A PROJECT REPORT ON

The Impact of Liquidity on Bank Profitability:

Post Crisis Evidence from European Banks

Submitted for partial fulfilment of the Requirements for the

Award of the Degree of Under Graduated of B.com

By- Shruti Mishra

GGV/21/04206

UNDER THE SUPERVISION OF DEPARTMENT OF

COMMERCE
SCHOOL OF MANAGEMENT& COMMERCE
GURU GHASIDAS VISHWAVIDYALAYA
BILASPUR (C.G)
MAY 2024 (Term -: 2021-24)

CANDIDATE'S DECLARATION
 I, Shruti mishra, a student of B.com (Semester IV) in the
Department of Commerce, Guru Ghasidas Vishwavidyalaya
hereby declare that the project report entitled Indian Stock is
“The Impact of Liquidity on Bank Profitability: Post Crisis
Evidence from European Banks” submitted for partial
fulfillment of the requirements for the award of the degree of
B.com of Commerce comprise only my original work and due
acknowledgment has been made in the text to all other
materials used. The results embodied in this project report
have not been submitted to any other university or institute
for the award of any other degree.

Date: July 02, 2021 Shruti Mishra

 Introduction
 Banks have a very important role in the modern globalized financial world. The
effective act of banks depends on a great extent on the financial environment in
which they operate and on their reliability.
 The simplest indication for the financial health of a bank is the value of its assets,
compared to the value of its liabilities but during the recent financial crisis another
type of buffer was underlined, which is the liquidity that the banks have in order to
cover any unexpected outflows.
 The financial crisis that started in 2007 resulted in many problems and failures in
the financial sector but also in many other business sectors.
 This is an example of the importance of the modern banking and financial system
and the effect they have on the real economy. In particular, the strength and the
financial health of the banking industry is an important factor that can determine
the economic stability and growth.
 Moreover, the significance of sound liquidity management was underlined during
the financial turmoil of 2007 when the credit crisis due to the subprime mortgage
lending led to a liquidity crisis, caused by the fall of housing prices and
delinquencies in mortgage lending. Furthermore, the crisis deteriorated the
performance of international stock markets and caused a drying liquidity in
interbank markets since banks and other financial institutions were reluctant to
make any transactions within these markets. Soon this adverse situation led to
several failures in the US where the crisis started but also caused problems in the
global financial system.
 Overall, the reasoning behind the importance of liquidity is that a bank can have
assets that exceed its liabilities, but it still faces the risk of a sudden bank run or any
other situation where it will not be able to liquidate its assets in order to cover
needs that may arise.
 This type of liquidity risk was examined by the Basel Committee which published
Liquidity Risk Management and Supervisory Challenges in 2008, where it supported
that many banks did neither have an adequate framework in order to assess
liquidity risk, nor a sound liquidity management process. Furthermore, the bad
condition of the overall financial system after 2007, revealed the need for the
incorporation of prudential liquidity measures in the banking regulatory framework,
since the importance of liquidity was not fully appreciated.
 The third Basel accord reviewed the banking risk management practices in order to
strengthen and make the financial and banking system more resilient to shocks. The
latest Basel accord incorporated and examined carefully the liquidity risk that
financial institutions face and introduced some new ratios:
 The “Liquidity Coverage Ratio” (LCR), the “Net Stable Funding Ratio” (NSFR) and the
“Tier 1 Leverage Ratio”. These regulatory standards ensure that banks will have
adequate liquidity for the next 30 days, that bank have a stable funding for their
long term assets and finally provide guidance on how much a bank can leverage its
capital base.
 On the other hand, even though the lack of liquidity was proved to be a significant
factor for bank failures, holding excess liquid assets can also have a negative impact
through the opportunity cost of higher returns. In previous literature there has
been found both a positive and a negative relationship between bank liquidity and
profitability. Thus regulators and bank managers have to take into consideration the
trade off between adequate liquidity and resilience to external shocks and the cost
of holding low return assets which limit and reduce banks’ profitability.
 That said, the aim of this dissertation is to examine the relationship between
liquidity measures that banks hold and other regulatory imposed capital with the
banks’ profitability.
 This liquidity measures are both from the asset and liability side of the balance
sheet. Furthermore, the period under examination is after the 2007 financial crisis,
thus the results are interesting in terms of evaluating the impact of liquidity levels
on the banks’ performance, after a crisis in which low liquidity had a significant role
in the deterioration of the whole financial system.
 The second chapter presents a review of previous studies that examined the impact
of liquidity on the profitability of banks and then, in chapter 3 there is a description
of the variables that are used in the empirical part, the source of the data, as well as
the research methodology. In chapter 4 there is the presentation of the empirical
results and finally, chapter 5 includes the conclusions of this study and suggestions
for further research.
 Chapter 1
Liquidity Risk, Regulation and recent developments in the
European Banking System
In this chapter, there is a presentation of the liquidity risk phenomenon in the
banking sector, as well as a theoretical approach to the interaction between liquidity risk
and the financial performance of banks. In addition, there is also a brief presentation of the
recent developments and changes in the European Banking system, since this study
examined the impact of liquidity on the profitability of European banks.

Liquidity Risk

 Liquidity in the banking system is defined by the Basel Committee on Banking

Supervision (BCBS) as the ability of a bank to have available cash or to readily find
cash in order meet its obligations when they come due, without incurring any
unexpected losses.
 The banks’ assets and the related liquidity obligations are very important because
they can determine the weaknesses and strengths that are related to the ability of
the institution to settle its obligations at a timely manner. In addition, a solvent
bank can settle its liabilities when they come due by selling off its assets.
 If a bank has high liquidity then these sells will not incur any unexpected losses but
on the other hand, if the bank has liquidity problems then the sale of these assets
could lead to insolvency.
 Cash holdings in currency or on accounts at any central bank can be sources of
liquidity and in addition another form of liquidity are highly creditworthy securities
like government bills and other securities with short-term maturities. Moreover,
short-term securities are relatively safer than other and can be traded in liquid
markets which mean that these securities can be sold at large volumes without
incurring losses due to price changes.
 Liquidity risk is the risk of being unable to liquidate a position at a timely manner
and a reasonable price and they divide liquidity risk into execution cost (cost of
immediacy) and opportunity cost (the cost of waiting).
 Moreover, it includes both the inability of funding a, the key elements in any
bank’s liquidity position are the maturity transformation which refers to the
 relative maturities of a bank’s assets and liabilities and the inherent liquidity of a
bank’s assets which refers to the ability of any asset to be sold without any
significant loss and under any market condition. In practice these two elements
that are mentioned above are intertwined.
 Apart from the aforementioned maturity mismatch that can result into less
liquidity, liquidity is also affected by the general economic conditions which can
cause less resource generation. For example, during recessionary economic
conditions the depositors’ demand increase creating liquidity risk. In particular, the
phenomenon of a bank run can cause the failure of a specific bank that is in trouble
or even the entire banking system due to contagion effects. Moreover, the problem
arises from the fact that banks finance illiquid assets with demandable claims,
which can cause a drying liquidity in cases of increased depositors’ demand
 The liquidity risk does not only threat the solvency and the financial performance
of banks but it also affects its reputation. A bank that is in financial trouble due to
liquidity risk may lose the confidence of its depositors regarding its ability to
provide investors’ funds at a timely manner. Thus the liquidity risk management is
among the most important activities conducted at banks because it ensures that a
bank will have the necessary liquidity reserves in order to meet any unexpected
need and thus prevent panic dispersion among depositors and investors.
 Recent financial and technological innovations have provided banks with new
opportunities of accessing funding resources, but the recent crisis indicated that
there are still many risks and challenges for liquidity management (Driga and Socol,
2009). In particular, the lack of funds that occurred due to the non-performing
credits, affected the banks’ ability to meet the increased obligations towards
depositors. Thus, despite the technological innovations, liquidity risk is still present
and depends on several factors.
 Another important factor regarding the liquidity position of a bank is its size. In
particular, as supported by the author, the size of the bank can affect the attitude
towards wholesale funding, including the access to the markets and the cost of the
funds that are obtained. Furthermore, the importance of the bank’s size is derived
by the economies of scope and scale that can be achieved.
 For example, a larger bank may have better access to financial markets and
interbank markets because of its because banks that accept on demand deposits
and give loan commitments, might need to hold higher liquidity buffers in order to
accommodate any unpredictable needs .
 Regarding the measurement of the liquidity risk and its effect, in the past,
researchers have been focusing on liquidity risk that arises from the liability side of
a bank’s balance sheet and less attention was given on the asset side.

