See discussions, stats, and author profiles for this publication at: https://www.researchgate.

net/publication/366950158

Sensitivity of banks' capital in the Federation of Bosnia and Herzegovina to

changes in exogenous and endogenous factors

Article in BH Ekonomski forum · January 2021

DOI: 10.5937/bhekofor2102053N

CITATIONS READS

2 14

3 authors, including:

Elman Nadzakovic Senada Dupovac

University of Zenica University of Zenica
4 PUBLICATIONS 5 CITATIONS 2 PUBLICATIONS 2 CITATIONS

SEE PROFILE SEE PROFILE

All content following this page was uploaded by Elman Nadzakovic on 16 February 2023.

The user has requested enhancement of the downloaded file.

 BH ECONOMIC FORUM 53

Elman Nadžaković1
Senada Dupovac2
Bedrija Hromić3

SENSITIVITY OF BANKS’ CAPITAL IN

THE FEDERATION OF BOSNIA AND HERZEGOVINA
TO CHANGES IN EXOGENOUS AND
ENDOGENOUS FACTORS

ABSTRACT
In order to ensure that risky assets are covered by the bank’s capital, the regulator
has prescribed a minimum capital adequacy ratio (CAR). The prescribed CAR, whi-
ch in the Federation of Bosnia and Herzegovina is 12%, represents the ability of
banks to absorb losses caused by bad placements, expressed as the ratio of capital
and risk assets of the bank. The sample of this research was performed at the level of
the whole banking system of the Federation of Bosnia and Herzegovina. The primary
goal of this paper is to try to identify exogenous and endogenous factors that affect
the capital sensitivity of banks operating in the Federation of Bosnia and Herze-
govina. The research is focused on the period from 2014-2020, every quarter, and
the research will explain descriptive statistics, correlation analysis and regression
analysis. The research estimated two regression models, assumptions were chec-
ked and the obtained results were interpreted. The results of the first model showed
that there is a statistically significant positive correlation between the leverage rate
and the capital adequacy ratio, and a statistically significant negative correlation
between the GDP growth rate and the capital adequacy ratio. The second model
confirmed that there is a negative statistically significant correlation between the
variables return on average assets (ROAA), total income / average assets (UPPA),
net interest income / average assets (NIM) with the capital adequacy ratio.
Keywords: capital adequacy ratio, GDP, leverage, banks, non - performing loans
JEL: G22, D24, E58

1 PhD Student, Faculty of Economics University of Zenica, e-mail: elman.nadzakovic@gmail.com

2 PhD Student, Faculty of Economics University of Zenica, e-mail: s.dupovac@gmail.com
3 MBA in Executive Management, Victoria, BC, Canada, hurija83@gmail.com
54 UNIVERSITY OF ZENICA FACULTY OF ECONOMICS

1. INTRODUCTION
As one of the significant market players, banks conduct their business within the
regulatory terms and conditions. The capital adequacy requirement represents one of
the most important themes for bank governance and financial regulators. Accordin-
gly, the Capital Adequacy Ratio (CAR), as a recommended measurement of a bank’s
available capital, indicates banks’ ability to absorb losses. The higher CAR implies
higher absorption of losses due to non-performing loans. The CAR is used to protect
the bank depositors while minimizing risk exposures.
The most important regulator is the Basel Committee on Bank Supervision (BCBS)
that provides a baseline for a stronger structure of the bank capital so to minimize
the risk of non-payments. The BCBS sets a framework for banks to calculate their
capital via CAR as the minimum requirement. The corresponding bank agencies set
the minimal CAR for Bosnia and Herzegovina.
This undertaking is structured via sections. After introductions, the authors researc-
hed and elaborated on the relevant literature. The influence of non-performing loans
on CAR is presented as a separate section and describes the relationship between
non-performing loans and CAR. The goal of this research, research questions, hypot-
hesis, data and methodology are provided in the fourth section as well as the analysis
and discussion about the findings and results of the research. The last section provi-
des a summary and final research notes.

