Previous Article in Journal
Perception of Corporate Governance Factors in Mitigating Financial Statement Fraud in Emerging Markets: Jordan Experience
 
 
Font Type:
Arial Georgia Verdana
Font Size:
Aa Aa Aa
Line Spacing:
Column Width:
Background:
Article

Impact of Macro Factors on NPLs in the Banking Industry of Kazakhstan

1
Higher School of Economics and Business, Al-Farabi Kazakh National University, Almaty 050051, Kazakhstan
2
JSC “The Fund of Problem Loans”, Almaty 050051, Kazakhstan
3
Transport and Telecommunication Institute, Faculty of Management and Logistics, LV-1019 Riga, Latvia
4
SIA StarBridge, LV-1050 Riga, Latvia
5
Institute of Life Sciences and Technologies, Daugavpils University, LV-5401 Daugavpils, Latvia
*
Authors to whom correspondence should be addressed.
J. Risk Financial Manag. 2025, 18(8), 431; https://doi.org/10.3390/jrfm18080431 (registering DOI)
Submission received: 3 July 2025 / Revised: 27 July 2025 / Accepted: 29 July 2025 / Published: 2 August 2025
(This article belongs to the Section Banking and Finance)

Abstract

The importance of non-performing loans (NPLs) for the stability of financial sectors is difficult to overestimate. The NPL level depends on numerous factors; this study’s goal is to determine the impact of macroeconomic factors on NPLs with the mediation effect of foreign, saving and social factors in Kazakhstan’s banking sector. To determine the affecting factors, the authors performed a systematic literature review. To determine the dependencies between constructs, the Partial Least Squares Structural Equation Modeling (PLS-SEM) method was used. Macroeconomic factors’ direct effect on non-performing loans (NPLs) was examined; a significant negative dependence was determined. The mediation effect of foreign, saving, and social factors was investigated. Foreign factors have a mediation effect, strengthening the dependence between macro factors and NPLs. Nevertheless, they do not have a mediating effect; moreover, they balance and make the effect of macro factors on NPLs statistically insignificant. These findings allow policy-makers to stabilize the situation on NPLs in the financial markets of developing countries like Kazakhstan by directly influencing not only the financial sector but also other sectors of the national economy.

1. Introduction

The stability of a financial system is one of the primary tasks for national governments. It works as a blood system for any country, affecting economic development (Athari, 2021a, 2021b; Menicucci & Paolucci, 2021; Saliba et al., 2023).
However, it is necessary to consider the “health” of the financial system, not only taking into consideration financial indicators but shifting the research view to economic indicators as factors depending on and, at the same time, affecting macroeconomic factors.
When a financial system malfunctions, a country’s economy can suffer significantly. There is also an inverse relationship—the financial system of any country is very sensitive to the state of the macroeconomic performance (Ariffin, 2012; Bucevska & Hadzi Misheva, 2017; Climent-Serrano, 2019; Dibar & Sridevi, 2021; Tanasković & Jandrić, 2015). The sectors of the economy also react to and, simultaneously, affect the risks and state of the financial system (Cernisevs & Popova, 2022; Melnyk et al., 2025; Onyshchenko et al., 2025; Popova & Cernisevs, 2023; Saksonova & Papiashvili, 2021; Saksonova & Vilerts, 2015). There are numerous risks which should be considered in the financial sector (Cernisevs et al., 2023). This article considers only one factor of instability of the financial system—non-performing loans (NPLs). NPLs serve as one of the visible manifestations of the deteriorating quality of credits and increased systemic risk within the banking sector. NPLs are widely discussed by the scientific community (Anastasiou et al., 2019; Boudriga et al., 2009; Irani et al., 2022; Saliba et al., 2023; Vouldis & Louzis, 2018).
NPLs not only affect bank profitability and capital adequacy but also lower credit availability and slow economic growth and development. In emerging economies like Kazakhstan, banking systems have the leading role in redirecting the financial resources into the real sectors of the economy; they are subjected to serious risks (Popova et al., 2024, 2025). Therefore, the impact of macroeconomic factors on loan performance becomes particularly visible.
The goal of this research is to determine the impact of macroeconomic factors on NPLs with the mediation effect of foreign factors, saving factors, and social factors in Kazakhstan. The factors for analysis were chosen on the basis of a substantial literature review (the methodology is presented in Section 2.1).
As we see, scientists pay great attention to the impact of macroeconomic indicators on NPLs. Therefore, the topic is not absolutely novel. However, researchers traditionally choose a wide range of indicators for the application of statistical analysis. The authors of this article have decided to concentrate on a rather small number of indicators to have a more exact and brighter vision of the factors’ interconnectedness. We tried to choose the minimum set of indicators capable of shaping the repayment capacity of borrowers and the overall confidence in the financial system, which we supposed to be the leading factors of credit deterioration. The chosen indicators reflect macroeconomic volatility, foreign financial pressures, the savings behavior of the population, and deposit dynamics and social inequality indicators, and this paper analyzes their effect on “bad loans” (NPLs) in the Kazakh banking sector.
The authors believe that this impact has not been investigated fully.
First of all, the authors have not found any articles considering the economies of central Asia in accordance with the inclusion criteria. There are analyses of European and Asian countries, BRICS countries, the US, Turkey and South American countries, and African countries. Nevertheless, the post-Soviet central Asian countries were not analyzed, though they are rich in resources, have well-developed mining and manufacturing industries, and are engaged in intensive financing of the real sector of the economy via the banking system. Therefore, we suppose that this unique combination of features should be studied separately. Certainly, articles not corresponding to the inclusion criteria can be found, for example, (Zaitenova & Baibulekova, 2016); however, they are excluded from the study. We have chosen the Republic of Kazakhstan, since this country is a typical representative of post-Soviet countries with a resource-based economy and developed banking system redirecting financial flows to real-economy sectors. The economies and banking systems of Kazakhstan and other central Asian countries (Uzbekistan, Turkmenistan, Tajikistan, and Kyrgyzstan) have many similarities, which appear to be due to historical, political, and geographical factors. All central Asian countries are at the stage of creating market economies (Kaser, 1997; Pomfret, 2021), although in most of them, the state continues to play a significant role, especially in the strategic sectors of the economy. The economies of these countries are resource-based, and they rely heavily on the export of natural resources. This dependence on raw-material exports affects the stability of national currencies and the banking sector (Yadav, 2016). Nevertheless, the development of digital banking services (mobile banking, online payments, e-wallets, fintech startups) is one of the priority tasks in these countries. There is growing interest in digitalization and financial technologies, and countries are actively developing their own electronic solutions (Begimkulov, 2025; Ure, 2021). Kazakhstan is one of the regional leaders in this area (e.g., the KASPI.kz platform).
In these countries, SMEs are actively supported through both banking and non-banking sectors (Prêtet & Klang, 2019); therefore, the microfinance sector is well developed.
Central Asian countries depend on foreign trade and external investments, and their banking systems rely on international reserves, foreign investments, and loans from international organizations (Dosmagambet et al., 2018). Consequently, these countries have currency restrictions or regulations. Central banks play a key role in monetary policy, exchange rate regulation, and supervision. In addition, banking systems in these countries have faced challenges such as poor credit portfolio quality and overdue debt, which require strong regulation (Čihák & Demirgüç-Kunt, 2013).
Kazakhstan stands out in this context with a higher level of economic development, liberalization, and digitalization of the banking sector, making it a relative leader in the region. Our choice of this country is based on these factors.
Another novelty is the approach to testing the indicators. We did not test the direct impact of these indicators, but they were considered as mediators for economic factors. We believe it can show a more complex structure of interconnections of factors when they interplay and change the final effect on NPLs.
One more novelty is the application of the PLS-SEM method for investigating the relationships between variables. The detailed explanation of the method and its advantages are shown in Section 2.2. The employment of this method is a good contribution to other statistical tools applied to this topic study, for example, autoregressive (AR), moving average (MA), and autoregressive integrated moving average (ARIMA) models.
We hope this study will contribute to the topic, and it will be interesting for the scientific community as well as for the practitioners from the industry interested in managing the financial flows, including loans.

