Monday, November 1, 2010

Betting On A 317-Year Pattern

Betting On A 317-Year Pattern
Today is start of 6-month seasonally favorable period


By Mark Hulbert, Marketwatch | 1 November 2010

CHAPEL HILL, N.C. (MarketWatch)— Did you sell in May and go away?

If so, and you're a follower of the seasonal pattern known as the Halloween Indicator, now is the time to get back into the stock market. But how strong is this pattern, really? How can we know for sure that it genuinely exists, and is not just a statistical fluke?

Those are the right questions that investors should be asking. Fortunately, just in time for Halloween and the beginning of the six-month seasonally favorable period, a new academic study has begun circulating with lots of data to help answer those questions. The study, "Are Monthly Seasonals Real? A Three Century Perspective," was written by Ben Jacobsen, a finance professor at Massey University in New Zealand, and Cherry Zhang, a Ph.D. student at the same institution.

The researchers carefully analyzed monthly returns for the U.K. stock market back to 1693, more than 300 years ago, looking for evidence of any seasonal patterns— not just the Halloween Indicator. Of all the well-known patterns— including the most famous of all, the so-called January Effect— the Halloween Indicator is the only one that persisted throughout the three-century sample. To be sure, they didn't find that the Halloween Indicator existed in each and every year. But that's hardly surprising, since no indicator works all the time.

But over progressively longer horizons, the success rate grew to very impressive and consistent levels. It beat the market in 71% of all two-year periods since 1693, and 82% of five-year periods. When the researchers focused on 10-year periods, they found that the strategy beat the market 92% of the time.

Consider the strategy's performance in the U.S. in recent years. According to the Hulbert Financial Digest, you would have beaten a buy-and-hold by 3.6 percentage points per year over the last five years if you had invested in riskless Treasury bills from May Day through Halloween and been 100% invested in stocks the other six months of the year. As always, however, the usual caveats apply: There are no guarantees. Even if this new study helps to convince you that the Halloween Indicator is the Real Deal, you still don't have any assurance that the future will be like the past.

And even if the future is like the past, you need to be prepared for the strategy failing in individual years. The most recent failure in the U.S. came during the recent bear market: From Halloween 2008 to May Day 2009, the Dow Jones Industrial Average (DJIA) fell 12.4%. But, especially when compared to almost all other popular indicators, which have no statistical significance whatsoever, it's comforting to come across one that's been working for more than 300 years.

No comments:

Post a Comment