Table 4 OLS regression models of AC and...

Table 4

OLS regression models of AC and earnings forecast error

	M1		M2		M3		M4
	Bias	Accuracy	Bias	Accuracy	Bias	Accuracy	Bias	Accuracy
Intercept	**52.20^a	***51.13^a	***46.99^a	***46.11^a	***53.67^b	***54.42^b	35.56	**51.39^b
AUDITCOM	***−25.92^a	***−26.74^b	**−28.26^a	*−26.69^b	*−18.13^c	*−20.45	**−22.59^b	**−21.60^b
LNSIZE							0.97	−2.15
LNAGE	**−10.26^b	*−7.10^c	**−10.23^b	*−8.13^c	**−10.53^b	*−7.44
LNDEBT	**−52.43^b	***−53.82^b	**−51.65^b	***−51.16^b	**−53.45^b	***−55.50^a	**−50.22^b	**−51.53^b
Auditor	***−34.44^b	**−24.07^b	**−36.87^b	**−23.12^b	***−39.64^a	**−27.19^b	***−32.60^a	*−20.93^b
Executives			−0.03	0.07
Property			20.38	17.77
Management					*−25.65	*−23.46
PPP	−5.61	−0.05
Standard							35.56	0.44
N	55	55	55	55	55	55	55	55
Adjusted R²	0.188	0.149	0.236	0.169	0.209	0.189	0.128	0.120
F-test statistic	**2.44	**2.30	**2.53	*1.94	**2.52	*2.07	*2.10	*2.06
VIF	[1.02–1.28]	[1.02–1.28]	[1.02–1.28]	[1.02–1.28]	[1.04–1.23]	[1.04–1.23]	[1.09–1.21]	[1.09–1.21]

	M1		M2		M3		M4
	Bias	Accuracy	Bias	Accuracy	Bias	Accuracy	Bias	Accuracy
Intercept	**52.20^a	***51.13^a	***46.99^a	***46.11^a	***53.67^b	***54.42^b	35.56	**51.39^b
AUDITCOM	***−25.92^a	***−26.74^b	**−28.26^a	*−26.69^b	*−18.13^c	*−20.45	**−22.59^b	**−21.60^b
LNSIZE							0.97	−2.15
LNAGE	**−10.26^b	*−7.10^c	**−10.23^b	*−8.13^c	**−10.53^b	*−7.44
LNDEBT	**−52.43^b	***−53.82^b	**−51.65^b	***−51.16^b	**−53.45^b	***−55.50^a	**−50.22^b	**−51.53^b
Auditor	***−34.44^b	**−24.07^b	**−36.87^b	**−23.12^b	***−39.64^a	**−27.19^b	***−32.60^a	*−20.93^b
Executives			−0.03	0.07
Property			20.38	17.77
Management					*−25.65	*−23.46
PPP	−5.61	−0.05
Standard							35.56	0.44
N	55	55	55	55	55	55	55	55
Adjusted R²	0.188	0.149	0.236	0.169	0.209	0.189	0.128	0.120
F-test statistic	**2.44	**2.30	**2.53	*1.94	**2.52	*2.07	*2.10	*2.06
VIF	[1.02–1.28]	[1.02–1.28]	[1.02–1.28]	[1.02–1.28]	[1.04–1.23]	[1.04–1.23]	[1.09–1.21]	[1.09–1.21]

Note(s): Multiple linear regression models estimated by cross-sectional OLS. Dependent variable is the Earnings Forecast Error (ERROR). Bias regressions use the signed EFE as dependent variable. Accuracy regressions use the absolute value of EFE as dependent variable. Heteroscedasticity has been corrected using White’s methodology

The variables are described in Table A1

VIF: Variance Inflation Factor. Maximum-minimum values are reported

***, **, * significant at the 1%, 5 and 10% levels, respectively

^{a, b, c} significant at the 1%, 5 and 10% levels, respectively, using the bootstrap methodology

Source(s): Table by authors

[ViewLarge]

Sharing Unavailable