Nghiên cứu tác động của dịch vụ ngân hàng quốc tế đến hiệu quả hoạt động của các ngân hàng thương mại việt nam (tt)

MINISTRY OF EDUCATION AND TRAINING
UNIVERSITY OF ECONOMICS OF HO CHI MINH CITY
-----------------------------

HUYNH THI HUONG THAO

RESEARCHING THE IMPACTS OF
INTERNATIONAL BANKING SERVICES
TO OPERATIONAL EFFICIENCY OF
VIETNAMESE COMMERCIAL BANKS

Major: Finance and Banking
Code: 62340201

SUMMARY OF ECONOMIC DOCTORAL THESIS

Ho Chi Minh City - 2017


The work was completed at:
University of Economics of Ho Chi Minh City
The scientific instructors:
Assoc. Prof. Dr. Tran Huy Hoang
Reviewer:
1.
2.
3.
The thesis will be protected in front of thesis scoring
council at
the University of Economics of Ho Chi Minh City
at ... hour ......day ......month ...... 2017

The thesis can be studied at:
- Synthesis Scientific Library in Ho Chi Minh City
- Library of University of Economics of Ho Chi Minh City


1
CHAPTER 1: INTRODUCTION
1. Reason for study of the thesis
Operational efficiency (OE) is the top concern of all
commercial banks. In the trend of international economic integration
becoming more and more widespread, commercial banks with
banking services including both domestic and international banking
services (IBS), the role of IBS has a strong influence on the bank's
OE increasingly. Some scientific studies on the impact of a banking
service or a specific indicator such as: bad debt, non-credit services,
income diversification ... for the bank's OE but the study of the
impact of IBS on OE of commercial banks is not found in previous
studies. From the new requirements and the practical needs in
Vietnam, the study of overall OE and the impact of IBS on OE indepth are very important and valuable, the author selected the topic:
"Researching the impacts of IBS to OE of VCB" as the thesis.
2. Objectives and research questions of the thesis
The research has the general objective of studying the
impacts of IBS to OE of VCB. Based on the general objective, the
dissertation is carried out with two specific objectives:
Objective 1: To study the impact of IBS on the profitability
of VCB according to their profitability indicators: return on assets
(ROA) and return on equity (ROE).
Objective 2: To study the impact of IBS on the marginal
efficiency of VCB according to the indexes of Data Envelopment
Analysis (DEA): technical efficiency (TE), pure technical efficiency

(PTE), scale efficiency (SE).
To achieve the research objectives, the thesis has to answer
two research questions:


2
Question 1: Does IBS affect to indicators of the profitability
of VCB?
Question 2: Does IBS affect to indicators of DEA: TE, PTE,
SE of VCB?
3. Object and scope of the research
The thesis chosen object of study is the impacts of IBS to OE
of VCB. Research’s data taken from annual reports or financial
statements were audited from 38 banks including state commercial
banks and joint stock commercial banks in period 2008-2014. The
macroeconomic indicators of Vietnam such as the rate of economic
growth, inflation rate from the General Statistics Office of Vietnam.
Within the scope of the study, the thesis utilized the two
most common indicators when evaluating the profitability: ROA and
ROE. For the method of evaluating the marginal efficiency analysis,
the thesis selects the DEA method as the representative with
indicators: TE, PTE, SE. The research in the dissertation does not
include international investment operations.
4. Research methodology
The research model is based on Trujillo-Ponce (2013),
Nguyen Thi Hong Vinh (2014), Pham Huu Hong Thai (2014), Lam
Chi Dung et al. (2015) ... to study the impacts of IBS to OE. In
addition, the thesis approached the nonparametric method DEA of
Avkiran (1999), Ngo Dang Thanh (2010), Lieu Thu Truc and Vo
Thanh Danh (2012), Truong Quang Thinh (2012), Le Phan Thi Dieu

Thao and Nguyen Thi Ngoc Quynh (2013) ... to determine TE, PTE,
SE.


