Descriptive Statistics
Table 1 presents the descriptive statistics for the variables included in the study. The variable PROFF, representing profitability of banks, has a sample size of 195 observations. The mean profitability is approximately 0.021, with a standard deviation of 0.033. The minimum profitability recorded in the sample is 0.001, while the maximum is 0.167. The variable OE, which measures operating efficiency, has a mean value of 1.48 and a standard deviation of 1.063. The minimum observed value for operating efficiency is 0.077, while the maximum value is 5.
The variable NI, representing non-interest income ratio, has a mean of 0.544 and a relatively large standard deviation of 2.449. The minimum non-interest income ratio is 0, indicating some banks in the sample do not generate non-interest income, while the maximum ratio is 23.438.
Leverage (LV), computed as the ratio of total debt to total assets, has a mean value of 0.615 and a standard deviation of 0.858. The minimum leverage recorded is 0, indicating some banks in the sample do not have any debt, while the maximum leverage is 5.167. Bank size (BnS), measured using the natural logarithm of total assets, has a mean value of 10.203 and a standard deviation of 3.379. The minimum bank size observed is -13.816, indicating the logarithm of assets for some banks is negative, while the maximum bank size is 13.218.
GDP growth (GDPg), representing the annual percentage change in Gross Domestic Product, has a mean value of 8.94 and a standard deviation of 2.048. The minimum observed GDP growth rate is 5.637, while the maximum is 12.551. Inflation (INF), measured by the annual percentage change in consumer prices, has a mean value of 16.378 and a standard deviation of 9.057. The minimum inflation rate recorded is 6.628, while the maximum is 33.25.
These descriptive statistics provide an overview of the central tendency, dispersion, and range of the variables in the study, offering a foundation for further analysis and interpretation of the data.
Table 1
Variable | Obs | Mean | Std. Dev. | Min | Max |
PROFF | 195 | .021 | .033 | .001 | .167 |
OE | 195 | 1.48 | 1.063 | .077 | 5 |
NI | 195 | .544 | 2.449 | 0 | 23.438 |
LV | 195 | .615 | .858 | 0 | 5.167 |
BnS | 195 | 10.203 | 3.379 | -13.816 | 13.218 |
GDPg | 195 | 8.94 | 2.048 | 5.637 | 12.551 |
INF | 195 | 16.378 | 9.057 | 6.628 | 33.25 |
Sources: Own survey 2023 |
Correlations
Table 2 presents the pairwise correlations among the variables in the study. Each cell in the table represents the correlation coefficient between two variables. Starting with the variable PROFF, which represents profitability, it has a perfect correlation of 1.000 with itself (as expected) in the diagonal. The variable OE (operating efficiency) shows a positive correlation of 0.467 with PROFF, indicating a moderate positive relationship between operating efficiency and profitability.
The variable NI (non-interest income ratio) exhibits a weak negative correlation of -0.046 with PROFF, suggesting a minimal negative association between non-interest income and profitability. The variable LV (leverage) demonstrates a strong positive correlation of 0.849 with PROFF, implying a significant positive relationship between leverage and profitability. Bank size (BnS), represented by variable 5, shows a weak negative correlation of -0.224 with PROFF, suggesting a slight negative relationship between bank size and profitability.GDP growth (GDPg), variable 6, exhibits a weak negative correlation of -0.029 with PROFF, indicating a minimal negative association between GDP growth and profitability. Lastly, inflation (INF), variable 7, displays a weak positive correlation of 0.021 with PROFF, implying a minimal positive relationship between inflation and profitability. These pairwise correlations provide insights into the direction and strength of the relationships among the variables in the study, forming a foundation for further analysis and understanding of their interdependencies.