2. Methodology

2.1. Literature Review

The goal of the study is to analyze the impact of macro indicators on NPLs. The preliminary literature review is a mandatory part of the study. The obtained results of the literature review are shown in Introduction section of this article. The literature review was completed according the methodology described in several research studies (Butler et al., 2021; Popova & Popovs, 2023; Popova, 2020; Wee & Banister, 2016).
There are certain stages in the literature review. It is necessary to determine the study goal, the corresponding keywords for the literature search, and inclusion and exclusion criteria, which are all part of the first research stage. The following are the requirements for inclusion: only English-language full-text publications that are accessible online and listed in the Scopus database. Items were excluded if they were books, not English-language full-text articles, or book chapters.
Scopus was chosen as a research database since it is one of the top databases with a large number of peer-reviewed social science studies. The wide selection of articles makes it possible to select the full collection of peer-reviewed papers in accordance with the stated objective. Even though Web of Science is a widely used scientific database, it does not offer as many articles in the social sciences as the Scopus database (Falagas et al., 2008; Mongeon & Paul-Hus, 2016; Vujković et al., 2022). The articles in Scopus and the social science research displayed in the Web of Science database interact.
The authors therefore made the decision to exclusively use the Scopus database because it enables the accomplishment of the objectives.
The next step involves using the identified keywords to conduct a search in the Scopus Database using the Boolean approach. The concept of search is demonstrated in Figure 1.
At the first stage, duplicate publications, books, book chapters, articles that are not in English, or that are not accessible online are all eliminated as part of the primary selection process.
The abstracts of the chosen papers are screened for relevance to the research goal in the third stage. Articles that are thought to be unrelated to the research are excluded.
Then, the authors read the full-text articles, excluding the ones not related to the research goal.
Next, the snowballing search was applied. The authors checked the references of the previously selected articles, and from the list of references in the articles, among those that seemed to be the most corresponding to the study goal, we chose articles for further study. These articles also satisfied the set inclusion criteria; however, there was one exception—at this stage, we also included the book chapters and articles not indexed in Scopus.
Then, qualitative analysis was used to analyze the selected articles.

2.2. Research Design

This paper uses Partial Least Squares Structural Equation Modeling (PLS-SEM) to explore the structural relationships between macroeconomic, financial, foreign, and social determinants and the level of non-performing loans (NPLs) in Kazakhstan’s banking sector. PLS-SEM is particularly suited for exploratory modeling with small to medium samples and when the goal is prediction and theory development (Hair et al., 2017).
There are two approaches to structural equation modeling (SEM): partial least squares (PLS) and covariance based (CB) (W. Chin et al., 2020; Dash & Paul, 2021; Hair et al., 2017).
Structural Equation Modeling allows analyzing the relations between latent variables. Each latent variable is composed of indicators—the measurable variables. The obtained complex variables comprising several indicators are tested for reliability and validity (W. Chin et al., 2020; Dash & Paul, 2021; Hair et al., 2017; Ringle et al., 2020).
CB-SEM is traditionally employed for verification, confirming the existing results and supporting the correct selection of factors. It is one of the methods that uses constructs (latent variables) as common factors (Hair et al., 2017). In contrast, PLS uses composite variables comprising the linear combination of measures (indicators) (Hair et al., 2017).
PLS-SEM is traditionally used for a causal-predictive analysis of the variance in the independent constructs. Practically, it is a combination of regression of path and components analysis; it is used for predicting how the specific construct (latent variable) will behave (W. Chin et al., 2020; Hair et al., 2021; Mateos-Aparicio, 2011; Ringle et al., 2020; Schlägel & Sarstedt, 2016). Moreover, PLS-SEM allows applying moderating and mediating factors via the calculation of direct and indirect relationships (Ringle et al., 2020). Therefore, this method is used by many researchers (Onyshchenko et al., 2025; Shujahat et al., 2018; Valaei et al., 2017; Wang et al., 2016).
The authors developed the set of hypotheses focused on proving the relationships between macroeconomic indicators and bad credits. The macroeconomic indicators traditionally used for predicting the NPLs were taken: real GDP, GDP growth rate, and inflation. The usage of these macroeconomic indicators was proven for other world regions (see Table 1), and the authors assume that the same indicators can be taken for the Kazakhstan economy. The same refers to all other indicators used for the formation of latent variables: we formed them using the indicators, and these were investigated by scholars for other world regions.
However, the goal of the article is to determine the impact of macroeconomic factors; we assumed that this impact can be affected by other factors, and we decided to consider the mediation effect of some factors that we considered to be of primary importance for NPLs. We tried to reduce the number of latent variables to avoid multicollinearity and to achieve more evident results.
Then, we have taken separate factors reflecting the foreign influence on the Kazakhstan economy. Kazakhstan is a resource-based economy, and it depends heavily on the foreign sector. Therefore, it is logical to consider foreign factors and their impact on NPLs. Saving behavior is very important for the formation of loans. Therefore, the authors use the latent variable saving factors for determining the mediation effect on NPLs. Then, we assume that NPLs are connected not only to the intentional misbehavior of people; the most evident case of the appearance of NPLs is social factors that impact the individuals’ behavior. Therefore, this latent variable was also included in the study.
The developed hypotheses consider the mediation effect of foreign, saving and social factors.
H1. 
Macroeconomic indicators have a direct impact on NPLs with an inverse relation.
H2. 
Macroeconomic indicators have an impact on NPLs with an inverse relation with the mediation of foreign factors.
H3. 
Macroeconomic indicators have an impact on NPLs with an inverse relation with the mediation of saving factors.
H4. 
Macroeconomic indicators have an impact on NPLs with a direct relation with the mediation of social factors.
The conceptual model, presented in Figure 2, was developed to visualize the hypotheses.
The methodological framework is grounded in the developed conceptual model.

2.3. Data Sources and Indicators

The model is constructed on the basis of five latent variables. The latent variables in their turn use the publicly available macroeconomic data for the Republic of Kazakhstan for 20 years—from 2004 to 2024 (Bureau of National Statistics, n.d.; Ministry of Finance, n.d.; National Bank of Kazakhstan, n.d.). The 20-year period allows enough inputs for statistical analysis, which is the final purpose of the data selection; another criterion of this choice is data availability. The earlier information is not fully available for all the indicators that were used in the study. The data were used as a time series without any exclusion due to positive or negative shocks in the economy of Kazakhstan. The data were standardized before application to analysis (see Appendix A).
The following indicators were initially used for constructing the model (see Table 2).
These indicators were applied to PLS-SEM analysis. After testing the outer model, the indicators with loadings below 0.5 were excluded (Cernisevs et al., 2023; Popova & Zagulova, 2022a). The indicators participating in further analysis are shown in Table 3.
The model was constructed and evaluated in SmartPLS 4.0 software. In the first stage, the outer model was estimated. The quality of the outer model affects the quality of the latent variables. The latent variables meeting all requirements and all estimation criteria allow the researcher to determine further relationships in the structural model with maximum accuracy.
The loadings of all indicators were estimated, and the indicators with loadings below 0.5 were excluded due to not significantly contributing to the model. The obtained model demonstrates very good loadings and p-values. The loadings are above 0.8, demonstrating the highly satisfactory quality of the constructs, and all p-values are less than 0.001 (see Table 2) (Hair et al., 2012, 2019). Then, the model was tested for reliability, convergent validity and discriminant validity. The obtained values were estimated in accordance with the criteria established for the PLS-SEM outer model assessments (Cernisevs et al., 2023; Hair et al., 2014; Henseler et al., 2015; Popova & Zagulova, 2022a, 2022b; Wong, 2013).
The model’s stability and convergence were estimated via the iterations number needed to finish the assessment with the maximum threshold of 300 iterations (Hair et al., 2021; Ringle et al., 2020).
Multicollinearity is measured via the Variation Inflation Factor (VIF). Usually, the threshold set for the models is 3; however, for some macroeconomic indicators, the threshold can be set at the level of 10 (Mooi & Sarstedt, 2011; Popova & Popovs, 2022).
The Average Variance Extracted (AVE) is used to estimate the convergent validity; the criteria of acceptance is having an AVE above 0.5, which means the explanation of more than 50% of the indicators’ variance. The ρA measure is used for internal consistency, which is considered as an “approximately exact reliability measure of the PLS-SEM composites” (Wang et al., 2016). Another criterion, which is supposed to be the most mild and “liberal” one for measuring reliability of the model, is the Composite Reliability (CR) (Garson, 2016). This study has exploratory purposes, and in this type of research, it is common to accept CR as more than or equal to 0.6 (W. W. Chin, 1998; Garson, 2016; Höck & Ringle, 2006). Some authors suppose that the CR should be ≥0.7 (Mateos-Aparicio, 2011) or even ≥0.8 (Daskalakis & Mantas, 2008; Wong, 2013). Usually, such values are used for the confirmation of existing theories. However, in this study, the CR is more than 0.799, so it satisfies the highest criteria. It is highly important to support the uniqueness of the variables. The measurement of Cronbach’s alpha is assumed by some authors to be a “conservative measure” (W. W. Chin, 1998; Garson, 2016; Rahi, 2017; Sanmukhiya, 2020).
Discriminant validity shows the degree of independence of the variables. The strictest criteria applied to the variables is the heterotrait–monotrait ratio of correlations (HTMT) measure. According to (Hair et al., 2019, 2021; Henseler et al., 2015), it should be <0.85 if the variables are different and <0.90 if the variables are closely interrelated. The authors also used the Fornell–Larcker criterion and cross-loadings (Henseler et al., 2015) for measuring discriminant validity, which are more “liberal” and seem to be applicable to this research.
The second step is the estimation of the internal or structural model.
The model tests the direct impact of macroeconomic factors on NPLs as well as the impact of these factors with the mediation of social, saving and foreign factors. Due to the goals of the study, the authors measured the total effect without division on direct and indirect impacts. It was decided that the principal interest is in discovering the impact of macroeconomic factors but not mediators.
The Standardized Path Coefficients (β) traditionally accepted are in the range from −1 to +1; the obtained betas are in this acceptable range.
F2 values (Effect Sizes) can be weak, moderate and strong within the determined paths. Weak are f2 ≥ 0.02, moderate are f2 ≥ 0.15, and strong are f2 ≥ 0.35 (Cohen, 1988). The higher the f2 values, the greater impact the exogenous variable has on the endogenous variable.
The model fit is determined by means of Standardized Root Mean Square Residual (SRMR), which demonstrates the level of discrepancy between the actual and estimated correlation matrix. According to (Henseler et al., 2015), a good fit with low discrepancy will have an SRMR value less than 0.08, demonstrating minimal discrepancy.

