Short-term Outlook:  25% Bearish  From Feb 25, 1091 SPX

Equity Indicators - Updates and Extremes

AAII Sentiment Survey

Earlier this month, we took a look at the surge in Neutral responses among the newsletter writers polled by Investor's Intelligence.  Technically, there isn't a Neutral category in that survey, but those who are generally bullish yet expect a short-term correction are considered so.

There wasn't too much we could conclude from that development, as it was so historically rare.  Apparently it isn't just newsletter writers who are confused; the latest poll from the American Association Of Individual Investors (AAII) shows a surge in Neutral responses as well.

Once again, there wasn't anything predictive about a multi-year high in Neutral responses.  Sometimes it led to a rally, sometimes a decline, and sometimes a flat market.

Looking at various combinations, though, one thing stood out:  when the S&P 500 was positive in the three months leading up to the high Neutral reading, then it tended to decline in the three months afterward; when the S&P had declined heading into it, then it tended to rally.

There were four weeks (not including the current one) when the Neutral responses hit a two-year high and the S&P was positive over the prior three months.  Over the next three months, the S&P showed the following returns:

07/29/94:  +1.4%

10/17/97:  -1.7%

06/05/98:  -7.8%

12/31/09:  -0.5% (so far)

And here are the instances when the S&P had declined over the prior three months, along with their returns over the next three months:

08/16/96:  +9.9%

09/06/96:  +15.5%

02/11/00:  +3.3%

01/31/03:  +5.0%

04/08/05:  +1.1%

01/30/09:  +4.9%

This study suffers from a small sample size and a little bit of data mining, so any conclusions are shaky, but it's at least interesting that there was such a stark difference between when the market had risen versus declined heading into the high Neutral readings.  If we do draw any conclusions from this, then it would suggest a modest downside bias going forward over the intermediate-term in our current case.

Equity-Only Put/Call Ratio

On a much (much!) shorter time frame, yesterday's reversal was remarkable.

While there are a number of different ways to quantify such a fakeout, and perhaps some of them will give conflicting signals (that's the problem with purely quantitative studies).  But to me the most intriguing aspect of yesterday's action was that the S&P 500 SPDR (SPY) gapped down more than -1%, closed above its open by more than +1%, yet still closed in negative territory (below the prior day's close).

In that fund's history, there have been 20 other similar days.  And over the next two days, it managed to add to the late-day gains 17 times (an 85% success rate) with an average return of +2.5%.  This was strictly a short-term phenomenon - looking out any longer than a few days, and the probability of a rising market moves back to random.

What was also notable about yesterday was the extreme registered in the Equity-Only Put/Call Ratio.  On the chart we update daily on the site, we use bands around the indicator to monitor extremes.  Yesterday, the put/call ratio closed more than 30% beyond its average over the past six months.

So we got a major price reversal and an indication of excessive pessimism.  What happens when we put them together on the same day?

As you might suspect, the results were rare but quite positive.  Over the next two days, the S&P 500 was positive 6 out of 7 times, and sported an average return of +4.0%.  That outsized return is due to the exceptional volatility we tended to see accompany such extreme readings.

Here are the dates along with the S&P's return over the next couple of days:

10/08/98:  +3.5%

10/18/00:  +4.2%

09/17/01:  -2.3%

09/21/01:  +4.6%

06/14/02:  +3.5%

03/17/08:  +1.6%

10/10/08:  +12.8%

Like the pure price reversal study above, this was best used as a very short-term indication only, as the longer-term the results were mixed.  Still, it suggests a continuation of yesterday's reversal heading into the new month.