Oggi si chiude il primo trimestre 2010
Timing is everything
How you can profit from seasonal patterns in stocks
SAN FRANCISCO (MarketWatch) -- Mark Twain famously observed that October is one of the most dangerous months to speculate in stocks. The others, he added, "are July, January, September, April, November, May, March, June, December, August, and February."
Maybe the humorist was using a different calendar. There's ample evidence that the stock market's performance is tied to the time of year and even the days of a month. Seasonal patterns persist not just in the U.S., but in other countries as well.
For instance, the quarter is coming to a close and the market's worst six-month stretch is on the horizon. Those calendar effects have influenced stocks before, and you can expect they will again. Stocks, as they always do, follow money flows and fluctuate in ways that echo the past.
What's behind this seasonality? Human nature, said Jeffrey Hirsch, editor-in-chief of the Stock Trader's Almanac, a weathervane for the market's calendar-based moves.
"There is a habitual nature to society and human activity," Hirsch said. "People's behavior and what they do with their money and time bears upon economics and the stock market."
If you recognize these patterns, you can increase the odds of matching or even outperforming the market with considerably less risk. Plus, transaction costs are no longer an issue nowadays using a discount broker and exchange-traded funds such as SPDR S&P 500 (SPY) , iShares Russell 2000 Index (IWM) , or any comparable fund tracking a broad-based benchmark.
One of the most visible calendar patterns is the so-called Halloween effect, also known as the "Sell in May" indicator, which holds that stocks typically are weaker during summer than winter. Another pattern appears in the final trading days of the month and the first trading days of the new month. A subset of that is the "first day of the month trade" -- buying and selling a market index on the first trading day of the month and not going back in until the first day of the next month.
Other market biases surface as well: Stocks tend to be stronger during the middle of the month, particularly over the five trading days before St. Patrick's Day, March 17. As part of that, Hirsch said, the ninth trading day of March has been positive for the Nasdaq index more than 70% of the time since 1986, including most recently its rise on Thursday. Stocks also tend to rise in the two or three trading days before a market holiday, such as July Fourth or Christmas.
To be sure, many skeptics dismiss calendar effects as random events, giving them about as much predictive credit as astrology. Naysayers have even more reason to disbelieve after the past couple of years. Going long U.S. stocks in November 2008 and holding through April 2009 would have cost you big money. You'd have compounded the injury by selling then and sitting out until November as stocks recovered.
Accordingly, use these indicators as a market guide, not a GPS. "These things don't happen every time; it's a general tendency," said Ed Clissold, senior global analyst at market strategists Ned Davis Research. "You're talking about odds that are modestly better than 50-50. You have to look at them in the context of what else is going on in the market."
Moreover, individual investors tend to lack the discipline such trading strategies demand. If you do attempt seasonally driven trades, venture just a small portion of your money. And be wary of "experts" peddling timing systems that purport to outperform the market year-in and year-out.
"It's like being a card counter; you have to play many rounds to get the numbers in your favor," said Mark Hulbert, editor of the Hulbert Financial Digest, which tracks the performance of investment advisory newsletters and is a service of MarketWatch, the publisher of this report. "There's never a guarantee that these systems will work every year. The merit is to come close to the market's return while incurring below-average risk."
1. Halloween effect
"Sell in May and go away" is a time-worn market adage, referring to the period from May through October that has been the weakest for U.S. stocks going back at least 60 years. May, June and August typically have been lackluster, with September especially treacherous, according to the Stock Trader's Almanac.
Meanwhile, the six months from November through April, with the exception of February, have marked the strongest period for the benchmark Standard & Poor's 500 Index (SPX) .
S&P 500 monthly average performance*(Jan. 1970 - Feb. 2010)
*Not including dividends
Data: Standard & Poor's Inc.
In fact, for the past 20 years or so, dreaded October also has been a generally winning month for the markets, suggesting that traders may be trying to front-run the traditional year-end buildup.
"There's a sprint to the finish," said Richard Ross, global technical strategist at Auerbach Grayson, a New York-based brokerage. Traders start to focus on bonuses, holidays, vacations, he said, adding that "There are a lot of tailwinds behind the market."
The pattern continues into January and through the spring, with the first month's performance tending to be a barometer for the rest of the year.
"The cycles of greed and fear happen to coincide with the seasons," Ross said. "This has been ingrained in the markets from the very beginning."
The Halloween effect's notoriety should have eliminated it as an opportunity long ago, or in the words of Yogi Berra: "Nobody goes there anymore; it's too crowded."
