One of the best short-term market timing systems of the 1980s and 1990s was given up for dead some years ago, but it’s started beating the market again.
The Seasonality Timing System, created in the 1970s by Norman Fosback, then head of the Institute for Econometric Research, is simple and mechanical. It says to invest in stocks around the turn of the month and immediately prior to stock-exchange holidays — and shelter in a money-market fund at all other times.
Fosback claimed it was the best stock market timing system ever created, and my Hulbert Ratings performance auditing firm found that in fact it had the best risk-adjusted performance of any of dozens of monitored market-timing strategies.
Its resurrection teaches a powerful lesson: Don’t be too quick to give up on a once-successful strategy because of market-lagging returns. It takes many years of underperformance to conclude that a strategy is more than just down on its luck.
The chart above plots the returns of the Seasonality Timing System since the early 1980s, as calculated by my performance-auditing firm. The returns were calculated on the assumption that, when bullish, a stock portfolio was fully invested in the Wilshire 5000’s total return index, and otherwise in 90-day Treasury bills
TMUBMUSD03M
.
Notice that the system was outperforming a buy-and-hold approach until about 15 years ago, but then began to falter. Beginning five or so years ago, the system once again returned to its winning ways, and since then has risen at a faster pace than buy-and-hold. The table below shows the specifics over these various sub-periods. Note that Fosback’s system calls for being in the market only about a third of the time, so even equaling the overall market’s raw return would be a significant achievement.
| Fosback’s Seasonality Timing System’s annualized return | Buy-and-hold’s annualized return | |
| 12/1981 to 12/2007 | 13.6% | 12.6% |
| 12/2007 to 12/2017 | 1.5% | 8.6% |
| 12/2017 to 3/2023 | 14.1% | 10.0% |
In retrospect, a follower of Fosback’s Seasonality Timing System in 2017 should have stayed the course, resisting the temptation to throw in the towel after a decade of lagging the market.
Was there any way to know at that time?
Yes. Consider a column I wrote in May 2017 about whether this timing model’s market-lagging decade was a reason to give up on it. For that column I interviewed David Aronson, a statistician who has authored several books on how to base investment decisions on a sound statistical foundation, including “Evidence-Based Technical Analysis and Statistically Sound Machine Learning for Algorithmic Trading of Financial Instruments,” (co-authored with Timothy Masters).
Aronson’s conclusion was that it’s “too early to conclude that the seasonality strategy has stopped working… I know it’s tempting for many investors to throw in the towel but the fact is that most investment strategies have widely varying performances around their average performance. It’s tough to stay the course but I think that’s the right thing to do with the Fosback strategy.”
How to stay the course
There are several rules of thumb you can use when analyzing a system to determine if its recent underperformance is severe enough to justify getting rid of it. The first is to determine if there is a plausible theoretical rationale for why the system should work in the first place. If there isn’t such a rationale, or if the world has changed so that the rationale no longer applies, then you should at least consider throwing in the towel. Otherwise you should give it the benefit of the doubt.
Fosback’s Seasonality Timing System does rest on a solid theoretical foundation. The turns of the month coincide with a typical worker’s payday, when 401(k) withdrawals get invested in the market. In addition, the trading sessions immediately preceding exchange holidays have an upward bias because, researchers have found, short sellers don’t want to be short during an extended period in which the markets are closed.
There’s no evidence that either of these factors has become any less important than when Fosback devised the strategy in the 1970s. (An email seeking comment from Fosback was not immediately answered.)
So long as the strategy you’re investigating can jump over these theoretical hurdles, then your second step is to subject its track record to statistical scrutiny: Can you, at the 95% confidence level, conclude that its recent performance is statistically different than its prior record?
While there are many sophisticated statistical tests you can use for this…
Read More: The best stock-timing system of the 1980s and 1990s is still beating the market