3. Results

3.1. Outer Model Evaluation

The first step is estimation of the outer model quality to create the basis for an accurate assessment of the structural model.
The measures of validity and reliability are made on the basis of the Average Variance Extracted (AVE), Composite Reliability (CR) and Cronbach’s alpha (see Table 4).
As we see, all the measures demonstrate the highest level of reliability and validity, and the quality of constructs is high, promising the good quality of the structural model. Even Cronbach’ alpha demonstrates the very high quality of model. All values are significantly above the determined thresholds.
The multicollinearity (VIF) was checked, and the values were not higher than 3.3, which demonstrates the absence of multicollinearity among the indicators. The indicators do not demonstrate a high level of correlation, and it means that the impact of each indicator can be determined separately, increasing the quality of the model and its reliability.
The discriminant validity was measured using all measures available in SmartPLS software. The authors used the Fornell–Larcker criterion and cross-loadings to state that the discriminant validity is satisfactory. Certainly, some indicators are interconnected logically, since they describe the economic development of the country, and they sometimes can show the similar dependencies, as demonstrated in (Popova & Popovs, 2022). These close relationships of variables are reflected by the HTMT ratio (see Table 5).
The effects are moderate and strong, and the impact of exogenous variables on endogenous ones ranges from medium to strong.
Further, it was established that 11 of the 300 iterations allowed finishing the assessment.
Therefore, it is possible to conclude that the outer model is of high quality, while the constructs (latent variables) can provide an accurate estimation of the structural model for testing hypotheses and achieving the set goal.

3.2. Structural (Inner) Model Results

The constructed model is shown in Figure 3.
The model comprises five latent variables:
  • Macroeconomic factors;
  • Foreign factors;
  • Savings factors;
  • Social factors;
  • Non-performing loans (NPLs).
According to the developed conceptual model of the research, the latent variables are used for constructing the following relationships:
  • Direct impact of macroeconomic factors on NPLs;
  • Impact of macroeconomic factors on NPLs with foreign factors variable as a mediator;
  • Impact of macroeconomic factors on NPLs with savings factors variable as a mediator;
  • Impact of macroeconomic factors on NPLs with social factors variable as a mediator.
Therefore, the model corresponds to the set hypotheses and research design.
The yellow rectangles demonstrate the indicators, which participated in the construction of the latent variables. The latent variables are shown as blue circles. The arrows from blue circles to yellow rectangles show the values of loadings of each indicator, and these values should not be lower than 0.5. As we can see, all loadings’ values are very high, demonstrating the strong relationship of indicators with the latent variables and importance of them as a part of latent variables. The values inside the blue circles mean the determination coefficient R2. These values are very high for the saving factor (0.708) and NPLs (0.740), demonstrating that 71% and 74% are explained by the chosen indicator, respectively. The determination coefficient for the latent variable foreign factors is 0.433, which means that 43.3% of variable is explained by the chosen indicators. Nevertheless, R2 is very low for the variable social factors, explaining only 17.4% of the variable.
The structural model was calculated via bootstrapping algorithm application. The values on the arrows show betas. The possible explanation of betas will be provided in the Discussion section.
The constructed model allows testing the set hypotheses. The relationships within the model were estimated, and these are presented in Table 6.
The structural model demonstrated several significant and policy-relevant relationships (see Table 7).
As we see, macroeconomic factors have strong negative relationships with all of the independent variables. However, theses relationship in this paper are not important, since the goal of the study does not include the determination of these relations, and they only serve for calculating the total effect on NPLs.
What is interesting for this study is the relations between macroeconomic factors and NPLs directly and with the mediating role of other latent variables.
As a result, two hypotheses out of the four were not rejected, while two were fully rejected. The main findings of the study are presented in Table 8.

4. Discussion

The authors propose the discussion of the research results, based on the developed PLS SEM model, presented within this paper to the scientific community. The research results show that Kazakhstan’s economy behaves like other economies in relation to the macroeconomics influence on the NPLs. The development of the national economy has a significant inverse relationship with NPLs: β = −0.800, T-statistics 9.729, p value < 0.001. Despite the fact that the dependencies between Kazakhstan’s economy and NPLs was not widely assessed within the academic articles, which was one of the motivations to conduct this research, we assume that in relation to the macroeconomic influence on NPLs, it is possible to rely on the results of other countries’ assessments by the academic society.
In addition to the above, the research shows the negative relations of macroeconomic factors on NPLs with the mediation of foreign factors (β = −4.088, T-statistics 2.524, p value 0.012). That proves that foreign factors work as a mediator and enforce the effect of the macroeconomic factors on NPLs. The mediator increases the impact of the independent variable. The amount of NPLs increases dramatically when economic growth is coupled with a rise in the proportion of foreign currency loans or a decline in the value of the national currency.
On the other hand, unsupported hypotheses are manifested differently. The research shows that the impact of macroeconomic factors on NPLs with saving factors as a mediating variable is positive, as the total effect results show, but this impact is insignificant. The other results of the research show that saving factors’ impact on NPLs is significant, but when they are used as a mediator for the macroeconomic factors, the significance of the effect disappears.
The authors propose to the academic society to discuss the saving and social factors’ neutralizing effect in relation to the macroeconomic factors and non-performing loans. We would like to propose our explanation of this effect.
Economic growth usually leads to inflation, which in a stable situation is not more than 3–5%, and to an increase in expenditures of households and businesses, while their savings do not rise significantly. At the same time, the financial system’s capacity to lend money may be limited in a period of rapid growth of the economy, which naturally increases the price of lending—the interest rate. In this situation, the demand for credit decreases and citizens’ habits switch to savings, increasing household and business deposits in commercial banks.
In view of the above, we may assume that saving factors can offset or weaken the effect of economic growth on NPLs; despite the fact that economic growth usually reduces NPLs, the increase in savings and the decrease in borrowing may reduce the negative effect.
Similarly, social factors like the Gini index and quintile coefficient, representing income inequality, may act as a balancing force. Nevertheless, the growing economy and the increase in income inequality may reduce the ability of a large group of citizens to cover their loans. This may create pressure in the opposite direction, which potentially may increase NPLs and weaken the improving beneficial effect of macroeconomic factors.
Therefore, the authors propose as the explanation, that savings and social factors in Kazakhstan may have neutralizing influences due to a switch of the households’ and businesses’ habits from borrowing to savings and an increase in income inequality.
It is very important to understand the mediation effects of various factors on NPLs. The policy-makers traditionally try to provide resilience for the financial sector using financial tools only. However, this paper, in line with the studies of other scholars, shows that it is absolutely necessary to consider other factors related to the national economy, and mediation effects allow changing the situation significantly without having any direct impact on the financial sector. It provides the governments of developing countries with certain flexibility regarding the resilience of the financial sector.

4.1. Implications

The obtained results have serious implications on practice. They can be applied to regulating the processes within the financial system of Kazakhstan. The obtained results are significant for policy-makers, financial regulators, and banking institutions in Kazakhstan and potentially in similar economies. According to the research findings, macroeconomic stability reduces NPLs, and any improvements in macroeconomic conditions lead to a decrease in NPLs. Therefore, the focus on maintaining macroeconomic stability can serve as a key strategy to improve the health of the financial system. A stable economic system can reduce credit risk and enhance loan repayment capacity. Policy-makers in Kazakhstan can use these findings to prioritize sustained economic growth and macroeconomic discipline as tools to reduce credit risk in the banking sector.
The external (foreign) factors like exchange rate stability and public debt levels amplify the impact of macroeconomic factors on NPLs due to the intensification of inverse relationships between macroeconomic indicators and NPLs. The level of non-performing loans (NPLs) rises significantly when economic expansion coincides with an increase in foreign currency-denominated loans or a depreciation of the local currency. Therefore, the financial regulators and banks can mitigate the risks within the national financial system by monitoring and managing the exposure to foreign economic shocks. Managing external vulnerabilities, for example, minimizing excessive foreign currency lending and maintaining a balanced public debt profile, can further mitigate the risk of rising NPLs.
Savings behavior, such as households’ deposits or accounts balance, does not mediate the relationship between macroeconomic indicators and NPLs. Moreover, the inverse effect of macroeconomic factors on NPLs is levelled off by these saving factors. Therefore, the policy-makers should keep in mind that while promoting savings is important for financial stability, it may not directly buffer the impact of macroeconomic downturns on NPLs. Therefore, policies focused on savings alone may be insufficient to reduce credit risk during economic slumps. In this way, the rejected hypotheses are even more important for understanding how to manage the simultaneous stability of economic growth and saving factors.
Social factors (e.g., unemployment, income inequality, etc.) do not significantly mediate the relationship between macroeconomic conditions and NPLs when considered alongside macroeconomic variables. A growing economy combined with rising income inequality can undermine the ability of a significant portion of the population to meet their loan obligations. This could create a counteracting pressure that offsets the positive impact of macroeconomic improvements, potentially leading to higher NPLs and weakening the overall beneficial effect of favorable macroeconomic conditions. Therefore, the policy-makers should consider not only development of national economy as a means of stabilizing the financial system but also think about indicators of income inequality and poverty within the country.
These results suggest that macroeconomic indicators should be included in Kazakhstan’s policies aimed at maintaining financial stability. Regulators could improve their stress testing models by considering how sensitive the financial system is to foreign factors. They should also consider stricter rules for foreign currency loans, especially when the tenge is losing value. In addition, tools like countercyclical capital buffers can be adjusted based on how fast the economy is growing and how exposed it is to external risks. This would help banks handle financial shocks better and support long-term stability in the lending market.
International best practices and regulatory models can be applied to Kazakhstan with confidence, and the country can be included in broader comparative analyses or regional policy frameworks.

