College of Finance, Management and Development

Department of Development Economics

Program: Extension

DETERMINANTS OF LIQUIDITY POSITION:

EVIDENCE FROM COMMERCIAL BANKS IN
ETHIOPIA

BY
NEGASEW WORKU

JUNE, 2023
ADDIS ABABA, ETHIOPIA

0
 ETHIOPIAN CIVIL SERVICE UNIVERSITY
COLLEGE OF FINANCE, MANAGEMENT AND
DEVELOPMENT
DEPARTMENT OF DEVELOPMENT ECONOMICS
PROGRAM: EXTENSION

DETERMINANTS OF LIQUIDITY POSITION:

EVIDENCE FROM COMMERCIAL BANKS IN
ETHIOPIA

By
Negasew Worku
ID No. ECSU1902068

Under the Supervision of

Tesfaye Chofana (PhD)

A THESIS SUBMITTED TO THE GRADUATE PROGRAM, COLLEGE OF

FINANCE, MANAGEMENT AND DEVELOPMENT OF ETHIOPIAN CIVIL
SERVICE UNIVERSITY, IN PARTIAL FULFILLMENT OF THE
REQUIREMENTS FOR THE AWARD OF A MASTER’S OF SCIENCE
DEGREE IN DEVELOPMENT ECONOMICS.

JUNE, 2023
Addis Ababa, Ethiopia
1
 Declaration

This is to declare that the thesis entitled “Determinants of Liquidity Position: Evidence
from Commercial Banks in Ethiopia”, submitted in partial fulfillment of the
requirements for the degree of Master of Science in the (Department of Development
Economics/College of Finance, Management and Development) of Ethiopian Civil
Service University, is a record of original work carried out by me and has never been
submitted to any other institution to get any other degree or certificates. The assistance
and help I received during the course of this investigation have been duly acknowledged.

Name of the Candidate: Negasew Worku Teshome Date __________ Signature______

i
 Advisor/Supervisor Approval Form

ETHIOPIAN CIVIL SERVICE UNIVERSITY

COLLEGE OF FINANCE, MANAGEMENT AND DEVELOPMENT
DEPARTMENT OF DEVELOPMENT ECONOMICS

Approval of Thesis for Defense

I hereby certify that I have supervised, read, and evaluated this thesis titled
“Determinants of Liquidity Position: Evidence from Commercial Banks in
Ethiopia”, by Negasew Worku Teshome prepared under my guidance. I recommend the
thesis be submitted for oral defense.

________________________________ ________________ _____________

Supervisor’s name Signature Date

_________________________________ ________________ _____________

Name of Department Head Signature Date

ii
 Acknowledgment

First of all, I want to express my gratitude to the Almighty God for providing me the
ability, strength, and resolve to finish my study in general and my thesis in particular.

Then, I would like to extend my sincere gratitude to Tesfaye Chofana, the PhD adviser of
my research paper, for his essential advice and comments beginning with the proposal
and continuing until the completion of the study.

I also want to express my gratitude to the personnel and divisions of the National Bank of
Ethiopia and the Ministry of Finance for supplying the study with the macroeconomic
data and the audited financial statements of all commercial banks.

Finally, I am also sincerely appreciative of the love, support, and encouragement I have
received from my family and friends, as well as the many ideas they have shared with
me.

iv
 Table of Contents
Content Page

Declaration ........................................................................................................................... i
Approval of Thesis for Defense .......................................................................................... ii
Approval of Thesis after Defense ...................................................................................... iii
Acknowledgment ............................................................................................................... iv
Table of Contents ................................................................................................................ v
List of Tables ................................................................................................................... viii
List of Figures .................................................................................................................... ix
List of Appendices .............................................................................................................. x
Abbreviations and Acronyms ............................................................................................ xi
Abstract ............................................................................................................................. xii
CHAPTER ONE ................................................................................................................. 1
INTRODUCTION .............................................................................................................. 1
1.1 Background of the Study ........................................................................................... 1
1.2 Statement of the Problem .......................................................................................... 5
1.3 Objectives of the Study ............................................................................................. 8
1.3.1 Main Objective ................................................................................................... 8
1.3.2 Specific Objectives ............................................................................................. 8
1.4 Research Questions ................................................................................................... 8
1.5 Significance of the Study .......................................................................................... 9
1.6 Scope and Limitation of the Study ........................................................................ 9
1.7 Organization of the Study ...................................................................................... 9
CHAPTER TWO .............................................................................................................. 11
LITERATURE REVIEW ................................................................................................. 11
2.1 Introduction ............................................................................................................. 11
2.2 Overview of Liquidity Risk..................................................................................... 11
2.3 Theories of Liquidity Position Management ........................................................... 13
2.3.1 Commercial Loan Theory ................................................................................. 13
2.3.2 The Shift ability Theory of Liquidity ............................................................... 14
2.3.3 The Income Anticipation Theory ..................................................................... 15

v
 2.3.4 The Liability Management Theory ................................................................... 16
2.4 Quantitative Method for Measuring Liquidity Position .......................................... 16
2.5 Determinants of Liquidity Position: Empirical Evidence ....................................... 18
2.5.1 Firm Specific Determinants of Liquidity Position in Banking Sector.............. 19
2.5.2 Macroeconomic Determinants of Liquidity Position in Banking Sector ......... 26
2.6 Research Gap........................................................................................................... 31
2.6.1 Knowledge Gap ................................................................................................ 31
2.6.2 Methodological Gap ......................................................................................... 32
2.7 Conceptual Framework ............................................................................................... 32
CHAPTER THREE .......................................................................................................... 33
RESEARCH METHODOLOGY...................................................................................... 33
3.1 Research Design ...................................................................................................... 33
3.2 Data Type and Source ............................................................................................. 34
3.3 Population, Sampling Method and Sampling Frame .............................................. 34
3.3.1 Population of the Study .................................................................................... 34
3.3.2 Sampling Method ............................................................................................. 34
3.3.3 Sampling Frame ................................................................................................ 35
3.4 Methods of Data Analysis and Tools ...................................................................... 36
3.5 Description of Variables and Hypothesis Development ......................................... 36
3.5.1 Liquidity Position ............................................................................................. 36
3.5.2 Independent Variables ...................................................................................... 37
3.5.2.1 Bank Specific Factor .................................................................................. 37
3.5.2.2 External (Macroeconomic) Variable.......................................................... 40
3.6 Model Specification ................................................................................................ 44
3.7 Chapter Conclusion ................................................................................................. 45
CHAPTER FOUR ............................................................................................................. 46
RESULTS AND DISCUSSIONS ..................................................................................... 46
4.1 Descriptive statistics ................................................................................................ 46
4.2 Correlation Analysis ................................................................................................ 52
4.3 Testing Assumptions of Classical Linear Regression Model (CLRM)................... 53
4.3.1 Test for normality assumption (ut ∼ N (0, σ2) ................................................. 53

vi
 4.3.2 Test for homoskedasticity assumption (Var (ut) = σ2) ..................................... 55
4.3.3 Test for absence of serious multicollinearity assumption ................................ 55
4.4 Other Post-estimation Diagnostic Testing............................................................... 57
4.4.1 Model specification test .................................................................................... 57
4.4.2 Testing for panel cross-sectional dependence .................................................. 58
4.4.3 Panel Unit Root Tests ....................................................................................... 59
4.4.4 Testing for serial correlation............................................................................. 60
4.4.5 Summary of diagnostic testing ......................................................................... 60
4.4.6 Sargan test of over identifying restrictions ....................................................... 61
4.4.7 Arellano-Bond test for zero autocorrelation in first-differenced errors ........... 61
4.5 Regression Result and Discussion........................................................................... 62
4.6 Discussion on Results of System Dynamic panel-data estimation ......................... 64
4.7 Summary of Hypothesis Tested .............................................................................. 69
CHAPTER FIVE .............................................................................................................. 70
CONCLUSION AND RECOMMENDATION ................................................................ 70
5.1 Conclusion............................................................................................................... 70
5.2 Recommendations ................................................................................................... 71
5.2.1 Recommendation for Management of Commercial Banks in Ethiopia ............ 72
5.2.2 Recommendation for supervisory bodies of Ethiopian commercial banks
(NBE)......................................................................................................................... 74
5.2.3 Further Research ............................................................................................... 74
References ......................................................................................................................... 76
Appendices ........................................................................................................................ 80

vii
 List of Tables

Table 3.1 List of Sampled Commercial Banks ................................................................. 36

Table 3.2 Summary of Research hypothesis ..................................................................... 42
Table 3.3 Dependent variables: Proxies, significance and studies ................................... 43
Table 3.4 Independent variables: Proxies, significance and studies ................................. 43
Table 4.1 Summary of Descriptive Statistics for Variables .............................................. 47
Table 4.2 Correlation matrix of dependent (LIQDR) and independent variables ............ 52
Table 4.3 Results of Normality Test ................................................................................. 54
Table 4.4 Results for Heteroskedasticity Test...................................................................55
Table 4.5 Results for Multicollinearity Test ..................................................................... 56
Table 4.6 Results of Ramsey RESET test.........................................................................57
Table 4.7 Result for Model Specification Test ................................................................. 58
Table 4.8 Results of Pasaran CD test for cross-sectional dependence ............................. 59
Table 4.9 Results of second generation Pesaran’s CADF panel unit root test ................. 59
Table 4.10 Heteroskedasticity-robust Born and Breitung (2016) HR-test as post
estimation .......................................................................................................................... 60
Table 4.11 Sargan test of over identifying restrictions ..................................................... 61
Table 4.12 Arellano-Bond test for zero autocorrelation in first-differenced errors .......... 62
Table 4.13 Result of System dynamic and FGLS panel-data estimation...........................63
Table 4.14 Summary of Hypothesis Test .......................................................................... 69

viii
 List of Figures

Figure 2.1: Conceptual Framework ......................................................................... .........32

Figure 4.1:Trends of Real GDP growth rate from 2000 to 2021 ...................................... 50
Figure 4.2: Trends of Inflation rate from 2000 to 2021 .................................................... 51
Figure 4.3: Histogram of Residual .................................................................................... 54

ix
 List of Appendices

Appendix I: Data for Liquidity Position Model................................................................ 80

Appendix II: Results for Pooled OLS ............................................................................... 87
Appendix III: Results for Fixed Effect Model .................................................................. 88
Appendix IV: Results for Random Effect Model ............................................................. 89
Appendix V: Results for Correlation Analysis ................................................................. 90
Appendix VI: Results for Normality test .......................................................................... 90
Appendix VII: Results for Heteroskedasticity Test .......................................................... 91
Appendix VIII: Results for Multicollinearity Test ........................................................... 91
Appendix IX: Results for Model Specification Test (Link test) ....................................... 92
Appendix X: Results for Crosses Sectional Dependence ................................................. 92
Appendix XI: Results for Serial Correlation .................................................................... 92
Appendix XII: Results for Over Identifying Restrictions (Sargan test) ........................... 93
Appendix XIII: Results for First and Second Order Autocorrelation (Arellano-Bond) ... 93
Appendix XIV: Results for System Dynamic Panel-Data Estimation.............................. 94
Appendix XV: Results for Cross-sectional time-series FGLS regression………….…....95

x
 Abbreviations and Acronyms
ABB Abay Bank S.C
AWB Awash Bank S.C
ADB Addis Bank S.C
BRB Birhan Bank
BIS Bank for International Settlement
BOA Bank of Abyssinia S.C
BUB Bunna International Bank
CAR Capital Adequacy Ratio
CBE Commercial Bank of Ethiopia
CBO Cooperative Bank of Oromia S.C
CBS Commercial Banks
CLRM Classical Liner Regression Model
DAB Dashen Bank S.C
ENB Enat Bank S.C
FGLS Feasible Generalized Least Square
RGDP Real Gross Domestic Product
IM Interest Margin
INFR Inflation Rate
LNTA Natural Logarithm of Total Asset
LB Lion Bank S.C
LCR Liquidity Coverage Ratio
LIQDR Liquidity Asset to Deposit Ratio
LP Liquidity Position
LR Liquidity Risk
NBE National Bank of Ethiopia
NB Nib Bank S.C
NSF Net Stable Funding Ratio
ORB Oromia Bank S.C
OLS Ordinary Least Square
OPEFF Operational efficiency
RGDP Real Gross Domestic Product Growth
ROA Return on Assets
TANG Tangibility
UB United Bank
VIF Variable Inflation Factor
WEB Wegagen Bank
ZEB Zemen Bank

xi
 Abstract

Risk is a fundamental component of all financial institutions, and the banking industry in
particular. The purpose of this study is to identify the key factors that affect liquidity position in
commercial banks in Ethiopia. Based on a theoretical and empirical literature analysis, the paper
examines some macroeconomic and bank-specific factors that influence the liquidity position of
commercial banks in Ethiopia. The study used a sample of fourteen private and one public
commercial bank, and it covered the years 2000 to 2021 with unbalanced panel data from
audited financial statements. Liquidity position was represented by the ratio of liquid assets to
deposits. The higher the liquid asset to deposit ratio indicates the higher the commercial banks'
liquidity position and vice versa. The econometric package of STATA 14 software was used to
analyze the data utilizing the system dynamic panel data estimation method and correlation
analysis. According to the results of the system dynamic panel data estimation, the independent
variables that are distinctive to each bank have a substantial impact on the liquidity position of
commercial banks in Ethiopia. These independent variables include bank size, capitalization,
loan loss provision, and operational efficiency. Furthermore, the real GDP growth rate and
inflation rate have a substantial positive impact on the liquidity position of Ethiopian commercial
banks from the macro-economic factors. The findings of this study also suggest that profitability,
tangibility, and interest margin do not fully explain how liquidity position in Ethiopian
commercial banks is affected. To summarize, each commercial bank should use its own liquidity
position management policy as a framework to monitor its liquidity position, taking into account
changes in its size, capitalization, loan loss provision, operational efficiency. Furthermore,
commercial bank in Ethiopia also considers the level of real GDP growth rate and inflation rate
when formulating and implementing the liquidity position management policies.
Keywords: Liquidity position, Ethiopian Commercial Banks, Macroeconomic, Bank Specific.

xii
 CHAPTER ONE
INTRODUCTION

1.1 Background of the Study

The history of banking has been evolved back in 1905 in Ethiopia by establishing Bank
of Abyssinia in effect of the memorandum of understanding between king Menelik-II and
the National Bank of Egypt. After that, the National Bank of Ethiopia was established as
the Central Bank in 1963 by a proclamation No. 206/1963. Ethiopia is following on
restricted domestic banking with the cluster of nationalized commercial banks comprising
of public and private banks under the direct supervision of the National Bank of Ethiopia
at present, and the country has not yet opened their room for foreign banks to operate in
the economy, even if the country decides to do this in the near future. Additionally, the
state-owned commercial banks Commercial Bank of Ethiopia and Development Bank of
Ethiopia, which were both state-owned, dominated the financial sector in Ethiopia until
the end of the 19th century. Since the 1990s, the market has been open to private
companies, and as of late, Ethiopia is home to two government-owned banks and more
than 25 private ones (NBE, 2021).

The engine of economic growth in emerging market economies is generally the banking
industry. Commercial banks in particular contribute significantly to the economy by
performing intermediation tasks. Receiving funds from the public with surplus funds
involves accepting demand, time, and savings deposits or borrowing from the public or
other banks. These funds are then partially or fully used to disburse loans, advances, and
credit facilities as well as to invest funds in other ways (Asteway, 2017).

A key indicator of a nation's overall economic development is the financial sector's

stability and soundness. The state of the economy as a whole is largely reflected in the
health of commercial banks, which is in turn largely reflected in the health of their
borrowers (Das & Ghosh, 2007). A solid and stable financial system enhances investment
by facilitating the trade of goods and services, mobilizing savings, supporting profitable
business prospects financially, and efficiently allocating resources. Additionally, the 2008
financial crisis teaches us the value of creating an ideal level of liquidity to handle
1
challenging financial and economic conditions, according to Ferrouhi and Lehadiri
(2014). Umar and Sun (2015) also pointed out that the 2008 financial crisis showed that
the economy suffers when banks do not operate effectively. As a result, they draw the
conclusion that bank liquidity is crucial for the efficient operation of the economy. Since
a bank failure might have disastrous effects on the entire financial system, banking in
modern economies is often all about developing effective risk management procedures.
Recent periods of financial turbulence were mostly caused by faulty risk management
techniques, such as insufficient follow-up and management of liquidity position and
improper regulation of bank lending (Thiagarajan, et al., 2014).

In 21st century business environment is added multifaceted and intricate than ever. The
majority of businesses have to operate with uncertainties and qualms in every dimension
of their performance. Without a doubt, in today date, unpredictable and explosive
atmosphere leads all banks in front of several risks like: credit risk, liquidity risk,
operational risk, market risk, foreign exchange risk, and interest rate risk, along with
others risks, which may possibly hinder the survival and success of the banks (Ali, et al.,
2011).

The 2008 global financial crisis has motivated many researchers to re-examine and
reconsider the area of liquidity position after being determined of being one of the major
escalators of the observed financial contagion and the credit crunch during that period. A
well-managed financial institution should have a clear system of identification for
monitoring and controlling of liquidity risks, according to the majority of studies carried
out by various researchers. Achieving and maintaining the optimum level of liquidity in
the banking sector will help to achieve a high level of financial stability. Accordingly,
Basel III (2008) recognized the significance of holding an optimum level of liquid assets
and suggested that banks improve their liquidity position in order to meet their financial
obligations and reduce the risks encountered during times of crises in order to avoid
suffering significant financial losses.

Therefore, the primary priority of commercial banks is the creation of liquidity because it
is essential to their continued operation as a going concern company. It is clear that the

2
banking industry is crucial to a nation's economic development. This is accomplished by
linking economic units with surpluses and deficits. However, because of their central
position in the maturity conversion of short-term deposits into long-term loans, banks are
intrinsically exposed to liquidity position risk, which can damage both individual
institutions and the markets as a whole (Liza, 2018).

According to Milic and Solesa (2017), liquidity refers to a bank's capacity to make liquid
money available for withdrawal, to finance asset expansion and company operations, as
well as to deal with other anticipated and unanticipated financial requirements. In
addition, Belaid et al. (2016) defined liquidity position risk as the possibility that a bank's
customers would demand more cash than the bank has on hand in the form of cash and
other liquid assets. Bank in Settlement (2008) defined liquidity as a bank's capacity to
finance asset growth and pay commitments when they become due without sustaining
unacceptably large losses.

By utilizing one of the most crucial techniques for quantifying liquidity position, the ratio
of liquid assets to deposits, this research focuses on examining the factors that affect
liquidity position in Ethiopian commercial banks. According to the available empirical
literature, there are two groups of factors that determine the liquidity position. One body
of research points to internal variables as an avoidable or preventable predictor of banks'
liquidity position. The other, however, is an external element that has an overall impact on
all banks, such as macroeconomic factors and banking regulations.

The Liquidity position of Ethiopian commercial banks can be analyzed by considering

both state owned banks like Commercial Bank of Ethiopia and various privately owned
commercial banks. The trends of liquidity position of commercial bank of Ethiopia as
measured by Liquid asset to deposit ratio is just between 10% and 78% between the year
2000 and 2021, with average liquidity level of 36% with a minimum level of liquidity
around 10% in the year of 2015 and maximum level of 78% recorded in the year of 2007,
which suggests that there is some evidence of liquidity problem encountered by the bank
as 15 percent of liquid asset to net current liability ratio is the minimum liquidity position
maintained by banks consistently as per NBE directive No.SBB/57/2014. From figure 1.1
below, we can observed that the liquidity position of Commercial bank of Ethiopia shows
3
a dramatic fall in their liquidity position starting from the year 2010 and it is just between
around 10 % and 16% from the year 2015 and 2021, which suggests that there is an
indication of liquidity problem during the recent period.

The trends of liquidity position of commercial bank of Ethiopia is about 44% during the
year 2000, and which reduces to 35% in the year 2001. Furthermore, their liquidity
position scales up to 69% in the year 2003 and it consistently registered satisfactory result
until the year 2007. Starting from the year 2008, CBE shows a significant fall in its
liquidity position on a year on year basis as 47% registered in the year of 2008 and its
average liquidity position is around 19% for the thirteen consecutive years until the year
2021.

Similarly, the liquidity position of some selected private commercial banks also shows
many difficulties from period to period as discussed in the following paragraph. On
average, the liquidity level of those sampled banks in this study, as measured by the
liquid asset to deposit level is just around 20% with a minimum and maximum level of 13
% and 40 % respectively from the year 2015 up to 2021 respectively, which spots there is
some indication of liquidity risk faced by those banks which requires great attention to
achieve optimum liquidity level by identifying the significant internal and external
factors of liquidity position.

For example, the liquidity position of united bank as measured by their liquid asset to
deposit ratio was 46.05 % in the year of 2000 and it could reasonably maintained its
position until the year 2011 with an average value of 57.7% from the year 2001 to 2011.
However, starting from the year 2012 it shows dramatic fall in its liquidity position as it
was registered an average liquidity position of 23.2% for ten consecutive years from the
year 2012 up to the year 2021. On the other hand, Wegagen bank has registered 63.54%
of liquidity position in the year of 2000 and its liquidity position starts dramatic fall in the
year of 2015 as registered 24.79% and starting from this year the banks liquidity position
cannot lift out from those deteriorated level of liquidity as its average liquidity position
was around 23% for six consecutive years until the year of 2021. Similarly the liquidity

4
position of Nib bank S.c was around 111.54% during the year 2000 which is significantly
higher than other banks during the year under investigation and their liquidity position
dramatically fall to 44.23% in the following year of 2001. However, starting from the
year 2014 the bank shows further deterioration in its liquidity position as 24.18% was
registered in this year. Additionally, Nib bank cannot lift out from those deteriorating
level of liquidity position as 18.2% is their average liquidity position for seven
consecutive years from 2015 until 2021. To sum up, the same problem observed in other
private commercial banks in Ethiopia and the above illustration is just provided for the
purpose of giving some insight about the trends of liquidity position of some commercial
banks in Ethiopia.

1.2 Statement of the Problem

As was previously mentioned, the banking industry plays a significant role in an
economy by acting as a bridge between the economic system's surplus and deficit units
and by carrying out a variety of important tasks on both sides of the balance sheet. The
audited statement of financial position gives interested parties crucial information
regarding the level of liquidity of commercial banks in Ethiopia. For instance, the assets
side of a balance sheet comprises loans from banks in addition to current and fixed assets,
whereas the liabilities side includes, among other things, client deposits (Shah et al.,
2018).

The world economy has been significantly harmed by the global financial crisis since
2007, which started in the United States. Financial firms at the time wrote off losses of
billions of dollars, forcing them to cut staff, and the government also lost money by
subsidizing these financially troubled financial organizations. Identification and control
of liquidity position factors are crucial for the smooth operation of financial markets in
general and the banking industry in particular, as evidenced by the liquidity issues that
some banks had during this global financial crisis (Vodova, 2013).

Banks acting as intermediaries must effectively control the supply and demand of their
funds in order to conduct commercial operations securely and successfully and to build
strong relationships with important stakeholders. The danger to maintain large or low

5
bank reserves, the risk of a high interest rate, and the risk to harm banks' reputation and
goodwill can all result from improper liquidity management by the bank (Ismael, 2010).
Banks may experience a liquidity issue as a result of mismanaged cash or erratic
withdrawals by depositors during periods of bad societal economic conditions. The global
financial crisis of 2007–2008 also results from banks' failure to provide depositors and
other parties, such as borrowers, with liquidity. Therefore, in the current atypical external
influencing environment as well as the competitive economic contexts, proper
management of the liquidity positions in the banks are incredibly difficult or complex and
considerably important (Ahmed, et al., 2011).

According to NBE Report on January 2020, cheques worth close to Six billion Birr are
pending between banks, while most of the banks mention system failure as a reason for
not processing the cheques while being cash-strapped. In recent period, some banks
including the commercial bank of Ethiopia (CBE) restricted the maximum amount of
money a depositor can withdraw from their saving or other accounts. In line with this, the
Central Bank of Ethiopia, which is worried about the unusual liquidity crisis affecting
multiple financial institutions, wrote a letter to all of the banks on January 17, 2020
inviting them to apply for the loan from NBE as it is a lender of last resort. To this end,
the National Bank of Ethiopia (NBE) availed a nine-billion-Birr loan to cash strapped
commercial banks at a competitive bidding interest rate of nine percent, which is 4.5
percentage points lower the average lending rate of the commercial banks. On the same
date 12 banks out of the total commercial banks, applied for the loan, which has a
maturity period of one month.

