Content uploaded by Mohd. Islam
Author content
All content in this area was uploaded by Mohd. Islam on May 14, 2019
Content may be subject to copyright.
Volume 10, Number 2, December 2018
Liquidity and Risk Taking Behavior of Commercial
Banks in Bangladesh
Mohd. Anisul Islam
Rownak Jahan**
Abstract: We use a sample of conventional banks to test the hypothesis on
the association between liquidity risk and risk-taking behavior of banks in
Bangladesh. Using suitable econometric model (pooled regression with
time fixed effect), findings reveal that greater amount of liquidity as well
as lower liquidity risk inspires banks to take more risk measured by risk-
weighted assets. However, high liquidity risk is associated with high risk
evidenced by Z-score and liquidity creation. Liquidity risk has been
proxied by deposits to assets ratio while bank risk has been indicated by
risk-weighted assets, loan loss provisions, Z-score, and liquidity creation.
Besides, banks with large asset and capital base take more risk because of
their absorbability. In contrast, profitable banks reduce their exposure to
overall risk. Moreover, economic expansion reduces the risk exposure of
banks while price level instability accelerates the risks of banks. Findings
of this study have implications for banking industry regulator prescribing
risk management guideline and advocating greater liquidity for banks
under Basel III.
Keywords: Liquidity, Risk exposure, Deposits, GDP, and BASEL III.
1.0 Introduction
Liquidity risk is one of the most discussed issues in banking and financial
environment as a whole. Nowadays liquidity crunch has become so visible in
current global crisis. Liquidity risk poses a lot of challenges for bank overall
risk management. Identifying the appropriate factors influencing liquidity risk
is very crucial for the banking sector to walk along the path of financial
stability. In order to apprising the dependability and consistency of current
banking sector, it is essential to accumulate precise information of liquidity
risk and it is also vital to recognize how to handle liquidity fluctuation.
Simply liquidity risk can be defined as the degree of inability to convert assets
of the bank in cash without degrading values, without significant loss or large
Corresponding Author, Lecturer, Department of Finance, University of Dhaka,
Dhaka-1000, Email: ai.fin@du.ac.bd
**Research Associate, Credit Rating Agency of Bangladesh Limited, E-mail:
rjreema3@gmail.com
208 Journal of Banking & Financial Services
price change. A vital challenge for banks is to maintain an optimal liquidity
point. Both high and low liquidity can generate grounds of risk for banking
sector. Liquidity problem arises when a bank’s current assets is not enough to
meet current liability, or when the depositors wants to redeem their savings,
banks cash balance is not enough to repay them. This problem is known as
liquidity risk. Liquidity risk is followed by solvency risk as a bank finds an
imbalance between their asset and liability side. Both excess liquidity and
shortage of liquidity create risk for bank. High liquidity by sourcing more
deposits may work as a cushion to the bank to meet the obligation of the
depositors very easily but at the same time it may make bankers confident to
take more risk in terms of risk weighted asset, loan loss provision, non-
performing loan and Z score. These parameters eventually increases overall
risk position of banks.
Lower liquidity risk reflects excess deposit or liquid asset. This motivates
bank manager to lend loan more aggressively without judging borrowers’
credit worthiness or repayment ability. This practice or tradition eventually
accumulates credit risk. Credit risk means failure of the borrowers to repay
their loan according to agreed payment schedule. The management of
liquidity risk involves estimating risk adjusted return by identifying the
exposure within adequate parameters. The lending decision of conventional
bank is taken based on asymmetric information flow from market. The entire
banking industry cannot extract same level of information. Some bank can dig
out the real clues other banks fail. The precise information flow helps taking
correct lending decision followed by superior profit margin. The incorrect
lending decision causes liquidity risk and bank faces losses as the borrower
fail to repay bank’s loan which in turn raises the extent of classified loan.
Banking industry cannot avoid the adverse selection, moral hazard and
asymmetric information regarding borrower’s attitude.
In the past liquidity risk had been and in future will remain one of the most
critical risks of conventional commercial banking sector. Different regulatory
and supervisory committee, including Basel committee suggests a global level
of maintaining a sustainable liquidity buffer to face against short term
liquidity crisis. Newly implemented BASEL III requires bank to fund their
funding need through long term financing and to maintain higher capital
requirement which efficiently help to control bank risk-taking. Liquidity risk
arises from both internal and external economic condition. Internal factor
deals with bank’s internal management quality and their expectation towards
profit and earning, bank asset and loan portfolio diversification, capital
composition along with leverage position, banking regulation. These factors
Liquidity and Risk Taking Behavior of Commercial Banks in Bangladesh 209
can be controlled but yet unknown to the policy makers. The external factor
includes economic recession, government policy regarding credit market,
GDP and inflation etc.
The present study seeks to find out the impact or influence of liquidity risk on
bank risk-taking tendency. That is, to see if there is any is the relationship
exists between liquidity position and overall risk position. All our background
studies and theoretical overviews indicate that there is relationship between
liquidly risk and overall risk but the direction of relationship varies. Through
this study we will observe if our findings are consistent to literature review or
we like to explore a new phase of learning.
The broad aim of this paper is to identify the significant contribution of
liquidity risk to overall risk of the bank covering conventional commercial
banks in Bangladesh. The rest of the paper is structured as follows: Section 2
gives an overview of the Bangladeshi banking industry for the period
covering 2013 to 2016. Section 3 presents literature review on the influence of
liquidity risk on overall risks of banks. Section 4 and 5 describe data
characteristics and formation of model respectively. Section 6 comprises
discussion of the findings. Section 7 includes conclusions.
2.0 Overview of Banking Industry in Bangladesh
Banking industry of Bangladesh plays a major role to accelerate the economic
activities of Bangladesh. There are four categories of bank in the industry.
State-owned commercial banks, specialized banks, private commercial bank,
and foreign commercial banks. The banking industry of Bangladesh has gone
through both challenges and development. From the year 1982 to 2000,
banking sector has undergone various reforms. The reforms were essential to
ensure sustainable growth of them. Regulator has started implementing Basel
IIIrecently for achieving better prudence and supervision. Some recent
parameters have been presented through following tables and graph.
