The Impact of Credit Risk on the Profitability of European
Commercial Banks:
Panel data evidence from Europe in the period 2000-2018
By
Natalia Rapti
Student number: S3155927
Email: n.rapti@student.rug.nl
University of Groningen
Faculty of Economics and Business Master’s program in Finance
1. Introduction
performance and structure of banks (Petria, Capraru, and Ihnatov, 2015). Taking into account both the crises and the major transformations the European economy has encounter, the banking industry’s environment has changed rapidly. This, in turn, creates great challenges to the banks operating within the EU and consequently affects their performance. We are interested in investigating European commercial banks’ performance, considering that banks’ profitability remains one of the main challenges banks faces due to its ability to amplify the banks' financial positions against risks related to globalization and transparency (Rekik and Kalai, 2018). The more profitable the banking sector is the more capable and stronger will be in resisting negative shocks and the higher its ability to assist in the financial systems’ stability (Rekik and Kalai, 2018). In the same line, Athanasoglou, Brissimis, and Delis (2008), and Garcia-Herrero, Gavila, and Santabarbara (2009), also support that the key to preserving stability within the financial system depends on the profitability of the banking sector which becomes extremely important during periods of negative shocks.
After the post-crisis period, commercial banks’ most vital risks became credit risk and liquidity risk (Anastasiou, Louri, and Tsionas, 2019). This is due to the tremendous effects credit risk caused during the crisis which was mainly driven by banks’ excessive lending to subprime lenders (Hull, 2015). When borrowers started defaulting on their loans several banks foreclosed and others faced liquidity issues (Hull, 2015). This, in turn, affected the financial systems throughout the world (Al-shakrchy, 2017), given that the entire financial system was selling and purchasing securities originated from subprime loans granted in the US (Hull, 2015). This makes clear the importance of having a strong credit risk management system due to its potential devastating effects to both the financial institutions and the entire financial system. Although all risks banks faces can negatively affect their performance, credit risk has been recognized as the dominant risk affecting the greatest performance (Sinkey, 1992). Hull (2015, p. 41) defines credit risk as “the risk that counterparties in loan transactions and derivatives transactions will default”, and also agrees that it is one of the most important risks banks face. Hull (2015), justifies this due to commercial banks’ nature of activities which involves mainly accepting deposits and providing loans (Hull, 2015). Given that commercial banks’ key source of income comes from granting lending, the manner credit risk is managed is of great importance due to its effects on the banks’ profitability, survival, and growth. Therefore, due to its high importance, it is of great interest to investigate how bank profits are affected by this type of risk.
international exposure and in turn great influence in the European economy. Due to their large market capitalization, these banks contain an inherent systemic risk. This is ‘’the risk that a default by one financial institution will create a ripple effect that leads to defaults by other financial institutions and threatens the stability of the financial system’’ (Hull, 2015, p. 328). Unlike other sectors, in the banking industry when one bank fails its effects may spread to other banks, causing a chain effect and potentially affecting the stability of the whole financial system at home or even abroad (Zhang et al., 2016). This could happen due to the several transactions occurring among banks within the interbank market. If one bank fails other banks could also take a loss on their transactions with the defaulted bank. This, in turn, could cause the bank to default if the loss is large and produces financial issues such as insolvency. For instance, during the GFC, several banks that there were considered as ‘’too big to fail’’ were bailed out by their governments instead of being allowed to default, due to governments worrying about systemic risk (Hull, 2015). Therefore, we believe that these banks are worth examining given that their distress could lead to major disruptions to both the financial system and the economy.
There is an extended list of studies examining this relationship such as Bourke (1989), Molyneux and Thornton (1992), Angbazo (1997), Godlewski (2005), Beck, Jakubik, and Piloiu (2013), Kaaya and Pastory (2013), and Kayode et al. (2015). The studies employ as a proxy for credit risk the non-performing loans ratio (NPLR) and find a negative association among the banks’ performance and credit risk, concluding that the higher the exposure to credit risk the lower the profits. Moreover, studies such as Mester (1996), Berger and DeYoung (1997), Fries and Taci (2005), and Podpiera and Weill (2008), discuss the importance of the NPLR and conclude that the non-performing loans are negatively associated with the banks’ profitability and stability due to causing deterioration in the banks’ asset quality. However, these studies focus on a restricted time horizon while our research takes place during 2000 to 2018, therefore taking into account major changes occurring over the last 18 years such as the GFC and SDC which could have an effect on the relation examined.
after the GFC. Finally, we research if the European banking sector is exposed to moral hazard behaviour. Consequently, this study attempts to investigate and answer the following research question: What is the relationship between credit risk and profitability of European commercial banks?
The remainder of the paper is structured by Section 2 which provides a theoretical and empirical literature review about the topic, as well as information about the hypotheses tested. Section 3 discusses the sample collected and methodology employed for the research by explaining the procedure used to collect the data and the model and tests exploited. Section 4 analyses the data and presents the findings in which the research question is answered, and Section 5 concludes the paper by providing a summary of the paper, as well as recommendations for further research. 2. Literature review
2.1. Theoretical review
Keeton and Morris (1987) discuss the ‘moral hazard’ hypothesis which is considered to be the leading theory for explaining bank failures and problem loans. Under this hypothesis, when banks' capital is low, bank managers respond to moral hazard incentives by taking excessive risk in the form of risky loans. This, in turn, increases the non-performing loans which eventually cause the overall portfolio risk to increase. Indeed, banks' excessive lending is commonly rendered as the main determinant for the increase of the NPLs (Keeton, 1999). The highlight of this theory is that as the NPL increases it leads to an increase in riskier lending, which may potentially cause further deterioration in the quality of the bank's loans as well in the instability of the financial system (Berger and DeYoung, 1997).
An alternative explanation of this behaviour is given by Jensen and Meckling’s (1976) theory of incentives. According to this theory, there are two types of moral hazard issues which motivate bank managers to take excessive risk. The first concerns managerial rent-seeking, which emerges when managers seek to satisfy their private benefits by either monitoring inadequately the loans' quality or by promoting ''pet projects''. The moral hazard problem arises when managers pursue their own incentives, which are usually not in line with the owners' interests. The second moral hazard issue emerges from a conflict of interest between creditors and shareholders. When stockholders accept risky loans, they ultimately transfer this risk to the depositors. As the amount of risky loans increases, the NPLs also increases which in turn creates incentives to both the bank managers and shareholders in transferring this risk (Zhang et al., 2016). What this theory implies, is that both moral hazard issues can provoke excessive risk-taking, thus leading to an increase in the NPLs and a decrease in the assets' quality. These, in turn, could ultimately cause the bank's failure. An additional explanation of the emergence of moral hazard behaviour is provided by Fiordelisi, Marques-Ibanez, and Molyneux (2011). In this case, the existence of information frictions and the presence of agency issues among the bank owners and managers could all cause the rise of moral hazard incentives. In contrast, banks with higher capital are less likely to engage in moral hazard activities and more likely to adopt cost-cutting procedures such as by allowing shareholders to have active participation in the capital allocation or in the control of the bank’s costs (Fiordelisi, Marques-Ibanez, and Molyneux, 2011). Bourke (1989) explains that high capitalized banks are able to attract cheaper capital, due to being perceived as less risky and consequently enjoy more profits, while low capitalized banks enjoy less, and thus, are more likely to engage in riskier behaviour to enhance their profits.
by the prospect theory (Ackert and Deaves, 2009) which argues that agents are risk-seekers when it comes to sure losses but become risk-averse when faced with sure gains. Therefore, it is reasonable to assume that during distressed situations bank managers have greater incentives to increase their excessive risk-taking activities (Zhang et al., 2016).
