Article

Elections and bank non-performing loans: Evidence from

developed countries
Peterson K. Ozili1

1 Central Bank of Nigeria, Abuja, Nigeria, email: [email protected]

Abstract: The existing literature has not examined how elections affect bank non-performing loans
and its determinants even though banks are often the largest borrowers to fund election campaigns
in many countries. This study investigates the determinants of bank non-performing loans (NPL)
during election years in 35 developed countries. The fixed effect regression methodology was used
to estimate the determinants of bank non-performing loans during election years. It was found that
the banking sector experienced high NPLs during election years. Efficient banks operating in robust
legal environments have higher non-performing loans during election years. It was also found that
capital adequacy ratio, real GDP growth, loan-to-GDP ratio, cost-to-income ratio, political stability,
and absence of terrorism are significant determinants of bank non-performing loans. The findings
imply that election matters for the persistence of bank non-performing loans in developed countries.

Keywords: non-performing loan; bank performance; election; banks; credit risk; efficiency; capital
adequacy ratio; GDP growth; political stability.

JEL Classification: G21, G28.

1. Introduction
Rising non-performing loans (NPLs) are a major cause of bank failure, and it is a
significant obstacle to the development of the banking sector. Prior studies show that
NPLs have a signaling effect because they convey valuable information about incurred
losses and the quality of banks’ loan portfolios (Ahamed, 2017; Laryea et al., 2016).
Citation: Ozili, P. K. (2024).
Although bank supervisors want fewer NPLs and possibly a single-digit NPL ratio in the
Elections and bank non-performing
banking sector, in reality, the level of NPLs in banks has increased significantly over the
loans: Evidence from developed
years, especially in Europe1—due to the COVID-19 pandemic, the Russia-Ukraine war
countries. Modern Finance, 2(2), 64-
and rising inflation (Kasinger et al, 2021). Elections are also known to introduce shocks in
79.
the business environment. The literature shows that elections affect business activities,
Accepting Editor: Adam Zaremba initial public offerings (IPOs), stock markets, corporate investment, and labor spending
Received: 25 August 2024 (Amore & Corina, 2021; Clarke, 2013; Kayser, 2006; Çolak et al., 2017). However, the effect
Accepted: 26 September 2024 of elections on the banking sector has not been examined extensively. The existing
Published: 27 September 2024 literature has not examined how elections affect bank NPL and its determinants, even
though banks are often the largest borrowers to fund election campaigns in many
countries.
Copyright: © 2024 by the authors.
This study investigates the determinants of NPLs during an election year, focusing
This article is an open-access article on the impact of the election year on NPLs while controlling for bank-specific factors and
distributed under the terms and the quality of legal institutions, political stability, and other factors. Understanding how
conditions of the Creative Commons elections affect bank NPLs is vital because bank lending to politicians may give rise to
Attribution (CC BY) license credit risk, which is the risk that borrowed loans for election campaigns may not be repaid
(http://creativecommons.org/licenses to banks. If the borrower wins the election, banks often respond by writing off the loan or
/by/4.0/). restructuring the loan for delayed repayment of interest and principal in exchange for the

1 https://blogs.lse.ac.uk/europpblog/2019/03/27/why-non-performing-loans-are-still-putting-the-european-banking-union-at-risk/

Modern Finance. 2024, 2(2). https://www.doi.org/10.61351/mf.v2i2.175 https://mf-journal.com/

Modern Finance. 2024, 2(2). 65

borrower’s political patronage to support the bank’s business interests in the country.
Banks can also write off the debt of borrowers affiliated with the ruling party with the
expectation that the ruling party will, in return for the kindness shown by the bank, create
a favorable business environment for the bank. Such loan write-offs have negative
implications for the stability of the banking sector. Furthermore, even if banks do not give
loans to fund election campaigns, banks may still experience non-performing loans that
arise from election uncertainties. The above arguments demonstrate the potential effect of
the election on bank non-performing loans and its overall effect on the banking sector.
I argue that the ‘election year’ is a type of country risk that banks will take into
account if they believe that a change in the current government following general elections
may affect their ability to recover loans from politically connected debtors, thereby
leading to an increase in credit risk and rising non-performing loans. Despite the
importance of elections on bank credit risk, the existing literature has not examined the
behavior of non-performing loans in the banking sector during an election year. I examine
this issue in this study. In the empirical analysis, I analyze data from 35 developed
countries.
The study reveals that the banking sector experiences high NPLs during an election
year. Efficient banks operating in countries with solid legal environments have higher
NPLs during an election year. It was also found that capital adequacy ratio, real GDP
growth, loan-to-GDP ratio, cost-to-income ratio, political stability, and absence of
terrorism are significant determinants of non-performing loans.
This study makes three contributions to the NPL literature. First, it contributes to the
literature examining non-performing loan behavior under various contexts and economic
conditions (see. Podpiera and Ötker, 2010; Zhang et al., 2016; Vithessonthi, 2016; Kauko,
2012). These studies show that the level of non-performing loans in banks is influenced by
economic factors and other unique conditions faced by banks. Second, this study
contributes to the literature investigating institutional factors' influence on non-
performing loans (e.g., Beck et al., 2015; Ghosh, 2015; Bardhan & Mukherjee, 2016). By
controlling for the election year effect, political stability, and legal system quality, this
study presents new insights to understand how unique political and legal institutions
across countries influence the size of non-performing loans during an election year. Third,
this study examines the behavior of non-performing loans in the banking sector of
developed economies with regular elections that are transparent and free from structural
barriers. This presents a natural setting to test the effect of election-year on non-
performing loans in this study.
The remainder of the paper is organized as follows. Section 2 provides an overview
of the relevant literature. Section 3 presents the data and methodology. Section 4 presents
the empirical results. Section 5 presents the conclusion of the study.

2. Literature Review
Many studies examine the bank-specific and external determinants of bank NPLs
(Louzis et al., 2012; Klein, 2013; Nkusu, 2011; Fofack, 2005). For instance, Beck et al. (2015)
examined the macroeconomic determinants of NPLs across 75 countries and found that
the following variables were significant determinants of NPLs: real gross domestic
product growth rate, share prices, the exchange rate, and the lending interest rate. Ozili
(2020) investigates the behavior of NPLs in European systemic and non-systemic banks
and finds that more profitable banks witness higher NPLs regardless of whether systemic
or non-systemic. Systemic banks have fewer NPLs during economic booms and periods
of increased lending, while non-systemic banks experience higher NPLs during increased
lending. Also, the NPLs of non-systemic banks are negatively associated with regulatory
capital ratios.
Ghosh (2015) investigates banks in the United States and shows that greater
capitalization, liquidity risk, poor credit quality, more significant cost inefficiency, and
banking industry size are major factors that increase NPLs, while greater bank
Modern Finance. 2024, 2(2). 66