Table 1: Financial Highlights of Banking Sector in 2017
Industry
Balance Sheet
Outstanding
Deposit
Outstanding
Loan
Bank Asset
to GDP
Ratios
Bank Account
Percentage of
Population
$145 Billon
$113 Billon
$89 Billon
67.2%
31%
Source: Bangladesh Bank
210 Journal of Banking & Financial Services
Table 1 can be presented as an evidence of this tremendous growth of modern
banking system. Over the years, technology has evolved at a significant
degree in the banking industry of Bangladesh has. New banking system has
given a commendable look to existing banking sector. Agent banking, plastic
money, and digital financial services can be great examples to represent the
remarkable development of banking industry.
2.1 Liquidity trends of Bangladeshi Banks
The data of the total asset, total deposit and deposit to asset ratio of banks in
Bangladesh from the year 2011 to 2016 are given in Table 2 and total deposits
to assets ratio has been graphically presented as bar chart in Figure 1.
Table 2: Total Assets, Total Deposits, and Deposits to Assets Ratio of
Banking Industry
Year
Total Deposits
(Billion BDT)
Total Assets
(Billion BDT)
Deposit to
Asset Ratio
2011
4509.7
5867.6
0.769
2012
5396
7030.70
0.767
2013
6273
8000.2
0.784
2014
6965.1
9143.1
0.762
2015
7928.6
10314.7
0.769
2016
8933.9
11626.6
0.768
Source: Bangladesh Bank
Over 2012-2016, the percentage of credit growth experienced an average
growth of 16%. The highest growth is presented by construction sector,
consumer finance and working financing sector capital sector. 56% of loan
was disbursed by keeping real-estate asset as collateral. Investment in
government bond has growth of 18% and deposit experienced 14% growth
over the five years from 2012 to 2016. During this five year period equity
capital of banking sector has grown up 10%. The amount of risk weighted
asset has decreased reflecting that banks are shifting to low risk asset. Most of
the banks have been able to keep their capital above regulatory position.
Liquidity and Risk Taking Behavior of Commercial Banks in Bangladesh 211
Figure 1: Deposits to Assets Ratio of Banking Industry
Deposits to assets ratio soared in the year of 2013 and declined to lowest level
in 2014 which can be attributed to the significant gap between interest rates of
deposits products of banks and savings certificate issued by the Bangladesh’s
government.
3.0 Literature review
Free cash flow theory: Jenson has established free cash flow theory in 1986.
The theory expresses that when a firm holds excess cash the manager of the
firm take weak investment (lower or negative NPV) decision. This is relevant
to banking decision, when a bank has high deposit as reflector of high
liquidity it disburses more loan which in turn increases credit risk.
Agency cost theory: Agency conflict arises from different objectives, goals or
interest of different stakeholders of the firm. There are 3 types of agency
conflicts exist between: 1. Managers vs. shareholders 2. Managers vs. debt
holders 3.Shareholders vs. debt holders. As loan disbursement is used as
performance benchmark for manger, managers are always motivated to take
large numbers of lending decision irrespective of risk measurement. Mangers
are less likely to consider long-term risk which they have been creating by
giving risky loan for short term high profit. Only high quality audit with deep
investigations can hold the manager responsible for their action. During
financial turmoil, shareholders influence manager to finance capital by debt
rather than launching new equity whereas bondholders demand high interests
charges as return, creditors have lawful right to get reimbursement if principal
212 Journal of Banking & Financial Services
and interest payment are failed. Managers need to pay attention in the firm as
professionally as possible in order to meet the interest payments to the
creditor and realize the shareholder wealth maximization.
Risk and return theory: Liquidity and profitability are treated as the opposite
sides of same coin. They have an inverse relationship according to trade off
theory. To insure liquidity or stability, bank need to hold more liquid asset. As
bank hold more liquid asset in form of cash and bank deposit, bank mislay the
chance of investing these liquid asset in other profitable sector. Moreover,
bank has to pay more maintenance cost to manage this high liquid asset. For
both these trouble, the profit margin of bank goes down. On the other hand, if
bank do not hold sufficient level of liquid asset, bank cannot meet its short
term liability as withdrawal demand by the customer. So, bank invite financial
instability. As a result bank faces always a dilemma between high liquidity
and high profitability. To find an optimum level for both liquidity and
profitability is a real challenge for bank.
Berger and DeYoung (1997) found that lagged risk-weighted asset is
significant and positively related to risk measured by non-performance loan
(NPL) as a percentage of total loans. They reasoned that a relatively risky
loan portfolio would result in Liquidity risk. Merton (1977) proved that
deposit works as put options when liability is less than the asset. In the time of
global financial crisis, people hold cash money as deposit than to invest in the
risky asset to avoid instability. Daimon and Dybvig (1983) showed deposit
exposes bank credit risk. Adrian and Shin (2010) describe that excess capacity
by depository banks allows them to give more loan by taking high risk, and
the borrower fails to pay loans during the financial crisis. Wagner (2007)
develops a theoretical model to express a relationship between asset liquidity
and stability. He found that high liquidity reduce stability in financial crisis
not during financial expansion. Lucchetta (2007) proved that when risk free
rate is high, bank take more risk. So, high liquidity gives confidence to do
more interbank lending and borrowing.
Acharya and Naqvi (2012) and Wagner (2007) described that short-term
liquidly has a risky impact on long-term stability and welfare cost of the
society. Drehmann and Nikolaou (2013) described liquidity as a risk indicator
as it leads to aggressive bidding. Funding liquidity risk is inversely related
with the liquidity position of banker according to their study because shortage
of liquid fund disrupts bank’s ability to conduct its regular transactions
smoothly. Liquidity risk is also supported by the free cash flow theory of
Jensen (1986). Imbierowicz and Rauch (2014) find that high liquidity risk is a
Liquidity and Risk Taking Behavior of Commercial Banks in Bangladesh 213
contributor to bank failure. Konishi and Yasuda (2004) describe the factors
affecting bank risk-taking based on data from Japan. This research paper
showed that if banks maintain adequate capital, they need to bear less risk; if
the retired government higher officials become the member of board of
directors they have no effect on bank risk-taking; there is a nonlinear
relationship between shareholders ownership and bank risk; high shareholding
reduces bank risk initially but due to asset substitution effect, the bank risk
raises again; and if the franchise value decreases this risk of bank increases.