2. Literature review
The sensitivity of the capital, as the main theme of this research, is demonstrated
through the observations of capital adequacy. The structure of the capital as the pillar
of capital adequacy represents a combination of different sources of funding. This
influential area of economic theory attracted numerous scientists so the Modiglia-
ni-Miller theorem 1958 formed the basis on the structure of the capital. Following
this theorem, numerous studies appeared and provided views on the optimal stru-
cture of the capital. However, it is important to note, that the majority of scientists
focused on the non-financial sector while a few scientists focused on the banking
sector exclusively.
Hewaidy and Alyousef (2018) researched the influence of seven (7) different
bank-specific and two (2) macroeconomic determinants on the adequacy of bank’s
capital based on the evidence from Kuwaiti banks. The findings pointed at a signi-
ficant negative relationship between the CAR and the bank size, profitability, asset
quality and liquidity. On the other hand, the quality of banks’ governance is in a
positive correlation with the CAR. The type of bank has no significant link with the
 BH ECONOMIC FORUM 55

CAR. The interesting finding in this study is that liquidity has a negative correlation
with the CAR measurement. In addition, it is found that there is no influence on pro-
fitability when using the Return on Asset and the Return on Equity (ROA & ROE)
measurements. With respect to macroeconomic determinants, the authors found that
the inflation rate is negatively correlated with the CAR, but other macroeconomic
variables do not have a significant influence on the CAR.
Dao and Nguyen (2020) found out that CAR and indicators of the banks’ perfor-
mance have, statistically significant and positive linkage with credit growth, GDP
growth, equity-to-deposit ratio and cost-to-income ratio in relation to CAR and ROE
variables. The findings suggest that commercial banks should control the credit
growth, GDP growth, equity-to-deposit ratio and cost-to-income ratio if they want to
maintain an adequate level of capital.
Giesecke, Dixon and Rimmer (2017) researched the effects of an increase in CAR
on an increase of potential expenditures related to the unanticipated macroeconomic
impacts as commercial banks reduce leverage. In that context, the authors believe
that study disclosed very small macroeconomic consequences due to an increase in
CAR.
El-Ansary, El-Masry and Yousry (2019) research encompassing 37 Islamic banks
and 75 conventional banks concludes the following: there is a significant correlation
between CAR and the size of the bank, operational efficiency and the growth rate of
GDP resulting in a retroactive impact on CAR. While study findings with respect to
the Islamic banks indicate a significant linkage between CAR and deposit-to-asset
ratio, the conventional banks indicate a linkage between CAR and profitability, cre-
dit and portfolio risks.
Aktas and others (2015) estimated the impact of banks variables and external fa-
ctors on banks’ coefficient of the adequacy of capital in the South Eastern European
region. The banks’ size, profitability (ROA), leverage, liquidity, net interest margin
(NIM), and risk are used as independent variables in the model, while an economic
growth rate, inflation, real interest rate, the stock market volatility index within Eu-
rozone, coverage of deposit insurance and banks’ governance indicators are added to
the original model for control of external factors.
For this study, during the 2007-2012 period, authors used annual data coming from
71 commercial banks from 10 different countries within the region. The study fin-
dings indicate that dimensional exploratory variables such as size, ROA, leverage,
liquidity, net interest margin, and risk have a statistically significant correlation with
CAR. At the same time, external factors such as economic growth rate, the stock
market volatility index within the eurozone, coverage of deposit insurance and go-
vernance have a statistically significant impact in determining CAR for the regional
banks.
56 UNIVERSITY OF ZENICA FACULTY OF ECONOMICS

Dreca (2013) explored the impact of capital structure, bank size, profitability indi-
cators, the portion of deposits and loans in total assets, and leverage on the CAR.
The study encompassed 10 banks with data collected over the 2005-2010 period.
According to the findings, SIZE, DEP, LOA, ROA, ROE and LEV have a significant
impact on the CAR. On the other hand, it appears that LLR and NIM do not have a
significant effect. The variables such as SIZE, DEP, LOA and ROA negatively impa-
ct CAR; while variables such as LLR, ROE, NIM and LEV are positively linked with
CAR. Except for LOA and ROA, all variables have expected signs.
Eldomiaty, Eldin and Azzam (2016) aimed to demonstrate empirical evidence expla-
ining the internal financial indicators of banks with influence on CAR and the stress
test of the capital adequacy requirement on Egyptian banks. The results indicate that
return on assets, return on equity and asset-based market share have a significant
impact on CAR. The ratio of loans-to-deposits and non-performing loans have a po-
sitive significant effect on CAR. The coefficient of liquidity expressed in US dollars
has a trivial effect on CAR. The coefficient of liquidity in Egyptian pounds has an
insignificant effect on CAR. The coefficient of loans-to-deposits, non-performing
loans and liquidity in US dollars are robust and significant determinants on CAR.
Benazic and Masic (2016) pointed out an existence of a stable co-integrative linkage
between macroeconomic variables and CAR. In the long run, the forecasted increase
in the GDP growth rate leads towards a decline of CAR of Croatian banks, while an
increase of the interest rate on credits in kuna with the currency clause and depre-
ciative real-effective currency index of kuna leads towards an increase in CAR of
the Croatian banks. On the other hand and short-term positive changes in the GDP
growth rate and interest rate on credits in kuna with a currency clause have a positive
effect on the Croatian banks CAR while a positive change in a real effective currency
value of kuna has a negative impact on the Croatian banks’ CAR.