profitability lowers NPLs. Ghosh (2015) also shows that higher real GDP, accurate
personal income growth rates, and housing price index change reduce NPLs, while
inflation, unemployment, and higher public debt significantly increase NPLs. Bardhan
and Mukherjee (2016) examine the role of bank-specific determinants in explaining the
dynamics of NPLs in Indian banks during the post-liberalization period from 1995 to 2011.
They find that, although there was a significant time persistence of NPLs in the Indian
banking system, larger banks were more prone to loan default than smaller banks. Al-
Khazali and Mirzaei (2017) examine whether oil price shocks have any impact on bank
NPLs and find that a rise in oil prices leads to a decrease in NPLs, and oil price shocks
have an asymmetric effect on bank problem loans, with adverse oil price movements
having a more significant impact than positive oil price movements.
A few studies examined NPLs under several contexts and certain events, but none
have considered the effect of the election event on bank NPLs. For instance, Podpiera and
Ötker (2010) investigated the determinants of credit default swaps during the initial phase
of the global financial crisis in selected European Large Complex Financial Institutions
(LCFIs) from 2004 to 2008. They use a dynamic panel data estimator and find that the
LCFIs’ business models, earnings potential, and economic uncertainty are among the most
significant determinants of credit risk. Zhang et al. (2016) examine the impact of NPLs on
bank behavior in China. They examine 60 city commercial banks, 16 state-owned and
joint-stock banks, and 11 rural commercial banks during the 2006 to 2012 period and find
that higher non-performing loans increase riskier lending, causing further deterioration
in loan quality for Chinese commercial banks and leading to financial system instability.
Osei-Assibey and Asenso (2015) investigate the influence of the central bank’s regulatory
capital on several indicators of bank performance, including NPLs, over the 2002 to 2012
period. They find that banks create more loans when they have excess capital over the
minimum requirement, and the loans subsequently lead to high NPLs. Vithessonthi (2016)
investigates the link between bank credit growth and NPLs in Japan which was
experiencing deflation during the 1993 to 2013 period. Vithessonthi (2016) shows that
bank credit growth is positively correlated with NPLs prior to the global financial crisis
of 2007 but negatively correlates with NPLs afterwards. In addition, credit growth and
NPLs do not affect bank profitability. Overall, their findings suggest that although an
increase in credit supply increases the level of NPLs, it does not lead to higher
profitability.
Furthermore, Kauko (2012) investigates the interrelationship between trade current
account deficits and the number of NPLs before the global financial crisis and finds that
the rapid credit growth from 2000 to 2005 combined with a current account deficit was a
significant determinant of NPLs for the EU countries. Barseghyan (2010) shows that the
existence of NPLs and a delay in government bailout lead to a persistent decline in
economic activity and output. The decline in economic output was caused by a fall in
investment, an endogenous decline in productivity, and the number of firms in Japan.
Cucinelli (2015) examines whether an increase in credit risk (measured as NPLs and loan
loss provisions) during the financial crisis reduces banks' lending activity in Italy. The
study analyze 488 listed and unlisted Italian banks during the 2007 to 2013 period. The
results show that higher credit risk (i.e., non-performing loans and loan loss provisions)
harms bank lending behavior. Ozili (2019) examines the size of NPLs contingent on the
level of financial development and finds a positive association between the level of
financial development and the size of bank NPLs. Bashir et al. (2017) examine the
influence of banking system transparency and competition in reducing the level of NPLs
for Chinese banks from 2000 to 2014. They find that high transparency in the Chinese
banking system decreased NPLs but not in the case of government-owned banks, while
higher competition increases NPLs.
Recent studies such as Ahiase et al. (2024) investigate the macro determinants of bank
NPLs in 53 African countries. They find that the debt-to-GDP ratio, unemployment rate,
regulatory quality, government effectiveness, and inflation are significant determinants
Modern Finance. 2024, 2(2). 67

of NPLs in African countries. Mamoon et al. (2024) examine the effect of central bank
independence and transparency on the occurrence of NPLs. They examine 39 countries
and find that NPLs are fewer in countries where the central bank is free from political
interference. They also find that NPLs are fewer in countries where the central bank is
transparent. Rehman et al. (2024) investigated the effect of corruption control on NPLs in
81 banks in Pakistan, India, and Bangladesh from 2000 to 2019. They used the fixed effect
model and found that control of corruption has a significant negative effect on NPLs. This
indicates that solid control of corruption would decrease NPLs in Pakistan, India, and
Bangladesh. Abdullah et al. (2024) examine the effect of NPLs on bank lending in six
Association of Southeastern Asian Nations (ASEAN). They find that NPLs and loan Loss
provisions negatively affect bank lending. This indicates that banks are cautious about
increasing lending with high NPLs and loan loss provisions. Nguyen (2024) examined the
effect of NPLs on bank profitability in Vietnam from 2005 to 2020. They find that bank
profitability, bank size, and economic boom reduce NPLs while operating costs, loan loss
provisions, and macro factors worsen bank NPLs. Ozili (2023) investigates the
correlation between banking sector NPLs and the level of sustainable development. The
author finds a significant positive correlation between banking sector NPLs and the level
of sustainable development among European countries and in countries in the region of
the Americas. Generally, the evidence for the determinants of NPLs is mixed in the
literature. However, the existing literature has not examined the effect of the election on
bank NPLs both at the bank level and country level.

3. Research Methodology

3.1. Data
Annual country-level data were collected for 35 developed countries, chosen from
the United Nations list of developed countries.2Developed countries were used in the
study because they have more regular elections than developing countries and transition
economies. Also, developed economies do not experience coups (overthrows of
government) or dictatorships that prevent elections, while all these are arguably more
prevalent in some transition economies and developing countries.
Financial statement annual data for each country were obtained from the World bank
database. The sample period is from 1998 to 2016 and is sufficient to cover at least six
election cycles. Real gross domestic product (GDP) growth rate data were collected from
the World Economic Forum archived in the World Bank database. In contrast, institutional
data were collected from the World Governance Indicators database of the World Bank.
Election data were obtained from public sources such as government websites, Wikipedia,
and other public sources. These public sources are the most reliable source of information
on elections. Finally, see Appendix 1 for the source of data and variable description.

3.2. Model
The baseline model used to estimate the determinants of non-performing loans
during the election year is specified below. It is a modified version of the models used in
previous NPL literature (see Louzis et al., 2012; Beck et al., 2015; Ozili, 2019; Ozili, 2020).
NPL , = 𝛽 + 𝛽 CI , + 𝛽 CAP , + 𝛽 CRISIS , + 𝛽 ΔGDP , + 𝛽 PS , + 𝛽 ELECT +
𝛽 LG , + 𝑒 (1)

where CI = ratio of total cost to total income in each country's banking sector. NPL = ratio
of non-performing loans to gross loans in each country's banking sector. ELECT = a binary
variable that equals one in an election year and zero in a non-election year. CRISIS = a

2 The UN List of developed countries can be found here: https://www.un.org/development/desa/dpad/wp-

content/uploads/sites/45/WESP2019_BOOK-ANNEX-en.pdf
Modern Finance. 2024, 2(2). 68

binary variable that equals one in a banking or financial crisis year and zero otherwise.
∆GDP = real domestic product growth rate of each country. PS = political stability and
absence of terrorism index; the higher, the better. CAP = ratio of bank capital to total assets
(%). LG = credit ratio provided to the private sector by banks as a share of GDP in each
country. t = year. i = country.
The model above expressed non-performing loans as a function of its bank-specific
and external determinants. The models are estimated using the panel fixed effect
regression method based on the Hausman test, which showed that the fixed effect method
is the preferred panel regression method. The choice of the fixed effect regression method
is also because the fixed effect model allows us to control for all time-invariant omitted
variables that are difficult to observe, and it reduces the potential sources of bias in the
estimations in comparison to the random effect model, thereby giving us a less biased
estimator.