The National Bank of Ethiopia requires banks to maintain liquid assets of not less than 15
percent of their current liabilities as per NBE directive No.SBB/57/2014. However, for
example, as discussed above if we consider state owned banks like commercial bank of
Ethiopia, their liquidity level as measured by liquid asset to deposit ratio ranges from
10% and 16 % from the year 2015 to 2021, which spots some practical liquidity problem
faced by those public banks. Similarly, the liquidity level of some private commercial
banks (united bank, Nib bank, Wegagen bank, Birihan bank…) ranges from 13 % and

6
40% from the year 2015 to 2021, which again requires great attention for achieving and
maintaining optimum liquidity level.

According to the aforementioned, if Ethiopia's banking industry lacks robust mechanisms

for identifying and monitoring liquidity position factors, it is only a matter of time before
a bank faces bankruptcy risk, which ultimately causes the collapse of the entire economy,
just as it did in Europe and America. Therefore, it is crucial for the stability of the
banking sector as well as the economy as a whole to do research on the variables that
affect the liquidity position in the case of Ethiopian commercial banks. The fundamental
driver of this interest is the possibility of systemic contagion and financial distress
spreading throughout the whole financial institution if there is a liquidity constraint at one
of the so-called "too big to fail" banking institutions. Additionally, a bank with an
adequate degree of liquidity will be able to fulfill its obligations, even during challenging
circumstances like bank runs. According to this viewpoint, having the right amount of
liquidity lowers the likelihood of bankruptcy, which might lower funding costs and boost
bank profitability (Belaid et al., 2016).

Several researchers, like Akhtar et al. (2011),, et al. (2016), Cucilenni, (2013), and
Yaacob, et al. (2016), have investigated the factors that influence liquidity position, but
few investigations, such as those by Samuel and Tseganesh (2015) and in Ethiopia
(2012), have been made. Since their goal was to research subjects like the relationship
between liquidity and financial profitability of banks, the majority of them neglected to
study the determinants of banks' liquidity position directly. Few studies have been done
on the determinants of liquidity position in Ethiopian commercial banks because studies
like Tirualem (2009) and Tsion (2015), which study the liquidity risk management
practice using primary data (questionnaires and interviews), ignore the determinants of
liquidity position among commercial banks in Ethiopia. As a result, the researcher
concluded that while there has been considerable research on liquidity risk management,
the connection between liquidity position and financial performance, and the like, there
has been relatively little in-depth research on the factors that directly influence liquidity
position across nations. Despite the fact that this is a current issue for many financial
institutions, notably commercial banks, the researcher identified few papers on the topic
7
of the determinants of liquidity position in Ethiopian setting. These and other causes drive
the researcher to investigate these internationally pressing financial concerns with a focus
on Ethiopian commercial banks.

In addition, from those few studies, most of them made focuses on sample of only private
commercial banks and ignoring the state owned banks like Commercial Bank of Ethiopia.
Therefore, this study tries to fill those gaps identified by including some young banks and
state owned banks like commercial bank of Ethiopia. The primary liquidity position
determinants have been objectively explored in this study because the financial sector
variables are affected by a variety of internal and external sources. These factors have
been divided into macroeconomic variables and ones particular to banks. These made it
possible to understand the situation surrounding the liquidity position of different
commercial banks in Ethiopia, given that this study covered a number of them, as well as
to identify the clear-cut factors that determine how such risks are managed in Ethiopian
commercial banks.

1.3 Objectives of the Study

1.3.1 Main Objective
Finding the factors that influence liquidity position in Ethiopian commercial banks is the
study's main goal.

1.3.2 Specific Objectives

The study specifically focuses on the following goals:

- Analyzing the extent of Ethiopia's commercial banks' liquidity condition.

- Determining the bank-specific factors that affect Ethiopian commercial banks' liquidity
position.

- Examining the macroeconomic factors that affect Ethiopian commercial banks' liquidity
position.

1.4 Research Questions

The following research questions are expected to be answered after completing the
research
8
 - How does the current liquidity position of Ethiopian Commercial banks?
- What are the significant bank specific determinants of commercial banks liquidity
position in Ethiopia?
- What are the significant macroeconomic determinants of commercial banks
liquidity position in Ethiopia?

1.5 Significance of the Study

Beyond serving an academic purpose, the study also made the following contributions: it
helped Ethiopian commercial banks identify the key variables that determine their
liquidity position, pay attention to variables that significantly affect liquidity level, and
take timely corrective action. Additionally, it offers useful data to the regulatory body
(NBE) for the formulation of policies relating to the liquidity position of commercial
banks. Finally, it will be used as a resource by other researchers who perform additional
research on the same or a similar subject.

1.6 Scope and Limitation of the Study

The study's focus has been narrowed to the evaluation of the macroeconomic and bank-
specific factors that affect the liquidity position of Ethiopian commercial banks registered
with the NBE. One state-run commercial bank and fourteen privately owned commercial
banks with operations in Ethiopia are included in the study's scope. For seven banks that
began operations prior to 2000, namely Commercial Bank of Ethiopia, Awash Bank,
Bank of Abyssinia, Dashen Bank, Wegagen Bank, United Bank, and Nib Bank, the
researcher took into account twenty-two fiscal years, or from 2000 to 2021. For the
remaining banks, unbalanced panel data were taken into account from the time that they
began operations until the year of 2021.Only six bank-specific and three macroeconomic
drivers are covered in this study because it is impossible to include all elements that
affect liquidity position in one study.

1.7 Organization of the Study

Five chapters make to the organization of the research. The background of the study,
problem statement, research objectives, research questions, significance of the
investigation, scope and limitation of the study, and organization of the research report
9
are all covered in the first chapter. The review of both theoretical and empirical literature,
gap analysis, and the creation of conceptual framework are the main topics of the second
chapter. The third chapter covers the study's methodology. The results and discussion are
covered in the fourth chapter. The study's conclusion and recommendations are the main
topics of the final chapter.

10
 CHAPTER TWO

LITERATURE REVIEW

2.1 Introduction

According to several literature sources, risk is an unanticipated event that has financial
repercussions that lead to loss or decreased revenues. Due to the unpredictability or
uncertainty of future trading activity, any action that has the potential to generate profits
or cause losses may be referred to as a risky venture. To put it another way, it can be
described as the unpredictability of the outcome of the future (Driga, 2012).

According to Sahlemichael (2009) and the Basele Committee on Banking Supervision,

risk is an unpleasant reality in any organization and, if not adequately detected and
managed, could have a negative impact on a company's ability to continue operating.
However, the harm might be more severe in the case of banks because the banking
industry involves not just the owners but also depositors, other banks, and ultimately the
entire economy. Credit risk, market risk, liquidity risk, and operational risk are the main
threats to banks.

Commercial banks are for-profit businesses that function as a middleman between

borrowers and lenders, luring temporarily excess funds from corporate and individual
clients and then disbursing loans to those in need of capital. From this angle, banks carry
out their functions while taking on a variety of risks in order to deal with the money that
belongs to people and other businesses. The possibility that a loss will occur is, thus, a
measure of risk, and for many firms, risk is a factor in all decisions. In fact, weighing
prospective risks and profits is a fundamental aspect of business decision-making (Driga,
2012).

2.2 Overview of Liquidity Risk

According to Bank in Settlement (2008), liquidity describes a bank's capacity to finance

asset growth and fulfill commitments as they become due without suffering intolerable
losses. According to Yaacob et al. (2016), liquidity risk can also be seen of as a part of

11
market risk that has to do with a bank's ability to meet depositor demands for money
withdrawals. Liquidity risk is a risk related to an investment that has limited
marketability or that is difficult to sell in order to safeguard against potential loss. When
certain investments with a focus on liquidity start to feel the pinch of a subordinate
composition against these investments, this is referred to as an investment subject to
liquidity risk. The entire company is compelled by liquidity risk to make alternative
decisions rather than turning investments into cash. As a result, by using scientific
techniques, liquidity risk components can be thoroughly explored (Iqbal et al., 2015).

The primary function of banks in the maturity transformation of short-term deposits into
long-term loans leads to liquidity risk. There are two different kinds of liquidity risk in it:
market liquidity risk and funding liquidity risk. The risk associated with funding liquidity
develops when the bank is unable to efficiently meet both anticipated and unforeseen
current and future cash flow needs and collateral requirements without impairing either
normal business operations or the firm's financial position. When a bank finds itself
unable to quickly offset or exit a position at market price due to insufficient market depth
or market disruption, market liquidity risk is present (Vodova, 2013).

Additionally, an asset's liquidity relates to how easily it may be converted into cash or a
cash equivalent asset. The inability to sell an asset right away without suffering large
losses causes liquidity risk. The danger of not being able to fund its portfolio of assets at
the right maturities and rates as well as the risk of not being able to liquidate a position
promptly at reasonable market prices both constitute liquidity risk for the financial
industry, particularly the banking institution. It can be described in terms of unpredictable
timing of cash inflows and cash outflows from the business operation, but it can also be
defined as a maturity mismatch between assets and liabilities (Ali, 2004).

A bank must be able to maintain appropriate liquidity since a lack of liquidity when
fulfilling obligations to depositors, other banks, and financial institutions can have
negative effects on the bank's reputation and share prices in the money market. A bank
run, which occurs when depositors rush to take their money from a bank, can
occasionally be caused by a liquidity risk (Perez, 2014). Regulatory and supervisory
organizations as well as international financial stability assurance entities have focused
12
on establishing and maintaining improved bank liquidity risk management practices. The
Basel III framework, which takes into account the vulnerability that led to the financial
crisis by increasing bank capital and liquidity levels with the aim of having a more stable
banking sector, has offered one of the most recent regulatory reforms (Belaid, 2016).

Numerous factors influence a bank's own liquidity, which in turn influences the amount
of liquidity they can generate. The balance between liquidity risk and liquidity
production, or a bank's liquidity management technique, is affected by these elements to
varying degrees. The fundamental factor in balancing a bank's liquidity risk and liquidity
position is its assets and liabilities. All of a bank's funding sources make up the bank's
liabilities. Deposit accounts, lent funds, and long-term funds are the three main sources of
funding for banks. How much liquidity risk a bank is exposed to and how much liquidity
it can generate are both directly impacted by the quantity and sources of cash. The less
risk a bank exposes itself to, the more readily it can access cash, and the more money it
can hold on to, the more liquidity it can generate. Banks need liquidity to cover expected
and unforeseen balance sheet swings as well as to offer capital for expansion (Tsion,
2015).

2.3 Theories of Liquidity Position Management

Either the management of assets or liabilities is the foundation of the theory of liquidity
management. Three theories have been produced on the management of assets, such as
the commercial loan theory, the shift ability theory, and the projected income theory,
whereas one theory has been developed on the management of liabilities (Shodhganga,
1997).

2.3.1 Commercial Loan Theory

This theory, developed by Adam Smith in 1776, is also known as the traditional or real
bills doctrine theory. According to this view, banks should avoid making long-term loans
and should only use short-term, self-liquidating productive loans and real bills as their
primary earning assets. It also requires the central bank to lend money to the bank by
using these assets as collateral with every short-term self-liquidating loan. As a result, the
bank is assured an adequate level of liquidity. According to Mugenyah (2015), rigorous

13
adherence to this theory would ignore long-term loans, which are crucial for funding
sizable investments. As a result, economic growth would be constrained.

Furthermore, according to this theory, commercial banks' liquidity would be guaranteed

as long as assets were held in short-term loans that would be repaid as part of routine
company operations. Banks are obligated to provide financing for the flow of
commodities through the subsequent production and consumption cycles, or for what are
known as working capital loans or inventories today. The hypothesis also predicts that
repayment from the bank's self-liquidating assets would be sufficient to ensure a superior
liquidity situation. This explanation ignored the possibility that transitory deposit
withdrawals and granting credit requests could have a negative impact on the generation
of liquidity. Additionally, the theory fails to take into account the typical stability of
demand deposits when taking liquidity into account. The bank's liquidity position could
eventually be impacted by this obvious viewpoint (Tsion, 2018).

2.3.2 The Shift ability Theory of Liquidity

Harold G. Moulton created this idea in 1915 to displace the conventional one. According
to this hypothesis, banks might safeguard their liquidity positions in the event of heavy
deposit withdrawals by holding credit instruments like commercial papers and treasury
bills. These securities might be easily traded on secondary markets without suffering
substantial capital losses. This theory's primary flaw is that it ignores the scenario of a
severe financial crisis in which every bank wants to sell their financial assets, which
results in a loss of market (Casu et al., 2006).

This notion is also supported by the fact that the assets owned by banks are either to be
sold to other lenders or investors or transferred to the central bank, which is always
prepared to buy assets that are offered for sale. If a commercial bank has assets to sell, it
can meet its clients' liquidity demands (Mugenyah, 2015). According to the shift ability
idea, banks could retain credit instruments with readily accessible secondary markets as a
sort of liquidity reserve to protect themselves against unexpectedly large deposit
withdrawals the most efficiently. Commercial paper, prime bankers acceptances, and, as
it turned out, Treasury bills were among the financial assets included in this liquidity

14
reserve. All of these securities, in a normal situation, would satisfy the marketability and
capital certainty requirements due to their short durations to maturity. The fast increase in
the volume of short-term U.S. Government liabilities throughout the 1930s and 1940s
improved the shift ability theory (Diamond, 1984). Because we lack capital markets to
raise money by selling financial instruments there during times of cash shortage, it is
challenging to apply the shift ability theory of liquidity in Ethiopia. As a result, the ease
with which those commercial assets can be converted into cash without suffering
significant losses is reduced.

2.3.3 The Income Anticipation Theory

According to Herbert V. Prochnow's income anticipation theory, which he devised in
1949, it is possible to pay off a long-term debt by setting up regular monthly installments.
These payments are determined by estimating the customer's future earnings and
evaluating his creditworthiness. This hypothesis outweighs the others because it
constantly ensures a high degree of safety and liquidity throughout the functioning of the
bank.

Furthermore, according to this theory, if scheduled loan repayments are made from the
borrower's future income, liquidity can be increased and assured. This theory links loan
repayment to income rather than relying only on security. According to this hypothesis, a
bank's liability may be influenced by the maturity distribution of its loan and investment
portfolios. The idea also took into account the fact that some loan classes are more liquid
than others (Ibe, 1991). According to this hypothesis, a bank's ability to monitor its
liquidity can be determined by the right phasing and structuring of the loan payments it
makes to its clients. As a result, if a customer bases the projected loan payments on the
borrower's future, the liquidity can be managed appropriately. Nzotta (1997) asserts that
the theory places a premium on a borrower's earning potential and credit quality as the
strongest guarantees for obtaining adequate liquidity. Ibe (1991) asserts that the theory
takes into account banks' shift toward self-liquidating commitments. Many commercial
banks have adopted a ladder effect in their investment portfolio as a result of this
hypothesis. If this theory is taken into account in the context of Ethiopia, commercial

15
banks can also preserve and enhance their liquidity position by planning amortized loan
repayment, which also enhances asset quality by reducing non-performing loans.

2.3.4 The Liability Management Theory

The liability management theory advocates that banks can satisfy their liquidity
requirement by bidding in the market for additional funds to meet loan demand and
deposit withdrawal. The fundamental ideas of the theory can be linked to the revival of
federal funds markets in the 1980s and the emergence of negotiable time deposits as a
significant money market instrument in the United States. Banks in the United States rely
on the federal funds market, the euro-dollar market, or the sale of loan participation
certificates for liquidity. Liability management is the term given to this type of borrowing
(Mugenyah, 2015). Since banks can arrange and negotiate interbank loans and direct
loans from NBE during times of operational liquidity shortfall, this idea is also valid in
the Ethiopian setting.

2.4 Quantitative Method for Measuring Liquidity Position

A financial organization, especially a bank, can use a variety of resources to meet its
liquidity needs, including new deposits, loans that are nearing maturity, borrowed funds,
and/or borrowing from the central bank through the discount window. As a result, most
commercial banks consider access to and management of liquidity to be crucial activities.
It is important to address the sources of liquidity position as well as potential
management strategies before looking at the methods for monitoring it.

There are three main sources of liquidity risk, according to Rochet (2008): on the liability
side, there is significant uncertainty regarding the volume of withdrawals of deposits or
the renewal of rolled-over interbank loans, particularly when the bank is at risk of
insolvency or when there is a general liquidity shortage. On the asset side, there is
uncertainty regarding the quantity of future loan requests that a bank will get, as well as
off-balance sheet activities such credit lines and other commitments, and positions made
by banks on derivative markets.

Aspach et al. (2005) state that there are a few mechanisms that banks can employ to
protect themselves from liquidity crises. First, banks build up a reserve of liquid assets on
16
the asset side of the balance sheet, such as cash, balances with other banks and central
banks, government-issued debt securities, and similar securities, or reverse repo trades, to
lessen the likelihood that the bank's viability will be threatened by liquidity demands. The
second mechanism has to do with the liabilities side of the balance sheet and involves
banks borrowing from other banks on the interbank market when they need liquidity.
However, the risk of market liquidity is closely related to this method. The final
technique dealt with the liability side of the balance sheet, where the central bank often
serves as a lender of last resort (LOLR) to give emergency liquidity assistance to specific
illiquid institutions and to supply aggregate liquidity in case of a systemic liquidity
shortfall.

In line with this, liquidity position of banks can be measured by using two basic
approaches. The difference between assets and liabilities at both present and future dates
is referred to as the liquidity gap or flow approach in the first method. A surplus between
assets and obligations at any time represents a hole that needs to be closed (Bessis 2009).
The second approach related with the Liquidity ratios or stock approach includes various
balance sheet ratios, which should identify main liquidity trends. These ratios indicate the
fact that bank should be sure that appropriate, low-cost funding is available in a short
time. This might involve holding a portfolio of assets that can be easily sold, which
includes cash reserves, minimum required reserves or government securities, holding
significant volumes of stable liabilities (especially deposits from retail depositors or fixed
time deposits) or maintaining credit lines with other financial institutions.

In conclusion, the stock approach looks for trends in liquidity using several balance sheet
ratios. The bank evaluates its liquidity position by analyzing the variability in inflows and
outflows to calculate the amount of reserves that will be required throughout a period of
time using the flow approach, which treats liquid reserves as a reservoir. The flow
strategy is more data driven and there is no established method to forecast inflows and
outflows, despite the fact that both approaches are conceptually appealing. Because of
this, the stock techniques are more widely used in academic research and practice (see
Crosse and Hempel 1980; Yeager and Seitz 1989; Hempel et al. 1994; Vodova 2011).

17
The loan-to-deposit ratio and the liquid asset to total deposit ratio are the two most
common stock ratios, according to Crosse and Hempel (1980). The higher the loan-to-
deposit ratio and the lower the liquid asset to total deposits ratio, respectively, the less
able a bank is to meet any additional loan demands or unexpected withdrawals of funds
by depositors. The loan-to-deposit ratio does not show other assets that can be converted
into cash to meet withdrawal or loan demands, whereas the liquid assets ratio does not
take into account the flow of money from loan repayments, increases in liabilities or
deposits, and the demand for bank funds. Thankfully, the ratios frequently move in
unison (Crosse and Hempel 1980). As a result, the stock approach's liquid asset to deposit
ratio was employed for the purposes of this study.

2.5 Determinants of Liquidity Position: Empirical Evidence

There are two types of risks that firms must deal with: financial risks and non-financial
risks. To safely operate any business, it is imperative to consider both of these categories
of risks (Ali et al., 2011).Commercial banks are for-profit businesses that function as a
middleman between borrowers and lenders, luring in seasonal resources from company
and individual clients and disbursing loans to individuals in need of money. From this
angle, banks carry out their operations by absorbing a variety of risks related to the
money held by individuals and other businesses. Therefore, risk is the possibility that a
loss will occur, and it affects every business action. In fact, weighing prospective risks
and profits is a fundamental aspect of business decision-making (Driga, 2012).

The banking industry's liquidity concerns are impacted by a number of variables. It can
be classified as a bank specific (diversifiable risk or unsystematic) and macroeconomic
(systematic risk), as shown by a variety of literatures, including Vodova (2013) and
Waemustafa & Sukri (2015). Determinants of liquidity position particular to a bank
include things like financial performance, loan loss provision, capitalization,
tangibleness, operational efficiency, and size of the bank relative to total assets.
Furthermore, macroeconomic factors like interest margin, inflation rate, real GDP as a
gauge of economic development, currency rate, lending interest rate, and financial

18
stability (financial crises in 2009 & 2010) heavily influence a company's liquidity
position.

To summarize, Woods and Dowd (2008) identified two primary categories of risks that
have an impact on a company's cash flows and cost of capital:

1. Factors unique to a given firm: These elements contribute to diversified or

unsystematic financial risk. These risks, such fraud, fire, and lawsuits, are specific
to the business's particular activity. By putting in place suitable internal controls
and other risk management techniques, the organization can reduce various
sources of these risks.
2. Market-wide or systematic risk: this kind of risk refers to unavoidable hazards
that are quantified by beta equity. It results from macroeconomic factors that have
an impact on every aspect of business, such as changes in interest rates, the rate of
real GDP growth, currency rates, and commodity prices. These risks can be
controlled using a variety of strategies, including the use of derivative contracts
and other financial risk management instruments.

2.5.1 Firm Specific Determinants of Liquidity Position in Banking Sector

Previous research by other scholars revealed a number of bank-specific factors. However,

the actual investigations carried out by various academics show distinct connections
between them and bank liquidity. These factors, which include Bank Size, Loan Loss
Provision, Capitalization, Financial Performance, Operational Efficiency, and
Tangibility, are primarily linked to the financial conduct and organizational structure of
banks.

i) Bank Size
Numerous researches on liquidity position have been undertaken, and these studies
attempt to relate bank size to liquidity position. In these studies, the size of the bank can
be regarded as bank specific parameters. A number of researchers, including Gilbert
(2013), Belaid, et al. (2016), Akhtar, et al. (2011), Abdulganiyy, et al. (2017), Cucinelli
(2013), Ahmed, et al. (2011), Iqbal, et al. (2015), Shaikh (2015), and Vodova (2013), use
bank size as a measure of liquidity position and attempt to relate it in some way to bank
19
size. The subsequent paragraphs below attempt to demonstrate a link between liquidity
risk and bank asset size using the findings of several studies.

The effect of bank size on bank liquidity positionhas been a hotly contested topic in
earlier literature. According to the "too big to fail" theory, a bank's size might negatively
affect its liquidity position. Some banks, according to (Lucchetta, 2007), believe they are
too large to fail, which deters them from maintaining higher liquidity ratios. In other
instances, certain sizable banks may not wish to expand their liquidity level since they,
like the Commercial Bank of Ethiopia, promise various forms of financial support in the
event of financial hardship. They view themselves as being too large, and they are aware
that the government has a strong incentive to shield them from failure, which would be
detrimental to the overall financial and economic health of the nation.

Additionally, a lot of academics have looked into the detrimental impact of bank size on
liquidity level. According to Vodova (2013), who examined the factors influencing
liquidity in the Hungarian banking industry between 2001 and 2010, the size of banks is
inversely related to liquidity levels. This finding—that bank size has a detrimental effect
on the liquidity situation of commercial banks—was further corroborated by studies by
Cucinelli (2013) in the context of European banks, Mugenyah (2015) in Kenya, and
Bunda (2008) in developing economies.

Bank size is determined by the natural logarithm of total assets in studies on the factors
that determine liquidity position. Understanding how risks change when a company
grows requires taking into account a number of factors, including bank size. Due to its
impact on both costs and how easily liquidity can be accessed, a bank's size can have a
significant impact on its position (Ali et al., 2011).According to Estrada (2011), as a
company grows, it may go bankrupt. However, as a company grows, it will allow them to
manage risk, similar to how economies of scale work. This is warranted because a
developing company frequently relies on debt financing, which increases risk and
ultimately endangers both the creditor and lender. Collapses can result from expanding in
size and accruing more debt. Additionally, vulnerabilities in the chain may result in
greater maintenance costs for industrial and financial institutions when the scale is

20
depending on the consistency of networks, such as huge supply chains. Therefore,
expanding in size may likewise be linked to rising hazards.

Ahmed, et al. (2011) conducted an empirical study on liquidity risk and Islamic banks in
Pakistan as part of their research on the factors that influence liquidity risk. The study
evaluates the factors that affect liquidity risk during a four-year period, from 2006 to
2009, and takes six Pakistani banks that offer services into account as pure Islamic banks.
The analysis comes to the conclusion that the association between bank size and liquidity
position is negligible. The size of the bank had a negligible negative impact on liquidity
position, according to Mugenyah (2015), who studied the factors that determine liquidity
risk in Kenyan commercial banks. In addition, Iqbal et al. (2015) conducted a study on
the impact of liquidity risk on firm-specific characteristics in the context of Pakistani
Islamic banks. Four Islamic banks are included in the study, which spans the thirteen-
year period from 2001 to 2013 and was examined using multivariate regression analysis.
The study also found a substantial positive association between bank size and liquidity
position.