Empirically, Keeton and Morris (1987) find evidence of the ‘moral hazard’ hypothesis. Specifically, the study shows a negative and significant association among the equity-to-assets ratio and the NPL ratio. Berger and DeYoung (1997) also find the capital ratio to be negatively associated with the NPLR. In particular, the study finds that one of the determinants of higher NPLs was banks’ low capital ratios indicating moral hazard behaviour. The study concludes that low capitalization evokes banks to take higher risk. Likewise, the findings of Salas and Saurina (2002), and Klein (2013) show a negative and significant relation between the equity-to-assets ratio and the NPLR, which is evidence of the ‘moral hazard hypothesis. In the same line, Louzis, Vouldis, and Metaxas (2012), also approximate moral hazard behaviour through the banks’ capitalization to investigate the ‘moral hazard’ hypothesis on the Greek banking system. As opposed to the previous studies, the findings show an insignificant association among the capital ratio and the NPLR, concluding that moral hazard does not exist on the Greek financial system. The authors explain this result by taking into account Greece's small-scale market which could create disincentives for bank managers to take excessive risk and short-termism for reputation reasons. Additionally, the authors explain that due to Greece’s few numbers of banks, the regulatory authorities are able to have a precise on-site overview of each bank’s riskiness on its loan portfolio. As a result, the regulatory authorities are able to interfere quickly and accordingly each time they observe risky behaviour. The authors conclude that this leads to a lower possibility for the bank managers to evoke higher NPLs due to moral hazard incentives.
2.2. Empirical review
The existing literature examining the impact of credit risk on banks has been widely investigated over the last decades. One of the earliest studies is of Bourke’s (1989) where the author analyses the performance of 90 banks in 12 countries or territories in Australia, Europe, and North America from 1972 to 1981. The study finds a negative association among credit risk and profitability and concludes that as a bank’s exposure to credit risk increases, its profitability deteriorates. Consistent with Bourke’s (1989) findings, is the study of Molyneux and Thornton (1992) on a sample of 18 European countries during the period 1986 to 1989. In the study of Miller and Noulas (1997), the authors estimate the determinants of profitability using a sample of 201 large US commercial banks from 1984 to 1990. The study examines the impact of credit risk and finds the banks’ profitability to worsen as their exposure to this type of risk increases. The authors explain this result by taking into account banks' high exposure to risky loans eventually leading to greater losses when these loans are not repaid. Consequently, the loan losses evoke the banks’ returns to decrease and subsequently the overall performance to decline. The study concludes that large banks in the US experienced low profitability due to deterioration in the quality of their loans portfolio. Likewise, Athanasoglou, Brissimis, and Delis (2008) examine the profitability determinants of Greek banks during 1985-2001. The study analyzes bank-specific determinants and finds credit risk to be negatively and significantly associated with Greek banks’ profits indicating that the higher a bank’s exposure to credit risk the worse it will perform. Similarly, Angbazo (1997) finds evidence that banks’ loan portfolios with lower exposure to credit risk contribute to higher profitability. In the same line, Dietrich and Wanzenried (2011) examine the determinants of profitability on a sample of 372 commercial banks in Switzerland during the period 1999-2009. Particularly, the study investigates the credit risk and finds that it is negatively related to the performance of Swiss banks.
environment. The study concludes that whether the banks’ risk management procedures are efficient depends on the manner credit risk is diversified to decrease the NPL ratio. Furthermore, Berger and DeYoung (1997) investigate the relation among commercial banks’ cost efficiency and problem loans (NPLs). The study finds an association between cost efficiency and NPLs, which in turn the authors conclude it affects the bank’s performance.
In the more recent literature, Aremu, Suberu, and Oke (2010) find evidence of the NPLs’ contribution to the inefficiency of banks' profits. Particularly, the authors identify NPLs as the most important factor threatening Nigerian banks' profitability. Park and Weber (2006) also find evidence that NPLs contributes to the inefficiency of Korean banks’ performance during 1992-2002. Kaaya and Pastory (2013) approximate bank profitability and credit risk through the ROA and NPLR respectively to investigate their association in Tanzanian commercial banks. The findings show a negative correlation among ROA and NPLR indicating that as the exposure to credit risk increases Tanzanian banks’ profits declines. Likewise, Kayode et al. (2015), examine the issue for a panel dataset of Nigerian banks over 2000 to 2013. The panel data analysis confirms that higher NPLs are associated with decreasing profits. The authors conclude that the NPL ratio reduces banks’ credit lending, liquidity, and slows down their growth. Similarly, Messai and Jouini (2013) find a negative correlation among the NPLs and the performance of 85 commercial banks in Spain, Greece, and Italy during 2004-2008. Furthermore, Ghosh (2015) also records a negative correlation among the profitability and the NPL ratio on a sample containing savings institutions and commercial banks over the District of Columbia and over 50 US states during the period 1984-2013. In the same line, Beck, Jakubik, and Piloiu (2013) present evidence of this relation in 75 countries, in both emerging and advanced economies, over 2000-2010. The paper also finds that in the last decade, in most of the sampled countries, the quality of credit portfolios was somewhat stable until the global financial crisis in 2007-2008. Since 2008, the quality of the banks' assets declined abruptly due to the global economic downturn. However, the deterioration in the performance of loan portfolios is not the same among countries.
In overall, these findings are robust to alternative measurements providing strong evidence of this relation. For instance, Klein (2013) employs the return on equity (ROE) as an alternative profitability indicator and finds the NPLs to be negatively and significantly correlated with the banks' performance. Similarly, Louzis, Vouldis, and Metaxas (2012), also finds the ROE to be negatively related to the NPLs of Greek commercial banks from 2003 to 2009.
indicators to capture the effect of credit risk and all conclude on the same results. First, credit risk poses a strong and negative effect on the banks’ financial health. Second, higher exposure to credit risk damages and contributes to the instability of banks’ profits.
2.3. Hypotheses development
During the global financial crisis, the multiple weaknesses of European banks were revealed including the issue of the NPLs (Anastasiou, Louri, and Tsionas, 2019), reflecting the important linkage among credit risk and profitability. Such linkage is important since an increase in banks’ credit risk raises the possibility of a crisis in both the banking sector and the financial market (Al-shakrchy, 2017). This reveals the significant threat credit risk poses on the financial institutions and the financial system as a whole. Therefore, finding a statistical association between credit risk and performance is essential for comprehending the manner credit risk affects the financial institutions and consequently the entire financial system. Based on the literature we expect to find a negative relation. Hence, our hypotheses are developed as followed:
H1: credit risk affects negatively the profitability of European commercial banks. H2: the relation between European commercial banks’ credit risk and profitability fluctuates over time.
Our next line of a hypothesis lies on the effects of the global financial crisis and the Sovereign debt crisis. What makes this issue so important is the fact that the GFC made clear that a financial crisis has the ability to influence strongly the stability of the global financial system as well as can produce devastating effects to the banking industry and the real economy (Zhang et al., 2016). Indeed, the effects of the crisis were rapidly spread throughout the European banking sector (Al-shakrchy, 2017), which in turn affected strongly the structure and performance of banks (Petria, Capraru, and Ihnatov, 2015). Therefore, given the severity of the crises, it is of vital importance to examine their effects on the European financial system. Hence, our hypotheses are formulated as:
H3: the global financial crisis has an effect on the profitability of European commercial banks.
H5: the relationship between credit risk and profitability of European commercial banks differs before and after the global financial crisis.
The last hypothesis concerns the examination of the ‘moral hazard’ hypothesis in the European banking industry. In particular, we investigate whether the sampled banks engage in riskier behaviours when faced with a low capital ratio, which in turn indicates the presence of a moral hazard. Consequently, the hypothesis is formulated as followed:
H6: European commercial banks engage in excessive risk-taking activities when faced with a low capital ratio.
