Unveiling the drivers of bank profitability: insights from Ethiopian banks  – Nature.com 05:54 

Humanities and Social Sciences Communications volume 12, Article number: 721 (2025) Cite this article

Abstract

Bank profitability analysis seeks to identify macroeconomic, industry, and bank-specific factors, helping management understand and assess key determinants of profitability. To this end, the objective the current study is to analyze the effects of bank specific, industry-specific and macroeconomic factors on fourteen Ethiopian commercial bank profitability for the period of 2011–2023 using a random effect (RE), generalized methods of moments (GMM), fully modified ordinary least square (FMOLS). The findings reveal that income diversification negatively impacts return on assets (ROA), while gross domestic product (GDP) and government effectiveness have substantial positive effects on return on assets (ROA). Return on equity (ROE) is positively influenced by productivity, government effectiveness, and innovation indexes, but negatively affected by capital adequacy and income diversification. Lastly, profitability evaluated by net interest margin (NIM) is positively affected by productivity, COVID-19, and GDP, whereas bank stability and government effectiveness have negative impacts. This study is the first to explore the influences of the innovation index, government effectiveness, and COVID-19 on the performance of Ethiopian banks. By addressing this novel area, the current study offers valuable insights for policymakers, regulators, bank managers, analysts, and other stakeholders, providing a foundation for strategies to enhance the profitability and performance of banks in Ethiopia

Introduction

With their crucial role as middlemen between savers and borrowers, banks form the financial system’s foundation in efficiently directing funds to investment, which contributes to economic growth in both advanced and emerging economies (Al-Harbi, 2019; Luo et al., 2016). They are essential architects of economic stability and prosperity in the developed world, going beyond simple financial transactions (Mishra et al. 2024). Banks support commerce, innovation, and investment, over and above protecting deposits and granting loans (Safiullah and Paramati, 2024; van Eeghen, 2021). These functions help to maintain the global economy. Their vital functions, including lending, investment management, and financial planning, make them the backbone of trade. They are essential to developed economies’ economic framework because they stabilize markets and influence monetary policy (Feng and Wang, 2018; Ofori-Sasu et al., 2022)

Building on the groundbreaking studies of (Al-Harbi, 2019; Chand et al., 2024; Ofori-Sasu et al., 2022; Siddique et al., 2022), academics, decision-makers, and financial executives have continued to pay close attention to the variables affecting bank profitability. Additionally, as stable and functional banking institutions are necessary for economic growth and development in emerging nations, knowledge of bank profitability has a noteworthy impact on macroeconomic policy and governance in those areas (Al-Harbi, 2019; Chand et al., 2024; Ofori-Sasu et al., 2022; Siddique et al., 2022).

Numerous studies have explored the factors influencing bank profitability across various environmental contexts, time periods, and regions, which are generally viewed through two lenses: a panel of countries or individual case studies. Studies conducted in group of countries includes but not limited to US and Asia (Abbas et al., 2019), Europe (Menicucci and Paolucci, 2016), South Asia (Sufian, 2012), GCC countries (Chowdhury and Rasid, 2016; S. Khan, 2022), Sub-Saharan Africa (Munyambonera, 2013), Organization of Islamic Cooperation (OIC) (Al-Harbi, 2019), Asia (Saif-Alyousfi, 2022), Middle East and North Africa(MENA) (Pan et al., 2023), Latvia and Lithuania (Titko et al., 2015), and United kingdom (O’Connell, 2023) and single country cases Brazil (Bernardelli and Carrasco-Gutierrez, 2024) Egypt (Kassem and Sakr, 2018),, China (Sufian, 2009), Pakistan (Ali and Puah, 2019), India (Al-Homaidi et al., 2018), Malawi (Chirwa, 2003), Bangladesh (Gazi et al., 2024), Tunisia (Bougatef, 2017), Nigeria (Akinkunmi, 2017), Switzerland (Dietrich and Wanzenried, 2014), Portugal (Garcia and Guerreiro, 2016), Tanzania (Kapaya and Raphael, 2016), Figi (Chand et al., 2024), Vietnam (Thanh et al., 2022), Kenya (Ongore and Kusa, 2013), Hungary (Serwadda, 2018). Most recently, a few studies have been conducted in Ethiopia, such as (Isayas, 2022) and (Bushashe, 2023).

Much of the existing research focuses on developed countries (Bernardelli and Carrasco-Gutierrez, 2024; Dietrich and Wanzenried, 2014; Garcia and Guerreiro, 2016; O’Connell, 2023), which are distinguished by well-regulated and efficient financial systems, mature banking sectors, and established capital markets. Nonetheless, findings derived from studies on developed economies often cannot be directly applied to developing countries like Ethiopia. The stark differences in corporate and legal environments, taxation systems, governance laws, political structures, government effectiveness, and innovation levels shape how firms operate and influence their decisions, including bank performance. Moreover, variations in interest rate environments, banking systems, and funding sources further underscore the uniqueness of each country’s financial landscape (Figs. 1–3).

figure 1
Fig. 1: Trends of ROA from some selected African nations.
figure 2
Fig. 2: Trends in key performance indicators of Ethiopian commercial banks.
figure 3
Fig. 3: Framework of the study: developed by the author (2025).

