

3.0640

717.1900

30.4765

63.2117

3.1470

514.0000

A

32

18.6040

15.5689

1.6960

76.2060

19.0323

26.8788

5.2040

153.1840

Total

282

25.6077

45.4093

1.6960

717.1900

23.2694

43.0184

3.1470

514.0000

IntIn_Loan:Interest income/Average total loan

IntIn_Ass:Interest income/Total earning interest assets

B

70

19.1616

17.7580

-41.8370

86.6200

15.5154

10.0329

3.8030

67.3100

BB

64

16.9513

14.6495

4.9450

116.6560

13.7887

5.9801

5.1230

39.7260

BBB

116

37.3368

90.0853

1.5400

520.3250

17.7913

39.6251

2.4600

413.7670

A

32

18.4976

32.5349

5.3740

191.8260

11.9574

10.6335

1.3040

64.2230

Total

282

26.0610

60.4427

-41.8370

520.3250

15.6560

26.3065

1.3040

413.7670


11
Bank’s
credit
rating

Number
of Obs

Standard
deviation

Average

Min

Max

Standard
deviation

Average

IntEx_Cap:Interest expense/Total interest bearing capital

B

Min

Max

NIM:Net interest margin

70

14.8549

7.8491

3.5600

54.4100

15.9974

8.3957

4.7900

52.1300

BB

64

11.2155

9.3518

2.9000

70.8000

11.3424

9.2202

0.6000

66.7300

BBB

116

9.8933

7.3001

3.2300

69.0200

9.9942

7.3708

-5.1700

68.0100

A

32

6.8878

2.4154

3.6800

13.2700

6.9241

2.7880

0.0000

13.7700

Total

282

11.0839

7.9578

2.9000

70.8000

11.4420

8.2354

-5.1700

68.0100

NetIntIn_Ass:Net interest income/Average total assets

OthIn_Ass:Other operation income/Average total assets

B

70

13.8780

7.6415

3.7000

61.9300

6.0989

3.3288

1.6300

22.4300

BB

64

12.2491

18.7090

3.5600

139.9600

4.3123

2.4593

1.0200

10.7000

BBB

116

9.1547

6.9236

3.1200

67.9000

4.3609

2.7360

0.1900

16.5400

A

32

6.2269

1.9516

3.4800

12.4600

2.4841

0.6486

1.1000

3.7800

Total

282

10.6972

10.9169

3.1200

139.9600

4.5683

2.8806

0.1900

22.4300

NonIntEx_Ass: Non interest expense/Average total assets

B

ROAA:Return on Average total assets

70

6.2806

3.4380

1.3900

16.2000

7.1035

5.7961

0.6933

40.5170

BB

64

4.4777

2.8568

1.2300

14.4600

5.6268

6.8561

1.1943

52.5063

BBB

116

4.7868

3.2113

0.0000

16.6100

4.7438

5.8730

0.7413

61.2250

A
Total

32
282

2.5263
4.8309

0.9753
3.1961

0.0000
0.0000

4.9600
16.6100

3.3689
5.3739

1.5905
5.8792

0.9227
0.6933

8.9373
61.2250

ROAE:Return on Average equity

B

Exp_Int:Total expense/Total Income

70

6.0141

4.7440

0.7347

32.3683

6.1262

5.3752

-0.4880

32.5820

BB

64

4.8491

5.1862

1.1607

39.2023

4.8953

5.0337

1.0670

36.9770

BBB

116

4.3054

5.5214

0.4237

58.0370

4.3207

5.4613

0.5420

58.1850

A
Total

32
282

3.0632
4.7120

1.2759
5.0017

0.9083
0.4237

6.9520
58.0370

3.0702
4.7574

1.2694
5.1023

1.0000
-0.4880

7.1600
58.1850

NetLoan_Ass:Net loan/Total asset

NetLoan_ShortCap:Net loan/ Total shorterm capital

B

70

3.5223

5.8696

-0.6480

35.7410

6.4434

4.7629

1.0610

23.1070

BB

64

1.9295

1.7895

-1.7550

9.8840

4.8889

5.9204

0.9370

44.4720

BBB

116

1.5113

1.2835

-2.7600

7.3990

3.9355

4.3881

0.1740

40.4890

A

32

1.3356

0.8671

-0.2360

2.8730

2.6124

1.5312

0.6230

6.1850

Total

282

2.0855

3.2642

-2.7600

35.7410

4.6243

4.8026

0.1740

44.4720

NetLoan_Debt:Net loan/Total debts

LiAss_ShortCap:Liquidity assets/Total shortterm capital

B

70

7.0169

5.4266

0.0300

29.0060

1.6594

4.1657

-9.4630

23.8630

BB

64

5.0173

5.8099

0.9400

43.5840

1.2533

1.7253

-6.5410

7.7230

BBB

116

4.0887

3.9986

0.2640

33.3720

1.3015

2.1374

-10.6100

13.1910

A
Total

32
282

2.5883
4.8560

1.4021
4.8536

0.4480
0.0300

5.6190
43.5840

1.3420
1.3840

0.5523
2.6187

0.2750
-10.6100

2.7100
23.8630

LiAss_Debt:Liquidity asset/Total debts

B

70

13.8842

12.0269

-12.1550

58.9220

BB

64

11.4613

12.6789

-74.5380

30.8730

BBB

116

13.2433

9.6325

-31.4700

44.0260

A
Total

32
282

15.4400
13.2472

5.7954
10.7057

4.2120
-74.5380

28.6060
58.9220

Source:Author’s caculating from thesis’ data sets


12
Table 4.2: The result of homogeneity of variance test and One way Anova analyzing on bank’s
financial ratios in emerging markets

Variance
LnAss

AssGrow
CreGrow
LoanLoss_Ln
LoanLoss_Equ
LoanPro_Loan
Equ_Ass
Equ_Loan
Equ_ShortCap
Equ_Debt
IntIn_Loan
IntIn_Ass
IntEx_Cap