They also argued that achieving a high amount of capital adequacy ratio
reduce bank risk talking ability.
King (2013) concluded a higher interest cost to hold high net stable fund ratio
will reduce bank profitability and increase bank risk. Repullo (2011) shows
through his model that the bank takes more risk when the lending rate of
Central bank is higher. Shim (2013) showed the as capital buffer reduces the
chance of default as banks holding more risky asset maintain a high capital
buffer. So, if a bank holds high risky asset, it cannot go for further risky
investment with a higher level. Jeitschko and Jeung (2005) showed
relationship between bank risk and capital buffer is influenced by other
factors like issuer, depositor and bank management. In case of well-
capitalized banks, owners prefer less risky investment to avoid the greater
amount of loss as he/she already involved with high capital, but manger like
to engage in risky lending to earn large profit to improve his performance.
According to Thakor (1996), capital buffer motivates banks to reduce risk.
According to Carmona (2007), smaller banks face more instability due to
illiquid asset and lead to failure. Calem and Rob (1999) showed that bank
capital and risk-taking have a U shaped relation. Small bank reduces its risk
by increasing their capital but a well-capitalized bank actually faces more risk
with high capital in the long run as it influences them to take an aggressive
lending decision.
According to Bertay et al. (2013) big bank not necessarily face more risk due
to low funding risk. Large bank has more capital and asset to fight against any
kind of systematic risk and smaller shocks. Demsetz and Strahan (1997)
showed the even bank-specific risk is also reduced by the larger asset holding
of the big bank. The large banks are less risky so they have higher stability
score evidenced by higher Z score established by Mercieca et al., (2007) and
Stiroh (2004). Delis et al., (2014) and Cornett et al. (2011) proved that illiquid
asset reduce aggressive lending and bank tend to hold more liquidity position
214 Journal of Banking & Financial Services
during financial turmoil. According to Ivashina and Scharfstein (2010), larger
banks take high credit risk. There is a positive relationship between
profitability and overall risk. They have found that off-balance sheet activities
positively impact both credit and overall risk. They also found that bank with
high liquidity can take high risk.
Khan et al. (2016) analyzed the relationship between liquidity risk and bank
risk-taking. Using quarterly data for US bank holding companies from 1986
to 2015, they concluded an increase in deposit funding increases banks overall
risk as evidenced by higher risk-weighted assets and higher negative Z-scores.
They also concluded that small bank size and high capital buffers restrict
banks to some extent from taking more risk even they face lower liquidity
risk.
4.0 Methodology
To fulfill the objective to test the relationship between banks’ liquidity risk
and overall risk we have used panel regression and run Panel OLS model for
testing hypotheses of the study. To capture the macroeconomic effect, we
have used year dummy variables. The analysis applied panel data regression
with fixed time effect and clustered robust Standard error. All the variables
have lag effect because based on review of literature, we have assumed that
impact of variables like change in degree of liquidity of this year will have
impact on the risk-taking propensity of bank in next year. Analytical
framework: the base line model in order to test the relationship between bank
risk and bank liquidity is-
Riski,t = αLiquidityi,t-1 + βControlsi,t-1 + δt + εi,t
Riski,t= Asset risk+ Overall risk,
Where Asset risk= Risk weighted asset + Loan loss provision
andOverall risk= Z score + Liquidity creation
α = Coefficient of independent variable t-1
β= Coefficient of control variable t-1
δ= Coefficient of time fixed effect
ε= error term
All the dependent and Independent variables are selected on the basis of
previous study and on which required data can be collected.
Liquidity and Risk Taking Behavior of Commercial Banks in Bangladesh 215
4.1 Proxies for banks’ risk measure
To run this study we need dependent variables that can reflect risk position of
bank including liquidity risk, asset risk and overall risk. Previous studies have
considered various variables as dependent one. For this study we have chosen
four dependent variables as proxy of bank risk.
RWA: Risk weighted asset can be described as minimum amount of asset
bank must need to hold to face risk of insolvency. It is used as proxy of bank
asset risk. RWA can be considered as a reflector of asset quality and risk as it
is measured according to Basel accord all around the banking industry. We
have taken a ratio form using risk weighted asset divided by total asset.
LLP: Loan loss provision is also used as proxy of asset risk. Loan loss
provision is the fund created by the bank to cover up loan default. Every year
this fund is keep aside as provision against loss. If LLP increases, it indicates
that overall risk of bank is rising. We have taken loan loss provision relative
to average assets. Cebenoyan & Strahan (2004) used standard deviation of
LLP to total loan, we have used ration of LLP to total asset to show the
aggressive lending assessment of banks.
Z-score: Usually Z score reflects the stability position of nay bank. The higher
the Z score the safer the bank is from instability or volatility. When overall
risk of bank increases Z score decreases. With an objective to keep our
calculation simple and sustain uniformity with other dependent variable we
have used – Z score. By the minus sign the whole interpretation has turn
reverse. Now the Z score reflects risk position rather than stability. The
higher the – Z score the riskier the bank is and vice versa. When overall risk
of bank increases Z score decreases. We have used natural logarithm of the Z-
scores.
LC: Liquidity creation: Liquidity creation measure, introduced by Berger &
Bouwman (2009), reflects how effectively bank can finance liquid liability by
liquid asset. If LC increases liquidity risk decreases but overall risk of bank
increases by intermediation risk. Following Berger & Bouwman (2009),
liquidity creation variable has been calculated as follows:
Liquidity Creation = (0.5×Illiquid Assets + 0.5 ×Liquid Liabilities −0.5
×Liquid Assets −0.5 ×Illiquid Liabilities −0.5 ×Equity)
216 Journal of Banking & Financial Services
4.2 Bank liquidity risk measure
We have taken deposit to total asset ratios as proxy of liquidity risk. The
higher the ratio the higher liquidity position bank is holding which gives
confidence to take more risk in asset portfolio.