Ethiopia has long been distinctive in its approach to the financial sector, maintaining a policy of restricting its banking business for foreign banking (Kozo et al., 2007). However, this stance recently shifted with the introduction of the banking business proclamation No. 1360/2024 (Zewdu and Bezabih, 2024), signaling a significant step toward opening its financial market. In contrast, neighboring countries such as Uganda, Kenya, and Tanzania liberalized their banking systems to foreign institutions many years ago, fostering greater integration with global financial networks (Ahmed et al., 2023). Over the past decade, the performance of Ethiopian banks, measured by key gauges such as return on assets ((ROA) for further reference, please see the figure below), return on equity (ROE), and net interest margin (NIM), has shown considerable variation. Recent research on Ethiopian banks has been limited in scope (Isayas, 2022), addressing only short time periods and a narrow range of variables (Bushashe, 2023; Isayas, 2022) and failed to addresses the impact of COVID-19 on Ethiopian banks performance. This study aims to address these gaps by incorporating more novel variables like income diversification, productivity, bank stability, government effectiveness, innovation indexes, and COVID-19 and which were not commonly featured in previous research conducted over a long period from 2011 to 2023 concurrently employing RE, GMM, and FMOLS as a means of data analysis.

The paper is structured as follows. Section “Literature review and formulation of hypotheses” examines relevant literature, highlighting key theories and identifying gaps in extant research. Section “Methods” outlines the data sources and research methodology used in this study, ensuring a clear understanding of the analytical framework. In Section “Data analysis and result”, the paper presents the econometric results and engages in a critical discussion of their implications. Finally, the section “Conclusion” synthesizes the findings and offers targeted recommendations, providing actionable insights for stakeholders on the case of the banking industry.

Literature review and formulation of hypotheses

A brief backdrop of the Ethiopian banking sector

The Ethiopian banking sector began with the establishment of the Bank of Abyssinia in 1905, serving primarily Italian interests during the regime of Emperor Menelik II. This marked the first step in formalizing Ethiopia’s financial system. The landscape transformed significantly in 1942 with the creation of the National Bank of Ethiopia (NBE), which took over the role of the central bank and laid the foundation for modern banking operations in the country. This pivotal moment set the stage for the sector’s development, but it wasn’t until the end of the Derg regime in 1991 that the sector truly began to evolve.

The fall of the Derg regime heralded a new era of economic reforms and liberalization. The 1994 Banking Business Proclamation broke the state’s monopoly, allowing for the establishment of private banks and foreign investments. This shift catalyzed a surge in the sector’s growth, ushering in a period of intense competition and innovation. The early 2000s saw further modernization with the introduction of the 2004 Banking Proclamation, which established a more robust regulatory framework and encouraged technological advancements, including mobile and online banking.

Following the report by the National Bank of Ethiopia NBE (2023) today, the Ethiopian banking sector comprises 31 commercial banks (of which one is state-owned and the others are private) is a vibrant and rapidly evolving part of the economy. With total assets exceeding 2 trillion Ethiopian Birr (ETB) and deposits surpassing 1.7 trillion ETB, the sector demonstrates remarkable growth, thereby raising the number of bank branches to 11,281. It serves over 20 million deposit accounts, and the expansion of ATMs to around 4000 underscores a commitment to improving accessibility and customer convenience. These metrics reflect a dynamic sector that adapts and thrives amidst ongoing economic and technological changes.

Underpinning theories

Structure conduct performance (SCP) model

Following Tan (2015), the SCP framework posits that the structure of a market shapes how firms behave, which in turn affects their overall performance. Market structure is assessed through factors such as the number of industry competitors, product diversity, and the barriers to entering or exiting the market. Conduct encompasses the strategies and behaviors of firms, including pricing strategies, product differentiation, implicit collusion, and the use of market power. On the other hand, performance is gauged by metrics such as operational efficiency, resource allocation effectiveness, and financial profitability.

The SCP model posits that the market structure shapes the behavior of firms, which subsequently impacts overall economic performance. This framework has been employed to theoretically support and justify competition policy (Bain, 1951; Ferguson and Ferguson, 1998; Mason, 1939). They were the first two to propose that the degree of market concentration affects a company’s profitability. Companies in highly concentrated industries typically make significantly more money than those in less concentrated markets, they found. Increased concentration in the banking industry allows banks to work together, raise prices, and ultimately gain significant profits, according to the SCP paradigm (Bain, 1951; Bourai et al., 2024; Heggestad and Mingo, 1977; Mason, 1939; Stigler, 1964).

Market power theory (MPT)

MPT examines how firms can influence or control prices, contrasting with perfect competition, where no single firm affects pricing (Berger, 1995). It argues that firms with substantial market dominance can set higher prices due to their control over demand and supply, leading to increased profits and reduced competition (Kariuki, 2024). Unlike price takers in a perfectly competitive market, these dominant firms use their market position to command premium prices and secure monopoly-like profits. Additionally, external pressures such as competition and regulation can further enhance the financial performance of these powerful firms, allowing them to leverage their market power for greater success (Horvath et al., 2016)

Efficiency structure theory (EST)

Grounded up on Demsetz (1973), EST delves into how market competition influences a firm’s efficiency. It asserts that in competitive markets, firms are driven to enhance their operational efficiency to survive and outperform rivals. This competitive pressure forces businesses to streamline processes, cut costs, and innovate, resulting in higher productivity and improved performance. In contrast, in markets with limited competition or monopolistic structures, firms face less incentive to optimize their operations, which can lead to inefficiencies and stagnant productivity. Essentially, the theory underscores the vital role of competitive dynamics in fostering business excellence and operational effectiveness, revealing how market structure directly impacts a firm’s efficiency and overall success (González et al., 2019).