In contrast, Shen et al.'s (2009) study on bank liquidity risk and performance in 12
advanced economies (Australia, Canada, France, Germany, Italy, Japan, Luxembourg,
Netherlands, Switzerland, Taiwan, United Kingdom, and United States) found that size
and liquidity risk have a significant positive relationship, whereas size and liquidity
position squared has a significant positive relationship. Researcher’s claim that major
banks believe they are too big to fail based on this finding. As a result, they are motivated
to take greater risks and hold more loans, which have left them with a substantial money
shortage when trying to finance an increase in their loan portfolio. Beyond the optimal
point, however, the impact of size is detrimental. Additionally, a number of other
academics have discovered that a bank's size has a favorable impact on its liquidity
status. Small banks have lower liquidity ratios because they concentrate more on
conventional intermediation and transformation activities (Chagwiza, 2014; Moussa,
2015).As a result, the relationship between size and liquidity positionis non-linear, which
suggests that, as was previously noted, there is debate among scholars about the
relationship between size and liquidity position.
21
 ii) Loan loss provision
When lending money, banks run the risk that the money won't be repaid in full if the
borrower defaults. A bank will make a charge to the profit and loss statement (a
"provision") to build a loan loss reserve that is reported on the balance sheet when a loan
loss becomes likely. The loan balance is decreased through a charge to the loan loss
reserve when the entire principal and interest amount of the loan becomes uncollectible
(Angklomkliew, et al., 2009).

In order to analyze the type of relationship that exists between liquidity position,
measured with the liquidity coverage ratio (LCR) and the net stable funding ratio
(NSFR), and some specific bank structure variables, Cucinelli (2013) studied the
determinants of bank liquidity risk within the context of the Euro area. The sample
consists of 1080 public and unlisted Euro zone banks, and OLS regression based on panel
data is the analysis methodology used. According to the results of the OLS analysis, loan
loss provision has a positive, significant influence on liquidity risk as measured by the
LCR, but a negative, minor influence on liquidity risk as evaluated by the NSFR.

Thong (2013) conducted yet another study on the variables influencing the liquidity risk
in the system of Vietnamese commercial banks. The study includes annual reports from
27 Vietnamese commercial banks for the years 2002 through 2011. Liquidity risk was
calculated using financing gap in the study. Finally, the regression analysis shows that
there is a negative but small correlation between the loan loss provision ratio and the
liquidity position.

iii) Capitalization
Capital adequacy ratio (CAR) is an indicator of the equity level in the banking sector.
Two hypotheses can be developed to explain the relationship between a bank’s CAR and
liquidity position.

On the premise that the capital level has a strong favorable impact on bank liquidity, the
first hypothesis can be constructed. A study by Mugenyah (2015) that found a high
capital adequacy ratio is a reliable predictor of a bank's stability and liquidity situation
lends support to this idea. Furthermore, Repullo (2004) has supported this finding by
22
showing that raising capital levels will encourage banks to enhance liquidity levels and
reduce liquidity risk. These conclusions can be supported by the fact that the bank's
capital adequacy ratio comprises reserves and equity capital that can be used to offset any
liquidity problems that may arise in the course of business.

Furthermore, El Khoury's (2015) research on the factors affecting liquidity in the

Lebanese banking sector, which used data from 23 commercial banks between 2005 and
2013, supported the risk absorption hypothesis and showed that capital level affects both
the ratio of liquid assets to total assets and the ratio of liquid assets to customer deposits
in a way that is both positive and statistically significant. Similar to this, a 2008 study by
Bunda and Desquilbet on the factors influencing bank liquidity risk in 36 emerging
nations found that high capital levels have a favorable effect on the liquidity level of
1107 banks. Furthermore, the study by Ferrouhi, E., and Lahadiri, A. (2014) supports the
notion that banks with stronger capital adequacy face a lower exposure to liquidity risk.
The ratio of liquid assets to total assets and a bank’s capital adequacy level has been
proven to be positively correlated by numerous previous researches (Vodova, 2013).

Additionally, Vodova (2013) included both bank-specific and macroeconomic variables

in a study on the factors affecting commercial banks' liquidity in Hungary from 2001 to
2010. The study discovered that the conclusion that determines capital adequacy has a
considerable favorable impact on liquidity. As anticipated, bank liquidity increases as
capital adequacy levels rise; solvent banks are also liquid. The capital adequacy ratio in
conventional banks has a significant positive relationship with liquidity risk at a 90%
level of confidence, whereas it has a positive but insignificant relationship in the case of
Islamic banks, according to a study by Ali et al. (2011) on the management of liquidity
risk.

A greater capital level may discourage banks from creating liquidity by making the
capital structure of banks more brittle, according to the second hypothesis, which was
derived by examining the relationship between capital level and bank liquidity. In light of
this, it appears that bank liquidity and capital adequacy level are negatively correlated.
The second theory has been supported by a number of research that show that banks'

23
CAR has a detrimental impact on bank liquidity as indicated by the ratio of liquid assets
to total deposits (Chagwiza, 2014; Moussa, 2015).

Ayele (2012)'s study, which Mugenyah (2015) cites, claims that capitalization as
measured by capital adequacy can be utilized as a sign of a bank's financial health in
terms of its capacity to pay operational expenditures and maintain fund liquidity.
According to the study, a crucial factor of liquidity risk is the regulatory requirement on
the minimum capital that banks must keep, as well as the ratio of core capital to total
assets as a measure of capital adequacy. A bank's capacity to take on new clients is also
gauged by capital sufficiency. Financial flexibility is provided by capital size for banks
and other financial institutions. In numerous research, including those by Abdulganiyy, et
al. (2017), Cucinelli (2013), Mugenyah (2015), and Vodova (2013), capital adequacy is
seen as a bank-specific indicator of liquidity risk.

Additionally, Abdulganiyy, et al.'s (2017) study on the factors influencing liquidity risk
in the dual and fully Islamic banking sector, which included information from Malaysia
and Sudan, produced conflicting findings. Using panel data analysis and Ordinary Least
Square Regression Analysis (OLS), three banks from each of the countries between 2004
and 2015 were chosen for the study. The authors draw the conclusion that the various
environments in which Islamic banks operate affect the significance of liquidity risk
determinants. The study's final finding is that capital sufficiency has a minimally
favorable impact on liquidity risk in Sudanese Islamic banks but a large negative impact
on liquidity risk in Malaysian Islamic banks.

A sample of 1080 listed and unlisted Euro zone banks were used in Cucinelli's (2013)
study on the factors that influence bank liquidity risk in the Euro area. OLS regression
based on panel data was used for the analysis. According to the study's findings, capital
adequacy and liquidity risk are positively related although not significantly so. As a
result, capitalization has a non-linear effect on liquidity risk, which is analogous to the
situation with bank size.

24
 iv) Profitability
Profitability is a manipulated measure of how effectively an institution could use
resources from its core business and turn a profit. It is an all-encompassing indicator of
an institution's long-term financial stability. The literature has taken into account the link
between financial performance and bank liquidity, as illustrated below.

Molyneux and Thornton (1992) found a negative correlation between liquidity in the
banking industry and financial performance as determined by return on asset (ROA). This
is because retaining liquid assets will result in a loss of opportunity for banks because
they will have lower returns than other sorts of hazardous long-term investments. A little
more, According to Hackethal et al. (2010), who investigated the factors affecting bank
liquidity in Germany's state-owned savings institutions and discovered that the degree of
financial performance had a sizable detrimental effect on bank liquidity. The detrimental
effect of performance on bank liquidity has been supported by numerous other research,
including those by Melese and Laximikantham (2015), Vodova (2011), and Moussa
(2015).

The financial success of banks and their liquidity, however, have not been found to
significantly correlate by numerous other academics, including Vodova (2013). In
contrast, Iqbal et al.'s (2011) study on the impact of liquidity risk on financial
performance in the context of Islamic banks in Pakistan aimed to create a model and
assess the association between these variables through its analysis. The conclusion was
that, at a 1% level of significance, financial performance as determined by return on asset
had a favorable and significant impact on liquidity risk.

Similar to this, a study by Ahmed et al. (2011) on the factors influencing liquidity risk in
Islamic banks in Pakistan identified profitability as an explanatory variable, assessed in
terms of return on asset, and found no statistically significant relationship between
profitability and liquidity risk. In addition, a comparison of the management of liquidity
risk between conventional and Islamic banks in Pakistan by Akhtar et al. (2011), which
took into account the years 2006 through 2009, found that, while insignificant in
conventional banks, liquidity risk has a significant positive relationship with return on
assets in Islamic banks with a 90% confidence level. As many academics in various
25
countries have found, the effect of financial performance on the liquidity risk of
commercial banks is not definitive.

v) Operational efficiency
These metrics can be used to evaluate the efficiency with which banks internally manage
non-interest income and expense to mitigate risk exposure (Gilbert, 2013). This variable
will show how effectively a bank manages its internal asset and liability workouts in
order to effectively manage their risk exposures.

According to Shakih (2015), who conducted a study on the estimation of liquidity risk
drivers by using panel data in the instance of five Islamic banks in Pakistan by taking into
account seven years' worth of data from 2007 to 2013, the study employed the ratio of
total cash to total asset for measuring liquidity risk. With a p-value of 0.058 and a 90%
level of confidence, the study's panel data technique, specifically the fixed effects
regression model, revealed that operational efficiency had a substantial impact on
liquidity risk.

In addition, Moussa (2015) conducted research on the factors affecting bank liquidity in
Tunisia for 18 chosen banks between the years of 2000 and 2010 using both static panels
and panel dynamic methods. The study's regression analysis demonstrates that
operational efficiency, as expressed by the operating expense to total asset ratio, has a
considerable detrimental effect on the liquidity of Tunisian banks. These results can be
explained by the fact that if banks' operational efficiency rises, their operating expense to
total asset ratio falls, their liquidity position improves and, as a result, their liquidity risk
is decreased. As a result, according to the results of the research mentioned above, it is
expected that operational efficiency will have a favorable impact on the liquidity situation
of Ethiopian commercial banks. This expectation can serve as the study's hypothesis.

vi) Tangibility
Tangibility is measures of how much of the bank’s total assets are fixed assets. The
proportion of fixed assets to total assets is the yardstick by which tangibility is measured.
According to Ahmed et al.'s study from 2011, tangibility has a negative and significant
effect on liquidity risk. Since total asset is the sum of current and fixed assets, a rise in

26
fixed asset as a percentage of total asset may however weaken banks' liquidity position
because a greater fixed asset level leads in a lower liquidity ratio, which raises the risk of
liquidity.

2.5.2 Macroeconomic Determinants of Liquidity Position in Banking Sector

This study takes into account external factors that affect bank liquidity in addition to
internal ones, particularly the macroeconomic conditions specific to Ethiopia. These
variables mainly affect the rate of inflation, interest margin, and economic growth. These
elements are particularly crucial in developing nations because of their distinctive traits,
such as excessive domestic government borrowing and rising bank funding costs. Banks
are forced to finance government spending at a discount to the market price, which
causes financial distress in various restricted contexts (Ali, 2004).

In addition to bank-specific issues, macroeconomic factors are important determinants of

liquidity risk, according to several scholars' analyses. Macroeconomic factors should be
considered in the analysis of banks' liquidity risk because they have a significant impact
on the level of that risk, according to Cucinelli (2013), Yaacob, et al. (2016), Vodova
(2013), Milic & Solesa (2017), Sheefeni & Nyambe (2016), and Abdulganiyy, et al.,
(2017). Inflation, employment, GDP growth, stock indexes, monetary policy, lending
rates, currency rate movements, and economic conjuncture variations are a few of the
macroeconomic factors that have an impact on the liquidity position of commercial
banks. From these factors, the researcher in this study chooses real GDP growth rate,
interest margin and inflation rate as determining factors of Ethiopian commercial banks
liquidity risk.

i) Economic growth
Economic growth is regarded as one of the most important elements that might affect the
liquidity position of banks since it quantifies a country's ability to generate products and
services. Business activities expand as a result of economic expansion, and loan requests
rise as a result. In light of this, banks will have more options to provide loans when they
have less liquid assets. Due to this, the position of banks' liquidity may be negatively
impacted by economic development.
27
Real Gross Domestic Product (GDP), which is used to quantify economic growth, is
defined as the monetary value of all finished products and services produced inside the
borders of a nation during a given time frame. GDP is made up of all personal and
governmental consumption, government spending, investment, personal inventories,
paid-in contraction costs, and the foreign trade balance. GDP is largely utilized as a
barometer for a country's standard of life and an economic health indicator.

In accordance with Trenca et al. (2015), who examined the macroeconomic factors
influencing 40 commercial banks in six countries in Southern Europe (Croatia, Greece,
Italy, Portugal, Spain, and Cyprus) from 2005 to 2011, economic growth as measured by
RGDP has a detrimental and statistically significant impact on bank liquidity.
Additionally, Vodova (2011) demonstrated that, for Czech commercial banks from 2001
to 2009, economic development had a detrimental impact on banks' liquidity.

However, other researchers, such as (Chagwiza, 2014; Moussa, 2015), argue that banks
prefer to maintain a high level of liquidity during an economic expansion since they are
less confident in their clients' capacity to make loan payments during an economic
downturn. Bank liquidity and economic growth are so positively correlated. Additionally,
a 2013 study by Cucinelli on the factors that affect bank liquidity risk in the context of
the Euro zone discovered that GDP had a positive and significant impact on liquidity risk
with a 95% level of confidence. Similar to this, a study by Vodova (2013) on the factors
affecting the liquidity of commercial banks in Hungary from 2001 to 2010 using panel
data regression analysis demonstrates that GDP and bank liquidity have a positive and
substantial association. Similar findings were made by Moussa (2015) and Vodova
(2013).

El Khoury (2015), however, was unable to demonstrate a meaningful connection between

bank liquidity and economic growth. Additionally, Abdulganiyy et al. (2017) did a study
to contrast the factors that affect Islamic banks' liquidity risk in two different
environments: those with a fully Islamic banking system and those with a dual banking
system. The study's sample of Islamic banks from Malaysia and Sudan served as a
representation of the two financial environments. he study employed information from
28
the Islamic Banks Information System, which was given by the Islamic Research and
Training Institute, for three banks in each of the nations between 2004 and 2015. The
authors get to the conclusion that the significance of liquidity risk determinants is
determined by the various environments in which Islamic banks operate by employing
Ordinary Least Squares (OLS) and panel data analysis methodologies. The analysis also
shows that in Sudan and Malaysia, respectively, there is no discernible association
between GDP and liquidity risk.

In actuality, as researched by several scholars in various countries, the influence of Real

GDP growth rate on banks' liquidity position is not definitive.

ii) Interest margin

The difference between the annual average lending rate and deposit rate was utilized as a
proxy for the interest margin (Brock and Suarez 2000).

The liquidity preference hypothesis states that in order to lend, lenders must have a high
interest rate, which includes an interest rate margin or liquidity premium. To persuade
lenders to grant them loans on a long-term basis, borrowers are willing to pay interest rate
margin or a liquidity premium. With longer time to maturity, the liquidity premium or
interest rate margin's magnitude grows. The price that balances the desire to hold wealth
in the form of cash with the amount of cash that is actually available, in addition to the
interest rate, is what lenders pay as a result of giving up their liquid money in exchange
for a greater premium.In order to increase lending and decrease their holdings of liquid
assets, larger interest margins or higher liquidity premiums will force banks to do so. As
a result, there was a bad correlation between banks' liquidity situation and interest rate
margin.

In effect, in this study the researcher supposes that Interest rate margins expected to have
a negative impact on commercial banks liquidity position.

iii) Inflation rate

Inflation rate refers to the general increase in price levels. The effect of inflation rate on
bank liquidity has been a subject of controversy in several empirical literatures. Some
29
researchers found that an increase of the inflation rate will lower the purchasing power of
individuals, who will then need more money to buy the same products. As a result, the
demand for loans will increase and thus, bank liquidity position will decrease (Trenca et
al., 2015). According to Moussa (2015), who investigated the factors influencing
liquidity in the Tunisian banking industry between 2000 and 2010 and discovered an
inverse association between inflation and banking liquidity. Also proven by Bunda and
Desquilbet (2008), a high inflation rate has a detrimental effect on the liquidity of banks
in emerging nations. Similar findings were made in the Czech banks by Vodova (2011),
the Pakistani banks by Malik and Rafique (2013), and the Lebanese banks by El Khoury
(2015).

A higher degree of inflation, on the other hand, would result in a lower real rate of return,
which would discourage banks from making more loans and encourage them to maintain
more liquid assets. As a result, the inflation rate and liquidity level have a positive
relationship (Trenca et al., 2015). Finally, neither Vodova (2013) nor Chagwiza (2014)
have been able to confirm any meaningful connection between liquidity and inflation.

In consequence, again the impact of inflation on liquidity position of commercial banks is

not conclusive as per the finding of different researchers discussed above.

30
2.6 Research Gap
2.6.1 Knowledge Gap
According to the aforementioned theoretical and empirical literature evaluation, the
banking industry must include liquidity position analysis as a key element. Additionally,
it demonstrated how a variety of variables, including macroeconomic indicators and
bank-specific characteristics, might affect the liquidity position of banks. This study,
however, focuses on some macroeconomic and bank-specific aspects. In light of earlier
studies that have been evaluated in the literature and the researcher's own ideas, the
variables for this study were chosen with a focus on bank-specific and macroeconomic
factors that affect the liquidity position of a few chosen commercial banks in Ethiopia.

The majority of empirical studies, including those by Belaid et al. (2016), Akhtar et al.
(2011), Cucinelli (2013), Abdulganiyy et al. (2017), and Vodova (2013), focus on the
identification of determinant variables with the aim of analyzing factors affecting bank
liquidity level in the case of the European Union and some Asian and African nations,
including Pakistan, the Philippines, Malaysia, Sudan, and Tunisia. The literature analysis
also reveals the existence of contentious findings that come from various investigations
that have been done so far. Furthermore, the assessment of the literature reveals that little
work has been done to identify the factors that influence the liquidity position of banks in
Sub-Saharan African nations, particularly Ethiopia.

Due to this, it is possible to say that there have been a limited number of studies on
liquidity position analysis conducted in Ethiopia, but the majority of them ignore
studying the determinants of liquidity level directly in favor of focusing on issues like the
connection between liquidity position and bank performance in Ethiopia, including
financial risk and profitability (Samuel, 2015; Sori, 2014); and risk management practices
(Tirualem, 2009).

Some researchers, such as Tseganesh (2015), studied the determinants of liquidity risk on
private Ethiopian commercial banks on variables affecting the liquidity of Ethiopian
commercial banks; the researchers' sole method of measuring liquidity risk was loan to
deposit ratio. Nevertheless, other measures of liquidity position, such as the ratio of liquid
31
assets to deposits for measurements of liquidity level, are required to achieve the goal of
this study, which is to analyze the determining factors of liquidity position of commercial
banks operated in Ethiopia taking into account 15 commercial banks.

2.6.2 Methodological Gap

Various earlier research conducted on the determinants of liquidity position mostly
employees balanced panel data analysis by ignoring young banks in their analysis.
Therefore, this study tries to fill those gaps to identify the determinants of liquidity
position of various internal and external factors by considering young banks in the
analysis and as such unbalanced panel data model are employed. Finally most studies
conducted in various countries in general and in Ethiopia in particular are not supported
by full disclosure of major diagnostic or misspecification and ignore the impact of the
lagged value of current period liquidity position on the current period liquidity level.
Hence, this study tries to fill this methodological gap by using dynamic panel data model
in the analysis with disclosure of all major panel data diagnostic tests.

2.7 Conceptual Framework

The following conceptual framework has been developed from the discussion made
above.

Figure 2.1: Conceptual Framework

Bank Specific factor
. Bank Size
. Loan loss provision
. Capitalization Liquidity Position
. Profitability Liquid Asset/Deposit ratio
. Operational efficiency
. Tangibility

Macro-Economic factor
. Economic growth
. Inflation rate
. Interest Margin
Source: Researchers own understanding

32
 CHAPTER THREE
RESEARCH METHODOLOGY
The research methodology can be defined as the process performed to collect data for
undertaking the empirical study. It is divided into five parts. The first part presents
research design, the second and third part considers the target population and sample and
the data sources respectively. The fourth specifies how variables are measured. The last
discusses the analysis techniques used.

3.1 Research Design

The positivism ideology was used to perform this study because it firmly supports
objectivism, measurability, scientific procedures, and value-free reasoning while ignoring
belief, emotion, and perception. This suggests that a quantitative research approach is
used. According to Abiy et al. (2009), quantitative research is a methodical and scientific
exploration of numerical qualities, phenomena, and their interactions using mathematical
models, theories, and hypotheses relevant to natural and/or social settings. This
quantitative investigation employed the econometrics analytical method and went beyond
the use of statistical and mathematical models. In order to see the association between
bank specific & macroeconomic variables with liquidity position of commercial banks in
Ethiopia, quantitative research approach is employed in this study.

A deductive research approach, according to Loose (1993), referenced by Nigist (2015),

necessitates the creation of a conceptual and theoretical framework before carrying out an
empirical observation test. A deductive (quantitative) method is recommended if the issue
is identifying the variables that affect an outcome, evaluating the usefulness of an
intervention, or discovering the most accurate predictors of outcomes. The greatest way
to test a theory or explanation is also by using this method. One dependent variable and
two or more independent variables are the foundation of a mathematical regression model
that is constructed. In order to convert this mathematical model into an econometrics
model, an error term (ε) was included. The developed hypothesis is then tested based on
the estimates made after the obtained data is entered into the econometrics model to
estimate the model's parameters. Therefore, this study can be termed as explanatory type
33
of research since the objective of the study is to analyze the impact of independent (bank
specific and macroeconomic) variables on the dependent variable liquidity position of
commercial banks in Ethiopia,

3.2 Data Type and Source

Only secondary data were employed in this study, and dynamic panel data analysis was
performed to examine the data. The panel's secondary data is quantitative in nature and is
based on bank statements of financial condition and profit & loss that have been audited.
Data from banks' audited financial statements was utilized to pinpoint bank-specific
variables affecting liquidity position. The National Bank of Ethiopia (NBE) and each
bank's audited financial accounts (statement of financial position and statement of profit
& loss) were used in the study in order to strengthen the validity and dependability of the
research. Additionally, yearly economic reports for macroeconomic factors obtained from
reports from the NBE, MOF, World Bank, and IMF were also included.

3.3 Population, Sampling Method and Sampling Frame

3.3.1 Population of the Study

The study population includes all commercial banks that existed in the fiscal year
2021/22. According to NBE (2021/22), there are more than twenty fivecommercial banks
in the year 2021/22. These are; Commercial Bank of Ethiopia (CBE), Awash Bank S.C
(AWB), Dashen Bank S.C (DAB), Bank of Abyssinia S.C (BOA), Wogagen Bank S.C
(WEB), United Bank S.C (UB), Nib Bank S.C (NIB), Cooperative Bank of Oromia S.C
(CBO), Lion Bank S.C (LIB), Oromia Bank S.C (ORB), Zemen Bank S.C (ZEB), Buna
Bank S.C (BUB), Birehan Bank S.C (BRB), Abay Bank S.C (ABB),Addis Bank S.C
(ADB), Debub Global Bank S.C (DGB), Enat Bank S.C (ENB), Amhara Bank S.C (AB),
Geda Bank S.C (GB), Ahadu Bank S.C (AHB), Tsehay Bank S.C (TSB), Tseday Bank
S.C (TSDB), Sinqee Bank S.C (SB), Goh Bank S.C etc

3.3.2 Sampling Method

A sample of Ethiopian banks was gathered between the years of 2000 and 2021 in order
to explore the factors that affect bank liquidity position in the Ethiopian setting. The

34
National Bank of Ethiopia, the World Bank, the International Monetary Fund (IMF), and
the websites of the investigated banks were used to gather the necessary banks as well as
macroeconomic data. The secondary data sources are shown in Table 3.1.

Purposive sampling was utilized in this study, where the service year of the banks was set
as the criterion and all commercial banks that satisfied it were chosen as a sample.
Therefore, a sample of the fifteen commercial banks listed below that were operational
prior to 2012 is chosen for this study. Since there were few observations, it was decided
not to include the remaining banks because it was not appropriate to draw conclusions
from such limited data.

3.3.3 Sampling Frame

According to the National Bank of Ethiopia's official website, there will be more than
twenty five commercial banks licensed and operating in Ethiopia as of December 2021.
Consequently, the study will use a sample of fifteen commercial banks that were
authorized and in operation before to 2012.The remaining banks were not included
because there was just a few years' worth of data available, most of them less than a year.
All of the ones that were considered for this study have been in operation for at least 10
years and have been located throughout the nation, allowing for a fair representation of
the banking sector in Ethiopia. The twenty-two fiscal years from 2000 to 2021 for the
first seven banks indicated in Table 3.1 and the first year of operation up to the calendar
year 2021 for the remaining eight banks are covered by this research. As a result, a link
between variables may be drawn using 257 observations and an imbalanced panel data
model.