3.3. Variable justification

For the explanatory variables, the ELECT variable is the primary variable of interest
because it captures the election year in the developed countries in the sample. A positive
sign for the ELECT coefficient is expected if the banking sector experiences higher non-
performing loans during the election year. This is because business uncertainty during the
election year can amplify banking sector credit risk, possibly giving rise to non-
performing loans. Furthermore, banks that lend to election campaign borrowers may
experience difficulty in compelling such borrowers to pay their debt, which could lead to
loan defaults.
For the control variables, the CAP variable is introduced into the model to control for
the role of bank capital in absorbing unexpected losses that arise from high non-
performing loans. Banks that have a low capital ratio will avoid risky lending to prevent
their capital from being depleted, while banks that have adequate capital can take more
risks and engage in risky lending, which may lead to higher non-performing loans; thus,
a positive relationship between bank capital ratio and non-performing loans is expected.
This expectation is supported by Zhang et al. (2016) and Osei-Assibey and Asenso (2015).
The cost-to-income ratio (CI) variable is introduced into the model to control for banks’
efficiency in managing costs relative to income. This ratio is also considered to reflect the
quality of bank management. Banking sectors with a lower efficiency ratio (i.e., high cost
and low income) are considered inefficient and tend to have higher non-performing loans;
thus, a positive relationship between CI and NPL is expected. This expectation is
supported by Zhu et al. (2015), Karim et al. (2010) and Ozili (2018).
The CRISIS variable is obtained from the World Bank database containing data on
countries that have experienced financial or banking crises. The CRISIS variable is
introduced into the model to capture the years a country experienced financial (or
banking) crises. Ideally, non-performing loans should be higher in a crisis year and lower
in a non-crisis year (Ozili, 2018); thus, a positive relationship between NPL and CRISIS is
expected. In the estimation procedure, the CRISIS variable is a binary variable that equals
one in a banking or financial crisis year and zero otherwise. The real gross domestic
product growth rate (ΔGDP) variable is introduced into the model to control the size of
non-performing loans along the business cycle. Prior studies show that banks tend to have
fewer non-performing loans in economic boom years because the level of employment is
often high in good times which increases the ability of debtors to earn income and repay
their debt, thus reducing the level of loan default. On the other hand, banks tend to have
larger non-performing loans in recession years because the level of unemployment is often
high in bad times; people would lose their source of income, which increases the
probability of loan default and leads to higher non-performing loans (Beck et al., 2013;
Klein, 2013; Ozili, 2020).
The total loan-to-GDP ratio (LG) is introduced into the model to reflect the extent to
which the economy is driven by private credit. The LG ratio measures the amount of credit
Modern Finance. 2024, 2(2). 69

banks provide to the private sector as a share of GDP. Banks in highly credit-driven
economies tend to experience higher non-performing loans than banks in economies that
are less driven by credit; thus, a positive relationship between NPL and LG is expected.
Using the formula below, the LG ratio is derived mathematically by multiplying the
credit-to-deposit ratio data with the bank deposit-to-GDP data in the World Bank
database.
𝐿𝑂𝐴𝑁 / 𝐺𝐷𝑃 = (𝐿𝑂𝐴𝑁 / 𝐵𝐴𝑁𝐾 𝐷𝐸𝑃𝑂𝑆𝐼𝑇) ∗ (𝐵𝐴𝑁𝐾 𝐷𝐸𝑃𝑂𝑆𝐼𝑇 / 𝐺𝐷𝑃) (2)
The PS variable is the political stability (PS) index. The PS variable is introduced into
the model to capture the effect of political stability (or instability) on bank NPLs. Higher
values of PS indicate more excellent political stability. A negative relationship between
NPL and PS is expected because Ghosh (2016) shows that banks in politically unstable
environments tend to perform poorly - they have low profitability and higher non-
performing loans. The LAW variable measures legal system quality or the rule of law. The
LAW variable is introduced into the model to capture the role of the rule of law in
reducing NPLs in the banking sector. A negative relationship between NPL and LAW is
expected because banks operating in countries with strong legal institutions can use the
power of the courts to compel debtors to repay the loans owed to banks, thereby reducing
bank NPLs. Table 1 presents a summary of the expected signs for the variables.

Table 1. Information about the variables

Variable Expected/Predicted Sign Description

NPL Dependent Variable Bank non-performing loans divided by gross loans
ELECT (+) Election year
CAP (+) Bank capital ratio
CI (+) Efficiency ratio
LG (+) Private credit supply to the economy as a percent of GDP
LAW (-) The rule of law/quality of legal system index
PS (-) Political stability and absence of terrorism index
ΔGDP (-) Annual change in real gross domestic product
CRISIS (+) Financial crisis years

4. Discussion of Results

4.1. Descriptive statistics and correlation analysis

Table 2 reports the summary of the descriptive statistics of the variables. The average
bank capital ratio (CAP) is 7.18% of total assets. CAP is higher in the banking sectors of
Iceland, Croatia, and Bulgaria and much lower in Belgium and Japan's banking sectors.
The NPL ratio averages 5.18%, a double-digit ratio in Cyprus, Croatia, and Italy. The high
double-digit NPLs indicate that the banking sector of these countries has low asset quality.
Comparatively, NPLs are much lower in the banking sector of Canada, Luxembourg, and
Australia. The bank efficiency ratio (CI) averages 59.61%. CI is higher in Germany, Greece,
and Switzerland's banking sector and much lower in Ireland and the Netherlands. ΔGDP,
on average, is about 2.42% and is lower in Greece and Italy and is higher in Lithuania and
Latvia. The PS variable is higher in Luxembourg and New Zealand's banking sector and
much lower in Spain and Romania. Overall, the result from the descriptive statistics
suggests that the banking sector and the institutional environment vary across developed
countries.
Modern Finance. 2024, 2(2). 70

Table 2. Descriptive statistics (mean)