Quiet life hypotheses (QLH)

QLH suggests that greater market power leads to reduced managerial effort, lowering bank profitability and efficiency (Berger and Hannan, 1998; Hicks, 1935). In banking, large institutions often prioritize market advantages over intermediation efficiency, resulting in higher loan costs and reduced credit availability (Coccorese and Pellecchia, 2010). While small banks typically show lower interest margins, large banks have been linked to financial allocation inefficiencies due to organizational diseconomies of scale, reliance on credit information agencies, and the pursuit of a “quiet life” rather than leveraging their market position to enhance financial access (Asongu and Odhiambo, 2019; Boateng et al., 2018; Mitchell and Onvural, 1996).

Competition-fragility and competition-stability hypotheses

Academically, the relationship between competition among commercial banks and their fragility remains one of the most debated topics in financial research. This debate centers on two competing views: the competition-fragility hypothesis and the competition-stability hypothesis. Proponents of the competition-fragility view (Ahnert and Martinez-Miera, 2021; Beck et al., 2006; Dam and Castillo, 2006; Hellmann et al., 2000; Horvath et al., 2016; Marquez, 2002) argue that increased competition leads to greater risk-taking as banks, under pressure to maintain profitability, may lower lending standards, reduce capital buffers, and engage in more speculative investments. These actions, in turn, heighten the potential for systemic risk and financial instability. Conversely, supporters of the competition-stability view suggest that competition fosters efficiency, reduces the dominance of too-big-to-fail institutions, and can disperse risks, thereby enhancing the overall stability of the banking sector.

Finance-led growth or growth-led finance theory

The association between economic growth and financial development (FD) has been widely studied, with mixed evidence emerging from the literature. Several studies, such as those by (Abbas et al., 2022; Aziakpono, 2011; Guru and Yadav, 2019) highlight a significant association between FD and economic growth. Furthermore, empirical research by (Arestis and Demetriades, 1997; King and Levine, 1993) provides evidence of a causal link between finance and growth. While much of the literature suggests that FD drives economic growth, other studies argue the reverse—that economic growth can lead to advancements in the financial sector. Additionally, some research has found that the relationship between finance and growth is nonlinear. For example, studies by (Berkes et al., 2012) suggest that FD boosts economic growth only up to a certain threshold, beyond which it may become counterproductive and even (Mengesha and Berde, 2023) argue that FD may not play a critical role in driving economic growth.

Hypotheses

Bank-specific variables

Credit risk (CR) and capital adequacy (CA)

According to Poudel (2012), credit risk (CR) is the risk that a bank could suffer financial loss if a borrower fails to pay back their loan on time or at all, potentially leading to bankruptcy if not managed correctly. There is disagreement among research scholars as to whether CR affects Profitability. On one hand, due to greater provision costs and default rates, CR can have a detrimental effect on profitability (Ali et al., 2022; Chand et al., 2024; Islam and Nishiyama, 2016). On the other hand, by lowering losses and enhancing financial stability, efficient CR management can have a favorable impact on profit (Adusei, 2015; Athari and Bahreini, 2021; Fadun and Silwimba, 2023). Capital adequacy (CA) can positively affect profitability by ensuring financial stability and reducing risk, which can lead to lower borrowing costs and increased investor confidence. Concurring with this idea, previous researcher has found that CA have a profound effect on Banks’ profitability (Chand et al., 2024; Chandrasegaran, 2020; Noor and Suhendra, 2024). However, on the negative side, maintaining high capital reserves can limit the funds available for investments and lending, potentially hindering growth and reducing overall returns.

H1: Credit risk affects bank profitability positively

H2: Capital adequacy affects banks’ profitability positively.

Income diversification and productivity

Banks can enhance their profitability in a competitive market by expanding revenue from non-core activities, such as commissions, service charges, and transaction charges. By distributing risk and lowering reliance on a solo source of income, income diversification can boost profitability (Addai et al., 2022; Ahamed, 2017; Li et al., 2021). However, it can also hurt profitability because of its propensity to dilute focus and add complexity (Brighi and Venturelli, 2016; Maudos, 2017). In economics, productivity measures the efficiency of production, defined as the ratio of output to input (Sain, 2014). It reflects the technical efficiency and resource allocation of enterprises, highlighting the relationship between output and the input factors used to achieve it (Pekuri et al., 2011). Increased productivity often leads to higher earnings, better customer service, and improved efficiency, ultimately boosting profitability (O’Connell, 2023; Radha and Aithal, 2023; Wanigasekara and Mendis, 2021). However, focusing too much on output can lead to issues like lower quality, burnout, and conflicting incentives, which can negatively impact profitability (İmrohoroğlu and Tüzel, 2014; Yousaf, 2023).

H3: Income diversification affects banks’ profitability positively

H4: Productivity affects banks’ profitability positively

Expense control

Expense control and profitability of banks, insurance companies, and microfinance institutions are regularly examined topics in economic literature. Previous research underscores that managing expenses is vital for improving bank profitability, as efficient cost control directly strengthens financial performance (Ali et al., 2022; Jeris, 2021; Mirzaei et al., 2024). However, excessive cost-cutting can undermine operational effectiveness, potentially harming long-term profitability by stifling growth and innovation (Garcia and Guerreiro, 2016; Stavárek and Polouček, 2004).