But it has persisted in the U.S., and many countries. "Even though it does not work every year, I think it is hard to find periods of, say, a decade when it would not have worked," said Ben Jacobsen, a finance professor at Massey University in New Zealand who has published seminal research on the Halloween indicator.
"Generally," he said, the November through April trading pattern "works often enough to make the believers happy and a bit richer on average and the skeptics happy as well but a bit poorer on average."
2. Turn of the month indicator
In this strategy, you buy on the last trading day of the month and sell after the first three or four days of the next month.
Why has this approach succeeded? Again, you're following the money. Money managers are "window-dressing" portfolios at the end of the month to improve returns, while pension funds and automatic retirement plans are also contributing to buying demand.
"Selling into the early month's strength is a good strategy," Ross said. "If you get a pickup in the first couple of days, money is definitely coming off the sidelines out of retirement programs, and money managers have a clean slate and you get a nice lift up."
The best illustration of this indicator's power comes from "The Seasonality Timing System," backed by research from veteran market strategist Norm Fosback, editor of Fosback's Fund Forecaster newsletter.
The system calls for being 100% invested on the last trading day of the month and selling after four trading sessions of the next month, and also being fully invested for the two trading days preceding a market holiday.
"The Seasonality Timing System has been superb on a risk-adjusted basis," Hulbert wrote in a recent MarketWatch article. In an interview, he added: "It's the best market timing system of any." See Mark Hulbert's column on this timing system.
According to Hulbert's research, a portfolio that switched between the Wilshire 5000 Index and 90-day T-Bills on the seasonality system's signals gained 4.1% annualized on average from the end of December 1999 through the end of February 2010, versus a 0.3% decline for a buy-and-hold investor. Importantly, you took only one-third of the market's risk.
That makes sense in a declining market, when missing the worst days would have been to your benefit. What about a bull run? From Dec. 31, 1989 through Dec. 31, 1999, the seasonality system gained 13.4% on average each year, compared to 17.6% for buying and holding.
While that's a smaller total return, the strategy carried only about 40% of the market's risk. On a risk-adjusted basis, that puts the timing system ahead of buying and holding, Hulbert said.
Fosback created the system in the mid-1970s, based on data going back to 1926.
"It's been 35 years in real time," Fosback said. "Over this 35-year period it has continued to demonstrate above-average returns. You wouldn't have beaten the market, but you would have earned a return pretty close to the market's average."
And you'd have captured this performance without suffering through the market's unpredictable swings. "You're exposed to the risk of market fluctuations just 30% of the time; 70% of the time you're absolutely risk-free," Fosback said.
If the turn of the month effect is due to month-end paycheck, pension contributions and other sources, what accounts for the bullish sessions leading into market holidays?
Fosback attributes this to the unwillingness of short-sellers to leave positions exposed to market-changing events over holidays. "My hypothesis is there was short covering the day before the holiday in particular, and traders put back their positions after," he said.
Fosback added that the seasonality system seems more valuable nowadays for smaller stocks, and suggested that investors starting out might consider a small-cap ETF that tracks the Russell 2000 Index (RUT) , for instance.
3. First day of the month trade
A subset of the turn of the month effect is the "first day of the month" trade, where you're in the market for one full day each month and in cash for the remainder of the month.
This strategy, not surprisingly, dominates in bearish periods For example, from the end of 1999 through March 1, an investor who followed the trade using the S&P 500 as a proxy would be up 28% on a cumulative basis, and a $10,000 investment would be worth about $12,800, according to S&P. A buy-and-hold investor, on the other hand, would have lost 25% cumulatively and $10,000 would be worth only about $7,600.
"In a bear market, it does better," said Howard Silverblatt, senior index analyst at Standard & Poor's.
Not so in bull markets. From the first trading day of 1990 until Dec. 1, 1999, the first day trade in the S&P 500 would have netted you about 3.5% annualized, excluding dividends, and a $10,000 investment would have been worth $10,406.
Buy and hold, meanwhile, would have delivered a yearly gain of 15.3% and that $10,000 would have grown to more than $41,000.
So be careful when utilizing this or any other trade and pay attention to broader market trends and technical analysis. "There's no indicator or strategy in isolation that generates a buy or sell signal," said Auerbach Grayson's Ross. "You need a cluster of evidence supporting or disproving your case."
Still, the odds are in the first-day trade's favor. From 1926, about 57% of these trades were up, versus 52% of all the days, S&P reports. Said Silverblatt: "More cash coming in pushes the market up, even in a declining market."
Powered by CBS MarketWatch