Sample Countries CAP NPL ELECT ∆GDP PS LAW CRISIS CI LG

1 Canada 4.64 0.88 0.31 2.35 1.07 11.35 0 59.26 105.55
2 United States 10.50 1.95 0.26 2.28 0.46 10.94 0.35 59.54 51.19
3 Japan 4.67 3.38 0.36 0.71 1.03 10.58 0.28 58.18 116.23
4 France 5.37 4.21 0.15 1.56 0.48 10.52 0.28 69.48 86.88
5 Germany 4.59 3.66 0.26 1.42 0.90 10.47 0.28 77.61 96.48
6 Italy 6.18 10.24 0.21 0.44 0.54 10.94 0.28 66.86 76.84
7 United Kingdom 6.50 2.33 0.21 2.03 0.47 10.35 0.35 61.75 -
8 Norway 6.67 1.21 0.21 1.76 1.29 10.29 0 54.25 88.56
9 Iceland 12.19 4.66 0.31 3.41 1.35 8.64 0.28 49.84 136.36
10 Switzerland 5.46 1.52 0.26 1.94 1.33 9.82 0.28 79.16 153.37
11 Sweden 4.95 1.10 0.26 2.54 1.21 10.35 0.28 62.84 110.81
12 Spain 6.87 3.47 0.31 2.05 0.02 10.52 0.28 56.13 127.55
13 Portugal 6.44 5.51 0.21 0.92 0.98 10.23 0.28 63.53 128.28
14 Netherland 4.48 2.51 0.31 1.77 1.09 10.17 0.28 64.28 114.77
15 Luxembourg 5.27 0.45 0.21 3.51 1.42 7.64 0.28 48.41 83.93
16 Ireland 6.98 7.88 0.11 5.37 1.15 9.82 0.28 36.41 104.95
17 Greece 7.46 14.83 0.26 0.50 0.23 10.52 0.28 74.62 79.34
18 Finland 6.52 0.54 0.16 1.90 1.45 10.06 0 51.20 72.87
19 Denmark 6.03 2.36 0.26 1.35 1.10 10.06 0.28 61.83 152.57
20 Belgium 4.35 2.86 0.26 1.69 0.83 10.17 0.28 63.76 61.51
21 Austria 6.14 2.65 0.21 1.76 1.16 10.17 0.28 61.81 90.91
22 Australia 5.80 0.97 0.36 3.19 0.98 10.05 0 51.36 105.79
23 New Zealand 6.52 1.15 0.31 2.81 1.33 10.23 0 59.88 122.01
24 Slovenia 8.50 7.21 0.16 2.37 1.03 13.05 0.28 57.63 54.62
25 Romania 8.20 14.76 0.21 3.32 0.18 14.52 56.13 22.80
26 Poland 8.16 8.89 0.21 3.75 0.71 14.41 0 62.24 37.27
27 Malta 8.09 7.32 0.21 3.71 1.26 6.88 0 49.04 101.74
28 Lithuania 10.24 9.21 0.32 4.07 0.75 12.94 0 65.08 34.13
29 Latvia 8.31 5.27 0.32 3.96 0.53 12.52 0.285 56.85 49.82
30 Hungary 8.60 6.98 0.21 2.34 0.84 13.94 0.28 67.83 43.94
31 Estonia 10.79 1.66 0.21 3.76 0.70 13.35 0 53.75 61.34
32 Cyprus 7.34 24.69 0.21 2.48 0.49 8.47 0 58.32 174.58
33 Czech Republic 6.15 8.22 0.21 2.55 0.93 13.70 0.21 54.36 42.94
34 Croatia 12.51 10.05 0.32 1.74 0.55 13.47 0.14 57.11 53.48
35 Bulgaria 11.37 9.74 0.21 3.19 0.27 14.11 0 54.30 41.68
Total mean 7.18 5.18 0.25 2.42 0.86 11.01 0.19 59.61 87.14
Total median 6.50 2.81 0.00 2.53 0.93 11.00 0.00 58.89 84.25
Observations 601 606 665 665 595 595 476 594 616
Source: Author’s calculations.

Table 3 reports the Pearson correlation coefficients and the associated t-statistics and
p-values. The NPL coefficient is positive and weakly correlated with the ELECT variable
(0.008). However, the correlation between NPL and ELECT is not significant. NPL is
negative and significantly correlated with ΔGDP (-0.188***), indicating that bank NPLs
are procyclical with fluctuations in the business cycle. The CRISIS coefficient is positive
and significantly correlated with NPLs. This indicates that non-performing loans are
higher in a financial crisis year. The PS coefficient is negatively correlated with NPLs. This
indicates that banks in politically stable environments have fewer non-performing loans.
The CI coefficient is positively correlated with NPLs. This indicates that non-performing
loans are higher in inefficient banking sectors. The LG coefficient is also negatively
correlated with NPLs. This indicates that non-performing loans are lower in highly credit-
driven economies. Overall, the correlation of the variables is sufficiently low to be
Modern Finance. 2024, 2(2). 71

concerned about multi-collinearity. Table 3 reports the Pearson correlation coefficients for
the country variables.

Table 3. Correlation Matrix

Variables CI ∆GDP LG CRISIS CAP PS LAW ELECT NPL

CI 1.000
-----
∆GDP 0.001 1.000
(0.01) -----
LG -0.155*** -0.370*** 1.000
(-2.92) (-7.39) -----
CRISIS 0.117** -0.423*** 0.290*** 1.000
(2.18) (-8.67) (5.63) -----
CAP -0.079 0.104** -0.360*** -0.048 1.000
(-1.48) (1.94) (-7.17) (-0.90) -----
PS -0.120** 0.079 0.181*** -0.161*** -0.265*** 1.000
(-2.24) (1.48) (3.41) (-3.03) (-5.10) -----
LAW -0.069 -0.049 -0.269*** 0.179*** 0.253*** -0.383*** 1.000
(-1.29) (-0.92) (-5.19) (3.38) (4.86) (-7.70) -----
ELECT -0.009 0.034 0.009 -0.005 -0.021 -0.055 0.012 1.000
(-0.17) (0.64) (0.17) (-0.11) (-0.38) (-1.03) (0.22) -----
NPL 0.103* -0.188*** -0.147*** 0.206*** 0.341*** -0.352*** 0.280*** 0.008 1.000
(1.93) (-3.56) (-2.75) (3.90) (6.73) (-6.98) (5.42) (0.14) -----
t-statistics are reported in parenthesis. ***, **, * indicate statistical significance at the 1%, 5% and 10%
levels. Source: Author’s calculations.

4.2. Regression results: Election year as a determinant of bank non-performing loans

The estimation result is reported in Table 4. The ELECT coefficient is positive and
insignificant in all three estimations, implying that the election year did not significantly
affect bank non-performing loans in developed economies. The CI coefficient is positively
significant for the control variables in columns (2) and (3). This suggests that an inefficient
banking sector has higher non-performing loans, which is consistent with the findings of
Zhu et al. (2015), Karim et al. (2010), and Ozili (2018). The CAP coefficient is positively
significant, as expected. This indicates that banks that have high capital ratios experience
higher non-performing loans. This result is consistent with Zhang et al (2016) and Osei-
Assibey and Asenso (2015). The ∆GDP and PS coefficients report the expected sign in the
three estimations. The result indicates that non-performing loans are higher in
recessionary years and politically unstable environments. The CRISIS coefficient is
positively significant, as expected in columns 1 and 3, and is consistent with the apriori
expectation that banks experience higher non-performing loans during financial (or
banking) crisis years. As expected, the LG coefficient is positively significant in columns
1 and 2, indicating that banks in highly credit-driven economies have higher
nonperforming loans.
Modern Finance. 2024, 2(2). 72

Table 4. Determinants of bank non-performing loans

(1) (2) (3)

Panel Fixed effect Panel GMM Panel OLS
Variables Coefficient Coefficient Coefficient
(t-statistic) (t-statistic) (t-statistic)
C 2.955* - -
(1.69)
NPLt-1 0.009***
(66.89)
ELECT 0.015 -0.050 0.188
(0.05) (-0.33) (0.41)
CI 0.012 0.026** 0.039***
(0.79) (2.58) (3.86)
CAP 0.377*** 0.233*** 0.546***
(3.70) (5.47) (8.48)
CRISIS 1.101* -3.152* 1.295**
(1.87) (-1.79) (2.35)
∆GDP -0.375*** -0.407*** -0.206***
(-5.04) (-6.39) (-3.12)
PS -4.217*** -2.839*** -1.776***
(-4.89) (-4.04) (-3.88)
LG 0.024*** 0.059*** -0.005
(2.70) (3.64) (-1.04)
R2 67.91 24.81
Adjusted R2 62.47 23.49
F-statistic 12.48
Prob (F-statistic) 0.000
J-statistic 11.65
Prob(J-statistics) 0.705
No of Observation (adjusted) 346 282 346
Model (1) is estimated using panel fixed effect regression (with country and year fixed effect). Model
(2) is estimated using GMM first-difference regression (this includes difference fixed effect and
period effect using the ordinary coefficient covariance method). Model (3) is estimated using
ordinary OLS with no fixed effect. NPL = ratio of non-performing loans to gross loans: the lower,
the better. ELECT = a binary variable that equals one in an election year and zero in a non-election
year. CRISIS = a binary variable that equals one in a financial or banking crisis year and zero
otherwise. ∆GDP = real domestic product growth rate; CI = cost to income ratio. CAP = ratio of total
capital to total asset. PS = political stability/absence of terrorism index: the higher, the better. LG =
credit ratio provided to the private sector by banks as a share of GDP. NPLt-1 = one-year lag of the
non-performing loans ratio. ***, **, * denote 1%, 5% and 10% significance levels. Source: Author