H5: Expense control affects banks’ profitability positively

Industry-specific

Banking stability

According to (Ali and Puah, 2019) various stakeholders including researchers, policymakers, and bank authorities’ have shown a great deal of interest and entered into a lengthy debate about the profitability and stability of the financial sector following the Great Recession of 2007–2008. In some previous research, bank stability improves bank profitability (Shahriar et al., 2023).

H6: There is a positive association between bank stability and profitability

Macroeconomic factors

Gross domestic product (GDP) growth rate

Macroeconomic variables commonly employed in research, such as GDP growth rate and remittances, are often analyzed to understand their impact on the success of commercial banks. However, empirical studies have yielded conflicting results regarding the relationship between bank profitability and GDP (Abel et al., 2023; Ali et al., 2022; Isayas, 2022; Yakubu and Bunyaminu, 2022) point to a positive relationship, while others suggest a negative one (Athari and Bahreini, 2021; Islam and Nishiyama, 2016; Jeris, 2021). The picture painted by empirical research on the correlation between GDP growth rate and profitability of banks is nuanced and multifaceted.

H7: GDP Growth rate affects banks’ profitability positively

Government effectiveness and innovation indexes

Effectiveness in the context of Government, according to Angelini et al. (2020), refers to the government’s ability to create and carry out the nation’s active policies and regulations. Conferring to certain research for example (Chand et al., 2024; Jagannath and Maitra, 2023; Ullah et al., 2024), the performance of the financial industry and government effectiveness are positively correlated. Robust innovation indexes drive the development of cutting-edge financial products and services (Medyawati et al., 2021; Zhu and Guo, 2024), which in turn positively contributes to the success of the business organization.

H8: Government effectiveness affects profitability positively

H9: Innovation indexes affect profitability positively

Covid-19

The COVID-19 pandemic has profoundly disrupted global economies, with significant implications for the banking sector. Banks, as financial intermediaries, faced unprecedented challenges, including economic contraction, declining asset quality, increased credit risk, and operational disruptions. The interrelation between COVID-19 and bank performance is complex and multifaceted. On the one hand, the economic slowdown and heightened uncertainty led to reduced loan demand, higher default rates, and compressed interest margins and finally in turn adversely affect the overall banks performance (Shabir et al., 2023; Susanti et al., 2023; Xiazi and Shabir, 2022). On the other hand, some previous empirical evidence shows banks are more resilient to the outbreak of the covid pandemic (Demir and Danisman, 2021).

H10: Covid-19 affects profitability negatively

Investigation’s framework

The investigation’s framework of the current research aims to build insights from previous research while addressing existing gaps in the literature. By analyzing various studies, the study has identified key variables and relationships crucial for understanding our research problem. This framework incorporates four novel variables—productivity, government effectiveness, innovation index, and bank stability. This integration allows us to comprehensively explore their interactions, enriching our understanding of the dynamics involved (Table 1).Table 1 Summary of the most recent literature.

Full size table

Methods

Data and sample selections

The current study sought to investigate firm-specific, industry-specific, and macroeconomic factors of profitability in Ethiopian commercial banks by using a quantitative research approach paired with an explanatory research design. This study embarks on an insightful exploration of a balanced panel of data, which enable researchers to examine the same entities across different time periods (Hill et al., 2018) drawn from fourteen prominent financial institutions (FIs) in Ethiopia, spanning from 2011 to 2023. The banks in focus include renowned names such as Awash International Bank, Dashen Bank, Cooperative Bank of Oromia, Bank of Abyssinia, Hibret Bank, Wogagen Bank, Nib International Bank, Oromia International Bank, Lion International Bank, Brehan International Bank, Buna International Bank, Zemen Bank, and Abay Bank. among others. The sample banks were selected based on the availability of 13 years of annual audited financial statements, including the statement of financial position and the profit and loss and other comprehensive income statement. Banks lacking these statements were completely excluded from the sample. Above all, data on firm-specific variables comes directly from the annual audited reports of these institutions, sourced from the National Bank of Ethiopia. In tandem, macroeconomic and industry-related data are meticulously gathered from trusted sources like the World Bank and TheGlobalEconomy.com (See Table 2).Table 2 Assumption and test for panel data.

Full size table

Tools for econometric analysis and model specification

While thoroughly examining all required assumptions for performing static and dynamic regression, this study made use of a combination of static regression (random effect) and dynamic regression (GMM and FMOLS).

In this study, to determine the most suitable model between fixed effects (FE) and random effects (RE) estimations, the Hausman test was conducted. This test helps to decide which model provides more consistent estimates based on the assumption that the random effects are uncorrelated with the explanatory variables. According to the test results, RE is selected across all panels (Table 3).Table 3 Test for choosing the appropriate model.

Full size table

Additionally, to select the appropriate model between pooled ordinary least squares (POLS) and RE/FE, the Breusch–Pagan Lagrange multiplier (LM) test was performed.

Based on the research of (Almaqtari et al., 2019; Bekele, 2023; Brooks, 2019; Chowdhury and Rasid, 2016; Masood and Ashraf, 2012; Menicucci and Paolucci, 2016), and others, the following random-effect regression model establishes the basic structure and framework of the panel data.