3. The influence of non-performing loans on car

life, today is unimaginable without loans to the extent that an individual’s growth is
assessed via credit rating. The societies, the nations, are assessed via credit rating.
Accordingly, the term credit presents an open field for further research. There are
numerous definitions of the term credit so Vuckovic defined it as ‘temporary lending
of determined buying power so the consumer of credit would act or perform on the
market with borrowed resources while stating the time-limited ownership. (Vucko-
vic, 1958)
 BH ECONOMIC FORUM 57

Based on this definition, Bjelica believes that the buying power is a manifestation
of the economic power and therefore, the term credit implies a temporary service of
the lender to the borrower by providing access to a pre-determined amount of money
with implied duties of the borrower to return the monetary resources within a given
timeframe almost the same amount of the borrower’s buying power. (Bjelica, 2001,
p. 227)
Simply stated, credit is a way of lending resources, mostly financial, which a creditor
or lender extend to a debtor or borrower, subject to fulfilling specific requirements,
while a debtor or borrower undertakes the duty of returning the same resources in
accordance with the pre-determined terms and conditions. Banks approve loans to
individuals and corporations. This makes lending a core activity in most banks. Lo-
ans represent the main revenue stream for banks as well as the main source of risk(s).
For that reason, it is of great importance to have adequate bank governance related
to the management of the lending risks. In the end, the stability of commercial banks
and the strengthening of the safety and the certainty of the banking sector as a whole
are the main dependencies for a successful lending business. (Agic & Dusanic Gacic,
2019)
The non-performing loans, as one of the main risks, imply financial assets that do
not provide interest or scheduled payments for a specific period and as such become
the problem known as economic stagnation. Therefore, the minimization of non-per-
forming loans represents also the main requirement for the economic growth. (Islam
& Islam, 2018) researched and introduced empirical evidence of non-performing
loans influencing, significantly, the capital adequacy of Bangladesh banks. The re-
sults point out that there is a significant negative correlation between non-perform-
ing loans and CAR as per the data of this research. Figure 1 depicts trends NPL rate
and CAR of the banks that are conducting their business within the Federation of
Bosnia and Herzegovina.

Figure 1: Trends NPL and CAR

Source: Data created and prepared by authors

58 UNIVERSITY OF ZENICA FACULTY OF ECONOMICS

Figure 1 depicts the timeline for the 4Q 2014 to the 4Q 2020 where is a noticeable
trend in the CAR increase while the NPL (total, corporations and individuals) de-
clines. The same trend shows a negative correlation between CAR and NPL, in-
dicating a weakening of the credit risk. Continuous improvement in the quality of
the credit portfolio within a given timeframe is due to the banks’ efforts to write off
non-performing loans and to the increased frequency in re-programming of credits.
In addition, an increase in efficiency of the credit risk management by banks as well
as opportunities to pay out credits via re-programs and imposed moratorium of credit
duties from clients who face difficulties in repaying credit dues, greatly contributed
to preventing further weakening of the quality of the credit portfolio. Despite a pos-
itive trend, it is important to emphasize that the NPL rate of the European Union at
the end of Q4 2020 recorded 2.63%, which represents the aim of the analyzed bank-
ing sector within the Federation of Bosnia and Herzegovina.