P-value of
P-value of
P-value of
homogeneity
One way
homogeneity
Variance
of variance
anova
of variance
test
analyzing
test
0.0110
0.0000 NIM
0.0300
0.0000

0.0000 NetIntIn_Ass
0.0070
0.4010
0.6690 OthIn_Ass
0.0000
0.0000
0.0060 NonIntEx_Ass
0.0000
0.0220
0.0570 ROAA
0.1300
0.0000
0.0060 ROAE
0.2360
0.4920
0.1810 Exp_Int
0.1360
0.1580
0.4760 NetLoan_Ass
0.0000
0.2060
0.6320 NetLoan_ShortCap
0.0870
0.0000
0.0210 NetLoan_Debt
0.0010
0.0000
0.0740 LiAss_ShortCap
0.0010
0.1470

0.6340 LiAss_Debt
0.1210
0.0540
0.0000
Source: Author’s caculating from thesis’ data sets

P-value of
One way
anova
analyzing
0.0000
0.0020
0.0000
0.0000
0.0100
0.0280
0.0230
0.0000
0.0000
0.0000
0.7890
0.3390

According to the average values of bank’s financial ratios in each bank’s credit ratings and the P-value
of One way Anova analyzing showed

in table 4.1 and 4.2, we can conclude that LnAss, AssGrow,

LoanLoss_Ln, LoanLoss_Equ, LoanPro_Loan, Equ_Debt, IntIn_Loan, IntEx_Cap, NIM, NetIntIn_Ass,
OthIn_Ass, NonIntEx_Ass, NetLoan_Ass, NetLoan_ShortCap và NetLoan_Debt have different average values

in each bank’s credit ratings.
Meanwhile, the P-values of

homogeniety of variance tests on CreGrow, Equ_Ass, Equ_Loan,

Equ_ShortCap, IntIn_Ass, ROAA, ROAE, Exp_Int and LiAss_Debt are not significant (>10%) so the
assumption about homogeneity of variance on these variables are invalid. So that we take a nonparameter
Kruskal – Wallis test on these proxies.
Table 4.3: The results of Krukal – Wallis test on bank’s financial ratios in emerging markets
Variable
CreGrow
Equ_Ass
Equ_Loan
Equ_ShortCap
IntIn_Ass

P-value of Kruskal –
P-value of Kruskal –
Variable
Wallis test
Wallis test
0.0050 ROAA
0.0000
0.0360 ROAE
0.0000
0.0250 Exp_Int
0.0000
0.0040 LiAss_Debt
0.2500
0.0220

Source: Author’s caculating from thesis’ data sets

According to the results of One way anova analyzing and Kruskal – Wallis test, we can conclude that
LnAss,

AssGrow,

CreGrow,

LoanLoss_Ln,

LoanLoss_Equ,

LoanPro_Loan,

Equ_Ass,

Equ_Loan,

Equ_ShortCap, Equ_Debt, IntIn_Loan, IntIn_Ass, IntEx_Cap, NIM, NetIntIn_Ass, OthIn_Ass, NonIntEx_Ass,
ROAA, ROAE, Exp_Int, NetLoan_Ass, NetLoan_ShortCap and NetLoan_Debt have different average values in


13
each bank’s credit ratings in emerging markets. In contrast, the average values of LiAss_ShortCap and
LiAss_Debt are not different in each bank’s credit ratings.
We apply one way anova analyzing on bank’s financial ratios in developed markets.
Table 4.4: Average values of bank’s financial ratios classified by bank’s credit ratings
in developed markets
Bank’s

