18 /274
公司盈餘平穩化行為與盈餘資訊性之關係－合格境外機構投資者角色之檢測
18
4.2 Prelimentary Association Analysis
This study conjectures that the earnings informativeness of income smoothing is
associated with high QFIIs ownership. It implies that managers know QFIIs’ trading
objectives and/or behaviors and consequently, a large amount of QFIIs’ participation results
in the reporting of distinctive income smoothing. To gain preliminary evidence to support
this conjecture, this study firstly constructs a zero-cost portfolio return to show the
informative component of income smoothing would depend on the monitoring or
opportunistic role of QFIIs played in the managerial earnings reporting. We carry out a two-
way sequential sorting. Firstly, we sort firms into three groups based on earnings (X
t3
), which
is denoted as low X
t3
, middle X
t3
, and high X
t3
groups, respectively. Secondly, within each
tercile, first sort income smoothing, then sequentially sort on QFIIs’ participation level. This
procedure results in nine subsamples (3*3) from the low income smoothing with low QFII
ownership to the high income smoothing with high QFII ownership in each earnings tercile.
We then form three types of zero-cost portfolio return for the high income smoothing with
high QFII ownership (the ninth subsample) to their counterpart subsamples. If the zero cost
portfolio return is negative (or positive) and statistically significant, then the results will
support the notion that QFIIs’ participation somewhat induces firms’ earnings manipulation
(or informativeness) behavior, resulting in information distortion (or reinforcement). The
result is presented in Table 4.
From Table 4, the annual stock returns (R
t
) for the high income smoothing with high
QFIIs ownership subsample are 0.0325, 0.0258, and 0.0608 in each earnings (X
t3
) tercile,
positive and statistically significant. In Panel A, the annual stock returns (R
t
) for the high
income smoothing with low QFII ownership subsample (the seventh subsample) are 0.0796,
0.0495, and 0.0777 in each earnings (X
t3
) tercile, positive and statistically significant. The
zero cost portfolio yields negative annual stock returns of -0.0471 (
p
= 0.0099), -0.0247 (
p
=
0.0729), and -0.0169 (
p
= 0.1011) in the earnings tercile, respectively. The significantly
negative returns to the zero-cost portfolio in the low X
t3
and middle X
t3
are consistent with
notion that QFIIs’ participation somewhat induces firms’ earnings manipulation behavior,
resulting in information distortion. In Panel B, it also revealed that the zero cost portfolio,
which is constructed for the high income smoothing with high QFII and low income
smoothing with low QFII subsamples, yields negative annual stock returns in the respective
earnings tercile. In Panel C, the high/low income smoothing with high QFII yields negative
annual stock returns in the respective earnings tercile, yet is only marginally significant in
the low earnings (X
t3
) tercile. These results to some extent support our conjecture, i.e., the
earnings informativeness of income smoothing is associated with QFII ownership.