Table 6 Endogeneity test using...

Table 6

Endogeneity test using propensity score matching (PSM) analysis

Panel A: First-stage logistic regression results
	Coefficient	z-stat	p-value
LNBSIZE	0.272	0.750	0.450
BINDP	3.467	2.200	0.028
FSIZE	−0.352	−1.530	0.125
ROA	1.311	0.680	0.495
LEV	0.046	1.670	0.096
LFAGE	−0.312	−1.340	0.180
Intercept	4.050	1.600	0.111
Year FE		Yes
Industry FE		Yes
Observations		390
Pseudo R-squared		0.028
Log-likelihood		−219.75

Panel A: First-stage logistic regression results
	Coefficient	z-stat	p-value
LNBSIZE	0.272	0.750	0.450
BINDP	3.467	2.200	0.028
FSIZE	−0.352	−1.530	0.125
ROA	1.311	0.680	0.495
LEV	0.046	1.670	0.096
LFAGE	−0.312	−1.340	0.180
Intercept	4.050	1.600	0.111
Year FE		Yes
Industry FE		Yes
Observations		390
Pseudo R-squared		0.028
Log-likelihood		−219.75

Panel B: Mean test between treatment and control groups
	Treatment	Control	t-test
LNBSIZE	2.286	2.228	0.375
BINDP	0.229	0.236	0.586
FSIZE	10.771	10.725	0.649
ROA	0.049	0.045	0.744
LEV	6.425	6.099	0.732
LFAGE	3.211	3.358	0.046

Panel B: Mean test between treatment and control groups
LNBSIZE	2.286	2.228	0.375
BINDP	0.229	0.236	0.586
FSIZE	10.771	10.725	0.649
ROA	0.049	0.045	0.744
LEV	6.425	6.099	0.732
LFAGE	3.211	3.358	0.046

Panel C. Second-stage regression results of association between women directors and women representation in senior management
	DV = WOMMCDUM			DV = WOMMC
	(1)	(2)	(3)	(4)	(5)	(6)
BGENDIV	6.728^***			0.192^***
BGENDIV	(3.33)			(2.81)
1WOMAN		−0.068			−0.010
1WOMAN		(−0.11)			(−0.67)
≥3WOMEN			2.273^***			0.084^***
≥3WOMEN			(3.84)			(4.06)
LNBSIZE	−0.282	−0.723	−1.515^**	0.023	0.009	−0.010
LNBSIZE	(−0.42)	(−1.02)	(−2.32)	(1.13)	(0.42)	(−0.54)
BINDP	0.426	0.479	1.421	−0.004	0.004	0.038
BINDP	(0.12)	(0.14)	(0.37)	(−0.05)	(0.05)	(0.49)
FSIZE	−1.028^**	−0.772^*	−0.801^*	−0.033^**	−0.029^*	−0.030^*
FSIZE	(−2.23)	(−1.77)	(−1.85)	(−1.99)	(−1.76)	(−1.96)
ROA	−1.315	−1.315	0.170	0.014	0.027	0.061
ROA	(−0.60)	(−0.59)	(0.07)	(0.16)	(0.30)	(0.68)
LEV	0.065	0.030	0.026	0.006	0.005	0.005
LEV	(0.76)	(0.34)	(0.28)	(1.39)	(1.16)	(1.28)
LFAGE	−0.814^*	−0.858^*	−1.090^**	−0.005	−0.009	−0.012
LFAGE	(−1.71)	(−1.87)	(−2.25)	(−0.31)	(−0.53)	(−0.71)
Intercept	11.841^**	11.397^**	13.559^***	0.322^**	0.345^**	0.386^**
Intercept	(2.50)	(2.44)	(2.72)	(2.05)	(2.18)	(2.54)
Observations	168	168	168	198	198	198
Pseudo/Adjusted R²	0.229	0.169	0.254	0.238	0.194	0.299

Panel C. Second-stage regression results of association between women directors and women representation in senior management
	DV = WOMMCDUM			DV = WOMMC
	(1)	(2)	(3)	(4)	(5)	(6)
BGENDIV	6.728^***			0.192^***
BGENDIV	(3.33)			(2.81)
1WOMAN		−0.068			−0.010
1WOMAN		(−0.11)			(−0.67)
≥3WOMEN			2.273^***			0.084^***
≥3WOMEN			(3.84)			(4.06)
LNBSIZE	−0.282	−0.723	−1.515^**	0.023	0.009	−0.010
LNBSIZE	(−0.42)	(−1.02)	(−2.32)	(1.13)	(0.42)	(−0.54)
BINDP	0.426	0.479	1.421	−0.004	0.004	0.038
BINDP	(0.12)	(0.14)	(0.37)	(−0.05)	(0.05)	(0.49)
FSIZE	−1.028^**	−0.772^*	−0.801^*	−0.033^**	−0.029^*	−0.030^*
FSIZE	(−2.23)	(−1.77)	(−1.85)	(−1.99)	(−1.76)	(−1.96)
ROA	−1.315	−1.315	0.170	0.014	0.027	0.061
ROA	(−0.60)	(−0.59)	(0.07)	(0.16)	(0.30)	(0.68)
LEV	0.065	0.030	0.026	0.006	0.005	0.005
LEV	(0.76)	(0.34)	(0.28)	(1.39)	(1.16)	(1.28)
LFAGE	−0.814^*	−0.858^*	−1.090^**	−0.005	−0.009	−0.012
LFAGE	(−1.71)	(−1.87)	(−2.25)	(−0.31)	(−0.53)	(−0.71)
Intercept	11.841^**	11.397^**	13.559^***	0.322^**	0.345^**	0.386^**
Intercept	(2.50)	(2.44)	(2.72)	(2.05)	(2.18)	(2.54)
Observations	168	168	168	198	198	198
Pseudo/Adjusted R²	0.229	0.169	0.254	0.238	0.194	0.299

Note(s): This table presents the results of the propensity score matching (PSM) analysis. Panel A shows the first-stage regression results where the board gender diversity is a binary variable that is regressed on several firm-specific characteristics. Panel B tests the differences in firm characteristics between the treatment (board with at least one woman) and control (board with no woman) groups of firms. Panel C presents the regression models estimated on PSM samples. The t-statistics reported in parentheses in Panel C are based on robust standard errors. Superscripts ***, **, and * denote significance at the 1%, 5%, and 10% levels, respectively. The definitions of variables are given in Appendix

Source(s): Authors’ own work

[ViewLarge]

Sharing Unavailable