3. Data and Methodology
3.1. Sample data
The sample includes data from the largest commercial banks in Europe in terms of total assets. SNL Financial (S&P Global Market Intelligence) provides a ranking list of the top 50 largest European banks by total assets as of December 31, 2017 (Appendix 1). The reason we choose these type of banks is due to their large market capitalization which may render them as systemic banks with international exposure and thus great influence in the economy. These banks all together constitute a higher share in the European banking sector and thus have a dominant effect on the economy throughout Europe. Therefore, they are worth examining given that their distress could lead to major disruptions to both the financial system and the economy. The sampled banks are investigated during a time horizon from 2000 to 2018. The reason we choose such a long period is to capture the effects of both the global financial crisis, for the pre-crisis and post-crisis period, and the Sovereign debt crisis which both has caused tremendous effects all over the European economy and potentially to the European banking industry.
and Wanzenried (2011), Beck, Jakubik, and Piloiu (2013), Kaaya and Pastory (2013), Ghosh (2015), Kayode et al. (2015), and Khan, Scheule, and Wu (2017). ROE has also been employed in a number of studies such as Bourke (1989), Fries and Taci (2005), Park and Weber (2006), Athanasoglou, Brissimis, and Delis (2008), Dietrich and Wanzenried (2011), and Louzis, Vouldis, and Metaxas (2012). This indicator reflects the bank’s ability to generate profits based on the amount of equity capital stockholders have invested in the bank (Kayode et al., 2015). Consequently, this study also employs ROA and ROE as indicators of profitability.
As an indicator of credit risk, the non-performing loan ratio (NPLR) is employed. We justify the choice of the NPLR based on its properties associated with credit risk and its frequency being used in previous studies. The NPLR is calculated as non-performing loans (NPLs) to total loans, where NPLs are loans overdue by more than 90 days (Louzis, Vouldis, and Metaxas, 2012). Additionally, an NPL can be defined as a loan in which the payments are overdue by less than 90 days but they are not expected to be paid, or more than 90 days' of interest has been capitalized, refinanced, or delayed after negotiations have taken place (IMF, 2009). A loan is rendered as an NPL when there is a late payment on the interest or principal rather than when the loan defaults (Choudhry, 2011). However, regardless of the state of the debtor, an NPL eventually will be written off as a default loss (Choudhry, 2011). Therefore, this ratio is closely associated with credit risk considering that it realizes the ratio of loans that have either defaulted or are close to default to the total loans. According to Beck, Jakubik, and Piloiu (2013), NPLs are commonly employed in credit risk models as a measurement of credit risk. Indeed, studies such as Mester (1996), Berger and DeYoung (1997), Fries and Taci (2005), Godlewski (2005), Brewer and Jackson (2006), Park and Weber (2006), Podpiera and Weill (2008), Aremu, Suberu, and Oke (2010), Louzis, Vouldis, and Metaxas (2012), Beck, Jakubik, and Piloiu (2013), Kaaya and Pastory (2013), Kayode et al. (2015), Klein (2013), Messai and Jouini (2013), and Ghosh (2015), employ on their research the NPLR as a proxy of credit risk. Consequently, this study also exploits this ratio as an approximation of a bank’s credit risk.
Balashanmugam, 2002; Athanasoglou, Brissimis, and Delis, 2008). Consistent with Guru, Staunton, and Balashanmugam (2002), Lehar (2005), Iannotta, Nocera, and Sironi (2007), Naceur and Kandil (2009), Khan, Scheule, and Wu (2017), and Rekik and Kalai (2018), we measure the bank’s size through the natural logarithm of the bank’s total assets. The existing literature provides a mixed opinion of the association between the bank’s size and its profitability. For example, Lehar (2005) supports that large in size banks have a higher number of products, participate more actively in the market, and thus have higher opportunities for risk diversification. Short (1979) also supports in favour of this relationship and explains that a bank's size is closely linked to its capital adequacy, given that large banks don’t have necessarily expensive capital and, thus, are more profitable. Likewise, Haslem (1968), Bourke (1989), and Molyneux and Thornton (1992) all relate positively bank size to capital ratios and argue that as the bank's size increases its profitability also increases. Indeed, studies such as Smirlock (1985), Iannotta, Nocera, and Sironi (2007), and Bertay, Demirguc-Kunt, and Huizinga (2013), find a significantly positive relationship between the bank’s size and its performance, concluding that the larger the size of a bank the higher its profitability. However, such as Short (1979), Tschoegl (1982), Guru, Staunton, and Balashanmugam (2002), Micco, Panizza, and Yanez (2007), and Naceur and Kandil (2009) find an insignificant association. Moreover, Berger, Hanweck, and Humphrey (1987) argue against this relationship and explain that a small cost reduction can be saved by increasing the size, suggesting that it might lead to scale inefficiencies. Similarly, Athanasoglou, Brissimis, and Delis (2008) argue that when banks increase extremely their size it could harm them due to bureaucratic and/or other reasons which could cause diseconomies of scale. Indeed, Benston, Hanweck, and Humphrey (1982) find evidence that large in size banks do not experience economies of scale. Therefore, the literature provides ambiguous results about this relationship and further research is required to have a clear view of their association.
Al-shakrchy (2017) find the reverse. Moreover, Naceur and Kandil (2009), finds an insignificant association between ROA, ROE and liquidity risk. Therefore, a final conclusion about the impact of liquidity risk continues to be vague and further research is required to have a clear view of this relationship. Consistent with Guru, Staunton, and Balashanmugam (2002), Williams (2004), and Al-shakrchy (2017), we indicate liquidity risk through the ratio of total loans to total deposits. Bourke (1989), and Athanasoglou, Brissimis, and Delis (2008) also support the use of this ratio as a measurement of liquidity risk.
The last group of profitability determinants is related to external macroeconomic control variables. Macroeconomic variables are primarily used as control variables and thus are treated as exogenous (Pain, 2003). The variables most commonly used in the literature are the gross domestic product (GDP) annual growth rate and inflation rate. In accordance with Iannotta, Nocera, and Sironi (2007), Valverde and Fernandez (2007), Naceur and Kandil (2009), Fiordelisi, Marques-Ibanez, and Molyneux (2011), Petria, Capraru, and Ihnatov (2015), and Khan, Scheule, and Wu (2017), we employ the GDP growth rate as a control variable. The GDP growth rate is employed in order to capture the effects of the business cycles (Fiordelisi, Marques-Ibanez, and Molyneux, 2011), and is considered the most significant indicator of a country’s performance showing the state of its economic cycle (Blanchard, Amighini, and Giavazzi, 2017). The literature suggests that there is a strong dependence between an economy’s GDP growth rate and the borrowers’ ability to repay their loans (Salas and Saurina, 2002; Louzis, Vouldis, and Metaxas, 2012; Beck, Jakubik, and Piloiu, 2013; Klein, 2013). That is so, because an expansion to the GDP leads to an increase in income (Blanchard, Amighini, and Giavazzi, 2017) which in turn improves borrowers’ debt servicing ability (Klein, 2013). In this case, non-performing loans are lower and bank profits are higher. Conversely, a contraction in the economic activity reduces income and weakens borrowers’ capacity to repay their loans (Klein, 2013), resulting in increased NPLs as more obligors past due to their debt obligations. Therefore, we expect the GDP to be positively related to the banks’ profitability. This assumption is proven empirically by Iannotta, Nocera, and Sironi (2007), Naceur and Kandil (2009), and Petria, Capraru, and Ihnatov (2015) where the studies conclude that as the economic activity increases, bank profits rise.
the inflation rate are correctly anticipated. If the bank expects inflation then it can adjust accordingly its interest rates to grow its revenues at a faster rate than its expenses and gain greater profits. Empirically, Bourke (1989), Molyneux and Thornton (1992), Claessens, Djankov, and Lang (2000), Guru, Staunton, and Balashanmugam (2002) find a positive relationship between bank profits and inflation. According to Guru, Staunton, and Balashanmugam (2002), this indicates that the bank’s management was anticipating inflation and adjusted accordingly its loan interest rates, most likely by increasing its rates, leading to an increase on its profitability. This, in turn, results in revenues increasing at a faster rate than expenses thus leading to higher profits. However, if the bank does not anticipate the inflation, and thus does not adjust its interest rates, it leads to an increase in its financing costs and in turn a decline in performance.
