21 / 372
臺大管理論叢
第
26
卷第
2
期
21
CETR
) is more negative when auditor independence is lower. The coefficient of the
interaction between
Spec
and
Fee
(
Tenure
) is positive and statistically significant in Model
(3), which means that auditor industry expertise is more positively associated with book-tax
difference (
BTD
) when auditor independence is poor. To summarize, the reported
coefficients of the interactive variables are uniformly consistent with H3 and indicate that
industry experts are more likely to use their industry expertise to help clients engage in tax
avoidance activity, especially when auditor independence is poor.
Results on control variables suggest the following patterns. There is a positive
association between audit firm size and clients’ tax avoidance in Table 8. Corporate efficient
tax rate (
BETR
and
CETR
) is positively associated with
Size
,
Roa
, and
CFO
, and negatively
associated with
Lev
. Corporate book-tax difference (
BTD
) is positively associated with
Size
and
Lev
, which means firms with bigger size and higher leverage are more likely to engage
in tax avoidance activity. The regression results of control variables are basically consistent
with previous researches.
(2)
Tax-avoidance = CETR
(3)
Tax-avoidance = BTD
Spec = IPS Spec = IPS_D Spec = IMS Spec = IMS_D Spec = IPS Spec = IPS_D
-0.2110***
-0.1982***
2.2271***
2.1474***
2.1766***
2.0427***
(-8.92)
(-8.61)
(6.70)
(6.56)
(6.54)
(6.30)
-0.0049
-0.0079
0.8818*
0.1997**
0.3796*
0.0144
(-0.35)
(-1.17)
(1.81)
(2.28)
(1.92)
(0.15)
0.0062
0.0097
0.0912
0.2051
0.0750
0.1711
(0.37)
(0.58)
(0.39)
(0.90)
(0.34)
(0.72)
-0.0001
-0.0004
0.0211***
0.0183***
0.0147**
0.0171***
(-0.28)
(-0.80)
(2.90)
(2.79)
(2.18)
(2.64)
-0.6320**
-0.1122***
9.4932***
1.3349*
12.5783***
0.9830*
(-2.48)
(-2.72)
(2.65)
(1.77)
(3.52)
(1.70)
-0.0086*
-0.0072*
0.1479*
0.0311*
0.0472*
0.0317*
(-1.90)
(-1.89)
(1.68)
(1.66)
(1.77)
(1.66)
0.0025
0.0003
0.1448***
0.1452***
0.1460***
0.1634***
(0.91)
(0.19)
(3.66)
(3.80)
(3.72)
(4.49)
0.0044
0.0043
0.0275
0.0269
0.0271
0.0256