sample of UAE banks as well as for the sample of CBs but
insignificant in the group of IBs. Again, we find strong
evidence that the effect of bank risk-taking on cost efficiency
is significantly different between the two groups of banks.
Hypothesis HA5 is thus confirmed. CBs strive to improve
their cost efficiency by reducing the risk more than IBs.
HA5: There is a significantly different effect of risk-taking
on the UAE banks’ cost efficiency between the
conventional and Islamic banking systems.
The risk effect on bank profitability as measured by EARTA
is presented in Table 7. We do not observe a significant
influence of bank risk-taking on profitability in both samples
(CBs and IBs). It means that risk-taking activities do not
improve bank profitability of either type of banks. Table 8
shows the risk effect on bank profitability as measured by
EARGL. We find no significant influence of bank risk on
profitability in the group of CBs but a significant negative
relationship between Log(Z) and IBs’ profitability (see
Model 11 and 12); this provides partial evidence in support
of the hypothesis (HA6) that risk-taking has a different
impact on bank profitability of the two banking systems in
the UAE.
HA6: There is a significantly different effect of risk-taking
in the UAE banks’ profitability between the
conventional and Islamic banking systems.
The effect of a bank’s ownership on cost efficiency as measured by CIR is presented in Table 5. We do not find evidence for a significant effect of ownership type on cost
efficiency. All the measures of ownership structure (Ownership concentration, Government ownership, and Foreign
ownership) are statistically insignificant for both CBs and
IBs. Thus, we are unable to confirm our hypothesis that the
effect of ownership structure is different between the two
types of banking systems. However, the results reported in
Table 6 show that both Government ownership and Foreign
ownership have a strong influence on a bank’s cost efficiency
(both coefficient estimates are negative and significant).
Furthermore, Foreign ownership effect is strongly significant in the sample of CBs but insignificant for IBs. This
result is in line with our findings reported in Table 4 that
shows a significant difference in ownership structure
between CBs and IBs. It seems that IBs have a higher percentage of ownership concentration (47.70% vs. 34.13%),
whereas the share of Government ownership is larger in CBs
(11.61% vs. 5.20%). Hypothesis HA7 is thus confirmed.
HA7: There is a significantly different effect of ownership
structure on the UAE banks’ cost efficiency between
the conventional and Islamic banking systems.
The effect of a bank’s ownership structure on bank profitability as measured by EARTA is reported in Table 7.
Similar to the results in Table 5 for cost efficiency effect, we
do not observe a significant influence of ownership structure
on bank profitability; all the measures of ownership structure
are statistically insignificant for CBs. There is a marginally
significant effect of Foreign ownership on bank profitability
only in the sample of CBs (see Model 8). The results for
ownership structure effect on bank profitability as measured
by EARGL are presented in Table 8. Once again, no significant relationship between different types of ownership
and bank profitability is observed. As a result, we may
conclude that the type of ownership structure does not affect
the bank profitability of either type of banks. Thus,
hypothesis HA8 is rejected.
H08: There is no significantly different effect of ownership structure on the UAE banks’ profitability
between the conventional and Islamic banking
systems.
5 Robustness Checks
In addition to the previous tests where we use alternative
measures of cost efficiency and profitability, we check the
robustness of our results by running the regression analysis
separately for three different periods: the full sample period
(2005–2015), the precrisis and crisis period (2005–2009),
and the postcrisis period (2010–2015). The results are
reported, respectively, in Tables 9, 10, 11, and 12. The effect
on a bank’s cost efficiency as measured by CIR before,
during, and after the crisis is presented in Table 9. We find
evidence that banks with a higher return on assists
(ROA) were able to achieve better efficiency in the precrisis
period, including the crisis years, and in the full sample
period; however, this effect is insignificant in the postcrisis
period. In opposite, the effect of Income Diversity seems to
be more pronounced in the postcrisis period. Bank size also
appears to be a relevant determinant of efficiency performance during the precrisis and crisis periods.
The capital adequacy measures are all significant in the
postcrisis period and the full sample period; however, they
do not affect a bank’s cost efficiency during the precrisis and
crisis period, this is a result of enhancing the capital
requirements after the crisis period (2008–2009). Islamic
bank’s dummy variable is significant in all models (except
for precrisis and crisis period), which shows that the efficiency effects are more pronounced in the sample of IBs than
in the sample of CBs. Similarly, the effect on a bank’s cost
efficiency as measured by NIM is presented in Table 10. The
Banking System in the MENA Region: A Comparative Analysis …
79
insignificant in the group of IBs. Again, we find strong
evidence that the effect of bank risk-taking on cost efficiency
is significantly different between the two groups of banks.