4.3. Interaction analyses

4.3.1. Interaction analysis of NPL determinants during election year
This section tests the effect of the election year on each determinant of non-
performing loans. The ELECT variable interacted with each NPL determinant to do this,
as shown in Equation 2 below.
𝐿 , = 𝛽 + 𝛽 CI , + 𝛽 CAP , + 𝛽 CRISIS , + 𝛽 ΔGDP , + 𝛽 PS , + 𝛽 ELECT +
𝛽 LG , + 𝛽 ELECT × CAP , + 𝛽 ELECT × CRISIS , + 𝛽 ELECT × ΔGDP , + 𝛽 ELECT ×
PS , + 𝛽 ELECT × LG , + 𝑒 (3)
The result is reported in Table 5. The ELECT*CI coefficient is negatively significant,
indicating that inefficient banking sectors have fewer non-performing loans during an
election year. Meanwhile, the ELECT*CAP, ELECT*LG, ELECT*GDP, ELECT*PS, and
ELECT*CRISIS coefficients are all statistically insignificant.
Modern Finance. 2024, 2(2). 73

Table 5. Interaction analysis of NPL determinants during an election year (Panel fixed effect
regression estimation)

(1) (2) (3) (4) (5) (6) (7) (8)

Variables Coefficient Coefficient Coefficient Coefficient Coefficient Coefficient Coefficient Coefficient
(t-statistic) (t-statistic) (t-statistic) (t-statistic) (t-statistic) (t-statistic) (t-statistic) (t-statistic)
C 2.955* 2.931* 2.425 3.035* 3.174* 2.856 2.967* 2.635
(1.69) (1.67) (1.39) (1.73) (1.79) (1.61) (1.69) (1.44)
ELECT 0.015 -0.054 -4.031** -0.777 -0.633 0.246 -0.008 2.164
(0.05) (-0.13) (-2.59) (-0.81) (-0.85) (0.32) (-0.02) (0.76)
CI 0.012 0.012 0.024 0.013* 0.012 0.013 0.012 0.024
(0.79) (0.78) (1.51) (1.92) (0.79) (0.80) (0.79) (1.46)
CAP 0.377*** 0.379*** 0.364*** 0.353*** 0.373*** 0.379*** 0.378*** 0.348***
(3.71) (3.71) (3.59) (3.35) (3.66) (3.71) (3.70) (3.24)
CRISIS 1.101* 1.105* 1.052* 1.133* 1.058* 1.106* 1.074* 1.022
(1.87) (1.87) (1.80) (1.92) (1.79) (1.87) (1.72) (1.63)
∆GDP -0.375*** -0.382*** -0.387*** -0.377*** -0.371*** -0.374*** -0.374*** -0.403***
(-5.04) (-4.83) (-5.25) (-5.07) (-4.99) (-5.03) (-5.03) (-5.06)
PS -4.216*** -4.195*** -4.294*** -4.200*** -4.271*** -4.147*** -4.220 -4.183***
(-4.89) (-4.84) (-5.03) (-4.87) (-4.94) (-4.66) (-4.88) (-4.70)
LG 0.024*** 0.025*** 0.025*** 0.024*** 0.023** 0.024*** 0.024*** 0.023**
(2.71) (2.72) (2.78) (2.74) (2.47) (2.71) (2.69) (2.55)
ELECT*∆GDP 0.027 0.081
(0.26) (0.67)
ELECT*CI -0.069*** -0.061***
(-2.64) (-2.17)
ELECT*CAP 0.117 0.093
(0.88) (0.62)
ELECT*LG 0.007 0.009
(0.98) (0.99)
ELECT*PS -0.262 -0.374
(-0.33) (-0.44)
ELECT*CRISIS 0.106 0.096
(0.13) (0.10)
R2 67.91 67.92 68.64 67.99 68.01 67.91 67.91 68.80
Adjusted R2 62.47 62.35 63.21 62.44 62.46 62.35 62.34 62.76
F-statistic 12.48 12.20 12.62 12.25 12.26 12.20 12.19 11.38
Prob (F- 0.000 0.000 0.0000 0.000 0.000 0.000 0.000 0.000
statistic)
No of 346 346 346 346 346 346 346 346
Observations
The models in Table 5 are estimated using panel fixed effect regression (with country and year fixed
effect). GMM estimations could not be run due to the near-perfect collinearity of the lag variables
with the instrumental variables. NPL = ratio of non-performing loans to gross loans: the lower, the
better. ELECT = a binary variable that equals one in an election year and zero in a non-election year.
CI = cost to income ratio. CAP = ratio of total capital to total asset. CRISIS = a binary variable that
equals one in a financial, banking, or economic crisis year and zero otherwise. ∆GDP = real domestic
product growth rate. PS = political stability/absence of terrorism index: the higher, the better. LG =
credit ratio provided to the private sector by banks as a share of GDP. ***, **, * denote 1%, 5% and
10% significance levels. Source: Author’s calculations.

4.3.2. Effect of bloc membership on non-performing loans

In this section, I test whether bloc membership positively or negatively impacts non-
performing loans in the election year. Some regional political blocs adopt a uniform
mechanism to deal with non-performing loan problems in the banking sector of member
Modern Finance. 2024, 2(2). 74