 Besides that, liquidity risk can also arise from the breakdown or delays in cash
flows by the borrowers. Furthermore, in the past, many researchers focused on
liquidity ratios in order to measure the risk.

 However, according to Poorman and Blake (2005) it is not enough to measure

liquidity only by using ratios and banks need to develop new techniques for
liquidity measurement.

 Therefore, the purpose of this study is to examine the effect of different liquidity
measures that are derived both from the asset side and the liability side of the
balance sheet and also to employ liquidity ratios, as well as a regulatory imposed
liquidity buffer.
 -Liquidity and Financial Performance of
Banks
 The financial performance and profitability of banks is a function of internal and
external factors. Some of the internal factors that can determine the profitability of
a bank are related with the management decisions, regulatory objectives like the
levels of liquidity, expense management and the size of the bank.
 It has also been proved that liquidity problems affect the banks’ earnings and
capital and in some extreme circumstances, such conditions may even result into
bankruptcies of otherwise solvent banks.
 One of the causes of such bank failures, is the fact that liquidity problems may
force banks to borrow from the markets even at an extreme high rate in order to
settle their obligations during a liquidity crisis which ultimately can cause a
decrease in their earnings. Another element that affects a bank’s earnings and is
related with the liquidity is the fire sale risk of assets which can result in an
impairment of the bank’s asset value.
 The fire sale risk arises when a financial institution has to sell a large number of its
assets in order to meet its obligations or to reduce its leverage in conformity with
the regulatory capital adequacy requirements.
 The maturity mismatch between demand in deposits and the corresponding
resources may force a bank to generate these funds from alternative sources and
at a higher cost, thus implying a negative impact on the bank’s financial
performance.
 Regulators reviewed the liquidity risk management in the banking sector and
proposed a stricter regulatory framework which required from banks to hold more
liquid assets in order to become more resilient against potential liquidity or funding
difficulties.
 On the other hand, assets that are highly liquid like government securities and
cash, usually have a relatively lower return and banks that hold such securities, face
an opportunity cost and constraints in their profitability potentials.
 Apart from the negative impact of these constraints that liquid assets impose on
the banks’ profitability, there is evidence that these holdings are beneficial during
adverse economic conditions.
 Thus, there is a dilemma that is faced by a bank’s management since the ultimate
objective is the maximization of profits, but preserving the liquidity of the bank is
equally important.
 On the other side, as the liquidity of a bank increases the opportunity cost of
forgone income due to the lower return assets, comes to predominate and
eventually high liquidity lowers profitability.
 There is a non-linear relationship between the liquidity levels of a bank and its
profitability (Growe et al., 2014). Overall the management of commercial banks is
responsible for estimating and controlling the liquidity levels and maintain an
appropriate level of liquid assets that will not reduce dramatically the profitability
but at the same time ensure that the bank has enough liquidity to overcome any
unexpected need.
 The results of previous studies regarding the liquidity risk are limited and most
studies usually investigate the determinants of banks’ profitability. In addition, the
results from these papers, regarding liquidity are mixed. In the literature review of
this study there is a more detailed presentation of previous studies and their
findings regarding the relationship between the liquidity risk and bank’s
performance.
The effect of Regulation and Supervision on
the Banking Industry
In this study, except from the liquidity measures that are derived both from the
assets the liabilities side of the balance sheet, there is also the examination of the
regulatory imposed capital (Tier 1 capital) on the banks’ performance.
Regulatory authorities have tried to introduce a new regulatory framework with
stricter rules regarding the liquidity risk, in order to make banks and other financial
institutions more resilient in such shocks. Thus, this chapter presents the impact of
regulation, on liquidity requirements and in turn, on profitability. The capital
adequacy of banks is one of the most important elements of this new framework
and regulators have paid much attention on it.
The Capital Adequacy Ratio is the ratio of a bank’s core capital to the assets and
off-balance sheet liabilities weighted by the risk . The minimum level of core capital
relative to the risk weighted assets that banks should hold is 8% and after the
recent crisis, banks are obliged to take into consideration the liquidity risk
requirements when they calculate the Capital Adequacy Ratio.
Regarding the impact of the Capital Adequacy Ratio on the performance of the
bank, found a clear illustration of the regulatory capital on the cost of
intermediation and the banks’ profits.
They supported that the ratio, internalizes the risk for shareholders and increases
the return on assets and equity. Apart from the return on assets and equity there is
no direct empirical evidence of capital adequacy on the net interest margin.
This conclusion is based on the fact that capital requirements can have a positive
impact on bank efficiency through the reduced likelihood of financial distress,
market power and reduced agency problems.
Moreover The reason for the reduced profitability is that the cost of an increase in
the capital requirements are fully transferred to the depositors. Earlier empirical
studies that examined the relationship between bank performance and official
supervision provide mixed results.
Moreover, banks’ risk-taking behaviour is also affected by the regulatory
framework via the moral hazard phenomenon. In particular, regulatory authorities,
except from imposing and examining regulations they also try to avoid the systemic
risk arising from any individual bank.
There are also some regulatory imposed rules that may create incentives for banks
to hold less liquid assets.
For example, the leverage ratio which is considered one of the most important
capital constraints provides an incentive for banks to move to riskier assets with
higher returns and low liquidity.
These incentives are derived from the fact that the leverage ratio is not a risk-
sensitive capital requirement. On the other hand, capital requirements that are
more risk-sensitive like those that are based on the risk weighted assets or capital
stress tests; provide incentives towards more liquid assets and higher overall
liquidity.
In addition, regarding the interaction between liquidity and capital requirements
supported that all else equal, when a bank holds high capital levels there is likely to
be less need for simultaneously high liquidity levels.
The higher capital levels are contributing towards the decrease of the likelihood
that there will be a loss of confidence by funders and at the same time ensures and
increases the likelihood and the ability of the central bank to perform the lender of
last resort functions in case of trouble, as it will be clear that the bank is solvent.
Of course this relationship is not guaranteed and banks need to maintain the
appropriate level of liquidity in order to meet unexpected obligations. On the other
hand, in the case of low levels of liquidity there is a greater need for banks to
maintain higher capital levels and thus be protected from a confidence shock by its
funders. Overall, despite the levels of liquidity that a bank holds, there will always
be the need for banks to maintain a certain amount of capital in order to be
protected under adverse conditions.
European Banking System and recent
developments
In this section of the study, there is a brief presentation of the European banking
system, since the region that is under examination for the relationship between
liquidity risk and bank performance is the European Union (28). Policymakers have
been sought the financial and political integration of the European Union through a
bank regulation framework, as a complement of the European internal market.
Eventually, the European Banking Union was formed in 2012 as a response of the
Eurozone crisis and the fragility of numerous of banks during that period.

After the creation of the banking union, the EU countries transferred the
responsibility for banking policy and supervision from a national level to the
European Union. From 2014, the European Banking Union consists of two main
mechanisms. The first is the Single Supervisory Mechanism (SSM) which is one of
the safety pillars of the banking union, and its aim is to grant the European Central
Bank (ECB) a supervisory role, to monitor the financial stability of systemically
important banks. The second mechanism is the Single Resolution Mechanism (SRM)
which is based on the regulatory framework and its objective is the establishment
of a Single Resolution Fund to finance the restructuring of bank that have
bankrupted.

Furthermore, the whole European Banking System is under the authority and
supervision of the European Central Bank, which consists of 19 member states and
administers the monetary policy of the Eurozone. At last, another important
institution is the European Banking Authority (EBA), which main task is to conduct
stress tests to increase transparency and identify any weaknesses in banks’ capital
structures.

Regarding some characteristics and recent developments of the European Banking

System, according to the 2015 report on the financial structures by the ECB, the
number of credit institutions declined on a consolidated basis from 6,054 at the
end of 2013 to 5,614 at the end of 2014. In 2008 when the crisis started there were
6,774 credit institutions. There was a mass decline in the number of solvent credit
institutions due to many bankruptcies and failures of European banks.
Furthermore, according to the ECB (2015) the market concentration as measured
by the Herfindahl3 index and the share of total assets, remained on an upward
path, since the pre-crisis period. This increased concentration in the industry is a
primarily result of the decline of the number of credit institutions and moreover
the main causes of this increased concentrations are developments in Germany,
Italy and Spain.

