Evaluation of the fintech era on the performance of Moroccan banks: analysis through non-performing loans

1. Introduction

The evolution of technology is sometimes dizzying. Innovative financial technologies, often called fintech, are making major breakthroughs around the world (Mahboub et al., 2023). This advancement of the fintech revolution has changed the traditional role of banking by introducing new financial services such as digital lending, peer-to-peer lending, online payments, virtual currency, crowdfunding, etc. These new electronic financial services have revolutionized the tasks of banking activities as service providers, and have challenged the banking profession (Sadok & El Maknouzi, 2021).

However, historical hindsight shows that the arrival of fintech does not constitute the first technological “disruption” to occur in the banking markets. The development of ATMs and online banking also constituted profound changes brought about by new technologies, which have little changed the structure of banking markets (Mahboub and Sadok, 2022). The emergence of online banking has been a factor of radical change in the structure of banking markets, allowing the appearance of new players offering substitute banking services (Sakka et al., 2022). The effects of this emergence have not been a radical change in the banking landscape, but an evolution in consumer practices, with the implementation of digital banking services including in conventional banks. Thus, in 2021, within traditional banks themselves, 445 fintechs were created and 2.8 billion euros raised to finance 155 operations (Sadok, 2023a).

Several researchers have described the fintech revolution for banks as a double-edged sword: on the one hand, it reduces banking costs by making it possible to better and quickly know customers, reduce the requirements of bank branches and therefore increase banking performance. On the other hand, fintech increases the operating costs necessary for setting up adequate infrastructure for the development of fintech technologies, particularly in emerging countries, and therefore deteriorates performance (Sadok & Assadi, 2023). Ozili (2018), Phan et al. (2020), Ozili (2021) and Gomber et al. (2018) explored the impact of the fintech era on banking performance through Non-Performing Loans (NPL). They concluded that the fintech era has a negative influence on banking sector NPLs. Zalan and Toufaily (2017) and Thakor (2019) also come to this conclusion. These authors also conclude that the fintech revolution negatively influences the performance of banks. It reduces market shares and revenues as well. Magee (2011) confirms this finding by stating that fintech loans are non-performing because the collateral requirements for loan recovery are low in fintech loans. However, Claessens et al. (2018) argued that the default risks in fintech loans are low because they have a short maturity: they are more oriented towards consumer credit compared to those of traditional banks, whose loans are intended to finance much more real estate and the commercial sector with longer maturities. Other researchers such as Ozili (2018) and Wang et al. (2021) have highlighted other positive impacts of the fintech revolution in poverty reduction, contribution to sustainable development, reduction of costs generated by intermediaries and financial inclusion (Ozili, 2018; Wang et al., 2021).

These arguments reveal the non-existence of a consensus on the impact of the fintech revolution on banking performance. Once these elements of perspective have been established, and ignoring the effects of macroeconomics such as the unemployment rate, the real interest rate, the total external debt/GDP and the inflation rate on the performance of the banking sector, the objective of this article is to focus on the impact of the introduction of financial technologies in the operational processes of Moroccan conventional banks on their performance. Digitalization offers banks substantial gains in operational efficiency. Automating processes, from customer data management to transaction completion, significantly reduces delays and errors. Indeed, intelligent systems enable optimal analysis and processing of information, allowing better assessment of the creditworthiness of credit applicants, thus optimizing decision-making and risk management. This increased efficiency translates into reduced operational costs and a better ability to respond quickly to customers’ financing needs.

To our knowledge, there is no specific study that has examined the role of the fintech era on the performance of the banking sector in Morocco. This article attempts to contribute to filling this gap and add to the growing literature on the link between the fintech era and its impact on banking performance. Thus, to ensure whether fintech makes it possible to better evaluate the granting of bank credits, and therefore makes it possible to reduce overdue credits, measured by NPL, this work will be presented as follows: Section 2 reviews the available literature on the above issue; Section 3 presents the methodology and empirical framework; Section 4 covers the analysis and discussion of the results; the last section highlights the concluding remarks.