countries. I consider the case of member countries in the European Union, the G7 (also
known as the most advanced economies), and the member countries of the Bank of
International Settlement. 3 (BIS). I estimate the differences in banking sector non-
performing loans during an election year for EU, non-EU, BIS, and G7 countries. Binary
variables were used to capture bloc membership in the European Union (EU), non-
European Union (non-EU), BIS member countries (BIS), and the G7 countries (MDE). The
‘EU’ binary variable is assigned a value of one if the developed country is a member of
the European Union and zero otherwise. The ‘non-EU’ binary variable is assigned a value
of one if the developed country is not a member of the European Union and zero
otherwise. The ‘MDE’ binary variable is assigned a value of one if the developed country
is a member of the G7 countries and zero otherwise. The ‘BIS’ binary variable is assigned
a value of one if the developed country is a Bank of International Settlement member and
zero otherwise. The estimation result in Table 5 is based on the model below.
NPL , = 𝛽 + 𝛽 ELECT + 𝛽 CI , + 𝛽 CAP , + 𝛽 CRISIS , + 𝛽 ΔGDP , + 𝛽 PS , +
𝛽 LG , + 𝛽 EU + 𝛽 Non-EU + 𝛽 MDE + 𝛽 BIS + 𝛽 ELECT × EU + 𝛽 ELECT ×
Non-EU + 𝛽 ELECT × MDE + 𝛽 ELECT × BIS + 𝑒 (4)
The estimation result is reported in Table 6. The EU coefficient is positively
significant at the 1% level in column 1, which suggests that the EU banking sector
generally experiences higher non-performing loans. On the other hand, the ‘non-EU’ and
‘BIS’ coefficients are negatively significant, which suggests that the non-EU banking sector
and the banking sector of the BIS member countries generally have fewer non-performing
loans. Furthermore, the ELECT*EU, ELECT*Non-EU, ELECT*MDE, and ELECT*BIS
coefficients are not statistically significant. This suggests no significant differences in non-
performing loans across the four blocs during an election year.
4.3.3. Effect of legal system quality on non-performing loans
This section tests the effect of legal system quality on each determinant of non-
performing loans. Developed countries have better legal systems than developing
countries. Banks in developed countries can use the power of the courts to compel debtors
to repay their debt, reducing the level of non-performing loans. In the estimation, the
ELECT and LAW variables interact with each NPL determinant. The result is reported in
Table 7.
As can be observed, the LAW*∆GDP coefficient is insignificant. Meanwhile, the
LAW*CAP, LAW*LG, LAW*CI, LAW*PS, and LAW*CRISIS coefficients are significant,
indicating that legal system quality significantly affects some NPL determinants.
Similarly, the LAW*ELECT*CI coefficient is negatively significant, indicating that
inefficient banks (i.e., banks with high cost relative to income) operating in robust legal
environments have fewer non-performing loans during an election year. Meanwhile, the
LAW*ELECT*CAP, LAW*ELECT*LG, LAW*ELECT*GDP, LAW*ELECT*PS, and
LAW*ELECT*CRISIS coefficients are not statistically significant.

5. Conclusion
This study examined the determinants of bank non-performing loans during election
year. The findings revealed that non-performing loans are higher in the banking sector of
EU member countries, while NPLs are significantly lower in the banking sector of non-
EU and BIS member countries. Regarding the election year effect, the findings reveal that
efficient banking sectors have higher non-performing loans during election year. It was
also found that efficient banks operating in robust legal environments have higher non-
performing loans during election years. The main message of this paper is that the

3 The mission of the BIS is to promote monetary and financial stability in the financial system of member countries. These countries
cooperate to deal with the non-performing loan problems in member countries, thereby promoting banking stability in member
countries. BIS member countries are represented by their central bank.
Modern Finance. 2024, 2(2). 75

‘election year’ is a significant country risk factor that banks should consider in their risk
management decisions because uncertainty about election outcomes may affect the loan
portfolio of banks, mainly when election campaigns are funded with loans from the
banking sector. Such loans may not be repaid when due, leading to the bank’s loan
portfolio impairment.

Table 6. Bank non-performing loan determinants during election year in EU, non-EU, BIS, and G7
countries (Panel fixed effect regression estimation)
(1) (2) (3) (4) (5) (6) (7) (8)
Variables Coefficient Coefficient Coefficient Coefficient Coefficient Coefficient Coefficient Coefficient
(t-statistic) (t-statistic) (t-statistic) (t-statistic) (t-statistic) (t-statistic) (t-statistic) (t-statistic)
C 2.612* 4.605*** 5.137*** 9.023*** -2.672* 4.601*** 5.155*** 9.752***
(1.69) (3.18) (3.42) (4.71) (-1.69) (3.17) (3.42) (4.46)
ELECT 0.033 0.033 -0.061 -0.009 -0.145 0.077 -0.108 1.739
(0.08) (0.08) (-0.14) (-0.02) (-0.15) (0.16) (-0.22) (0.63)
CI 0.016 0.016 0.015 0.012 0.016 0.016 0.015 0.012
(1.11) (1.11) (0.98) (0.84) (1.10) (1.10) (0.99) (0.84)
CAP 0.358*** 0.358*** 0.344*** 0.327*** 0.357*** 0.357*** 0.343*** 0.329***
(4.51) (4.51) (4.15) (4.06) (4.51) (4.51) (4.15) (4.08)
CRISIS 0.259 0.259 0.525 1.056* 0.258 0.258 0.529 1.073*
(0.42) (0.42) (0.84) (1.66) (0.42) (0.42) (0.85) (1.68)
∆GDP -0.356*** -0.356*** -0.373*** -0.358*** -0.357*** -0.357*** -0.373*** -0.364***
(-4.49) (-4.49) (-4.50) (-4.48) (-4.48) (-4.48) (-4.48) (-4.52)
PS -2.548*** -2.548*** -2.706*** -2.560*** -2.544*** -2.544*** -2.715*** -2.550***
(-5.36) (-5.36) (5.47) (-5.35) (-5.34) (-5.34) (-5.46) (-5.32)
LG -0.007 -0.007 -0.012*** -0.017*** -0.007 -0.007 -0.012** -0.016***
(-1.36) (-1.36) (-2.23) (-3.08) (-1.37) (-1.37) (-2.23) (-3.11)
EU 1.994*** 1.929***
(4.01) (3.28)
Non-EU -1.994*** -1.929***
(-4.01) (-3.28)
MDE -0.794 -0.855
(-1.55) (-1.44)
BIS -3.783*** -3.496***
(-3.45) (-2.95)
ELECT*EU 0.222
(0.21)
ELECT*Non-EU -0.222
(-0.21)
ELECT*MDE 0.220
(0.21)
ELECT*BIS -1.793
(-0.64)
R2 39.60 39.60 37.08 38.86 39.61 39.61 37.09 38.93
Adjusted R2 36.08 36.08 33.41 35.29 35.89 35.89 33.1 35.18
F-statistic 11.25 11.25 10.11 10.90 10.66 10.66 9.580 10.36
Prob (F-statistic) 0.000 0.000 0.0000 0.000 0.000 0.000 0.000 0.000
No of Observations 346 346 346 346 346 346 346 346
The models in Table 6 are estimated using fixed effect regression (with country and year fixed
effect). GMM estimations could not be run due to near-perfect collinearity of the lag variables with
the instrumental variables. NPL = ratio of non-performing loans to gross loans: the lower, the better.
ELECT = a binary variable that equals one in an election year and zero in a non-election year. CI =
cost to income ratio. CAP = ratio of total capital to total asset. CRISIS = a binary variable that equals
one in a financial, banking, or economic crisis year and zero otherwise. ∆GDP = real domestic
product growth rate. PS = political stability/absence of terrorism index: the higher, the better. LG =
credit ratio provided to the private sector by banks as a share of GDP. EU = a binary variable that
equals one if the country is a member of the European Union and zero otherwise. Non-EU = a binary
variable that equals one if the country is not a member of the European Union and zero otherwise.
MDE = a binary variable that equals one if the country is a member of the G7 countries according to
Modern Finance. 2024, 2(2). 76

the UN’s classification. ***, **, * denote 1%, 5% and 10% significance levels. Source: Author’s
calculations