Table 1
Definition of the variables employed in the study.
Measurement Description
Bank profitability variables
Return on assets Net income/Total assets ROA Return on equity Net income/Total equity capital ROE
Bank-specific variables
Non-performing loans ratio Non-performing loans/Total loans
NPLR
Bank size Natural logarithm of total assets
LNTA
Liquidity Total loans/Total deposits LIQ Capitalization Equity capital/Total assets CAP Credit growth Total loans/Total assets CRG
Macroeconomic variables
Gross domestic product GDP annual growth rate GDP Inflation rate Annual percentage change in
the consumer price index
INF
Unemployment rate % of total labour force UNE
from the last ten years, and due to missing annual reports for specific years. Some banks haven’t published yet the annual report for 2018 and other banks have archival documents of annual reports from 2005 and onwards. Additionally, there are two banks on our sample that were founded in earlier years, one in 2006 and one in 2010, and thus we don’t have a complete dataset for these banks. Consequently, for the reasons discussed, we are not able to construct a complete dataset and thus we employ an unbalanced panel dataset. Finally, we collect the data related to the macroeconomic variables from Thomson Reuters Eikon, OECD data and the World Bank Data.
3.2. Empirical model construction
The majority of the literature reviewed on bank profitability agrees that the linear functional form is the most appropriate function to employ for analysis (Guru, Staunton, and Balashanmugam, 2002). Studies such as Short (1979), Bourke (1989), Molyneux and Thornton (1992), Molyneux, Lloyd-Williams, and Thornton (1994), Demirguc-Kunt and Huizinga (2000), and Kaaya and Pastory (2013), employ a linear regression model in their examination of banks’ performance. Accordingly, this study also employs a linear function and exploits the Ordinary Least Squares (OLS) method. Consistent with Bourke (1989), Molyneux and Thornton (1992), and Kayode et al. (2015), we employ a panel data analysis given that our sample contains several variables over different periods of time.
Beltratti and Stulz, 2012; Acharya and Mora, 2015; Khan, Scheule, and Wu, 2017) to avoid inference issues evoked by outliers (Zhang et al., 2016). This approach addresses the outlier problems by limiting the extreme values of the sample which could cause bias on the results if not treated properly (Brooks, 2014; Khan, Scheule, and Wu, 2017).
Following the same approach as Bertay, Demirguc-Kunt, and Huizinga (2013), we include one period lagged value of all bank-specific explanatory variables, and the contemporaneous values of the macroeconomic variables. The lagged values are included in order to reduce the likelihood of reverse causality (Bertay, Demirguc-Kunt, and Huizinga, 2013). In particular, we want to avoid the possibility of ROA causing the NPLR while our main interest is to examine the NPL's ratio effect on the banks’ performance. Messai and Jouini (2013) explain that high profitable banks have fewer incentives to engage in risky behaviours, while inefficient banks are more likely to grant risky loans and subsequently obtain higher levels of NPLs. This is due to the trade-off between risk and returns which indicates that a riskier loan produces higher returns which in turn boosts profits. This causes the NPLs to increase which is a result of low profitability. The inclusion of the lagged NPLR eliminates this possibility and produces the causality of NPLR on ROA. Econometrically, the setup of the regression model employed is as followed:
(2)
where i=bank, N=number of banks, so i=1,2,..., N, t=time period, T=number of time periods, so t=1,2,..., T, α is the constant term, β is a coefficient estimate reflecting the degree to which the relative explanatory variable contributes to the change of the dependent variable, and uit is the random disturbance term which can be either a fixed or a random effect term depending on the panel data model chosen.
To decide which model to use, the Hausman test is employed (Brooks, 2014) to Eq. (1) and Eq. (2).
According to Brooks (2014), fixed effects models allow the intercept in the regression model to differ for each cross-sectional unit while simultaneously it is time invariant. In such a model, all slope coefficients are fixed both over time and cross-sectionally. The entity-fixed effects model is structured in the same way as Eq. (1) but with the additional inclusion of the entity/cross-section fixed effect μi, and the random disturbance term vit. Additionally, we can employ a time-fixed effects model where years are included as additional explanatory variables to capture the effect from different time periods. For a time-fixed effects model, μi is replaced by the time-fixed effect λt. Moreover, Brooks (2014) suggests an alternative to the fixed effect, the random effects model where the relationship between the dependent and independent variable is assumed to be the same both temporally and cross-sectionally. This approach, as with the fixed effects model, uses different intercepts for each entity which are constant over time. However, the difference between those two models is in terms of the intercept for each entity, which is the same for all entities over time and assumed to arise from a mutual intercept. Additionally, this model includes a random
variable which is time-invariant but varies cross-sectionally. Furthermore, we use as a benchmark the linear models (1) and (2) to examine
further the relationship between credit risk and profitability. Particularly, we are interested in examining whether this relation is stable or fluctuates over time. To capture this effect we use the interaction term between the lagged NPL ratio and the GDP growth rate. The rationale behind this interaction term lies in the literature’ suggestion that the borrowers’ capacity to honour their debt obligations is strongly related to the economic activity (Salas and Saurina, 2002; Louzis, Vouldis, and Metaxas, 2012; Beck, Jakubik, and Piloiu, 2013; Klein, 2013). A decrease in the economic activity reduces borrowers’ income and, in turn, deteriorates their ability to pay back their debts (Klein, 2013) resulting in an increase in the banks’ NPLs and a decline in profits. Consequently, this indicates that the fluctuation or stability between the banks’ performance and credit risk depends on whether the economy grows, shrinks or stays constant.
Finally, the last model concerns the examination of the ‘moral hazard’ hypothesis. Consistent with Salas and Saurina (2002), Louzis, Vouldis, and Metaxas (2012), and Klein (2013) we include the lagged value of the capital ratio due to its impact not being straightforward. Additionally, to avoid the reverse causality issue between the NPLR and the independent variables, we include one period lagged value of the bank-specific control variables. The macroeconomic variables are included in their contemporaneous values. Consequently, the regression model employed is:
(3)
4. Empirical Results
4.1. Descriptive statistics and empirical model
differently their credit lending. This could happen either due to different risk appetites or lack of ability to grant a higher credit amount. Moreover, the indicator of liquidity risk amounts to 135.4% on average indicating that the European commercial banks during the last two decades suffered from high liquidity risk. Concerning the macroeconomic variable GDP, on average each bank’s economy grows at a rate of 1.640% with the value ranging from -2.828% to 4.321%. The negative GDP growth rate comes from the GIIPS countries Italy and Spain, which are considered as one of the weakest countries in the EU (Poon, Shen, and Burnett, 2017). The second macroeconomic variable inflation rate amounts on average to 1.737%. This variable reflects the price consumers pay for goods across the different European countries, which may differ from each country significantly. For example, the deflation rate of -15.1% is recorded by Spain. Finally, during the period 2000 to 2018, the average unemployment rate on the European sampled countries is 7.484% of the total labour force. Spain reports the highest rate of 15.490% while Germany the lowest with a rate of 3.425%.
Table 2
Summary statistics of the regression variables. The variables concern the largest European commercial banks during 2000-2018. All observations for all variables are winsorized at the top and bottom 5% to limit the extreme values.
