Macroeconomic Determinants of Credit Risk on the Example of Non-performing Loans

The aim of this research study is to investigate the determinants of the non-performing loan ratio in 2019, which can be considered a proxy for the condition of the banking sector. While the factors with a material impact can be internal (bank-specific) or external in nature, this study focuses only on macroeconomic determinants, including real GDP growth, unemployment rate, public debt ratio, domestic credit to private sector ratio, currency depreciation, inflation rate, interest rate and a dummy variable representing OECD membership.

The analysis covers data points from 106 economies over the period from 2008 to 2019. The covered period was chosen to focus on the consequences of the global crisis caused by the subprime credit market crash. Due to extraordinary measures taken by governments and central banks, the year 2020 was therefore omitted on purpose. The NPL ratio, GDP growth, unemployment, inflation, and domestic credit to private sector ratio were extracted from the World Bank database, whereas interest rates, public debt, and currency depreciation were taken from the International Monetary Fund (IMF) dataset.

The originality of the article resides in the author's concentration on panel data from countries around the globe, encompassing all continents, economies of varying sizes, and both developed and developing nations. The majority of scholars have explored the NPL ratio through the lens of time series. Panel-based studies have advantages compared to time-series analyses, and these benefits include (1) increasing degrees of freedom while reducing problems of data multicollinearity, (2) building more realistic behavioural models, (3) eliminating or limiting estimation bias, (4) obtaining more precise estimates of relationships and producing more accurate predictions, (5) providing information on the appropriate level of aggregation, and (6) simplifying cross-sectional or time series data (Hsiao, 2005, p. 12). However, there were a limited number of panel studies that included countries from multiple continents and focused on the effects of the 2008 financial crisis. Moreover, participation in the OECD has not been considered a variable influencing NPLs.

Four sections comprise the remainder of the article: The Literature Review section begins the article, and is followed by the Methods section, which includes the variable definitions. The Results section contains descriptive statistics, a correlation matrix, and tests for normality, heteroscedasticity, collinearity, and endogeneity. The final section, Discussion and Conclusions, includes a discussion of the findings, study limitations, concluding remarks, and policy implications.

According to the IMF, a non-performing loan (NPL) is a loan when payments of interest and/or principal are past due by at least 90 days; interest payments equal to or greater than 90 days have been capitalized, refinanced, or delayed by agreement; or payments are less than 90 days overdue. Other circumstances, however, can also call into question whether payments will be made in full. Classification as a non-performing loan means that the debtor is in default, i.e., at risk of bankruptcy. If the debtor starts paying off liabilities again, the loan returns to the performing category. The NPL ratio, expressed as the ratio between the value of non-performing loans and the total loans in the bank, is the fundamental measure of credit risk. It is one of the most important measures affecting the bank's financial results (IMF, 2019, p. 59).

Consequently, the value of the NPL ratio can be viewed not only as a measure of credit risk, but also as an indicator of the overall state of the economy. Non-performing loans may determine the overall weakness of the banking system within a country (Giammanco, Gitto & Ofria, 2022, p. 2). Other studies prove that economic changes manifest themselves more quickly in the banking sector than in other sectors of the economy (Skikiewicz & Garczarczyk, 2018, p. 128). NPLs can be also used to mark the onset of a banking crisis (Reinhart & Rogoff, 2010, p. 41). A number of stakeholders, including the public as a user of banking services and as a potential investor in the banks’ equity, the financial markets, the banks’ management, shareholders, the banking supervisors, and in terms of ensuring the stability of the financial system and academic circles, regard information on the bank's loan quality as crucial.

The financial system was profoundly affected by the 2008 financial crisis (Batrancea, 2021, p. 2). The reduction of economic activity caused by the global crises lead to a decrease in liquidity across different sectors (Piosik, 2022, p. 339). Further, liquidity problems lead to increased levels of non-performing loans for banks. The 2008 financial crisis’ implications have changed banks’ principles of operation (Amuakwa-Mensah, Marbuah & Marbuah, 2017), (Kuzucu, 2019), (Tarchouna, Jarraya -& Bouri, 2017), (Yüksel, 2017). In result, the Basel Committee on Banking Supervision announced the third Basel Accord in November 2010 with the aim of enhancing the stability of the banking system as a reaction to the weaknesses in financial regulation that had been uncovered. According to Duffie, the reforms, particularly bank capital rules, have contributed significantly to the stability of the financial sector (Duffie, 2018, p. 2).

The topic of NPLs has been investigated by numerous authors. Two groups of factors influence the value of this indicator: internal (at the company level) and macroeconomic. Bank-specific determinants can be generally summarized as poor management; however, a lower-quality credit portfolio can sometimes be compensated for by a higher loan margin. Internal factors include an inefficient process of credit scoring, insufficient control of borrowers, and the absence of collateral (Podpiera & Weil, 2008), a high level of efficiency, bank size (Abid, Ouertani & Zouari-Ghorbel, 2014), and banks’ low capitalization (Keeton & Morris, 1987).

Klein examined the banking sector in the countries of Central, Eastern, and Southeastern Europe from 1998 to 2011 in terms of macroeconomic factors. The level of non-performing loans rises with an increase in currency depreciation, the unemployment rate, and inflation, but NPLs exhibit negative correlation with real GDP growth (Klein, 2013).