credit
rating

Number of
Obs

Standard
deviation

Average

Min

Max

Average

Standard
deviation

Min

Max

AssGrow:Average grow rate of bank’s assets
over 3 years

LnAss:Logarit bank’s total assets

B

8

11.4577

.6713

10.0440

12.3110

0.0017

0.0920

-0.0945

0.1964

BB

25

10.7804

1.5263

5.2490

12.8230

0.0339

0.1207

-0.1151

0.3602

BBB

52

10.4119

1.7275

6.6830

13.8380

0.0218

0.0884

-0.1968

0.3364

A

152

11.4706

1.7368

7.3850

14.7240

0.0143

0.1047

-0.4042

0.5461

AA

48

11.8124

1.6266

7.7130

14.2390

0.0557

0.0472

-0.0594

0.1627

AAA

11

11.9109

1.3402

10.1530

14.4920

0.0322

0.0577

-0.0453

0.1623

Total

296

11.2977

1.7253

5.2490

14.7240

0.0243

0.0951

-0.4042

0.5461

CreGrow:Average grow rate of loan
over 3 years

LoanLoss_Ln:Overdue loan/total loan

B

8

0.0098

0.1309

-0.1826

0.2727

24.7640

13.5608

0.8800

44.8630

BB

25

-0.0093

0.1130

-0.1708

0.2753

13.2698

8.8345

0.2180

37.9730

BBB

52

0.0361

0.1170

-0.1491

0.5168

6.5633

8.5192

0.0410

44.6330

A

152

0.0443

0.1859

-0.4088

1.4227

3.6221

4.6317

0.0000

44.6490

AA

48

0.0558

0.0648

-0.0759

0.3202

1.7663

1.7369

0.0090

8.1780

AAA

11

0.0800

0.1715

-0.0318

0.5829

1.0203

1.6900

0.0000

5.8160

Total

296

0.0406

0.1534

-0.4088

1.4227

5.1274

7.4089

0.0000

44.8630

LoanLoss_Equ:Overdue loan/Equity

LoanPro_Loan:Loan provision/Average total loan

B

8

306.8720

186.7417

28.7310

690.0910

2.8763

1.8082

0.0100

5.9100

BB

25

130.6217

89.4631

4.7470

381.3450

1.6124

1.1304

-0.3100

4.5700

BBB

52

59.6979

86.3918

0.2420

414.3900

0.9592

1.6735

-0.2900

10.6800

A

152

37.0226

55.7724

0.0000

464.9560

0.4932

1.0274

-0.4900

10.8700

AA

48

23.2292

63.8459

0.0400

433.0000

0.3258

0.3950

-0.0700

1.8300

AAA

11

73.3240

232.5974

0.0000

774.5700

0.0900

0.2509

-0.1800

0.6500

Total

296

55.3169

96.9361

0.0000

774.5700

0.6919

1.2257

-0.4900

10.8700

Equ_Ass:Equity/Total assets

Equ_Loan:Equity/Net loan

B

8

5.8810

3.1806

2.1000

11.3540

9.2836

4.4860

3.0860

16.1920

BB

25

6.6860

3.7176

0.2520

16.6180

12.9957

10.5310

0.6050

53.5830

BBB

52

8.4437

3.8266

1.4640

23.1450

19.7328

37.4397

3.2320

276.9500

A

152

7.1024

3.4055

1.1450

20.4280

16.7668

22.8025

1.9740

198.1360

AA

48

8.0283

2.9046

0.0060

16.0840

20.1603

29.4661

0.0090

151.4190

AAA

11

8.3670

6.5926

0.6530

20.1630

17.4925

15.9759

0.7620

57.3420

Total

296

7.4670

3.6177

0.0060

23.1450

17.3444

25.8997

0.0090

276.9500


14
Bank’s
credit
rating

Number of
Obs

Average

Standard
deviation

Min

Max

Average

Equ_ShortCap:Equity/Short term capital

Standard
deviation

Min

Max

Equ_Debt:Equity/Total debts

B

8

7.1263

3.6462

2.3740

13.6510

6.3853

3.6620

2.1450

12.8410

BB

25

10.2912

6.1303

0.2880

29.5010

7.4736

4.4889

0.2610

19.9290

BBB

52

12.6163

5.6406

1.7230

33.6150

9.5709

4.9256

1.4860

30.1150

A

152

14.0868

19.5748

2.0100

171.4090

7.9145

4.1861

1.1620

26.2490

AA

48

12.0495

8.4045

0.0080

63.6210

8.9642

3.5694

0.0060

19.7250

AAA

11

23.8286

21.0270

5.8390

60.7440

9.7279

8.2934

0.6570

25.2560

Total

296

13.3514

15.4219

0.0080

171.4090

8.3645

4.4883

0.0060

30.1150

IntIn_Loan:Interest income/Average total loan

IntIn_Ass:Interest income/Total earning interest assets

B

8

4.1388

1.2694

2.5400

6.3800

3.9250

1.4714

1.5800

5.8500

BB

25

3.6928

2.5023

1.5200

11.8800

3.9576

2.5821

0.8700

11.6800

BBB

52

3.7648

1.5208

0.0900

9.3700

3.5096

1.3593

1.0600

8.8300

A

152

3.9412

1.5957

0.7200

10.9100

3.4638

1.5443

0.1900

9.9400

AA

48

3.8719

1.4801

1.7300

7.4400

3.3244

1.4050

1.2300

6.3300

AAA

11

5.2309

6.7548

0.0000

24.6100

3.0355

1.2925

1.0800

4.5000

Total

296

3.9313

2.0545

0.0000

24.6100

3.4875

1.5942

0.1900

11.6800

IntEx_Cap:Interest expense/Total interest
bearing capital

NIM:Net interest margin

B

8

2.1900

0.4121

1.7300

2.9500

1.7205

1.1500

-0.1980

3.3640

BB

25

2.4952

2.2075

0.7400

11.6300

1.6064

1.1379

0.1360

5.2070

BBB

52

1.8075

0.9530

0.2800

4.5900

1.8282

0.9111

0.3460

4.6810

A

152

1.9437

1.3209

0.0300

7.3800

1.6478

1.3925

-0.0370

10.3710

AA

48

1.5529

1.1935

0.1500

4.8000

1.8591

0.7391

0.7560

3.7010

AAA

11

1.9327

1.4846

0.3300

4.4600

1.2684

1.2307

0.3440

3.5380

Total

296

1.9092

1.3441

0.0300

11.6300

1.6981

1.1946

-0.1980

10.3710

NetIntIn_Ass:Net interest income/Average total assets

OthIn_Ass:Other operation income/Average
total assets

B

8

1.5258

1.0146

-0.1950

2.9270

0.6049

0.4338

-0.0780

1.2280

BB

25

1.4659

1.0459

0.1340

4.5890

1.2422

0.8313

-0.2310

3.4960

BBB

52

1.7070

0.8400

0.3420

4.3510

0.7673

0.6303

-1.3640

2.7220

A

152

1.4897

1.1239

-0.0310

8.2190

1.2219

2.3249

-0.1810

21.1570

AA

48

1.6610

0.6182

0.7230

3.0580

1.1403

0.9118

0.0070

4.1040

AAA

11

1.2160

1.1900

0.3390

3.3870

0.2618

0.2869

-0.0980

