10/29/2008
This formula suggests a possible market bottom occurred this week
The best investment book I’ve read this year is Ken Fisher’s The Only Three Questions That Count. It's fairly technical, but in a nutshell, the book lays out suggestions and examples for how an investor can know things that other don’t. In the final analysis, that is really the only way to gain a real edge in the markets.
One of the generalizations he proposes is the following: for most bear markets, the first 1/3 of the loss takes place in the first 2/3 of the time. For example, if a hypothetical bear market lasted 18 months and lost 30%, the charts would show that after the first 12 months (2/3 of the time), the total loss would be 10% (1/3 of the total loss). Of course, the final 1/3 of the time plays out the final 20% loss. The implication is that tops tend to be rounded, while bottoms are fast and furious.
Clearly this is not an exact formula, and the only real way to measure it is after the bear market completes itself, but it got me thinking. How would this structure match up against the bear market we’ve been in so far? To test this idea, I assumed that the end of the bear market happened on 10/27/2008, 2 days ago. That was the lowest point on the chart before the 11% up move yesterday.
Below you can see the chart.
Chart 1 Twothirds rule playing out in reallife
The top of the market was on 10/09/07. 263 days later, the market bottomed on 10/27/08. Twothirds of this time, 175 market days, comes to June 20, 2008. The closing prices for these 3 days are:
10/9/07: 156.48 6/20/08: 131.58 10/27/08: 83.95
The total loss from 156.48 to 83.95 is 46.35%. The loss in 2/3 of the time (until June) is 15.91%. Therefore, the portion of loss in the first 2/3 of the down move is 15.91/46.35 = 34.3%! This is about as close to 1/3 (33.3%) as you can get in real life. The suggestion here is that it would make perfect sense if we put in the market bottom on 10/27. I plan on staying long and adding to my market ETFs, such as SSO, QLD, UYG (equivalent to roughly 2 times the moves of SPY, QQQQ, and XLF).
Take care, Dan