4.3 Control Variables (Bank Specific and Macroeconomic)
We have divided control variables in two part bank specific and macro
variables. These variables have lag impact, meaning change in amount of any
of these control variables current year will have impact on bank risk position
next year. If described other way to find out the influential factors behind the
risk of current year we have to consider the variable of t-1 year (previous
year).
Asset: The size of the bank is very crucial factor for credit risk. The bank with
larger size can take more risk as it already has huge financial support to act as
back up for financial downturn. Larger bank can disburse more risky loan.
We have converted total asset with LN function to have linear relation.
Loan: We have taken the ratio of loan to total asset. The higher the ratio, the
higher the risk position. The higher the amount of loan disbursement, the
higher the probability of default, so credit risk surges.
ROA: Return on asset (ROA) indicates how much net profit is earned from
banks asset. It reflects bank management expectation of earning. If the
management expects to earn a higher ROA it will provide higher amount of
loan reflecting higher risk. So there is positive relation between ROA and
overall risk- taking position.
Equity: Equity has been measured as the amount of total equity as a percent of
total asset. The higher the ratio the safer the bank is. At the same time bank
feels more confidant to make risky lending decision. The bank with higher
capital is less vulnerable to financial crises or economic downturn and vice
versa.
IB Spread: The higher the spread, the bank has to face higher cost of fund, so
profit position is downgraded expressing high risk.
GDP: High GDP is an indicator of prosperous economic condition. During
economic development employment level is high, so earning of the people is
high. With the higher earning ability people is more able to take loan and
repay the interest and principal amount. So creditworthiness of borrower
Liquidity and Risk Taking Behavior of Commercial Banks in Bangladesh 217
improves better due to enhanced financial strength. As a result overall risk
goes down. During economic expansion bank risk is low.
Inflation: High inflation reduces the value of money as a result financial
strength of borrower deteriorates. So banks face high credit as well as interest
rate risk. The higher the inflation rate, the higher the probability of default,
higher the bank risk. During contraction of economy bank risk is high.
We have applied panel ordinary least square regression method. We have
used lag effect of time (T-1) for all explanatory variables. We have considered
time fixed effect in the first model (Table 5) where we have not used
macroeconomic variables.OLS for panel data has some sort of limitations, as
it does not consider qualitative impact as divergence effect of management
strategies among the banks of Bangladesh, also ignore year wise business
cycle effect or does not considers macroeconomic effect to overcome these
limitations we have use year dummy to incorporate time effect. As in
Bangladesh there is no significant contradiction regarding management
strategy we do not consider bank dummy.
4.4 Hypotheses of the study
Whether liquidity risk has impact on overall risk for banking industries is one
of the most argued questions in risk management. There is a general principle
that liquidity position has a certain positive impact on overall performance of
banks. After Acharya and Naqvi (2012), and Cheng et al., (2015) has
empirically showed that liquidity can be a dangerous weapon of threat when
become excess.
By reviewing all theories and literatures we can draw a line of judgment that
liquidity position has impact on overall risk position of bank. Now the main
question of this study is whether there is any relationship between liquidity
risk and overall risk-taking tendency of bank. If relationship exists, what is
the direction of the relationship and magnitude of the relationship reflecting
how strong or profound the relation is?
H1: There is significant relationship between liquidity risk and overall risk of
conventional commercial banks in Bangladesh.
H2: Bank facing lower liquidity risk will have propensity to take more risk.
H3: Economic growth increases and price level stability decreases overall
risk and asset risk of banks.
218 Journal of Banking & Financial Services
5.0 Data
We have collected financial data of 22 conventional commercial banks over 5
years (2013-2017). We have collected data from each year specific annual
reports of sample banks. We have gone through balance sheet, income
statement and related detail notes. To avoid outlier problem all variables
except macro variable have been winsorized. We have winsorized data in
95% level to make dataset free from outliers because they can distort the
actual relationship between variables. We have derived the mode and
magnitude of dependency of overall risk of the bank on various influential
independent and controlling factors. These variables include bank specific and
macro factors. The variables are described in detail below in section 5 as well
as in Appendix A.
Table 3reports the summary statistics of the yearly data for 22 conventional
banks from 2013to 2017. The top and bottom 5% of all observations for all
variables have been winsorized to limit the extreme values. Most of the
conventional banks of Bangladesh hold almost 80% risk-weighted assets.
Banks are keeping 1.67% of their asset as loan loss provision. Bank having
average Z-score of -3.5 reflects relatively less stable position of banking
industry but standard deviation 0.93 means that while a few banks are facing
risk, most of them are not. Statistics shows that average liquidity position in
terms of liquid assets to deposits is around 80%. Banks owners are investing
equity capital which is only 8% of the total assets, showing high reliance on
leverage fund. Bangladesh has successfully maintains average real GDP
growth rate of 6.62% in the 2013-2017 period but average inflation rate of
6.34% was above the target level of policy makers.
Table 3: Summary Statistics
Variable
Observations
Mean
Std dev.
Minimum
Maximum
RWA
110
0.7954
0.1058
0.5656
0.9859
LLP
110
0.01663
0.0125
0.0013
0.0761
Z SCORE
110
-3.4544
0.9269
-5.079
-1.220
LC
110
0.0360
0.0948
13740
0.3194
DEPOSITS
110
0.7822
0.0527
0.6384
0.8707
Liquidity and Risk Taking Behavior of Commercial Banks in Bangladesh 219
ASSET
110
26.0061
0.3219
25.3357
26.6185
LOAN
110
0.6626
0.0545
0.5004
0.7529
EQUITY
110
0.0834
0.0175
0.0507
0.1256
ROA
110
0.0101
0.0038
0.0028
0.0244
IBSPREAD
110
0.0119
0.0088
0.0011
0.0265
INFLATION
110
0.0634
0.0055
0.0566
0.0717
GDP
110
0.0662
0.0054
0.0601
0.073
Note: Risk weighted Asset is measured by risk weighted activity divided by total asset, Loan
Loss Provision is measured by annual loan loss amount of percentage of total asset, Z score is
measured by log of return on asset added to equity by asset divided by standard deviation of
asset (multiply the values for banks’ Z- scores by −1 to facilitate a more consistent
interpretation amongst risk proxies), Liquidity creation is measured by liquidity creation
divided by to total asset, Deposit is measured by total deposit by total asset, Asset is measured
by log of total asset, Loan is measured by total loan by total asset, Equity is measured by total
equity by asset, Return of total asset is measured by net interest income divided by total asset,
Interbank spread is measured by Scheduled Banks weighted average deposit rate subtracted
from weighted average call money market rates, GDP is real GDP growth rate, Inflation has
been proxied by CPI index.
