SEE HOW THIS IS PUT TO USE ---> | Coupling Market Timing with Seasonality |
                                                  | How to use the Seasonality Actuarial Chart & Table (SATC) |

There is an old Wall Street adage that goes "Sell in May and go away". It works on the premise that stock market returns during the period June to October are historically on average far lower than from November to May. In addition, following the rule allows you to miss the stock-market crash prone September/October months. The investing strategy around this adage involves selling your portfolio at the end of May and keeping your funds in a money market or fixed deposit account until November where you re-invest in the stock market again.

Though this effect is often cited informally, it has largely been ignored in academic circles. Nonetheless analysis by Bouman and Jacobsen (2002) shows that the effect has indeed occurred in 36 out of 37 countries examined. However research by Standard and Poors' Equity Research that examined the S&P 500 index growth in the May-Oct periods after the last 14 major stock market corrections going back to 1932 show that the adage does not apply toward the end of big bear markets. The S&P 500 only declined in 2 of the 14 periods assessed, and 9 of the 14 periods showed strong growths as the markets clawed out of their troughs.

Unfortunately the JSE has been left out of any published research regarding this phenomenon, and thus we set about our own research to see if the effect applies to our local markets, and if a viable mechanical investment strategy can be derived from any seasonality we see inherent in the market.

The graphs below reflects the average monthly growth on the JSE ALSH for the last 20-year and 10 year periods respectively.

We notice that the JSE has similar but not identical seasonality than international markets. Looking at the 20 year historical data we can see that our "return-to-equities" date is one month earlier (October) than the November mooted in international markets. May definitely seems to be the right month to be "going away" from the markets as evidenced from both the 20 year and 10 year data, however since we are "toward the end" of a major correction, we probably do not want to act on it this year. We note that the JSE seasonality has changed a bit in the last 10 years as opposed to the long term trend. In the last 10 years, August has shown up as an incredibly strong month, as opposed to a "flat" month in the 20 year seasonality chart, and February has been a negative month recently as opposed to a "marginally up" month in the 20 year trend.

If one looked purely at the strategy of investing in the JSE ALSH index for just a single month of the year, for the last 10 years, (and sticking your money under your mattress for the balance of the time) then the total % growth achieved from such a strategy is shown below on the chart on the left. However, when including a deemed buy and sell brokerage cost of 1% the chart on the right shows that except for August and December, brokerage consumes most of the likely gains to be had from such a strategy.

The charts above are just looking for the returns power of the various individual months and thus only look at the accumulated returns from the months your money is vested in the JSE and exclude interest received during the other 11 months your money is not being used in the market. The chart below shows your total returns if you earned interest at the Prime rate less 4% on your funds whilst you were not invested in the JSE and assumes 1% buy and sell brokerage costs.

We observe that only the December and August trends are powerful enough to overcome the eroding effects of brokerage and beat a "money in the bank" strategy, but they could still not beat a ALSH buy-and-hold strategy. All the months to the right of "BANK" did not manage to beat the returns from a simple bank account.

We now attempt to find a multi-month seasonal strategy than can beat a simple JSE ALSH "buy-and hold" strategy. Obviously by adding more "contiguous" months to the strategy we can eliminate the erosion caused by brokerage. The best performing combinations are shown below:

"ALSH" is the buy-and-hold strategy. "BANK" is a strategy that stays completely out the share market and sits in a FNB call account for the entire year earning prime less 4%. The black numbers within the horizontal bars represent the percentage time the strategy was invested in the JSE.

We note that the top strategy invests in the JSE for all 8 months of the year (67% of the time) that showed a positive average growth over the last 10 years (March-to-May, August and October-to-Jan.) It requires 3 trades per year (6 transactions in total) and delivered 337% growth versus the ALSH's 324%. Although this strategy did not comprehensively out-perform the ALSH index it managed to literally match its performance by only being exposed to stock market risk 67% of the time. Another interesting strategy is the 6 months one of May, August and October through to end January. It marginally underperformed the ALSH index, but was only exposed to stock market risk 50% of the time!

For the risk-averse investor, you could keep your funds in a money-market account for 10 months of the year, and invest in the stock market in August and December, and you would improve your returns of a pure money-market strategy from 10.24% compound per annum to 13.13% compound per annum.

All these strategies require 2-3 trades per year for a total of 20-30 trades over the 10 year period. The periods the top strategy was invested in the JSE versus the respective growths since October 2007 is shown below. We see that the strategy can be rather volatile, up one trade and down the next and this is a characteristic of seasonal-only strategies that make them unsuitable for the risk averse investor. Note in the chart how the seasonal strategy was saved partially from the great 2008 crash by being out the market during June and July 2008 (when the JSE fell 13%) and September 2008 (when the JSE fell 11.3%).

If you deployed a high-yielding cash instrument for the non-vested periods (such as a money market account or anything else that beats the FNB call rate) you will manage to out-perform the ALSH marginally. 

You can deploy a low cost and convenient  JSE Top-40 ETF such as SATRIX40 as proxy for the ALSH. We tested the daily JSE index % differential between the ALSH and the TOPI (top-40 trade-able) over the last 5 years below and you can see that SATRIX would make an excellent proxy. (On an unrelated topic, notice how the TOPI pulled away from the ALSH in the final-gasp bull charge from Feb08 to JUL08, just before the great crash. More research on this effect as crash predictor later!)

An additional benefit of going with a SATRIX40 ETF  is that you can halve your brokerage costs by purchasing/selling the ETF shares from the fund manager directly, which is about half the costs of doing the trade through your on-line broker. As we have shown, brokerage costs of 2% per trade can eat away at your compounded returns. Since you know the dates you want to be invested in advance you have plenty of time to set-up the trade and specify the date to SATRIX when you want it to be effected.

Be warned though, that the strategy can be volatile, with you missing bull runs whilst sitting on the sidelines, getting into the JSE during a slide, and being up one trade and down the next. However, with time, the strategy is designed to outperform as time allows the favourable probabilities inherent in the strategy to come to fruition.

We can confirm the adage seen on international markets with some small exceptions such as the more recent AUG anomaly and the fact that historically OCT seems a good month for the JSE.

Whilst May is the month to "go away" research has shown this is not the case in bear markets and so you are not advised to heed this advise this year.

After-tax returns of the strategy are only likely to beat the ALSH for investment periods less than three years. After 3 years the ALSH buy-and-hold delivers superior after-tax returns due to lower capital gains taxes been applied as opposed to standard income tax for investments held for less than three years. So this strategy only makes sense if you already are invested in money market funds or cash deposits but wish to seek a higher return with minimal risk.

As always, our standard disclaimer applies : "Historical returns are no guarantee for future performance." As is always the case with the markets, in the short term they can surprise but if you stick to your strategy in the long term it becomes more likely that history will repeat itself.

UPDATE 27 JUL 2009 : We have incorporated the power of JSE Seasonality into our famous SUPERModel Composite Market Timing system. This lifted 31 year returns from 113,000% to 208,000%, by only adding 20 extra trades on top of the original 10 over the 31 year period. Combining seasonality with a market timing system significantly improves the volatility characterisitic of seasonal-only strategies. Go read how we did this by clicking on the below link.

SEE HOW THIS IS PUT TO USE ---> | Coupling Market Timing with Seasonality |
How to use the Seasonality Actuarial Chart & Table (SATC) |
Make a Free Website with Yola.