3
5. Structure of the subject
In addition to the introduction, conclusion, appendices,
references, structure of the thesis consists of 5 chapters: Introduction;
Theoretical foundations and empirical researches; Methods, data and
models of research; Research results; Conclusions and policy
implications.
6. Practical significance of the thesis
- The thesis has presented the theory related to IBS and OE,
identifying the indicators of IBS that impact on OE of VCB, thereby
determining the model of the impacts of IBS to OE.
- The thesis has quantified TE, PTE, SE of VCB in the
period 2008-2014, the results of this research is a scientific reference
to help banks recognizing the status of business activities, evaluating
the level of completion of the plan and the level of banking growth.
- The thesis analyzes the positive factors, the negative factors
of IBS affecting OE and suggests some implications to promote
positive factors, controls negative factors belongs to IBS to
constantly improve OE.


4
CHAPTER

2:

THEORETICAL

FOUNDATIONS

AND

EMPIRICAL RESEARCHES
2.1. International banking services of banks
2.1.1. The concept of international banking services
According to Nguyen Thi Cam Thuy (2012), IBS includes all
financial services requested by customers related to international
trade, international investment, international tourism ... In other
words, the bank provides financial services in the international
market for profit purposes. According to Tran Huy Hoang et al.
(2006), a bank providing IBS is a bank providing IBS related to
foreign exchange or nonresidents. In spacial scope, IBS is not only
between countries but also within each country's territory, regardless
of whether the currency of that country is freely convertible into
another currency. On the basis of inheriting the concept of IBS and
legal documents in Vietnam, concepts of IBS used in this thesis is the
banking services related to foreign currency provided by banks.
2.1.2. Indicators of international banking services
Two indicators: the ratio of foreign currency liabilities to
total liabilities and capital (Truong Quang Thong, 2010) and the ratio
of foreign currency loans to total foreign currency assets are the two
most reflective indicators of IBS, are independent variables,
explaining the effect of IBS on banking OE. The ratio of foreign
currency liabilities to total liabilities and capital reflects the funding
mobilization operations of commercial banks. Meanwhile, the ratio
of foreign currency loans to total foreign currency assets reflects the

operations of using funds of commercial banks.


5
2.2. Operational efficiency of commercial banks
2.2.1. The concept of operational efficiency of commercial banks
According to Berger and Mester (1997), commercial banks'
OE represents in the relationship between revenues and the cost of
using resources, is the ability to turn inputs into the best outputs in
business.
2.2.2. The methods of evaluating the operational efficiency of
commercial banks
According to Berger and Humphrey (1997), Heffernan and
Fu (2008), the analysis of commercial banks’ OE often uses two
main methods: profitability indicators and the margin efficiency
analysis method.
2.2.2.1. The method of using profitability indicators
The most common used indicators are ROA and ROE
(Tarawneh, 2006, Zeitun, 2012, Naceur, 2003). Besides the above
two coefficients of profit, OE of banks is also represented by the
marginal rate of income.
2.2.2.2. Marginal efficiency analysis method
The research using DEA method, DEA’s results includes:
technical efficiency (TE) or whole TE, pure technical efficiency
(PTE) and scale efficiency (SE) (Charnes, Cooper and Rhodes, 1978;
Banker, Charnes and Cooper, 1984).
2.3. The relationship between international banking services and
OE of banks
IBS can not stand individually that can only grow well in
horizontal and vertical linkages with other services of banks,

together with support and the rise of mutual control. In terms of the
nature and history of the development, IBS is always a useful part,


6
tied to domestic banking services and created the entire business
operations of a bank. IBS is an integral part of the banks' business
activities and the development of IBS will impact on OE of banks
certainly and effectively. Therefore, the study of the impact of IBS
to OE is truly meaningful and essential.
2.4. The empirical studies of IBS and OE
2.4.1. The empirical studies of IBS and OE
With in-depth studies of IBS in Vietnam: Le Thanh Lan
(2004), Tran Huy Hoang et al. (2006), Nguyen Thi Cam Thuy
(2012), the authors have just used statistical or inductive, deductive
method without quantitative method to study specially about OE and
IBS.
With researches of Nguyen Viet Hung (2008), Truong
Quang Thong (2010), Nguyen Thi Loan and Tran Thi Ngoc Hanh
(2013), Trinh Quoc Trung and Nguyen Van Sang (2013), Pham Huu
Hong Thai (2014), Nguyen Thi Hong Vinh (2014), Vo Xuan Vinh
and Tran Thi Phuong Mai (2015), Lam Chi Dung et al. (2015), OE
can assess by profitability indicators such as ROA, ROE or by
indicators of DEA. Pooled OLS, FEM, REM or Tobit regression
models are often used to analyze the factors affecting on banks' OE.
There have been some studies on the impact of a banking servise or
a specific indicator of banks such as bad debt, non-credit services,
bank's income diversification ... on OE, but a separate research
considering the impact of IBS on banks’ OE using quantitative
methods is not available in Vietnam.