2. The relationship between fintech and banking performance: literature review

In the contemporary world, nations and companies are devoting more and more resources to to digital technology. Although this asset contributes to increasing productivity, its consequence is to produce or improve a set of goods or services, to implement better techniques applicable to business processes, and more generally on all dimensions of entrepreneurship (Mohsin et al, 2023).

The study of the effects of fintech on banking performance has been the subject of multiple studies in recent years. Several researchers have attempted, through different measures and approaches, to empirically examine the influence of fintech on financial services, and, consequently, its impact on the performance of the banking sector. By analyzing the opportunities and threats of fintech services on the traditional banking sector, Romānova and Kudinska (2016), concluded that the non-adoption of digital is an existential risk given the fierce competition and adaptation challenges imposed by newcomers (fintech). In the same vein, Wewege et al. (2020) confirm this finding and argue that innovation in financial services improves the functioning of processes in the banking sector. The latter have an influence on the adoption of fintech products, particularly in the face of competition and growing customer demands (Bani Atta, 2024). This alignment with customer needs and satisfaction allows for an increase in productivity in the operationalization of processes and, consequently, the improvement of the profitability of financial operations (Sadok et al., 2022). Studies conducted by Haas et al. (2015), Setia et al. (2013), Mahboub and Sadok (2023a) corroborate these results and highlight the positive aspects of the fintech era on the performance of the banking sector.

For their part Navaretti et al. (2017) concludes that traditional banks, especially large commercial banks, take digital investment seriously, and that the alignment of conventional banks with the demands and constraints of the fintech era is a necessity to ensure sustainability (El Alami et al., 2015).

In this context, Tang (2019) explored whether fintech constitutes a substitute or complement to traditional banks. He wonders if they respond more to the needs of customers not covered by traditional banks, and in this case, it would be a complement, or rather to the new needs of customers of traditional banks.

The main reason why fintech may pose an existential risk to traditional banks if they do not integrate fintechs into their operational processes lies in the distortions of fintechs’ lower operating costs compared to traditional banks (Sadok, 2023a).

The use of technology makes the operationalization of banking processes less costly: the customer who wants a loan, the main and lucrative activity of the bank, can now fill out an online application. The application is accepted or rejected based on the elements provided, such as income, assets, analysis of the bank statement, or retrieved via other data sources, including the Internet and social networks (Benkhayat et al., 2015). This process of evaluating credit applications, based on the speed of more precise processing of a mass of data more representative of what was previously required (Fuster et al. 2019) allows for better selection of borrowers and a reduction in the volume of non-performing loans (NPLs). These loans represent the largest and most costly financial risk in the banking sector. Therefore, the investment of traditional banks in digital (fintech) is therefore supposed to improve the overall performance of the bank.

It is precisely to determine whether this fintech era has made it possible to improve the performance of traditional banks, thanks to a more detailed assessment of credit risks (Sadok, 2023b) and savings on structural costs (Mahboub & Sadok, 2023b), that a series of studies have been carried out.

It is with this in mind that the work of Welltrado (2018) estimates that operating costs represent 2.7% of outstanding loans at Lending Club (a type of fintech bank), compared to 7% at traditional banks. Cortina and Schmukler (2018) estimate that due to this competitive pressure, 1% of banking revenue losses have been observed in North America. The very significant gap in operating costs can thus offer a major competitive advantage to banks that use fintechs to gain competitive advantages in credit markets. Hau et al. (2019) come to the same conclusion for Chinese banks that extend credit using fintechs.

On this basis, Philippon (2015) concludes that there is room for significant improvement in banking performance by reducing financial intermediation costs through increasingly massive use of fintech. In this perspective, Pierri and Timmer (2020) investigated the role fintech on banking NPLs during the global financial recession. They concluded that banks that use more IT services in banking transactions have fewer NPLs. Ozili (2021), who explored the impact of the fintech on Non-Performing Loan (NPL) banking sectors in 35 countries, concluded that the fintech-led transition has a negative impact on banking NPLs. In the same line, Buchak et al. (2018) conducted a study on the United States where they concluded that fintech benefits from an advantage in terms of convenience for borrowers, with the added benefit of lower service costs. They even note that the ease of processes appears to allow fintech lenders to charge higher rates to less risky borrowers, which ultimately improves the bank’s performance.