35
Table 3.1 List of Sampled Commercial Banks
No. Banks Name Year of Establishment
1 Commercial Bank of Ethiopia 1963
2 Awash Bank S.C 1994
3 Dashen Bank S.C 1995
4 Bank of Abyssinia 1996
5 Wegagen Bank 1997
6 United Bank S.C 1998
7 Nib Bank S.C 1999
8 Cooerative Bank of Oromia 2005
9 Lion Bank S.C 2006
10 Oromia Bank S.C 2008
11 Zemen Bank S.C 2009
12 Bunna Bank S.C 2009
13 Berhan Bank S.C 2010
14 Abay Bank S.C 2010
15 Addis International Bank S.C 2011

3.4 Methods of Data Analysis and Tools

To accomplish the stated general and specific objectives of the study, data analysis comes
next after the necessary data have been acquired. Two different forms of analysis will be
used in this investigation. Descriptive statistics are the first, while system dynamic panel
data analysis is the second. The study's descriptive analysis section focused on providing
a brief summary of the variables. Each variable's mean, maximum, minimum, and
standard deviation are included. On the other hand, the most crucial portion of the
analysis—the system dynamic panel data model regression analysis—helps to look at the
factors that influence liquidity position and establish relationships between dependent and
independent variables using the STATA 14 statistical package.

3.5 Description of Variables and Hypothesis Development

Through testing the relationships between liquidity position and bank-specific and
macroeconomic factors affecting it in the case of commercial banks in Ethiopia, the
researcher tries to determine the impact of various independent variables on the
dependent variable in this study. Bank size, profitability, capitalization, loan loss
provision, tangibility, and operational effectiveness are among the aspects that are
distinctive to banks. Furthermore, the macroeconomic factors include the rate of inflation,

36
interest margin, and economic growth. Therefore, the following hypotheses were
generated with detailed explanations of the variables and their assessment under the study
in order to fulfill the study's purpose.

3.5.1 Liquidity Position

According to Woods & Dowd (2008), liquidity position is the company's capacity to exit a
position at little to no cost and also relates to the availability of sufficient funds to meet
financial commitments when they become due. When long-term assets are financed by
short-term obligations, which results in financing risk, liquidity risk develops in banks.

According to the Basel Committee on Banking Supervision's definition from 1997,

liquidity position refers to a bank's capacity to fund increases in assets, which are loans, or
to accommodate decreases in liabilities, which are deposits. The Basel Committee on
Banking Supervision's recommendation to quantify liquidity position using the ratio of
liquid assets to deposits is used in this study, according to Shen et al. (2009). Cash on
hand, cash in the bank, reserve accounts with NBE, and Treasury bills acquired are all
considered liquid assets for banks in this study. Savings deposits, demand deposits, and
fixed deposits are also included in bank deposits.

3.5.2 Independent Variables

3.5.2.1 Bank Specific Factor

Evidence from existing research shows that there is a significant correlation between
liquidity position and several bank-specific characteristics. As a result, the bank-specific
factors that were considered in this study are described in the section of this paragraph
that follows.

i) Bank size
One of the bank-specific characteristics that affect liquidity position is bank size. It
measures the impact of bank asset size on liquidity position and reflects the size of the
bank's assets. Natural logarithm of total asset (LNTA) is the proxy used to estimate bank
size, according to (Ahmed, et al., 2011; Iqbal, et al., 2011). The natural logarithm of the

37
total asset was also employed by the researcher to approximate the size of Ethiopia's
commercial banks in this study.

According to the literature thus far reviewed, there are two opposing theories and
empirical findings about the connection between bank size and liquidity position.
Therefore, the size effect might not be linear, which indicates that by obtaining
economies of scale, the liquidity position is likely to rise up to a certain level and then fall
after that level. The following will be the hypothesis examined:

H1: Bank size has a significant negative influence on liquidity position of commercial
banks.

ii) Loan loss provision

A study on the determinants of bank liquidity position in the context of the Euro area was
conducted by Cucinelli (2013) to examine the type of relationship that exists between the
liquidity position, as measured by the liquidity coverage ratio (LCR) and the net stable
funding ratio (NSFR), and some particular bank structure variables. OLS regression
based on panel data was used in the investigation, and the sample consists of 1080 listed
and non-listed Euro zone banks. According to the OLS regression result, loan loss
provision has a positive, significant influence on liquidity position as assessed by the LCR
but a negative, minor effect on liquidity position as evaluated by the NSFR.

H2: Loan loss provision to total loan ratio has a significant positive effect on liquidity
position of commercial banks.

iii) Capitalization

Another element unique to banks that affects liquidity position is capitalization. It displays
the bank's capital adequacy ratio and serves as a gauge for the impact of capitalization on
liquidity position. The capital of the bank divided by its total assets, according to
Menicucci and Paolucci (2016), serves as a proxy for determining the capital adequacy
ratio. There are two competing hypotheses on the connection between banks liquidity and
capitalization has been presented in the literature so far. Therefore, the effect of
capitalization may not be linear, which implies that the risk of running out of liquidity is

38
likely to rise up to a given level by giving banks a reserve to deploy during the crisis and
fall from a certain level by discouraging banks from creating liquidity. The following
hypothesis will be examined:
H3: Capitalization has a significant positive influence on liquidity position of commercial
banks.

iv) Profitability
In order to develop a model and test the relationship between liquidity position and firm
performance through its facets, Iqbal et al. (2011) studied the impact of liquidity level on
firm specific factors in the case of Islamic banks in Pakistan. They found that return on
asset can be used as a stand-in for measuring an asset's earning potential, which is
financial performance or profitability. Additionally, return on asset can be determined by
dividing net income by the banks total asset. Last but not least, the study discovered that
return on asset has a positive and significant impact on liquidity position. Profitability
and financial success have a favorable, albeit small, impact on liquidity position,
according to another study by Ahmed et al. (2011). The profitability and liquidity status
of commercial banks in Ethiopia are negatively correlated, according to a study on
liquidity determinants by Nigist (2016). The following hypothesis has been formed in
light of this;

H4: Return on asset has a significant negative effect on liquidity position of commercial
banks.

v) Operational efficiency
The ability of the bank to successfully manage its non-interest income and expense while
hedging against its risk dimensions is referred to as operating efficiency. Shakih (2015)
conducted a study in which the determinants of liquidity position in Islamic banks in the
context of Pakistan were estimated using panel data. The results of the fixed effects
regression show that, with a p-value of 0.058 and a 90% level of confidence, operational
efficiency has a substantial positive impact on liquidity position. Moussa's (2015)
findings concur with this outcome. The ratio of noninterest expenses to total assets is

39
utilized in this study as a proxy for operational efficiency. The following is how the
hypothesis has been formed:

H5: Operational efficiency has a significant positive impact on liquidity position of

commercial banks in Ethiopia
vi) Tangibility
Tangibility indicates the proportion of fixed asset from the total asset of the bank’s
balance sheet. The proxy used for measuring tangibility is the ratio of fixed asset to total
asset. The study conducted by Ahmed, et al., (2011) reveals that tangibility has a negative
and significant impact on liquidity position. This indicates, a rise in fixed asset as
proportion to total asset will reduce the liquidity position of banks since total asset is the
summation of both current and fixed assets, as the higher fixed asset lower liquidity ratio
which results increase in liquidity risk.

H6: Tangibility has a significant negative influence on liquidity position of commercial

banks.

3.5.2.2 External (Macroeconomic) Variable

vii) Economic growth

A study conducted by Vodova (2011) and Shen et al. (2009) on determinants of liquidity
risk the proxy used to indicate economic growth is that of real GDP growth rate of the
country. The total value of goods and services a nation produces is represented by their
GDP growth. It is therefore utilized as a stand-in for economic cyclicality. According to
the findings of Vodova (2011) and Shen et al. (2009), loan demand rises in tandem with
economic growth, which reduces banks' liquidity reserves and raises their exposure to
liquidity risk.

Furthermore, during periods of economic expansion, bank deposits are less appealing
than investments, which contribute to a development of the funding gap. In a downturn,
lending possibilities aren't as attractive, thus banks keep a bigger proportion of liquid
assets, according to Vodova (2011). The validity of both arguments is strong. Also
supporting this claim is the outcome of both researches. According to Nigist (2016), real

40
GDP growth rate has a positive correlation with liquidity risk while having a strong
negative correlation with liquidity. And the following will be the development of the
tested hypothesis:

H7: Real GDP growth rate has a significant negative effect on liquidity positionof
commercial banks in Ethiopia.

viii) Interest Margin

The difference in interest rates borrowers pay that compels holders of liquid assets to part
with their money is known as the interest rate margin. The liquidity preference hypothesis
postulates that lenders need to borrow money at a high interest rate, which includes the
interest rate margin or liquidity premium. In order to persuade lenders to extend them a
long-term loan, borrowers are willing to pay an interest rate margin or a liquidity
premium. When the time to maturity increases, so does the magnitude of the interest rate
margin or liquidity premium. Thus, as a result of receiving a greater premium, lenders
give up their available cash. In addition to the interest rate, this means that the price must
balance the desire to hold wealth in the form of cash with the amount of cash that is
actually accessible. Therefore, increased interest margins or liquidity premiums will
compel banks to provide more credit and decrease the amount of liquid assets they have
in reserve.

As a result, the link between the interest rate margin and the banks' liquidity position was
negative. According to Brock and Suarez (2000), the difference between the annual
average loan rate and deposit rate was utilized as a proxy for the interest rate margin. In
essence, the researcher in this study assumes that the impact of interest rate margin on
commercial banks' liquidity situation is negative.

H 8: Interest margin has a positive and significant impact on commercial banks liquidity
position.

ix) Inflation rate

In earlier literature, there has been discussion over how the rate of inflation affects bank
liquidity. According to Moussa (2015), who examined the factors affecting banking

41
liquidity in Tunisia between 2000 and 2010 and discovered a negative correlation
between inflation and banking liquidity. The detrimental impact of high inflation rates on
the liquidity of banks in emerging nations was proven by Bunda and Desquilbet (2008).
Vodova (2011b) in Czech banks, Malik and Rafique (2013) in Pakistani banks, and El
Khoury (2015) in Lebanese banks all came to the same conclusion.

Contrarily, according to other analysts, increased inflation would result in a decline in the
real rate of return, discouraging banks from making more loans and encouraging them to
maintain more liquid assets instead. According to Trenca et al. (2015), there is a direct
correlation between inflation rate and liquidity level. Last but not least, neither Vodova
(2013) nor Chagwiza (2014) have discovered any meaningful connection between
liquidity and inflation. As a result, we expect that inflation will positively affect liquidity
in this study, and the following is how the hypothesis is being tested:
H9: Inflation has a negative impact on commercial bank’s liquidity position.

Table 3.2 Summary of Research hypothesis

Determinants Independent Variables Hypothesis Expected Relationship with
Liquidity Position
Bank Size H1 Negative
Loan loss provision H2 Positive
Internal Capitalization H3 Positive
Profitability H4 Negative
Operational efficiency H5 Positive
Tangibility H6 Negative
External Economic growth H6 Negative
Interest margin H7 Positive
Inflation rate H8 Negative

42
Table 3.3 Dependent variables: Proxies, significance and studies
Dependent Indicator Proxy Significance References
Variable
Banks Liquidity Asset to Liquid Asset Higher LIQDR Vodova, 2013;
Liquidity Deposit Ratio (LQDR) TotalDeposit indicates Chagwiza,
Position higher 2014; Moussa,
liquidity 2015; El
position Khoury, 2015;
Roman and
Sargu, 2015.

Table 3.4 Independent variables: Proxies, significance and studies

Independent Indicator Proxy Significance References
variable
Internal Loan loss provision loan Loss provision Higher LLP Tseganesh, 2012;
(LLP) Total loans & 𝑎𝑑𝑣𝑎𝑛𝑐𝑒𝑠 indicates higher Vodova, 2013; El
level of loan Khoury, 2015.
loss provision
Capitalization(CAR) Total capitals & 𝑅𝑒𝑠𝑒𝑟𝑣𝑒𝑠 Higher CAR Al Khouri, 2012;
Total assets ratio indicates Tseganesh, 2012;
higher bank’s Vodová, 2013;
capitalization Moussa, 2015;
Roman and Sargu,
2015; El Khoury,
2015; Abdul
Rahman and
Saeed, 2015.
Profitability (ROA) Net income Higher ROA Al Khouri, 2012;
Total assets indicates higher Tseganesh, 2012;
profitability. Vodova, 2013;
Moussa, 2015;
Abdul Rahman
and Saeed, 2015.
Operational efficiency Non Interest Income Higher OPEFF KinfeTuga (2016)
(OPEFF) Non − Interest Expense indicates higher
level of
operating
efficiency
Bank size (Size) Ln(Total assets) Higher SIZE Al Khouri, 2012;
indicates higher Tseganesh, 2012;
bank’s size Vodova, 2013; El
Khoury, 2015;
Moussa, 2015.
Economic growth Real GDP growth rate Higher real Bordeleau and
External/ (ECG) GDP growth Graham, 2010; Al
Macroeconomic rate indicates Khouri, 2012;
43
 higher Vodova, 2013;
economic Moussa, 2015.
growth
Inflation rate (INFR) GDP deflator variation rate Higher GDP Demiurgic, Kunt
deflator growth and Huizinga,
indicates higher 1999; Peters et al.,
inflation 2004; Growe et
al., 2014; Petria et
al., 2015.
Interest margin (IM) Higher IM Bordeleau and
indicates Graham, 2010;
Excess interest Munteanu, 2012;
earned Vodova, 2013.

3.6 Model Specification

The following econometric model was constructed for investigating the factors that
determine liquidity position in order to meet the research objective. Using the dynamic
panel data estimation technique, the dependent variable liquidity position is regressed
with its corresponding bank-specific and macroeconomic independent factors to
determine the liquidity level determinate in Ethiopian commercial banks.

In this work, the most important explanatory factors affecting the liquidity position of
Ethiopian commercial banks have been determined using a system dynamic panel data
model. The following is the general model for this investigation.
𝒀𝒊𝒕 = 𝛃𝐨 + 𝛃𝐱𝐢𝐭 + 𝛆𝐢𝐭
The symbols “i" and "t" stand for the cross-sectional and time-series dimensions
respectively. The model's dependent variable, or liquidity risk, is shown by the left-hand
side equation. The estimated model's independent variables are shown on the right side,
while the error term is shown in brackets.
Specifically, Liquidity risk (LR) model will be
LPit = β0 + β1 (LPit-1) β2 (LNTAit) +β3 (LLPit) + β4 (CARit) + β5 (ROAit) + β6 (OPEFFit) β7 (TANGit) +β8
(RGDPit) + + β9 (INFit) +β10 (IMit) + ԑ it
Where, LPit: liquidity position measures for bank “i” in period “t”. It is calculated as the
ratio of liquid asset to deposit ratio.
β0: Y-intercept.
βi: Coefficient of variable where “i” ranges from 1 to 7.
LPit-1: The lagged values of liquidity position of bank “i” in period “t”.
LNTA: Size of bank “i” in period “t”.
44
 LLP: Loan loss provision of bank “i” in period “t”.
CAR: Capitalization of bank “i” for period “t”.
ROA: Financial performance of bank “i” for period “t”.
OPEFF: Operating efficiency of bank “i” for period “t”.
TANG: Tangibility of bank “i” for period “t”.
RGDP: Real GDP growth rate for period “t”.
INFR: Inflation rate for period “t”.
IM: Interest margin of bank “i” for period “t”.

3.7 Chapter Conclusion

The methodology used to conduct data analysis in the next Chapter is covered in Chapter
3. Quantitative and secondary data from the banks audited financial statements and
macro-economic data are used in this study. Additionally, the system dynamic panel data
estimation methods are used to examine the relationship between bank-specific and
macroeconomic factors that influence liquidity position in Ethiopian commercial banks.

45
 CHAPTER FOUR

RESULTS AND DISCUSSIONS

4.1 Descriptive statistics

This section explores descriptive statistics of dependent and independent variables, which
aid in understanding the broader context of the variables under investigation. The mean,
standard deviation, minimum and maximum values, and other descriptive data are
included. While the minimum and maximum statistics display the maximum and
minimum values of each variable from all samples, the mean value displays the average
value of all samples in each variable for a specific period under study. The standard
deviation, which is the square root of variance, displays how far the sample deviates from
the mean.

Table 4.1 displays descriptive statistics for data collected from fifteen commercial banks
in Ethiopia over a period of twenty-two years, from 2000 to 2021, with 257 observations
and unbalanced panel data because eight of the fifteen banks began operations after 2004.
An example of a commercial bank is the Commercial Bank of Ethiopia, along with the
Awash Bank, Dashen Bank, Bank of Abyssinia, United Bank, Wegagen Bank, Nib Bank,
Cooperative Bank of Oromia, Lion Bank, Oromia Bank, Zemen Bank, Birhan Bank,
Bunna Bank, Abay Bank, and Addis Bank. There are ten variables listed in the table, of
which liquidity risk (LR) is a dependent variable. The other independent variables are
bank size (LNTA), loan loss provision (LLP), capital adequacy ratio (CAR), return on
asset (ROA), operational efficiency (OPEFF), and inflation rate from macroeconomic
variables. Here is a more thorough explanation of the model's summary statistics based
on Table 4.1:

46
Table 4.1 Summary of Descriptive Statistics for Variables
Variable Mean Std. Dev. Min. Max. Obs.
LIQDR Overall 0.4050638 0.2023 0.1036 1.1154 N= 257
Between 0.0458 0.3550 0.4867 n=15
Within 0.1977 0.0683 1.1128
LNTA Overall 8.862957 1.7104 4.8600 13.8100 N= 257
Between 0.8689 7.8250 11.6272 n=15
Within 1.4485 5.3400 11.7900
LLP Overall 0.0310 0.0372 0.0000 0.2900 N= 257
Between 0.0212 0.0106 0.0970 n=15
Within 0.0300 -0.0480 0.2239
CAR Overall 0.1449 0.0794 0.0372 0.8682 N= 257
Between 0.0445 0.0607 0.2425 n=15
Within 0.0672 0.0277 0.8172
ROA Overall 0.0256 0.0133 -0.0360 0.0672 N= 257
Between 0.0054 0.0163 0.0366 n=15
Within 0.0124 -0.0387 0.0582
OPEFF Overall 0.9849 0.5218 0.0150 2.8800 N= 257
Between 0.2651 0.6563 1.6207 n=15
Within 0.4612 -0.2568 2.5031
TANG Overall 0.0235 0.0148 0.0050 0.0720 N= 257
Between 0.0058 0.0102 0.0324 n=15
Within 0.0136 0.0011 0.0716
RGDP Overall 0.0885 0.0302 -0.0216 0.1357 N= 257
Between 0.0031 0.0857 0.0959 n=15
Within 0.0301 -0.0196 0.1376
INFR Overall 0.1398 0.1054 -0.0824 0.4436 N= 257
Between 0.0139 0.1297 0.1687 n=15
Within 0.1045 -0.0723 0.4536
IM Overall 0.0494 0.0168 0.0050 0.0926 N= 257
Between 0.0104 0.0277 0.0683 n=15
Within 0.0131 0.0043 0.0797
Source: STATA 14 Results for Descriptive statistics

Liquidity position

Liquidity position is measured by using stock method particularly the liquid assets to
deposit ratio. The higher the value of liquid asset to deposit ratio indicates the higher the
liquidity position since liquidity position is the result when banks can accommodate
unexpected withdrawals of fund by their customer.

47
According to Table 4.1, the mean liquid asset to deposit ratio is 40.50% and its value
range from 10.36%in the year of 2015 for CBE up to 111.54% for NIB bank in the year
2000 with a standard deviation of 20.23%. From this we can gauge that the lower
liquidity position is just observed in the case of CBE which is state owned bank and the
highest of liquidity position as measured by liquid asset to deposit ratio is observed in the
case of NIB which is private commercial banks during the period under investigation.
Furthermore, the liquidity position with in commercial banks in Ethiopia has a standard
deviation of 19.77%, which indicates on average, the liquidity position of commercial
banks shows a 19.77% deviation from period to period under investigation. Besides, the
liquidity position of commercial banks in Ethiopia shows a standard deviation of 4.58%
between themselves during the period under investigation.

Natural logarithm of total asset (LNTA)

It is a proxy’s measure of bank size and has a mean value of 8.863. An extremely high
standard deviation from the mean of 171.04% indicates a variance from the mean.
Therefore, 4.86 and 13.81, respectively, are the minimum and highest values. The lowest
value belongs to CBO in the first operating year, and the greatest value belongs to CBE
in the year 2021, both of which are more than 100% over the lowest value of CBO.

Loan loss provision (LLP)

It is the ratio of loan loss provision to total loan. The higher the value of loan loss
provision to total loan indicates the higher the loan loss provision maintained by the
selected commercial banks during the period. According to Table 4.1, loan loss provision
has a mean value of 3.1% from the period 2000 to 2021 for the selected sample
commercial banks with a standard deviation of 3.72%. The minimum and maximum
value ranges from 0% to 29% respectively. The minimum value belongs to CBO in their
first year of operation, which indicates that the bank’s loan loss provision is higher
among other selected commercial banks during the period under study. Additionally, the
maximum value registered in the year of 2003 for CBE, which suggests that the banks
hold the highest level of loan loss provision as compared with other selected commercial
banks during the period under investigation.

48
Capital Adequacy Ratio (CAR)

It is a proxy used for measuring the level of capitalization of commercial banks, which is
the ratio of total capital and reserves to total asset of commercial banks. The average
capital adequacy ratio of selected commercial banks over a period between 2000 up to
2021 is 14.49%. This indicates that sample Ethiopian commercial banks that were
included in this study able to finance 14.49cents of their one birr asset growth out of their
capital and reserve in the years from 2000 up to 2021 with a standard deviation of 7.94
percent. The least capitalized observation is -3.72% which belongs to CBE in the year of
2016, whereas highly capitalized bank during the period under investigation is CBO with
86.82% of capital adequacy ratio in their first year of operation 2005.

Return on Asset (ROA)

A ratio of net income after taxes to total assets is used as a proxy for assessing banks'
profitability. Over the years from 2000 to 2021, the selected commercial banks' average
return on assets was 2.56%. This demonstrates that from 2000 to 2021, a one birr
investment in the assets of the sample Ethiopian commercial banks included in this
analysis generated 2.56 cents of net income on average. The least profitable commercial
bank is Lion International Bank ((LIB)) with a ROA of -3.6% observed in the year of
2007, which is extremely low even below zero cents (loss) per one birr invested, whereas
the most profitable commercial bank is Zemen Bank with a ROA of 6.72% where a bank
generates 6.72 cents return per one birr investment made during the period of 2010.
According to the standard variation statistics, the difference in financial performance
between samples of commercial banks is 1.33%.
Operational efficiency

Operational effectiveness reveals how well banks control their internal non-interest
revenue and costs in order to hedge against their risk dimensions. The ratio of total non-
interest income to non-interest expense is used in this study as a proxy for measuring
operational efficiency, and the greater the value of the ratio, the more operational
efficiency there is. Selected commercial banks' mean operational efficiency for the period
49
from 2000 to 2021 is 98.49%, with a standard deviation of 52.18%. The operational
efficiency value ranges from 1.5% for CBO observed in the year of 2005—the least
efficient bank throughout the study period—to 288% registered for CBE in the year of
2009—the most efficient year of operation.
Tangibility
The percentage of fixed assets in a bank's total assets can be used as a proxy for the
amount of tangibility of commercial banks. For the sample commercial banks between
2000 and 2021, the ratio of fixed assets to total assets has a mean value of 2.35%,
indicating that on average 2.17% of their total assets are held by fixed assets, with a
standard deviation of 1.48%. While its value in the observation spans from 0.5% recorded
at CBE in 2013 to 7.20% at Nib Bank S.C. in 2018.
Real GDP growth rate (RGDP)
The real GDP growth rate is used to measure the macro economic performance of a given
country and it is the same for all of the sample commercial banks during the period. The
mean value of Real GDP growth rate during the period under study is 8.85% with a
standard deviation of 3.02%. The highest value of RGDP growth rate was registered in
the year of 2004 with a growth rate of 13.57%, whereas the lowest level of RGDP growth
rate is -2.16% registered in the year of 2003.

Real GDP Growth Rate

0.16
0.14
0.12
0.1
0.08
0.06 Real GDP Growth Rate
0.04
0.02
0
-0.02 2000 2002 2004 2006 2008 2010 2012 2014 2016 2018 2020
-0.04

Figure 4.1 Trends of Real GDP growth rate from 2000 to 2021

50
Inflation Rate (INFR)
The mean value of inflation rate during the period under study is 13.98% with a standard
deviation of 10.54%. The highest inflation rate was registered in the year of 2008 with an
average rate of 44.36%, whereas the lowest level of inflation rate is -8.24% registered in
the year of 2001.