Yit=∝+Xitβ+ui+εit

(1)

The aforementioned model mentioned can be represented in the following practical and operational format: where Yit denotes the dependent variable (Profitability), α represents the intercept term for the explanatory variables, β is a k × 1 vector of parameter to be estimated, and the of observations vector Xnt, is 1 × k, t = 1 …, T; n = 1, …, N. The above RE can be transformed into three vector models as follows:

PROit=∝+β1Xit+β2Yit+β3Zit+ui+εit

(2)

Where PROit is the dependent variable (profitability) for bank i at time t, Xit is the vector of bank-specific variables including credit risk (CR), capital adequacy (CA), income diversification (IND), productivity (PRO), expense control (EC), Yit is the vector of bank-specific variable that include bank stability (BS), Zit is the vector of macroeconomic variables including government effectiveness (GE), Gross domestic Product (GDP), innovation index (II), and Covid-10 (COV), ui represents the random effects (unobserved heterogeneity specific to each bank), and εit is the error term.

This study also employed the generalized method of moments (GMM) estimators, designed for dynamic panel data models, over random effects (RE). GMM was preferred because of its ability to address specific issues related to endogeneity, heterogeneity, and dynamic relationships in panel data (Arellano and Bond, 1991; Arellano and Bover, 1995; Holtz-Eakin, 1994)

PROit=∝+λPROit−1+β1Xit+β2Yit+β3Zit+ui+εit

(3)

Where PROit is the dependent variable (profitability) for the bank i at time t, PROit−1 is the lagged dependent variable (Profitability at time t-1), Xit is the vector of bank-specific variables, Yit is the vector of bank-specific variable, and Zit is the vector of macroeconomic variables, ui represents the unobserved individual effects (random effects). ϵit is the idiosyncratic error term, ∝ is the constant term, λ,β1, β2, and β3 are the coefficients to be estimated.

To enhance the robustness of the analysis, the paper also conducts several diagnostic tests, including first-generation panel unit root tests, panel autocorrelation tests, panel heteroskedasticity tests, and panel cointegration tests (see Appendices 1 through 5). These rigorous tests ensure the reliability of the chosen models and address potential issues such as non-stationarity, serial correlation, heteroskedasticity, and long-run relationships between variables, thereby strengthening the overall findings of the study Table 4.Table 4 Variable definition, notations, and measurement.

Full size table

The choice of estimation method depends on the presence of cointegration between variables. When cointegration exists, OLS becomes unreliable due to unobserved panels and time effects. This study adopts panel cointegration estimation models, including fully modified ordinary least squares (FMOLS). Alongside conventional panel models, FMOLS is a nonparametric method that provides optimal estimates in the presence of serial autocorrelation, heteroskedasticity, and endogeneity (Dzingirai and Dzingirai, 2024; Kao et al., 1999; Seo et al., 2019; Seo and Shin, 2016). It yields reliable results even with cross-sectional dependency. Following (Khan et al., 2019), the FMOLS estimator can be written as follows.

βNT∗−β=(∑i=1N⁡L22i−2∑i=1T⁡(Xit−X¯it)2)=∑i=1N⁡L11i−1L22i−1(∑I=1T⁡(Xit−X¯i)μit∗−Tγ^i)

(4)

Where, μit∗=μit−L^21iL^22iΔxit,γ^it=Γ^21iΩ^21i0−L^21iL^22i(Γ22i∗+Ω^22i0 and Li^ was the lower triangulation of Ω^i.

Variable notation, measurement, and expected sign

Data analysis and results

Descriptive analysis

The financial performance of the sampled banks reveals significant insights into their profitability, highlighted by key metrics such as ROA, NIM, and ROE. With an average ROA of 2.72, banks generate varying returns, from losses of −1.1638 to impressive gains of 6.3477, indicating disparities in asset utilization. The average NIM of 5.4899 reflects substantial interest income, though it ranges from 1.4640 to 9.26, suggesting diverse strategies in interest management. Remarkably, the average ROE is 22.236, showcasing extraordinary returns for shareholders, despite considerable variation among banks.

Table 5 also reveals crucial insights into the financial dynamics of 182 observations, highlighting the interplay of various metrics. The average CR stands at a mere 0.02, indicating effective management of potential loan defaults, with a range from 0 to 0.09. CA averages 0.14, suggesting a solid buffer against economic fluctuations, although the maximum of 1.04 shows that some institutions are exceptionally well-capitalized. IND, averaging 1.08, points to a strategic approach in revenue generation, while the standard deviation of 0.80 indicates significant variability, suggesting opportunities for some firms to enhance their revenue sources. Pro is at 0.80, with a max of 3.39, signaling diverse efficiency levels across the sample.Table 5 Summary of descriptive statistics Obs 182.

Full size table

Furthermore, BS averages a robust 10.05, reflecting a sound financial environment, though the standard deviation of 1.20 suggests differences in stability among institutions. In contrast, the GE score of −0.62 indicates challenges in governance, potentially impacting operational performance. The GDP growth rate of 8.48 suggests a vibrant economic backdrop, which is essential for financial institutions to thrive. The Innovation Index, averaging 21.95, underscores a commitment to innovation, vital for maintaining competitiveness. Overall, the mean, min, max, and standard deviations provide a comprehensive picture of the sector’s resilience and adaptability, highlighting both strengths and areas for improvement in financial practices. The COVID-19 variable has a mean value of 0.231, indicating a relatively low average level in the sample. With a standard deviation of 0.422, there is sensible variability around the mean, and the values range from 1.000 to 2.000.

Correlation analysis

Table 6 presents a correlational matrix highlighting the significant relationships among various financial metrics, providing valuable insights into their interdependence. Notably, the ROA shows a strong positive association with ROE at 0.52, signaling that firms achieving higher asset returns also tend to provide better returns to their shareholders. This relationship underscores the importance of asset efficiency in driving overall profitability.Table 6 Correlation matrix.