Variable Mean Std. Dev Min Max
ROA 0.0041 0.0038 -0.0042 0.0116 ROE 0.0796 0.0753 -0.1140 0.2050 NPLR 0.0396 0.0399 0.0035 0.1530 LNTA 14.1100 2.1690 11.5900 19.3000 CAP 0.0510 0.0175 0.0237 0.0869 CRG 0.5130 0.1660 0.1760 0.7690 LIQ 1.3540 1.5550 0.1600 7.4280 UNE 7.4840 2.9710 3.4250 15.4900 INF 1.7370 1.1190 -0.1510 4.0710 GDP 1.6400 1.7570 -2.8280 4.3210
5, 6). The results show that our data suffer from heteroskedasticity and autocorrelation which in turn we solve by employing robust standard errors clustered at the bank level. The normality test reveals that the residuals are non-normally distributed. However, the normality assumption is not required to be satisfied considering that for large sample size, violation of this assumption is virtually inconsequential (Brooks, 2014). Therefore, the non-normality of the residuals does not pose a violation of our model. Moreover, the Hausman test shows the fixed effects model, along with year fixed effects, to be the most appropriate model for analysing the regressions of ROA and ROE. For testing the moral hazard hypothesis, the Hausman test reveals that we should employ the random model.
4.2. Bank performance and credit risk
Table 3 reports the empirical results for our main profitability indicator ROA. Model (1) examines the effect of credit risk on banks’ performance while controlling for internal and external profitability determinants as well for year fixed effects. We use as benchmark model (1) to further examine whether the relationship between the NPLs ratio and ROA fluctuates over time, as indicated by the model (2). The investigation of the global financial crisis and the Sovereign debt crisis are indicated by models (3) and (4). Finally, the last model (5) takes into account whether the relationship examined differs before and after the GFC.
Table 3
Fixed effects estimation results for the return on assets (ROA) as the dependent variable. The coefficients in bold indicate statistical significance.
All regressions include year fixed effects.
Variables Model 1 Model 2 Model 3 Model 4 Model 5
NPLRt-1 -0.0164** -0.0221*** -0.0168** -0.0175** -0.0110
(0.0067) (0.0065) (0.0067) (0.0065) (0.0114) LNTA t-1 -1.97e-06 -5.59e-07 -1.53e-05 1.16e-05 1.08e-06
(0.0002) (0.0002) (0.0002) (0.0002) (0.0002) LIQ t-1 9.77e-05 0.0001 0.0001 7.85e-05 8.88e-05
(0.0002) (0.0002) (0.0002) (0.0002) (0.0002)
GDP 0.0004*** 0.0002** 0.0005*** 0.0004*** 0.0004***
(0.0001) (0.0001) (7.56e-05) (8.49e-05) (8.71e-05) INF -0.0002 -0.0002 -0.0004** -0.0002* -0.0002
(0.0002) (0.0002) (0.0001) (0.0001) (0.0001) NPLRt-1*GDP 0.0033*
Table 3 (continued)
Model 1 Model 2 Model 3 Model 4 Model 5 Year 2001 -0.0006 -0.0006 -0.0003 -0.0006 -0.0005 (0.0006) (0.0006) (0.0006) (0.0006) (0.0006) Year 2002 -0.0017** -0.0017** -0.0014** -0.0018** -0.0016** (0.0007) (0.0006) (0.0007) (0.0007) (0.0006) Year 2003 -0.0009 -0.0009 -0.0006 -0.0010 -0.0009 (0.0006) (0.0006) (0.0006) (0.0006) (0.0006) Year 2004 -0.0008 -0.0007 -0.0007 -0.0008 -0.0007 (0.0005) (0.0005) (0.0006) (0.0005) (0.0005) Year 2005 -0.0005 -0.0004 -0.0003 -0.0005 -0.0004 (0.0006) (0.0006) (0.0006) (0.0006) (0.0006) Year 2006 -0.0001 -7.94e-07 -4.81e-05 -0.0001 -4.91e-06
(0.0005) (0.0005) (0.0005) (0.0005) (0.0005) Year 2007 -0.0009 -0.0009 -0.0009 -0.0008 (0.0006) (0.0006) (0.0005) (0.0005) Year 2008 -0.0036*** -0.0038*** -0.0036*** -0.0035*** (0.0008) (0.0008) (0.0008) (0.0008) Year 2009 -0.0025** -0.0028** -0.0028*** -0.0024** (0.0011) (0.0011) (0.0009) (0.0009) Year 2010 -0.0027*** -0.0028*** -0.0026*** (0.0006) (0.0006) (0.0006) Year 2011 -0.0041*** -0.0041*** -0.0036*** (0.0006) (0.0006) (0.0005) Year 2012 -0.0032*** -0.0032*** -0.0026*** Year 2013 -0.0031*** -0.0030*** -0.0027*** (0.0008) (0.0008) (0.0006) Year 2014 -0.0039*** -0.0038*** -0.0038*** (0.0008) (0.0007) (0.0007) Year 2015 -0.0031*** -0.0031*** -0.0032*** (0.0009) (0.0009) (0.0008) Year 2016 -0.0036*** -0.0036*** -0.0036*** -0.0037*** (0.0008) (0.0008) (0.0008) (0.0007) Year 2017 -0.0023*** -0.0023*** -0.0021*** -0.0023*** (0.0007) (0.0007) (0.0007) (0.0006) GFCdummy -0.0019*** (0.0006) SDCdummy -0.0034*** (0.0006) GFCpostcrisisdummy -0.0028*** (0.0006) NPLRt-1* GFCpostcrisisdummy -0.0090 (0.0113) Constant 0.0066* 0.0068* 0.0067* 0.0067* 0.0063* (0.0036) (0.0036) (0.0036) (0.0035) (0.0035) R-squared 0.356 0.361 0.337 0.346 0.340 Number of Banks 50 50 50 50 50
All estimated equations, except for model (5), produce a negative and statistically significant impact of the NPL ratio on ROA. Economically, this implies that there exists a negative association between credit risk and the profitability of European commercial banks. This result is expected considering that the NPLR represents loans that are past due interest and principal payments and which eventually are written off as default losses (Choudhry, 2011). As the number of loan losses increases, the financial losses incurred are covered by the bank directly (Al-shakrchy, 2017) which, in turn, causes its profits to decline severely. Considering that commercial banks’ main source of income comes from granting credit when the expected cash flow fails to arrive, the loss is major. Additionally, the NPLs not only affect the quality of the banks’ assets but also the entire balance sheet items (Ghosh, 2015). Therefore, the overall performance declines and if the deterioration is irreversible, it could threaten the bank’s entire financial health and potentially lead to failure. Consequently, we conclude that the higher the exposure to credit risk, the worse the performance of European commercial banks.
Comparing the results with those of earlier studies, the negative significant effect of NPLR on ROA is consistent with the findings of Kaaya and Pastory (2013), and Kayode et al. (2015). In overall, the negative relationship between credit risk and performance is supported by a number of studies employing alternative indicators, thus providing strong and robust evidence for the relation established. In particular, studies such as Bourke (1989), Molyneux and Thornton (1992), Mester (1996), Angbazo (1997), Berger and DeYoung (1997), Miller and Noulas (1997), Fries and Taci (2005), Godlewski (2005), Park and Weber (2006), Athanasoglou, Brissimis and, Delis (2008), Podpiera and Weill (2008), Aremu, Suberu, and Oke (2010), and Dietrich and Wanzenried (2011) conclude that the deterioration in banks’ profits can be attributed to higher exposure to credit risk.
Al-shakrchy, 2017). Nevertheless, we conclude that liquidity risk does not seem to be an issue in the performance of European commercial banks but further research would contribute to the establishment of a clear association among them.