Inflation Rate
0.5

0.4

0.3

0.2
Inflation Rate
0.1

0
2008
2000
2001
2002
2003
2004
2005
2006
2007

2009
2010
2011
2012
2013
2014
2015
2016
2017
2018
2019
2020
2021
-0.1

-0.2

Figure 4.2 Trends of Inflation rate from 2000 to 2021

Interest Margin (IM)
The mean value of the interest margin during the period under study is 4.94% with a
standard deviation of 1.68%. The minimum and maximum interest margin was 0.5% and
9.26% respectively. The minimum interest margin of 0.5% belongs to Oromia bank in the
year of 2009, which indicates the lowest level of interest earned on their average interest
bearing assets during the year of 2009. Whereas, the maximum level of interest margin
belongs to Nib Bank with 9.26% during the year of 2021, which implies that the Nib
Bank generates the highest level of interest among other commercial banks during the
year of 2021. Overall, the finding suggest that there was a significant difference between
the interest margins of Oromia bank in 2009 and Nib bank in 2021, with Nib bank
achieving a much higher level of interest earnings.

51
4.2 Correlation Analysis
It is crucial to demonstrate the correlation between the variables utilized in the regression
analysis before examining the regression results. The statistical method used to examine
how closely two variables are related is called correlation analysis. The correlation
analysis of the variables under inquiry is the focus of this section of the study. Correlation
analysis is primarily used to determine whether the model has a multicollinearity issue
and to determine whether the variables are moving in unison or not. The correlation
coefficient shows the strength of a linear relationship between two variables. The
correlation coefficient lies within the -1 to +1 range. The perfect positive correlation
between two variables is represented by a value of +1, and the perfect negative
correlation is represented by a value of -1. The stronger the correlation of the data it
reflects, the closer the coefficient is to either of these numbers. Correlation values closer
to zero highlight weaker/poorer association than those closer to +1/-1 on this scale, where
0 denotes no correlation.

Table 4.2 Correlation matrix of dependent (LIQDR) and independent variables

LIQDR LNTA LLP CAR ROA OPEFF TANG RGDP INFR IM
LIQDR 1.0000
LNTA -0.6819 1.0000
LLP 0.1959 0.0958 1.0000
CAR 0.4572 -0.558 -0.32 1.0000
ROA -0.1926 0.1832 -0.152 -0.289 1.0000
OPEFF 0.3337 -0.027 0.2226 -0.198 0.6326 1.0000
TANG -0.3056 0.2003 -0.295 0.0595 -0.165 -0.3833 1.0000
RGDP 0.2428 -0.084 -0.117 0.1250 0.2273 0.3413 -0.184 1.0000
INFR 0.0202 0.2065 -0.129 0.0211 0.1216 0.1136 0.0443 0.0113 1.0000
IM -0.4924 0.2290 -0.350 0.0528 0.1305 -0.4553 0.2603 -0.132 0.1016 1.000
Source: STATA14 Result for correlation matrix
The relationship between the independent variables utilized in this study and the
dependent variables used to calculate liquidity position using LIQDR is shown in the
correlation matrix in table 4.2. The liquidity position (LP) correlation analysis with
independent variables is displayed in the correlation matrix. Based on the correlation
matrix liquidity position (LP) as measured by liquid asset to deposit ratio (LIQDR) have
52
a positive correlation with loan loss provision (LLP), capitalization(CAR), Operational
efficiency (OPEFF), Real GDP growth rate (RGDP) and Inflation rate(INFR).The
positive relationship between loan loss provision as measured by loan loss provision as a
ratio of total loans and advances and liquidity position indicates if the loan loss provision
as a ratio of total loans and advances increase, it increases the banks liquidity position as
its liquid asset to deposit ratio increases. Similarly speaking, the positive correlation
between capitalization as measured by the ratio of capital and reserves to total asset
increases, the liquid asset to deposit ratio also increases and which ultimately improves
the banks liquidity position and vice versa. Finally, the positive correlation between
operational efficiency, Real GDP growth rate and inflation rate indicates the increase in
those variables leads to an increase in the liquid asset to deposit ratio of the bank, which
intern implies an improve in the liquidity position of commercial banks in Ethiopia.

Conversely, bank size as measured by LNTA, profitability as measured by ROA,

tangibility (TANG) and interest margin are negatively correlated with liquidity
position(LP) as measured by liquid asset to deposit ratio. The positive correlation results
indicate that as those independent variables rise, the ratio of liquid assets to deposits falls
and the liquidity position declines.

4.3 Testing Assumptions of Classical Linear Regression Model (CLRM)

The paper's prior chapter's use of an econometric model was essential to achieving the
study's goals. An econometric model should theoretically pass pre and post estimation
tests or diagnostic tests, the researcher is also aware of this. The researcher used the
following tests to make sure the model is trustworthy, consistent, and valid.

4.3.1 Test for normality assumption (ut ∼ N (0, σ2)

It is crucial to do a test on the residuals for normal distribution before using statistical
analysis methods that presuppose normality, such as ordinary least square regression. The
researcher makes the assumption that the data has a normal distribution of the residuals
and will only reject this assumption if there is compelling evidence to support the
opposite, i.e., if the test is significant and the distribution is non-normal. In this study, the
residuals skewness-kurtosis test is employed to examine residual normality. The
53
 skeweness-kurtosis test for normal data states that if the p value is less than 0.05, the data
are not normal. The null hypothesis will not be accepted if the p value is greater than
0.05, indicating that the model's error term is regularly distributed.

Table 4.3 Results of Normality Test

Skewness /Kurtosis tests for Normality
Joint
Variable Obs Pr Pr (Kurtosis) adj chi2 (2) Prob>chi2
(Skewness)
Uhat 257 0.0505 0.3190 4.85 0.0884

Source: STATA 14 Results for tests of Normality

2.5
2
1.5
Density

1
.5
0

0 .2 .4 .6 .8 1
Fitted values

Figure 4.3 Histogram of Residual

Source: STATA 14 Results for histogram of residual

According to table 4.3.1 the data used in this study is normally distributed since the p-
value is greater than 0.05, we fail to reject that the data is normally distributed.

54
4.3.2 Test for homoskedasticity assumption (Var (ut) = σ2)
The variance of each disturbance term should be a fixed value, according to Gujarati
(2004), who lists this as one of the main presumptions of the classical linear regression
model. Homoskedasticity is predicated on this, whereas heteroskedasticity results from a
violation of this presumption. When heteroskedasticity presents, there are two main
consequences of the least square estimators: first the least square estimator is still linear
and unbiased estimator, but it is no longer best i.e. there is another estimator with a small
variance. Second, the standard errors calculated for the least squares estimators are
incorrect which affect confidence intervals and hypothesis testing that use those standard
errors, which could ultimately lead to misleading conclusion.

Table 4.4 Results for Heteroskedasticity Test

Modified Wald test for group wise heteroskedasticity
H0: sigma(i)^2 = sigma^2 for all i
chi2 (15) 69.87
Prob > chi2 0.0000
Source: STATA14 Result for heteroskedasticity test using Breusch –Pagan test

The modified Wald test supports the model's inclusion of heteroskedasticity, as shown in
the above table. The null hypothesis of homoskedasticity/constant variance could be
rejected because the p value for the modified Wald test of heteroskedasticity is 0.0000,
which is significantly less than 0.05. The problem of heteroskedasticity can be solved by
using either of the feasible generalized least square (FGLS) or generalized methods of
moment (GMM) estimation method.
4.3.3 Test for absence of serious Multicollinearity assumption

When all or some of the explanatory variables in a regression model have a linear
connection with one another, this is referred to as multicollinearity (Gujarati, 2004).
Variable inflation factor (VIF) or tolerance values are employed to determine the level of
multicollinearity. Tolerance value and variable inflation factor (VIF) are both used
interchangeably. According to Ahmad and Ariff (2007) and Gujarati (2004), a major
multicollinearity issue is present if the variable inflation factor (VIF) between various
variables is greater than 10 and the tolerance level is less than 0.10. The tolerance value
55
ranges from 0 to 1, with 0 indicating collinearity problems and 1 indicating no
multicollinearity issues. Table 4.3.4 shows that the mean variable inflation factor (VIF)
for the model and each independent variable are both less than 10. This test verifies that
there is less collinearity among the explanatory factors.

Table 4.5 Results for Multicollinearity Test

Variable VIF 1/VIF
OPEFF 4.12 0.242975
ROA 3.33 0.300209
IM 2.14 0.466573
CAR 1.94 0.515139
LNTA 1.83 0.545583
AQ 1.67 0.598886
TANG 1.34 0.746041
RGDP 1.24 0.804324
INFR 1.15 0.86745
Mean VIF 2.09
Source: STATA 14 Result for Variable inflation factor test

From table 4.3.4 we can conclude that there is lower degree of collinearity among
explanatory variables since the VIF of the variables is less than 5; there is no problem of multi-
collinearity (Gujarati, 2005).

4.6 Discussion on Endogeneity

The problem of endogeneity occurs if the explanatory variable is correlated with the
unknown random error term in a regression model due to measurement error, auto
regression with auto correlated errors, reverse causality and omitted variable bias
(Wooldridge, 2002).

4.6.1 Omitted variable and endogeneity

As per Jarkowsky (2002), omitted variable bias is the difference between the expected
value of an estimator and the true value of the underlying parameter as a result of failures
to control for a relevant explanatory variables i.e. when one or more explanatory
variables which is supposed to be considered is missed, our estimation is likely to be an
error. In other words, omitted variable bias occurs when we misspecify the model by
including an irrelevant variable (over specification), or more seriously excluding relevant
variable from the model. Therefore if we miss relevant variable from our model that

56
variable correlated with the error term and which leads to endogeneity and as a result our
estimation becomes biased and inconsistent.

Table 4.6 Results of Ramsey RESET test using powers of the fitted values of liqdr
Ho: model has no omitted variables

F(3, 244) 10.33

Prob > F 0.0000
Source: STATA 14 Result for Variable inflation factor test

From the result of table 4.6, we reject the null hypothesis of the model has no omitted
variable bias has been rejected and there is some evidence of inclusion of irrelevant
variables or exclusion of relevant variables and this leads to endogeneity problem.

4.6.2 Reverse causality and endogeneity

Reverse causality occurs when there is a condition that describes the association of two
variables in which an explanatory variables causing or affecting the dependent variable
and also the dependent variable also affects the explanatory variables. In this research,
the researcher suspects that there is a problem of reverse causality between liquidity and
profitability and the like.

In finance literature, there is a trade-off between liquidity and profitability of banks in

which as profitability affects liquidity of banks liquidity position also affects their level
of profitability. This indicates that if commercial banks hold a significant level of liquid
assets, their profitability erodes as much of their funds tie in liquid assets than investing
in long-term assets like loans and advances. Similarly, if commercial banks grant a
significant level of loans and advances, this reduces their liquidity position but increases
their profitability. From this we can understand that there is a reverse causality between
liquidity and profitability and this leads to the problem of endogeneity which makes our
estimates biased and inconsistent.

4.5 Other Post-estimation Diagnostic Testing

4.5.1 Model specification test
The assumption of CLRM that the regression model is accurately stated or that there was
no specification bias or error is known as the "model specification test," which is another

57
assumption of CLRM. This may be the result of the model being under fitted or over
fitted due to the inclusion of irrelevant variables or the exclusion of relevant variables; a
flawed functional form and a mistake in the measurement of a proxy will also break this
assumption. This study employs a link test to determine whether the issue is present.
Table 4.7 Result for Model Specification Test

Source SS df MS Number of obs=257

F(2,254)=393.27
Model Prob>F=0.0000
Residual 2.55921399 254 0.010075646 R-squared=0.7559
Adj R squared=0.7540
Total 10.4841941 256 0.040953883 Root MSE = 0.10038
Liqdr Coef. Std. Err. t P>|t| [95% Conf. Interval]
hat 0.821311 0.124433 6.60 0.000 0.57625 1.066364
hatsq 0.207262 0.1382613 1.50 0.135 -0.06502 0.47954
cons 0.032000 0.0265348 1.21 0.229 -0.0202 0.0842
Source: STATA14 Result for model specification tests
With a p value of 0.000, the link test in the aforementioned table clearly shows that the
model predictive value (_hat) is significant. The square prediction (_hatsq), on the other
hand, is not significant at the 5% level of significance with a p value of 0.135.
Consequently, the link test verifies that there is no proof of a model misspecification
issue.

4.5.2 Testing for panel cross-sectional dependence

According to Baltagi, cross-sectional dependence is a problem in macro panels with long
time series (over 20-30 years). This is not much of a problem in micro panels (few years
and large number of cases). Pasaran CD (cross-sectional dependence) test is used to test
whether the residuals are correlated across entities or not. Cross-sectional dependence can
lead to bias in tests results also called contemporaneous correlation. The null hypothesis
is that residuals are not correlated across entities. The problem of cross sectional
dependence can also be solved by using either of the feasible generalized least square
(FGLS) or generalized methods of moment (GMM) estimation method (Sarafidis etal,
2008).

58
Table 4.8 Results of Pasaran CD test for cross-sectional dependence
Pesaran's test of cross sectional independence:
Average absolute value of the off-diagonal elements= 8.743, Pr= 0.0000
Average absolute value of the off-diagonal elements = 0.340
Source: STATA14 Result of testing cross-sectional dependence

Based on the above table, since the p-value is statistically significant and the null
hypothesis is the residuals are not correlated and we reject the null and concludes that the
residuals are correlated across entities.

4.5.3 Panel Unit Root Tests

Panel data estimators are inefficient if the variables in a panel data are non-stationary
unless they are not co-integrated. Although Pesaran (2015) classifies first and second
generations of the panel unit as root test models, panel co-integration models also have
become popular in recent years. Within the panel unit root tests methods, there are two
generation of tests. The first generation of tests assumes that cross section units are cross-
sectionally independent, whereas the second generation of panel unit root relaxes this
assumption and allows for cross sectional dependence. And therefore, since in this study
the cross units are cross-sectionally dependent, the researcher employs the second
generation panel unit root tests methods of Pesaran’s CADF test (2007) to overcome the
problem of cross-sectional dependence in the first generation tests.

Table 4.9 Results of second generation Pesaran’s CADF panel unit root test
HO: Null hypothesis assumes that all series are non-stationary
Variable LNTA LLP CAR ROA OPEFF TANG RGDP* INFR* IM
Z[t-bar] -5.763 -0.671 -5.254 -4.871 -3.677 -3.208 -11.289 -12.27 -7.441
P-value 0.000 0.0251 0.000 0.000 0.000 0.001 0.000 0.000 0.000
Source: STATA14 output for panel unit root tests
*Their first differences are considered for testing the second generation panel unit
root test.
Based on the above table, we reject the null hypothesis of all panel series are non-
stationary and we can concluded that all of the panel series are stationary.

59
4.5.4 Testing for serial correlation
Serial correlation tests apply to macro panels with long time series over 20-30years. Not
a problem in micro panels with very few years. Serial correlation causes the standard
errors of the coefficients to be smaller than they actually are and higher R-squared.
The hypothesis for this test was as follows;
H0: No first-order serial correlation.
Ha: Some first order serial correlation.

Table 4.10 Heteroskedasticity-robust Born and Breitung (2016) HR-test as post estimation
Variable HR-stat p-value N max T balance?
Post Estimation 1.84 0.066 15 22 unbalanced
Source: STATA14 Results for serial correlation tests

Based on the above table, the null hypothesis is no serial correlation up to order p and we
fail to reject the null since the p-value is not significant at 5% level of significance and
we can concluded that there is no serial correlation in the model.

4.5.5 Summary of diagnostic testing

Based on various diagnostic testing conducted elsewhere in this thesis, there is some
problem of heteroskedasticity, cross-sectional dependence and endogeneity which are
common in panel data framework. Therefore due to those diagnostic problems the GMM
are used since it has the ability to correct those problems detected in panel data (Sarafidis
etal, 2008).

In addition to the above diagnostic problems, since the lagged value of liquid asset to
deposit ratio obviously affects the current year liquidity level of commercial banks in
Ethiopia it can be reasonable to use dynamic panel data model for considering the effect
of the lagged values of liquidity risk as a regressor for analyzing the determinants of
liquidity risk in the case of commercial banks in Ethiopia. To sum-up, the presence of a
lagged dependent variable as a regressor and cross section-specific unobserved
heterogeneity are the most important factors for deciding to use the system dynamic
panel data model in this study.

60
Furthermore, feasible generalized least square (FGLS) allows panel data estimation in the
presence of serial autocorrelation within panels and cross-sectional correlation and
heteroskedasticity across panels. Hence, the FGLS can be used in addition to System
GMM by using the lagged values of liquid asset to deposit ratio as one of theregressor.

Finally, before applying the dynamic panel data model of system GMM there is a need to
perform both Sargan test of over identifying restrictions and Arellano-Bond test for zero
autocorrelation in first-differenced errors as follows.

4.5.6 Sargan test of over identifying restrictions

The Sargan test of over identifying restrictions is applied to check the validity of the
instruments used in simultaneous equation models. As the name suggests, it is applicable
to over identified model equations. Over identification indicates that the number of
endogenous regressor is less than the number of instruments used in the given equation. It
can be easily determined using order and rank conditions of identification. Identification
is one of the essential and initial steps in estimating simultaneous equation models.
Furthermore, if an equation or model is over identified, tests of over identifying
restrictions must be applied to check the validity of instruments. When an equation or
model is just or exactly-identified, such tests are not applicable.

Table 4.11 Sargan test of over identifying restrictions

Sargan test of over identifying restriction
H0: over identifying restrictions are valid
chi2(194) 227.13
Prob > chi2 0.0518
Source: STATA14 Results for sargan test of over identifying restriction

From the above table, since the p-value of the sargan test over identifying restriction is
insignificant, we fail to reject the null hypothesis of over identifying restrictions are valid
and therefore, over identifying restrictions are valid.

4.5.7 Arellano-Bond test for zero autocorrelation in first-differenced errors

Linear dynamic panel data models include lags of the dependent variable as covariates
and contain unobserved panel level effects, fixed or random. By construction, the

61
unobserved panel-level effects are correlated with the lagged dependent variables,
making standard estimators inconsistent. Arellano and Bond (1991) derived a consistent
generalized method of moment’s estimator for the parameters of this model. This
estimator requires that there be no autocorrelation in the idiosyncratic errors and it can be
tested by using Arellano-Bond test for zero autocorrelation in first-differenced errors as
described in the following table.

Table 4.12 Arellano-Bond test for zero autocorrelation in first-differenced errors

Order z Prob > z
First -1.8256 0.0679
Second 1.2604 0.2075
Source: STATA14 Results for Arellano-Bond test for zero autocorrelation

From The output of the above table, we can conclude that there is no significant evidence
of serial correlation in the first-differenced errors at both order 1 and 2. Therefore, we can
conduct the dynamic panel data model of system GMM for analyzing the determinants of
liquidity position of commercial banks in Ethiopia.

4.6 Regression Result and Discussion

The drivers of liquidity position in Ethiopian commercial banks between the years of
2000 and 2021 are examined in this study using unbalanced panel data. The unbalanced
panel data has a total of 257 observations and comprises cross-sectional data from 15
banks representing both private and public commercial banks throughout a period of ten
to twenty-two years, depending on the bank's year of operation.

In this section, the relationship between the dependent variable and independent variables
discussed above has been analyzed on the basis of the findings on this empirical study of
system dynamic panel data estimation method. The dependent variable, liquidity position
of Ethiopian commercial banks, was measured by using the banks Liquid asset to deposit
ratio.

62
Table 4.13 Result of System dynamic and FGLS panel-data estimation

System dynamic panel-data estimation Cross-sectional time-series FGLS

Obs per group: regression
Min= 9 Coefficients: generalized least
Avg = 16.13333 squares
Max = 21 Panels: heteroskedastic
Number of obs = 242 Correlation: no autocorrelation
Number of groups = 15
Coefficient Std. Error Z P>|z| Coefficient P>|z|
Variable
CONS 0.2551182 0.0765201 3.33 0.001 0.7569 0.000
LIQDR 0.3850038*** 0.0496213 7.76 0.000 0.2585*** 0.000
L1.
LNTA -0.0309239*** 0.0064289 -4.81 0.000 -0.0620*** 0.000
LLP 1.453846*** 0.208651 6.97 0.000 0.8464*** 0.000
CAR 0.7304497*** 0.1714562 4.26 0.000 0.3358*** 0.001
ROA -0.8985094 0.9778536 -0.92 0.358 -3.5063*** 0.000
OPEFF 0.0612041** 0.024486 2.50 0.012 0.1786*** 0.000
TANG -0.317566 0.4961249 -0.64 0.522 -0.2946 0.487
RGDP 0.4271104** 0.1835239 2.33 0.020 0.5175*** 0.006
INFR 0.1770032*** 0.0543281 3.26 0.001 0.2196*** 0.000
IM 0.1926836 0.6974923 0.28 0.782 -0.3859 0.429
*, ** and *** denotes the level of significance at 90%, 95% and 99% level of confidence respectively.

Source: STATA 14 Regression result

As per the above result, the model can be written as follows by using the result of
System dynamic panel-data estimation and since there is a positive relationship between
liquid asset to deposit ratio and liquidity position, the sign of coefficient doesn’t changes
as the model explained by using liquidity position (LPit) as dependent variable.

LPit = 0.2551+0.3850(LP L1.) - 0.0309 (LNTAit) + 1.4538 (LLPit) + 0.7304 (CARit) -

0.8985 (ROAit) +0.0612 (OPEFFit) - 0.3175 (TANGit) + 0.4271 (RGDPit) + 0.1770
(INFRit) + 0.1926 (IMit) + ԑ it

As per the result of table 4.13 of the System dynamic panel-data estimation of the model
which has been performed on nine independent variables, out of which seven of them
including the lagged values of liquidity position are statistically significant even at five
percent level of significance and three variables are statistically insignificant even at ten
percent level of significance. From the bank specific factors the lagged value of liquid
asset to deposit ratio (LIQDR L1), Bank size (LNTA), loan loss provision (LLP),
63
Capitalization (CAR) and Operational efficiency (OPEFF) have a significant impact on
liquidity position of Ethiopian commercial banks at 95% level of confidence. Besides
tangibility and financial performance have insignificant impact on liquidity position of
Ethiopian commercial banks even at 90% level of confidence. On the other hand, from
the macroeconomic variables Real GDP growth rate and inflation rate have a significant
effect on liquidity position of commercial banks in Ethiopia at 95% and 99% level of
confidence respectively. Besides, interest margin have insignificant impact on liquidity
position of Ethiopian commercial banks even at 10% level of significance.

In examining coefficients of the regression result, the lagged values of liquid asset to
deposit ratio (LIQDR. L1), loan loss provision, capitalization, operational efficiency, real
GDP growth rate and inflation rate have a significant positive impact on liquidity
position of commercial banks as measured by liquid asset to deposit ratio, which
indicates increase in the value of these independent variables will leads to an increase in
liquidity position, since the highest liquid asset to deposit ratio indicates the higher the
liquidity position of commercial banks and decrease in these explanatory variables will
result in decrease in liquidity position of Ethiopian commercial banks. In contrast, bank
size has a significant negative effect on liquidity position of commercial banks. Whereas
the rest factors like; profitability, tangibility and interest margin have insignificant effect
on liquidity position of commercial banks in Ethiopia.

4.7 Discussion on Results of System Dynamic panel-data estimation

i) Lagged value of Liquid asset to deposit ratio (LIQDR. L1)

From the results of table 4.13, the lagged value of liquid asset to deposit ratio of the
current year can be statistically and positively affected by the lagged values of liquid
asset to deposit ratio. The result can be interpreted as, on average, if the lagged value of
commercial banks liquidity position increases/decreases by 1 birr, the current year liquid
position also increases/decreases by 0.385 birr respectively. In other words, if the
previous year liquidity position as measured by liquid asset to deposit ratio
increases/decreases by a percent, the current year liquidity position also increases by
0.385 percent, by holding other factors remain constant. Therefore, the current year

64
liquidity position of commercial banks of Ethiopia have also significantly affected by
their previous year level of liquidity position.

ii) Bank size (LNTA)

When examining the impact of bank size on liquidity position for commercial banks in
Ethiopia, one method of measuring bank size is by utilizing the natural logarithm of total
assets. This method can be used to determine the liquidity level, which is unstable as a
business grows. With a p-value of 0.0000 and a negative coefficient (0.0309), table 4.13's
findings show that bank size has a statistically significant negative impact on the liquidity
position of Ethiopian commercial banks. This finding confirmed the negative impact of
bank size on the liquidity level of Ethiopian commercial banks, which implies that if
bank size increases the liquidity position as measured by liquid asset to deposit ratio
decreases which intern increases the liquidity risk of commercial banks in Ethiopia. The
result can be interpreted as if bank size increases/decreases by one percent, the liquidity
position of the commercial banks decreases/increases by 0.03 units/birr, by holding other
factors remain constant. Therefore the result reveals that there is a negative effect of bank
size on liquidity position of commercial banks in Ethiopia.

