Tuesday, October 13, 2009

Timing The Market

Many participants in the stock market base their buy and sell decisions on attempts to time the market. The idea is to buy into the market just before prices rise, and sell before they decline. Many studies have shown that it is very difficult to correctly time the market. But assuming you had superior foresight, how often would you have to be right in order to beat a buy and hold investor?

Chua, Woodward and To took an interesting approach to this question. Using a mean market return of 12.95% and a standard deviation of 18.30%, the authors ran 10,000 simulations of market years where an investor has an assigned probability of correctly determining a bull or bear market. If said investor guessed a bull market, he was credited with the market's return. If the investor guessed a bear market, he was given the T-Bill rate of return. The buy and hold investor always received the market return.

The results were surprising. In order for a market timer to beat the buy and hold investor on average, he had to correctly predict a bull or bear market a full 80% of the time! The authors concluded that the cost of an investor missing out on bull markets was quite high. Timers lost much ground in the years that the market did well where the timers incorrectly guessed a bear market.

However, the standard deviation between the timer and the holder was quite wide, suggesting that many market timers will indeed outperform the holders, and this is the case even at a 50% probability of a correct prediction. As such, timers who are successful will be promoted and we are sure to continue to hear about timers who have beaten the markets. Whether the outperformance as a result of timing is sustainable or simply the result of random distribution is another story. On average, being right 80% of the time would appear to be quite a stretch.

2 comments:

Retirement Savior said...

I would suggest that for market timing, most investors don't make intermediate term forecasts, but work off of the recent past.

Considering the fact that recency bias looms large among investors, I wonder if market timing would be better shown by a simple "if year 1 is positive, then long the market for year 2, if year 1 is negative, then cash for year 2."

It would show even worse results than the study you mentioned.

Jonathan Goldberg said...

Very interesting research indeed. In my view it just confirms my belief in value investing and in NOT trying to time the market. Sure there will be people who outperform for certain periods as market timers, but is it sustainable as you asked? I wouldn't think so. One or two missed calls and they can see all their hard efforts wiped out. Great post Saj!