276 /304
沙賓法
404
條及審計準則第
5
號是否會減少內部控制揭露錯誤？
276
Table 7 Logistic Regression for Restatement and Non-restatement Samples
(Dependent Variable =
EFFECTIVE
)
Predicted Sign
Restatement Sample
Non-restatement Sample
Coefficient
p-value
Coefficient
p-value
SOX404
？
-0.932
0.046
-0.682
0.204
SIZE
＋
0.075
0.306
0.008
0.918
ROA
＋
0.224
0.697
-0.466
0.351
LEV
－
-0.225
0.395
-0.146
0.579
PE
－
0.000
0.914
-0.003
0.077
MB
－
0.010
0.565
0.046
0.013
BIGN
？
0.283
0.259
0.592
0.003
RCP
－
-0.006
0.062
-0.007
0.023
FT
－
-0.356
0.051
-0.285
0.096
AGLOSS
－
-0.503
0.010
-0.715
<0.001
MARKETCAP
＋
0.000
0.246
0.000
<0.001
CONSTANT
2.689
0.034
18.164
0.982
YEAR
(include)
(include)
INDUSTRY
(include)
(include)
LR chi squared
271.78
<0.001
272.12
<0.001
Pseudo R
2
0.172
0.132
Sample size
1853
10774
Note: Variables are defined in Table 2. P-values are based on two-tailed tests.
Table 8 presents the results of applying Model 2 to the subsample of the SOX 404
era. For the restatement sample, the coefficient on
AS5
is 1.413, with a p-value < 0.001,
indicating that the odds of concluding that internal controls are effective for restatement
companies adopting AS5 increases by 411.16%, compared with their restatement
counterparts adopting AS2. This result suggests that the implementation of AS5 has
resulted in a higher Type II error rate. For the non-restatement sample, the coefficient on
AS5
is 1.418, with a p-value < 0.001, indicating that the odds of concluding that internal
controls are effective for non-restatement companies adopting AS5 increased by 412.79%,
compared with the odds of their non-restatement counterparts concluding the same while
adopting AS2. This finding is consistent with our H2b, and it suggests that implementation
of AS5 could improve ICFR-audit efficiency by reducing Type I errors. To sum up, our
evidence shows that even though the more flexible and less prescriptive AS5 can enhance
the efficiency of ICFR audits, it inadvertently lowers public ICFR-disclosure quality,
measured as Type II errors.