As I am writing this, we are in the midst of a market correction that began in March. Earlier in the year, after I had completed my backtesting, I posted my results in another forum, and these results predicted that a correction in the upcoming weeks was quite likely, which in hindsight did turn out to be the case. The original post can be found at this link:
techstocks.com
This resulted in a couple of responses, which can be found at these links:
techstocks.com
techstocks.com
techstocks.com
techstocks.com
I would like to move the discussion here so that the various threads can remain on topic.
Here is an excerpt from the original post (1583) mentioned above:
<< Using the AAII archive data, and comparing to the S&P 500, I have found that the percentage bulls has exceed 60% only 3 times since 1987. Once was just weeks before the 87 crash. Another was at the end of May 1996, just weeks before the big small-stock correction. And a third was at Christmas 1992, when the market didn't correct but just stayed flat for a couple of months.
The 63% bullish sentiment reported a week ago is only the fourth time since polling began that it exceeded 60%. Based on the three previous incidences, it looks like we have a 67% chance of a correction.
Since the sample size of the surveys is a little small (typically 100 respondents) I've also tested a two-week moving average. Only on the three occasions mentioned above did the two-week average number of bulls reach 55%. Two additional times it reached 53%, and they were not followed by a correction during the next few weeks (the market kept going up). So at 53%, the historical record over a few weeks following hitting the 53% level: 2 corrections, 1 flat market, and 2 up markets. So this second criterion makes it look as if there is a 40% chance of having a correction in the upcoming weeks, a 20% chance of nothing happening, and a 40% chance of the bull market continuing.
So for the two criteria: one says there's a 67% chance of correction in upcoming weeks, the other a 40% chance.>>
The American Association of Individual Investors survey that I used is posted every Friday morning by the AAII on their America Online site, which can be found by going to keyword AAII, then reference shelf, then surveys, then investor sentiment survey. They have text and Excel versions there that are updated each week, along with the S&P 500 data. While AAII does have a web site at aaii.org that duplicates most of the AOL site, it unfortunatly does not have the survey results. If anyone would like to to look at the data themselves, possibly to try their own backtesting using a spreadsheet, I can post it here on this thread in text format. |