Table 7. Effect of legal system quality on the NPL determinants during election year
(1) (2) (3) (4) (5) (6) (7) (8)
Variables Coefficient Coefficient Coefficient Coefficient Coefficient Coefficient Coefficient Coefficient
(t-statistic) (t-statistic) (t-statistic) (t-statistic) (t-statistic) (t-statistic) (t-statistic) (t-statistic)
C 2.955* -0.812 4.531 8.288** 2.888 -3.979 -0.198 25.151
(1.69) (-0.25) (0.94) (2.03) (0.83) (-1.18) (-0.06) (3.53)
ELECT 0.046 -0.026 3.114** -0.344 -0.761 0.553 0.014 1.450
(0.14) (-0.06) (2.47) (-0.46) (-1.01) (0.68) (0.04) (0.86)
CI 0.014 0.014 -0.089 0.017 0.011 0.019 0.015 -0.142**
(0.91) (0.88) (-1.28) (1.09) (0.72) (1.19) (0.94) (-2.02)
CAP 0.407*** 0.411*** 0.404*** -0.550* 0.405*** 0.452*** 0.442*** -0.757**
(3.91) (3.91) (3.93) (-1.78) (3.92) (4.31) (4.23) (-2.29)
CRISIS 1.069* 1.095* 0.639* 1.497** 1.096* 0.797 -4.824* -3.601
(1.81) (1.82) (1.02) (2.52) (1.87) (1.34) (-1.80) (-1.19)
∆GDP -0.369*** -0.319 -0.378*** -0.361*** -0.339*** -0.332*** -0.349*** -0.739**
(-4.96) (-1.17) (-5.13) (-4.93) (-4.51) (-4.40) (-4.69) (-2.53)
PS -4.597*** -4.580*** -4.659*** -4.863*** -4.101*** 0.563 -4.436*** -2.539
(-5.09) (-5.04) (-5.21) (-5.44) (-4.38) (0.24) (-4.91) (-0.99)
LG 0.025*** 0.025*** 0.027*** 0.022** -0.035 0.025*** 0.029*** -0.076**
(2.80) (2.81) (3.04) (2.52) (-1.15) (2.76) (3.15) (-2.36)
LAW*∆GDP -0.005 0.032
(-0.23) (1.37)
LAW*CI 0.012* 0.017**
(1.67) (2.47)
LAW*CAP 0.096*** 0.119***
(3.25) (3.74)
LAW*LG 0.006** 0.011***
(1.98) (3.30)
LAW*PS -0.499*** -0.129
(-2.30) (-0.54)
LAW*CRISIS 0.511** 0.374
(2.23) (1.44)
LAW 0.324 0.340 -0.263 -0.546 -0.077 0.585** 0.216 -2.475***
(1.38) (1.37) (-0.58) (-1.54) (-0.25) (2.27) (0.90) (-3.55)
LAW*ELECT*∆GDP 0.003 0.013
(0.33) (1.44)
LAW*ELECT*CI -0.008** -0.006***
(-2.52) (-2.78)
LAW*ELECT*CAP 0.005 0.015
(0.61) (1.59)
LAW*ELECT*LG 0.001 0.001
(1.19) (1.13)
LAW*ELECT*PS -0.053 -0.029
(-0.67) (-0.37)
LAW*ELECT*CRISIS 0.032 0.067
(0.46) (0.82)
R2 68.11 68.13 69.03 69.28 68.70 68,76 68.67 72.34
Adjusted R2 62.58 62.34 63.40 63.71 63.02 63.09 62.98 66.15
F-statistic 12.31 11.78 12.28 12.43 12.09 12.13 12.07 11.71
Prob (F-statistic) 0.000 0.000 0.0000 0.000 0.000 0.000 0.000 0.000
No of Observations 346 346 346 346 346 346 346 346
The estimations in Table 7 are estimated using fixed effect regression (with country and year fixed
effect). GMM estimations could not be run due to near-perfect collinearity of the lag variables with
the instrumental variables. NPL = ratio of non-performing loans to gross loans: the lower, the better.
ELECT = a binary variable that equals one in an election year and zero in a non-election year. CI =
cost to income ratio. CAP = ratio of total bank capital to total asset. CRISIS = a binary variable that
Modern Finance. 2024, 2(2). 77

equals one in a financial or banking year and zero otherwise. ∆GDP = real domestic product growth
rate. PS = political stability/absence of terrorism index: the higher, the better; LAW = the rule of law
index; the higher, the better; LG = ratio of credit provided to the private sector by banks as a share
of GDP. ***, **, * denote 1%, 5% and 10% significance levels. Source: Author’s calculations.

The findings have several implications. One implication of the findings is that
political events, such as elections, can significantly affect banks' lending decisions. Two,
bank supervisors should consider the ‘election-year effect’ in assessing industry-wide
non-performing loans in the banking sector. One idea is to assess the growth of NPLs in
the years before the election, during the election, and in the years after the election. Such
assessment can help bank supervisors understand whether high non-performing loans in
an election year are caused by loans granted to politicians for election campaigns, which
banks write off, or whether the NPLs are caused by general business uncertainty
surrounding an election event. Thirdly, legal solid institutions can help to reduce the
negative effect of elections on non-performing loans; therefore, policy makers should
develop policies that strengthen existing legal institutions with appropriate enforcement
powers.
Future research could investigate other cross-country macro-events and political
events that influence the level of non-performing loans in the banking sector of developed
countries. Secondly, future research can investigate the effect of election on bank non-
performing loans in developing and transition countries. Future studies should consider
that most developing and transition countries may have irregular elections or experience
political events (such as military takeovers) that may prevent elections.

Funding: This research received no external funding.

Data Availability Statement: The data is available in the World Bank database. Data can be found
at https://databank.worldbank.org/.

Appendix

Table A.1. Variable Description

Indicator Name Short definition Source

CI Bank cost-to-income ratio (%) Financial Soundness Indicators Database (fsi.imf.org),
International Monetary Fund (IMF)
∆GDP Annual gross domestic product growth rate World Bank national accounts data, and OECD National
Accounts data files.
LG Credit to the private sector by banks as a share of Bankscope, Bureau van Dijk (BvD)
GDP.
CAP The ratio of total bank capital to total asset
NPL Bank non-performing loans to gross loans (%) Financial Soundness Indicators Database (fsi.imf.org),
International Monetary Fund (IMF)
PS Political stability and absence of terrorism index World Governance Indicator in the World Bank database
LAW The rule of law/quality of legal system index World Governance Indicator in the World Bank database
ELECT Election year variable Constructed by author
CRISIS The financial crisis indicator Global financial development indicator in the World Bank
database

References
Abdullah, W., Rahmati, S., Mohammad, A., Al-Hajj, N., & Khalaf, B. A. (2024). Do Non-performing Loans Affect Bank Lending Behavior?
In The AI Revolution: Driving Business Innovation and Research: Volume 2 (pp. 919-933). Cham: Springer Nature Switzerland.
https://doi.org/10.1007/978-3-031-54383-8_70
Ahamed, M. M. (2017). Asset quality, non-interest income, and bank profitability: Evidence from Indian banks. Economic Modelling,
63, 1-14. https://doi.org/10.1016/j.econmod.2017.01.016
Modern Finance. 2024, 2(2). 78