Considering the external macroeconomic variables, all estimation outputs produce a positive and significant association among the GDP growth rate and the performance of European banks. This finding is consistent with our expectations considering that an expansion in the economic activity increases borrowers’ income which in turn enhances their capability to meet their debt obligations (Salas and Saurina, 2002; Louzis, Vouldis, and Metaxas, 2012; Beck, Jakubik, and Piloiu, 2013; Klein, 2013; Blanchard, Amighini, and Giavazzi, 2017). Consequently, as obligors pay back their loans on the predetermined day, revenues grow faster than expenses leading to higher profits (Athanasoglou, Brisimis, and Delis, 2008). Another explanation for this finding is that the banks' profits appear to be procyclical, as the demand for credit increases during economic booms resulting in higher profits. (Athanasoglou, Brissimis, and Delis, 2008; Albertazzi and Gambacorta, 2009). The positive effect of the GDP on the profitability of bank is in line with Iannotta, Nocera, and Sironi (2007), Naceur and Kandil (2009), and Petria, Capraru, and Ihnatov (2015) providing support to the argument that higher economic activity leads to improved performance. Furthermore, we see the bank profits to be negatively related to the inflation rate, although this result is supported only from estimation outputs (3) and (4). This finding contrasts the positive effect of this variable on ROA found by Bourke (1989), Molyneux and Thornton (1992), Claessens, Djankov, and Lang (2000), and Guru, Staunton, and Balashanmugam (2002). An explanation for this result could be that the bank management failed to forecast adequately future inflation and thus did not adjust accordingly its loan interest rates (Perry, 1992; Guru, Staunton, and Balashanmugam, 2002). As a consequence, financing costs increased at a faster pace than revenues causing a decline in profits (Perry, 1992; Guru, Staunton, and Balashanmugam, 2002). After we established an association among the bank’s profitability and its credit risk, we examine whether this relation is stable or fluctuates over time. The results are presented in the model (2) where the interaction term between the lagged NPLR and GDP growth rate is positive and statistically significant indicating a fluctuating relationship over time. Economically, this implies that banks do not face constant decreasing profits, as a result of higher credit risk, but instead, it depends on the current phase of the economic cycle. In other words, an economy that grows indicates a lower credit risk and higher profits, while the opposite holds for when the economy shrinks.
crises period as a result of the crises contagion effects. These findings are expected considering that a financial crisis has the ability to influence strongly the financial systems’ stability throughout the world, as well harm the financial institutions and the entire economy (Zhang et al., 2016). Our results are supported by Petria, Capraru, and Ihnatov (2015), Zhang et al. (2016), and Al-shakrchy (2017), providing evidence to the argument that the GFC influenced strongly the performance of banks and produced tremendous effects throughout the European banking industry. However, this in turn, raises the question whether the banks were profitable enough to confront such negative shocks taking into account that the more profitable the banking industry is the higher its ability to resist and overcome negative shocks such as a financial crisis (Athanasoglou, Brissimis, and Delis, 2008; Garcia-Herrero, Gavila, and Santabarbara, 2009; Rekik and Kalai, 2018). Furthermore, the last estimation output (5) considers whether the established association among bank profits and the NPL ratio differs before and after the GFC. The results report an insignificant interaction term between the credit risk and the post-crisis period indicating that the manner credit risk affects the profitability of European commercial banks is the same as it was prior to the financial crisis. However, the dummy variable representing the post-crisis period is negative and statistically significant. This implies that the GFC’ effects persisted for a prolonged period of time and contributed to the inefficiency of the European banks’ profits. However, further research is required to generate a clear view of the crisis’s persistence in the European financial market.
Finally, all estimation equations generate a consensus conclusion about the year dummies. We observe that until the year 2007 there is no impact on the performance of banks, except in the year 2002 where profits decline due to an external event affecting all banks at the same time. However, from 2008 and onwards all year dummies report a negative and statistically significant effect on European banks’ performance. These findings are expected considering that in 2008 the GFC peaked while shortly later in 2010 the SDC crisis emerged. Taking into account that the European banking industry has been struggling over a decade with two major crises occurring over a short period of time from each other, the deterioration in the banks’ performance is expected and well justified.
4.3. Moral hazard hypothesis
Lozano-Vivas (2015). However, it matches the findings of Louzis, Vouldis, and Metaxas (2012), and Ghosh (2015). One possible explanation is the large market capitalization of our sampled banks which renders them as systemic banks with international exposure and, thus, great influence in the European economy. If one bank fails it could lead to devastating effects to both the financial system and the entire economy. Therefore, due to their high importance, they are subject to stricter regulations and to continuous monitoring from the regulatory authorities. As a consequence, this makes it more difficult for the bank managers to engage in excessive risk-taking activities without the immediate interference from the regulatory authorities. Moreover, it could also create disincentives for riskier behaviours considering the stricter regulatory environment the banks operate.
Table 4
Random effects estimation results for the non-performing loans (NPLs) ratio as the dependent variable. The coefficients in bold indicate statistical significance.
Variables Model 1 Model 2 Model 3 Model 4 CAPt-1 0.1860 0.1050 0.1410 0.0064 (0.1410) (0.1360) (0.1360) (0.1920) CRGt-1 -0.0007 -0.0012 -0.0007 0.0008 (0.0211) (0.0212) (0.0213) (0.0224) ROAt-1 -2.0260*** -1.7480*** -1.7340*** -1.4680*** (0.4700) (0.4680) (0.4590) (0.3900) LIQt-1 -0.0039** -0.0038** -0.0040** -0.0046** (0.0016) (0.0015) (0.0017) (0.0018) LNTAt-1 0.0004 0.0009 0.0001 -0.0001 (0.0015) (0.0015) (0.0014) (0.0014) GDP -0.0010 -0.0018** -0.0010 -0.0012* (0.0006) (0.0007) (0.0006) (0.0006) INF -0.0013 -0.0011 -0.0013 -0.0009 (0.0013) (0.0013) (0.0013) (0.0012) UNE 0.0047*** 0.0044*** 0.0042*** 0.0042*** (0.0011) (0.0010) (0.0011) (0.0009) GFCdummy -0.0087*** (0.0026) SDCdummy 0.0070*** (0.0023) GFCpostcrisisdummy 0.0104 (0.0094) CAPt-1* GFCpostcrisisdummy -0.0080 (0.2030) Constant 0.0062 0.0083 0.0130 0.0196 (0.0228) (0.0230) (0.0231) (0.0229) R-squared 0.2900 0.3070 0.3082 0.3200 Number of Banks 50 50 50 50
Notes: in parentheses are the robust standard errors clustered at the bank level. *** Denote significance at 1%. ** Denote significance at 5%. * Denote significance at 10%.
4.4. Robustness tests
losses. A further analysis examining this observation would provide us with a better understanding of the sensitivity among ROA and ROE. Furthermore, Table 6 examines the robustness of the ‘moral hazard’ hypothesis results by replacing the control variable ROA with ROE. We find consistent estimates with our initial findings and conclude that moral hazard does not seem to be an issue on the European banking industry.
Table 5
Robustness test.
The setup is similar to the one in Table 3, but now the dependent variable is the return on equity (ROE).
The coefficients in bold indicate statistical significance.
