McGraw-Hill

146 Selectine Securities
FIGURE 4-3
The Holiday Effect (Average Preholiday Returns)
New Year's
PresidenYs Day
Good Friday
Memorial Day
Fourth of July
Labor Day
Thanksgiving
Christmas -
Avg. Nonholiday
I
1 I I I I
0 0.125 0.25 0.375 0.5 0.625 0.7
Return (percent)
Saurce: Data horn R. Adel. Hlph stak rem befom holidays: Sloan Workitq Paper, Massach-
Technology. Cambridge. 1984.
InsUWe of
Blue Chip issues toward year In end. any case, the dollar magnitude
of this year-end, large-capitalization stock rally is several times the
magnitude of the more well-known January small-size effect.
The holiday anomaly appears fairly stable over time. In the
most recent decade, however, preholiday returns not have been exceptional.
Nevertheless, the effect does not appear be a to statistical
artifact. For instance, it is not driven by outliers, as 75 percent of
preholiday days are up, versus only 54 percent of all trading days
[Ariel(1984)].
The settlement process, discussed as a potential explanation
for the day-of-the-week effect, has complex implications for fluc
tions in value around holidays [Lakonishok and Levi (1982)l. For
example, this theory predicts a high Thursday return preceding a
Friday holiday, which is what occurs. But it predicts a lower-thanaverage
Friday return preceding a Monday holiday, this and is not
consistent with empirical results. Moreover, the magnitude of any
value changes occurring because of settlement procedures is much
too small to account for the holiday effect.