Ahiase, G., Andriana, D., Agbemava, E., & Adonai, B. (2024). Macroeconomic cyclical indicators and bank non-performing loans:
Does country governance matter in African countries? International Journal of Social Economics, 51(1), 62-80.
https://doi.org/10.1108/IJSE-11-2022-0729
Al-Khazali, O. M., & Mirzaei, A. (2017). The impact of oil price movements on bank non-performing loans: Global evidence from oil-
exporting countries. Emerging Markets Review, 31, 193-208. https://doi.org/10.1016/j.ememar.2017.05.006
Amore, M. D., & Corina, M. (2021). Political elections and corporate investment: International evidence. Journal of International
Business Studies, 52(9), 1775-1796. https://doi.org/10.1057/s41267-021-00421-6
Bardhan, S., & Mukherjee, V. (2016). Bank-specific determinants of nonperforming assets of Indian banks. International Economics and
Economic Policy, 13(3), 483–498. https://doi.org/10.1007/s10368-016-0344-4
Barseghyan, L. (2010). Non-performing loans, prospective bailouts, and Japan's slowdown. Journal of Monetary Economics, 57(7), 873-
890. https://doi.org/10.1016/j.jmoneco.2010.08.002
Bashir, U., Yu, Y., Hussain, M., Wang, X., & Ali, A. (2017). Do banking system transparency and competition affect nonperforming
loans in the Chinese banking sector? Applied Economics Letters, 24(21), 1519–1525. https://doi.org/10.1080/13504851.2017.1305082
Beck, R., Jakubik, P., & Piloiu, A. (2013). Non-performing loans: What matters in addition to the economic cycle? ECB Working Paper
No. 1515. https://doi.org/10.2139/ssrn.2214971
Beck, R., Jakubik, P., & Piloiu, A. (2015). Key determinants of non-performing loans: new evidence from a global sample. Open
Economies Review, 26(3), 525-550. https://doi.org/10.1007/s11079-015-9358-8
Clarke, L. (2013). Doing business for democracy: the impact of corporate activity on the Kenyan elections. Africa Conflict Monthly
Monitor, 2013(03), 40–43.
Çolak, G., Durnev, A., & Qian, Y. (2017). Political uncertainty and IPO activity: Evidence from US gubernatorial elections. Journal of
Financial and Quantitative Analysis, 52(6), 2523-2564. https://doi.org/10.1017/S0022109017000862
Cucinelli, D. (2015). The impact of non-performing loans on bank lending behavior: evidence from the Italian banking sector. Eurasian
Journal of Business and Economics, 8(16), 59-71. https://doi.org/10.17015/ejbe.2015.016.04
Fofack, H. L. (2005). Nonperforming loans in Sub-Saharan Africa: causal analysis and macroeconomic implications. The World Bank.
Washington D.C. https://doi.org/10.1596/1813-9450-3769
Ghosh, A. (2015). Banking-industry specific and regional economic determinants of non-performing loans: Evidence from US states.
Journal of Financial Stability, 20, 93–104. https://doi.org/10.1016/j.jfs.2015.08.004
Karim, M. Z. A., Chan, S. G., & Hassan, S. (2010). Bank efficiency and non-performing loans: Evidence from Malaysia and Singapore.
Prague Economic Papers, 2(1), 118–132. https://doi.org/10.18267/j.pep.367
Kasinger, J., Krahnen, J. P., Ongena, S., Pelizzon, L., Schmeling, M., & Wahrenburg, M. (2021). Non-performing loans-new risks and
policies? NPL resolution after COVID-19: Main differences to previous crises (No. 84). SAFE White Paper.
Kauko, K. (2012). External deficits and non-performing loans in the recent financial crisis. Economics Letters, 115(2), 196–199.
https://doi.org/10.1016/j.econlet.2011.12.018
Kayser, M. A. (2006). Trade and the Timing of Elections. British Journal of Political Science, 36(3), 437-457.
https://doi.org/10.1017/S0007123406000238
Klein, N. (2013). Non-performing loans in CESEE: Determinants and impact on macroeconomic performance. International Monetary
Fund. IMF Working Paper, No. 13/72 https://doi.org/10.5089/9781484318522.001
Laryea, E., Ntow-Gyamfi, M., & Alu, A. A. (2016). Nonperforming loans and bank profitability: Evidence from an emerging market.
African Journal of Economic and Management Studies, 7(4), 462–481. https://doi.org/10.1108/AJEMS-07-2015-0088
Louzis, D. P., Vouldis, A. T., & Metaxas, V. L. (2012). Macroeconomic and bank-specific determinants of non-performing loans in
Greece: A comparative study of mortgage, business and consumer loan portfolios. Journal of Banking & Finance, 36(4), 1012-1027.
https://doi.org/10.1016/j.jbankfin.2011.10.012
Mamoon, A., Kwabi, F., Ezeani, E., & Hu, W. (2024). The impact of central bank independence and transparency on banks' non-
performing loans and economic stability. Journal of Banking Regulation, 1-16. https://doi.org/10.1057/s41261-024-00237-y
Nguyen, P. D. (2024). Non-performing loans and bank profitability: evidence from Vietnam. Macroeconomics and Finance in Emerging
Market Economies, 1-21. https://doi.org/10.1080/17520843.2024.2318927
Nkusu, M. M. (2011). Nonperforming Loans and Macrofinancial Vulnerabilities in Advanced Economies. IMF Working Paper No
11/161. https://doi.org/10.5089/9781455297740.001
Osei-Assibey, E., & Asenso, J. K. (2015). Regulatory capital and its effect on credit growth, non-performing loans, and bank efficiency:
Evidence from Ghana. Journal of Financial Economic Policy, 7(4), 401-420. https://doi.org/10.1108/JFEP-03-2015-0018
Ozili, P. K. (2018). Banking stability determinants in Africa. International Journal of Managerial Finance, 14(4), 462–483.
https://doi.org/10.1108/IJMF-01-2018-0007
Ozili, P. K. (2019). Non-performing loans and financial development: new evidence. The Journal of Risk Finance, 20(1), 59–81.
https://doi.org/10.1108/JRF-07-2017-0112
Ozili, P. K. (2020). Non-performing loans in European systemic and non-systemic banks. Journal of Financial Economic Policy, 12(3),
409–424. https://doi.org/10.1108/JFEP-02-2019-0033
Ozili, P.K. (2023). Sustainable development and bank non-performing loans: Are they correlated? Arab Gulf Journal of Scientific
Research. https://doi.org/10.1108/AGJSR-01-2023-0028
Podpiera, J., & Ötker, M. I. (2010). The fundamental determinants of credit default risk for European large complex financial
institutions. International Monetary Fund, IMF Working Paper 10/153. https://doi.org/10.5089/9781455201365.001
Modern Finance. 2024, 2(2). 79

Rehman, A., Mehmood, W., Alsmady, A. A., & Sharif, A. (2024). Corruption's impact on non-performing loans of banks in emerging
markets: Empirical insights. Research in Globalization, 9, 100241. https://doi.org/10.1016/j.resglo.2024.100241
Vithessonthi, C. (2016). Deflation, bank credit growth, and non-performing loans: Evidence from Japan. International Review of
Financial Analysis, 45, 295-305. https://doi.org/10.1016/j.irfa.2016.04.003
Zhang, D., Cai, J., Dickinson, D. G., & Kutan, A. M. (2016). Non-performing loans, moral hazard and regulation of the Chinese
commercial banking system. Journal of Banking & Finance, 63, 48–60. https://doi.org/10.1016/j.jbankfin.2015.11.010
Zhu, N., Wang, B., & Wu, Y. (2015). Productivity, efficiency, and non-performing loans in the Chinese banking industry. The Social
Science Journal, 52(4), 468-480. https://doi.org/10.1016/j.soscij.2014.10.003

Disclaimer: All statements, viewpoints, and data featured in the publications are exclusively those of the individual author(s) and
contributor(s), not of MFI and/or its editor(s). MFI and/or the editor(s) absolve themselves of any liability for harm to individuals or
property that might arise from any concepts, methods, instructions, or products mentioned in the content.