Variables Model 1 Model 2 Model 3 Model 4 Model 5
Table 5 (continued)
Model 1 Model 2 Model 3 Model 4 Model 5 Year 2012 -0.0833*** -0.0835*** -0.0597*** (0.0169) (0.0169) (0.0144) Year 2013 -0.0811*** -0.0794*** -0.0632*** (0.0160) (0.0158) (0.0134) Year 2014 -0.0934*** -0.0924*** -0.0850*** (0.0144) (0.0144) (0.0135) Year 2015 -0.0768*** -0.0762*** -0.0706*** (0.0149) (0.0148) (0.0149) Year 2016 -0.0891*** -0.0884*** -0.0821*** -0.0855*** (0.0141) (0.0139) (0.0143) (0.0126) Year 2017 -0.0685*** -0.0687*** -0.0593*** -0.0657*** (0.0131) (0.0132) (0.0125) (0.0115) GFCdummy -0.0386*** (0.0129) SDCdummy -0.0765*** (0.0105) GFCpostcrisisdummy -0.0726*** (0.0117) NPLRt-1* GFCpostcrisisdummy -0.0507 (0.2290) Constant 0.1180 0.1220* 0.1120 0.1110 0.1080 (0.0709) (0.0716) (0.0726) (0.0715) (0.0704) R-squared 0.377 0.382 0.343 0.369 0.366 Number of Banks 50 50 50 50 50
Table 6
Robustness test.
The setup is similar to the one in Table 4, but now ROA is replaced with ROE. The coefficients in bold indicate statistical significance.
Variables Model 1 Model 2 Model 3 Model 4 CAPt-1 0.0530 -0.0134 0.0267 -0.1270 (0.1520) (0.1470) (0.1500) (0.2100) CRGt-1 -0.0041 -0.0041 -0.0037 -0.0007 (0.020) (0.0202) (0.0202) (0.0219) ROEt-1 -0.0830*** -0.0707*** -0.0697*** -0.0599*** (0.0200) (0.0195) (0.0199) (0.0176) LIQt-1 -0.0037** -0.0036** -0.0039** -0.0045** (0.0016) (0.0016) (0.0017) (0.0018) LNTAt-1 0.0005 0.0009 0.0002 -0.0001 (0.0015) (0.0015) (0.0014) (0.0014) GDP -0.0009 -0.0018** -0.0010 -0.0011* (0.0006) (0.0007) (0.0006) (0.0006) INF -0.0013 -0.0011 -0.0014 -0.0008 (0.0013) (0.0013) (0.0013) (0.0012) UNE 0.0049*** 0.0045*** 0.0044*** 0.0042*** (0.0011) (0.0011) (0.0011) (0.0010) GFCdummy -0.0092*** SDCdummy (0.0025) 0.0072*** (0.0024) GFCpostcrisisdummy 0.0078 (0.0094) CAPt-1* GFCpostcrisisdummy 0.0548 (0.2050) Constant 0.0107 0.0121 0.0168 0.0255 (0.0231) (0.0234) (0.0232) (0.0234) R-squared 0.2835 0.3022 0.3026 0.3170 Number of Banks 50 50 50 50
Notes: in parentheses are the robust standard errors clustered at the bank level. *** Denote significance at 1%. ** Denote significance at 5%. * Denote significance at 10%.
5. Conclusions and recommendations
particular, banks with a higher level of the NPLs ratio are more likely to face loan losses and subsequently default loans causing their overall performance to decline. We find this relationship to fluctuating over time depending on the economy’s business cycle. We further examine bank-specific and macroeconomic control variables and find evidence that the bank size and liquidity risk are unimportant in explaining European commercial banks’ profitability while the external macroeconomic factors GDP and inflation rate have a deterministic role in structuring banks’ profits. The findings also show that during the global financial crisis and Sovereign debt crisis, European banks faced a decrease in their profits. However, we find that the manner in which bank profits are affected by credit risk remains the same as it was prior to the global financial crisis.
Furthermore, we investigate the ‘moral hazard’ hypothesis and find no evidence of moral hazard on the European banking industry indicating that banks do not engage in excessive risk-taking activities. We explain this result by taking into account banks’ large size and, in turn, a systemic risk which could raise the necessity to be under active and constant monitoring. When banks operate under a strict regulatory environment, it becomes more difficult to take risky decisions without the immediate alarm and intervention of the regulatory authorities. However, further research is required to have a clear conclusion of the relationship between regulation and bank behaviour.
In sum, this research concludes that the profitability of European commercial banks during the period 2000 to 2018 has been affected negatively by their exposure to credit risk. The robustness of our results to alternative empirical methods strongly supports these findings. Particularly, our results support the argument that banks should allocate carefully credit in order to improve their assets’ quality and reduce their riskiness. Additionally, we believe that the findings of this study could assist bank managers to restructure their credit risk management procedures or regulatory authorities to update the banking regulations for better monitoring and elimination of excessive loan losses.
References
Acharya, V.V, Mora, N., 2015. A crisis of banks as liquidity providers. Journal of Finance 70(1), 1-43.
Ackert, L., Deaves, R., 2009. Behavioral Finance, Psychology, Decision-Making, and Markets. 1st edition. Cengage Learning.
Albertazzi, U., Gambacorta, L., 2009. Bank profitability and the business cycle. Journal of Financial Stability 5 (4), 393–409.
Al-shakrchy, E.J., 2017. The impact of credit risk managing on bank profitability an empirical study during the pre-and post-subprime mortgage crisis: The case of Swedish commercial banks. Journal of Business and Finance 3 (1), 31-42.
Anastasiou, D., Louri, H., Tsionas, M., 2019. Nonperforming loans in the euro area: Are core–periphery banking markets fragmented? International Journal of Finance and Economics 24 (1), 97-112.
Angbazo, L., 1997. Commercial bank net interest margin, default risk, interest rate risk, and off-balance sheet banking. Journal of Banking and Finance 21, 55-87.
Aremu, O.S., Suberu, O.J., Oke, J.A., 2010. Effective credit processing and administration as a panacea for non-performing assets in the Nigerian banking system. Journal of Economics 1(1), 53-56.
Athanasoglou, P.P., Brissimis, S. N., Delis, M. D., 2008. Bank-specific, industry specific and macroeconomic determinants of bank profitability. Journal of International Financial Markets, Institutions and Money 18 (2), 121-136.
Beck, R., Jakubik, P., Piloiu, A., 2013. Non-performing loans: What matters in addition to the economic cycle? Working Paper No. 1515. European Central Bank. Beltratti, A., Stulz, R.M., 2012. The credit crisis around the globe: Why did some banks perform better? Journal of Financial Economics105 (1), 1-17.
Benston, G. J., Hanweck, G. A., and Humphrey, D. B., 1982. Scale economies in banking. Journal of Money, Credit, and Banking 14 (4), 435-456.
20 (3), 501-520.
Berger, A. N., DeYoung, R., 1997. Problem loans and cost efficiency in commercial banks. Journal of Banking and Finance 21 (6), 849-870.
Berger, A.N., Bouwman, C.H.S., 2009. Bank liquidity creation. The Review of Financial Studies 22(9), 3779–3837.
Bertay, A., C., Demirguc-Kunt, A., Huizinga, H., 2013. Do we need big banks? Evidence on performance, strategy and market discipline. Journal of Financial Intermediation 22(4), 532-558.
Blanchard, O., Amighini, A., Giavazzi, F., 2017. Macroeconomics: A European Perspective. 3rd Edition. Pearson Education Limited.
Bourke, P., 1989. Concentration and other determinants of bank profitability in Europe, North America, and Australia. Journal of Banking and Finance 13(1), 65-79. Brewer III, E., Jackson III, W. E., 2006. A note on the “risk-adjusted” price-concentration relationship in banking. Journal of Banking and Finance 30 (3), 1041-1054.
Brooks, C., 2014. Introductory Econometrics for Finance. 3rd Edition. Cambridge University Press, United Kingdom.
Chirwa, E.W., 2003. Determinants of commercial banks' profitability in Malawi: a cointegration approach. Applied Financial Economics 13(8), 565-571.
Choudhry, M., 2011. An Introduction to Banking: Liquidity Risk and Asset-Liability Management. John Wiley & Sons, Chichester, U.K.