Regarding structural developments in the banking industry, in the previous years

there was a gradual shift towards deposit funding which eventually came to a halt,
but banks reduced the use of wholesale funding and their reliance on the central
bank funding. These developments, indicated a trend towards a more traditional
banking business model for banks in the euro area.

Despite the efforts towards a more stable financial system, the recent crisis still
affects the financial health of banks. In particular, since 2008 there has been
deterioration in the loan quality which has resulted in a steady increase of Non
Performing Loans (NPLs) .
This situation is faced by many banks and causes liquidity problems which then
affect their overall performance and the economy as a whole, since banks cannot
proceed with mores credits, because they have tied up capital in order to resolve
the problem with the NPLs.
More recently, the new regulatory requirements that were imposed after the crisis
had a profound impact on the banks’ activities and business model (ECB, 2016).
Banks were forced to quit some profitable but riskier business strategies in
conformity with the new regulatory framework.
This fact in combination with the Herfindahl-Hirschman Index (HHI) is defined as
the sum of the squares of the market shares of all firms within the industry, where
the market shares are expressed as fractions. When the index is below 1000 it
indicates low concertation. On the other hand, an index above 1800 indicates high
concentration and an index between 1000 and 1800 is considered to be as a
moderate concentration in the industry.
Weak macroeconomic and financial condition deteriorated their financial
performance during the latest years and thus the stability of the banking sector
depends on their ability to adapt their business models to the new operating
environment (ECB, 2016).

Chapter 2: Literature Review

 In this section there is a review of the relevant literature related with liquidity risk
and banks’ performance. There is a very limited number of previous studies that
specifically investigated the particular relationship between liquidity risk and bank
performance. Most of the studies that are relevant to this study’s topic, were
mainly focused on examining the determinants of banks’ profitability, and liquidity
was usually one of the examined determinants

Literature Review

In the literature, banks’ profitability was usually measured by Return on Assets

(ROA) or Return on Equity (ROE) and in most of the studies it was expressed as a
function of internal and external determinants. The internal factors were,
profitability determinants, level of liquidity, capital adequacy, expense
management, bank size and others, and they were mainly factors that are
influenced by the banks’ management decisions and policy objectives. On the other
hand, the external factors were both industry related and macroeconomic
determinants and overall they were variables that reflected the general economic
and legal environment of the region

under examination. Liquidity risk can be generally calculated using balance sheet
positions. In the past, researchers focused on the use of liquidity ratios in order to
measure liquidity risk. However, Poorman and Blake (2005), indicated that
researchers should not rely only on liquidity ratios when they try to measure the
liquidity in banks.

According to the Committee of European Banking Supervisors (2008), during the

past years banks did not pay the required attention through their risk management
process on the liquidity risk. In addition, as it was mentioned above there were not
many studies that examined the direct impact of liquidity on banks’ performance.
The first studies on liquidity risk were mainly focused on bank runs (Diamond and
Dybvig, 1983). Since then many researchers and practitioners were interested in
the relationship between bank performance and liquidity risk. Bourke (1989)
examined the internal and external determinants of bank profitability in Europe,
North America and Australia and included a liquidity measure in his analysis.
The liquidity ratio that he employed was the ratio of liquid assets to total assets
and he supported that there was a positive relationship between the ratio and
banks’ profitability.
In Europe, Molyneux and Thorton (1992), used a sample of eighteen European
countries from 1986 to 1989 and examined the determinants of their bank
performance. Their results demonstrated that the ratio of liquid assets to total
assets is negatively related to return on assets (ROA).
The same negative relationship was also supported by Guru et al. (1999), who
investigated the determinants of commercial banks profitability in Malaysia. They
supported that liquid assets are often associated with lower returns and thus high
levels of liquid assets would be expected to be associated with lower profitability.
Their results verified their expectations and in addition they supported that the
difference between Bourke’s (1989) results could be due to different elasticities of
demand for loans in the two samples. In addition, Barth et al., (2003) examined the
impact of the structure, the scope and the independence of bank supervision on
the bank profitability.
They employed a sample of 2300 banks from 55 countries. In their study, liquidity
risk was measured again by the ratio of liquid assets to total assets. They found
that this liquidity ratio had a negative and highly significant relationship with the
profitability,
Indicating a negative relationship between liquidity risk and profitability as
measured by ROA. Another study that employed the ratio of liquid assets to total
assets is the study of Demirguc-Kunt et al., (2003).
They investigated the impact of bank regulation, concentration and institutions on
bank profitability (Net Interest Margin) using a sample of 1400 banks from 72
countries. Their results demonstrated that high liquid asset holdings are related
with lower net interest margins. They also supported that high liquid securities may
receive lower interest income and thus there is a negative relationship with the
bank profitability.
Moreover, some researchers used the ratio of loans to total assets which is similar
to the aforementioned studied the determinants of banks’ Net Interest Margins
and profitability in 80 OECD countries. Their results showed that liquidity which
was measured with the ratio of loans divided by total assets, is negatively related
to profitability as measured by ROA and positively related to Net Interest Margin.
In addition, Athanasoglou et al. (2006), examined the profitability behavior of bank-
specific, macroeconomic determinants and industry related factors, using an
unbalanced panel dataset. The banks that they investigated were from the South
Eastern Europe and the period under examinations was from 1998-2002.
The ratio of loans to total assets which was used as a proxy of liquidity, was found
to have a positive but insignificant relationship with profitability (ROA & ROE). This
was an unexpected result for the authors and the explanation given was related
with the lack of resources of the SEE banking system to meet the liquidity
standards of more developed banking systems, and thus these banks prevent
failures by maintaining an illiquid position.
Except from the ratios of liquid assets to total assets and loans to total assets, other
studies used different liquidity measures. Furthermore, despite the dominance of
the negative relationship between liquidity ratios and profitability, there are also
studies that found a positive relationship.
Kosmidou et al. (2005) examined the impact of bank’s characteristics, financial
market structure and macroeconomic conditions, on the bank’s profitability of UK
domestic commercial banks, during the period 1995-2002. Their results were
mixed, supporting a positive relationship between the ratio of liquid assets to
customers and short-term funding and ROAA and a negative relationship with Net
Interest Margin.
This positive relationship is consistent with Bourke (1989) and Kosmidou (2008) in
which study there was also a positive relationship. In particular, the researcher’s
objective was to examine the determinants of performance in the Greek banking
system during the EU financial integration period (1990-2002).
An unbalanced pooled time series sample of 23 banks was employed and the
results were mixed. The ratio of loans to customers and short term funding was
negatively and significant related with ROAA but when financial structure and
macroeconomic variables were employed in the equation the relationship became
positive but insignificant. The negative and significant relationship of the ratio with
ROAA, implied that less liquid banks have lower ROAA which was inconsistent with
the author’s expectations, but consistent with Bourke’s (1989) research.
Other studies that measured liquidity using a ratio of assets to customer and short
term funding, are, Pasiouras and Kosmidou (2007), who used the ratio of loans to
customer and short term funding. In their research, they studied the effect of
bank’s specific characteristics and the overall banking environment on the
profitability (ROAA) of commercial domestic and foreign banks operating in the 15
EU countries, for the period 1995-2001.
The results showed that liquidity was statistically significant and positively related
to the profitability of domestic banks which indicates a negative relationship
between the level of liquid assets and bank profitability. On the other hand, in the
case of foreign banks, liquidity is also significant but it is negatively related which
indicates a positive relationship between liquid assets and profitability.
In addition to the previous study, Naceur and Kandil (2009), investigated the
effects of capital regulation on the cost of intermediation and profitability of banks
in Egypt. Their sample contained 28 banks observed over the period 1989 to 2004.
They found that liquidity had a positive and significant effect on the cost of
intermediation, indicating that the increased liquidity imposed by regulations,
induces higher cost of intermediation to increase earnings. They also concluded
that banks’ liquidity does not determine return on assets or equity (ROA and ROE)
significantly.
Some other studies that used alternative measures for liquidity are the following.
Shen et al. (2009), studied the determinants of bank performance in terms of
liquidity risk management measured by the financing gap ratio and the ratio of net
loans to customers and short term funding.
The unbalanced panel dataset of banks from 12 advanced economies for the
period 1994-2006 provided results showing that the liquidity risk is negatively
associated with return on assets average (ROAA) and return on average equity
(ROAE). This indicated that banks with larger financing gap, lack stable and cheap
funding and they depend on liquid assets and external funding to meet their
obligations.
On the other hand, there was a positive relationship between liquidity risk and Net
Interest Margin, which in contrary indicated that banks with high levels of illiquid
assets, may receive higher income through interest than more liquid banks.
Moreover, Ariffin (2012), analyzed the relationship between liquidity risk and
financial performance of Islamic banks in Malaysia.
The period under examination was during the crisis and in particular the period
2006-2008. The author measured the liquidity risk with the ratio of total assets
over liabilities and found that in time of crisis, liquidity risk, return on assets (ROA)
and return on equity (ROE) tend to behave in an opposite way and in particular,
liquidity risk may lower the banks’ profitability.
An alternative study of David and Samuel (2012), examined the effect of liquidity
management on the profitability of commercial banks. Their research methodology
was based on structured and unstructured questionnaire on the management of
banks. Moreover, they formulated a hypothesis which was then statistically tested
through Pearson correlation data analysis.
Their results, which were derived direct from the banks’ management, indicated
that profitability in banks is significantly influenced by liquidity and vice versa.
In this study apart from the impact of liquidity of banks on their financial
performance, there is also the examination of the impact of regulatory imposed
capital (Tier 1 Core Capital) on the banks’ performance.
All the aforementioned studies regarding liquidity risk, were mainly focused on
finding the determinants of banks’ profitability and they used different proxies of
liquidity in order to examine its effect.
Apart from these studies, there are some research papers that directly examined
the impact of liquidity on the banks’ performance. In particular, a recent study of
Bordeleau and Graham (2010), in a working paper for the Bank of Canada, analyzed
the impact of liquid assets holdings on bank profitability.
Their sample consisted of US and Canadian banks and the period under
examination was 1997-2009. This study used ROA and ROE as dependent variable
of profitability, which was then regressed against a non-linear expressions of liquid
asset holdings and a set of bank specific and macroeconomic control variables.
Results suggested that banks that hold some liquid assets, have improved
profitability. However, according to the authors, there is a cut off point where
further liquid asset holdings, diminish the profitability.
In addition, results provided evidence that the aforementioned relationship
between liquid assets and profitability, depends on business model of each bank
and risk of funding market difficulties. Moreover, Arif and Anees (2012),
investigated the liquidity risk in the banking system of Pakistan and evaluated its
effect on banks’ profitability.
 In their research, they employed different liquidity measures that were derived
from the banks’ balance sheets, like deposits, liquidity gap and NPLs.
The sample included 22 banks and the period under examination was 2004-2009.
Their findings, demonstrated that the liquidity risk significantly affects the
profitability of banks, with liquidity gap and NPLs being the two factors that
exacerbate the risk. However, according to the authors, this liquidity risk can be
mitigated by raising the deposit base, maintaining sufficient cash reserves and
decreasing the liquidity gap and the NPLs.
Another research of Ferrouhi (2014), evaluated the effect of banks’ liquidity
positions on their profitability in Marocco which was measured by ROA, ROE, ROAA
and NIM. In order to specify the relationship between liquidity risk and profitability,
the author used the aforementioned profitability ratios, six liquidity ratios and
other macroeconomic and bank specific variables for the period 2001-2012.
The results were mixed and the relationship derived between profitability and
liquidity risk was dependent on the model used. Overall, according to the results
the authors could not determine whether a liquid bank is more efficient than an
illiquid bank.
Finally, a recent study that examined the impact of liquidity is the study of Marozva
(2015). This study examined the impact of liquidity on bank performance for South
African banks and for the period 1998-2014. In particular, in this study liquidity was
measured in the context of funding liquidity risk and market liquidity risk.
According to the results, there is a negative significant relationship between net
interest margin and funding liquidity risk. Besides that, there is an insignificant co-
integrating relationship between NIM and the two liquidity measures.
Chapter 3: Variable Description, Data and Empirical
Methodology