Maier (2016) reaches the same conclusion in an analysis of SMEs borrowing from a European fintech platform. The convenience and transparency of the process play much more role than other classic criteria related to customer relations such as loyalty or proximity. Thus, the flexibility and transparency of credit application processing offered by fintech would play as decisive a role as the cost of financing in attracting SME borrowers (El Maknouzi and Sadok, 2021), especially with the rise of distrust towards traditional banks (Bröstrom et al., 2018; Fungacova et al., 2019).

The expansion of fintech is therefore promised to positively impact the bank’s performance through the improvement of credit file processing, and consequently the reduction of non-performing loans (NPLs). The experiences of implementing financial technology in certain southern countries such as Kenya and China, through pilot experiences such as that of Taoboa and M-Pesa, revealed that the success of fintech has made possible not only accessibility to banking services for the excluded, but more generally social inclusion and inclusive growth (Sadok, 2021).

The field of fintech and its impact in the Moroccan context are still little studied. This analysis of the scientific literature on this topic allows us to outline certain avenues and study variables to help fill this gap. More precisely, it allows us to shed light on this assertion which consists in considering that the implementation of fintech in the Moroccan banking sector makes it possible to better evaluate the granting of credit, and therefore, reduce the risk of insolvency of borrowers- measured by the reduction of Non-Performing Loan (NPL)- and consequently improve the performance of traditional banks. The following empirical approach will be devoted to this issue.

3. Methods

To analyze the existence of the relationship between the emergence of fintech in the operational and business processes of banks and the performance of the latter, we estimate a panel data model from a representative sample of data from Moroccan banks of 2007 to 2022. Thus, we divided the period of this study into 2 sub-periods: the first period where fintech was not sufficiently present in the digitalization of banking processes (fintech) was not sufficiently present in the business processes of Moroccan banks, namely the period 2007-2014. This period would serve as a reference to compare it to that of the second period studied, that of the fintech era, 2015-2022, to be able to determine whether the banks studied were more or less efficient during this period, ignoring other macroeconomic variables. This segmentation of the period studied into 2 sub-periods, considering the fintech era as that beginning from 2015, is justified if we take into consideration the emergence of fintech and their generalization in the banking sector of developing countries from the second decade of this millennium. We justify this temporal demarcation by referring to the work that raised the same theme in the literature review made above, and more particularly that of Ozili (2021) and Arner et al. (2016). These works generally consider that conventional banks in developing countries are showing a growing interest in online banking and the digitalization of their processes from the beginning of the second decade of this millennium. And to have a more precise anchor point for this growing interest in the digitalization of Moroccan banks, we examined the management reports of these five banks from 2010, and this interest proved to be more concrete in terms of investment, and especially the deployment of digital processes and applications from 2015.

At the end of 2022, the number of credit institutions and similar organizations approved in Morocco stands at 90 establishments, including: 19 banks, 5 participatory banks, 29 financing companies, 6 offshore banks, 11 micro-credit associations, 18 payment institutions, the Deposit and Management Fund and the Central Guarantee Fund. Our sample includes five Moroccan banks listed in the Table 1 below, including the average size of their balance sheet, the average of their net banking income (NBI), as well as the average of their Return on Equity (ROE) during the 2 periods studied. The choice of these banks comes from the fact that the share of these first 5 banks in the total assets of the banking sector in Morocco stood at 76.4% in 2022, at 75.4% in 2014 against 78.1 % in 2007 (Mouline and Sadok, 2021a, 2021b). Therefore, the level of concentration of the banking activity of these 5 banks is high, and remains approximately the same throughout the period studied between 2007 and 2022. This leads us, for sampling purposes, to focus only on these five largest banks, considering that the latter constitute the bulk of the banking market in Morocco.

In this study aimed at investigating the influence of the fintech era on the performance of the banking sector, we opted for proxies in order to capture the variable to be explained, which is performance, from the explanatory variables specific to the banking sector during the 2 sub-periods. Thus, referring to previous works such as those of Dwumfour (2017) and Fernández et al (2016), we found that the overall value of non-performing loans (NPLv) and the coverage rate of these loans (Cnpl) as indicators to measure the performance of the banking sector than other indicators such as net banking income. In addition, the rate of variation of the interest margin (IMv), the variation of interest charges divided by net banking income(I/NBI), the variation margin on commissions (MCv), the operating coefficient (OC), the value of provisions for non-performing loans (Pnpl), are included as explanatory variables. These explanatory variables were chosen on the basis of the literature analysis carried out by Phan et al. (2020) and Ghosh (2015).