4.2. Areas for Further Research

The study opens up questions about the role of saving and social factors in credit risk dynamics. Future research could explore non-linear relationships, time-lagged effects, or the sector-specific impacts of these factors. Additionally, qualitative studies could help uncover deeper mechanisms behind these statistical relationships.

4.3. Limitations of the Research

The research limitations are determined by the type of investigation.
First of all, the authors consider Kazakhstan as a representative of post-soviet central Asian countries; however, it does not consider other post-soviet central Asian countries—Uzbekistan, Turkmenistan, Tajikistan, or Kyrgyzstan. The choice of the same variables but for another country can change the results of the study.
The choice of latent variables and indicators for the analysis was restricted purposefully; however, it is also a limitation of the study. The choice of other variables can change the results of the study.
The type of implemented analysis—PLS-SEM—predetermines the certain statistical analyses; however, it could be possible to employ other types of statistical analyses.

5. Conclusions

The article is devoted to the relationship between various macro indicators and NPLs. The authors do not use isolated indicators but rather combine them in complex latent variables, which then participate in PLS-SEM.

Hypotheses Testing

H1, “Macroeconomic indicators have a direct impact on NPLs with an inverse relation,” was confirmed. Any increase in the national economy results in a decrease in NPLs. This finding is similar to other researches devoted to this topic, although we used complex latent variables containing several macro indicators.
The test of the direct impact of latent variables (foreign factors, social factors and saving factors) on NPLs was beyond the framework of this study; however, in the process of testing other hypotheses, we obtained results which allow concluding that the Kazakh financial system has completely the same dependencies as other countries, and it can be considered not individually but rather in line with other countries.
H2, “Macroeconomic indicators have an impact on NPLs with an inverse relation with the mediation of foreign factors,” was confirmed. The mediating effect of foreign factors strengthens the negative relationships between macro factors and NPLs.
H3, “Macroeconomic indicators have an impact on NPLs with an inverse relation with the mediation of saving factors,” was not confirmed. The effect of saving factors levels off, balancing the inverse impact of macro factors. Separately, these variables have significant relations with NPLs, but when they act together, the total effect is insignificant.
H4, “Macroeconomic indicators have an impact on NPLs with a direct relation with the mediation of social factors,” was not confirmed, either. The situation is similar to the H3 situation—separately, the impacts are significant, but together, the mediator makes impact of macro factors insignificant.
The findings of the study are quite interesting for scientific discussion, since they allow various speculations. The authors have proposed their interpretations of results as a point for such discussion. Moreover, the evident implications of the research results will contribute to the development of the banking sector in Kazakhstan with the simultaneous stabilization of the situation within the industry.

Author Contributions

Conceptualization, A.K. and O.C.; methodology, Y.P.; software, S.P.; validation, A.K., Y.P. and O.C.; formal analysis, S.P.; investigation, A.K.; resources, A.K.; data curation, Y.P.; writing—original draft preparation, A.K.; writing—review and editing, O.C.; visualization, S.P.; supervision, Y.P.; project administration, A.K. All authors have read and agreed to the published version of the manuscript.

Funding

This research received no external funding.

Institutional Review Board Statement

Not applicable.

Informed Consent Statement

Not applicable.

Data Availability Statement

The data are available in Appendix A.

Conflicts of Interest

Author A.K. was employed by the company JSC “The Fund of Problem Loans”. Author O.C. was employed by the company SIA StarBridge. The remaining authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.

Appendix A

Table A1. Sample of data for PLAS-SEM analysis (the full set of data is available under request).
Table A1. Sample of data for PLAS-SEM analysis (the full set of data is available under request).
YearGDP (USD Billion)GDP Growth (%)Inflation Rate (%)Unemployment Rate (%)Exchange Rate (KZT/USD)Current Account Balance (% of GDP)Public Debt (% of GDP)Quintile CoefficientGini IndexTotal Credits (USD Billion)Household Deposits (USD Billion)Total Deposits (USD Billion)Bad Credits—Private Individuals (USD Billion)Bad Credits—Legal Companies (USD Billion)
200426.39.66.88.41362.212.34.232.70.85260.2
200532.99.77.58.1133.82.111.94.132.50.96.22.57.20.3
200644.39.68.67.8126.12.810.5432.317.83.18.50.4
200758.48.510.87.3122.55.28.73.9321.29.13.8100.5
200870.23.218.66.6120.33.49.13.831.81.4114.211.50.6
200975.91.27.36.3147.5−3.58.53.831.51.813.5513.80.7
201093.97.36.25.8147.3−2.98.83.931.22.115.85.8160.9
2011149.17.58.35.4145.5−5.49431320.26.518.51.2
20122024.85.15.3149.1−2.811431.54.5247.8201.5
2013236.664.85.1153.6−2.113.24.131.84.826.58.522.51.6
2014221.44.25.75188.9−3.115.14.132.14.2299.224.81.4
2015184.41.27.85222.2−6.5174.232.53.731.51027.21.3
2016137.34.113.65.1342.2−3.419.84.332.83.53210.828.51.2
2017160.84.2114.9326−0.121.34.3333.23311.2301.1
2018179.34.164.8344.7−0.520.94.233.22.834.51231.81
2019181.74.55.34.8382.6−2.821.54.233.52.5361333.50.9
2020171−2.56.85412−3.623.44.333.82.33814.535.20.85
2021197484.9421.12.5244.334240.515.5370.8
2022220.63.5144.8447.73.425.24.434.21.9421638.50.75
20232374.2134.8460.54.126.54.434.51.743.517400.7