The results of this study are consistent with those of Shen et al. (2009) and Iqbal et al.
(2011), who argue that the size of the bank has a negative, significant impact on the
liquidity position. The size of the bank had a considerable effect on the liquidity position
of Ethiopian commercial banks, according to Liza (2018) and (2015). But the results
contradict those of Estrada (2011), who found that when a firm grows, it may go
bankrupt. When a firm grows, it will allow them to manage risk, similar to how
economies of scale work. As a result, the study's findings support the idea that some
companies are "too big to fail," and commercial banks in Ethiopia should exercise
caution when it comes to the risk of liquidity by controlling the amount of assets they
have, such as loans and advances, and by adhering to a sound policy of maintaining a
balanced loan portfolio in their statement of financial position.

iii) Loan loss provision

These variables can be measured by using the ratio of loan loss provision to total loans
and advances for identifying the impact of loan loss provision on liquidity position of
65
commercial banks in Ethiopia. According to the result of the above regression table, loan
loss provision has found to have statistically significant positive effect on liquidity
position of Ethiopian commercial banks with the p-value of 0.000 and positive coefficient
of 1.4538,whichconfirmed a positive effect of holding loan loss provision on liquidity
level of commercial banks in Ethiopia. The result can be interpreted as, if the bank’s loan
loss provision as gross loans and advances increases/decreases by one birr, the
commercial banks liquidity position as measured by liquid asset to deposit ratio
increases/decreases by 1.45 birr respectively, holding other things remain constant.
Therefore, the result indicates that there is a positive effect of banks loan loss provision
on liquidity position of commercial banks in Ethiopia.

The finding of this study are in line with Roman and Sargu (2015) they found in the same
study a significant positive relationship between banks loan loss provision and bank
liquidity position in the Czech Republic, Lithuania and Romania. In Ethiopian context,
this result can be justified due to the fact that that the commercial banks in Ethiopia holds
some provision when granting loans and advances to their customer and when the loan
becomes default those provision can be used to protect the bank from liquidity crisis.

iv) Capitalization
The capital adequacy ratio (CAR), which may be used to gauge capitalization, was found
to be statistically significant in the regression analysis with a p-value of 0.000 and
positive coefficients of 0.7304. The positive coefficient of 0.7304, when held constant,
means that, on average, a 1 Birr increase or reduction in the capital adequacy of
commercial banks in Ethiopia will result in a 0.7304 Birr increase or decrease in their
liquidity position. As a result, CAR significantly reduces Ethiopia's commercial banks'
liquidity position.

A study by Menicucci and Paolucci (2016) found that a high capital adequacy ratio is a
positive predictor of a bank's stability and liquidity position and helps to improve the
liquidity position of commercial banks. This study's findings are consistent with their
findings. In addition, this study's findings support Repullo's findings from 2004 by
showing that banks will improve their liquidity levels and reduce their liquidity risk when
capital levels rise. The outcome can be explained by the fact that since retained earnings
66
and bank reserves are included in commercial banks' capitalization, raising their levels
will lessen their vulnerability to serious liquidity position risk.

v) Operational Efficiency
Operational efficiency gauges how effectively a bank controls internal operational costs
in order to successfully control risk exposure. Operating efficiency can be approximated
by the ratio of total noninterest income to noninterest expense. Operational efficiency is a
significant factor at the 5% level of significance, according to the above regression's
findings, with a p-value of 0.012 and a coefficient of 0.0612. Accordingly, if we assume
citrus paribus, a one Birr increase or decrease in operational efficiency will result in a
0.061 Birr rise or decrease in the liquidity position of Ethiopia's commercial banks,
respectively. As a result, operational efficiency is a significant explanatory factor that
may be used to assess the liquidity position in commercial banks in Ethiopia. The
findings of other research like Shakih (2015) and Moussa (2015), which discovered a
substantial positive link between the two variables of operational efficiency and liquidity
position, lend support to this result. According to Shakih (2015), increasing operational
efficiency increases the liquidity position by freeing up resources; conversely, operational
inefficiency has the opposite effect. And according to Tilahun & Dugassa, poor
management of banking institutions will make banks less efficient and have an impact on
their liquidity situation. In conclusion, bank efficiency improves the liquidity position
faced by commercial banks in Ethiopia.

vi) Real GDP Growth Rate

According to the result of regression in table 4.13, real GDP growth rate has a positive
and significant effect on liquidity position with p-value 0.020 and a positive coefficient of
0.4271. This result can be interpreted as, on average, if the Ethiopians Real GDP growth
rate increases/decreases by one unit, the liquidity position of commercial banks
increases/decreases by 0.42 units, which intern reduced/increased the liquidity risk of
commercial banks in Ethiopia by the same unit respectively. The outcome shows that
banks are better able to manage their liquidity position even though loan demand is rising
as the GDP grows year over year. This result differs from those of Vodova (2013) and
Cucilleni (2013), who discovered a strong correlation between GDP and liquidity risk in
67
commercial banks in Hungary and the Euro area, respectively. However, this study's
findings are comparable to those of Abdulganiyy, et al. (2017) who examined the factors
that determine liquidity position in the Islamic banking systems of Malaysia and Sudan
and found that real GDP growth rate had a positive impact on commercial banks'
liquidity position. With regard to this study's findings on the factors affecting liquidity
position at Ethiopian commercial banks, Liza (2018) also came to similar conclusions.

vii) Inflation Rate

The result of System dynamic panel regression in table 4.13 reveals that inflation rate has
a positive and significant effect on liquidity position with p-value 0.001 and a coefficient
of 0.1770. The result indicates; if the inflation rate increases the liquidity position of
commercial banks in Ethiopia also increases and it is interpreted as on average, if
inflation rate increases/decreases by one birr, the liquidity position of commercial banks
also increases/decreases by 0.17birrrespectively; by holding other factors remain
constant. This result can be justified if the inflation rate increases, the real rate of return
earned by commercial banks reduced and the banks become reluctant to grant loans and
advances to their borrowers, which intern increases their liquid assets. This result is in
line with Moussa's (2015) research, which examined the factors affecting liquidity in the
Tunisian banking industry between 2000 and 2010 and discovered a correlation between
inflation and the risk of banking liquidity. Furthermore, this study's findings show the
detrimental effects of Ethiopia's high inflation rate on the liquidity of commercial banks,
which strengthen the findings of a previous study by Bunda and Desquilbet (2008). The
results are also in line with those obtained by Vodova (2011b) in Czech banks, Malik and
Rafique (2013) in Pakistani banks, and El Khoury (2015) in Lebanese banks.

viii) Other variables

As per the result of both system dynamic and FGLS panel model regression in table 4.13
tangibility and interest margin has insignificant effect on liquidity position even 90%
confidence level.

68
4.8 Summary of Hypothesis Tested
Table 4.14 Summary of Hypothesis Test
Independent variables Expected effect on Findings Decision
liquidity position
Bank Size Negative and Negative and Fail to Reject the hypothesis
Significant significant
Loan loss provision Positive and Positive and Fail to reject the hypothesis
Significant significant
Capitalization Positive and Positive and Fail to reject the hypothesis
significant significant
Profitability Negative and Negative and Reject the hypothesis
Significant insignificant
Operational Efficiency Positive and Positive and Fail to reject the hypothesis
significant Significant
Tangibility Negative and Negative and Reject the hypothesis
Significant Insignificant
Real GDP Growth Rate Negative and Positive and Reject the hypothesis
Significant Significant
Inflation Rate Negative and Positive and Reject the hypothesis
significant Significant
Interest Margin Positive and Positive and Reject the hypothesis
Significant insignificant

69
 CHAPTER FIVE

CONCLUSION AND RECOMMENDATION

The analysis of the research's findings and discussions of it are the main topics of the
preceding chapter that has been discussed so far. On the basis of the findings of the
analysis covered in the previous chapter, conclusions and suggestions are offered in this
chapter. The chapter is organized into two pieces in accordance with this; the first section
presents the study's conclusions, and the second section presents recommendations for
commercial banks, their supervisory body, and more research based on the analysis of the
study.

5.1 Conclusion
The analysis of macroeconomic and bank-specific factors affecting the liquidity position
of commercial banks in Ethiopia from 2000 to 2021 was the main goal of this study. For
the purpose of this study, unbalanced panel data from fifteen commercial banks with 257
observations were used. Using descriptive statistics and system dynamic panel data
regression, the sample data from both public and private commercial banks were studied.
Liquid asset to deposit ratio of commercial banks is the dependent variable used to
measure liquidity position. The dependent variable was regressed with independent
variables, including real GDP growth rate, inflation rate, and interest margin from
macroeconomic explanatory variables, as well as bank size, loan loss provision,
capitalization, return on asset, operational efficiency, and tangibility from bank-specific
explanatory variables.

The results of this study show that, based on bank-specific characteristics, bank size has a
considerable negative impact on Ethiopian commercial banks' liquidity positions. This
indicates that the liquidity position of Ethiopian commercial banks is significantly
negatively impacted by bank size. This data leads us to the logical conclusion that
Ethiopia's largest bank, or banks holding the largest number of assets, are extremely
exposed to liquidity deterioration. The result backs up the notion of "too big to fail" that
has been covered in other sections of this study.

70
The commercial banks' holding of loan loss provisions also has a positive impact on the
Ethiopian commercial banks' liquidity position. As a result, the data indicates that
commercial banks in Ethiopia are less vulnerable to liquidity deterioration the bigger the
loan loss provision they hold. This suggests that the bank should set aside a portion of the
gross loans and advances at the time the borrower receives loans and advances in order to
manage the liquidity position.

Capitalization as measured by capital as a ratio of total assets also has a positive impact
on liquidity position of commercial banks in Ethiopia, which indicates that if the bank’s
capital and reserve improves, their vulnerability to liquidity deterioration reduces, by
keeping other factors constant. Similarly, operational efficiency also has a significant
positive impact on liquidity position of commercial banks in Ethiopia in which the
improvement in operational efficiency of commercial banks leads to the improvement in
the liquidity position of those banks in Ethiopia.

On the other hand from the macro economic factors, the Real GDP growth rate and
inflation rate also has a significant positive impact on liquidity position of commercial
banks in Ethiopia in which if there is an improvement in Economic performance of the
country there is an improvement in liquidity position of those banks in Ethiopia. Besides,
if the countries inflation rate increases, its impact on increasing the demand of their
customers to withdraw their fund is less than the reluctance of commercial banks to grant
loans and advances to their borrowers and overall the liquidity position of commercial
banks improved as their long-term loans and advances reduces from their portfolio of
assets.

The finding also reveals that profitability, tangibility and interest margin of commercial
banks in Ethiopia has insignificant effect on liquidity position of those banks by keeping
other factors remain constant.

5.2 Recommendations
In Ethiopian commercial banks from 2000 to 2021, the study's findings indicated that
factors such as bank size, capitalization, loan loss provisions, operational efficiency, real
GDP growth rate, and inflation rate were the major contributors to liquidity position.
71
Hence, focusing and taking the necessary action on these indicators could improve the
likelihood and impact of liquidity position in Ethiopian commercial banks. Based on the
findings of the study the following possible recommendations are forwarded for both the
commercial banks management and the supervisory board i.e. National bank of Ethiopia.
Furthermore, the study also gives some recommendation for future study.

5.2.1 Recommendation for Management of Commercial Banks in Ethiopia

❖ The recommendation forwarded from this finding can be explained by
considering the impact of both bank specific and macro-economic factors on
liquidity position of commercial banks as measured by their ability to fulfil the
unexpected withdrawal of funds and their ability to finance the increased demand
for loan.
❖ The capitalization of Ethiopian commercial banks is a significant bank-specific
element influencing their liquidity position. Therefore, Ethiopian commercial
banks should take seriously managing their capital and other equity balances.
Private commercial banks specifically need to control their dividend policy in
order to maintain a sufficient level of retained profit during a given accounting
period.
❖ Secondly, because operational efficiency has such a large and favourable impact
on the liquidity position of Ethiopian commercial banks, concentrating on and
reengineering the banks' business processes in accordance with those factors
could increase their operational efficiency and improve their liquidity position.
Since the commercial banks operating efficiency measured as a ratio of non-
interest income to non-interest expense, there are at least two measures, which
helps to become operationally efficient, i.e. improving their non-interest income
sources and reducing their non-interest expense as much as possible. For example,
the commercial banks should implement different mechanisms of business
process improvement like business process reengineering (BPR), total quality
management (TQM) approach to efficiently manage their non- interest expense
like employee salary, benefits, and general expenses in order to improve their
liquidity position.
72
❖ Bank size also has a significant negative impact on commercial banks liquidity
position and this finding suggests that the larger the size of the bank as measured
by their total asset, the lower the liquidity position exposure of those commercial
banks in Ethiopia. Therefore, the bank tries to balance their asset portfolio with
due consideration of their quality. Furthermore, the banks should implement
efficient and effective credit management policy to improve the quality of their
asset and to reduce the amount of long-term assets financed through short term
liabilities like customers deposit since it leads those banks under severe liquidity
crisis.
❖ Finally, loan loss provision is one of the powerful determinants of commercial
banks liquidity position determinants in Ethiopia. This result noted that
commercial banks in Ethiopia tries to hold some amount of gross loans and
advances as a provision in order to minimize the negative impact on liquidity
position which arises from the likelihood of those loans and advances become
default. To this end, bank should actively manage its collateral positions,
differentiating between encumbered and unencumbered assets. A bank should
monitor the legal entity and physical location where collateral is held and how it
may be mobilized in timely manner in case when the loan becomes default.
❖ Among the macroeconomic factors examined in this research commercial banks
in Ethiopia is advisable to pay attention for the macro-economic factors of general
economic performance and inflation rate. From the results, the real GDP growth
has a positive impact on the banks liquidity position in Ethiopia and therefore the
banks management tries to implement scenario planning to anticipate the
Ethiopian economic condition in developing their strategies for managing their
liquidity position. Besides, inflation rate also have a significant positive effect on
liquidity position of commercial banks in Ethiopia and the banks management
tries to consider the impact of inflation on their key business decisions like
provision of loans and their final decision should consider inflation rate as one of
the fundamental factors in their liquidity position. To sum up, commercial banks
in Ethiopia should not only be concentrated on bank specific determinants, but

73
 also macroeconomic factors must be incorporated in developing strategies to
effectively manage their liquidity position effectively.

5.2.2 Recommendation for supervisory bodies of Ethiopian commercial banks

(NBE)
The supervisors of Ethiopian commercial banks (NBE), should assess the adequacy of
both a bank's liquidity level management framework and its liquidity position and should
take prompt action if a bank is deficient in either area in order to protect depositors and to
limit potential damage to the financial system in particular and the whole economy in
general. To this end, NBE should strictly follow and revise the reserve requirement of
those commercial banks held at NBE as their capital adequacy ratio is one of the most
significant factors of determining the liquidity position of commercial banks in Ethiopia.

The Supervisors should also supplement their regular assessments of a bank’s liquidity
position management framework by monitoring a combination of internal reports,
prudential reports and market information. In this case, Communication should occur
regularly with the banks senior management during normal times, by increasing the
nature and frequency of the information sharing as appropriate during times of liquidity
crisis.

NBE should also intervene to require effective and timely remedial action by a bank to
address deficiencies in its liquidity position management processes. In order to achieve
this measures the NBE should have robust and flexible banking supervision policies like
having a culture of revising the banks’ reserve requirement, initial paid up capital
requirement for new commercial banks, the ceiling on the ratio of non-performing loan as
gross loans and advances, minimum liquidity requirement, the balances maintained on
the NBEs interbank settlement transaction etc. depending on the current and recent
market conditions.

5.2.3 Further Research

Using the stock technique of the liquid asset to deposit ratio as a proxy for measuring
liquidity position and taking into account some bank-specific and macroeconomic

74
factors, this study examines the factors that determine liquidity position for both private
and public commercial banks. Future research into the measurement of liquidity position
using various techniques, such as the liquidity coverage ratio, net stable funding ratio,
and other measurements for the explanatory variable, will be intriguing. Other studies
will also take into account other aspects, such as government regulation, industry
competitiveness, etc., that affect the liquidity position of Ethiopian commercial banks. It's
also advised to research the factors that influence operational and market risk in addition
to liquidity position, which are two additional banking concerns.

75
 References
Abiy, Z., Alemayehu, W., Daniel, T., Melese, G., & Yilma, S.(2009). Introduction to research
methods: Preparatory module for Addis Ababa University graduate programs. Graduate
Studiesand Research Office. Addis Ababa University
Abdulganiyy, A., Zainol Z., & Ahmed N.H.(2017). Determinants of liquidity risk in dual and
fully Islamic banking system: Evidence from Malaysia and Sudan. Islamic Business
School ISSN: 2321-8819
Ahmed, K., Ahmed,Z.,& Naqvi,I.H. 2011. Liquidity Risk and Islamic Banks: Evidence from
Pakistan. Inter disciplinary Journal of Research in Business, 1(9), pp.99- 102.
Akhtar,M.F., Ali,K., & Sadaqat,S. (2011). Liquidity risk management: A comparative study
between conventional and Islamic banks of Pakistan. Inter disciplinary Journal of
Research inBusiness, 1(1), pp.35-44.
Ali, S.S.( 2004). Islamic mode of finance and associated liquidity risks.
Ali, K., Akhtar, M.F., & Sadaqat, S.(2011). Financial and non-financial business risk
perspectives: Empirical evidence from commercial banks. Middle Eastern Finance
andEconomics ISSN: 1450-2889 1(1).
Angklomkliew,S., George,J., & Packer,F. (2009). Issues and developments in loan
lossProvisioning: The case of Asia. BIS Quarterly Review.
Arellano, M., and S. Bond. 1991. Some Test of Specification for Panel Data: Review of Economic
Studies 58: 277-297
Aspachs, O, Nier, E and Tiesset, M (2005). Liquidity, Banking Regulation and then Macro
economy: Evidence on bank liquidity holdings from a panel of UK-resident banks, Bank
of England Working Paper.
Asteway, F.(2018). The effect of income diversification on bank performance: Evidence from
Ethiopian commercial banks. MSc thesis in Accounting and Finance, Addis Ababa
University.
Baltagi, B.H. (2005), Economic Analysis of Panel data. 3rd Edition, John Wiley & Sons Inc., New
York.
Basel International Settlement (2008).Principles for sound liquidity risk management
andsupervision. Bank for International Settlements. Basel, Switzerland.
Belaid, F.,Bellouma, M., & Omir, A.(2016). Determinants of liquidity risk: Evidence from
Tunisian bank. International Journal of Emerging Research in Management and
Technology ISSN: 2278-9359, 5(9), pp. 95-104.
Bessis, J (2009). Risk Management in Banking, 3rd ed., John Wiley & Sons: Chichester
Błach, J.(2010). Financial risk identification based on the balance sheet information.
Brock, P and Suarez, R (2000). Understanding the behavior of bank spreads in Latin America;
Journal of Development Economics, Vol. 63, pp. 113–134
Brooks, C.(2014). Introductory econometrics for finance.3rd ed., The ICMA Center, University
of Reading, Cambridge University press.
Bunda, I and Desquilbet, J (2008).The bank liquidity smile across exchange rate
regimes’,International Economic Journal, Vol. 22, No. 3, pp. 361-386
Castro, V. (2012). Macroeconomic determinants of the credit risk in the banking system: the case
of GIPSI.
Chagwiza, W (2011). Zimbabwean Commercial Banks Liquidity And Its Determinants,
International Journal of Empirical Finance, Vol. 2, No. 2, PP.52-64

76
Cucilleni, D.(2013). Determinants of bank liquidity risk within the context of Euro
area.Interdisciplinary Journal of Research in Business ISSN: 2046-7141, 2(10), pp. 51-
64
Crosse, H and Hempel, G (1980).Management Policies for Commercial Banks, 3rd ed., Prentice
Hall Inc., Englewood Cliffs, New Jersey
Das, A., & Ghosh, S. (2007). Determinants of credit risk in Indian state owned banks: An
empirical investigation. Economic Issue, 12(2).
Diamond, D. W. (1984). Financial intermediation and delegated monitoring, Review of Economic
Studies, 5(1), pp.393-414
Diamond, D. W. (1983). Bank runs, deposit insurance, and liquidity. The Journal of Political
Economy, 91(3), pp. 401-419.
Driga, I. (2004). Means of reducing credit risk. Annals of the University of Petroşani, Economics,
IV, pp. 69-74.
Drigă, I.(2012). Financial risks analysis for a commercial bank in the Romanian banking system.
Estrada, F.(2011).Theory of financial risk.
Garr, D.K., 2013. Determinants of credit risk in the banking industry of Ghana.
DevelopingCountry Studies.ISSN 2225-0565, 3(11).
Ferrouhi, E., Lahadiri, A. (2014), Liquidity determinants of Moroccan banking industry.
International Research Journal of Finance and Economics, 118, 103-112.
Gilbert, A.T.(2013). Analyzing financial risk management in banks: Evidence of liquidity risk,
credit risk and capital risk in South Africa.Msc thesis in Banking and Business, Eastern
Mediterranean University.
Gujarati, D.N.(2004). Basic Econometrics, 4th edn, McGraw Hill, Boston
Harvey, C.(1996). Banking reform in Ethiopia. Retrieved from www.ids.ac.uk
Hempel, G.H, Simonson, D.G and Coleman, A.B (1994).Bank Management, 4th ed., John Wiley
& Sons: New York
Holton,G.A.( 2004). Defining risk. Financial Analysts Journal Volume 60
Horcher, K.A.(2005). Essentials of financial risk management.Hoboken, New Jersey.
Ibe, S. O. (2013). The impact of liquidity management on the profitability of banks in
Nigeria.Journal of Finance and Bank Management 1(1), pp. 37-48.
Iqbal, S., Chaudry, S.N., Iqbal, N., Din, M.Z. (2015). Impact of liquidity risk on firm specific
factors: A case of Islamic banks of Pakistan. Journal of Business and Management
Research ISSN: 2356-5764 pp. 256-260.
Ismael, R.(2010). Assessment of liquidity management in Islamic banking industry.International
Journal of Islamic and Middle Eastern Finance and Management, 3(2) pp. 147-167.
Kelliher, P.O.J., Wilmot, D., VI.J, & Klumpes, P.J.M.(2011). A common risk classification
system for the actuarial profession.Risk classification working party.
Kinfe, T (2019). Determinants of Ethiopian private banks liquidity risk, MSc in Accounting and
finance research report: Addis Ababa University
Kothari, C.R. (2004). Research methodology: methods and techniques (2nded.). New age
international (p) limited, New Delhi.
Liza, B. (2018). Determinants of bank liquidity risk of commercial banks of Ethiopia. MSc thesis
in Accounting and Finance, Addis Ababa University.
Loose, J (1993) A Historical Introduction to the Philosophy of Science. 3rd ed., Opus: Oxford