In this section there is a presentation of the dependent and independent bank

specific and macroeconomic variables that were used in this study. Regarding the
bank specific dependent variables, they consist of both liquidity indicators from the
banks’ balance sheet as well as liquidity ratios. The sample that was used in the
empirical part, consists of 50 banks. Detailed presentation of the sample is
reported in part 3.3 Data.

3.1 Dependent variables

In this study, the banks’ liquidity will be examined on ROAA, ROAE, NIM and PBT in
order to get an overall and robust indication regarding the relationship between
liquidity and profitability. Next there is a presentation of the three performance
measures, together with three graphs that depict the differences of these
indicators during 2009-2015. The value of each year is derived as an average value
of the 50 banks that were used in the study’s sample.

3.1.1 Return on Average Assets (ROAA)

The main model that is going to examine the impact of liquidity on the financial
performance of banks will have as a dependent variable the ROAA. This ratio is
defined as the net profit after tax divided by the average total assets. It reflects the
ability of any bank’s management to generate profits from the value of assets.
Return on average assets is used instead of return on assets, in order to control for
differences in the value of assets that occur within the fiscal year. Moreover, ROAA
is considered the most important profitability measure, when it comes to compare
the financial performance of banks.

In the literature, many researchers have used ROA and ROAA in their models, like
Molyneux and Thorton (1992), Demirguc-Kunt and Huizinga (1999), Barth et al.
(2003), Kosmidou et al. (2005), Pasiouras and Kosmidou (2007), Kosmidou (2008),
Naceur and Kandil (2009), Shen et al. (2009), Bordeleau and Graham (2010), Ariffin
(2012), Ferrouhi (2014).
Figure 1- shows that the average ROAA of the sample (50 banks) increased from
2009-2010 and then it sharply decreased until 2012 when it started to increase
again.

.4%

.3%

.2%

.1%

.0%
-.1
%
2009 2010 2011 2012 2013 2014 2015

Figure 1. ROAA, Source: Bankscope, Software: Eviews 9

3.1.2 Return on Average Equity (ROAE)

Another measure that is employed to depict the performance of banks, is the

Return on Average Equity which is defined as Net Income over the average
shareholders’ equity. As it was mentioned before, the average value of
shareholders’ equity is used in order to capture any differences in the equity value
during the fiscal years (or season effects). This ratio indicates the profitability of a
financial institution or corporation by demonstrating the percentage of profit that
was generated compared to the invested money that shareholders contributed. In
Figure 2 there is the corresponding graph, regarding ROAE.

10%
5%
0%
-5%
-
10%
-
15%
2009 2010 2011 2012 2013 2014 2015
Figure 2. ROAE, Source: Bankscope, Software: Eviews 9
The ROAE graph shows a similar trend as the ROAA.
Net Interest Margin (NIM)

We also use the Net Interest Margin as a proxy of profitability, which is calculated by the ratio
of net interest income over the total earning assets and thus it shows the profitability of a
bank’s interest earning business. This performance metric, demonstrates how successful a bank
manages its investments decision (mainly regarding its loan portfolio), compared to its debt
obligations. Figure 3 shows how the average NIM of the sample changed over the period 2009-
2015.

1.42
%
1.40
%
1.38
%
1.36
%
1.34
%
1.32
%
1.30
%
2009 2010 2011 2012 2013 2014 2015

Figure 3. NIM, Source: Bankscope, Software: Eviews 9

3.1.4 Profit Before Tax (PBT)

Finally, the last dependent variable is Profit Before Tax. This profitability measure is
taken directly from the banks’ financial statements and reflects the profits before the
banks have to pay any corporate tax. This is because, each bank has different debt
obligations and thus different tax shields. Previous studies mostly used the three
aforementioned profitability ratios and Profit Before Tax was used by Arif and Anees
(2012) to directly examine the impact of liquidity on profitability. Next, Figure 4
demonstrates the changes on the average PBT of the sample. This graph has a similar
trend with ROAA and ROAE during the years, with slight differences.
4,000$

3,000
$

2,000
$

1,000
$
2009 2010 2011 2012 2013 2014 2015

Figure 4. PBT (mil), Source: Bankscope, Software: Eviews 9

3.2 Independent Variables
There are two categories of independent variables that are used in this study. First of
all, there are seven bank specific variables that are related with the banks’ core
activities and management’s decisions and in addition there are two macroeconomic
variables which capture the effect of the overall economy on the banks’ performance.
The macroeconomic variables correspond to the country in which each bank has its
headquarters.