0.8160

Total

296

1.5444

1.0019

-0.1950

8.2190

1.0782

1.7586

-1.3640

21.1570

NonIntEx_Ass: Non interest expense/Average
total assets

ROAA:Return on Average total assets

B

8

3.7551

1.7927

0.0520

6.1600

0.3375

2.4722

-2.7960

4.4290

BB

25

2.7567

1.4319

0.1470

6.3080

-0.2208

1.0151

-3.4080

1.2870

BBB

52

2.1230

1.9067

0.1950

13.9980

0.1170

1.8315

-11.2020

1.8410

A

152

2.0315

2.1603

0.0160

16.3690

0.4275

1.1498

-9.8850

6.6490

AA

48

1.7656

0.9176

0.1470

4.6780

0.7194

0.7254

-3.2490

2.0640

AAA

11

0.4705

0.4257

-0.0390

1.2000

1.0657

0.9620

0.0120

2.4610

Total

296

2.0543

1.9050

-0.0390

16.3690

0.3868

1.2965

-11.2020

6.6490


15
Bank’s
credit
rating

Number of
Obs

Standard
deviation

Average

Min

Max

ROAE:Return on Average equity

Average

Standard
deviation

Min

Max

Exp_Int:Total expense/Total Income

B

8

-0.3178

61.2013

-88.0050

81.8900

74.4116

35.5491

15.1590

133.5020

BB

25

-9.7504

34.8778

-150.1230

28.1520

66.1058

18.6566

24.3430

102.1420

BBB

52

2.6575

17.4151

-68.9970

23.8510

57.2777

20.5838

14.7550

145.2210

A

152

4.9554

11.9207

-74.3900

57.2300

64.3338

24.0560

9.4820

200.0000

AA

48

15.9109

46.3791

-53.1800

321.7950

54.1946

15.2158

14.5110

101.8980

AAA

11

51.0976

132.3486

0.0910

449.3400

29.6526

16.8780

7.4290

62.7300

Total

296

6.6585

36.9528

-150.1230

449.3400

60.5832

23.0990

7.4290

200.0000

NetLoan_Ass:Net loan/Total asset

NetLoan_ShortCap:Net loan/ Total shorterm capital

B

8

63.1024

7.9104

46.6140

70.1220

77.4719

11.1113

60.0990

95.6470

BB

25

57.7573

14.9626

4.0300

73.1660

87.2260

24.5990

4.6560

126.1080

BBB

52

63.8140

20.3189

2.5610

91.2770

95.8951

36.5762

3.0680

193.1080

A

152

57.4293

21.0607

5.1490

92.0240

98.7197

66.7776

9.6570

488.7750

AA

48

62.4362

21.3904

5.5320

97.0130

94.5649

48.3778

6.8200

283.3760

AAA

11

61.8165

25.8793

6.7990

94.1260

209.4969

203.4991

38.1260

766.4160

Total

296

59.7069

20.5274

2.5610

97.0130

100.1214

69.4688

3.0680

766.4160

LiAss_ShortCap:Liquidity assets/Total short-term
capital

NetLoan_Debt:Net loan/Total debts

B

8

69.6515

9.0654

51.4010

81.3050

10.0890

7.1723

3.7010

24.2990

BB

25

66.5232

17.6029

4.1020

84.0970

15.3523

19.1648

2.5500

89.0300

BBB

52

73.8234

23.1714

2.6510

104.0820

25.3165

30.2986

1.4260

159.0970

A

152

66.5928

23.9872

7.4470

147.4430

38.4403

42.6849

0.3020

391.6890

AA

48

72.3578

24.9982

6.5170

133.5730

23.0709

15.9403

0.8560

61.5100

AAA

11

70.0112

31.3312

7.2010

112.6920

67.9500

104.4219

7.7610

377.7170

Total

296

69.0017

23.5972

2.6510

147.4430

32.0229

40.7363

0.3020

391.6890

LiAss_Debt:Liquidity asset/Total debts

B

8

8.8375

5.9283

3.1700

21.4620

BB

25

12.5558

17.1223

1.7010

78.4480

BBB

52

17.4041

16.4145

1.3710

86.9570

A

152

24.0424

19.1224

0.0630

111.3150

AA

48

18.5450

12.3313

0.4960

54.2490

AAA

11

22.5233

20.4505

2.5690

71.3440

Total

296

20.5472

17.7427

0.0630

111.3150

Source: Author’s caculating from thesis’ data sets


16
Table 4.5: The result of homogeneity of variance test and One way Anova analyzing on bank’s

Variance
LnAss

financial ratios in developed markets
P-value of
P-value of
P-value of
homogeneity
One way
homogeneity
Variance
of variance
anova
of variance
test
analyzing

test
0.0060
0.0000 NIM
0.4900

P-value of
One way
anova
analyzing
0.6520

AssGrow

0.0250

0.1700 NetIntIn_Ass

0.2990

0.5800

CreGrow

0.3730

0.5100 OthIn_Ass

0.2980

0.3200

LoanLoss_Ln

0.0000

0.0000 NonIntEx_Ass

0.2850

0.0020

LoanLoss_Equ

0.0000

0.0000 ROAA

0.0020

0.0140

LoanPro_Loan

0.0000

0.0000 ROAE

0.0000

0.0000

Equ_Ass

0.0010

0.0770 Exp_Int

0.4650

0.0000

Equ_Loan

0.6030

0.7700 NetLoan_Ass

0.0060

0.0750

Equ_ShortCap

0.0210

0.1460 NetLoan_ShortCap

0.0000

0.0000

Equ_Debt

0.0000

0.0830 NetLoan_Debt

0.0190

0.4150

IntIn_Loan

0.0000

0.3930 LiAss_ShortCap

0.0000

0.0000

IntIn_Ass

0.3820

0.5230 LiAss_Debt

0.0190

0.0040

IntEx_Cap

0.0310
0.6520
Source: Author’s caculating from thesis’ data sets

The same as the above analyzing, basing on average values of bank’s financial ratios in each bank’s
credit ratings and the p-value of one way anova analyzing presented in table 4.4 and 4.5, we can conclude that
LnAss, LoanLoss_Ln, LoanLoss_Equ, LoanPro_Loan, Equ_Ass, Equ_Debt, ROAA, ROAE, NetLoan_Ass,
NetLoan_ShortCap, LiAss_ShortCap và LiAss_Debt have different average values in each bank’s credit ratings.
Meanwhile, the P-values of homogeniety of variance tests on CreGrow, Equ_Loan, IntIn_Ass, NIM,
NetIntIn_Ass, OthIn_Ass, NonIntEx_Ass and Exp_Int are not significant (>10%) so the assumption about
homogeneity of variance on these variables are invalid. So that we take a nonparameter Kruskal – Wallis test on
these proxies.
Table 4.6: The results of Krukal – Wallis test on bank’s financial ratios in emerging markets
Variable
CreGrow
Equ_Loan
IntIn_Ass
NIM