4. Data and variables

The research was conducted on the whole population meaning the Federation of Bo-
snia and Herzegovina banking system as a whole. Data are based on the sample of
15 banks operating within the Federation of Bosnia and Herzegovina. At the end of
2020, the banking system operated with 531 organizational units, which employed
a total of 6522 workers. The total net asset at the level of the banking sector is 24.4
billion, and the total capital is 3.1 billion BAM.
All data are collected from the reports coming from the Banking Agency of the Fe-
deration of Bosnia and Herzegovina, reports from the Central Bank of Bosnia and
Herzegovina, and Federal Office of Statistics for the Entity of Federation of Bosnia
and Herzegovina. The research was performed during the 2014-2020 period. The
dependent variable for the research model is CAR representing the relationship of
the bank’s capital and risk-weighted assets. Independent variables, with their in-
fluence on the dependent variable as part of this research, are Leverage Rate (LR),
GDP Growth Rate, Return on Average Assets (ROAA), Total Revenue/Average asset
(UPPA) and Net Interest Margin (NIM). The following table shows the expected
effects of independent variables on a dependent variable (CAR).
Table 1: Influencing factors on CAR
Short Name of Variable Factor Trend
LR Leverage Rate +
GDP GDP Growth Rate -
ROAA Return on Average Assets -
UPPA Total Revenue / Average Asset -
NIM Net Interest Margin -
Source: Authors
 BH ECONOMIC FORUM 59

DESCRIPTIVE STATISTICS:
For the purpose of calculating the influence of exogenous factors on a dependent
variable CAR, the model included variables such as LR rate and GDP growth rate.
In order to establish the influence of endogenous factors, the model included ROAA,
UPPA and NIM. All calculations are done while using STATA 12 software.
Table 2: Descriptive statistics of variables
Variable Obs Mean Std.Dev Min Max
CAR 25 16.44 1.221338 15.1 19.1
LR 25 9.858 .419305 9.2 10.54
GDP Growth 25 2.036 3.058987 -9.3 4.5
ROAA 25 .8028 .3177331 .3 1.3
UPPA 25 2.8836 1.82234 .5 5.6
NIM 25 1.9372 .9350914 .6 3.5

Source: Data created and prepared by authors

Within an observable period the CAR increased from 15.1% to 19.1%, which is
greater than the legal requirement minimum of 12%. In addition, the LV rate in-
creased from 9.2% to 10.54% while the GDP Growth rate increased from -9.3% to
4.5%. The standard deviation as the indicator of variation or dispersion of a set of
values, indicate the different performance of variables so the GDP Growth rate and
UPPA recorded a significant downfall at the end of 2020, resulting in a high standard
deviation. Unlike GDP Growth and UPPA, the other variables didn’t display high
oscillations in a given period with the exception of the continual gentle growth of
CAR, as supported by the standard deviation of the CAR trend.

5. The influence of exogenous factors on car

For the purpose of estimating the influence of exogenous factors on the CAR, the
linear regression model, according to which the parameters of the linear function
were selected as they explain the variables based on the principle of the smallest
quadrants. In order to establish the best influence of exogenous factors on CAR, the
researchers used the following regression model:
Yi = (b0 + b1Xi) + εi (1)
Where Yi = i-th dependent variable, Xi = i-th the value of independent variable,
b0 = constant (free member), b1 = ascent, a εi = random error. Including all indepen-
dent and dependent variables in the formula, the following model was created:
CAR = b0 +b1LR + b2GDP Growth + ε (2)
where the CAR represents the dependent variable, b0, represents a constant or free
member, b1LR and b2GDP Growth represent the ascent parameter of independent
variables, and ε represents a random error.
60 UNIVERSITY OF ZENICA FACULTY OF ECONOMICS

The LR rate, as the first independent variable indicates concerns such as: is it worth
using NPL in financing the banking business model as long as the lending activi-
ty results in the higher lending rate from the lending interest rate. The sensitivity
indicator of profitability is especially useful in changing performance between the
capital structure and financial debt. The GDP Growth rate, as the second independent
variable, represents the measure of the total product on the accounts of national re-
venue. It is used as the measure of the total economic power of one state or country
and represents the most frequently used macroeconomic variable for the purpose of
researching the influence on a dependent variable.
The forecasting model is shown via the following formula:

Table 3: Results of regression

Independent Variable Coef. t-value p-value
LR 2.215603 7.41 0.000
GDP Growth -.1009896 -2.46 0.022
_cons -5.195802 -1.74 0.096
F-value 0.0000
R2 0.8146
Adjusted R2 0.7978
Source: Data created and prepared by authors
In search for an answer to the question: „Can independent variables reliably fore-
cast dependent variable? “, testing was conducted and the following hypothesis was
posed:
H0: The model is significantly better in forecasting dependent variable when only
the arithmetic mean was used for forecasting purposes.
H1: The model is not significantly better in forecasting dependent variable when
only the arithmetic mean was used for forecasting purposes.
Upon hypothesis testing completion, the result of the F-test (p-value = 0.0000)
shows that hypothesis 0 is refuted, therefore, both tested variables have a significant
influence on CAR as dependent variable affirming that the model is significantly
better in forecasting dependent variable when only the arithmetic mean is used for
forecasting purposes.
The results of the analysis of variance (ANOVA), used to test whether or not the
model is significantly better in forecasting the value of a dependent variable when
only the arithmetic mean is used for forecasting, is significant (F(2,22) = 48.34, p<
.001). As the p value is 0.0000 it is reasonable to conclude that dependent variables
such as LR and GDP Growth rates reliably forecast the dependent CAR variable.
 BH ECONOMIC FORUM 61