The bank specific variables that are used in this analysis are the following: 1) Cash and
Due from Banks divided by Total Assets (CTTA), representing the liquidity from the asset
side of the balance sheet. The variable of cash was transformed into a ratio due to the
high correlation with the variable of Total Customer Deposits. 2) Total Customer
Deposits (DEP), representing the liquidity that is derived from the liability side of the
balance sheet. 3) The ratio of Impaired Loans to Gross Loans (ILTGL) which shows the
effect of credit risk on the banks’ performance. 4) The ratio of Net Loans to Total Assets
(LTTA), representing liquidity and the annual growth of the loan portfolio. 5) The ratio
of Loans less Customer Deposits to Total Assets (FGAPR) which represents a bank’s
financing gap (standardized by Total Assets). 6) The ratio of Tier1 Capital to Total Assets
(T1TTA), representing the regulatory imposed level of liquidity. 7) The Cost to Income
ratio (COST), which is included as a bank specific control variable.

The profitability of banks in not determined only by internal factors, but also by
external macroeconomic factors that affect the whole economy. Thus the research of
this study incorporates two macroeconomic variables, the annual real Gross Domestic
Product Growth (GDPGR) and the Inflation Rate (INF), measured as the percentage
annual change of consumer prices including all items.

Regarding the first variable, of Cash holdings to Total Assets, it is used as a proxy of
liquidity, since any bank has to keep sufficient funds in order to meet any unexpected
demand, mainly from depositors. Thus increased cash holdings and in turn increased
liquidity indicates a more resilient bank. On the other hand, maintaining high cash
reserves can be expensive (Holmstrom and Tirole, 2000) since banks that maintain high
cash levels, face the opportunity cost of forgone income from other alternative
investments.
 Due to this opportunity cost, the relationship of cash holdings with profitability is
expected to have a negative sign. The other variable that is also derived from the banks’
balance sheet, is the Total Customer Deposits can provide a natural hedge for banks,
against liquidity risk, because deposit inflows provide funding for loan demand shocks
or for cases of mass withdrawals. In addition, banks can increase their profitability by
increasing their loan portfolio and other activities with the use of deposits.
The ratio of Impaired Loans to Gross Loans, demonstrates the annual growth of non-
performing loans (NPLs). The NPLs directly affect the liquidity and the profitability of a
bank, by reducing its cash inflows from its loan portfolio which is usually the biggest
asset of a commercial bank.
Thus as this ratio increases, profitability is expected to decrease. Furthermore, the
ratio of Net Loans to Total Assets, represents liquidity in the form of the percentage of
total assets, that is invested in the loan portfolio.
The management of a bank should have established a maximum goal for this ratio in
order to avoid liquidity problems. An increase of this ratio could result into increased
profitability for banks but on the other hand, according to Staikouras and Wood
Banks which increase this ratio may face higher cost for their funding requirements.
Moreover, the standardized Financing Gap ratio that is employed in this study, is used
to proxy liquidity risk. Banks with a high financing gap, must use their cash, or even sell
some liquid assets in order to fund this gap. It consequently increases the cost of
funding and reduce the banks’ profitability. In previous studies, Shen et al. (2009) found
a negative relationship of this ratio with ROAA and ROAE and a positive relationship
with NIM. Found a negative relationship with ROA, indicating that banks with high
financing gap ratio, lack stable and cheap funding.
In addition to the previous liquidity measures, this study examines also the effect of
regulatory imposed capital on the banks’ profitability.
The ratio of Tier1 Capital over Total Assets is employed which also consists one of the
regulatory standards regarding the liquidity risk (leverage ratio). Thus there is a positive
sign expected in this study. Finally, the last independent bank-specific variable is the
Cost to Income ratio, which is used as a control variable to the banks’ characteristics. It
is defined as operating cost over the total generated revenues and it shows how
efficient a bank manages its expenses. The anticipated effect of Cost to Income ratio as
supported by Kosmidou, is negative since higher expenses mean less profits.
 Regarding the external variables which are related to the performance of the overall
economy, this study employs two measures. First, the real Growth rate of Gross
Domestic Product, which shows the annual percentage change of each country’s GDP,
depicts the state of the economic cycle.
When there is growth in the economy, it is usually supported by a credit expansion
which enhances this growth and also increases banks’ profitability. Overall, a positive
relationship is expected, which is also supported and . On the other hand, the Inflation
Rate is another important macroeconomic variable, affecting the real value of revenues
and costs.
Table 1. Variable Summary. Source: Bank Specific variables were obtained from
Bankscope; Macroeconomic Variables were obtained from the OECD database.

Variable Notation Expected effect

Return on Average ROAA
Assets
Vari Return on Average ROAE
able Equity
sVar Net Interest Margin NIM
iabl
es Profit Before Tax PBT

Cash to Total
Ma Assets CTTA Negative
cro
ec Total Customer DEP Positive / Negative
on Deposits
om Impaired Loans to ILTGL Negative
ic Gross Loans
Ba Net Loans to Total LTTA Negative
nk Assets
Sp Financing Gap Ratio FGAPR Negative
eci
fic Tier1 Capital to T1TTA Positive
In Total Assets
de Cost to Income
pe ratio COST Negative
nd
ent Real Growth Rate of GDPGR Positive
Gross Domestic
Product

Inflation Rate INF Negative/Positive

 3.3 Data
As it was mentioned above, there is a separation in the variables between bank
specific and macroeconomic variables. Bank specific variables were derived from
the Bankscope Database of Bureau and Dijk’s company.
On the other hand, the data for the macroeconomic variables were derived from
the OECD database.
Considering the recent financial crisis, and the important role of the liquidity risk in
the deterioration of the crisis, the period under examination is during the crisis and
in particular the recent years from 2009 to 2015. In addition, the sample consists of
50 large European banks. In order to form the sample, the banks were sorted by
the value of their total assets in the Bank scope database.
The selection criteria of this study are the following:
1) The institutions have to be active commercial banks.
2) In addition, only banks that had continuously data during the period under
examination were selected.
3) These continuously data have to be reported under the Banknote’s C2 consolidation
code, which consists of the financial statement of a mother bank which in turn integrates the
statements of its controlled subsidiaries or branches with an unconsolidated companion.

$850,000,000

$800,000,000

$750,000,000

$700,000,000

$650,000,000

$600,000,000
2009 2010 2011 2012 2013 2014 2015

Figure 5. Average Total Assets, Source: Bankscope, Software: Eviews 9

Descriptive Statistics
The descriptive statistics of the internal and external variables that are used in this study.
The CTTA, the FGAPR, the T1TTA and the LTTA variables are expressed as a percentage of
total asset. Deposits and Profit Before Tax, are expressed in millions of USD dollars. Next
there is a presentation of the most significant inferences from Table 2.

Table 2. Descriptive Statistics

Mean Median Maximum Minimum Std. Dev

ROAA 0.176629 0.27 1.06 -6.83 0.6018

ROAE 1.606086 6.01 27.2 -724.67 40.74147

NIM 1.362029 1.265 2.97 0.27 0.595805

PBT 2,165.975 1,438.354 22,565.0 -33,035.89 5,252.402

CTTA 3.285771 2.27 16.14 0.0042 2.832466

DEP 291,788.40 193,762.70 1,361,297 357.57 265,764

ILTGL 5.419714 3.915 33.84 0.38 4.896033

LTTA 51.1472 52.76 87.7 14.7 16.78778

FGAPR 12.07229 10.79251 81.47418 -68.17835 20.26572

T1TTA 4.363654 4.35923 6.942073 1.776051 1.12419

COST 64.68391 63.2 195.35 24.88 17.23353

GDGGR 0.549212 1.19665 26.2761 -8.269 2.815848

INFL 1.281231 1.1898 4.5 -4.4819 1.221251

Regarding the profitability measures of ROAA and ROAE, we can observe that the
average values for our sample are 0.176% and 1.606% accordingly. These values
are smaller than the median values which are 0.27% and 6.01% accordingly,
indicating that there are significant profitability differences among the banks in the
sample (Dietrich and Wanzenried ,2011). In addition, ROAE has a high standard
deviation indicating high variance among the banks’ data, which is also observed by
the difference between the maximum and minimum value. The same high standard
deviation holds also for Profit Before Tax and Total Customer Deposits, which are
 expressed in millions of USD dollars. In addition, the ratio of impaired loans has a
mean value of 5.41% which is quite low and close to the European average in 2015
(Mesnard et al., 2016), but it is still higher than other developed countries (USA,
Japan). Furthermore, Net Loans consist on average the 51.14% of total assets but
there is a high standard deviation, indicating differences in the banks’ business
models. Next, the Financing Gap ratio has a positive mean, indicating that on
average banks have more loans than deposits, but there exist banks that also have
a negative gap. Finally, the Tier 1 ratio, has a positive mean, higher than 3% which
is the minimum requirement imposed by the Basel Committee (BCBS, 2014).