P-value of Kruskal –
P-value of Kruskal –
Variable
Wallis test
Wallis test
0.0060 NetIntIn_Ass
0.0250
0.3330 OthIn_Ass

0.0010
0.8420 NonIntEx_Ass
0.0000
0.0140 Exp_Int
0.0000
Source: Author’s caculating from thesis’ data sets

The results of Krukal – Wallis test presented in table 4.6 show that CreGrow, NIM, NetIntIn_Ass,
OthIn_Ass, NonIntEx_Ass và Exp_Int have different average values in each bank’s credit ratings.
Basing on the results of One way anova analyzing and Kruskal – Wallis test, we can conclude that
LnAss, CreGrow, LoanLoss_Ln, LoanLoss_Equ, LoanPro_Loan, Equ_Ass, Equ_Debt, NIM, NetIntIn_Ass,


17
OthIn_Ass, NonIntEx_Ass, ROAA, ROAE, Exp_Int, NetLoan_Ass, NetLoan_ShortCap, LiAss_ShortCap and
LiAss_Debt have different average values in each bank’s credit ratings in developed markets. In contrast, the
average values of AssGrow, Equ_Loan, Equ_ShortCap, IntIn_Loan, IntIn_Ass, IntEx_Cap và NetLoan_Debt
are not different in each bank’s credit ratings.
4.2 The result of variable selection for ordered logic model and accessing model’s creditability
4.2.1 The result of variable selection for ordered logic model
To select varaibles for ordered logit model and avoid model’s over-fitting, we pick up 5 sub data sets for
the initial data set. Each sub data sets’ number of observations equal 80% of the number of observations in the
initial data set. We again resgress the model on each sub data sets. We choose the sub data sets from the initial
data set by applying systematical sample selection methold with 5 steps.
Tabble 4.7: The ordered logit models on 5 sub data sets and the initial data set
in emerging markets

Sub data set 1
Variable
Country_rating

Bicra
Government
Group
LnAss
AssGrow
CreGrow
LoanLoss_Ln
Equ_Debt
IntEx_Cap
NIM
NetIntIn_Ass
OthIn_Ass
NonIntEx_Ass
ROAE
Exp_Int
NetLoan_Ass

Coefficient
2.2670
0.6750
0.9698
4.0543
0.6060
-8.0183
3.1831
-0.1034
0.0159
0.2371
-0.1957
-0.0301

-0.5666
0.4862
-0.9325
0.9262
-0.1387

Number of obs =
225
LR chi2(17) = 289.9800
Prob > chi2 = 0.0000
Pseudo R2
= 0.5010
Log likelihood = -144.4212

Sub data set 2
P-value
0.0000
0.0020
0.0190
0.0000
0.0000
0.0090
0.1860
0.0000
0.0010
0.1480
0.2040
0.1120
0.0080
0.0140

0.0310
0.0230
0.1900

Variable

Coefficient

Country_rating
1.8320
Bicra
0.7465
Government
0.9815
Group
4.0830
LnAss
0.5704
AssGrow
-5.3294
CreGrow
0.5879
LoanPro_Loan
-0.0142
Equ_Loan
-0.0859
Equ_ShortCap
-0.0129
Equ_Debt
0.0191

IntIn_Ass
0.0181
NetIntIn_Ass
-0.0149
OthIn_Ass
-0.2888
NonIntEx_Ass
0.1787
ROAE
0.1182
NetLoan_ShortCap
-0.1072
LiAss_Debt
-0.0215
Number of obs =
225
LR chi2(18) = 284.4600
Prob > chi2 = 0.0000
Pseudo R2
= 0.4890
Log likelihood = -148.6402

P-value
0.0000
0.0010
0.0270
0.0000
0.0000
0.0040
0.2340

0.0010
0.1530
0.2210
0.0550
0.1300
0.3780
0.1370
0.2660
0.2480
0.2510
0.2710


18

Sub data set 3
Variable

Coefficient

Country_rating
Bicra
Government
Group
LnAss
AssGrow
LoanLoss_Ln
LoanLoss_Equ
Equ_ShortCap
Equ_Debt

NIM
OthIn_Ass
NonIntEx_Ass
ROAA
NetLoan_Ass
NetLoan_Debt

2.0111
0.8798
1.2492
4.0705
0.5726
-6.1673
-0.1178
0.0214
-0.0047
0.0179
-0.0897
-0.3846
0.2826
0.1410
0.0576
-0.1013
Number of obs =
226
LR chi2(16) = 295.7900
Prob > chi2 = 0.0000
Pseudo R2
= 0.5081
Log likelihood = -143.1579

Sub data set 4
P-value
0.0000
0.0000
0.0040
0.0000
0.0000
0.0010
0.0020
0.5760
0.4570
0.0000
0.2460
0.0280
0.0820
0.1720
0.5920
0.2460

Variable

2.1079
Country_rating
0.8533
Bicra
1.1090
Government
4.4270
Group

0.5170
LnAss
-6.8707
AssGrow
0.7967
CreGrow
-0.0967
LoanLoss_Ln
-0.0167
Equ_ShortCap
0.0179
IntIn_Loan
-0.1847
NetIntIn_Ass
-0.4910
OthIn_Ass
0.4132
NonIntEx_Ass
0.6560
ROAA
-0.4994
Exp_Int
0.1545
LiAss_ShortCap
Number of obs =
226
LR chi2(16) = 304.2700
Prob > chi2 = 0.0000
Pseudo R2
= 0.5154