The model explains 81.46% variability of the dependent variable (CAR), meaning
79.78% variance for CAR variable is explained by independent variables such as
LR and GDP Growth rates. The research studies conducted for the Croatian banking
system indicate that the GDP Growth rate causes a long-term drop in the CAR rate.
(Benazic & Masic, 2016) It is very important to understand how a macroeconomics
affair, especially in instances of this research (GDP Growth), influences the every-
day life of individuals. The research shows that the negative value of GDP Growth
leads to an increase in CAR, meaning as GDP Growth decline results in an increase
of interest rates, lending rates…(the list can go on), which leads to an increase in the
rate of NPL and in the long-run leads to an increase in CAR trends. Even though, the
model is utilized from the total population as opposed to the sample, almost the same
explanation would follow in relation to the variance of the dependent variable. Given
the model significance, it is reasonable to proceed with the further interpretation of
collected results.

Table 4: Correlation matrix

LV rate GDP Growth rate
LR rate 1.0000
GDP Growth rate -0.4471 1.0000
Source: Data created and prepared by authors
If amongst the predictors exist a perfect linear relation, the forecast for regression
model cannot be calculated singularly, and given the increasing level of multicolline-
arity, the forecast of coefficients of regression models become unstable and standard
errors can be unstable too.
From the designed correlation matrix it is visible that there is no high correlation
between LR and GDP Growth rate. As a supplemental review, the variation inflation
(VIF) and tolerance (1/VIF) factors were used. VIF indicator means the factor of
variance inflation. As a rule, variables with VIF values greater than five (5) represent
a basis for further exploration, while VIF values greater than ten (10) definitely, and
show the existence of very high multicollinearity. The tolerance, defined as 1/VIF,
is used for checking the levels of multicollinearity. The tolerance lower than 0.1 is
comparable with the VIF 10 indicating that there is a significant problem. The tole-
rance 1/VIF lower than 0.2 indicates that there is a potential problem.

Table 5: Multi-correlation matrix: variance inflation factor (VIF)

Variable VIF 1/VIF
GDP Growth 1.25 0.800076
LR 1.25 0.800076
Mean VIF 1.25
Source: Data created and prepared by authors
62 UNIVERSITY OF ZENICA FACULTY OF ECONOMICS

The output shows that all individual values are significantly below 10, while all va-
lues for tolerance are greater than 0.2. Simultaneously, the average VIF is 1.00 the-
refore, confidently can be stated that within indicators there is no multicollinearity.
The coefficient b1 that is attached to the LR variable tells that if LR increases by 1%
it is expected an average increase of CAR in the amount of 2.21% points. The coeffi-
cient b2 that is attached to the GDP Growth variable tells that if GDP Growth increa-
ses by 1%, the decrease in CAR in the amount of .100% points is expected. Based on
findings, it is reasonable to say that both coefficients are statistically significant. The
negative correlation between GDP Growth and CAR tells that an increase in GDP
leads to a decrease in CAR and vice versa. The negative correlation and higher CAR
than the regulated minimum requirement, lead to the conclusion that banks increase
CAR because of NPL resulting in lower growth of GDP.
In order to compare the relative strength of different predictors in the model, the opti-
on beta, used to calculate standardized regression coefficient (z-score), was included.
Given the fact that all beta coefficients are expressed through standard deviation,
their value is standardized. This enabled the comparison of the forecasts of the rela-
tive strength of each predictor. Comparing the achieved standardized values led to
the conclusion that LR (z-score=.7606522) has higher relative importance within a
model in relation to the GDP Growth rate (z-score= -.2529404). This way all features
were tested and all assumptions of regression analysis were completed.