Full size table

Furthermore, the NIM is negatively allied with both ROA (−0.24) and ROE (−0.27), suggesting that as banks increase their interest income relative to their assets, the overall return might decrease, possibly due to higher risk profiles or competitive pressures. CA presents a significant negative correlation with ROE (0.32), indicating that higher capital reserves may be associated with lower returns on equity, potentially reflecting conservative risk management strategies.

Among the other variables, the IND exhibits a negative correlation with both ROA, ROE, and NIM, suggesting that firms with diversified income streams may face challenges in maintaining high profitability. Additionally, the II is negatively correlated with PROD (−0.64), indicating that increased innovation efforts might not directly translate to enhanced operational efficiency. Overall, these significant correlations highlight the intricate balance financial institutions must navigate between profitability, risk management, and innovation, providing a roadmap for strategic decision-making in the sector. Overall, all the correlation coefficients between the independent variables are below 0.80, indicating that the variables are largely free from multicollinearity (Lind et al., 2012).

Stationarity test

Table 7 displays the outcomes of the unit root tests, which provide crucial insights into the stationarity of various financial variables, highlighting their long-term properties. For ROE, the Levin, Lin and Chu t* statistics of −6.20 and the Im, Pesaran, and Shin W-stat of −4.06, both significant at p < 0.001, confirm that ROE is stationary at levels. In contrast, the ROA exhibits similar behavior, with a t-statistic of −4.51, indicating stationarity at the level as well. Notably, the NIM and CR show differing results. While NIM’s levels test reveals non-stationarity (t-statistic of −0.91), its first difference is significant, suggesting that it becomes stationary when differenced. Conversely, CR is stationary at levels with a t-statistic of −4.21. Other variables, such as CA and IND, exhibit mixed results, with CA showing non-stationarity at levels and significance upon differencing. The EC variable also indicates stationarity only after first differencing, while BS shows strong stationarity at levels. Overall, these results imply that while some financial metrics exhibit stable long-term behavior, others require differencing to achieve stationarity. This is critical for further time-series analyses and regression modeling, as non-stationary data can lead to misleading inferences. Understanding these dynamics will enhance the robustness of future analyses in the financial sector. without the risk of spurious results and enables accurate forecasting, thereby making conclusions more credible in time-series contexts.Table 7 Panel unit root test and panel cointegration test.

Full size table

According to the set rules, if we accept the null hypothesis (H0) with a p value higher than 5%, it implies that there may be a unit root, which would indicate non-stationarity. On the other hand, if the alternative hypothesis (H1) is accepted and the p value is less than 5%, the dataset is assumed to be stable because there is no unit root. Since every variable in this study, including COV, has a probability value of less than 0.05 in both level and first difference, the present study’s statistical analysis shows that every variable is devoid of unit roots. Because of this, we end up accepting the alternative hypothesis and rejecting the null hypothesis. Thus, the current study can proceed with OLS-based regression, FE and RE with assurance.

Panel cointegration test

After confirming stationarity, the next step is to assess long-term cointegration between the outcome variable and the regressors. One common method for this is the residual-based Pedroni test. The Pedroni test includes three statistics: the modified Phillips–Perron (MPP) t-test, the Phillips–Perron t (PP)-test, and the augmented Dickey–Fuller (ADF) t-test. For the panel ROA, ROE, and NIM, two of the Pedroni tests (modified Phillips–Perron and Phillips–Perron) indicate cointegration, while the Augmented Dickey–Fuller t-test does not. This suggests the need for both dynamic and linear regression models.

Discussion

The empirical findings from the models are summarized in Tables 8–10, showcasing RE, GMM and FMOLS. Each model categorizes factors influencing profitability into bank-specific, industry-specific, and macroeconomic groups, using ROA, ROE, and NIM as dependent variables.Table 8 Panel 1-return on assets (ROA).

Full size tableTable 9 Panel 2-return on equity (ROE).

Full size tableTable 10 Panel 3-net interest margin (NIM).

Full size table

Panel 1- ROA

A random effect (RE) analysis of the variables affecting ROA is shown in Table 8. Key variables—IND, EC, GE, II, GDP, and COV—show statistically significant impacts on ROA, with IND, GE, and COV at the 1% level, and EC and II at the 5% level, and lastly GDP at 10% level. Positive coefficients for these variables indicate that diverse income sources contribute to higher bank profitability, as evidenced by antecedent research like (Addai et al., 2022; Kwaku et al., 2023; Li et al., 2021). Studies, including those by (Bernardelli and Carrasco-Gutierrez, 2024; Chand et al., 2021; Mirović et al., 2023; Yuan et al., 2022), show that GDP growth boosts bank profitability through increased demand for loans and services. This demand results in higher transaction volumes and lower default rates, linking economic expansion to improved performance metrics. A higher II signifies a bank’s ability to innovate, attract customers, and improve service delivery. Ultimately, prioritizing innovation is associated with greater financial success, supported by multiple studies demonstrating its positive impact on profitability (Ashiru et al., 2023; Atukunda et al., 2024; Kebede, 2024).