References

  1. Ahiase, G., Andriana, D., Agbemava, E., & Adonai, B. (2023). Macroeconomic cyclical indicators and bank non-performing loans: Does country governance matter in African countries? International Journal of Social Economics, 51(1), 62–80. [Google Scholar] [CrossRef]
  2. Anastasiou, D., Bragoudakis, Z., & Malandrakis, I. (2019). Non-performing loans, governance indicators and systemic liquidity risk: Evidence from Greece. International Journal of Finance & Economics, 24, 97–112. [Google Scholar] [CrossRef]
  3. Anita, S. S., Tasnova, N., & Nawar, N. (2022). Are non-performing loans sensitive to macroeconomic determinants? An empirical evidence from banking sector of SAARC countries. Future Business Journal, 8, 7. [Google Scholar] [CrossRef]
  4. Arham, N., Salisi, M. S., Mohammed, R. U., & Tuyon, J. (2020). Impact of macroeconomic cyclical indicators and country governance on bank non-performing loans in Emerging Asia. Eurasian Economic Review, 10(4), 707–726. [Google Scholar] [CrossRef]
  5. Ariffin, N. M. (2012). Liquidity risk management and financial performance in Malaysia: Empirical evidence from islamic banks. Aceh International Journal of Social Sciences, 1(2), 26309. [Google Scholar] [CrossRef]
  6. Athari, S. A. (2021a). Domestic political risk, global economic policy uncertainty, and banks’ profitability: Evidence from Ukrainian banks. Post-Communist Economies, 33, 458–483. [Google Scholar] [CrossRef]
  7. Athari, S. A. (2021b). The effects of institutional settings and risks on bank dividend policy in an emerging market: Evidence from Tobit model. International Journal of Finance & Economics, 26, 4493–4515. [Google Scholar] [CrossRef]
  8. Ayhan, F., & Kartal, M. T. (2021). The macro economic drivers of non-performing loans (NPL): Evidence from selected countries with heterogeneous panel analysis. MANAS Sosyal Araştırmalar Dergisi, 10(2), 986–999. [Google Scholar] [CrossRef]
  9. Begimkulov, E. (2025). Bank digitalization and financial stability in Central Asia: Assessing risk and resilience. Journal of Eastern European and Central Asian Research (JEECAR), 12(1), 1. [Google Scholar] [CrossRef]
  10. Boudriga, A., Boulila Taktak, N., & Jellouli, S. (2009). Banking supervision and nonperforming loans: A cross-country analysis. Journal of Financial Economic Policy, 1, 286–318. [Google Scholar] [CrossRef]
  11. Boussaada, R., Hakimi, A., & Karmani, M. (2020). Is there a threshold effect in the liquidity risk–non-performing loans relationship? A PSTR approach for MENA banks. International Journal of Finance & Economics, 27(2), 1886–1898. [Google Scholar] [CrossRef]
  12. Bucevska, V., & Hadzi Misheva, B. (2017). The determinants of profitability in the banking industry: Empirical research on selected Balkan countries. Eastern European Economics, 55, 146–167. [Google Scholar] [CrossRef]
  13. Bureau of National Statistics. (n.d.). Agency for strategic planning and reforms. Available online: https://stat.gov.kz/en/standard/national/ (accessed on 21 May 2025).
  14. Butler, L., Yigitcanlar, T., & Paz, A. (2021). Barriers and risks of Mobility-as-a-Service (MaaS) adoption in cities: A systematic review of the literature. Cities, 109, 103036. [Google Scholar] [CrossRef]
  15. Cernisevs, O., & Popova, Y. (2022). ICO as Crypto-Assets Manufacturing within a Smart City. Smart Cities, 6(1), 40–56. [Google Scholar] [CrossRef]
  16. Cernisevs, O., Popova, Y., & Cernisevs, D. (2023). Business KPIs based on compliance risk estimation. Journal of Tourism and Services, 14(27), 222–248. [Google Scholar] [CrossRef]
  17. Chin, W., Cheah, J.-H., Liu, Y., Ting, H., Lim, X.-J., & Cham, T. H. (2020). Demystifying the role of causal-predictive modeling using partial least squares structural equation modeling in information systems research. Industrial Management & Data Systems, 120(12), 2161–2209. [Google Scholar] [CrossRef]
  18. Chin, W. W. (1998). The partial least squares approach for structural equation modeling (G. A. Macoulides, Ed.; pp. 295–336). Lawrence Erlbaum Associates. [Google Scholar]
  19. Climent-Serrano, S. (2019). Effects of economic variables on NPLs depending on the economic cycle. Empirical Economics, 56(1), 325–340. [Google Scholar] [CrossRef]
  20. Cohen, J. (1988). Statistical power analysis for the behavioral sciences. Lawrence Erlbaum. [Google Scholar]
  21. Čihák, M., & Demirgüç-Kunt, A. (2013). Rethinking the state’s role in finance. Available online: https://papers.ssrn.com/abstract=2245477 (accessed on 16 May 2025).
  22. Dash, G., & Paul, J. (2021). CB-SEM vs PLS-SEM methods for research in social sciences and technology forecasting. Technological Forecasting and Social Change, 173, 121092. [Google Scholar] [CrossRef]
  23. Daskalakis, S., & Mantas, J. (2008). Evaluating the impact of a service-oriented framework for healthcare interoperability (pp. 285–290). IOS Press. [Google Scholar]
  24. Dibar, O. Y., & Sridevi, M. S. (2021). Investigation on macroeconomic determinants of non-performing loans: The case of ethiopian commercial banks. Available online: https://journal-innovations.com/assets/uploads/doc/10df4-1582-1595.23667.pdf (accessed on 16 May 2025).
  25. Dosmagambet, Y., Oskenbayev, Y., Taghizadeh-Hesary, F., & Mukan, M. (2018). Financial system, and SME financing in Kazakhstan (Working Paper No. 904). ADBI Working Paper Series. ADBI. Available online: https://www.econstor.eu/handle/10419/222671 (accessed on 1 June 2025).
  26. Fakhrunnas, F., Nugrohowati, R. N. I., Haron, R., & Anto, M. B. H. (2022). The determinants of non-performing loans in the indonesian banking industry: An asymmetric approach before and during the pandemic crisis. SAGE Open, 12, 21582440221102421. [Google Scholar] [CrossRef]
  27. Falagas, M. E., Pitsouni, E. I., Malietzis, G. A., & Pappas, G. (2008). Comparison of PubMed, Scopus, Web of Science, and Google Scholar: Strengths and weaknesses. The FASEB Journal, 22(2), 338–342. [Google Scholar] [CrossRef]
  28. Foglia, M. (2022). Non-performing loans and macroeconomics factors: The Italian case. Risks, 10, 21. [Google Scholar] [CrossRef]
  29. Garson, D. (2016). Partial least squares (PLS-SEM): Regression & structural equation models. North Carolina State University. [Google Scholar]
  30. Gjeci, A., & Marinc, M. (2018). Corruption and non-performing loans. SSRN Electronic Journal, 24(4). [Google Scholar] [CrossRef]
  31. Hair, J. F., Hollingsworth, C. L., Randolph, A. B., & Chong, A. Y. L. (2017). An updated and expanded assessment of PLS-SEM in information systems research. Industrial Management & Data Systems, 117(3), 442–458. [Google Scholar] [CrossRef]
  32. Hair, J. F., Hult, G. T. M., Ringle, C. M., Sarstedt, M., Danks, N. P., & Ray, S. (2021). An Introduction to Structural Equation Modeling. In J. F. Hair, G. T. M. Hult, C. M. Ringle, M. Sarstedt, N. P. Danks, & S. Ray (Eds.), Partial least squares structural equation modeling (PLS-SEM) using R (pp. 1–29). Springer International Publishing. [Google Scholar] [CrossRef]
  33. Hair, J. F., Risher, J. J., Sarstedt, M., & Ringle, C. M. (2019). When to use and how to report the results of PLS-SEM. European Business Review, 31(1), 2–24. [Google Scholar] [CrossRef]
  34. Hair, J. F., Sarstedt, M., Hopkins, L., & Kuppelwieser, V. G. (2014). Partial least squares structural equation modeling (PLS-SEM): An emerging tool in business research. European Business Review, 26(2), 106–121. [Google Scholar] [CrossRef]
  35. Hair, J. F., Sarstedt, M., Ringle, C. M., & Mena, J. A. (2012). An assessment of the use of partial least squares structural equation modeling in marketing research. Journal of the Academy of Marketing Science, 40(3), 414–433. [Google Scholar] [CrossRef]
  36. Hakimi, A., Boussaada, R., & Karmani, M. (2020). Is the relationship between corruption, government stability and non-performing loans non-linear? A threshold analysis for the MENA region. International Journal of Finance & Economics, 27(4), 4383–4398. [Google Scholar] [CrossRef]
  37. Henseler, J., Ringle, C. M., & Sarstedt, M. (2015). A new criterion for assessing discriminant validity in variance-based structural equation modeling. Journal of the Academy of Marketing Science, 43(1), 115–135. [Google Scholar] [CrossRef]
  38. Höck, M., & Ringle, C. M. (2006, September 28–30). Strategic Networks in the software industry: An empirical analysis of the value continuum. IFSAM VIIIth World Congress, Berlin, Germany. [Google Scholar]
  39. Irani, F., Athari, S. A., & Hadood, A. A. A. (2022). The impacts of country risk, global economic policy uncertainty, and macroeconomic factors on the Turkish tourism industry. International Journal of Hospitality & Tourism Administration, 23, 1242–1265. [Google Scholar]