6. The influence of endogenous factors on car

On the basis of gathered data, the following regression model for determining the
influence of endogenous factors on CAR was created:
CAR = b0 +b1ROAA + b2UPPA + b3NIM + ε (3)
The forecasting model is expressed in the following formula:

Table 6: Results of regression

Independent variables Coef. t-value p-value
ROAA -.2525663 -0.22 0.827
UPPA .810234 2.61 0.016
NIM -1.473914 -2.08 0.050
_cons 17.16164 26.94 0.000
F-value 0.0935
R2 0.2579
Adjusted R2 0.1518
Source: Data created and prepared by authors
 BH ECONOMIC FORUM 63

In search for an answer to the question “Can independent variables reliably forecast
dependent variable?”, testing was conducted and the following hypothesis posed:
H0: The model is significantly better in forecasting dependent variable when only
the arithmetic mean was used for forecasting purposes.
H1: The model is not significantly better in forecasting dependent variables when
only the arithmetic mean was used for forecasting purposes.
The result of the analysis of variance (ANOVA), used to test whether or not the
model is significantly better in forecasting the value of the dependent variable when
only the arithmetic mean is used for forecasting, is significant (F(3, 21) = 20.43, p<
.10). It is worth noticing that determined variables are significant and that the coe-
fficient of NIM -1.47 is significantly different from 0 while using alpha 0.05 due to
its p-value of 0,050. Like the UPPA, the ROAA (-0.2525663) is statistically insigni-
ficant given that the p-value (0,827) is significantly higher than 0.05. With respect
to the unified model, p-value is 0.0935 and therefore, it is reasonable to conclude
that independent variables such as ROA, UPPA and NIM reliably forecast CAR as
a dependent variable. The model provides 25.79% variability from the dependent
variable (CAR), meaning 15.18% variance for the CAR variable is explained by
independent variables ROA, UPPA and NIM. If the model is taken out of the total
population as opposed to the sample, it would be the same explanation of variance
values for dependent variables.

Table 7: Correlation matrix

ROAA UPPA NIM
ROAA 1.0000
UPPA 0.6672 1.00001
NIM 0.7701 0.9125 1.0000
Source: Data created and prepared by authors
Table 7 indicates that there is no high correlation between dependent variables
ROAA, UPPA and NIM. As an additional supplemental review, the VIF and toleran-
ce (1/VIF) indicators were used.

Table 8: Multi-correlation matrix: Variance Inflation Factor (VIF)

Variable VIF 1/VIF
UPPA 8.30 0.120499
NIM 6.09 0.164316
ROAA 2.50 0.399345
Mean VIF 5.63
Source: Data created and prepared by authors
64 UNIVERSITY OF ZENICA FACULTY OF ECONOMICS

The output shows that all individual values are below 10 while all values for toleran-
ce are greater than 0.1. Simultaneously, the average VIF is 5.63, therefore, confiden-
tly can be stated that within indicators there is no multicollinearity.
The coefficient b1 that is attached to the ROAA variable tells that if the profit on ave-
rage assets increases by 1% then it is expected to have the average lowering CAR for
0.252% points. Given that the ROAA indicator puts in a relationship between profit
and average asset, then it can be stated that NPL, part of the assets, with its increase
negatively influence this coefficient (lowering coefficient). The coefficient speaks
about the governance efficiency in the management of assets as to accomplish higher
returns, meaning the management of investments in generating profits. Hence, the
bank governance and management must continuously work on the prevention and
minimization of non-performing loans in order to have successful banking and reali-
zed profit. The coefficient b2 that is attached to the UPPA variable tells that if the re-
lationship between the total return and average asset lowers by 1% then it is expected
that CAR increases by .810% points. The coefficient b3 that is attached to the NIM
variable tells that if relationship between the net interest income and average assets
lowers by 1% then it is expected that CAR decreases by 1.473% points. Based on the
results, it can be concluded that variables UPPA and NIM are statistically significant
at the level of 5%.
Observing the correlation between independent and dependent variables, it is reaso-
nable to conclude that an increase in the value of all estimated variables attached to
the income and profit positively influence lowering the capital adequacy. Specifica-
lly, it is the case with the UPPA and NIM variables that are statistically meaningful
when at the level of 5%. As previously stated, the higher CAR leads to an increase in
the banks’ capacity to absorb losses that occurred due to poor placements of assets.
Given that the CAR value is higher than the legal minimum requirement, it is advi-
sable for banks that with an increase in the higher-quality asset placements the banks
can expect an increase in the NIM value. Ultimately, this would result in an increase
in income and profit while lowering the CAR value.
Comparing the standardized coefficient (beta) it was concluded that the relationship
between net interest income and average asset (z-score=1.208) has the highest re-
lative significance inside the model in relation to the ROAA (z-score= -.0657) and
UPPA (z-score=-1.128). This way all features were tested and all assumptions of
regression analysis were completed.
 BH ECONOMIC FORUM 65