Log likelihood = -143.0191

Sub data set 5
Variable
Country_rating
Bicra
Government
Group
LnAss
AssGrow
CreGrow
LoanLoss_Ln
Equ_ShortCap
Equ_Debt
IntEx_Cap
OthIn_Ass
NonIntEx_Ass
ROAE

Coefficient
2.0842
0.8235
1.1284
3.8271
0.4918
-3.5649
0.6086
-0.0666
-0.0127
0.0152

-0.0527
-0.3986
0.2538
0.4154

Coefficient

P-value
0.0000
0.0000
0.0120
0.0000
0.0000
0.0000
0.1030
0.0000
0.1110
0.0000
0.2900
0.0150
0.0170
0.0900
0.0910
0.1260

Initial data set
P-value
0.0000
0.0000
0.0090

0.0000
0.0000
0.0320
0.1640
0.0050
0.1700
0.0020
0.4460
0.0360
0.0830
0.1770

Variable
Country_rating
Bicra
Government
Group
LnAss
AssGrow
LoanLoss_Ln
Equ_Debt
OthIn_Ass

Coefficient
1.8164
0.8881
1.1377
3.9542
0.5979
-4.5942

-0.0832
0.0157
-0.1744

P-value
0.0000
0.0000
0.0020
0.0000
0.0000
0.0010
0.0000
0.0000
0.0010


19
-0.2973
0.2780
Exp_Int
-0.0845
0.2310
NetLoan_Debt
Number of obs =
226
Number of obs =
282
LR chi2(16) = 295.4400
LR chi2(9) = 353.8200
Prob > chi2 = 0.0000

Prob > chi2 = 0.0000
Pseudo R2
= 0.5036
Pseudo R2
= 0.4845
Log likelihood = -145.6301
Log likelihood = -188.2233
Source: Author’s caculating from thesis’ data sets
Basing on the regression results of ordered logit models on 5 sub data sets and the initial data set in
emerging markets presented in table 4.7, we calculate the frequencies of variables having significant cofficients
in above models.
Table 4.8: The frequency of variables having significant cofficients in ordered logit models
in emerging markets
Variable
Country_rating
Bicra
Government
Group
LnAss
AssGrow
CreGrow
LoanLoss_Ln
LoanLoss_Equ
LoanPro_Loan
Equ_Ass
Equ_Loan
Equ_ShortCap
Equ_Debt
IntIn_Loan

Frequency (%)
Variable
Frequency(%)
100 IntIn_Ass
0
100 IntEx_Cap
0
100 NIM
0
100 NetIntIn_Ass
0
100 OthIn_Ass
83
100 NonIntEx_Ass
67
0 ROAA
17
83 ROAE
17
0 Exp_Int
33
17 NetLoan_Ass
0
0 NetLoan_ShortCap
0
0 NetLoan_Debt
0
0 LiAss_ShortCap
0
83 LiAss_Debt

0
17
Source: Author’s caculating from thesis’ data sets

The empirical results in table 4.7 and 4.8 show that Country_rating, Bicra, Government, Group, LnAss,
AssGrow, LoanLoss_Ln, Equ_Debt và OthIn_Ass are variable having significant cofficients in the model on the
initial data set. Moreover, these proxies have the frequencies greater than 80% in ordered logit models on sub
data sets and initial data set. Besides that, the results of one way anova analyzing presented in table 4.1 also
indicate that LnAss, AssGrow, LoanLoss_Ln, Equ_Debt and OthIn_Ass having different average values in each
bank’s credit ratings. So that, we can conclude that these proxies have impacts on the bank’s credit ratings in
emerging markets.
We regress the ordered logit models with these variables on the initial data set in emerging markets.


20
Table 4.9: The ordered logit model with selected variables on initial data set in emerging markets
Standard
Variable
Coefficient
P-Value
Confident interval
Deviation
Min
Max
1.8164
0.2897
0.0000
1.2487
2.3842
Country_rating

0.8881
0.1857
0.0000
0.5241
1.2520
Bicra
1.1377
0.3722
0.0020
0.4082
1.8672
Government
3.9542
0.5847
0.0000
2.8082
5.1003
Group
0.5979
0.1049
0.0000
0.3923
0.8035
LnAss
-4.5942
1.4117
0.0010
-7.3612
-1.8273
AssGrow

-0.0832
0.0220
0.0000
-0.1264
-0.0400
LoanLoss_Ln
0.0157
0.0038
0.0000
0.0083
0.0232
Equ_Debt
-0.1744
0.0536
0.0010
-0.2794
-0.0695
OthIn_Ass
16.6202
19.3397
24.3097

/cut1
/cut2
/cut3

1.7691
1.9176
2.1752

13.1528
15.5812
20.0465

20.0876
23.0982
28.5730

Number of obs =
282
LR chi2(9) = 353.8200
Prob > chi2 = 0.0000
Pseudo R2
= 0.4845
Log likelihood = -188.2233
Source: Author’s caculating from thesis’ data sets
We apply the same analyzing process with the variables having impact on bank’s credit ratings in
developed markets.
Table 4.12: The ordered logit model with selected variables on initial data set
Variable
Country_rating
Government
Group
LnAss
LoanLoss_Ln
Equ_Ass
Equ_Loan
IntIn_Loan
NIM
ROAE

Exp_Int

Coefficient
1.3263
2.2145
1.3725
0.6562
-0.1043
0.1149
0.0131
0.1019
-0.5620
0.0121
-0.0232

in developed markets
Standard
P-Value
Deviation
0.1564
0.4154
0.3041
0.0939
0.0233
0.0485
0.0059
0.0740
0.1659
0.0050
0.0061

0.0000
0.0000
0.0000
0.0000
0.0000
0.0180
0.0270
0.1690
0.0010
0.0150
0.0000

Confident interval
Min
1.0197
1.4003
0.7765
0.4721
-0.1500
0.0197
0.0015
-0.0432
-0.8871
0.0023
-0.0352