Based on a system GMM model, the positive coefficients for GE, GDP, and II, Lag (1) ROA indicate that improved technology and innovation, robust economic growth, effective government policies, and last year’s performance measured by ROA collectively enhance asset returns. These findings are consistent with previous research, such as (Almaskati, 2022; Batten and Vo, 2019; Isayas, 2022; Mirović et al., 2023; O’Connell, 2023), which similarly identifies the positive impact of these factors on financial performance. Conversely, the negative coefficients for CR and IND reveal potential inefficiencies; more credit risks may lead to underutilization of assets, while excessive diversification can dilute focus and impede overall performance. This nuanced interplay emphasizes the critical balance firms must achieve between prudent capital management and strategic focus to optimize their ROA, aligning with earlier studies like (Jigeer and Koroleva, 2023; Le and Nguyen, 2024; Saif-Alyousfi, 2022) that highlight the importance of strategic alignment in achieving financial efficiency.

Using an FMOLS model, Table 8 further analyzes the variables affecting ROA and shows the important links and implications between the variables. The positive coefficients for GE and GDP indicate that these elements work synergistically to enhance financial performance. Robust economic growth (GDP) fosters a favorable investment environment, and GE ensures supportive regulations and policies. (Chand et al., 2021; Mashamba and Chikutuma, 2023; Medyawati et al., 2021) reinforcing the established link between these factors and financial success. Conversely, the negative coefficients for IND and CA reveal potential pitfalls; excessive diversification can dilute strategic focus, leading to scattered resources and ineffective decision-making, higher capital ratio enhances financial stability and reduces risk, it can also result in a lower ROA. This is because a larger capital base leads to fewer assets being used to generate profits, effectively diluting the return generated from those assets. This interplay underscores the critical balance firms must strike between prudent capital management and strategic alignment, echoing insights from earlier studies like (Bortoluzzo et al., 2024; Gazi et al., 2024; Madugu et al., 2020; Mutianingsih et al., 2023) that highlight the importance of coherent strategies for optimizing financial efficiency and enhancing ROA.

Panel 2-ROE

Table 9 reveals a captivating analysis of ROE as it intertwines with bank-specific, industry-specific, and macroeconomic factors. Across three robust models, the results consistently highlight that productivity (PROD), government effectiveness (GE), innovation indexes and first lagged of ROE have a significant positive impact on profitability as gauged by ROE. Notably, capital adequacy (CA) and income diversification (IND), emerges as a key player, showing significant negative effects in both RE, system GMM, and FMOLS,

IND significantly and negatively influences profitability measured by ROE at a striking 1% significance level (Prob = 0.01). This resonates with earlier studies (Edirisuriya et al., 2015; Kwaku et al., 2023; Mercieca et al., 2007), which argue that an overly diversified income stream can actually dampen a bank’s profitability. In a contrasting perspective (Chand et al., 2021; Li et al., 2021; Mashamba and Chikutuma, 2023; Mirović et al., 2023) claims that income diversification either positively influences profitability

Moreover, productivity (PROD) emerges as a noteworthy player, significantly influencing ROE at the 5% level (P value < 0.10), albeit with a negative coefficient. This finding aligns with the research of (Hale and Caliskan, 2020; Llorens et al., 2020), suggesting that heightened productivity may not necessarily translate to improved profitability for Ethiopian banks. In sharp contrast, other studies (Athanasoglou et al., 2008; Kosmidou, 2008; O’Connell, 2023) (Kosmidou, 2008) paint a different picture, highlighting a positive relationship between productivity and ROE.

The findings reveal that II has a positive effect on ROE. This result aligns with previous research (Citterio et al., 2024; González et al., 2019; Lee et al., 2020), which found similar outcomes, indicating that innovation boosts bank profitability by improving efficiency, reducing costs, and creating new revenue streams.

The results for government effectiveness (GE) meet expectations across the RE and system GMM, showcasing a positive significance at the 5% level (P value = 0.05) the finding is in line with by prior research (Chand et al., 2024; Hajer and Anis, 2018; Jagannath and Maitra, 2023; Ullah et al., 2024). In a compelling parallel, the capital adequacy (CA) ratio reveals a strong negative influence across all models, aligning with findings from studies by (Mashamba and Chikutuma, 2023; Mulbah et al., 2024; Mutianingsih et al., 2023; Radovanov et al., 2023).

Panel-3 NIM

An informative examination of NIM is shown in Table 10, which also examines the effects of the various determinants and profitability as measured by NIM.

For Ethiopian commercial banks, EC significantly and positively influences profitability measured by NIM at a striking 10% significance level (Prob = 0.10). This resonates with earlier studies (Islam and Nishiyama, 2016; Jima, 2023) which argue that cutting unnecessary costs and expenses improves banks’ profitability. In a contrasting perspective (Saksonova, 2014) claims that income diversification either positively influences profitability

COVID-19 (COV) shows a strong positive influence on NIM in FE and FMOLS, presenting a compelling parallel. It also demonstrates a significant positive impact on ROE at the 5 and 10% levels, respectively. This finding concurs with (Susanti et al., 2023; Xiazi and Shabir, 2022) and contradicts with (Ho et al., 2023; Mateev et al., 2024; Shabir et al., 2023).

Moreover, BS emerges as a noteworthy player, significantly influencing NIM at the 1, 5, and 10% level of significant albeit with a negative coefficient. This finding aligns with the research of (Mkadmi et al., 2021; Tan, 2015) suggesting that heightened productivity may not necessarily translate to improved profitability for Ethiopian banks. In sharp contrast, other studies (Kebede, 2024; O’Connell, 2023; Shahriar et al., 2023) paint a different picture, highlighting a positive relationship between productivity and ROE.

