31 /274
31
臺大管理論叢
第
27
卷第
4
期
5.3 Alternative QFIIs Volatility Measure Examinations
This study uses the QFIIs’ shareholdings turnover rate to measure QFIIs’ shareholdings
volatility and reruns Equation (3).
10
The turnover rate is measured as the average of the total
number of shares purchased and the sale of sample firms by the QFIIs scaled by the total
number of shares outstanding each year. In this setting, year 2000 is chosen as the starting
year because the turnover rate is available for that year from the TEJ. Consequently, the
observations are reduced to 4,535 in this further examination. In the 5% cutoff models, the
variable QFII_HL (QFII_HH) is denoted as one for a firm in which the QFIIs’ ownership is
at least 5% and their turnover rate is less than or equal to (more than) the industry-year
median turnover rate. Alternatively, in the industry-year median model, the variable QFII_
HL (QFII_HH) is denoted as one for a firm in which QFIIs’ ownership are more than (or
equal to) the industry-year median and their turnover rate is less than or equal to (more than)
the industry-year median turnover rate. The extracted results are presented in Panel C of
Table 7.
In the post-deregulation subperiod, the coefficients of QFII_HH
t
*IS
t
*X
t3
in the 5% and
industry-year median model are -0.083 (
t
= -1.81) and -0.134 (
t
= -3.37), respectively, and
they are negative and statistically significant at the 10% and 1% levels. It is found that the
coefficients of QFII_HL
t
*IS
t
*X
t3
are -0.067 (
t
= -1.59) and -0.102 (
t
= -2.89) in the 5% and
industry-year median model. However, the coefficients of QFII_HH
t
*IS
t
*X
t3
are statistically
insignificant in the pre-deregulation subperiod. These results are approximately the same as
the initial findings. It is interesting to find that the coefficients of QFII_HL
t
*IS
t
*X
t3
are
positive and statistically significant in the 5% cutoff setting. We note that only three years
(2000~2002) of data was available to regress the models in the pre-deregulation subperiod.
Thus, the inconsistency found in the above results may be triggered by the limitation of
sample years.
This study also uses the change of monthly QFIIs’ ownership (ΔQFIIs) in calculating
the coefficient of variation (CV) of ΔQFIIs ownership to proxy QFIIs shareholdings
volatility (ΔQFII (CV)) and reruns Equation (3). A firm is classified as long-term (short-
term) oriented QFIIs which is proxied by low (high) QFIIs shareholdings volatility (i.e.,
10 Institutional investors have different investment goals and philosophies (Bushee, 1998). As the number and
type of institutions increase, the chance for goal and strategy conflicts increases, and the synergy that
institutional investors experience in monitoring management may be lost (Khan, Dharwadkar, and Brandes,
2005). This study also uses the change in the total number of QFIIs to measure QFIIs’ shareholdings volatility
and reruns Equation (3). The results do not qualitatively change the primary findings.