Max
1.6328
3.0287

1.9684
0.8403
-0.0586
0.2100
0.0247
0.2470
-0.2368
0.0218
-0.0113


21
NetLoan_Ass
/cut1
/cut2
/cut3
/cut4
/cut5

0.0194
11.2014
14.3689
17.1884
21.2567
23.7446

0.0087
1.8856
1.9165
2.0324

2.1849
2.2434

0.0260

0.0024
7.5057
10.6126
13.2049
16.9744
19.3476

0.0365
14.8970
18.1251
21.1718
25.5390
28.1415

Number of obs =
296
LR chi2(12) = 287.7800
Prob > chi2 = 0.0000
Pseudo R2
= 0.3545
Log likelihood = -262.0579
Source: Author’s caculating from thesis’ data sets
4.2.2 The result of accessing model’s crediability
We apply the BIC (Bayesian information criteria) to compare the crediability of the selected model to
random models.

The results of BIC show that the selected models are the finest models.
4.3 Test of the ordered logit model’s assumptions
4.3.1 Test of multicollinearity in the models
To access the affect of high multicollinearity on the variables in the model, we calculate VIF ratios of
each proxies of the model. The variables cause the high multicollinearity in the model if their VIF ratios greater
than 10. We calculate the VIF ratios of the variables of the ordered logic model in developed markets and in the
emerging markets.
The results of VIF ratios show that the VIF ratios of all variables in 2 models are below 10. So that we
can conclude that the ordered logic models on the data sets in developed market and in emerging markets are not
affected by the multicollinearity.
4.3.2 Test of heteroskedasticity in the models
One of the most important assumptions of models is that there is no heteroskedasticity in the model. The
violation of this assumption can impact the standard deviations and p-values of variables in the models. We add
option Robust in the regression command to regress the model without heteroskedasticity assumption and
compare to the above regression results.
The results of model’s regression without heteroskedasticity assumption show that there changes in
standard deviations and p-value of the variables in the models. But the differences are not much and do not
change p-value and the direction of cofficients. So that, the thesis’models are not affected by the
heteroskedasticity.
4.3.3 Test of missing essential variables in the models
The test of missing essential variables in the models, proposed by Chen et al (2015), is applied in this
study. The test’s results show that there is no missing essential variables in the thesis’ models.


22
4.4 Access the marginal affects of the model’s variables.
We calculate the marginal affects of the variables having significant p-values in 4.2.1. The results show
that Country_rating, Government and Group have strong marginal affects on the banks’ classification. Besides,
that marginal affects of these variables in the data set of emerging markets are greater these affects in the
developed markets.

4.5 Analyzing the differences in impact of affecting factors on bank’s credit ratings between developed
markets and emerging markets
As presented in 3.4 to indentify the differences in impact of affecting factors on bank’s credit rating
between developed markets and emerging markets, we merge the data set in developed markets and the data set
in emerging markets together. Moreover, we add Emer, a dummy variable, in the model. This proxy take 0 value
if banks locate in emerging markets. Otherwise, it takes 1 value. After that, we create interact proxies between
Emer and each variables in the model. We add Emer and these interact proxies in the model
The model’ resgression result show that the cofficients of Country_rating_Emer and Emer are
significant and have directions as expected. Besides, the cofficient of Bicra_Emer is significant but Bicra’s is
not. The empirical results also indicate that the cofficients of Government_Emer and Group_Emer are
significant. The cofficient of Government_Emer is negative. But the cofficient of Group_Emer is positive.
Moreover, we notice that LnAss, AssGrow, LoanLoss_Ln, NIM, ROAE, Exp_Int and the interact
proxies between these variables and Emer have significant cofficients and oppositve directions.
4.6 Result analyzing
4.6.1 Analyzing the empirical model of affecting factors on bank’s credit ratings in emerging markets
4.6.1.1 The impact of systematical factors on bank’s credit ratings in emerging markets
The systematical factors include economic risk and industry risks of countries where banks locate have
positive affects on bank’s credit ratings. Besides that, the coefficients of these variables are significant at 1%
level.
These results are similar to bank’s rating criteria of international rating agencies. These results are also
coincident with the empirical results of Bellotti et al (2011a, 2011b) and Caporale et al (2012).
4.6.1.2 The impact of ownership factor on bank’s credit rating in emerging markets
From the regression results showed in table 4.9, we notice that ownership factor has a strong impact on
bank’s credit ratings. Government variable has a positive impact on bank’s credit ratings. This indicates that the
commercial banks belonging to government of countries where banks locate has a larger change to receive
better ratings. Similarly, Group variable has a positive affect on bank’s credit ratings. This show that the
commercial banks belonging big financial groups have a better changes to be classified into A or BBB rating
than the others.
These results are agreed with bank’s rating criteria of international rating agencies. Because these
agencies suppose that governments or international financial groups tend to support to their belonging

commercial banks.
4.6.1.3 The impact of scale factor on bank’s credit ratings in emerging markets


