4.4. Index effects and volatility
Cooper and Woglom (2001) argue that the increase in trading that
occurs around the time of an index addition will cause its volatility, a
measure of its total risk, to rise. A combination of increased demand
by index funds, and speculation by arbitrageurs and noise traders
(see also De Long, Shleifer, Summers, & Waldman, 1990) can cause
a temporary increase in volatility. However, provided that inclusion
in the index, or removal from it, does not affect a firm's underlying
operating performance and therefore its cashflows, there should be
no long-run effect on volatility.
We examine in Table 9 whether the volatility of abnormal returns
changes after index recompositions for additions and deletions. We
again employ the sample where outlying observations have been deleted
from the database. In order to minimise the noise that occurs
around the event period itself, we focus on long-run volatility and
measure the standard deviation of returns from 250 days before the
event to 250 days after. When we compare the actual levels of volatility
before and after the event, we find that both fall modestly — by
1.2% for additions and by 6.1% for deletions. If we compute the ratio
of variances of the pre- over post-event periods, this constitutes a
test statistic to examine the null hypothesis that the variances are
equal. Considering the p-values for these tests in the final row of
Table 9, it is evident that there is no significant change in volatility
following index inclusion or removal. This lends support to the notion
that index changes do not have discernible impacts on the underlying
performance of the firm over the long run.