7. CONCLUSION
In order to arrange capital requests for banks operating within the Federation of Bo-
snia and Herzegovina, the unified rules were established in accordance with Basel
III Accord, an international regulatory accord of agreed upon standards for capi-
tal adequacy. It is of importance to understand that the stability within the banking
system represents a foundation for the avoidance of eventual economic shocks. Bro-
adly looking, capital adequacy shows the ability to absorb losses so the higher capi-
tal adequacy banks have the higher potential to absorb losses due to non-performing
loans. Therefore, it is very challenging to decide whether or not to have a higher or
lower CAR. The stability perspective tells that it is better to have the higher CAR;
however, the profitability perspective tells that is desirable to have the lower CAR.
This research tested endogenous and exogenous factors impacting capital sensitivity.
For that reason, the research was focused on the 2014-2020 period. The tested va-
riables are capital adequacy ratio (dependent), leverage rate, GDP growth, return on
average assets (ROAA), total income/average asset (UPPA) and net interest income
/ average asset (NIM) (independent).
For the purpose of maintaining the optimal CAR, banks must control and monitor
non-performing loans with applicable credit forecasting and measuring that would
contribute to the decline of the NPL rate. Therefore, if banks want to be ready for
new challenges, they must focus their activities on controlling and monitoring the
NPL rate.
Given the research theme attractiveness and significance of the banks’ capital adequ-
acy for the entire state and society, the future research model should include the
banks operating in the pockets of RS as well as some other determinants that would
better explain the impact of ever-changing exogenous and endogenous factors on the
sensitivity of the capital in Bosnia and Herzegovina.

REFERENCES
1. Adams, Z. & Füss, R. (2010). “Macroeconomic determinants of international
housing markets”, . Journal of Housing Economics, No.19, pp.38-50.
2. Agić, Z. & Dušanić Gačić, S. (2019). “Kreditni potencijal bankarskog sektora
Bosne i Hercegovine”, Transition/Tranzicija, No. 21, pp.1-10.
3. Aktas, R., Acikalin, S., Bakin, B. & Celik, G. (2015). “The Determinants of
Banks’ Capital Adequacy Ratio: Some Evidence from South Eastern European
Countries”. Journal of Economics and Behavioral Studies, No.7, pp.79-88.
4. Benazić, M. & Mašić, A. (2016). “Utjecaj odabranih makroekonomskih varija-
bli na stopu adekvatnosti jamstvenog kapitala hrvatskih banaka”. EFZG Occa-
sional Publications (Department of Finance), pp.63-80.
66 UNIVERSITY OF ZENICA FACULTY OF ECONOMICS

