In 2004, two professors, Laura Frieder (Purdue) and Avanidhar Subrahmanyan (UCLA), published the results of a study of stock returns surrounding three holidays: St. Patrick's Day, Rosh Hashana (Jewish New Year) and Yom Kippur (Jewish Day of Atonement, 10 days following Rosh Hashana). They looked at both volume and price effects, finding a positive price effect for St. Patrick's Day and Rosh Hashana and a negative one for Yom Kippur. The data covered 1946-2000. The original study didn't really address an important issue: Lunar Cycles. The Jewish Holidays typically begin in mid-to-late September, though they can begin in early September or as late as early October. The Jewish New Year begins on a new moon.
The bullishness around Rosh Hashana, which is well documented over time, could be psychological, as a subset of traders becomes more optimistic (or the lunar cycle influences traders!), while the weakness surrounding Yom Kippur could be either reversion or a reflection of the solemn nature of the holiday. Here is a table of the last few years describing the returns of the S&P 500 for the days between (from the eve of Rosh Hashana to the eve of Yom Kippur) and the next five days:
I included the pre-Holiday YTD returns to get a sense of the type of market we were experiencing. I was kind of surprised by how little the market had moved from 2004-2007 up to that point of the year. Maybe 2010 isn't so atypical!
In any event, it's not clear that the market sells off between the two holidays, though it appears to bounce afterwards, sell-off or not (except for after the second "Yom Kippur Massacre" that took place in 2008!). I am not sure what happened last year, but no indicator works all the time. The fact that the market was up so much already may have played a role.
In addition to the fact that there is an overiding factor of the lunar calendar, another one should be mentioned: Quadruple Witching (options expirations). I checked to see when they were:
- 2010: Last Day of Holidays
- 2009: During the Holidays
- 2008: Before the Holidays
- 2007: Last Day of Holidays
- 2006: Eve of Rosh Hashana
- 2005: Before the Holidays
- 2004: During the Holidays
- 2003: After the Holidays
I didn't see anything from this timing that would add a lot to understanding when the trade works or doesn't work. In any event, this calendar effect doesn't appear to be having as much impact as it might have had in the past, which isn't surprising given all the computers arbitraging away these days.
Disclosure: None
Comments