The Market Timer

PORTFOLIO POINT: Many investors waste a lot of energy wondering when to sell but obsessing over things you can’t control

Over the past month, to some the sky has been about to fall. Global sharemarkets have pulled back amid concerns that the debt crisis in Ireland and Greece is worsening, China is likely to act to slow its economy and QE2 isn’t going to work – all compounded more recently by tensions on the Korean Peninsula.

After rallying by 11% between late August and early November, the ASX 200 has since slipped by as much as 4.5%. On top of that retreat, the ASX 200 has dipped below its 80-day moving average (see the top panel in the following chart) and momentum has reversed (as shown in the bottom two panels below):

And to cap things off, last week the 4600 support level cited throughout October and the early part of November was breached.

So why hasn’t our Active Strategy gone to a 'sell’?

The key to a successful market timing strategy is a good set of exit conditions. These exit rules must strictly control losses, but at the same time not sacrifice too many potentially profitable trades. Frequent and hasty exits produce small losses which sacrifice potentially profitable trades while also cutting trades short before they reaching their full profit potential.

A reversal in trend is obvious in hindsight. But market timers have to make decisions in real time. A market timer always stays in the market past the peak price and so misses some of the investment gains that are possible. The trick is to spot as early as possible when a trend has ended.

It is not uncommon for retracements of the size seen recently (of 4% or so) to be insufficient to trigger 'sell’ signals. While our Active Strategy model is driven by the ASX All Ordinaries index rather than the ASX 200, and the maximum pullback since November 5 in the All Ordinaries (at 4.1%) is not quite as much as for the ASX 200 (4.5%), the percentage fall to date from the early November peak is around the average for the percentage falls triggering past 'sell’ signals.

It is true that exiting a wrong signal (and a losing position) based upon a predetermined (and fixed) percentage price fall makes a lot of sense. In fact, a trailing stop exit places a limit on the loss (or drawdown) of equity as prices reverse.

However, exiting a winning position on such a basis makes no sense. Currently, our Active Strategy is in a winning position, with the last 'buy’ signal still generating a 2% gain at the most recent lows.

The issue now is: when do we get out of a winning position?

Exiting a winning position is a particular challenge for market timers because we have to be comfortable with letting a trend run as far as it can, crest, and then begin to decline before considering an exit with profits. If we don’t, we run the risk of missing out on the big wins necessary to at least offset our frequent small whipsaw losses.

At MarketTiming we use a signal exit. The exit occurs when our model gives a signal that is contrary to the currently held position. The model, which is only approximated by the indicators tracked in the above chart, includes proprietorial adaptive (or variable) moving averages. These effectively adjust the length of the moving average used according to the recent volatility and pattern of the underlying price data. Over time, the model produces a better investment return than one based on conventional moving averages.

Our model suggests that the ASX 200 would need to fall below 4550 for a major retracement to be on the cards and a 'sell’ signal issued. On Thursday, December 2 the ASX 200 closed at 4676.

Rather than using a signal exit as we do, some rely instead on a critical threshold exit which involves exiting when the market approaches or crosses a theoretical barrier (e.g. a trend line, a Fibonacci retracement, or a support or resistance level).

While this is not our methodology, we do make reference to support and resistance levels as a double-check, particularly when the market seems to be going sideways. In particular, a 'sell’ would be warranted if share prices have fallen below an established support level.

The fact that our Active Strategy hasn’t gone to a 'sell’ can be vindicated on this basis. While it is true that on November 23 the ASX 200 retreated through a support level set by some at 4600, careful consideration reveals that this breakout may have been more apparent than real. The 4600 support level wasn’t clearly established, with 4580 being the lower extreme of the support zone (see the dashed horizontal line in the top panel of the above chart).

Any breaches of this lower extreme were not significant in the sense that they were not deep or sustained. The conclusions one can draw from these breaches are usually false.

This assessment is reinforced too by the fact that the NYSE’s S&P 500 has stayed above support levels over the past month. Since early November the S&P 500 has twice respected support around 1180. It does not seem likely that the ASX would go into a major correction in advance of Wall Street.

In conclusion, when exiting becomes a real possibility (as it clearly has over the past week or so), it is best just to stick to a predetermined plan (and hold our nerve). Otherwise we run the risk of exiting too early. All trends must be allowed to play out without making unnecessary trades because of volatile short-term conditions.

And there can be no more timely evidence of the risk of exiting too early than Thursday’s 2% surge in the ASX 200. A premature exit could have been costly.

The following chart shows the performance of our market timing portfolio since the Active Strategy’s signals were first published by MarketTiming on December 15, 2009.

While the Active Strategy’s most recent 'buy’ signal remains a winning one, the benchmark buy and hold (B&H) strategy has also experienced the same proportionate rise.

As a result, our Active Strategy portfolio remains behind B&H for now after a whipsaw in late-August put our strategy over 2% behind B&H after being at least 2% ahead for the three months prior to that.

Alan Tregilgas is a director of Market Timing Pty Ltd. For a free trial click here.