Source: Authors’ Estimations
Table 4 represents the correlation coefficients of all 14 variables used for 22
conventional banks over the year from 2013-2017. The top and bottom 5% of
all observations for all variables have been winsorized to limit the extreme
values.
From the displayed pair wise correlation we can derive that risk weighted
asset is positively correlated with loan loss provision, deposit, asset, loan,
equity and inflation and negatively with Z score ROA and GDP. LLP is
positively related with loan, equity, IB Spread, Inflation, and GDP and
negatively related with Deposit, asset, ROA.
From the above pair-wise correlation table we have seen that no correlation
exceeds more that 55%. So it can be declared that the data set is free from
multicollinearity problem. VIF calculates the magnitude of severity of
multicollinearity in OLS model. VIF result has been presented in Appendix B.
As average VIF of all variables is 1.52, it is clear that the variables are free
from multicollinearity.
220 Journal of Banking & Financial Services
Table 4: Correlation Matrix
Note: Risk weighted Asset is measured by risk weighted activity divided by total asset, Loan
Loss Provision is measured by annual loan loss amount of percentage of total asset, Z score is
measured by log of return on asset added to equity by asset divided by standard deviation of
asset (multiply the values for banks’ Z- scores by −1 to facilitate a more consistent
interpretation amongst risk proxies),Liquidity creation is measured by liquidity creation divided
by to total asset, Deposit is measured by total deposit by total asset, Asset is measured by log of
total asset, Loan is measured by total loan by total asset, Equity is measured by total equity by
asset, Return of total asset is measured by net interest income divided by total asset, Interbank
spread is measured by Scheduled Banks weighted average deposit rate subtracted from
weighted average call money market rates GDP is GDP is real GDP growth rate, Inflation has
been proxied by CPI index.
Source: Authors’ Estimations
6.0 Results and Discussions
6.1 Liquidity risk and bank risk (Table: 5)
6.1.1 Risk weighted asset
We have run our OLS model with 95% confidence level. The result of F test
is 0.000 which is less than 0.05 significance level, reflecting the validity of
this model. The model is significant at 1% significance level which reflects
the excellent level of fitness of the estimation model. In other way we can say
that we can reject our null hypothesis (No impact of liquidity risk on overall
risk position on banking sector of Bangladesh). The value of adjusted R
square 52% reflects that 52% variation in bank risk proxied by RWA can be
explained by the independent (liquidity risk) and other bank specific control
RWA
LLP
Z-
Scores
LC
Deposits
Asset
Loan
Equity
ROA
IBSpread
Inflation
GDP
RWA
1.0000
LLP
0.0376
1.000
Z-Scores
-0.237
-0.17
1.0000
LC
-0.420
0.096
0.1352
1.0000
Deposits
0.0484
-0.11
-0.467
-0.225
1.0000
Asset
0.1376
-0.19
0.4438
0.0931
-0.442
1.00
Loan
0.3343
0.134
-0.090
0.1327
-0.153
0.16
1.00
Equity
0.4750
0.101
-0.034
-0.373
-0.304
0.05
-0.9
1.000
ROA
-0.049
-0.01
0.1538
-0.073
-0.126
0.02
0.04
0.255
1.000
IBSpread
-0.036
0.010
0.0190
0.1431
-0.170
0.19
0.11
0.021
0.129
1.0000
Inflation
0.0712
0.009
-0.057
-0.142
0.1054
-0.1
-0.0
-0.02
-0.085
-0.1115
1.0000
GDP
-0.148
0.059
0.1067
0.2141
-0.277
0.54
0.41
-0.26
-0.049
0.3975
0.0548
1.0000
Liquidity and Risk Taking Behavior of Commercial Banks in Bangladesh 221
variables. P value is significant for deposit, asset, loan and equity at 1%
significance level. So deposit, asset, loan and equity have a significant
relationship with risk weighted asset. ROA is significant variable at 10% to
explain the variation in asset risk - RWA.
The regression model shows that there is a significant positive relation
between deposit to asset ratio and risk weighted asset (asset risk) at
1%significance level. Banks having more liquidity (lower funding risk)
intends to take more risk. When bank has more liquidity, it works as a
motivation to disburse more loans without judging credit worthiness of the
borrower precisely. The compensation of the bank manager is positively
related to growth of loan disbursement. So with high amount deposits, bank
managers become overconfident to take aggressive loan disbursement
decision. As a result asset risk of the bank increases with high liquidity (lower
liquidity risk) which is consistent with the findings of the existing theories.
However, there is a negative relationship between liquidity risk and overall
risk-taking tendency of banking indicated by –Z score and liquidity creation
which will be discussed shortly.