Table 2.4: Statistics of the independent variables affecting
commercial banks' OE used in empirical researches


7
Independent

Author

Correlation

variables
Yudistira (2004), Nguyen Viet Hung
(2008), Sufian and Chong (2008),
Equity / Total
assets

+

Sanchez et al. (2013), Raphael
(2013), Ayadi (2014) ...
Gul et al. (2011), Aremu et al. (2013)

-

…
Yudistira (2004), Nguyen Viet Hung
Total assets

(2008), Gul et al. (2011), Alper and

+

Anbar (2011), Raphael (2013) ...
Sufian and Chong (2008) …

-

Nguyen Viet Hung (2008), Alper and

-

Total loans /

Anbar (2011) ...

Total assets

Gul et al. (2011), Sufian (2011),

+

Garza-Garcia (2012) …
Funds

Nguyen Viet Hung (2008), Nguyen

mobilized /

Thi Loan and Tran Thi Ngoc Hanh

Lending
Economic
Growth

(2013) …
Gul et al. (2011), Garza-Garcia

+

(2012) …
Gul et al. (2011), Sufian (2011) …

Inflation rate

-

+

Sufian and Chong (2008), GarzaGarcia (2012), Ongore and Kusa

-

(2013), Sanchez et al. (2013) …
2.4.2. Issues to study in Vietnam
The study is based on previous studies and has some new


8
features based on the following grounds:

(1) Economic integration to create conditions for the
banking system and the development of banking services as
Claessens and Horen (2009), Sufian and Habibullah (2011).
(2) Diversification of banking services both domestic
banking services and IBS will help banks maximize profits
(Chiorazzo et al., 2008, Gurbuz et al., 2013; Vo Xuan Vinh and
Tran Thi Phuong Mai , 2015 ...). One of the directions for VCB to
quickly stand up with commercial banks in the region and
international integration are the development and diversification of
IBS (Tran Thi Van Anh, 2016).
(3) The fact that there are many studies on the impact of
banking services on OE but still do not have any researches in
Vietnam and around the world gives a quantitative model of the
impact of IBS on OE of commercial banks.
There are many works in the world that research the
factors affecting banks’ OE and from those results, the authors
have proposed appropriate policies to improve banks’OE (Sufian,
2011, Garza- Garcia, 2012; Aremu et al., 2013; Ayadi, 2014 ...).
However, these studies only deal with banking services in general,
without any specific studies on IBS.
Based on the inheritance of these studies, the thesis differs
from previous studies on research content and research models
when using Pooled OLS, FEM, REM, FGLS to evaluate the
impacts of IBS to OE of VCB in the period 2008-2014. This also
serves as a basis for finalizing a suitable policy framework in the
management of commercial banks in Vietnam in the integration
period.


9

CHAPTER 3: METHODS, DATA AND RESEARCH MODELS
3.1. Research data
To carry out the research, the author has collected data on
the financial statements or annual reports of 38 commercial banks in
the period 2008 - 2014, macroeconomic data of Vietnam such as
economic growth rate, the inflation rate were collected from the
General Statistics Office of Vietnam. The research was classified
into two groups of banks based on the criteria of equity and total
assets, of which group 1 comprises 11 banks (Vietinbank,
Vietcombank,

BIDV,

Agribank,

Sacombank,

Militarybank,

Techcombank, Eximbank, SCB, ACB, SHB) and Group 2 includes
27 banks (PVcombank, Maritimebank, VPbank, HDbank, VIB,
Lienvietpostbank,