Furthermore, to make the variables of this study more explicit, we have drawn up figures in the appendix (Figures 1 and 2 below) summarizing the variables to be explained, namely the performance of banks, measured essentially by non-performing loans and the coverage rate of these loans, and explanatory variables namely, the rate of variation of the interest margin, the variation of interest charges divided by net banking income, the variation margin on commissions, the operating coefficient, the value of provisions for non-performing loans.

Figure 1. Data on financial variables from AWF Bank, BCP Bank and BMCE Bank in the fintech era.

Figure 2. Data on financial variables from BMCI Bank and SGMB Bank in the fintech era.

In the current analysis, we have included five explanatory variables, namely, the rate of variation of the interest margin (IMv), the variation of interest charges divided by net banking income (I/NBI), the variation margin on commissions (MCv), the operating ratio (OC) and the value of provisions for non-performing loans (Pnpl):

The rate of variation of the interest margin (IMv) is included because previous literature concludes that expansion in banking activities results in increasing banking profitability and capital buffer, which eventually decreases nonperforming loans (NPLv) and the amount allocated to cover these loans (Cnpl). Hence, we assume a negative relationship between the rate of variation of the interest margin and banking performance.

The variation of interest charges divided by net banking income (I/NBI) was chosen because countries facing higher interest charges have a higher chance of banking underperformance. Significant amounts of interest charges lead to defaults and poor bank performance.

The variation margin on commissions (MCv) also included as an explanatory variable because the increase in commissions promotes organizational efficiency and productivity, and therefore we can assume that the increase in this variable varies in the opposite direction as non-performing loans.

Then, we also took the operating ratio (OC) as an independent variable because previous studies concluded that the increase in the operating ratio is the consequence of a reduction in non-performing loans. Therefore, we can assume a negative relationship between the OC and NPL. The value of provisions for non-performing loans (Pnpl) have been selected because previous literature exhibits that the value of provisions for doubtful debts is a determining criterion for the quality of outstanding debts: the higher this rate, the more the bank is convinced by the fact that these debts are irrecoverable even after recourse and trial. We assume a positive relationship between the value of provisions for non-performing loans and the variables to explain banking performance, namely, the nonperforming loans and the coverage rate of these loans.

Finally, BIN variable is considered as a binary variable that takes the values 1 from 2015 to 2022 and 0 from 2007 to 2014. BIN is used to describe the influence of the fintech era on the performance of the banking sector. We assume here that BIN has a positive impact on the performance of the banking sector. Technological innovation and the growth of fintech products with adequate infrastructure can increase banking performance. The Table 2 in annexes shows the variable description and their expected relationship.

4. Model framework and variables analysis

This section highlights the model framework used to analyze the determinants of the performance of the Moroccan banking sector in the fintech era. We used the financial data of the 5 banks in our sample from the fintech period compared to the pre-fintech period to seek the meaning and the “intelligibility”, of the differences in bank performance during these two periods. The basic idea that linear structural equation modeling (LSEM) seeks to address below is to infer causal relationships from data of variables considered by the current state of knowledge to be explanatory. The objective here is to know whether the association observed between banking performance, measured by the good management of non-performing loans, is, or not, the result of a causal relationship induced by the arrival of fintech allowing better knowledge and a better evaluation of credit applicant files, and consequently an improvement in the explanatory variables of the model.

Equations 1 and 2 above of the LSME model are a formalized representation of the relationships between the variables considered from existing knowledge. It is possible to draw valid causal inferences from observing covariations between these variables, except that we will extend the model to include analysis of interactions over time. Thus, the equation formulated after including the interaction variable of the fintech period (BIN = 2015-2022) is as follows:

Where NPLv represent nonperforming loans, Cnpl is used to measure coverage rate of these loans, IMv measure the rate of variation of the interest margin, I/NBI is used to describe the variation of interest charges divided by net banking income, MCv is represent the variation margin on commissions, OC denotes the operating coefficient, Pnpl measure the value of provisions for non-performing loans, BIN is used to describe the fintech era, t represents the time period, $\mathrm{\epsilon }$ is used to show the error term, i denotes the country and c is the constant.