According to Table 10, GDP plays a vital role in influencing NIM demonstrating a significant positive in RE and FMOLS 1% and 5% level of significance level respectively. This finding echoes earlier research (Chand et al., 2024; Isayas, 2022; Mirović et al., 2023; Negash, 2023; Yuan et al., 2022), which underscores GDP as a key driver of bank profitability. However, the narrative is not without its contradictions; some studies (Ameur and Mhiri, 2013; Lamothe et al., 2024; Naceur and Omran, 2011; Tarus et al., 2012) suggest a statistically significant negative impact of GDP on ROE, with others noting a negative but insignificant relationship

Conclusion

The purpose of this study was to examine the key factors influencing the profitability of banks in Ethiopia between 2011 and 2023. The study used ROA, ROE, and NIM as profitability measures, analyzing them against a range of bank-specific, industry-specific, and macroeconomic variables through both static (RE) and dynamic (GMM and FMOLS) regression.

The analysis reveals that when profit is measured using ROA, variables such as credit risk, income diversification, expense control, and COV are negatively correlated with ROA. Conversely, GDP, government effectiveness (GE), and the lagged value of ROA (Lag(1) of ROA) are positively correlated with ROA. Theoretically, this research aligns with Growth-Led Finance and Competition-Stability Theory, which propose that enhanced access to finance drives growth, and a stable competitive environment encourages efficiency. These factors are expected to influence organizational practices, productivity and resource allocation, supporting the overall profitability of banks.

The regression analysis highlights some interesting patterns. For example, while CA and IND show negative influence, factors such as GE, II, and the lagged ROE appear to be positive drivers of profitability, as measured by ROE.

The findings also indicate that bank stability, as per the competition-fragility theory, plays an indispensable role in profitability, particularly when gauged by NIM. However, a certain variable IND and lagged of NIM shows inconsistent effect. For instance, while they show a positive impact on the RE and FMOLS, they exhibit a negative effect under the GMM model. Interestingly, COV unexpectedly shows a positive effect on bank performance, in contrast to the proposed hypothesis, suggesting that further investigation is needed to understand its role.

The study also finds that, according to Efficiency Structure Theory, the negative correlation between the profitability and productivity of Ethiopian banks may be due to the potential efficiency over-investment costs that lead to increased operation complexity and cost, thereby eroding profit margins. The over-focus on productivity also reduces flexibility to market and regulatory changes, which in turn impacts profitability.

Some independent variables—such as credit risk, expense control, bank stability, and the innovation index—show insignificant coefficients across different models employed. This might be due to the differences in the sizes of sampled banks, which might cause variation or perhaps outliers. Also, measurement errors together with data inaccuracies could undermine the validity of the findings, making it impossible to draw solid conclusions regarding these variables.

Theoretical and practical implications

Based on the findings, the study offers the subsequent theoretical and practical recommendations:

Limitations and future research directions

The current study only considers 14 banks taking as a sample, thoughs this limited sample could not potentially affect the generalizability of the findings, the study acknowledges that a larger sample of banks from a different nation would strengthen the robustness and wider applicability of the findings. As this study focuses solely on commercial banks, forthcoming research will explore profitability determinants in other financial institutions, such as insurance companies, microfinance institutions, and non-financial organizations, as these sectors may present distinct profitability dynamics, and their inclusion in future research could offer a more comprehensive understanding of financial performance across different industries. Furthermore, utilizing advanced methodologies like ARDL, DOLS, and threshold first difference generalized methods of moments (TFDGMM) estimation methods. These techniques would provide a better-balanced examination by overcoming potential endogeneity and increasing estimation accuracy. Furthermore, an integration of qualitative methods and quantitative methods would provide insights into the underlying facets of factors determining profitability across diverse institutional settings.

Data availability

The data that support this study’s findings are sourced from the National Bank of Ethiopia (NBE) (https://nbe.gov.et/) and the GlobalEconom.com (https://www.theglobaleconomy.com), but restrictions apply. These data were used under license for the current study and are not publicly available. However, they are available from the corresponding author upon reasonable request and with the permission of the National Bank of Ethiopia and GlobalEconom.com.

References

Download references

Acknowledgements

Not Applicable.

Author information

Authors and Affiliations

  1. Hawassa University, Hawassa, EthiopiaTekalign Negash Kebede

Contributions

The author bears exclusive responsibility for the idea, planning, analysis, and interpretation of the data; writing the paper and carefully reviewing it for intellectual substance; and approving the final version that will be published.

Corresponding author

Correspondence to Tekalign Negash Kebede.

Ethics declarations

Competing interests

The authors declare no competing interests.

Ethical approval

This article does not contain any studies with human participants performed by any of the authors.

Informed consent

This article does not contain any studies with human participants performed by any of the authors.

Additional information

Publisher’s note Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.

Open Access This article is licensed under a Creative Commons Attribution-NonCommercial-NoDerivatives 4.0 International License, which permits any non-commercial use, sharing, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons licence, and indicate if you modified the licensed material. You do not have permission under this licence to share adapted material derived from this article or parts of it. The images or other third party material in this article are included in the article’s Creative Commons licence, unless indicated otherwise in a credit line to the material. If material is not included in the article’s Creative Commons licence and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder. To view a copy of this licence, visit http://creativecommons.org/licenses/by-nc-nd/4.0/.

Reprints and permissions

About this article

Cite this article

Kebede, T.N. Unveiling the drivers of bank profitability: insights from Ethiopian banks. Humanit Soc Sci Commun 12, 721 (2025). https://doi.org/10.1057/s41599-025-05031-3

Download citation