77
Lucchetta, M (2007). What Do Data Say About Monetary Policy, Bank Liquidity and Bank Risk
Taking?, Economic Notes by Banca Monte dei Paschi di Siena SpA, Vol. 36, No. 2,
pp.189-203
Malik, M.F and Rafique, A (2013). Commercial Banks Liquidity in Pakistan: Firm specific and
Macroeconomic Factors‟, The Romanian economic Journal, No.48, pp. 144-146
Mazur, A, W., & Szajt, M.(2015). Determinants of liquidity risk in commercial banks in the
European Union. Argumenta oeconomica 2(35) ISSN 1233-5835.
Milic D., & Solesa, D., 2017. The analysis of macroeconomic determinants of the banking sector
liquidity with role in financing agricultural sector.Economics of Agriculture.
Molyneux, P and Thornton, J (1992). Determinants of European Bank Profitability; Journal of
Banking and Finance, Vol. 16, No.6, pp.173-178
Moussa, M. A. B.(2015). The determinants of bank liquidity: Case of Tunisia. International
Journal of Economics and Financial Issues 5(1) pp. 249-259.
Mugenyah, L.O.(2015). Determinants of liquidity risk of commercial banks in Kenya. MBA
thesis, University of Nairobi.
Muranaga, J., & Ohsawa, M. (1997).Measurment of liquidity risk in the context of market risk
calculation.Institute for Monetary and Economic Study. Bank of Japan.
Nigist, M.(2015). Determinants of Banks Liquidity: Empirical evidence on Ethiopian commercial
banks. MSc thesis in Accounting and Finance, Addis Ababa University.
Nzotta , S. M. (2004.) Money, Bankung and Finance: Theory and Practice.
Pesaran, M.H., 2004. General Diagnostic Tests for Cross Section Dependence in Panels,
Cambridge Working Papers in Economics 0435, Faculty of Economics, University of
Cambridge.
Pestova, A., & Mamonov, M. (2014). Macroeconomic and bank specific determinants of credit
risk: Evidence from Russia.
Repullo, R (2004). Capital requirements, market power, and risk taking in banking‟, Journal of
Financial Intermediation, No 13, pp.156-182
Rochet, JC (2008). Liquidity regulation and the lender of last resort; Financial Stability Review,
Special issue on liquidity, No.11, pp. 45-52
Sahlemichael, M. (2009).Credit risk management system of Ethiopian commercial banks.MSc
thesis in Accounting and Finance, Addis Ababa University.
Samuel, A. (2015). Determinants of commercial banks profitability: The case Ethiopian
commercial banks. MSc thesis in Accounting and Finance, Addis Ababa University.
Sarafidis et, al (2007). A test of cross sectional dependence for a linear dynamic panel model with
regressors.
Saurina,J. (n.d).Loan Loss Provisions in Spain ; A working macro prudential tool. Banco de
Saunders, A. (1997). Financial institutions management: A modern perspective (2nd ed).
New York: Irwin.España.
Shakih, A.S. (2015). Panel data estimation of liquidity risk determinants in Islamic banks: a case
study of Pakistan.
Shah, S. Q. A., Khan, I., Shah, S. S. A., & Tahir, M. (2018). Factors affecting liquidity of banks:
Emperical evidence from the banking sector of Pakistan. Colombo Business Journal.9(1)
pp.1- 18.
Sheefeni, P.S., & Nyambe, N.J.(2016). Macroeconomic determinants of commercial bank‟s
liquidity risk in Namibia.European Journal of Business, Economics & Accountancy.
ISSN 2056-6018, 4(5)

78
Shen,C.H., Chen,Y.K., Kao,L.F., & Yeh ,C.Y.( 2009). Bank liquidity risk and performance.
Singh, A., & Sharma, A. K.(2016). An empirical analysis of macroeconomic and bank-specific
factors affecting liquidity of Indian banks.Future Business Journal 2(1) pp. 40 53.
Sori, T.(2014). Factors affecting banks profitability: An empirical study on Ethiopian private
commercial banks. MSc thesis in Accounting and Finance, Addis Ababa University.
Tewodros, M.(2013). Determinants of financial risk: Empirical evidence on private commercial
banks of Ethiopia. MSc thesis in Finance and Investment, Mekele University.
Thiagarajan, S., Ayyappan, S. & Ramachandran A.(2011). Credit risk determinants of public and
private sector banks in India.European Journal of Economics, Finance and
AdministrativeSciences ISSN 1450-2275 Issue 34.
Thong, T. Q.(2013). Factors affecting liquidity risk in the system of Vietnamese commercial
banks. Economics Article University of Economics HCMC.
Tirualem, M.(2009). Assessment of liquidity risk management of commercial banks of
Ethiopia.MSc thesis in Accounting and Finance, Addis Ababa University.
Tseganesh, T (2012). Determinants of Banks Liquidity and their Impact on Financial
Performance, MSc research report: Addis Ababa University
Tsion, F. (2015).Risk management practice of Ethiopian commercial banks.MSc thesis in
Accounting and Finance, Addis Ababa University.
Umar, M., Sun, G. (2016), Determinants of different types of bank liquidity: Evidence from
BRICS countries. China Finance Review International, 6(4), 380-403.
Vodova, P.(2013). Determinants of commercial bank‟s liquidity in Hungary.Financial Internet
Quarterly, ISSN 1734-039X, 9(3), pp. 64-71.
Vodová,P.(2010) .Determinants of commercial banks‟ liquidity in Poland. Proceedings of 30th
International Conference Mathematical Methods in Economics.
Vodova.(2011). Liquidity of Czech commercial banks and its determinants.International Journal
Of Mathematical Models And Methods In Applied Sciences,5(6)
Vodova, P.(2013). Determinants of commercial banks liquidity in Slovakia.
Waemustafa, W., & Sukri, S.(2015). Macroeconomic and bank specific dynamic determinants
credit risk in Islamic banks and conventional banks. International Journal of Economics
andFinancial Issues, 5(2), pp. 476-481.
Wooldridge, J. M. (2009). Introductory Econometrics: A modern approach (4ed). Michigan State
University.
Woods, M., & Dowd, K.(2008). Financial risk management for management accountants.The
society of management accountants of Canada, the American Institute of Certified Public
Accountants and The Chartered Institute of Management Accountants.
Yaacob, F.S., Rahman, A.A., & Karim, Z.A.(2016). The determinants of liquidity risk: A panel
study of Islamic banks in Malaysia. Journal of Contemporary Issues and Thought.
Zelalem, T.(2013). Determinants of non-performing loans: Evidence study on Ethiopian
commercial banks. MSc thesis in Accounting and Finance, Addis Ababa University.
Zribi,N., & Boujelbène,Y.(2011) .The factors influencing bank credit risk: The case of Tunisia.
Journal of Accounting and Taxation, 3(4), pp. 70-78.

79
 Appendices

Appendix I Data for Liquidity Position Model

Bank YEAR LIQDR LNTA LLP CAR ROA OPEFF TANG RGDP INFR IM
CBE 2000 0.4400 9.89 0.139 0.0650 0.0221 1.01 0.012 0.0607 0.0066 0.0400
CBE 2001 0.3500 9.98 0.184 0.0605 0.0009 0.51 0.011 0.083 -0.082 0.0300
CBE 2002 0.4300 10.01 0.246 0.0374 -0.0216 0.37 0.010 0.0151 0.0068 0.0100
CBE 2003 0.6900 10.09 0.29 0.0528 0.0235 1.9 0.009 -0.022 0.1367 0.0200
CBE 2004 0.7400 10.24 0.244 0.0535 0.0128 1.15 0.008 0.1357 0.0333 0.0200
CBE 2005 0.6900 10.41 0.212 0.0431 0.0187 2.42 0.007 0.1182 0.0997 0.0200
CBE 2006 0.7500 10.49 0.177 0.0420 0.0232 2.6 0.006 0.1083 0.123 0.0200
CBE 2007 0.7800 10.68 0.142 0.0971 0.0218 1.66 0.006 0.1146 0.1724 0.0200
CBE 2008 0.4700 10.83 0.061 0.0905 0.0290 2.51 0.006 0.1079 0.4436 0.0300
CBE 2009 0.3600 10.99 0.031 0.0848 0.0350 2.88 0.007 0.088 0.0848 0.0400
CBE 2010 0.2900 11.21 0.018 0.0749 0.0295 1.86 0.008 0.1255 0.0815 0.0300
CBE 2011 0.3600 11.65 0.025 0.0548 0.0304 1.78 0.007 0.1118 0.3325 0.0300
CBE 2012 0.2153 11.98 0.022 0.0486 0.0398 2.48 0.006 0.0865 0.236 0.0370
CBE 2013 0.2306 12.18 0.027 0.0471 0.0331 1.5 0.005 0.1058 0.0746 0.0400
CBE 2014 0.1676 12.41 0.027 0.0454 0.0313 1.32 0.007 0.1026 0.0689 0.0390
CBE 2015 0.1036 12.63 0.027 0.0437 0.0319 1.15 0.008 0.1039 0.0957 0.0440
CBE 2016 0.1108 12.86 0.039 0.0372 0.0242 0.68 0.019 0.0943 0.0663 0.0447
CBE 2017 0.1468 13.10 0.044 0.0909 0.0219 0.6 0.019 0.0956 0.1069 0.0401
CBE 2018 0.1151 13.26 0.036 0.0823 0.0101 0.12 0.019 0.0682 0.1383 0.0453
CBE 2019 0.1580 13.48 0.042 0.0705 0.0179 0.47 0.016 0.0836 0.1581 0.0419
CBE 2020 0.1644 13.62 0.052 0.0608 0.0124 0.45 0.016 0.0606 0.2036 0.0415
CBE 2021 0.1593 13.81 0.05 0.0543 0.0147 0.54 0.013 0.0564 0.2684 0.0376
AWB 2000 0.4653 6.63 0.031 0.1238 0.0232 0.67 0.030 0.0607 0.0066 0.0467
AWB 2001 0.4075 6.81 0.034 0.1147 0.0132 0.68 0.032 0.083 -0.082 0.0376
AWB 2002 0.4330 7.01 0.038 0.1178 0.0120 0.61 0.034 0.0151 0.0068 0.0380
AWB 2003 0.4770 7.24 0.055 0.0978 0.0110 0.75 0.041 -0.022 0.1367 0.0260
AWB 2004 0.5080 7.48 0.077 0.0876 0.0160 0.95 0.042 0.1357 0.0333 0.0270
AWB 2005 0.4460 7.71 0.062 0.1024 0.0190 0.92 0.034 0.1182 0.0997 0.0330
AWB 2006 0.3620 7.99 0.049 0.1029 0.0300 1.32 0.030 0.1083 0.123 0.0360
AWB 2007 0.3620 8.25 0.043 0.1132 0.0420 1.67 0.026 0.1146 0.1724 0.0520
AWB 2008 0.4770 8.48 0.046 0.1239 0.0330 1.52 0.027 0.1079 0.4436 0.0430
AWB 2009 0.6420 8.77 0.055 0.1168 0.0250 1.3 0.023 0.088 0.0848 0.0440
AWB 2010 0.6620 8.98 0.047 0.1184 0.0340 2.12 0.029 0.1255 0.0815 0.0290
AWB 2011 0.5230 9.22 0.036 0.1293 0.0400 2.5 0.025 0.1118 0.3325 0.0270
AWB 2012 0.3430 9.39 0.027 0.1349 0.0360 1.5 0.027 0.0865 0.236 0.0400
AWB 2013 0.2850 9.69 0.023 0.1251 0.0310 1.12 0.03 0.1058 0.0746 0.0400
80
AWB 2014 0.3360 9.90 0.023 0.1261 0.0340 1.35 0.032 0.1026 0.0689 0.0400
AWB 2015 0.2100 10.08 0.017 0.1295 0.0290 1.05 0.038 0.1039 0.0957 0.0430
AWB 2016 0.2540 10.30 0.015 0.1289 0.0280 0.85 0.039 0.0943 0.0663 0.0500
AWB 2017 0.2290 10.64 0.015 0.1111 0.0280 0.79 0.028 0.0956 0.1069 0.0520
AWB 2018 0.2680 10.92 0.008 0.0881 0.0310 0.61 0.044 0.0682 0.1383 0.0640
AWB 2019 0.1910 11.22 0.009 0.0939 0.0380 0.8 0.034 0.0836 0.1581 0.0620
AWB 2020 0.2050 11.40 0.017 0.1011 0.0320 0.59 0.033 0.0606 0.2036 0.0700
AWB 2021 0.1710 11.77 0.017 0.0955 0.0310 0.68 0.027 0.0564 0.2684 0.0590
DB 2000 0.5388 6.76 0.034 0.0890 0.0143 0.66 0.02 0.0607 0.0066 0.0405
DB 2001 0.3984 7.00 0.032 0.0845 0.0214 0.93 0.016 0.083 -0.082 0.0426
DB 2002 0.4274 7.30 0.031 0.0821 0.0186 0.95 0.015 0.0151 0.0068 0.0335
DB 2003 0.4004 7.60 0.039 0.0648 0.0155 0.78 0.013 -0.022 0.1367 0.0313
DB 2004 0.4004 7.89 0.037 0.0643 0.0240 1.03 0.015 0.1357 0.0333 0.0345
DB 2005 0.3604 8.14 0.032 0.0711 0.0233 0.85 0.013 0.1182 0.0997 0.0407
DB 2006 0.3112 8.42 0.027 0.0849 0.0334 1.09 0.013 0.1083 0.123 0.0468
DB 2007 0.3438 8.71 0.025 0.0901 0.0353 1.23 0.016 0.1146 0.1724 0.0487
DB 2008 0.4739 8.97 0.023 0.0933 0.0345 1.43 0.012 0.1079 0.4436 0.0492
DB 2009 0.5934 9.18 0.023 0.0934 0.0285 1.58 0.011 0.088 0.0848 0.0458
DB 2010 0.5180 9.42 0.022 0.0909 0.0293 1.87 0.013 0.1255 0.0815 0.0268
DB 2011 0.5258 9.59 0.02 0.0953 0.0334 2.07 0.013 0.1118 0.3325 0.0275
DB 2012 0.4105 9.77 0.021 0.1043 0.0405 1.96 0.015 0.0865 0.236 0.0369
DB 2013 0.3824 9.89 0.022 0.1036 0.0326 1.55 0.016 0.1058 0.0746 0.0351
DB 2014 0.3700 10.00 0.019 0.1183 0.0342 1.63 0.027 0.1026 0.0689 0.0354
DB 2015 0.2791 10.12 0.017 0.1181 0.0312 1.25 0.028 0.1039 0.0957 0.0395
DB 2016 0.3019 10.26 0.017 0.1175 0.0273 1.16 0.028 0.0943 0.0663 0.0372
DB 2017 0.1891 10.45 0.02 0.1153 0.0239 0.89 0.024 0.0956 0.1069 0.0403
DB 2018 0.1957 10.72 0.01 0.1291 0.0232 0.64 0.062 0.0682 0.1383 0.0508
DB 2019 0.1362 10.94 0.006 0.1218 0.0200 0.49 0.054 0.0836 0.1581 0.0505
DB 2020 0.1634 11.13 0.002 0.1218 0.0247 0.47 0.059 0.0606 0.2036 0.0601
DB 2021 0.1572 11.46 0.005 0.1069 0.0212 0.26 0.045 0.0564 0.2684 0.0613
BOA 2000 0.3340 6.58 0.015 0.1713 0.0217 0.91 0.008 0.0607 0.0066 0.0393
BOA 2001 0.2734 6.80 0.026 0.1641 0.0235 0.63 0.013 0.083 -0.082 0.0580
BOA 2002 0.4790 7.04 0.057 0.1235 -0.0020 0.4 0.011 0.0151 0.0068 0.0410
BOA 2003 0.4710 7.20 0.077 0.1118 0.0050 0.41 0.01 -0.022 0.1367 0.0370
BOA 2004 0.4930 7.37 0.076 0.1218 0.0260 0.65 0.012 0.1357 0.0333 0.0540
BOA 2005 0.4670 7.63 0.049 0.1235 0.0330 1.27 0.017 0.1182 0.0997 0.0500
BOA 2006 0.3590 7.95 0.031 0.1418 0.0350 0.97 0.013 0.1083 0.123 0.0570
BOA 2007 0.3760 8.13 0.047 0.1187 0.0220 0.58 0.012 0.1146 0.1724 0.0510
BOA 2008 0.4150 8.36 0.089 0.0983 0.0004 0.41 0.015 0.1079 0.4436 0.0510
BOA 2009 0.6000 8.61 0.098 0.0948 0.0210 0.88 0.014 0.088 0.0848 0.0530

81
BOA 2010 0.5764 8.75 0.074 0.0932 0.0239 1.43 0.012 0.1255 0.0815 0.0355
BOA 2011 0.4767 8.89 0.033 0.0908 0.0267 1.26 0.012 0.1118 0.3325 0.0441
BOA 2012 0.3726 9.02 0.026 0.1100 0.0279 1 0.012 0.0865 0.236 0.0468
BOA 2013 0.2849 9.23 0.02 0.1090 0.0288 1.1 0.026 0.1058 0.0746 0.0399
BOA 2014 0.3019 9.33 0.018 0.1356 0.0253 0.81 0.026 0.1026 0.0689 0.0490
BOA 2015 0.5642 9.52 0.015 0.1325 0.0234 0.76 0.064 0.1039 0.0957 0.0490
BOA 2016 0.2276 9.73 0.014 0.1262 0.0236 0.75 0.064 0.0943 0.0663 0.0529
BOA 2017 0.1661 10.14 0.013 0.1147 0.0271 0.75 0.049 0.0956 0.1069 0.0511
BOA 2018 0.1741 10.37 0.012 0.1327 0.0196 0.38 0.056 0.0682 0.1383 0.0668
BOA 2019 0.1391 10.58 0.013 0.1260 0.0218 0.45 0.05 0.0836 0.1581 0.0625
BOA 2020 0.1335 10.95 0.012 0.0998 0.0175 0.3 0.071 0.0606 0.2036 0.0675
BOA 2021 0.1360 11.55 0.015 0.0833 0.0167 0.43 0.059 0.0564 0.2684 0.0605
UB 2000 0.4605 4.96 0.011 0.2797 0.0274 0.86 0.042 0.0607 0.0066 0.0748
UB 2001 0.5349 5.37 0.007 0.2944 0.0280 0.9 0.033 0.083 -0.082 0.0818
UB 2002 0.7510 5.75 0.012 0.2803 0.0150 0.64 0.025 0.0151 0.0068 0.0750
UB 2003 0.6030 6.15 0.024 0.1940 0.0130 0.67 0.019 -0.022 0.1367 0.0530
UB 2004 0.5450 6.51 0.039 0.1424 0.0120 0.76 0.013 0.1357 0.0333 0.0470
UB 2005 0.5600 6.98 0.039 0.1165 0.0350 1.45 0.01 0.1182 0.0997 0.0520
UB 2006 0.4860 7.38 0.029 0.1194 0.0330 1.49 0.009 0.1083 0.123 0.0610
UB 2007 0.4920 7.69 0.03 0.1649 0.0340 1.08 0.015 0.1146 0.1724 0.0650
UB 2008 0.5670 8.09 0.027 0.1439 0.0340 1.19 0.01 0.1079 0.4436 0.0760
UB 2009 0.6870 8.44 0.031 0.1118 0.0240 1.09 0.009 0.088 0.0848 0.0760
UB 2010 0.6930 8.68 0.036 0.1081 0.0330 1.64 0.007 0.1255 0.0815 0.0450
UB 2011 0.5900 8.95 0.028 0.1167 0.0340 1.79 0.008 0.1118 0.3325 0.0470
UB 2012 0.4200 9.08 0.023 0.1254 0.0360 1.38 0.011 0.0865 0.236 0.0670
UB 2013 0.2600 9.21 0.019 0.1203 0.0300 1.07 0.011 0.1058 0.0746 0.0700
UB 2014 0.3600 9.38 0.014 0.1326 0.0250 0.81 0.015 0.1026 0.0689 0.0740
UB 2015 0.2300 9.57 0.012 0.1174 0.0210 0.65 0.025 0.1039 0.0957 0.0760
UB 2016 0.2200 9.76 0.013 0.1200 0.0210 0.62 0.028 0.0943 0.0663 0.0750
UB 2017 0.1900 9.99 0.012 0.1149 0.0190 0.52 0.033 0.0956 0.1069 0.0710
UB 2018 0.2000 10.24 0.013 0.1054 0.0230 0.53 0.044 0.0682 0.1383 0.0780
UB 2019 0.1300 10.48 0.005 0.1080 0.0240 0.39 0.044 0.0836 0.1581 0.0760
UB 2020 0.1500 10.67 0.007 0.1245 0.0230 0.34 0.053 0.0606 0.2036 0.0850
UB 2021 0.1600 10.90 0.009 0.1198 0.0210 0.41 0.063 0.0564 0.2684 0.0790
WB 2000 0.6354 6.24 0.027 0.0973 0.0068 0.79 0.018 0.0607 0.0066 0.0322
WB 2001 0.5033 6.37 0.044 0.0995 0.0109 0.76 0.015 0.083 -0.082 0.0505
WB 2002 0.4430 6.47 0.049 0.0991 0.0110 0.67 0.022 0.0151 0.0068 0.0470
WB 2003 0.4460 6.79 0.051 0.1046 0.0140 0.71 0.017 -0.022 0.1367 0.0350
WB 2004 0.4670 7.04 0.058 0.1132 0.0320 0.94 0.014 0.1357 0.0333 0.0520
WB 2005 0.4810 7.39 0.051 0.1114 0.0350 1.08 0.013 0.1182 0.0997 0.0430

82
WB 2006 0.3720 7.72 0.048 0.1129 0.0370 1.1 0.011 0.1083 0.123 0.0460
WB 2007 0.4850 8.15 0.044 0.1158 0.0390 1.21 0.009 0.1146 0.1724 0.0460
WB 2008 0.6080 8.32 0.059 0.1468 0.0370 1.28 0.01 0.1079 0.4436 0.0500
WB 2009 0.7820 8.54 0.061 0.1634 0.0390 1.8 0.011 0.088 0.0848 0.0590
WB 2010 0.7739 8.66 0.04 0.1832 0.0411 1.85 0.014 0.1255 0.0815 0.0454
WB 2011 0.6951 8.99 0.045 0.1659 0.0468 1.95 0.014 0.1118 0.3325 0.0441
WB 2012 0.4847 9.03 0.024 0.1922 0.0410 1.62 0.037 0.0865 0.236 0.0481
WB 2013 0.3675 9.25 0.022 0.1761 0.0366 1.12 0.035 0.1058 0.0746 0.0509
WB 2014 0.3585 9.35 0.017 0.1860 0.0291 0.96 0.047 0.1026 0.0689 0.0512
WB 2015 0.2479 9.53 0.016 0.1761 0.0279 0.81 0.047 0.1039 0.0957 0.0522
WB 2016 0.2796 9.69 0.016 0.1733 0.0251 0.71 0.046 0.0943 0.0663 0.0537
WB 2017 0.2785 9.95 0.014 0.1602 0.0287 0.82 0.045 0.0956 0.1069 0.0552
WB 2018 0.1974 10.22 0.017 0.1397 0.0328 0.76 0.049 0.0682 0.1383 0.0630
WB 2019 0.1818 10.30 0.022 0.1442 0.0217 0.48 0.048 0.0836 0.1581 0.0634
WB 2020 0.2115 10.55 0.02 0.1338 0.0245 0.56 0.043 0.0606 0.2036 0.0654
WB 2021 0.1539 10.59 0.046 0.1265 0.0033 0.32 0.04 0.0564 0.2684 0.0720
NB 2000 1.1154 5.06 0.00 0.2532 0.0063 0.67 0.019 0.0607 0.0066 0.0177
NB 2001 0.4425 5.82 0.00 0.1845 0.0486 1.44 0.009 0.083 -0.082 0.0530
NB 2002 0.4840 6.28 0.14 0.1854 0.0300 1.23 0.009 0.0151 0.0068 0.0450
NB 2003 0.4150 6.79 0.04 0.1412 0.0180 0.81 0.007 -0.022 0.1367 0.0380
NB 2004 0.3980 7.13 0.038 0.1387 0.0330 1.28 0.006 0.1357 0.0333 0.0430
NB 2005 0.3790 7.46 0.041 0.1293 0.0310 1.16 0.006 0.1182 0.0997 0.0440
NB 2006 0.3000 7.61 0.039 0.1406 0.0310 1.15 0.015 0.1083 0.123 0.0440
NB 2007 0.3700 7.87 0.034 0.1630 0.0330 1.02 0.016 0.1146 0.1724 0.0510
NB 2008 0.5400 8.20 0.038 0.1639 0.0360 1.12 0.012 0.1079 0.4436 0.0640
NB 2009 0.7080 8.48 0.046 0.1516 0.0360 1.31 0.012 0.088 0.0848 0.0730
NB 2010 0.7434 8.69 0.039 0.1535 0.0373 1.6 0.012 0.1255 0.0815 0.0480
NB 2011 0.7066 8.87 0.041 0.1646 0.0377 1.68 0.011 0.1118 0.3325 0.0570
NB 2012 0.5106 9.02 0.027 0.1846 0.0372 1.49 0.012 0.0865 0.236 0.0580
NB 2013 0.3388 9.12 0.025 0.1822 0.0344 1.02 0.014 0.1058 0.0746 0.0763
NB 2014 0.2418 9.28 0.021 0.1828 0.0299 1.02 0.02 0.1026 0.0689 0.0688
NB 2015 0.1839 9.49 0.015 0.1642 0.0281 0.69 0.023 0.1039 0.0957 0.0823
NB 2016 0.2397 9.67 0.018 0.1591 0.0268 0.53 0.025 0.0943 0.0663 0.0879
NB 2017 0.1999 9.95 0.016 0.1405 0.0241 0.61 0.025 0.0956 0.1069 0.0836
NB 2018 0.1797 10.19 0.015 0.1267 0.0216 0.43 0.072 0.0682 0.1383 0.0870
NB 2019 0.1421 10.43 0.010 0.1308 0.0239 0.4 0.069 0.0836 0.1581 0.0844
NB 2020 0.1586 10.66 0.008 0.1363 0.0274 0.42 0.065 0.0606 0.2036 0.0844
NB 2021 0.1705 10.90 0.008 0.1311 0.0252 0.27 0.068 0.0564 0.2684 0.0926
CBO 2005 0.8060 4.86 0 0.8682 -0.0170 0.02 0.008 0.1182 0.0997 0.0830
CBO 2006 0.9080 5.41 0.008 0.5446 -0.0240 0.02 0.022 0.1083 0.123 0.0350