Table 3 below presents the statistical description of these explanatory variables and to be explained during the first study period (2007-2014), and during the second study period (2015-2022), considered as the fintech era.

The descriptive statistic shows that during the fintech Eera (2015-2022), the mean value of nonperforming loans NPLv (9.48), coverage rate of these loans Cnpl (7.22), supposed to approximate banking performance remained the same compared to the pre-fintech period (2007-2014), if not even a deterioration in the capacity to cover non-performing loans.

All other explanatory variables in the model evolved in a favorable direction, and yet performance did not follow this momentum: banking indicators increased from one period to another, but non-performing loans remained almost stable with a reduction in the capacity to cover non-performing loans. This first observation leads us to a second stage of analysis at the following point to ensure, or not, that the arrival of technologies in the fintech era has stimulated banking activity, but has not enabled better knowledge of customer credit applications, and consequently a reduction in non-performing loans.

5. Result and analysis

After analyzing the descriptive statistics, we evaluated the correlation relationship between the variables. Table 4 below presents the results of the correlation analysis. The Pearson correlation coefficient shows that the NPLv are weakly correlated with the variables supposed to reflect an impact on banking indicators induced by the wave of financial technologies. This implies that the NPLv remained constant as well as their provision value which decreased during the second period of the study which symbolizes the fintech revolution.

Furthermore, the results show that the rate of variation of the interest margin (IMv), the variation of interest charges divided by the net banking product (I/NBI), and the variation margin on commissions (MCv), are negatively correlated with NPLv, even if this correlation is not statistically significant enough.

Conversely, the operating ratio (OC) and provisions for non-performing loans (Pnpl) are positively associated with NPLv. This shows that the levels of non-performing loans impact the operating result (OC) and reciprocally increase the amount allocated for the provision of NPLv (Pnpl).

In terms of the coverage rate of the non-performing loans (Cnpl), correlation results show that the latter is negatively correlated to the rate of variation of the interest margin (IMv), the variation of interest charges divided by the net banking product (I/NBI) and the variation margin on commissions (MCv): when the coverage of the bad debt provision increases, it has a negative impact on banking performance indicators. Likewise, the result also reveals the operating result (OC) and the amount allocated for the provision of NPLv (Pnpl) are positively associated with the coverage rate of the non-performing loans (Cnpl). These results generally mean that the fintech period was characterized by an improvement in operating income (OC) and in the level of the amount allocated to the provision of NPLv (Pnpl) thus improving banking performance measured by the value of non-performing loans (NPLv) and the coverage rate of these loans (Cnpl).

This considered fintec period was marked by an improvement in the interest margin (IMv), the variation in interest charges divided by net banking income (I/NBI) and the variation margin on commissions (MCv), however, these improvements in banking indicators have not had a positive impact on the performance of the Moroccan banking sector.

After the correlation analysis done above, Table 5 below presents the results of the regression analysis. The results show that rate of variation of the interest margin (IMv) has a negative impact on value of non-performing loans NPLv (-0.618) and on the coverage rate of these loans (Cnpl) (-0.541). This implies that the value of non-performing loans decreases, as well as the coverage rate of these loans (Cnpl), with the increase in the level of the variation of the interest margin.

Based on the coefficient value of explanatory variables, we can infer that:

In addition to the above analysis between the determinants of the value of non-performing loans (NPLv) and the coverage rate of these loans (Cnpl), we included the results of the interaction analysis. The latter describes how the determinants of NPLv and banking stability, measured by the coverage rate of these loans, react with the expansion of the wave of the fintech revolution (2015-2022) compared to the first period of the study (2007-2014).