Claessens, S., Djankov, S., Lang, L. H. P., 2000. The separation of ownership and control in East Asian Corporations. Journal of Financial Economics 58, 81-112. Demirguç‐Kunt, A., Maksimovic, V., 1998. Law, finance, and firm growth. Journal of finance 53(6), 2107-2137.
Dietrich, A., Wanzenried, G., 2011. Determinants of bank profitability before and during the crisis: Evidence from Switzerland. Journal of International Financial Markets, Institutions and Money 21, 307–327.
Duran, M.A., Lozano-Vivas, A., 2015. Moral hazard and the financial structure of banks. Journal of International Financial Markets, Institutions and Money 34, 28–40.
Eisdorfer, A., 2008. Empirical evidence of risk shifting in financially distressed firms. The Journal of Finance 63 (2), 609–637.
European Central Bank (ECB)., 2015. Financial stability review. [online]. Available at:https://www.ecb.europa.eu/pub/pdf/fsr/financialstabilityreview201505.en.pdf?4c44 9c6c5ac622f7b679b21a1ba30252. [Accessed: 7 May 2019].
Fiordelisi, F., Marques-Ibanez, D., Molyneux, P., 2011. Efficiency and risk in European banking. Journal of Banking & Finance 35, 1315–1326.
Fries, S., Taci, A., 2005. Cost efficiency of banks in transition: Evidence from 289 banks in 15 post-communist countries. Journal of Banking and Finance 29(1), 55-81. Garcia-Herrero, A., Gavila, S., Santabarbara, D., 2009. What explains the low profitability of Chinese banks? Journal of Banking and Finance 33(11), 2080-2092. 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.
Goddard, J., Molyneux, P., Wilson, J.O.S., Tavakoli, M., 2007. European banking: An overview. Journal of Banking and Finance 31, 1911–1935.
Godlewski, J.C., 2005. Bank capital and credit risk taking in emerging market economies. Journal of Banking Regulation 6(2), 128-145.
Grier, W.A., 2007. Credit Analysis of Financial Institutions. 2nd Edition. Euromoney Institutional Investor PLC, London.
Haslem, J. A., 1968. A statistical analysis of the relative probability of commercial banks. Journal of Finance 23(1), 167-176.
Hull, J., 2015. Risk Management and Financial Institutions. 4th Edition. John Wiley & Sons, New Jersey.
Iannotta, G., Nocera, G., Sironi, A., 2007. Ownership structure, risk and performance in the European banking industry. Journal of Banking and Finance 31, 2127–2149. International Monetary Fund (IMF)., 2002. Financial soundness indicators: Analytical aspects and country practices. Occasional Paper No. 212. Washington.
International Monetary Fund (IMF)., 2009. Annual Report. [online]. Available at: https://www.imf.org/. [Accessed: 10 May 2019].
Jensen, M.C., Meckling, W.H., 1976. Theory of the firm: managerial behaviour, agency costs and ownership structure. Journal of Financial Economics 3(4), 305–360.
Kaaya, I., Pastory, D., 2013. Credit risk and commercial banks performance in Tanzania: A panel data analysis. Research Journal of Finance and Accounting 4(16), 55-62.
Kaminsky, G.L, Reinhart, C.M, 1999. The twin crises: the causes of banking and balance of payments problems. The American Economic Review 89(3), 473–500.
Kayode, O. F., Obamuyi, T. M., Owoputi, J. A., Adeyefa, F. A., 2015. Credit risk and bank performance in Nigeria. IOSR Journal of Economics and Finance 6 (2), 21-28.
Keeley, M.C., 1990. Deposit insurance, risk, and market power in banking. American Economic Review 80 (5), 1183–1200.
Keeton, W.R., Morris, C., 1987. Why do banks’ loan losses differ? Economic Review 72 (5), 3-21.
Keeton, W.R., 1999. Does faster loan growth lead to higher loan losses? Economic Review 84(2), 57-75.
Klein, N., 2013. Non-performing loans in CESEE: Determinants and impact on macroeconomic performance. IMF Working Papers No. 13/72, International Monetary Fund.
Lehar, A., 2005. Measuring systemic risk: A risk management approach. Journal of Banking and Finance 29(10), 2577-2603.
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 and Finance 36 (4), 1012-1027.
Mester, L. J., 1996. A study of bank efficiency taking into account risk-preferences. Journal of Banking and Finance 20(6), 1025-1045.
Messai, A.S, Jouini, F., 2013. Micro and macro determinants of non-performing loans. International Journal of Economics and Financial Issues 3(4), 852–860.
Micco, A., Panizza, U., Yanez, M., 2007. Bank ownership and performance. Does politics matter? Journal of Banking and Finance 31 (1), 219–241.
Miller, S.M., Noulas, A.G., 1997. Portfolio mix and large-bank profitability in the USA. Applied Economics 29, 505-512.
Molyneux, P., Thornton, J., 1992. Determinants of European bank profitability: A note. Journal of Banking and Finance 16, 1173-1178.
Molyneux, P., Lloyd-Williams, D.M ., Thornton, J., 1994. Competitive conditions in European banking. Journal of Banking and Finance 18(3), 445-459.
Naceur, S. B., Kandil, M., 2009. The impact of capital requirements on banks’ cost of intermediation and performance: The case of Egypt. Journal of Economics and Business 61, 70–89.
Nier, E., Baumann, U., 2006. Market discipline, disclosure and moral hazard in banking. Journal of Financial Intermediation 15 (3), 332–361.
Park, H.K., Weber, W.L., 2006. A note on efficiency and productivity growth in the Korean banking industry, 1992-2002. Journal of Banking and Finance 30(8), 2371-2386.
Perry, P., 1992. Do banks gain or lose from inflation? Journal of Retail Banking 14 (2), 25-30.
Petria, N., Capraru, B., Ihnatov, I., 2015. Determinants of banks’ profitability: evidence from EU 27 banking systems. Procedia Economics and Finance 20, 518 – 524.
Podpiera, J., Weill, L., 2008. Bad luck or bad management? Emerging banking market experience. Journal of Financial Stability 4(2), 135-148.
Poon, W.P.H., Shen, J., Burnett, J.E., 2017. An empirical study of international spillover of sovereign risk to bank credit risk. The Financial Review 52 (2), 281–302. Rekik, M., Kalai, M., 2018. Determinants of banks’ profitability and efficiency: Empirical evidence from a sample of banking systems. Journal of Banking and Financial Economics 1(9), 5-23.
Revell, J., 1979. Inflation and Financial Institutions. Financial Times, London.
Rinaldi, L., Sanchis-Arellano, A., 2006. Household debt sustainability: What explains household non-performing loans? An empirical analysis. Working Paper. No. 570. European Central Bank.
Salas, V., Saurina, J., 2002. Credit risk in two institutional regimes: Spanish commercial and savings banks. Journal of Financial Services Research 22(3), 203– 224.
Short, B.K., 1979. The relation between commercial bank profit rates and banking concentration in Canada, Western Europe and Japan. Journal of Banking and Finance 3, 209–219.
Sinkey, J.F., 1992. Commercial Bank Financial Management in the Financial-Services Industry. 4th Edition. Macmillan, New York.
Stern, G.H, Feldman, R.J, 2004. Too Big to Fail: The Hazards of Bank Bailouts. Brookings Institution Press, Washington, DC.
Tschoegl, A. E., 1982. Concentration among international banks: A note. Journal of Banking and Finance 6(4), 567-578.
Valverde, S.C, Fernandez, F.R, 2007. The determinants of bank margins in European banking. Journal of Banking and Finance 31, 2043–2063.
Williams, J., 2004. Determining management behaviour in European banking. Journal of Banking and Finance 28 (10), 2427-2460