Hypothesis HA5 is thus confirmed. CBs strive to improve
their cost efficiency by reducing the risk more than IBs.
HA5: There is a significantly different effect of risk-taking
on the UAE banks’ cost efficiency between the
conventional and Islamic banking systems.
The risk effect on bank profitability as measured by EARTA
is presented in Table 7. We do not observe a significant
influence of bank risk-taking on profitability in both samples
(CBs and IBs). It means that risk-taking activities do not
improve bank profitability of either type of banks. Table 8
shows the risk effect on bank profitability as measured by
EARGL. We find no significant influence of bank risk on
profitability in the group of CBs but a significant negative
relationship between Log(Z) and IBs’ profitability (see
Model 11 and 12); this provides partial evidence in support
of the hypothesis (HA6) that risk-taking has a different
impact on bank profitability of the two banking systems in
the UAE.
HA6: There is a significantly different effect of risk-taking
in the UAE banks’ profitability between the
conventional and Islamic banking systems.
The effect of a bank’s ownership on cost efficiency as measured by CIR is presented in Table 5. We do not find evidence for a significant effect of ownership type on cost
efficiency. All the measures of ownership structure (Ownership concentration, Government ownership, and Foreign
ownership) are statistically insignificant for both CBs and
IBs. Thus, we are unable to confirm our hypothesis that the
effect of ownership structure is different between the two
types of banking systems. However, the results reported in
Table 6 show that both Government ownership and Foreign
ownership have a strong influence on a bank’s cost efficiency
(both coefficient estimates are negative and significant).
Furthermore, Foreign ownership effect is strongly significant in the sample of CBs but insignificant for IBs. This
result is in line with our findings reported in Table 4 that
shows a significant difference in ownership structure
between CBs and IBs. It seems that IBs have a higher percentage of ownership concentration (47.70% vs. 34.13%),
whereas the share of Government ownership is larger in CBs
(11.61% vs. 5.20%). Hypothesis HA7 is thus confirmed.
HA7: There is a significantly different effect of ownership
structure on the UAE banks’ cost efficiency between
the conventional and Islamic banking systems.
The effect of a bank’s ownership structure on bank profitability as measured by EARTA is reported in Table 7.
Similar to the results in Table 5 for cost efficiency effect, we
do not observe a significant influence of ownership structure
on bank profitability; all the measures of ownership structure
are statistically insignificant for CBs. There is a marginally
significant effect of Foreign ownership on bank profitability
only in the sample of CBs (see Model 8). The results for
ownership structure effect on bank profitability as measured
by EARGL are presented in Table 8. Once again, no significant relationship between different types of ownership
and bank profitability is observed. As a result, we may
conclude that the type of ownership structure does not affect
the bank profitability of either type of banks. Thus,
hypothesis HA8 is rejected.
H08: There is no significantly different effect of ownership structure on the UAE banks’ profitability
between the conventional and Islamic banking
systems.
5 Robustness Checks
In addition to the previous tests where we use alternative
measures of cost efficiency and profitability, we check the
robustness of our results by running the regression analysis
separately for three different periods: the full sample period
(2005–2015), the precrisis and crisis period (2005–2009),
and the postcrisis period (2010–2015). The results are
reported, respectively, in Tables 9, 10, 11, and 12. The effect
on a bank’s cost efficiency as measured by CIR before,
during, and after the crisis is presented in Table 9. We find
evidence that banks with a higher return on assists
(ROA) were able to achieve better efficiency in the precrisis
period, including the crisis years, and in the full sample
period; however, this effect is insignificant in the postcrisis
period. In opposite, the effect of Income Diversity seems to
be more pronounced in the postcrisis period. Bank size also
appears to be a relevant determinant of efficiency performance during the precrisis and crisis periods.
The capital adequacy measures are all significant in the
postcrisis period and the full sample period; however, they
do not affect a bank’s cost efficiency during the precrisis and
crisis period, this is a result of enhancing the capital
requirements after the crisis period (2008–2009). Islamic
bank’s dummy variable is significant in all models (except
for precrisis and crisis period), which shows that the efficiency effects are more pronounced in the sample of IBs than
in the sample of CBs. Similarly, the effect on a bank’s cost
efficiency as measured by NIM is presented in Table 10. The
Banking System in the MENA Region: A Comparative Analysis …
79