Table 5 The impacts of intellectual...

Table 5

The impacts of intellectual capital on bank performance (Using System GMM)

Variables	Model-1		Model-2
Variables	Coefficient	Robust S.E.	Coefficient	Robust S.E.
ROA_t−1	0.19*	0.090	0.17**	0.086
VAIC	0.011**	0.005	–	–
CEE	–	–	0.227***	0.066
HCE	–	–	0.001	0.001
SCE	–	–	0.006	0.007
CAR	0.049***	0.011	0.024*	0.012
BS	−0.021*	0.011	−0.019*	0.010
NPLTL	−0.027**	0.012	−0.029**	0.013
CINEFF	−0.028***	0.006	−0.016**	0.008
IDIV	0.015***	0.004	0.007	0.005
LEV	−0.036***	0.010	−0.012	0.012
GDP	0.008***	0.001	0.012**	0.006
INF	−0.001	0.001	−0.007	0.008
F-Test	244.78***		415.16***
Hansen Test¹	P = 0.139		P = 0.179
AR(1)²	Z = −4.02	P = 0.000	Z = −3.99	P = 0.000
AR(2)³	Z = 0.27	P = 0.68	Z = 0.24	P = 0.72
No. of instruments	11		13
Observations	286		286

Variables	Model-1		Model-2
Variables	Coefficient	Robust S.E.	Coefficient	Robust S.E.
ROA_t−1	0.19*	0.090	0.17**	0.086
VAIC	0.011**	0.005	–	–
CEE	–	–	0.227***	0.066
HCE	–	–	0.001	0.001
SCE	–	–	0.006	0.007
CAR	0.049***	0.011	0.024*	0.012
BS	−0.021*	0.011	−0.019*	0.010
NPLTL	−0.027**	0.012	−0.029**	0.013
CINEFF	−0.028***	0.006	−0.016**	0.008
IDIV	0.015***	0.004	0.007	0.005
LEV	−0.036***	0.010	−0.012	0.012
GDP	0.008***	0.001	0.012**	0.006
INF	−0.001	0.001	−0.007	0.008
F-Test	244.78***		415.16***
Hansen Test¹	P = 0.139		P = 0.179
AR(1)²	Z = −4.02	P = 0.000	Z = −3.99	P = 0.000
AR(2)³	Z = 0.27	P = 0.68	Z = 0.24	P = 0.72
No. of instruments	11		13
Observations	286		286

Note(s): A two-step system GMM dynamic panel estimators is the estimation technique. The dependent variable is bank performance measured by ROA. ^{*, **}and ^*** denote significance at 10%, 5% and 1% levels, respectively. ¹Test of overidentifying restrictions (H_o: overidentifying restrictions are valid). The tests accept the null hypothesis that overidentifying restrictions are valid. ²Arellano-Bond test for the first-order autocorrelation (Ho: no autocorrelation). ³Arellano-Bond test for the second-order autocorrelation (Ho: no autocorrelation). The AR(1) and AR(2) test findings show that autocorrelation exists in the first-order but not in the second-order. All variables are winsorised at the 5% level

Source(s): Table created by authors

[ViewLarge]

Sharing Unavailable