Beck and the team who analysed a sample of 75 countries between 2006 and 2015 discovered that the following factors have a significant impact on NPL ratios: real GDP growth, share prices, the exchange rate, and the lending interest rate. In particular, changes in economic activity are the primary driver of the credit portfolio's quality. In addition, exchange rate depreciations may lead to a deterioration of banks’ assets in nations with a high proportion of lending in foreign currencies, such as Swiss franc-denominated credits in Poland, Hungary, and Croatia (Beck, Jakubik & Piloiu, 2015).

Kjosevski and Petkovski assert that the NPLs in the Baltic States are subject to comparable macroeconomic determinants. Their empirical research indicates that the most influential macroeconomic determinants of NPL are changes in GDP, public debt, inflation, and unemployment. (Kjosevski & Petkovski, 2021).

Using the generalized method of moments, researchers covering banks in Southeastern European countries from 2003 to 2010 reached comparable conclusions. According to their studies, higher NPL values were correlated with lower economic growth, higher inflation, and higher interest rates (Curak, Pepur & Poposki, 2013).

The Polish banking sector was covered by Wdowiński, who used quarterly statistical data from 1997 to 2013. According to the study, the GDP growth rate, interest rates, unemployment rate, and exchange rate fluctuations are the primary macroeconomic determinants of the non-performing loan ratio in banks. The author claims that a deep recession would result in a material deterioration of credit portfolios both in the corporate and consumer segments (Wdowiński, 2014).

Petkovski conducted more recent evaluations of the Polish banking sector on a panel of 18 Polish banks using annual data from 2005 to 2018. The results indicate that GDP growth, domestic lending to the private sector, public debt, and unemployment have the largest influence on the value of non-performing loans (Petkovski, Kjosevski & Jovanovski, 2021).

Other authors also include public debt as a variable that degrades the quality of the loan portfolio. A model constructed using a sample of 85 banks in three southern European countries (Greece, Italy, and Spain) from 2004 to 2008 confirms that the NPL ratio is negatively correlated with GDP growth and positively correlated with unemployment and the real interest rate (Messai & Jouini, 2013).

Similar conclusions are confirmed in the Greek banking system. The authors demonstrate that value of nonperforming loans can be primarily attributed to macroeconomic variables including GDP growth, unemployment, interest rates, and public debt (Louzis, Vouldis & Metaxas, 2015).

Batrancea gave an alternative viewpoint on the question of interactions between macroeconomic variables and NPLs. Researchers view GDP growth as a dependent variable explicable by the NPL ratio, bank capital to assets ratio, bank liquid reserves to assets ratio, interest rate spread, and inflation. The authors demonstrated, however, that bank capital-to-asset ratios were the primary drivers of economic development (Batrancea, Rathnaswamy & Batrancea, 2021).

The purpose of model validation is to ensure that a model is designed adequately to solve the problem it was planned to solve. In particular, it should provide accurate predictions. In order to achieve this, the author first developed a model using explanatory factors selected on the basis of the literature review. The second step focused on econometric analysis and tests including descriptive analysis, testing of specification, collinearity, endogeneity.

The findings reveal that macroeconomic variables, including the unemployment rate, banking system development, public state debt, and OECD membership debt, have a substantial effect on the ratio of non-performing loans.

The results validate the hypothesis that unemployment affects households’ ability to pay their loans resulting in a worsening of the loan portfolio of the entire economy. This work concurs with the findings of earlier studies on the NPL being correlated with the unemployment rate (Klein, 2013), (Louzis et al., 2015), (Messai & Jouini, 2013), (Wdowiński, 2014).

The significance of the introduction of the banking sector development index proxied by the relation between total loans granted to the private sector (generally, the size of bank assets) indicates that banking system maturity leads to accurate creditworthiness assessment. Banks in developed systems are also generally more risk averse as (Abid, Ouertani & Zouari, 2014), (Petkovski et al., 2021), (Keeton & Morris, 1987). However, this study does agree with the work of both Pesola and Nkusu, who claimed that high levels of debt make borrowers much more vulnerable to shocks, which may directly negatively influence their income and, therefore, their ability to service their obligations (Pesola, 2005), (Nkusu, 2011).

Furthermore, high state debt limits the availability of financing and may result in asset allocation inefficiencies that manifest as higher NPLs. This phenomenon is highly visible in Greece, which places second in the NPL's highest ranking and is characterized by the highest public ratio. This study confirms the conclusions reached by Louzis, who covered the Greek banking system (Louzis et al., 2015).

In contrast, the inflation rate, GDP growth rate, interest rate, and national currency depreciation are not significant in the constructed model. The lack of significance of GDP change and interest rates in the panel data model does not necessarily mean that those measures have no influence on the NPL ratio. Researchers demonstrated their significance when employing models based on time series. In practice, expansive lending policies fuelled by low interest rates lead to an increase in the value of loans, followed by a rise in GDP. Consequently, after an increase in interest rates, borrowers’ ability to repay may diminish, resulting in a rise in the nonperforming loan ratio.

The study does not take into account the country's geographical location and deliberately ends in 2019 in order to focus on the effects of the 2008 financial crisis and not on extraordinary actions implemented by governments and central banks in response to worldwide pandemics. The inclusion of the aforementioned would be an advantageous addition to the subject.

The practical implications are advantageous for governments, supervisory agencies, regulators, and international investors, who can either influence the NPL ratio through the implementation of pertinent economic policies or make informed business decisions. From the perspective of international investors, there may be an expectation of higher returns to account for the anticipated non-performing loans in indebted nations. In terms of economic policy, however, a study reveals that established banking systems have higher portfolio quality. In addition, effective public debt management contributes to a sustainable development policy. By facilitating sustainable growth and reducing administrative impediments, authorities can improve the economy's operation.