17 /274
17
臺大管理論叢
第
27
卷第
4
期
smoothing measure (IS
t
) is statistically insignificant and suggests few earnings
informativeness of income smoothing in our analysis. Consequently, this study divides the
entire observation into two exclusive subsamples basing on the median of QFII ownership,
and re-calculates the correlation coefficients. The untabulated result reveals that IS
t
is
marginally positively-associated with the annual stock returns (R
t
) in the low QFII ownership
subsample (
r
= 0.022, 10%), yet is significantly negatively-associated with R
t
in the high
QFII ownership subsample (
r
= -0.038, 5%). This result provides the preliminary evidence to
support the conjecture that QFIIs play an important role in firm’s earnings informativeness of
income smoothing, yet is conditional on the magnitude of QFIIs ownership. Consistent with
prior research (e.g., Jiambalvo et al., 2002; Velury and Jenkins, 2006; among others), we find
that QFIIs’ ownership is positively correlated with firm size. While most of the independent
variables are highly correlated with the others, the variance inflation factors (VIF) of the
pivotal interactive explanatory variables, i.e., QFII
t
*IS
t
*X
t3
, QFII_HH
t
*IS
t
*X
t3
, and QFII_
HL
t
*IS
t
*X
t3
in the specified regressions are 1.398, 1.122, and 1.197, respectively, and do not
suggest severe multi-collinearity problems (Neter, Wasserman, and Kutner, 1989).
Table 3 Pearson/Spearman Correlation Matrix (N = 5,766)
Variable R
t
X
t-1
X
t
X
t3
R
t3
IS
t
QFII%
t
LEV
t
MB
t
SIZE
t
R
t
-0.056
a
-0.079
a
-0.028
b
-0.126
a
-0.003 -0.036
a
-0.066
a
0.079
a
-0.015
X
t-1
0.026
b
0.462
a
0.403
a
-0.060
a
0.046
a
0.079
a
-0.263
a
-0.065
a
0.219
a
X
t
0.019 0.547
a
0.434
a
-.100
a
-0.022 0.071
a
-0.249
a
0.117
a
0.239
a
X
t3
0.100
a
0.389
a
0.557
a
-.082
a
0.023 0.051
a
-0.155
a
0.026
b
0.148
a
R
t3
-0.176
a
-0.123
a
-0.111
a
-0.065
a
-0.038
a
-0.045
a
0.041
a
-0.127
a
-0.142
a
IS
t
0.001 0.075
a
0.035
a
0.045
a
-0.047
a
-0.046
a
0.087
a
-0.096
a
-0.107
a
QFII%
t
-0.023 0.085
a
0.103
a
0.059
a
-0.040
a
-0.062
a
-0.099
a
0.206
a
0.438
a
LEV
t
0.036
a
-0.178
a
-0.167
a
-0.102
a
0.012 0.098
a
-0.100
a
-0.085
a
-0.152
a
MB
t
0.055
a
0.239
a
0.365
a
0.198
a
-0.197
a
-0.075
a
0.252
a
-0.136
a
0.397
a
SIZE
t
0.013 0.226
a
0.280
a
0.177
a
-0.106
a
-0.101
a
0.484
a
-0.127
a
0.517
a
Legends:
1. The definition of variables see Table 2.
2. The upper triangular of matrix presents Pearson correlation coefficients, and the lower triangular of
matrix presents Spearman correlation coefficients.
3. “a” and “b” denote the significance on 1% and 5% levels respectively, based on two-tailed tests.