  40. Jenkins, H., Alshareef, E., & Mohamad, A. (2021). The impact of corruption on commercial bank’ credit risk: Evidence from a panel quantile regression. International Journal of Finance & Economics, 28(2), 1364–1375. [Google Scholar] [CrossRef]
  41. Karadima, M., & Louri, H. (2021). Economic policy uncertainty and non-performing loans: The moderating role of bank concentration. Finance Research Letters, 38, 101458. [Google Scholar] [CrossRef]
  42. Kartal, M. T., Ayhan, F., & Altaylar, M. (2023). The impacts of financial and macroeconomic factors on financial stability in emerging countries: Evidence from Turkey’s nonperforming loans. Journal of Risk, 25, 25–48. Available online: https://www.risk.net/journal-of-risk/7955870/the-impacts-of-financial-and-macroeconomic-factors-on-financial-stability-in-emerging-countries-evidence-from-turkeys-nonperforming-loans (accessed on 16 May 2025). [CrossRef]
  43. Kartikasary, M., Marsintauli, F., Serlawati, E., & Laurens, S. (2020). Factors affecting the non-performing loans in Indonesia. Accounting, 6, 97–106. [Google Scholar] [CrossRef]
  44. Kaser, M. (1997). Economic transition in six Central Asian economies. Central Asian Survey, 16(1), 5–26. [Google Scholar] [CrossRef]
  45. Khan, M. A., Siddique, A., & Sarwar, Z. (2020). Determinants of non-performing loans in the banking sector in developing state. Asian Journal of Accounting Research, 5(1), 135–145. [Google Scholar] [CrossRef]
  46. Krasniqi-Pervetica, A., & Ahmeti, S. (2022). The effect of macroeconomic indicators on non-performing loans: The case of balkan countries. International Journal of Applied Economics, Finance and Accounting, 14(1), 42–49. [Google Scholar] [CrossRef]
  47. Kumar, V., & Kishore, M. P. (2019). Macroeconomic and bank specific determinants of non-performing loans in UAE conventional bank. Journal of Banking and Finance Management, 2(1), 1–12. [Google Scholar] [CrossRef]
  48. Kuzucu, N., & Kuzucu, S. (2019). What drives non-performing loans? Evidence from emerging and advanced economies during pre-and post-global financial crisis. Emerging Markets Finance and Trade, 55(8), 1694–1708. [Google Scholar] [CrossRef]
  49. Mance, D., Žiković, S., & Mance, D. (2015). Econometric analysis of croatia’s proclaimed foreign exchange rate. The South East European Journal of Economics and Business, 10(1), 7–17. [Google Scholar] [CrossRef]
  50. Mateos-Aparicio, G. (2011). Partial least squares (PLS) methods: Origins, evolution, and application to social sciences. Communications in Statistics—Theory and Methods, 40(13), 2305–2317. [Google Scholar] [CrossRef]
  51. Mehmood, A., Hidthiir, M. H., & Mat Nor, A. (2019). A conceptual paper for macroeconomic determinants of non-performing loans (NPLs) in banking sector of Pakistan. Asian Journal of Multidisciplinary Studies, 7(3), 6–15. [Google Scholar]
  52. Melnyk, O., Onishchenko, O., Fomin, O., Yarovenko, V., Pavlova, N., Drozhzhyn, O., & Vishnevsky, D. (2025). Smart port concept: A case study of port-integrated autonomous networks. Journal of Maritime Research, 22(1), 315–323. [Google Scholar] [CrossRef]
  53. Menicucci, E., & Paolucci, G. (2021). Gender diversity and bank risk-taking: An empirical investigation in Italy. Corporate Governance: The International Journal of Business in Society, 22, 317–339. [Google Scholar] [CrossRef]
  54. Ministry of Finance. (n.d.). Available online: https://www.gov.kz/memleket/entities/minfin?lang=en (accessed on 16 May 2025).
  55. Mohamad, A., & Jenkins, H. (2021). Corruption and banks’ non-performing loans: Empirical evidence from MENA countries. Macroeconomics and Finance in Emerging Market Economies, 14(3), 308–321. [Google Scholar] [CrossRef]
  56. Mongeon, P., & Paul-Hus, A. (2016). The journal coverage of Web of Science and Scopus: A comparative analysis. Scientometrics, 106(1), 213–228. [Google Scholar] [CrossRef]
  57. Mooi, E., & Sarstedt, M. (2011). A concise guide to market research. Springer. [Google Scholar] [CrossRef]
  58. Morakinyo, A. E., & Sibanda, M. (2016). The determinants of non-performing loans in the MINT economies. Journal of Economics and Behavioral Studies, 8(5), 39–55. [Google Scholar] [CrossRef]
  59. Nadham, V., & Nahid, B. (2015). Determinants of non-performing loans in commercial banks: A study of NBC Bank Dodoma Tanzania. International Journal of Finance & Banking Studies, 4(1), 70. [Google Scholar] [CrossRef]
  60. National Bank of Kazakhstan. (n.d.). Available online: https://www.centralbanking.com/directory (accessed on 1 June 2025).
  61. Onyshchenko, S., Bychkovsky, Y., Melnyk, O., Onishchenko, O., Zaporozhets, A., & Bedrii, D. (2025). Integrating human factors in project-oriented risk management for maritime safety. In O. Melnyk, O. Onishchenko, & A. Zaporozhets (Eds.), Maritime systems, transport and logistics I (Vol. 580, pp. 183–195). Springer Nature. [Google Scholar] [CrossRef]
  62. Pomfret, R. (2021). Central Asian economies: Thirty years after dissolution of the soviet union. Comparative Economic Studies, 63(4), 537–556. [Google Scholar] [CrossRef] [PubMed]
  63. Popova, Y. (2020). Economic or financial substantiation for smart city solutions: A literature study. Economic Annals-XXI, 183(5–6), 125–133. [Google Scholar] [CrossRef]
  64. Popova, Y., & Cernisevs, O. (2023). Smart city: Sharing of financial services. Social Sciences, 12(1), 8. [Google Scholar] [CrossRef]
  65. Popova, Y., Cernisevs, O., & Popovs, S. (2024). Impact of geographic location on risks of fintech as a representative of financial institutions. Geographies, 4(4), 753–768. [Google Scholar] [CrossRef]
  66. Popova, Y., Cernisevs, O., & Popovs, S. (2025). Contribution to modern economic region theory: Factor of intangible digital resources. Geographies, 5(1), 8. [Google Scholar] [CrossRef]
  67. Popova, Y., & Popovs, S. (2022). Impact of smart economy on smart areas and mediation effect of national economy. Sustainability, 14(5), 2789. [Google Scholar] [CrossRef]
  68. Popova, Y., & Popovs, S. (2023). Effects and externalities of smart governance. Smart Cities, 6(2), 1109–1131. [Google Scholar] [CrossRef]
  69. Popova, Y., & Zagulova, D. (2022a). Aspects of E-scooter sharing in the smart city. Informatics, 9(2), 36. [Google Scholar] [CrossRef]
  70. Popova, Y., & Zagulova, D. (2022b). UTAUT model for smart city concept implementation: Use of web applications by residents for everyday operations. Informatics, 9(1), 27. [Google Scholar] [CrossRef]
  71. Prêtet, A., & Klang, T. (2019). Enhancing access to finance for SMEs: The case of Central Asia. In Unlocking SME finance in Asia. Routledge. [Google Scholar]
  72. Rahi, S. (2017). Structural equation modeling using SmartPLS. Available online: https://www.researchgate.net/publication/328560499_Structural_Equation_Modeling_Using_SmartPLS (accessed on 16 May 2025).
  73. Rehmana, A., Adzisb, A. A., & Mohamed-Arshadc, S. B. (2020). The relationship between corruption and credit risk in commercial banks of Pakistan. International Journal of Innovation, Creativity and Change, 11(1), 701–715. [Google Scholar]
  74. Ringle, C. M., Sarstedt, M., Mitchell, R., & Gudergan, S. P. (2020). Partial least squares structural equation modeling in HRM research. The International Journal of Human Resource Management, 31(12), 1617–1643. [Google Scholar] [CrossRef]
  75. Saksonova, S., & Papiashvili, T. (2021). Micro and small businesses access to finance and financial literacy of their owners: Evidence from Latvia, Estonia and Georgia. In I. Kabashkin, I. Yatskiv, & O. Prentkovskis (Eds.), Reliability and statistics in transportation and communication (pp. 667–677). Springer International Publishing. [Google Scholar] [CrossRef]
  76. Saksonova, S., & Vilerts, K. (2015). Measuring Returns to Education: The Case of Latvia. Analele Stiintifice Ale Universitatii Al I Cuza Din Iasi—Sectiunea Stiinte Economice, 62, 251–261. [Google Scholar] [CrossRef]
  77. Saliba, C., Farmanesh, P., & Athari, S. A. (2023). Does country risk impact the banking sectors’ non-performing loans? Evidence from BRICS emerging economies. Financial Innovation, 9(1), 86. [Google Scholar] [CrossRef]
  78. Sanmukhiya, C. A. (2020). PLS–SEM approach to the Utaut model: The case of Mauritius. Annals of Social Sciences & Management Studies, 6(1), 101–115. [Google Scholar]
  79. Schlägel, C., & Sarstedt, M. (2016). Assessing the measurement invariance of the four-dimensional cultural intelligence scale across countries: A composite model approach. European Management Journal, 34(6), 633–649. [Google Scholar] [CrossRef]