Table 5: Panel OLS Regression for Liquidity Risk and Bank Risk
Asset risk
Overall risk
1
2
3
4
RWAt
LLPt
-Z-
scorest
LCt
Deposits t-1
0.8897***
(0.2038)
-0.0252
(0.0366)
-9.7174**
(4.3116)
-0.4954*
(0.2891)
Assets t-1
0.1571***
(0.0407)
-0.0187*
(0.0103)
1.1181**
(0.4994)
-0.0400
(0.0539)
Loan t-1
0.8850***
(0.1955)
0.0192
(0.0380)
-2.8148
(2.0883)
-0.0129
(0.2498)
Equity t-1
3.0221**
(0.6887)
0.1546
(0.1181)
-15.9385*
(8.6956)
-
2.0310***
(0.7245)
ROA t-1
-3.7040*
(1.9488)
-0.1026
(0.3454)
24.4468
(20.3146)
-0.08651
(2.4190)
Constant
-4.7571***
(1.0996)
0.4929
(0.2402)
-21.4622
(14.37)
1.6478
(1.389)
Time fixed
effect
Yes
Yes
Yes
Yes
222 Journal of Banking & Financial Services
Firm fixed
effect
No
No
No
No
Observations
88
88
88
88
Adjusted R-sq
0.52085
0.07575
0.3699
0.0904
Test of
Probability
F(8,
21)=27.62
(0.0000)
F(8,
21)=0.96
(0.4894)
F(8, 21)=4.52
(0.0026)
F(8,
21)=3.74
(0.0074)
Note: This table represents the panel regression results where the dependent variables are the
measures of banks’ asset risk as proxied by the ratios of risk-weighted assets to total assets
(RWA) and loan loss provisions to total loans (LLP) and banks’ overall risk as proxied by the
natural logarithm of the Z-scores (Z-scores) and liquidity creation The independent variable of
interest is banks’ liquidity risk as proxied by the ratio of total deposits to total assets (Deposit).
Control variables used are natural logarithm of total assets (Asset), return on assets (ROA) and
the ratios of total loans to total assets (Loan) and total equity to total assets (Equity). The
sample is based on the annual data of conventional commercial banks of Bangladesh. Time
fixed effects are considered and bank fixed effects are not considered in the regressions. P-
values are computed using heteroscedasticity-robust standard errors clustered for banks and are
presented in parentheses. *, **, and *** represent statistical significance at the 10%, 5%, and
1% level, respectively.
Source: Authors’ estimates
The regression model shows that there is a significant positive relationship
between size of the bank proxied by LNASSET and risk of the bank proxied
by RWA at 1%significance level. Banks with large asset base have more
ability to absorb risk by its large amount of diversified asset size compared to
small bank. Smaller banks are more risk averse than larger banks in terms of
risk-weighted assets.
The regression model shows that there is a significant positive relation
between loan relative to assets and risk of the bank proxied by RWA at
1%significance level. Bank holding more loan amount definitely face more
risk of default. Banks with large loan portfolio incur high credit risk compared
to banks with small loan portfolio. Higher growth of loan disbursement
volume adds more risk to the risk portfolio of banks.
The regression model shows that there is a significant positive relationship
between equity capital and risk of the bank proxied by RWA at
1%significance level. Banks investing higher of its own capital as fund are
likely to take more risk because it gives confidence to take more risky
initiatives. Our findings is consistent with theory established earlier. We have
found the negative relationship between ROA and RWA which establishes the
fact that profitable banks tend to take less risk to keep up their profitability.
Liquidity and Risk Taking Behavior of Commercial Banks in Bangladesh 223
6.1.2 Loan loss provision
The result of F test is almost 50% which is far away from accepted range 5%
questioning the validity of the model.
6.1.3 – Z score
Z score usually represents stability but in this model we have used –Z score
which is used as indicator of risk. Higher the – Z score reflects higher the risk
and vice versa. The result of F test is 0.0026 which is less than 0.05, reflects
that the model has reliability. In other way we can say that we can reject our
null hypothesis. The value of adjusted R square is 37%which reflects that37%
of variation in bank risk proxied by –Z score can be explained by the
independent (liquidity risk) and other bank specific control variables. P value
is significant for deposit, asset and equity at 5% significance level. So deposit,
asset, equity have a significant relationship with – Z scores. At 95%
confidence level we have found negative relationship between deposit and – Z
score. Banks, with low liquidity risk, tend to reduce overall risk.
Significant positive relationship between deposit and – Zscore reflects that
banks with available deposit fund tend to reduce the degree of overall risk by
managing each risk component prudently. Lower liquidity risk does not
encourage banks to add more credit risk, market risk, foreign exchange risk,
interest rate risk and other risk in their risky portfolio which can potentially
raise the overall risk of bank and in turn cause decline in reliability of banks
to the bank customers. The negative relationship between equity and – Z score
establishes the fact that the owners tend to protect their equity capital by
minimizing overall risk exposure of banks.
6.1.4 Liquidity Creation
Liquidity creation decreases the liquidity risk but with the high liquidity
creation, banks face more risk for aggressive loan disbursement. The negative
relationship between deposits to assets and risk measure – liquidity creation
reflects that availability of liquidity does not encourage banks to increase the
overall risk level of firm. The result of F test is 0.0074 which is less than 0.05,
reflecting the model has good explanatory power. Equity has a negative
relationship with LC. Banks with greater equity investment are likely to take
less risk.
224 Journal of Banking & Financial Services
6.2 Liquidity risk and bank risk considering impact of macroeconomic
factors
6.2.1 Risk-weighted assets
GDP has significant negative relationship with RWA. Higher GDP reflects
economic expansion and higher income. Higher income increases credit
quality of banking industry. Borrowers become more capable to repay loan
and interest amount. So overall risk position decreases when GDP is high.
Our findings are consistent to theory as well.
6.2.2 Loan loss provision
The result of F test is almost 50% which is far away from accepted range 5%
reflecting the model has no explanatory power.
6.2.3 Z score
GDP has negative relation with –Z score. During economic expansion bank
become more stable due to low default risk as higher income and greater
investment opportunities reduce the probability of default of borrowers.
Inflation has significant positive relationship with –Z score at 1 %
significance level. The result reflects that higher inflation increases risk
position of banks as value of –Z score is increased. During high inflation
value of money is decreased. So people’s ability of repay loan of bank also
deteriorated. As a result credit risk of bank boost up expressing high risk for
bank.
6.2.4 Liquidity creation
The greater the economic expansion in terms of GDP, the higher the
probability that overall risk of bank will increase because excess liquidity
supply from liquidity creation in the economy during economic expansion can
disrupt the financial wellbeing of banks.