3.5 Correlation Matrix

In order to test our sample data for high correlation that could affect the results,
Table 3 presents the correlation matrix for the independent bank specific and
macroeconomic variables. Considering the nature of the data, we can observe that
the correlations are quite low. That’s because some variables (e.g. Cash and Due
from Banks) were transformed into ratios. The highest correlation exists between
LTTA and FGAPR (0.61), but it still is lower than the threshold of 0.7. Overall,
according to Table 3. there is nothing important than could affect the results of this
analysis.

Table 3. Correlation Matrix

Correlation CTTA DEP ILTGL FGAPR LTTA T1TTA COST GDGGR IN

CTTA 1

DEP 0.271434 1

ILTGL -0.12921 -0.07469 1

FGAPR -0.19874 -0.38016 0.064403 1

LTTA -0.05995 -0.30927 0.138634 0.616102 1

T1TTA 0.340641 0.047532 0.255217 0.034454 0.349098 1

COST -0.00847 0.064725 -0.01277 -0.26941 -0.26842 -0.30983 1

GDGGR 0.159717 0.005515 -0.07134 -0.11984 -0.08261 0.011208 0.005551 1

INFL -0.00837 0.169426 0.079968 0.014238 0.065822 -0.00545 0.058891 0.078046

 3.6 Research Methodology
In order to examine the impact of the bank specific and
macroeconomic factors on the banks’ profitability, we employ
the following econometric models. These are the extended
equations that reflect all the variables that are used:

1) ROAAit = C + β1 + β2 + β3 + β4 + β5 + β6 1 + β7 + β8 + β9 +
δi + εit
2) ROAEit = C + β1+ β2+ β3+ β4+ β5+ β6 1 + β7
+ β8+ β9+ δi + εit
3) NIMit = C + β1+ β2+ β3+ β4+ β5+ β6 1 + β7
+ β8+ β9+ δi + εit
4) PBTit = C + β1+ β2+ β3+ β4+ β5+ β6 1 + β7
+ β8+ β9+ δi + εit

The dependent variable refers to the profitability (ROAA, ROAE, NIM, PBT) measure
of bank i at time t, δi reflects the fixed effects in the bank i and εit is the error term.
All the calculations and the regressions were performed with the Eviews 9
software.

Each of the three models was initially tested for the cross section effects. In order
to do so, the Hausman test (Baltagi, 2001) was performed in the four models.
According to the results, the null hypothesis which indicates that random effects
are more appropriate, is rejected and thus fixed effects are used instead (See
Appendix 2.).
In addition, in order to test the significance of each model’s effects, we conduct
the Redundant Fixed Effects – Likelihood Ratio. This ratio tests the joint significance
of the cross – section effects that are used. The p-values, obtained by the test
strongly reject the null hypothesis that the cross – section effects are redundant.
Furthermore, in order to get robust results regarding the standard errors and the
significance of each variable, we performed two residual diagnostic tests. First, we
employed the Breusch-Pagan test to check for heteroscedasticity in the residuals. T
Moreover, a manual process was employed in Eviews 9, to check the residuals for
serial correlation. The test was performed with the help of an auxiliary regression
which was estimated using the variables in their first differences. Then the residuals
series of this regression were regressed on the lagged residuals (residuals on
residuals
Under the null hypothesis which suggests that the original idiosyncratic errors are
serial uncorrelated, the autocorrelation coefficient should be -0,5. Using a Wald
Coefficient Restriction test, we can conclude that in the models that have as
dependent the ROAA, ROAE the null hypothesis cannot be rejected, while regarding
PBT there is not enough evidence to reject the null hypothesis, thus there is no
serial correlation among the residuals. In the model that has NIM as dependent the
null hypothesis is rejected indicating the presence of serial correlation.
Finally, according to the Redundant Variable Test, which tests the joint significance
of the variables that are included in the equation, the null hypothesis that assumes
that the bank specific variables are redundant, is rejected in all models.
Regarding the estimation of the models, all models were estimated using fixed
effects. Moreover, to get robust results, we estimated the models using the White
Cross –Section coefficient covariance method to control for the presence of
heteroscedasticity in the residuals. The third model (NIM) was estimated, using the
White Period method, which assumes that the errors are heteroscedastic and
serially correlated.
Chapter 4: Empirical Results
The empirical results regarding the estimation of the four models are presented in Table 4,
Table 5, Table 6 and Table 7 (at the end of this chapter). First of all, the explanatory power,
measured by the R2 is quite high in models 1), 3) and 4) which examine the impact of
liquidity on ROAA, NIM and PBT respectively, but it is not as high when the dependent
variable is ROAE. Moreover, we can observe that the models 1), 2) and 4) (ROAA, ROAE and
PBT) have similar results concerning the sign of each variable while there are differences in
their significance. On the other hand, model 3) which examines NIM has differences in the
variables’ signs and also in their significance.
The variable of Cash to Total Assets is negatively related with the banks’ profitability,
represented by ROAA, ROAE, PBT. This result is opposite to the positive relationship,
supported by the research of Arif and Anees, (2012), but consistent with Holmstrom and
Tirole, (2000) who supported that high cash reserves can be expensive due to the
opportunity cost of forgone income from other alternative investments. Moreover, this
variable is statistically insignificant in the models 2) and 4) and significant in the first model
(ROAA) at 10%. Thus, even though cash reserves increased as a percentage of total assets
during 2009-2015 (See Appendix 3.), this increased liquidity did not enhance the banks’
profitability. On the other hand, the third model (NIM) has a positive sign in this variable
but it is statistically insignificant.

In addition, Deposits and Due from Banks is also negatively related with
profitability in all four estimated models which is consistent with Dietrich and
Wanzenried,

In the models that have ROAA and NIM as dependent, Deposits are statistically
significant at 1%, in the model with PBT as dependent it is significant at 5% while in
the second model (ROAE) it is insignificant. After 2013 there is on average, a steep
reduction in total customer deposits (See Appendix 3.) which contributed to the
decreased profitability by decreasing the resources that were available for
investments.
First model it was significant at 5%. The results show that as the percentage of
impaired loans increased (See Appendix 3.) during 2009-2015, banks suffered by
reduced cash inflows from their loan portfolio which then caused a reduction in
their profits.
Furthermore, the ratio of Net Loans to Total Assets has also a negative impact on
the banks’ profitability, when it is measured by ROAA, ROAE and PBT, but it is
positively related with NIM.
This variable is statistically significant at 1% only in the third model (NIM and
indicates that banks which have a high ratio may face higher cost for their funding
requirements. In addition, the increased percentage of NPLs during this crisis
reduced the returns that banks expected from their loan portfolio.
On the other hand, the positive sign on NIM can be explained by the fact that this
ratio is depended on interest income. Thus, when this ratio increases, NIM also
increases.
Regarding the FGAPR variable, we found a negative relationship with profitability in
all models. The results are statistically significant (at 1%), in model 1) (ROAA) and
(at 5%) in model 4 (PBT). This negative sign indicates that illiquid banks that have a
high financing gap ratio, may face funding difficulties, which in turn reduce their
profitability. These results are consistent with etc regarding ROAA and ROAE, but
inconsistent with their result regarding NIM where they found a positive impact.
The variable that was used as a proxy for the regulatory imposed capital (T1TTA),
has a positive but insignificant impact on profitability when it is measured by ROAA,
ROAE and PBT and a negative and insignificant impact when the depended variable
is NIM.
The positive impact is consistent with Berger (1995), Goddard et al., (2004) and
Bandt et al., (2014) and according to Berger (1995), it shows that well capitalized
banks face lower expected bankruptcy costs and thus they can have access to
cheaper funding. In contrast, the negative impact of this capital ratio is consistent
with the aforementioned characteristic of NIM.
Since it depends on the interest income, higher capital requirements may shrink
the interest earning business of a bank and thus decrease NIM.
The Cost to Income ratio, which was used as a control variable to the bank specific
characteristics, was found to be negatively related with the four banks’ profitability
measures.
The variable was statistically significant at 1% in model 1 (ROAA), 2 (ROAE) and 4
(PBT), while in model 3 it is significant at 10%
Regarding the macroeconomic variables that were used in the regression, the
results are the following. The annual growth rate of GDP was found to be positively
related with the banks’ performance, in all models. Moreover, it was statistically
significant at 1% in the first model (ROAA) and at 5% in the fourth model (PBT),
Finally, the results the second macroeconomic variable that was used (INFL) were mixed.
Inflation, was positively related with ROAA, ROAE and PBT but negatively related with NIM.
Moreover, this relationship was insignificant in all four models.
The positive association is supported by previous studies, It also supports the theory that
inflation was anticipated and thus banks had the opportunity to adjust their interest rate
accordingly and eventually increase their profitability.
Table 4. DEPENDENT VARIABLE: ROAA, MODEL 1