Anbinhbank,

SeAbank,

DongAbank,

TienPhongbank, BacAbank, MDbank, OCB, VietAbank, MHB,

Saigonbank, Kienlongbank, PGbank, NamAbank, Vietcapitalbank,
NCB,

Phuongnambank,

Oceanbank,

Baovietbank,

VNBC,

Westernbank, GPbank).
3.2. Research models
Derived from the model of domestic and foreign authors on
factors affecting the OE of commercial banks, the regression models
used the studies: Trujillo-Ponce (2013), Nguyen Thi Hong Vinh
(2014), Pham Huu Hong Thai (2014), Vo Xuan Vinh and Tran Thi
Phuong Mai (2015), Lam Chi Dung et al. (2015) ...., the author has
developed a common model for the impact of IBS on banks’ OE as
follows:
OE = α + βIBSi, t + γCONTROLi, t + εi, t.
In which: OEi, t is the dependent variable with the
representative variables are ROA, ROE, TE, PTE, SE of bank i in


10
year t.
IBSi, t is the independent variable of IBS of bankk i in year t
with two representative variables: the total foreign currency loans to
foreign currency assets ratio (FLFA) and foreign currency liabilities

to total liabilities and capital ratio (or foreign currency liabilities on
total assets (FLTA).
CONTROLi, t is the control variables for the internal factors
of banks (equity to total assets ratio (ETA), the size of bank assets
(SIZE), loans to total assets ratio (LTA), deposits to loans ratio
(DTL)) and macroeconomic factors (economic growth (GDP), rate
of inflation (IFL)) impacting banks’ OE in year i; α is the blocking
factor; β and γ are estimated parameters; ε is the random number
error.
3.3. Describe the research variables
3.3.1. For the impact of IBS on ROA, ROE of Vietnamese
commercial banks
- Dependent variable: ROA, ROE.
- Independent variables includes: the total foreign currency
loans to foreign currency assets ratio (FLFA) and foreign currency
liabilities to total liabilities and capital ratio (or foreign currency
liabilities on total assets (FLTA).
- Control variables includes: equity to total assets ratio
(ETA), the size of bank assets (SIZE), loans to total assets ratio
(LTA), deposits to loans ratio (DTL), economic growth (GDP), rate
of inflation (IFL).
Hypotheses related to ROA and ROE dependent variables as
follows:
Hypothesis 1a: There are opposite correlations between


11
FLFA and ROA of all banks, banks in group 1 and banks in group 2.
Hypothesis 1b: There are opposite correlations between
FLFA and ROE of all banks, banks in group 1 and banks in group 2.

Hypothesis 2a: There are same correlations between FLTA
and ROA of all banks, banks in group 1 and banks in group 2.
Hypothesis 2b: There are same correlations between FLTA
and ROE of all banks, banks in group 1 and banks in group 2.
3.3.2. For studying the impacts of IBS on banks’ OE according
to the indicators of the DEA method: TE, PTE, SE
Based on studies by Leightner and Knox Lovell (1998),
Avkiran (1999), ..., the study uses a revenue and cost approach to
have input and output variables, including:
Input variables: includes interest expenses (X1), wage costs
(X2) and other expenses (X3).
Output variables: consists of interest income (Y1) and other
income (Y2).
- Dependent variables: technical efficiency (TE), pure
technical efficiency (PTE), scale efficiency (SE).
- Independent variables includes: the total foreign currency
loans to foreign currency assets ratio (FLFA) and foreign currency
liabilities to total liabilities and capital ratio (or foreign currency
liabilities on total assets (FLTA).
- Control variables includes: equity to total assets ratio
(ETA), the size of bank assets (SIZE), loans to total assets ratio
(LTA), deposits to loans ratio (DTL), economic growth (GDP), rate
of inflation (IFL).
Hypotheses related to TE, PTE, SE dependent variables as
follows:


12

Hypothesis 1c: There are opposite correlations between

FLFA and TE of all banks, banks in group 1 and banks in group 2.
Hypothesis 1d: There are opposite correlations between
FLFA and PTE of all banks, banks in group 1 and banks in group 2.
Hypothesis 1e: There are opposite correlations between
FLFA and SE of all banks, banks in group 1 and banks in group 2.
Hypothesis 2c: There are same correlations between FLTA
and TE of all banks, banks in group 1 and banks in group 2.
Hypothesis 2d: There are same correlations between FLTA
and PTE of all banks, banks in group 1 and banks in group 2.
Hypothesis 2e: There are same correlations between FLTA
and SE of all banks, banks in group 1 and banks in group 2.
3.4. Research methodology and verification
The study using panel data was regressed in four ways:
Pooled OLS, FEM, REM, FGLS. The tests used were: Hausman
testing for FEM or REM modeling, Modified Wald testing and
Breusch-Pagan Lagrangian testing for verify variance (Green, 2012),
Wooldridge for testing Self-correlation phenomena (Wooldridge,
2002 and Drukker, 2003). FGLS models can control self-correlation
and variance (Beck and Katz, 1995). The DEA method is used to
analyze TE, PTE, SE of VCB for the period 2008-2014 by using the
DEAP 2.1 software.


13

CHAPTER 4: RESEARCHING RESULTS
4.1. Studying the impact of IBS on OE according to ROA, ROE
of Vietnamese commercial banks
4.1.1. Descriptive statistics about variables
Data are presented in descriptive statistics, each variable is

described as: variable name, observation number, minimum value,
maximum value, mean value, standard deviation, therefore to
perform the most general analysis of variables introduced into the
model.
4.1.2. The coefficient of correlation between variables
Correlation analysis was used to examine the relationship
between FLFA, FLTA, ETA, SIZE, LTA, DTL, GDP, INF and
dependent variables: ROA, ROE for all banks, banks in group 1 and
banks in group 2.
4.1.3. Multi-collinear test results
The correlation coefficient between independent variables is
less than 0.8, which suggests the possibility of low multi-collinearity
in regression analysis.
Table 4.6: Multivariate test results of variables
Variable

VIF

VIF

(All banks) (Banks in group 1)

VIF
(Banks in group 2)

FLFA

1,12

1,37

1,07

FLTA

1,26

1,54

1,18

ETA

2,02

1,64

2,36

SIZE

2,47

2,81

2,98

LTA

2,22

4,57

2,13


14
DTL

2,27

3,05

2,20

GDP

1,24

1,26

1,29

INF

1,32

1,35

1,51

Medium VIF

1,74

2,20

1,84

Source: Results from Stata 11.1 with data of 38 selected VCB
4.1.4. Research results for ROA dependent variables
Table 4.8: Summary of research results for ROA dependent variables
Variable

Impact on ROA

All banks
FLFA

Banks in group 1

Same direction but

Same

not

with

statistically

significant
FLTA

Same
with

1% but not statistically
direction

1% with

significance level
Regression ROA
model

direction Negative direction

significance level
direction Same

Banks in group 2

1%

significance level

significant
Same
with

direction
1%

significance level

= ROA = 0,0304053 ROA

=

0,0192177*FLTA

+ 0,00867*FLFA + 0,0132762*FLTA

+ 0,0401921*ETA

0,0238037*FLTA - + 0,0372443*ETA
0,0016915*SIZE + + 0,202777*GDP
0,0137062*INF

Source: Results from Stata 11.1 with data of 38 selected VCB
FLFA effects in the same direction (+0,00867) to ROA at 1%
statistically significant for banks in group 1 and not statistically
significant for all banks and banks in group 2, ie, the increase in
FLFA will bring higher efficiency for banks in group 1, is a group of
banks with foreign currency lending have a significant proportion of
total loans, which have shown the advantages of this banks’ large