Liquidity and Risk Taking Behavior of Commercial Banks in Bangladesh 225
Table 6: Panel OLS Regression for Liquidity Risk and Bank Risk
(With Macroeconomic Factors)
Asset risk
Overall risk
1
2
3
4
RWAt
LLPt
-Z-scorest
LCt
Deposits t-1
0.8897***
(0.2038)
-0.0252
(.0366)
-9.7174**
(4.3116)
-0.4954*
(0.2891)
Assets t-1
0.1571***
(0.0407)
-0.0187*
(.0103)
1.1181**
(0.4994)
-0.0400
(0.0539)
Loan t-1
0.8850***
(0.1955)
0.0192
(0.0380)
-2.8148
(2.0883)
-0.0129
(0.2498)
Equity t-1
3.0221**
(0.6887)
0.1546
(0.1181)
-15.9385*
(8.6956)
-2.0310***
(0.7245)
ROA t-1
-3.7040*
(1.9488)
-0.1026
(0.3454)
24.4468
(20.3146)
-0.08651
(2.4190)
IB Spread t-1
-0.1634
(0.8438)
-0.0494
(0.0693)
-3.4996
(6.3267)
-1.455
(1.0060)
Inflation t-1
0.5605
(1.3346)
-0.1237
(0.1560)
18.9647***
(5.2170)
-1.4195
(1.2356)
GDP t-1
-5.1419**
(2.4519)
0.7196
(0.4479)
-38.0108*
(22.0944)
3.7089*
(2.0892)
Constant
-4.4873**
(1.0637)
0.4589*
(0.2402)
-20.46
(13.87)
1.6478
(1.389)
Time fixed effect
No
No
No
No
Firm fixed effect
No
No
No
No
Observations
88
88
88
88
Adjusted R-sq
0.52085
0.07575
0.3699
0.0904
Test of
Probability
F(8, 21) =
27.62
(0.0000)
F(8, 21) =
0.96
(0.4894)
F(8, 21) =
4.52
(0.0026)
F(8, 21)=3.74
(0.0074)
Note: This table represents the panel regression results where the dependent variables are the
measures of banks’ asset risk as proxied by the ratios of risk-weighted assets to total assets
(RWA) and loan loss provisions to total loans (LLP) and banks’ overall risk as proxied by the
natural logarithm of the Z-scores (Z-scores) and liquidity creation The independent variable of
interest is banks’ liquidity risk as proxied by the ratio of total deposits to total assets (Deposit).
Control variables used are natural logarithm of total assets (Asset), return on assets (ROA) and
the ratios of total loans to total assets (Loan), total equity to total assets (Equity), interbank
spreads (IB Spread), growth rate of real GDP (GDP), and inflation. The sample is based on the
226 Journal of Banking & Financial Services
annual data of conventional commercial banks of Bangladesh. Time fixed effects are
considered and bank fixed effects are not considered in the regressions. P -values are computed
using heteroscedasticity-robust standard errors clustered for banks and are presented in
parentheses. *, **, and *** represent statistical significance at the 10%, 5%, and 1% level,
respectively.
Source: Author’s estimates
7.0 Conclusions
This study examines the impact of liquidity risk on risk-taking tendency of
commercial banks. We have found evidence that more availability of deposit
funding reduces liquidity short fall risk but motivates bank managers to take
aggressive lending decision which raises the asset risk reflected by risk-
weighted asset. Though, higher deposits insures banks from liquidity risk, it
can fuel the risk-taking behavior of lending banks. Our results reflect that
increases in bank deposits accelerates the bank risk measured by risk-
weighted asset, which is consistent with the findings of Acharya and Naqvi
(2012) that banks disburse loan aggressively at lower lending rate when they
have availability of deposits. However, we could not affirm the positive
relationship between deposits ratios and bank’s overall risk captured by –Z
score and liquidity creation. It should be noted that banks have to earn
reliability to generate deposit fund from depositors. If banks’ risk increases, it
reduces the credibility of banks to the depositors. To keep up the reliance of
depositors, banks tend to manage the risk in a way which ensures that
excessive risk will not be added to the risk portfolio. Incorporating
macroeconomic variables with bank specific variables in extended model, we
have found that GDP and inflation has a negative and positive relationship
with risk-weighted asset respectively. During economic crisis bank incurs
high risk, so bank should avoid short term funding to improve liquidity.
The findings of our study reestablishes the fact that there is a considerable
linkage between deposits funding and risk-taking behavior which will help the
regulators to design the prudential credit policies for financial market.
Incentives of bank managers to take more risk, when bank deposits is more
available, can be better monitored and supervised. Examining the effect of
liquidity risk on more advanced risk measure of banks can be an important
direction for future research on this research idea. Advanced risk measure
should be a comprehensive measure of bank risk which will incorporate
technology risk, foreign exchange risk, interest rate risk, and credit risk etc.
more precisely.
Liquidity and Risk Taking Behavior of Commercial Banks in Bangladesh 227
References
1. Acharya, V. and Naqvi, H., (2012). The seeds of a crisis: A theory of bank
liquidity and risk taking over the business cycle. Journal of Financial
Economics, 106(2), pp. 349-366.
2. Adrian, T. and Shin, H.S., (2010). Liquidity and leverage. Journal of financial
intermediation, 19(3), pp. 418-437.
3. Berger, A.N. and DeYoung, R., (1997). Problem loans and cost efficiency in
commercial banks. Journal of Banking & Finance, 21(6), pp. 849-870.
4. Bertay, A.C., Demirgüç-Kunt, A. and Huizinga, H., (2013). Do we need big
banks? Evidence on performance, strategy and market discipline. Journal of
Financial Intermediation, 22(4), pp. 532-558.
5. Calem, P. and Rob, R., (1999). The impact of capital-based regulation on bank
risk-taking. Journal of Financial Intermediation, 8(4), pp. 317-352.
6. Carmona, G., (2007). Bank failures caused by Large withdrawals: An
explanation based purely on liquidity. Journal of Mathematical Economics, 43(7-
8), pp. 818-841.
7. Cebenoyan, A. S. and Strahan, P. E., (2004). Risk management, capital structure
and lending at banks. Journal of Banking & Finance, 28(1), pp. 19-43.