Independent Coefficient Std. Error Prob.

Variables
TESTS Statistic Prob.

C
Breusch-Pagan 1.689296 0.518863 0.0013
CTTA
Test (LM) -0.015369
1659,52 0.008781
0.0000 0.0811
DEP -9.22E-07 3.59E-07 0.0106
Redundant
ILTGL Fixed -0.044975 0.018605 0.0162
Effects
LTTA Test -0.004316 0.009237 0.6407
1) Cross-Section
FGAPR F 1) 1.856954
-0.013948 1) 0.0010
0.004982 0.0055
2) Cross-Section
T1TTA 0.215085 0.140699 0.1274
Chi- 2) 95.225611 2) 0.0001
COST -0.023435 0.001479 0.0000
Square
GDPGR 0.024370 0.005405 0.0000
Redundant
INFL 0.002845 0.015974 0.8588
Variable Test 27.34281 0.0000
R2 0.6048
(F-statistic)
Adjusted R2 0.5261
Table 5. DEPENDENT VARIABLE: ROAE, MODEL 2

Independent Coefficient Std. Error Prob.

Variables

C 62.19509 39.86366 0.0198

CTTA -0.952510 0.621859 0.1267
DEP -5.13E-06 2.51E-05 0.8385
ILTGL -1.869336 0.547125 0.0007
LTTA -0.695584 0.535727 0.1952
FGAPR -0.225884 0.435217 0.6041
T1TTA 21.46071 14.04158 0.1275
COST -1.601372 0.470560 0.0008
GDPGR 0.441291 0.423993 0.2988
INFL 1.690464 1.343184 0.2092

R2 0.4815
Adjusted R2 0.3781

Tests Statistic Prob.

Breusch-pagan
test (LM) 1980.28 0.0000

Redundant Fixed
Effects Test
1) Cross-Section F 1) 2.8676 1) 0.0000
2) Cross-Section
Chi- 2) 137.8908 2) 0.0000
Square

Redundant
Variable Test 22.6413 0.0000
(F-statistic)
Table 6. DEPENDENT VARIABLE: NIM, MODEL 3

Independent Coefficient Std. Error Prob.

Variables

C 0.957930 0.243090 0.0001

CTTA 0.003975 0.005507 0.4710
DEP -7.24E-07 2.05E-07 0.0005
ILTGL -0.018650 0.003955 0.0000
LTTA 0.020474 0.004245 0.0000
FGAPR -0.008376 0.005302 0.1152
T1TTA -0.030887 0.026027 0.2363
COST -0.001608 0.000959 0.0946
GDPGR 0.007099 0.005549 0.2018
INFL -0.006147 0.011080 0.5794

R2 0.9458
2
Adjusted R 0.9350

Tests Statistic Prob.

Breusch-Pagan
Test (LM) 1982.41 0.0000

Redundant Fixed
Effects Test
1) Cross-Section F 1) 42.436 1) 0.0000
2) Cross-Section 2) 734.121 2) 0.0000
Chi-square

Redundant
Variable test 14.665 0.0000
(F-statistic)
Table 7. DEPENDENT VARIABLE: PBT, MODEL 4

Independent Coefficient Std. Error Prob.

Variables

C 18295.99 4039.693 0.0000

CTTA -135.3482 84.34066 0.1096
DEP -0.009830 0.004941 0.0476
ILTGL -326.7948 64.38973 0.0000
LTTA -60.12604 62.91371 0.3400
FGAPR -93.98454 40.11103 0.0198
T1TTA 372.6691 811.1249 0.6463
COST -132.8939 16.62812 0.0000
GDPGR 213.2950 91.25838 0.0201
INFL 13.05280 108.5171 0.9043

R2 0.7053
2
Adjusted R 0.6466

Tests Statistic Prob.

Breusch-Pagan
Test (LM) 1612.86 0.0000

Redundant Fixed
Effects Test
1) Cross-Section F 1) 4.5323 1) 0.0000
2) Cross-Section 2) 198.4918 2) 0.0000
Chi-square

Redundant
Variable test 14.4593 0.0000
(F-statistic)
 Chapter 5: Conclusions
This study investigated the impact of liquidity, which was measured by several balance
sheet measures (e.g. cash, deposits), liquidity ratios (e.g. Loans to Total Assets, Impaired
Loans to Gross Loans), capital ratio, and external macroeconomic factors, on the
profitability of 50 European banks, measured by ROAA, ROAE, NIM and PBT. In addition,
since there is a limited number of previous studies that addressed this particular topic, this
study is intended to contribute to the relevant literature, through an overall and robust
assessment of liquidity on banks’ profitability, after the recent financial crisis, where
liquidity played a very important role.
In order to do so, a balanced panel data set was used, including 50 large European banks
for the period 2009-2015, resulting into 350 observations.
The results of this study, showed that the liquidity measures, had the same impact
regarding the sign of each variable on ROAA, ROAE and PBT, while there were some
differences in the NIM equation. In particular liquidity measures that were used directly
from the banks’ balance sheet, Cash and Due from Banks and Total Customer Deposits,
were found to have a negative relationship with ROAA, ROAE and PBT, providing support
that the opportunity cost of holding low yield assets and on the other hand holding
deposits which can not be invested appropriately or are invested in high risk assets, comes
to dominate the increased resilience of the banks due to increased liquidity. In the
equation were NIM was depended variable, Cash and Due from Banks was found to be
positively but insignificantly related with profitability.

Furthermore, regarding the liquidity ratios that were used as different proxies of
liquidity, they all had the expected outcome on the banks’ profitability. The ratio of
Impaired Loans to Gross Loans was found to have a negative and statistically
significant relationship with all profitability measures, indicating that the problem
of increased NPLs during the financial crisis, affected the liquidity and eventually
the banks’ profitability. Moreover, the ratio of Loans to Total assets, was negatively
related with ROAA, ROAE and PBT, but positively related with NIM.
The variable that was used as a proxy to the regulatory imposed liquidity was the
ratio of Tier 1 Capital to Total Assets. The results showed that there was a positive
relationship between this variable and profitability, when measured with ROAA,
ROAE and PBT.
 The relationship with NIM was negative, because the more capital is held as a
safety buffer the less resources are invested in the banks’ interest earning business.
Finally, regarding the bank specific factors, Cost to Income ratio which was used as
a control variable, was negatively related. The macroeconomic variables of annual
Growth Rate of GDP and Inflation, had a positive sign when we examined ROAA,
ROAE and PBT while in the NIM equation, inflation had a negative sign.