5. Bjelica, V., 2001. Bankarstvo, teorija i praksa. Novi Sad: Štamparija Budućnost.
6. Cellmer, R., Cichulska, A. & Bełej, M. (2020). “Spatial Analysis of Housing
Prices and Market Activity with the Geographically Weighted Regression”.
ISPRS International Journal of Geo-Information, No.9, pp.2-19.
7. Dao, B.T.T. & Nguyen, K.A. (2020). “Bank Capital Adequacy Ratio and Bank
Performance in Vietnam: A Simultaneous Equations Framework”. Journal of
Asian Finance, Economics and Business, pp.39-46.
8. Dreca, N. (2013). “Determinants of capital adequacy ratio in selected Bosnian
banks”. Dumlupınar Üniversitesi Sosyal Bilimler Dergisi EYİ 2013 Özel Sayısı,
pp.149-62.
9. El-Ansary, O., El-Masry, A.A. & Yousry (2019). “Determinants of Capital
Adequacy Ratio (CAR) in MENA Region: Islamic vs. Conventional Banks”. In-
ternational Journal of Accounting and Financial Reporting, pp.287-313.
10. Eldomiaty, T.I., Eldin, A.B. & Azzam, I., 2016. Determinants of Capital Adequ-
acy Ratios Under Basel III: Stress Testing and Sensitivity Analysis on Egyptian
Banks. [Online] Available at: https://ssrn.com/abstract=2788482 [Accessed 12
June 2021].
11. Agencija za bankarstvo Federacije Bosne i Hercegovine. 2017. Godišnji iz-
vještaj, [Online] Available at:https://www.fba.ba/images/Publikacije_Banke/
informacija_o_subjektima_bank_sist_fbih_31_12_2017.pdf [Accessed 05 June
2021].
12. Agencija za bankarstvo Federacije Bosne i Hercegovine. 2020. Godišnji iz-
vještaj, [Online] Available at: https://www.fba.ba/bs/informacija-o-subjekti-
ma-bankarskog-sistema-federacije-bih-sa-stanjem-na-dan-31122020godine-4
[Accessed 13 May 2021].
13. Giesecke, J.A., Dixon, P.B. & Rimmer, M.T. (2017). “The Economy-wide Im-
pacts of a Rise in the Capital Adequacy Ratios of Australian Banks”. Economic
record, No.93, pp.13-37.
14. Hewaidy, A.M. & Alyousef, H.Y. (2018). “Bank-Specific and Macroeconomic
Determinants of Capital”. European Journal of Economics, Finance and Admi-
nistrative Sciences, No.99, pp.5-20.
15. Islam, M.S. & Islam, M.Z., 2018. Non-performing loan as eroding factor of
capital adequacy: evidence from banking industry in bangladesh. Romanian
economic and business review, Available at: https://www.researchgate.net/
publication/328354054_non-performing_loan_as_eroding_factor_of_capi-
tal_adequacy_evidence_from_banking_industry_in_Bangladesh [Accessed 13
June 2021].
 BH ECONOMIC FORUM 67

16. US National Research Council, 2013. Acute Exposure Guideline Levels for Se-
lected Airborne Chemicals: Volume 14. Washington, DC: The National Acade-
mies Press.
17. Ridker, R.G. & Henning, J.A. (1967). “The Determinants of Residential Property
Values with Special Reference to Air Pollution”. The Review of Economics and
Statistics, No.49, pp.246-57.
18. Rosen, S. (1974). “Hedonic Prices and Implicit Markets: Product Differentiati-
on in Pure Competition”. Journal of Political Economy, No.82, pp.34-55.
19. Vučković, M., 1958. Bankarstvo. Beograd: Naučna knjiga.
68 UNIVERSITY OF ZENICA FACULTY OF ECONOMICS

Elman Nadžaković
Senada Dupovac
Bedrija Hromić

OSJETLJIVOST KAPITALA BANAKA U

FEDERACIJI BOSNE I HERCEGOVINE NA
PROMJENU EGZOGENIH I ENDOGENIH FAKTORA

SAŽETAK
Radi osiguranja da rizična aktiva bude pokrivena kapitalom banke, regulator je pro-
pisao minimalnu stopu adekvatnosti kapitala. Propisana stopa adekvatnosti kapi-
tala, koja u Federaciji BiH iznosi 12% predstavlja sposobnost banke da apsorbira
gubitke nastale lošim plasmanima iskazana kao odnos kapitala i rizične aktive ban-
ke. Uzorak ovog istraživanja je izvršen na nivou populacije cjelokupnog bankarskog
sistema Federacije Bosne i Hercegovine. Primarni cilj ovog rada je da pokuša pre-
poznati egzogene i endogene faktore koji utječu na osjetljivost kapitala banaka koje
posluju u FBiH. Istraživanje je fokusirano na period OD 2014. do 2020. godine,
na kvartalnom nivou, a istraživanje će objasniti deskriptivnu statistiku, korelacio-
nu analizu i regresionu analizu. Tokom istraživanja, estimirana su dva regresiona
modela, provjerene su pretpostavke i interpretirani su dobiveni rezultati. Rezultati
prvog modela su pokazali da postoji statistički značajna pozitivna korelacija između
stope finansijske poluge i stope adekvatnosti kapitala, te statistički značajna nega-
tivna korelacija između stope rasta GDP-a i stope adekvatnosti kapitala. Drugi mo-
del je potvrdio da postoji negativna statistički značajna korelacija između varijabli
dobit na prosječnu aktivu (ROAA), ukupan prihod/prosječna aktiva (UPPA), neto
kamatni prihod/prosječna aktiva (NIM) sa stopom adekvatnosti kapitala.
Ključne riječi: stopa adekvatnosti kapitala, GDP, finansijska poluga, banke, nekva-
litetni krediti
JEL: G22, D24, E58

View publication stats