When do I buy back in?

At the moment I am only about 30% invested in the stock market.

The question that is exercising my mind is "how will I know when to buy back in?

I am looking at using a cross of the 30 day Simple Moving Average (SMA) by the 10 day SMA from below (subject to the price being higher than the cross). The market has recently fallen up to 14% and is currently still down 10%. Unless we have a dramatic fall on Monday in response to the 3.4% fall on Wall Street S&P500 last Friday, the cross could happen this coming week.

There are risks to this approach. The risk is that the signals reverse several times without the market moving much between the crosses. That is called being whipsawn. It can cause losses and each transaction also has costs. The market is described as "trending sideways""or "in transition".

However, you need something to help you judge when to buy back in, or increase your weighting of stocks. Some would say the 200 SMA is more reliable, but you have much bigger losses in a major downturn or lost opportunities in a major upturn using such a long average.

So why would I use the 10 and 30 cross? Because it can be quite a profitable indicator. (If the market was down 25% I would likely use a much longer SMA than 30, maybe 100 - I feel that the deeper the fall, the longer the SMA for a buy signal should be to avoid whipsawing, but I have no backtesting to support this feel):

It gave a 10% rise from October 2001 to February 2002 with no whipsaws, and a 17% rise from March 2003 to March 2004 with 5 whipsaws. Using a 50 day SMA instead of a 30 day SMA would have given a 16% rise with no whipsaws and would have been a better choice on this example as the rise was gradual with 5 pullbacks. A sudden rise with no pullbacks would have been well suited to a shorter SMA like 30 Trading Days.

The chart above also shows how you can get whipsawn as the market goes through a 20% decline. If you used a 120 SMA instead of 30 you would have been whipsawn once on the way down and not at all on the way up. you would have been down 3% before your first sell signal, have lost 2% in the whipsaw and then picked up 14% for a net 9% gain compared to a zero gain over the same period if you had just continued to hold your stocks all through the period.

Here is a chart with two significant rises where there was no whipsawing using the 10 and 30 SMA cross. Note that we are down a similar amount now as was the case from both of the bottoms shown in this chart.



But what of the risks of multiple whipsaws with losses and transaction costs? Right click and open the next chart in a new window.


From Jan 91 to July 92 you would have had 7 roundtrip transactions for a total market movement of only +22%. a number of the round trips would have been loss making, reducing the 22% gain to one where on a net basis you captured only between 1/2 and 3/4 of the 22%. A 200 day SMA instead of a 30 would have given less whip sawing, but a higher first buy price, so there may not have been much benefit in using a longer SMA.

So not buying would have led to missing a profit of between 10 and 15 % in 18 months compared to being whipsawn by the 10 and 30 sma while a perfect buy at the bottom would have given 22% return - but how would you pcik the absolute bottom?

No one rings a bell at the top or bottom of the market! You have to have some basis on which to buy and sell, or adjust your portfolio weightings.

Choose the lengths of the SMA's you will use based on your own costs of transaction, loss aversion, ability to replenish funds and consider other factors as well, including fundamentals.

Before embarking on any strategy involving moving averages you should have a look at some recent work by Doug Short::

http://dshort.com/articles/2010/Nikkei-monthly-moving-averages.html

http://dshort.com/articles/2010/SP500-monthly-moving-average-history.html