15

scale of foreign currency lending. According to a study conducted by
Truong Quang Thong (2010), the foreign currency lending on total
loans has a positive impact on the group of state-owned commercial
banks as the banks in group 1, the author's suggestion policy should
increase the foreign currency lending to the group of state
commercial banks. So, the research results are similar to the study of
Truong Quang Thong (2010) but contrary to the theory 1a.
FLTA impacts on all banks (+0,0192177), banks in group 1
(+0,0238037) and banks in group 2 (+0,0132762) are all significant
1%, coefficient of banks in group 1 is higher than all banks and banks
in group 2. Positive correlation shows that if the banks increase the
mobilization of foreign currency and make good use of this mobilized
funds, they will increase the OE, have more investment opportunities
for banks. According to research results of Truong Quang Thong
(2010), FLTA has negative impact on ROA for both state-owned
commercial banks and joint stock commercial banks, therefore, the
results are not consistent with the results of Truong Quang Thong
(2010). Based on the above results, we accept the hypothesis 2a for
the sample all banks, banks in group 1 and banks in group 2.
4.1.5. Research results for ROE dependent variables
Table 4.10: Summary of research results for ROE dependent variables
Variable

Impact on ROE

All banks
FLFA

Banks in group 1

Banks in group 2

Negative direction

Same direction with Negative direction

but not statistically

1%

significant

level

significance but not statistically
significant


16
FLTA

Same direction with Same direction with

Same direction with

1%

1%

significance 1%

level

level

Regression ROE = -0,1931715 +
model

significance

significance

level
ROE

=

ROE

=

0,2122217*FLTA + 0,093209*FLFA

+ 0,1292072*FLTA

-

0,0116989*SIZE

- 0,0789531*ETA

+

+ 1,749651*GDP

-

0,9082031*GDP

+ 0,3059712*FLTA
0,8065711*ETA
0,1777064*INF

0,0827546*IFL

Source: Results from Stata 11.1 with data of 38 selected VCB
FLFA has the opposite effect on ROE for all banks and banks
in group 2 but not statistically significant, having the same effect with
1% significance for banks in group 1 (+0.093209), this result is in
contrast to hypothesis 1b. FLFA has the same effect on ROE as
statistically significant 1%, similar to the ROA, which means that the
increase in foreign currency lending of banks in group 1 will be more
effective for this banks, proving that these banks have managed well
the risks of foreign currency lending such as import-export finance,
international payment, international guarantee ….
FLTA impacts on all banks (+0,2122217), banks in group 1
(+0,3059712) and banks in group 2 (+0,1292072) are all significant
1%. The regression coefficient of banks in group 1 is higher than all
banks and banks in group 2. With their ability to manage of banks in
group 1, the more they mobilize foreign currency, the more they

increase OE. FLTA has a positive correlation with ROE, which is
similar to ROA, indicating that if banks use the foreign currency
mobilizing funds well, they will increase OE, creating more
investment opportunities for banks. This clearly shows that to


17
increase the bank's profitability, banks must increase their debt ratio
to increase leverage. With this result, we accept the hypothesis 2b for
the sample all banks, banks in group 1 and banks in group 2.
4.2. Studing the impacts of IBS on OE according to TE, PTE, SE
of Vietnamese commercial banks.
4.2.1. Study the OE of VCB with the DEA method
4.2.1.1. Descriptive statistics the research sample data variables
15.000.000
Y1
10.000.000

Y2

X1

5.000.000

X2
X3

0
2008

2009

2010

2011

2012

2013

2014

Figure 4.5: Average input and output variables 2008-2014
Source: Compiled from the annual report of 38 selected VCB

4.2.1.2. Results of TE estimation by DEACRS
The average TE of the samples over the years 2008, 2009,
2010, 2011, 2012, 2013, 2014 is 0,862; 0,917; 0,893; 0,946; 0,943;
0,931; 0,92 respectively, this shows that TE of VCB have reached
only 86,2% in 2008; 91,7% in 2009; 89,3% in 2010; 94,6% in 2011;
94,3% in 2012; 93,1% in 2013 and 92% in 2014.
4.2.1.3. Results of PTE estimation by DEA VRS
The average PTE over the years 2008, 2009, 2010, 2011,
2012, 2013, 2014 is 0,918; 0,959; 0,938; 0,966; 0,971; 0,951; 0,94
respectively. This suggests that PTE of VCB is only 91,8% in 2008;
95,9% in 2009; 93,8% in 2010; 96,6% in 2011; 97,1% in 2012;


18
95,1% in 2013 and 94% in 2014.