83
CBO 2007 0.6320 6.05 0.013 0.3066 0.0070 0.34 0.019 0.1146 0.1724 0.0480
CBO 2008 0.6710 6.52 0.012 0.2187 0.0210 0.5 0.019 0.1079 0.4436 0.0800
CBO 2009 0.4590 6.93 0.014 0.1528 0.0030 0.28 0.029 0.088 0.0848 0.0500
CBO 2010 0.6210 7.48 0.025 0.1069 0.0180 0.82 0.02 0.1255 0.0815 0.0520
CBO 2011 0.6100 7.82 0.02 0.0983 0.0220 1.2 0.025 0.1118 0.3325 0.0400
CBO 2012 0.4400 8.21 0.014 0.1137 0.0330 1.26 0.024 0.0865 0.236 0.0500
CBO 2013 0.7700 8.79 0.017 0.1065 0.0370 1.46 0.014 0.1058 0.0746 0.0400
CBO 2014 0.3200 8.90 0.018 0.1483 0.0490 1.5 0.018 0.1026 0.0689 0.0600
CBO 2015 0.3300 9.35 0.026 0.1231 0.0330 0.86 0.017 0.1039 0.0957 0.0600
CBO 2016 0.2500 9.28 0.053 0.1142 0.0004 0.31 0.023 0.0943 0.0663 0.0700
CBO 2017 0.2400 9.78 0.032 0.0856 0.0150 0.43 0.022 0.0956 0.1069 0.0500
CBO 2018 0.3100 10.31 0.024 0.0795 0.0180 0.51 0.019 0.0682 0.1383 0.0500
CBO 2019 0.2600 10.64 0.034 0.0787 0.0180 0.47 0.018 0.0836 0.1581 0.0500
CBO 2020 0.1500 10.87 0.028 0.0974 0.0250 0.57 0.019 0.0606 0.2036 0.0600
CBO 2021 0.2000 11.31 0.017 0.0873 0.0200 0.61 0.02 0.0564 0.2684 0.0500
LIB 2007 0.9549 5.58 0.013 0.5075 -0.0360 0.14 0.045 0.1146 0.1724 0.0289
LIB 2008 0.9720 6.35 0.012 0.2979 -0.0002 0.39 0.027 0.1079 0.4436 0.0618
LB 2009 0.6292 6.86 0.01 0.2013 0.0034 0.48 0.018 0.088 0.0848 0.0401
LB 2010 0.7277 7.22 0.016 0.1773 0.0345 1.33 0.014 0.1255 0.0815 0.0430
LB 2011 0.7000 7.50 0.014 0.1952 0.0280 1.25 0.011 0.1118 0.3325 0.0500
LB 2012 0.6000 7.81 0.015 0.1793 0.0350 1.39 0.009 0.0865 0.236 0.0500
LB 2013 0.4700 7.99 0.013 0.1842 0.0410 1.41 0.01 0.1058 0.0746 0.0500
LB 2014 0.4200 8.19 0.013 0.1738 0.0290 0.94 0.013 0.1026 0.0689 0.0600
LB 2015 0.3400 8.68 0.017 0.1403 0.0320 1.02 0.011 0.1039 0.0957 0.0500
LB 2016 0.2895 9.00 0.02 0.1318 0.0280 0.83 0.011 0.0943 0.0663 0.0566
LB 2017 0.3044 9.30 0.02 0.1320 0.0280 0.62 0.01 0.0956 0.1069 0.0632
LB 2018 0.2594 9.57 0.025 0.1263 0.0310 0.55 0.01 0.0682 0.1383 0.0685
LB 2019 0.2201 9.92 0.019 0.1255 0.0310 0.6 0.01 0.0836 0.1581 0.0657
LB 2020 0.2638 10.37 0.024 0.1095 0.0250 0.32 0.028 0.0606 0.2036 0.0704
LB 2021 0.1500 10.38 0.048 0.1130 0.0100 0.09 0.031 0.0564 0.2684 0.0783
OB 2009 0.9400 5.77 0.009 0.3364 -0.0190 0.29 0.04 0.088 0.0848 0.0050
OB 2010 0.7660 7.02 0.011 0.1895 0.0270 1.15 0.033 0.1255 0.0815 0.0230
OB 2011 0.5570 7.58 0.011 0.1509 0.0290 1.43 0.025 0.1118 0.3325 0.0240
OB 2012 0.5230 7.93 0.013 0.1570 0.0210 0.98 0.029 0.0865 0.236 0.0320
OB 2013 0.3940 8.27 0.015 0.1400 0.0200 0.77 0.027 0.1058 0.0746 0.0440
OB 2014 0.3730 8.72 0.013 0.1217 0.0310 0.92 0.018 0.1026 0.0689 0.0570
OB 2015 0.2090 9.16 0.013 0.1041 0.0280 0.83 0.019 0.1039 0.0957 0.0510
OB 2016 0.2300 9.33 0.018 0.1168 0.0150 0.47 0.045 0.0943 0.0663 0.0660
OB 2017 0.2470 9.70 0.019 0.1022 0.0210 0.74 0.032 0.0956 0.1069 0.0500
OB 2018 0.2920 10.08 0.008 0.1089 0.0360 0.84 0.027 0.0682 0.1383 0.0610

84
OB 2019 0.1930 10.37 0.016 0.1168 0.0270 0.6 0.024 0.0836 0.1581 0.0600
OB 2020 0.1920 10.43 0.013 0.1359 0.0260 0.44 0.024 0.0606 0.2036 0.0710
OB 2021 0.2060 10.64 0.032 0.1312 0.0230 0.37 0.024 0.0564 0.2684 0.0710
ZB 2009 0.7968 6.14 0.01 0.1957 -0.0277 0.55 0.066 0.088 0.0848 0.0053
ZB 2010 0.8830 6.96 0.016 0.1502 0.0672 2.44 0.032 0.1255 0.0815 0.0123
ZB 2011 0.6082 7.39 0.018 0.1492 0.0635 2.83 0.014 0.1118 0.3325 0.0168
ZB 2012 0.5020 7.78 0.018 0.1172 0.0431 2.12 0.02 0.0865 0.236 0.0188
ZB 2013 0.4483 8.09 0.085 0.1519 0.0334 1.41 0.018 0.1058 0.0746 0.0182
ZB 2014 0.4928 8.28 0.088 0.1673 0.0513 2.08 0.019 0.1026 0.0689 0.0319
ZB 2015 0.3019 8.49 0.055 0.1568 0.0348 1.45 0.014 0.1039 0.0957 0.0307
ZB 2016 0.4025 8.91 0.044 0.1359 0.0331 1.48 0.01 0.0943 0.0663 0.0262
ZB 2017 0.4201 9.18 0.046 0.1360 0.0293 1.6 0.052 0.0956 0.1069 0.0238
ZB 2018 0.3958 9.43 0.032 0.1399 0.0245 1.2 0.042 0.0682 0.1383 0.0297
ZB 2019 0.2174 9.59 0.022 0.1588 0.0356 1.37 0.049 0.0836 0.1581 0.0388
ZB 2020 0.3029 9.83 0.017 0.1688 0.0445 1.23 0.046 0.0606 0.2036 0.0580
ZB 2021 0.3169 10.13 0.016 0.1783 0.0437 1.31 0.04 0.0564 0.2684 0.0497
BRB 2010 0.8271 5.94 0.01 0.2670 -0.0221 0.19 0.022 0.1255 0.0815 0.0195
BRB 2011 0.7619 6.82 0.011 0.1641 0.0328 1.41 0.015 0.1118 0.3325 0.0344
BRB 2012 0.6098 7.16 0.012 0.1838 0.0306 1.49 0.012 0.0865 0.236 0.0350
BRB 2013 0.4644 7.69 0.015 0.1736 0.0213 1.1 0.009 0.1058 0.0746 0.0314
BRB 2014 0.4879 7.94 0.016 0.1971 0.0180 0.74 0.01 0.1026 0.0689 0.0604
BRB 2015 0.4052 8.34 0.014 0.1742 0.0297 1 0.013 0.1039 0.0957 0.0479
BRB 2016 0.2939 8.88 0.015 0.1473 0.0468 1.09 0.009 0.0943 0.0663 0.0581
BRB 2017 0.3161 9.26 0.014 0.1797 0.0373 1.02 0.015 0.0956 0.1069 0.0586
BRB 2018 0.2464 9.55 0.017 0.1565 0.0267 0.6 0.015 0.0682 0.1383 0.0661
BRB 2019 0.2043 9.86 0.018 0.1458 0.0276 0.65 0.019 0.0836 0.1581 0.0608
BRB 2020 0.1683 9.97 0.016 0.1604 0.0273 0.59 0.043 0.0606 0.2036 0.0744
BRB 2021 0.1390 10.20 0.043 0.1392 0.0081 0.42 0.047 0.0564 0.2684 0.0692
BUB 2010 0.9658 6.17 0.01 0.3522 0.0001 0.64 0.022 0.1255 0.0815 0.0180
BUB 2011 0.7700 6.66 0.011 0.2976 0.0310 1.17 0.02 0.1118 0.3325 0.0440
BUB 2012 0.4470 7.22 0.011 0.2103 0.0260 0.99 0.013 0.0865 0.236 0.0380
BUB 2013 0.3750 7.66 0.012 0.1759 0.0260 0.79 0.01 0.1058 0.0746 0.0510
BUB 2014 0.4150 8.01 0.012 0.1716 0.0310 0.87 0.021 0.1026 0.0689 0.0570
BUB 2015 0.2340 8.41 0.011 0.1506 0.0360 0.83 0.02 0.1039 0.0957 0.0610
BUB 2016 0.2330 8.83 0.017 0.1409 0.0330 0.79 0.017 0.0943 0.0663 0.0570
BUB 2017 0.2760 9.19 0.017 0.1378 0.0240 0.72 0.019 0.0956 0.1069 0.0480
BUB 2018 0.2680 9.47 0.015 0.1523 0.0280 0.63 0.013 0.0682 0.1383 0.0650
BUB 2019 0.2160 9.58 0.016 0.1772 0.0340 0.72 0.021 0.0836 0.1581 0.0700
BUB 2020 0.2190 9.85 0.018 0.1629 0.0260 0.48 0.033 0.0606 0.2036 0.0740
BUB 2021 0.1600 10.16 0.019 0.1468 0.0300 0.48 0.029 0.0564 0.2684 0.0770

85
AB 2011 0.7910 6.12 0 0.3449 -0.0116 0.5 0.025 0.1118 0.3325 0.0150
AB 2012 0.6000 7.12 0 0.2135 0.0290 1.09 0.029 0.0865 0.236 0.0340
AB 2013 0.3870 7.58 0.013 0.1732 0.0240 0.91 0.023 0.1058 0.0746 0.0470
AB 2014 0.3420 8.07 0.012 0.1414 0.0220 0.84 0.02 0.1026 0.0689 0.0460
AB 2015 0.2460 8.43 0.013 0.1563 0.0320 0.96 0.023 0.1039 0.0957 0.0500
AB 2016 0.2330 8.73 0.014 0.1553 0.0270 0.74 0.024 0.0943 0.0663 0.0570
AB 2017 0.2680 9.07 0.011 0.1504 0.0230 0.79 0.028 0.0956 0.1069 0.0510
AB 2018 0.3080 9.42 0.017 0.1463 0.0300 0.82 0.021 0.0682 0.1383 0.0510
AB 2019 0.2810 6.05 0.015 0.1627 0.0370 1.07 0.02 0.0836 0.1581 0.0540
AB 2020 0.2480 9.91 0.012 0.1529 0.0280 0.64 0.014 0.0606 0.2036 0.0580
AB 2021 0.2010 10.31 0.015 0.1409 0.0340 0.62 0.018 0.0564 0.2684 0.0650
ADB 2012 0.7511 6.05 0.01 0.3824 0.0344 1.1 0.02 0.0865 0.236 0.0221
ADB 2013 0.6843 6.82 0.01 0.2456 0.0406 1.54 0.024 0.1058 0.0746 0.0344
ADB 2014 0.5443 7.14 0.011 0.2498 0.0410 1.42 0.038 0.1026 0.0689 0.0437
ADB 2015 0.4412 7.45 0.012 0.2595 0.0391 1.27 0.03 0.1039 0.0957 0.0448
ADB 2016 0.4911 7.81 0.013 0.2585 0.0395 1.2 0.024 0.0943 0.0663 0.0510
ADB 2017 0.4065 8.14 0.012 0.2219 0.0314 1.04 0.018 0.0956 0.1069 0.0451
ADB 2018 0.3489 8.35 0.009 0.2141 0.0296 0.97 0.024 0.0682 0.1383 0.0499
ADB 2019 0.3360 8.62 0.011 0.2016 0.0327 1.02 0.018 0.0836 0.1581 0.0490
ADB 2020 0.3198 8.78 0.009 0.2090 0.0355 1.05 0.017 0.0606 0.2036 0.0563
ADB 2021 0.2954 9.09 0.009 0.1828 0.0353 1.17 0.014 0.0564 0.2684 0.0473

86
 Appendix II: Results for Pooled OLS

. reg liqdr lnta llp car roa opeff tang rgdp infr im

Source SS df MS Number of obs = 257

F( 9, 247) = 84.00
Model 7.90233815 9 .878037572 Prob > F = 0.0000
Residual 2.58185592 247 .010452858 R-squared = 0.7537
Adj R-squared = 0.7448
Total 10.4841941 256 .040953883 Root MSE = .10224

liqdr Coef. Std. Err. t P>|t| [95% Conf. Interval]

lnta -.0597748 .0050578 -11.82 0.000 -.0697368 -.0498128

llp .8710408 .2214082 3.93 0.000 .434952 1.30713
car .5355643 .1121062 4.78 0.000 .3147584 .7563703
roa -4.745047 .8721905 -5.44 0.000 -6.462926 -3.027168
opeff .1636702 .024843 6.59 0.000 .114739 .2126014
tang -.3086439 .497902 -0.62 0.536 -1.289319 .6720311
rgdp .6804423 .2352035 2.89 0.004 .217182 1.143703
infr .2696421 .0650881 4.14 0.000 .1414435 .3978407
im -1.121629 .5551129 -2.02 0.044 -2.214988 -.0282708
_cons .7553057 .061677 12.25 0.000 .6338258 .8767856

87
 Appendix III: Results for Fixed Effect Model

. xtreg liqdr lnta llp car roa opeff tang rgdp infr im, fe robust

Fixed-effects (within) regression Number of obs = 257

Group variable: id Number of groups = 15

R-sq: within = 0.7716 Obs per group: min = 10

between = 0.3239 avg = 17.1
overall = 0.7318 max = 22

F(9,14) = 127.02
corr(u_i, Xb) = -0.1751 Prob > F = 0.0000

(Std. Err. adjusted for 15 clusters in id)

Robust
liqdr Coef. Std. Err. t P>|t| [95% Conf. Interval]

lnta -.054596 .0092675 -5.89 0.000 -.0744729 -.0347192

llp .9162413 .3487158 2.63 0.020 .1683201 1.664162
car .6437205 .1994378 3.23 0.006 .215969 1.071472
roa -2.798235 1.724128 -1.62 0.127 -6.496121 .8996514
opeff .1362648 .057637 2.36 0.033 .0126457 .2598839
tang .4542178 .7834274 0.58 0.571 -1.226067 2.134503
rgdp .5150913 .2350766 2.19 0.046 .0109021 1.01928
infr .2678177 .0615894 4.35 0.001 .1357216 .3999138
im -3.376567 1.040203 -3.25 0.006 -5.60758 -1.145554
_cons .7778882 .0895268 8.69 0.000 .5858724 .9699041

sigma_u .05018753
sigma_e .09903428
rho .204338 (fraction of variance due to u_i)

88
 Appendix IV Results for Random Effect Model

. xtreg liqdr lnta llp car roa opeff tang rgdp infr im, re

Random-effects GLS regression Number of obs = 257

Group variable: id Number of groups = 15

R-sq: within = 0.7644 Obs per group: min = 10

between = 0.5443 avg = 17.1
overall = 0.7531 max = 22

Wald chi2(9) = 765.77

corr(u_i, X) = 0 (assumed) Prob > chi2 = 0.0000

liqdr Coef. Std. Err. z P>|z| [95% Conf. Interval]

lnta -.0592524 .0054177 -10.94 0.000 -.0698709 -.0486338

llp .866606 .2243485 3.86 0.000 .4268911 1.306321
car .5462725 .114623 4.77 0.000 .3216156 .7709293
roa -4.391499 .8977784 -4.89 0.000 -6.151113 -2.631886
opeff .1570042 .0253276 6.20 0.000 .107363 .2066455
tang -.1764217 .5156452 -0.34 0.732 -1.187068 .8342244
rgdp .6589924 .2338841 2.82 0.005 .200588 1.117397
infr .2714469 .0647594 4.19 0.000 .1445209 .3983729
im -1.529949 .5955201 -2.57 0.010 -2.697147 -.3627514
_cons .764443 .0630762 12.12 0.000 .640816 .88807

sigma_u .01668377
sigma_e .09903428
rho .02759712 (fraction of variance due to u_i)

89
 Appendix V Results for Correlation Analysis

. corr
(obs=257)

id year liqdr lnta llp car roa opeff tang rgdp infr im

id 1.0000
year -0.1290 1.0000
liqdr 0.1125 -0.6112 1.0000
lnta -0.1851 0.6649 -0.6819 1.0000
llp -0.1177 -0.3969 0.1959 0.0958 1.0000
car 0.1472 -0.0874 0.4572 -0.5586 -0.3202 1.0000
roa -0.1024 0.2049 -0.1926 0.1832 -0.1529 -0.2891 1.0000
opeff -0.2116 -0.2387 0.3337 -0.0278 0.2226 -0.1987 0.6326 1.0000
tang -0.0190 0.4201 -0.3056 0.2003 -0.2954 0.0595 -0.1654 -0.3833 1.0000
rgdp 0.0281 -0.0290 0.2428 -0.0844 -0.1175 0.1250 0.2273 0.3413 -0.1848 1.0000
infr 0.0192 0.3111 0.0202 0.2065 -0.1295 0.0211 0.1216 0.1136 0.0443 0.0113 1.0000
im 0.3915 0.4146 -0.4924 0.2290 -0.3501 0.0528 0.1305 -0.4553 0.2603 -0.1323 0.1016 1.0000

Appendix VI Results for Normality test

. predict uhat
(option xb assumed; fitted values)

. sktest uhat

Skewness/Kurtosis tests for Normality

joint
Variable Obs Pr(Skewness) Pr(Kurtosis) adj chi2(2) Prob>chi2

uhat 257 0.0505 0.3190 4.85 0.0884

90
 Appendix VII Results for Heteroskedasticity Test

. xttest3

Modified Wald test for groupwise heteroskedasticity

in fixed effect regression model

H0: sigma(i)^2 = sigma^2 for all i

chi2 (15) = 69.87

Prob>chi2 = 0.0000

Appendix VIII Results for Multicollinearity Test

. vif

Variable VIF 1/VIF

opeff 4.12 0.242975

roa 3.33 0.300209
im 2.14 0.466573
car 1.94 0.515139
lnta 1.83 0.545583
llp 1.67 0.598886
tang 1.34 0.746041
rgdp 1.24 0.804324
infr 1.15 0.867453

Mean VIF 2.09

91
 Appendix IX Results for Model Specification Test (Link test)

. linktest

Source SS df MS Number of obs = 257

F( 2, 254) = 393.27
Model 7.92498008 2 3.96249004 Prob > F = 0.0000
Residual 2.55921399 254 .010075646 R-squared = 0.7559
Adj R-squared = 0.7540
Total 10.4841941 256 .040953883 Root MSE = .10038

liqdr Coef. Std. Err. t P>|t| [95% Conf. Interval]

_hat .8213113 .1244336 6.60 0.000 .5762584 1.066364

_hatsq .2072624 .1382613 1.50 0.135 -.0650221 .4795468
_cons .0320004 .0265348 1.21 0.229 -.0202559 .0842567

Appendix X Results for Crosses Sectional Dependence

. xtcsd, pesaran abs

Pesaran's test of cross sectional independence = 8.743, Pr = 0.0000

Average absolute value of the off-diagonal elements = 0.340

. ssc install xthrtest

checking xthrtest consistency and verifying not already installed...
all files already exist and are up to date.

Appendix XI Results for Serial Correlation

. xthrtest

Heteroskedasticity-robust Born and Breitung (2016) HR-test as postestimation

Panelvar: id
Timevar: year

Variable HR-stat p-value N maxT balance?

Post Estimation 1.84 0.066 15 22 unbalanced

Notes: Under H0, HR ~ N(0,1)

H0: No first-order serial correlation.
Ha: Some first order serial correlation.

92
Appendix XII Results for Over Identifying Restrictions (Sargan test)
. estat sargan
Sargan test of overidentifying restrictions
H0: overidentifying restrictions are valid

chi2(194) = 227.13
Prob > chi2 = 0.0518

Appendix XIII Results for First and Second Order Autocorrelation (Arellano-Bond)

. estat abond

Arellano-Bond test for zero autocorrelation in first-differenced errors

Order z Prob > z

1 -1.8256 0.0679
2 1.2604 0.2075

H0: no autocorrelation

93
Appendix XIV Results for System Dynamic Panel-Data Estimation

. xtdpdsys liqdr lnta llp car roa opeff tang rgdp infr im

System dynamic panel-data estimation Number of obs = 242

Group variable: id Number of groups = 15
Time variable: year
Obs per group: min = 9
avg = 16.13333
max = 21

Number of instruments = 225 Wald chi2(10) = 1028.29

Prob > chi2 = 0.0000
One-step results

liqdr Coef. Std. Err. z P>|z| [95% Conf. Interval]

liqdr
L1. .3850038 .0496213 7.76 0.000 .2877478 .4822599

lnta -.0309239 .0064289 -4.81 0.000 -.0435242 -.0183236

llp 1.453846 .208651 6.97 0.000 1.044898 1.862794
car .7304497 .1714562 4.26 0.000 .3944018 1.066498
roa -.8985094 .9778536 -0.92 0.358 -2.815067 1.018048
opeff .0612041 .024486 2.50 0.012 .0132124 .1091957
tang -.317566 .4961249 -0.64 0.522 -1.289953 .654821
rgdp .4271104 .1835239 2.33 0.020 .0674102 .7868106
infr .1770032 .0543281 3.26 0.001 .0705222 .2834843
im .1926836 .6974923 0.28 0.782 -1.174376 1.559743
_cons .2551182 .0765201 3.33 0.001 .1051414 .4050949

Instruments for differenced equation

GMM-type: L(2/.).liqdr
Standard: D.lnta D.llp D.car D.roa D.opeff D.tang D.rgdp D.infr D.im
Instruments for level equation
GMM-type: LD.liqdr
Standard: _cons

94
Appendix XV Results for Cross-sectional time-series FGLS regression

. xtgls liqdr liqdr1 lnta llp car roa opeff tang rgdp infr im, igls
Iteration 1: tolerance = 0

Cross-sectional time-series FGLS regression

Coefficients: generalized least squares

Panels: homoskedastic
Correlation: no autocorrelation

Estimated covariances = 1 Number of obs = 257

Estimated autocorrelations = 0 Number of groups = 15
Estimated coefficients = 11 Obs per group:
min = 10
avg = 17.13333
max = 22
Wald chi2(10) = 1023.68
Log likelihood = 252.8108 Prob > chi2 = 0.0000

liqdr Coef. Std. Err. z P>|z| [95% Conf. Interval]

liqdr1 .2682062 .0351018 7.64 0.000 .199408 .3370044

lnta -.0625999 .0044913 -13.94 0.000 -.0714027 -.0537971
llp .8137491 .1960837 4.15 0.000 .4294322 1.198066
car .2742664 .1049394 2.61 0.009 .0685889 .4799439
roa -3.335135 .7936148 -4.20 0.000 -4.890591 -1.779678
opeff .1672105 .0219903 7.60 0.000 .1241103 .2103107
tang -.4664611 .4411135 -1.06 0.290 -1.331028 .3981056
rgdp .6102484 .2083513 2.93 0.003 .2018873 1.01861
infr .2560569 .0576287 4.44 0.000 .1431068 .369007
im -.39995 .5002571 -0.80 0.424 -1.380436 .5805358
_cons .7569149 .0545829 13.87 0.000 .6499344 .8638953

95