The results indicate that the financial technology (fintech) wave did not contribute significantly to the decrease in the value of non-performing loans. These remain the same from one period to another. But if we take into consideration the financial variables of the banking sample before and during the fintech era, such as the value of assets, the net banking product (NBI), we can conclude that the arrival of fintech in this sector made it possible to contain non-performing loans in values similar to the previous period despite the increase in turnover of the five banks in our sample. In other words, the proportion of non-performing loans in relation to the sum of credits distributed has decreased between the 2 periods. This explains why the value of provisions for non-performing loans (Pnpl) has decreased when we compare the period 2007-2014 with the fintech period (2015-2022).

The values of the BIN coefficient are negative for NPLv and positive for the coverage rate of these loans (Cnpl) because the expansion of fintech services has led to an increase in the number of fintech lenders, without leading to an increase in bad borrowers generating an increase in non-performing loans (NPLv). It appears that the increase in fintech services has improved customer management, file assessment and loan monitoring services, which ultimately reduced non-performing loans and improved the coverage rate of non-performing loans (Cnpl).

Furthermore, the above interaction results also conclude that the fintech era has had the effect of reducing the rate of variation of the interest margin (IMv), improving the variation of interest charges divided by the net banking product (I/PNB), and promote the rate of variation of the interest margin (IMv) and the variation margin on commissions (MCv), which cumulatively contributed to stabilizing the NPLv and reducing the value provisions for non-performing loans (Pnpl).

The result shows that the coefficient value of the explanatory variable (BIN* IMv, BIN* I/NBI, BIN* MCv, BIN* OC, BIN* Pnpl) is negative in the case of NPLv and positive for the Cnpl. Hence, we can conclude that during the fintech era, the Morrocan banking condition has improved slightly: the non-performing loans did not increase despite the significant increase in net banking income and the size of bank assets during the fintech era compared to the first period. This increase in business volume did not produce an increase in provisions for non-performing loans. The results of this study are consistent with the results conducted by Ozili (2021) and those of Pierri and Timmer (2020).

6. Conclusion, recommendations and limitations

This study explores the influence of the fintech era on Moroccan banking performance addressed through non-performing loans and their provisions. The question is whether the implementation of fintechs in the Moroccan bank’s credit evaluation process makes it possible to reduce overdue credits. Experiences carried out in certain countries of the South, such as Kenya and China, have revealed that these technological innovations in the banking sector have made it possible not only to reduce risk, but also to facilitate the inclusion of those excluded from banking services through pilot experiments such as Taoboa and M-Pesa (Sadok, 2021).

To measure the impact of the fintech era, we separated the data between two periods. The first covers the period 2007-2014 and the second covers the period 2015-2022, considered the period of the fintech revolution (Ozili, 2021; Arner et al., 2016). The results of our study show that the arrival of fintech in this sector made it possible to contain non-performing loans in values similar to the previous period despite the increase in turnover of the five banks in our sample. In other words, the proportion of non-performing loans in relation to the sum of loans granted decreased between the two periods studied. This explains why the value of provisions for non-performing loans (Pnpl) decreased when comparing the period 2007-2014 with the fintech period (2015-2022). This result corroborates those obtained by Buchak et al. (2018), Phan et al. (2020) and Ozili (2021).

The findings from the interaction analysis conclude that the second era of the fintech revolution had an effect in reducing the rate of variation of the interest margin (IMv), improving the variation of interest charges divided by the net banking product (I/PNB), and promote the rate of variation of the interest margin (IMv) and the variation margin on commissions (MCv), which cumulatively contributed to stabilizing the non-performing loans (NPLv) and reducing the value provisions for non-performing loans (Pnpl).

However, the findings of the present study suggest the following policy recommendations. First, the expansion of financial technologies is a progressive measure for developing countries that contributes to improving banking stability and a smooth social revolution for greater transparency and of inclusion (Sadok, 2023b). Then the use of fintech makes it possible to improve credit assessment services to control and monitor non-performing loans, and therefore better gauge the solvency of borrowers.

If the originality of this work consists in the analysis of the interactions between the era of financial technologies and the determinants of the performance of the Moroccan banking sector through non-performing loans, this study suffers, however, from some limitations: the first limit is the size of the data and the sample, and the second limit relates to the choice of variables to characterize banking performance. These limits will serve in the future as new potential directions for future research to better explore the problem studied.

Ethics and consent

Ethical approval and consent were not required.