Table 2
Variables | (1) | (2) | (3) | (4) | (5) | (6) | (7) |
(1) PROFF | 1.000 | | | | | | |
(2) OE | 0.467 | 1.000 | | | | | |
(3) NI | -0.046 | -0.171 | 1.000 | | | | |
(4) LV | 0.849 | 0.327 | -0.031 | 1.000 | | | |
(5) BnS | -0.224 | 0.088 | -0.128 | -0.226 | 1.000 | | |
(6) GDPg | -0.029 | -0.509 | 0.164 | 0.162 | -0.244 | 1.000 | |
(7) INF | 0.021 | 0.188 | 0.076 | -0.065 | -0.014 | -0.456 | 1.000 |
Sources: own using STATA 17 |
Hausman Specification Test
Table 3 presents the results of the Hausman (1978) specification test. The test value of the Chi-square statistic is 55.738, indicating the magnitude of the test result. The corresponding p-value is reported as 0.0000, indicating statistical significance at a high confidence level. The Hausman specification test is employed to assess the validity of the fixed effects model versus the random effects model in panel data analysis. The test compares the efficiency and consistency of the estimators under these two models. A significant result, such as the one observed here with a p-value of 0.0000, suggests that the random effects model is not appropriate and the fixed effects model should be employed instead.
The significance of the Chi-square test value and the extremely low p-value provide evidence in favor of using the fixed effects model in this study. It indicates that the fixed effects model offers a better fit to the data and produces more reliable estimates of the relationships among the variables. These findings underscore the importance of using appropriate statistical models and highlight the robustness of the chosen fixed effects model for analyzing the data in this study.
Table 3
Hausman (1978) specification test
| Coef. |
Chi-square test value | 55.738 |
P-value | 0.0000 |
Results of Fixed Effect Model
Table 4 presents the results of the fixed effects model for the variable PROFF (profitability), and the findings align with previous studies discussed. The variable OE (operating efficiency) shows a positive coefficient of 0.001, supporting the importance of operating efficiency in driving bank profitability (Sihotang, Hasanah, & Hayati, 2022). The variable NI (non-interest income ratio) has a statistically significant negative coefficient of -0.001, confirming previous research that a higher proportion of non-interest income relative to total assets is associated with lower profitability (Ahamed, 2017). The variable LV (leverage) exhibits a statistically significant positive coefficient of 0.034, indicating that higher leverage is associated with higher profitability (Bashiru et al., 2023).
Bank size (BnS) also demonstrates a statistically significant positive coefficient of 0.002, corroborating previous findings that larger banks tend to have higher profitability (Merin, 2016). The variables GDPg (GDP growth) and INF (inflation) do not show statistically significant coefficients, suggesting that their impact on bank profitability may not be significant in the Ethiopian banking sector, consistent with previous studies (Lv, Du, & Liu, 2022; Serwadda, 2018). The constant term does not exhibit statistical significance, indicating that its role in explaining the variation in profitability is not significant in this study.
The overall model's R-squared value of 0.773 implies that the independent variables in the fixed effects model explain approximately 77.3% of the variation in bank profitability, further supporting the importance of operating efficiency, non-interest income, leverage, and bank size as determinants of profitability.
These findings corroborate and confirm the relationships between operating efficiency, non-interest income, leverage, and bank size with profitability, as observed in previous studies conducted in different contexts. However, it is essential to consider the specific characteristics and dynamics of the Ethiopian banking sector in interpreting these findings.
Table 4
Results of Fixed Effect Model
VARIABLES | Fixed Effect for PROFF |
OE | 0.001 |
| (0.001) |
NI | -0.001*** |
| (0.000) |
LV | 0.034*** |
| (0.002) |
BnS | 0.002*** |
| (0.000) |
GDPg | -0.001 |
| (0.001) |
INF | 0.000 |
| (0.000) |
Constant | -0.013 |
| (0.010) |
Observations | 195 |
Number of c_id | 15 |
R-squared | 0.773 |
Note: Standard errors in parentheses *** p < 0.01, ** p < 0.05, * p < 0.1 |