WHAT PRACTITIONERS NEED TO KNOW
Mark P. Kritzman
Event studies measure the relationship between
event that affects securities and the return
securities. Some events, such
as a regulatory
or an economic shock, affect many securi-
ties contemporaneously; other events, such
in dividend policy or a stock split, are
to individual securities.
are often used to test the effi-
cient market hypothesis.
For example, abnormal
returns that persist after
an event occurs or abnor-
mal returns that
are associated with an
the efficient market hypothesis.
Aside from tests
of market efficiency, event studies
in gauging the magnitude of an
A classic event study published
in 1969 by
Fama, Fisher, Jensen,
and Roll examined the im-
of stock splits on security prices.^ The authors
found that abnormal returns dissipated rapidly
the news of stock splits, thus lending
to the efficient market hypothesis.
How to Perform An Event Study in Seven
The following steps describe
one of several
for conducting an event study of a
Define the event and identify the timing
of the event is not necessar-
the period during which the event occurs.
it may be the investment period imnnedi-
the announcement of the event.
Arrange the security performance data relative to
of the event.
If information about the
is released fully on a specific day with time
for traders to react, the day of the
is period zero. Then, measurement
and following the event are
For example, if the 90 trading days pre-
the event and the 10 days following the
are designated as the pre- and post-event
the pre-event trading days would be la-
Mark P. Kritzman, CFA.
a Partner of Windham Capital Manage-
ment in Cambridge, Massachusetts.
f - 90, f - 89, f - 88, . . . , f - 1; the event
0; and the post-event trading days,
t + I,
+ 2r t + 3, . . . , t + 10.
the event is
to each security, these days will differ
in calendar time,
Separate the security-specific component
turn from the security's total return during the pre-
event measurement period.
is to use
the market model
to isolate security-specific re-
turn. First, each security's daily returns during
the pre-event measurement period from
1 are regressed on the market's
the same period. The security-
are defined as the differences be-
the security's daily returns and the daily
returns predicted from
the regression equation
(the security's alpha plus
its beta times the mar-
ket's daily returns). This calculation is described by