Overall, we can conclude that our study demonstrated that in European Banks,
during the period 2009 – 2015, liquidity as measured by the balance sheet
measures, mostly had a negative impact on profitability, measured as a return on
invested funds or as absolute profits (ROAA, ROAE, NIM, PBT). The Non Performing
Loans phenomenon contributed towards this result by deteriorating the
performance of the loan portfolio of each bank. According to this study, banks
should maintain their liquidity levels mostly though their capital reserves (e.g. Tier
1 Capital) and take actions to mitigate the credit risk of their investments, as well as
their financing gap which imposes constraints in their funding procedure. At the
end, further research could be conducted in this particular topic, regarding the
impact of liquidity before and after the crisis on profitability and whether the
liquidity condition of European banks after 2007 is adequate to prevent a similar
crisis.
Internet Sources

Europa-European Union Website (2016) https://europa.eu/

European Central Bank (2016) http://www.ecb.europa.eu/home/html/index.en.html
Investopedia (2016) http://www.investopedia.com/
ResearchGate (2016) https://www.researchgate.net/
Appendix

1.
The following table presents the banks that are included in this study’s sample:

Banks Total Assets Ba Total Assets

2015 nk 2015
s
1 HSBC Holdings Plc 2,409,656,000 26 ABN AMRO Group N.V. 424,950,470
2 BNP Paribas 2,171,141,028 27 Svenska Handelsbanken 298,802,598
3 Crédit Agricole Group 1,849,601,556 28 Nationwide Building 297,337,403
Society
4 Deutsche Bank AG 1,773,685,387 29 Skandinaviska Enskilda 295,702,299
Banken AB
5 Barclays Bank Plc 1,660,828,428 30 Crédit Industriel et 276,511,709
Commercial SA
6 Banco Santander SA 1,459,183,477 31 KBC Group 274,747,963
7 Société Générale SA 1,452,793,711 32 Landesbank Baden- 254,779,536
Wuerttemberg
8 BPCE Group 1,270,043,571 33 Swedbank AB 254,579,539
9 RBS 1,208,369,913 34 La Banque Postale 238,114,321
10 Lloyds Banking Group 1,195,447,565 35 BFA Tenedora de 232,660,646
Plc Acciones SAU
11 UBS AG 950,767,093 36 Banco de Sabadell SA 227,139,688
12 UniCredit SpA 936,781,072 37 Erste Group Bank AG 217,466,961
13 ING Bank NV 912,931,970 38 Nykredit Realkredit A/S 202,604,532
14 Credit Suisse Group AG 827,341,023 39 Norddeutsche 197,058,251
Landesbank NORD/LB
15 BBVA SA 816,633,656 40 Belfius Banque SA 192,664,239
16 Credit Mutuel 805,453,471 41 Banca Monte dei Paschi 184,008,713
di Siena SpA-Gruppo
17 Intesa Sanpaolo 736,522,604 42 HSBC France SA 183,405,556
18 Cooperatieve Rabobank 729,856,300 43 Banco Popular Espanol 172,727,167
SA
19 Nordea Bank AB 704,265,663 44 Deutsche Postbank AG 163,959,720
20 Standard Chartered Plc 640,483,000 45 Raiffeisen Zentralbank 150,708,549
Oesterreich AG - RZB
21 Commerzbank AG 579,903,113 46 Le Crédit Lyonnais SA 148,927,602
22 KfW Group 547,602,622 47 Crédit Foncier de France 146,403,922
SA
23 Bank of Scotland Plc 505,828,404 48 Bank of Ireland 142,580,296
24 Danske Bank A/S 482,119,749 49 OP Financial Group 136,249,322
25 DZ Bank AG 444,573,769 50 Banco Popolare - 131,202,615
Società Cooperativa-
Banco Popolare
 The following tables show the results from the Hausman test and the manual
autocorrelation test that were performed for each model. The tables are derived
from Eviews 9 software.

MODEL 1: ROAA

Correlated Random Effects - Hausman Test

H0: Random effects are more appropriate

Chi-Sq.
Chi-Sq.
Test Summary Statistic d.f. Prob.

Cross-section random 40.435638 9 0.0000

Manual Autocorrelation Test Results (Auxiliary Regression)

Prob
Variable Coefficient Std. Error t-Statistic .

0.00
DRESIDROAA(-1) -0.439682 0.058914 -7.463089 00

DRESIDROAA(-1): The lagged value of residuals that were derived from the initial
equation (1), using the variables in their first differences. Following there is also the
Wald test that was employed.

Wald Test:
H0: Coefficient = -0.5

Probabilit
Test Statistic Value Df y

t-statistic 1.023833 249 0.3069

F-statistic 1.048234 (1, 249) 0.3069
Chi-square 1.048234 1 0.3059
MODEL 2: ROAE

Correlated Random Effects - Hausman Test

H0: Random effects are more appropriate

Chi-Sq.
Chi-Sq.
Test Summary Statistic d.f. Prob.

Cross-section random 50.494340 9 0.0000

Manual Autocorrelation Test Results (Auxiliary Regression)

Prob
Variable Coefficient Std. Error t-Statistic .

0.00
DRESIDROAE(-1) -0.405108 0.057852 -7.002543 00

DRESIDROAE(-1): The lagged value of residuals that were derived from the initial
equation (2), using the variables in their first differences. Following there is also the
Wald test that was employed.

Wald Test:
H0: Coefficient = -0.5

Test Statistic Value df Probability

t-statistic 1.640258 249 0.1022

F-statistic 2.690445 (1, 249) 0.1022
Chi-square 2.690445 1 0.1010
MODEL 3: NIM

Correlated Random Effects - Hausman Test

H0: Random effects are more appropriate

Chi-Sq.
Chi-Sq.
Test Summary Statistic d.f. Prob.

Cross-section random 81.556678 9 0.0000

Manual Autocorrelation Test Results (Auxiliary Regression)

Prob
Variable Coefficient Std. Error t-Statistic .

0.72
DRESIDNIM(-1) -0.019968 0.056957 -0.350579 62

DRESIDNIM(-1): The lagged value of residuals that were derived from the initial
equation (3), using the variables in their first differences. Following there is also the
Wald test that was employed.

Wald Test:
H0: Coefficient = -0.5

Test Statistic Value df Probability

t-statistic 8.428007 249 0.0000

F-statistic 71.03131 (1, 249) 0.0000
Chi-square 71.03131 1 0.0000
MODEL 4 PBT

Correlated Random Effects - Hausman Test

H0: Random effects are more appropriate

Chi-Sq.
Chi-Sq.
Test Summary Statistic d.f. Prob.

Cross-section random 25.495720 9 0.0025

Manual Autocorrelation Test Results (Auxiliary Regression)

Prob
Variable Coefficient Std. Error t-Statistic .

0.00
DRESIDPBT(-1) -0.388528 0.055836 -6.958432 00

DRESIDPBT(-1): The lagged value of residuals that were derived from the initial
equation
(4), using the variables in their first differences. Following there is also the Wald test
that
was employed.

Wald Test:
H0: Coefficient = -0.5
H : Coefficient =
Test Statistic Value df Probability

t-statistic 1.996440 249 0.0490

F-statistic 3.985774 (1, 249) 0.0470
Chi-square 3.985774 1 0.0459
Next there are seven graphs that present the evolution of the liquidity measures
that were used in this study, over the period 2009 – 2015. Each year’s value is
calculated as an average of the 50 sample banks data.

Cash and Due from Banks

$40,
000,
000

$35,
000,
000

$30,
000,
000 200
9 2010 2011 2012 2013 2014 2015
$25,
000,
000

$20,
000,
000

$15,
000,
000

Total Customer Deposits

$320,
000,0
00

$310,
000,0
00

$300,
000,0
00 200
9 2010 2011 2012 2013 2014 2015
$290,
000,0
00

$280,
000,0
00

$270,
000,0
00
$36,0
Impaired Loans to Gross Loans
00,00
0 6.5%

$35,0 6.0%
00,00
0 5.5%

5.0%
$34,0
00,00
4.5%
0
4.0%
$33,0 2009 2010 2011 2012 2013 2014 2015
00,00 Cost to Income
0 Ratio

$32,0 70%
00,00
0 68%

66%
$31,0
00,00 64%
0
62%

60%
2009 2010 2011 2012 2013 2014 2015

Net Loans to Total Assets

49%

48%

47%

46%

45%
2009 2010 2011 2012 2013 2014 2015
 Financing Gap Ratio
.16%

.14%

.12%
 Financing Gap Ratio
.16%

.14%

.12%

.10%

.08%
2009 2010 2011 2012 2013 2014 2015