23
Bank’s size, presented by the total asset of the banks, has positive affect on bank’s credit ratings. This
implies that the bigger size do the banks have the better change do they have to receive the better ratings.
Goddard et at (2004) explain that big commercial banks have economic scale advantages. They benefit from
their market power to generate an above average profit rate.
4.6.1.4 The impact of financial ratios on bank’s credit ratings in emerging markets
The empirical results show that the average assets growth rate over 3 years and the overdue loan/Total
loan ratio have negative impacts on bank’s credit ratings. These variable’s coefficients are significant at 1%
level. This indicates that banks having a rapid assets growth rate may be classified into low credit ratings. This is
agreed with the empirical results of Köhler (2015). Besides, the impact of the overdue loan/Total loan ratio
coincides with the bank rating criteria of banking regulatory authorities. Moreover, it is also proved by Caporale
et al (2012).
Next, among capital capacity ratios, we notice that the equity/total debt ratio has positive affect on
bank’s credit rating at 1% significant level. This implies that banks having high equity/total debt ratios can
receive better bank’s credit ratings. Pasiouras and Kosmidou (2007) explain that a strong equity help to increase
bank’s creditworthiness, reduce capital expense. These banks have more capability to expend their business and
deal with operation risks.
Finally, among earning ratios, we find that the other operation income/Average total asset (OthIn_Ass)
has negative impact on bank’s credit ratings. Berger et al (2010) indicate that bank’s diversification in China
cause an increase in operation expenditure and reduce bank’s profitability.
4.6.2 Analyzing the empirical model of affecting factors on bank’s credit ratings in developed markets
4.6.2.1 The impact of systematical factors on bank’s credit ratings in developed markets
Among systematical factors, we notice that the economic risk factor has positive impact on bank’s credit
rating at 1% significant level. The coefficient of Bicra, industry risk, is not significant in the model.
4.6.2.2 The impact of ownership factor on bank’s credit rating in developed markets
The results indicate that all independent variables presenting these factors have positive impact on

bank’s credit rating at 1% significant level. This result is the same as the result in the ordered logic model in
emerging markets.
4.6.2.3 The impact of scale factor on bank’s credit ratings in developed markets
The scale factor has positive impact on bank’s credit ratings. This result is also the same as the result in
the ordered logic model in emerging markets.
4.6.2.4 The impact of financial ratios on bank’s credit ratings in developed markets
Among bank’s asset quality ratios, we notice that the overdue debt/total debt ratio has negative impact
on bank’s credit rating at 1% significant level.
Next, the model’s results show that the equity/total asset ratio and the equity/total net debt ratio have
positive impact on bank’s credit rating at 5% significant level. This result is coincident with the result of the
empirical model in emerging market and empirical studies presented above.
In the group of earning ratios, we notice that there are 3 ratios include NIM ratio, ROAE ratio and the
total expenditure/total income ratio having significant coefficients. NIM ratio has negative impact on bank’s
credit ratings. This result is the same as the result of Matousek and Stewart (2009). In the emerging markets, the


24
levels of deposit interest rate and lending interest rate are lower than these of developed markets. So that the
negative impact of NIM on bank’s credit ratings may be caused by the difference in risk and level of interest
rates.
We also notice that the net loan/total assets ratio has positive impact on bank’s credit ratings. This result
is agreed with many previous empirical studies. Matousek and Stewart (2009) indicate that the liquidity
assets/total assets has negative impact on bank’s credit ratings. Because the bank which face difficuties in capital
mobilization must maintain more liquidity assets. This also reduce the bank’s profitabilies.
4.6.3 Analyzing the difference in impact of affecting factors on bank’s credit rating between developed
markets and emerging markets
4.6.3.1 The difference in impact of systematical factors on bank’s credit ratings between developed markets
and emerging markets.
The empirical results of the models indicate that credit rating of the country where banks locate have a
stronger impact on bank’s credit rating in emerging markets than in developed markets. Among the factors, the

industry risk only has impact on bank’s credit ratings in emerging markets. Liu and Ferri (2001) also conclude
that economy risk strongly impact on corporation’s credit ratings in emerging markets. But this factor does not
impact on the credit ratings of corporations in developed markets. Williams et al (2013) prove that the countries’
credit ratings are the ceilings of corporations’ credit ratings in these countries.
4.6.3.2 The difference in impact of ownership factor on bank’s credit ratings between developed markets and
emerging markets
The thesis’ results show that international financial group’s ownership has stronger impact on bank’s
credit rating in emerging markets than in developed markets. Meanwhile, the positive impact of government
ownership on bank’s credit ratings in emerging markets is lower than in the developed markets. Mirzaei et al
(2013) indicate that commercial banks belonging to international financial groups have a better profitability in
emerging markets than in developed markets. The reason is when these banks penetrate in the emerging markets,
they can use their advantages in banking technologies and innovation services to compete with domestic
commercial banks in order to attain a better profitability. In contrast, when these banks enter the developed
markets their profitability may reduce due to competion with domestic banks in these countries.
4.6.3.3 The difference in impact of scale factor and financial ratios on bank’s credit ratings between
developed markets and emerging markets
The thesis’s result suggest that there is no difference in impact of scale factor on bank’s credit ratings
between developed markets and emerging markets. Otherwise, the average asset growth rate over 3 years has
negative impact on bank’s credit rating in emerging markets. But this ratio has no impact on bank’s credit ratings
in developed markets. The reason of this is discussed in detail in section 4.6.1
Besides, the empirical results show that the impact of the overdue loan/total loan ratio on bank’s credit
ratings is lower in the emerging markets than in developed markets. The cause of this may come from
information asymmetry as explained in previous sections. Suarez (2001) explain that bank’s quality assets ratios
do not reflect the true credit risk of commercial banks due to difference in bank’s accounting standards.
Moreover, bank financial reports in emerging markets are not reliable due to inexact classification of overdue
loans.


25
Next, among the capital capacity ratios, we notice that there is no difference in impact of the equity/total

assets ratio and the equity/net loan ratio on bank’s credit rating between developed markets and emerging
markets. Shen et al (2012) explain that the rating agencies access carefully the impact of capital capacity ratios
on bank’s credit ratings in both high information asymmetry countries and low information asymmetry
countries. Although these ratios are usually not transparent in countries with high information asymmetry. But
the ratings agencies have no other choice and have to classify banks with high capital capacity ratios into high
credit ratings.
Moreover, the impact of bank’s earning ratios (include the total expenditure/total income ratio, Net
interest margin and net return/average total equity ratio) on bank’s credit ratings also reduce in emerging
markets. The information asymmetry in these countries is also the cause of this problem.
Finally, we notice there is no difference in impact of the net loan/total assest ratio on bank’s credit rating
between developed markets and emerging markets.