8. Cornett, M. M., McNutt, J. J., Strahan, P. E. and Tehranian, H., (2011). Liquidity
risk management and credit supply in the financial crisis. Journal of Financial
Economics, 101(2), pp. 297-312.
9. Delis, M. D., Hasan, I. and Tsionas, E. G., (2014). The risk of financial
intermediaries. Journal of Banking & Finance, 44, pp. 1-12.
10. Demsetz, R. S. and Strahan, P. E., (1997). Diversification, size, and risk at bank
holding companies. Journal of money, credit, and banking, pp. 300-313.
11. Diamond, D. W. and Dybvig, P. H., (1983). Bank runs, deposit insurance, and
liquidity. Journal of political economy, 91(3), pp. 401-419.
12. Drehmann, M. and Nikolaou, K., (2013). Funding liquidity risk: definition and
measurement. Journal of Banking & Finance, 37(7), pp. 2173-2182.
13. Furlong, F. T. and Keeley, M. C., (1989). Capital regulation and bank risk-
taking: A note. Journal of banking & finance, 13(6), pp. 883-891.
14. Hyun, J. S. and Rhee, B. K., (2011). Bank capital regulation and credit
supply. Journal of Banking & Finance, 35(2), pp. 323-330.
15. Imbierowicz, B. and Rauch, C., (2014). The relationship between liquidity risk
and credit risk in banks. Journal of Banking & Finance, 40, pp. 242-256.
16. Ivashina, V. and Scharfstein, D., (2010). Bank lending during the financial crisis
of 2008. Journal of Financial economics, 97(3), pp. 319-338.
17. Jeitschko, T. D. and Jeung, S. D., (2005). Incentives for risk-taking in banking–A
unified approach. Journal of Banking & Finance, 29(3), pp. 759-777.
228 Journal of Banking & Financial Services
18. Jensen, M. C., (1986). Agency costs of free cash flow, corporate finance, and
takeovers. The American economic review, 76(2), pp. 323-329.
19. Khan, M. S., Scheule, H. and Wu, E., (2017). Funding liquidity and bank risk
taking. Journal of Banking & Finance, 82, pp.203-216.
20. King, M. R., (2013). The Basel III net stable funding ratio and bank net interest
margins. Journal of Banking & Finance, 37(11), pp. 4144-4156.
21. Konishi, M. and Yasuda, Y., (2004). Factors affecting bank risk taking: Evidence
from Japan. Journal of Banking & Finance, 28(1), pp. 215-232.
22. Lucchetta, M., (2007). What do data say about monetary policy, bank liquidity
and bank risk taking?. Economic Notes, 36(2), pp. 189-203.
23. Mercieca, S., Schaeck, K. and Wolfe, S., (2007). Small European banks: Benefits
from diversification?. Journal of Banking & Finance, 31(7), pp. 1975-1998.
24. Merton, R. C., (1977). An analytic derivation of the cost of deposit insurance and
loan guarantees an application of modern option pricing theory. Journal of
Banking & Finance, 1(1), pp. 3-11.
25. Repullo, R., (2011). 11 Liquidity, Risk Taking, and the Lender of Last
Resort. Liquidity and Crises, p. 270.
26. Shim, J., (2013). Bank capital buffer and portfolio risk: The influence of business
cycle and revenue diversification. Journal of Banking & Finance, 37(3), pp. 761-
772.
27. Stiroh, K. J., (2004). Diversification in banking: Is noninterest income the
answer?. Journal of money, Credit and Banking, pp. 853-882.
28. Thakor, A.V., (1996). Capital requirements, monetary policy, and aggregate bank
lending: theory and empirical evidence. The Journal of Finance, 51(1), pp. 279-
324.
29. Wagner, W., (2007). The liquidity of bank assets and banking stability. Journal
of Banking & Finance, 31(1), pp. 121-139.
Liquidity and Risk Taking Behavior of Commercial Banks in Bangladesh 229
Appendix A. Names of the variables and construction of variables
Variable
(Expected
Sign)
Concept
Measurement
Data
Source
Dependent variables
RWA
Risk-weighted assets
Risk-weighted asset/total
assets
Financial
Statements
LLP
Loan loss provision
Loan loss provision/ total
assets
Financial
Statements
Z- Scores
Distant to default
Log [{ROA +
(Equity/Assets)}/Standard
Deviation of ROA)
Financial
Statements
LC
Liquidity creation
Liquidity creation/total
assets. Liquidity creation =
0.5 ×illiquid assets + 0.5
×liquid liabilities −0.5
×liquid assets −0.5 ×illiquid
liabilities −0.5 ×equity
Financial
Statements
Independent variables (Liquidity Risk)
Deposits
Risk-taking
propensity increases
when liquidity risk
decreases
Total deposits/ total assets
Financial
Statements
Explanatory control (Bank specific and macroeconomic variables)
Asset
Banks with larger
asset base are likely
to take more risk
Natural logarithm of total
assets
Financial
Statements
Loan
High amount of loan
growth raise default
risk
Total loans/total assets
Financial
Statements
Equity
Banks with large
equity capital are
likely to take less
risk
Total equity/total assets
Financial
Statements
ROA
Profitable banks
should take more
risk
Net income/total assets
Financial
Statements
IB spread
IB spread raises
bank risk through
higher cost of fund
(Weighted Average deposit
rate- Weighted Average Call
Money Market Rates)
Bangladesh
Bank
GDP
Economic expansion
reduces the degree
of risk
Real GDP growth rate
Bangladesh
Bank
Inflation
Price stability
increases risk
through uncertainty
CPI Index
Bangladesh
Bank
230 Journal of Banking & Financial Services
Appendix B. Diagnostic Test Results
i. Multicollinearity Test
Variables
VIF
1/VIF
lag deposits
1.44
0.695940
lag asset
1.12
0.894746
lag loan
1.08
0.926606
lag equity
1.80
0.555159
lag_ROA
1.44
0.695940
year
1.12
0.894746
2015
1.08
0.926606
2016
1.80
0.555159
2017
1.44
0.695940
lag deposits
1.12
0.894746
Mean VIF
1.51