  80. Shujahat, M., Ali, B., Nawaz, F., Durst, S., & Kianto, A. (2018). Translating the impact of knowledge management into knowledge-based innovation: The neglected and mediating role of knowledge-worker satisfaction. Human Factors and Ergonomics in Manufacturing & Service Industries, 28(4), 200–212. [Google Scholar] [CrossRef]
  81. Staehr, K., & Uusküla, L. (2020). Macroeconomic and macro-financial factors as leading indicators of non-performing loans: Evidence from the EU countries. Journal of Economic Studies, 48(3), 720–740. [Google Scholar] [CrossRef]
  82. Syed, A., & Tripathi, R. (2020). Macroeconomic vulnerabilities and its effect on nonperforming loans in indian commercial banks. Indian Journal of Finance, 14(2), 34–49. [Google Scholar] [CrossRef]
  83. Tanasković, S., & Jandrić, M. (2015). Macroeconomic and Institutional Determinants of Non-performing Loans. Journal of Central Banking Theory and Practice, 4(1), 47–62. [Google Scholar] [CrossRef]
  84. Umar, M., & Sun, G. (2016). Non-performing loans (NPLs), liquidity creation, and moral hazard: Case of Chinese banks. China Finance and Economic Review, 5(3), 51–75. [Google Scholar] [CrossRef]
  85. Ure, J. (2021). Digital solutions centre in central Asia. Available online: https://repository.unescap.org/items/ad6a85b8-265a-40b6-9aee-8d2103ac0325 (accessed on 16 May 2025).
  86. Valaei, N., Nikhashemi, S. R., & Javan, N. (2017). Organizational factors and process capabilities in a KM strategy: Toward a unified theory. Journal of Management Development, 36(4), 560–580. [Google Scholar] [CrossRef]
  87. Vouldis, A. T., & Louzis, D. P. (2018). Leading indicators of non-performing loans in Greece: The information content of macro-, micro- and bank-specific variables. Empirical Economics, 54(3), 1187–1214. [Google Scholar] [CrossRef]
  88. Vujković, P., Ravšelj, D., Umek, L., & Aristovnik, A. (2022). Bibliometric analysis of smart public governance research: Smart city and smart government in comparative perspective. Social Sciences, 11(7), 293. [Google Scholar] [CrossRef]
  89. Wang, Z., Sharma, P. N., & Cao, J. (2016). From knowledge sharing to firm performance: A predictive model comparison. Journal of Business Research, 69(10), 4650–4658. [Google Scholar] [CrossRef]
  90. Wee, B. V., & Banister, D. (2016). How to write a literature review paper? Transport Reviews, 36(2), 278–288. [Google Scholar] [CrossRef]
  91. Wong, K. K. K. (2013). Partial least squares structural equation modeling (PLS-SEM) techniques using SmartPLS. Marketing Bulletin, 24(1), 1–32. [Google Scholar]
  92. Yadav, R. (2016). Economic transformation in central Asia: A journey of twenty-five years. International Studies, 53(3–4), 286–304. [Google Scholar] [CrossRef]
  93. Zaitenova, N. K., & Baibulekova, L. A. (2016). Evaluation of influence of macroeconomic shocks on the banking sector of Kazakhstan. International Journal of Economics and Financial Issues, 6(2S), 188–194. [Google Scholar]
Figure 1. The conceptual model of the literature search in the Scopus database. Source: generated by the authors.
Figure 1. The conceptual model of the literature search in the Scopus database. Source: generated by the authors.
Jrfm 18 00431 g001
Figure 2. Conceptual model of the study. Source: generated by the authors.
Figure 2. Conceptual model of the study. Source: generated by the authors.
Jrfm 18 00431 g002
Figure 3. Structural model. Source: generated by the authors on the basis of SmartPLS 4 software.
Figure 3. Structural model. Source: generated by the authors on the basis of SmartPLS 4 software.
Jrfm 18 00431 g003
Table 1. Results of literature review on the topic. Source: generated by the authors.
Table 1. Results of literature review on the topic. Source: generated by the authors.
IndicatorResources
Real GDP(Anita et al., 2022; Ayhan & Kartal, 2021; Foglia, 2022; Hakimi et al., 2020; Karadima & Louri, 2021; Mance et al., 2015; Rehmana et al., 2020; Umar & Sun, 2016; Arham et al., 2020; Krasniqi-Pervetica & Ahmeti, 2022; Mehmood et al., 2019)
Industrial Production Index(Mance et al., 2015; Vouldis & Louzis, 2018)
Unemployment Rate(Anastasiou et al., 2019; Foglia, 2022; Khan et al., 2020; Kuzucu & Kuzucu, 2019; Mance et al., 2015; Nadham & Nahid, 2015; Ahiase et al., 2023; Arham et al., 2020; Kartal et al., 2023; Krasniqi-Pervetica & Ahmeti, 2022; Mehmood et al., 2019)
Inflation(Anita et al., 2022; Boussaada et al., 2020; Fakhrunnas et al., 2022; Hakimi et al., 2020; Karadima & Louri, 2021; Khan et al., 2020; Kuzucu & Kuzucu, 2019; Mance et al., 2015; Nadham & Nahid, 2015; Rehmana et al., 2020; Syed & Tripathi, 2020; Tanasković & Jandrić, 2015; Ahiase et al., 2023; Arham et al., 2020; Krasniqi-Pervetica & Ahmeti, 2022)
Interest Rate(Fakhrunnas et al., 2022; Mance et al., 2015; Nadham & Nahid, 2015; Rehmana et al., 2020; Tanasković & Jandrić, 2015; Arham et al., 2020; Mehmood et al., 2019)
GDP Growth(Anastasiou et al., 2019; Fakhrunnas et al., 2022; Khan et al., 2020; Kuzucu & Kuzucu, 2019; Nadham & Nahid, 2015; Syed & Tripathi, 2020; Tanasković & Jandrić, 2015) (Ahiase et al., 2023; Kartal et al., 2023; Staehr & Uusküla, 2020)
Lending Activities(Nadham & Nahid, 2015)
Capital Adequacy(Gjeci & Marinc, 2018; Morakinyo & Sibanda, 2016)
Profitability(Anastasiou et al., 2019; Karadima & Louri, 2021; Kartikasary et al., 2020; Kumar & Kishore, 2019; Morakinyo & Sibanda, 2016; Rehmana et al., 2020; Umar & Sun, 2016)
Liquidity(Anastasiou et al., 2019; Boussaada et al., 2020; Morakinyo & Sibanda, 2016; Umar & Sun, 2016)
Bank Total Credit(Ayhan & Kartal, 2021; Kuzucu & Kuzucu, 2019; Morakinyo & Sibanda, 2016; Kartal et al., 2023)
Interbank Offered Rate(Umar & Sun, 2016)
Bank Size(Boussaada et al., 2020; Gjeci & Marinc, 2018; Hakimi et al., 2020; Jenkins et al., 2021; Kartikasary et al., 2020; Rehmana et al., 2020; Umar & Sun, 2016)
Bank Capital(Boussaada et al., 2020; Hakimi et al., 2020; Jenkins et al., 2021; Kartikasary et al., 2020; Kumar & Kishore, 2019; Kuzucu & Kuzucu, 2019; Kartal et al., 2023)
Earning Volatility(Umar & Sun, 2016)
Leverage(Umar & Sun, 2016)
Corruption(Anastasiou et al., 2019; Gjeci & Marinc, 2018; Hakimi et al., 2020; Jenkins et al., 2021; Mohamad & Jenkins, 2021; Rehmana et al., 2020; Mehmood et al., 2019)
Bank Efficiency(Gjeci & Marinc, 2018; Jenkins et al., 2021; Kartikasary et al., 2020; Kumar & Kishore, 2019)
Exchange Rate(Ayhan & Kartal, 2021; Fakhrunnas et al., 2022; Khan et al., 2020; Kuzucu & Kuzucu, 2019; Tanasković & Jandrić, 2015; Kartal et al., 2023)
Tax Rate(Khan et al., 2020)
Governance(Anastasiou et al., 2019)
Political Stability(Anastasiou et al., 2019; Hakimi et al., 2020; Mehmood et al., 2019)
Government Effectiveness,(Anastasiou et al., 2019)
Bank Concentration(Anastasiou et al., 2019)
Bank Deposits(Anastasiou et al., 2019)
Foreign Direct Investment(Kuzucu & Kuzucu, 2019)
Income Diversification(Kumar & Kishore, 2019; Rehmana et al., 2020)
Loans-To-Deposit Ratio(Kartikasary et al., 2020)
Performance(Boussaada et al., 2020)
Liquid Assets(Boussaada et al., 2020; Hakimi et al., 2020)
Capitalization(Rehmana et al., 2020)
Saving Rate(Syed & Tripathi, 2020; Ayhan & Kartal, 2021)
Economic Policy Uncertainty(Karadima & Louri, 2021)
Bank Concentration(Karadima & Louri, 2021)
Government Budget Balance (Anita et al., 2022)
Public Debt(Anita et al., 2022; Foglia, 2022; Ahiase et al., 2023; Arham et al., 2020; Krasniqi-Pervetica & Ahmeti, 2022)
Money Supply(Anita et al., 2022)
Domestic Credit(Foglia, 2022)
Foreign Currency
Loans Ratio
(Tanasković & Jandrić, 2015)
Volume of import(Vouldis & Louzis, 2018)
Energy consumption(Mehmood et al., 2019)
Table 2. Credit risk drivers (source: generated by the authors).
Table 2. Credit risk drivers (source: generated by the authors).
Latent VariableIndicatorDescription of Indicator
Macroeconomic Factors (M)GDPGDP structure affects the credit risk: industrial sectors are frequently highly leveraged and capital-intensive. Corporate NPLs, especially related to mining and oil, have historically increased in Kazakhstan in response to changes in industrial GDP. Climate change and the volatility of commodity prices have an impact on agricultural lending. Large-scale defaults may result from droughts or subpar harvests, especially for rural banks with heavily weighted agricultural portfolios. Although borrowers in the service sector are typically more varied and smaller, they are also more exposed to crises such as recessions and pandemics. When GDP services declined, there may be a rise in defaults on retail and SME loans in urban areas. Therefore, the hypothesis tests the negative correlation with NPLs.
GDP Growth RateGDP growth and credit quality are strongly positively correlated. The negative consequences of recessionary pressures make it more difficult for borrowers to repay their loans, increasing NPLs in emerging economies. Therefore, the hypothesis tests the negative correlation with NPLs.