4.2.1.4. Scale efficiency of VCB
The average SE over the years 2008, 2009, 2010, 2011,
2012, 2013, 2014 is 0,939; 0,956; 0,952; 0,979; 0,972; 0,979; 0,98
respectively. This shows that SE of VCB is 93,9% in 2008; 95,6%
in 2009; 95,2% in 2010; 97,9% in 2011; 97,2% in 2012; 97,9% in
2013 and 98% in 2014.
4.2.2. Studying the impact of IBS on TE, PTE, SE of Vietnamese
commercial banks
4.2.2.1. Descriptive statistics about variables
The average TE was 91,58%, the lowest TE was
Kienlongbank in 2008 with 64,6%; The average PTE was 94,91%,
the lowest was Kienlongbank (68,5%). The average SE was 96,5%;
which is higher than the average TE and average PTE, indicating
that SE has a great impact on the OE.
4.2.2.2. The coefficient of correlation between variables
Correlation analysis was used to examine the relationship
between FLFA, FLTA, ETA, SIZE, LTA, DTL, GDP, INF and
dependent variables: TE, PTE, SE for all banks, banks in group 1
and banks in group 2.
4.2.2.3. Research results for TE dependent variables
Table 4.24: Summary of research results for TE dependent variables
Variable

Impact on TE

All banks
FLFA

Banks in group 1

Negative direction

Negative

but not statistically

with

significant

significance level

Banks in group 2

direction Same direction but
5% not

statistically

significant


19
FLTA

Same direction with Same direction but

Same direction with

5%

10%

significance not

level

statistically

significant

significance

level

Regression TE = 0,738668 + TE = 0,9456188 - TE = 0,3921823 +
model

0,1115416*FLTA + 0,1214874*FLFA

0,1230141*FLTA +

0,2398684*ETA

+

0,3038906*ETA

+

0,0222701*SIZE

-

0,0414814*SIZE

-

0,1926906*LTA

-

0,1511831*LTA

-

0,0300667*DTL

-

2,440381*GDP

1,735985*GDP

Source: Results from Stata 11.1 with data of 38 selected VCB
FLFA has the opposite effect on TE at statistically significant
level of 5% for banks in group 1 (-0,1214874) but the model is not
statistically significant (Prob> chi2 = 0,11161, Appendix 17), FLFA
is not statistically significant for all banks and banks in group 2.
Based on the above results, we reject the 1c hypothesis for the sample

all banks, banks in group 1 and banks in group 2 because regression
coefficients or regression models were not statistically significant.
FLTA impacts on all banks (+0,1115416) and banks in group
2 (+0,1230141) at 5% and 10% respectively. Based on the above
results, we reject the hypothesis 2c for the sample banks in group 1
because the regression coefficient is not statistically significant,
accepting hypothesis 2c for the sample all banks and banks in group
2. The results of the study on dependent variable of TE are similar to
the results when studying with the dependent variable ROA and ROE.
4.2.2.4. Research results for PTE dependent variables
Table 4.26: Summary of research results for PTE dependent variables


20
Variable

FLFA

FLTA

Impact on PTE

All banks

Banks in group 1

Banks in group 2

Negative direction

Negative direction

Negative direction

but not statistically

but not statistically

but not statistically

significant

significant

significant

Same direction with Same direction with Same direction but
5%

significance 5%

level

significance not

level

statistically

significant

Regression PTE = 0,66676 + PTE = 0,5903455 + PTE = 0,8495037 +
model

0,0671865*FLTA + 0,0847568*FLTA + 0,1759197*ETA

-

0,1634918*ETA

-

0,0182934*SIZE

+ 0,0199237*SIZE

0,0963768*LTA
2,085647*GDP

Source: Results from Stata 11.1 with data of 38 selected VCB
FLFA has the opposite effect on OE, which shows that
foreign currency lending does not bring efficiency to banks. This
shows that the Vietnamese banks are not well managing the risks of
foreign currency lending such as import-export finance. international
payment, international guarantee .... However, FLFA effect is not
statistically significant. Therefore, the 1d hypothesis of the study was
rejected.
FLTA impacts on all banks (+0,0671865) and banks in group
1 (+0,0847568) is 5% statistically significant level. This proves that
in order to increase banks’ OE, the banks expand their funding

mobilization activities in foreign currencies. With this result, we
accept the 2d hypothesis for the sample of all banks and banks in
group 1, rejecting the 2d hypothesis for banks in group 2.


