Short-term Outlook:  Neutral  From Jan 27, 1090 SPX

Equity Indicators - Updates and Extremes

Double 1% Up Day Failure

This is when it gets scary for people.  All recent support levels are broken, we just got a failed bounce, and the S&P had its worst loss in six months.  I'm already seeing more than a few references to the possibility of Black Friday.

The thing is, history is chock-full of these scenarios - the S&P having its worst loss in six months to a new multi-month low, when just coming off a 52-week high.  The fact of the matter is that about half the time, the fears were justified.  Not "Black Friday" kind of fears, but new intermediate-term decline kind of fears (see August 1990, March 1966, August 1957 as examples).

But half the time the fears were completely out of line, as the market bottomed that day or very soon thereafter (see February 2007, October 1997, November 1963 as examples).

Let's revisit the pattern we looked at in a Research Report on Tuesday and the Morning Report on Wednesday.  That was where the S&P hit a multi-month low, then enjoyed consecutive 1% up days.  The reports suggested that it was probably not the bottom, and we would most likely hit a new multi-month low before a meaningful rally, particularly if we saw short-term downside follow-through.

Well, we got the downside and we've already hit a new low.  So now let's look at every one of those patterns that failed almost immediately.  What we're going to see below is every instance where we got the double 1% up days, then rolled over a sunk to a new multi-month low within two weeks.

The results are very interesting.

 

S&P 500 Performance After Failed Bottom Coming Off Of Two-Month Low and Consecutive 1% Up Days
				After New Low...
Date Of Double 1% Days	Date Of New Low	Loss	Days 'Til New Low	1 Week Later	1 Month Later	3 Months Later	Days 'Til Important Low
05/31/62	06/13/62	-6.9%	9	-1.3%	4.2%	6.0%	9
02/12/73	02/23/73	-2.5%	8	-0.8%	-1.4%	-5.3%	2
09/17/74	09/27/74	-3.6%	8	-4.0%	14.4%	3.4%	4
10/01/90	10/10/90	-4.6%	7	-0.5%	2.4%	4.7%	1
06/05/98	06/15/98	-3.3%	6	2.4%	9.1%	-4.4%	0
08/18/98	08/27/98	-5.3%	7	-5.8%	0.6%	13.8%	2
03/26/01	04/03/01	-4.0%	6	5.6%	12.8%	11.6%	1
01/31/02	02/04/02	-3.2%	2	1.6%	6.2%	-3.8%	3
09/26/02	09/30/02	-4.6%	2	-3.7%	8.2%	7.9%	7
09/19/08	09/29/08	-11.8%	6	-4.5%	-15.0%	-21.4%	20
	Median	-4.3%	7	-1.1%	5.2%	4.1%	3

It took a median of 7 days, and a loss of -4.3%, before the S&P hit a new multi-month low.  A couple of times, it happened as quick as our current instance and did it in only 2 days.

But look at the other columns.  While short-term performance was mostly bad, in every case but one (a big failure during the 2008 distress) we were within days of an important low that lasted for several weeks at least.

While the logical assumption would be that such a quick failure of the 1% up days is an incredibly ominous sign, it was almost without exception the exact opposite.  By the time we hit a new low, the selling pressure was already nearly completed.

I've looked at a multitude of studies overnight, viewing the current selling in various lights.  The conclusions were almost universal - we should be within 1-3 days of a tradeable bounce, but the "tail risk" of a painful short-term washout is higher than average.  While a huge gap down and major selling pressure during the day would likely lead to at least a short-term bottom by Monday, we'll have to wait and see how the market reacts to the Payroll report and possibly even weekend news about sovereign credit risk.