Inflation RateInflation reduces real incomes and makes fixed-rate loans more burdensome, negatively impacting the credit quality. It also restricts long-term investment and makes borrowing more uncertain. Currency depreciation brought on by inflation can make debt loads worse in emerging countries like Kazakhstan that have a large amount of lending measured in foreign currencies. Therefore, the hypothesis tests the positive correlation with NPLs.
Social Factors (Soc)Unemployment RateUnemployment lowers household disposable income, affects a person’s capacity to repay loans, raises personal loan defaults, and increases the NPLs in emerging markets. Therefore, the hypothesis tests the positive correlation with NPLs.
Quintile CoefficientThe stratification of society assumes different access to financial resources and repayment capacities; higher income inequality correlates with higher credit risk. The lower-income population is more sensitive to economic shocks, and they have a high chance of not repaying the credits. Therefore, the hypothesis tests the positive correlation with NPLs.
Index GiniThe Gini index emphasizes differences in wealth distribution and financial resilience. It is a more comprehensive indicator of income inequality compared to the quintile coefficient. The higher the value of the Gini index, the higher the positive impact on NPLs. Therefore, the hypothesis tests the positive correlation with NPLs.
Saving Factors (S)Households DepositsResilient household savings (deposits) serve as a safeguard against defaults in difficult times, allowing mitigating risks of non-payment of the loans, allowing the amount of NPLs to decrease. Therefore, the hypothesis tests the negative correlation with NPLs.
Households Accounts BalanceHouseholds’ accounts balance is a measure of household liquidity and financial resilience, presenting the ability of individuals to manage short-term financial obligations, including loan repayments. A stronger accounts balance typically reflects higher savings capacity and lower vulnerability to income fluctuations, reducing the likelihood of default. Therefore, the hypothesis tests the negative correlation with NPLs.
Total Bank DepositsGrowing deposits promote the increase in the total number of loans, serve as a sign of rising public confidence in the banking system, and improve the ability to repay loans, decreasing the risks of NPLs. Therefore, the hypothesis tests the negative correlation with NPLs.
Foreign Factors (F)Public DebtExcessive public debt can discourage private sector borrowing and raise fiscal risks, affecting the creditworthiness of government-dependent industries and state-affiliated businesses. Private companies can suffer from reduced subsidies and debt servicing pressures. Therefore, the hypothesis tests the positive correlation with NPLs.
Exchange RateExchange rate fluctuations is a significant factor in the rise in NPLs in countries with big foreign exchange loan portfolios. For example, historically, periods of tenge depreciation in Kazakhstan have resulted in increases in corporate and household defaults, particularly for businesses with unhedged foreign exchange liabilities. Therefore, the hypothesis tests the positive correlation with NPLs.
Foreign Direct InvestmentsPositive FDI inflows generally indicate economic optimism and lower corporate segment credit risk. However, capital-intensive industries like mining or energy (typical for Kazakhstan) may become unstable due to drops in FDI, which could result in loan defaults. Therefore, the hypothesis tests the positive correlation with NPLs.
NPLnon-performing loans of private individuals
non-performing loans of legal entities
Table 3. Model indicators with loadings and p-values (source: generated by the authors).
Table 3. Model indicators with loadings and p-values (source: generated by the authors).
Indicators of the ModelLoadings and p-ValuesExcluded Indicators
Macroeconomic Factor (M)
GDP Growth (M1)0.875; p < 0.001GDP
Inflation (M2)0.943; p < 0.001
Savings Factor (S)
Household Deposits (S1)0.999; p < 0.001Households Accounts Balance
Total Bank Deposits (S2)0.999; p < 0.001
Foreign Factors (F)
Exchange Rate (KZT/USD) (F1)0.994; p < 0.001Public Debt;
Foreign Direct Investments
Public Debt as a Percentage of GDP (F2)0.992; p < 0.001
Social Factors (Soc)
Gini Index (Soc1)0.963; p < 0.001Unemployment Rate
Quintile Coefficient (Soc2);0.966; p < 0.001
NPLs
NPL 10.993; p < 0.001-
NPL 20.991; p < 0.001
Table 4. Outer model measurements: reliability and validity. Source: generated by the authors on the basis of SmartPLS 4 software.
Table 4. Outer model measurements: reliability and validity. Source: generated by the authors on the basis of SmartPLS 4 software.
pAAVECRCronbach’ Alfa
NPL0.9880.9840.9920.984
F0.9880.9870.9930.986
M0.8800.8280.9060.799
S1.0000.9970.9980.997
Soc0.9260.9300.9640.925
The use of green, as common in statistical analysis, shows the acceptable quality of the indicator.
Table 5. Heterotrait–monotrait ratio (HTMT) results. Source: generated by the authors on the basis of SmartPLS 4 software.
Table 5. Heterotrait–monotrait ratio (HTMT) results. Source: generated by the authors on the basis of SmartPLS 4 software.
FMNPLSSoc
F 0.996
M0.764 0.2900.4220.211
NPL
S 0.382
Soc 0.408
The use of green, as common in statistical analysis, shows the acceptable quality of the indicator.
Table 6. Results of hypotheses testing with bootstrapping algorithm. Source: generated by the authors on the basis of SmartPLS 4 software.
Table 6. Results of hypotheses testing with bootstrapping algorithm. Source: generated by the authors on the basis of SmartPLS 4 software.
Hypotheses Beta βSample MeanSTDEVNPLSSoc
H1M -> NPLs−0.800−0.6760.1320.0689.7290.000
H2M -> F -> NPLs−4.088−3.8601.4271.3592.5240.012
H3M -> S -> NPLs0.8960.5571.0781.0421.6660.096
H4M -> Soc -> NPLs1.0320.9941.0660.9151.5830.114
The use of green, as common in statistical analysis, shows the acceptable quality of the indicator. The use of red traditionally shows the bad quality of the indicator.
Table 7. Interpretation of the obtained results of bootstrapping algorithm. Source: generated by the authors on the basis of SmartPLS 4 software.
Table 7. Interpretation of the obtained results of bootstrapping algorithm. Source: generated by the authors on the basis of SmartPLS 4 software.
βT-Statisticsp ValuesInterpretation
M -> NPLs−0.8009.729<0.001Strong negative relations
M ->F -> NPLs−4.0882.5240.012Strong negative relations
M -> S -> NPLs0.8961.6660.096Insignificant direct relations
M -> Soc -> NPLs1.0321.5830.114Insignificant direct relations
The use of green, as common in statistical analysis, shows the acceptable quality of the indicator. The use of red traditionally shows the bad quality of the indicator.
Table 8. The main findings of the study.
Table 8. The main findings of the study.
H-s T-Statisticsp ValuesResults of Hypotheses Testing
H1M -> NPLs9.7290.000Not rejected (confirmed)
H2M -> F -> NPLs2.5240.012Not rejected (confirmed)
H3M -> S -> NPLs1.6660.096Rejected
H4M -> Soc -> NPLs1.5830.114Rejected
The use of green, as common in statistical analysis, shows the acceptable quality of the indicator. The use of red traditionally shows the bad quality of the indicator.
Disclaimer/Publisher’s Note: The statements, opinions and data contained in all publications are solely those of the individual author(s) and contributor(s) and not of MDPI and/or the editor(s). MDPI and/or the editor(s) disclaim responsibility for any injury to people or property resulting from any ideas, methods, instructions or products referred to in the content.

Share and Cite

MDPI and ACS Style

Kalimoldayev, A.; Popova, Y.; Cernisevs, O.; Popovs, S. Impact of Macro Factors on NPLs in the Banking Industry of Kazakhstan. J. Risk Financial Manag. 2025, 18, 431. https://doi.org/10.3390/jrfm18080431

AMA Style

Kalimoldayev A, Popova Y, Cernisevs O, Popovs S. Impact of Macro Factors on NPLs in the Banking Industry of Kazakhstan. Journal of Risk and Financial Management. 2025; 18(8):431. https://doi.org/10.3390/jrfm18080431

Chicago/Turabian Style

Kalimoldayev, Almas, Yelena Popova, Olegs Cernisevs, and Sergejs Popovs. 2025. "Impact of Macro Factors on NPLs in the Banking Industry of Kazakhstan" Journal of Risk and Financial Management 18, no. 8: 431. https://doi.org/10.3390/jrfm18080431

APA Style

Kalimoldayev, A., Popova, Y., Cernisevs, O., & Popovs, S. (2025). Impact of Macro Factors on NPLs in the Banking Industry of Kazakhstan. Journal of Risk and Financial Management, 18(8), 431. https://doi.org/10.3390/jrfm18080431

Article Metrics

Back to TopTop