21
4.2.2.5. Research results for SE dependent variables
Table 4.28: Summary of research results for SE dependent variables
Variable

FLFA

FLTA

Impact on SE

All banks

Banks in group 1

Banks in group 2

Negative direction

Negative direction

Same direction but

but not statistically

but not statistically

not

significant

significant

significant

Negative direction

Negative direction

Same direction but

but not statistically

but not statistically

not

significant

significant

significant

statistically

statistically

Regression SE = 0,981067 - HQQM = 1,251849 - HQQM = 0,7489874
model

0,082662*LTA

- 0,3542843*LTA

- + 0,0803187*ETA +

0,0187937*DTL

- 0,0606122*DTL

- 0,0137401*SIZE

0,0696219*INL

0,3777144*INF

0,0326891*LTA

Source: Results from Stata 11.1 with data of 38 selected VCB
FLFA has the opposite effect on all banks and banks in group
1 showed that increasing foreign currency lending did not increase
SE. Conversely, for banks in group 2, when increasing foreign
currency lending, SE will increase. However, the results of this study
are not statistically significant, so the hypothesis 1e is rejected.
FLTA has the opposite effect on all banks and banks in

group 1 showing that when banks increase foreign currency
mobilization activities, they do not increase SE. In contrast, for banks
in group 2, when the increase in foreign currency mobilization will
increase SE. However, this result is not statistically significant, so
hypothesis 2e is rejected.

-


22
CHAPTER 5: CONCLUSION AND POLICY IMPLICATIONS
5.1. Conclusions from the research model
With the dependent variables representing the OE are: ROA
and ROE, the results show that there is a similarity in the results of
research on FLFA and FLTA for both the dependent variable is ROA
and ROE.
As the dependent variables representing the OE are: TE,
PTE, SE, the research results show that FLFA is not statistically
significant, FLTA has a positive effect on the result of TE and PTE
(ie, the effectiveness of input management) and not statistically
significant with the dependent variable SE (which indicates the
ability of the bank to choose the optimal scale with the bank's
resources).
5.2. The main contribution of the thesis
In addition to the theoretical contributions, the thesis has
some practical contributions, the results of the thesis are expected to
support banks’ managers.
5.3. Some policy implications
5.3.1. With foreign currency mobilization service
According to research results, FLTA variable have

correlated positively with ROA, ROE, TE and PTE with all banks,
this result demonstrates that banks should increase in foreign
currency deposits to raise OE. To improve the ability to mobilize in
foreign currency, in addition to a reasonable interest rate policy,
VCB should consider operation scale through branches are allocated
according to appropriate geographic area nationally as well as
countries in the region.


23
5.3.2. With foreign currency credit service
According to the research results, FLFA has the same
relationship with ROA, ROE for banks in group 1, are the banks
with large foreign currency lending but not statistically significant
for the sample all banks. However, LTA have the opposite effect on
TE, SE proves that if banks expand their general lending activities
including local currency lending and foreign currency lending, this
will reduce OE. The current policy of the Vietnamese State Bank is
to strengthen control and limit foreign currency credit in order to
create stability in the foreign exchange market. This is in line with
the Vietnamese State Bank's management policies, and most
recently, Circular 24/2015/TT-NHNN and Circular 31/2016/TTNHNN on foreign currency lending.
5.3.3. With the other IBS
With foreign currency deposits raised, in order to maximize
profits required commercial banks to use these funds for other IBS’
diversification as develop foreign exchange business, international
payment services, international card services and correspondent
banks.
5.4. Some limitations of the study and subsequent research
The DEA model used in the study did not mention levels of

significance or reliability. In addition, due to the limitation of young
data, when the conditions and data are available, the next study on
the impacts of IBS on commercial banks’ OE are: longer study time,
broadening the scope of research, making research for all
commercial banks in Vietnam and some commercial banks in the
region.