Friday, August 24, 2007

Have we a bottom?

The Kirk Report has a good link fest of negative outlook on the market. One of his links is to Doug Kass at, who suggests a cumulative decline of 25% should not be seen as surprising, or impossible (this would put the S&P at 1,166).

My measure of intraday buying as measured by my Trade Ideas scan would seem to support a theory for further weakness. In my historical look at the rate of intraday buying from what I term the "Top-8" stocks, rallies from major reaction lows are supported by broad intraday buying - especially into the close of business - giving a "Top-8" list of stocks which is selected well inside 10 minutes, and often in a minute or less. Although the sample is restricted to only one major reaction low (July of last year) there is reason to suggest the very weak buying for the rally from last Thursday's lows is not what bottoms are made of and new lows are necessary to 'reset' the count. Indeed, the strength of buying we have seen this week from my scan has much in common with the buying seen during the March-May phase of 2006 prior to that summer's correction: are we looking at another 10% trim in the markets (calculated from whatever reaction high this rally makes)?

However, market internals like the Nasdaq Summation Index ($NASI), the NYSE Summation Index ($NYSI), the S&P 500 Bullish percents ($BPSPX), and the S&P Percent of Stocks Above 50 Day Moving average ($SPXA50R) all point to a significant bottom in place - so which measure is right?

I decided to take a closer look at the market internals and look at potential 'failures' - or at least - early signals for a bottom. I will only focus on one indicator, the Nasdaq Summation Index, but there are likely similar stories to be found in the other aforementioned market internals.

The Nasdaq Summation Index has always been a favorite of mine at it switches to extremes relatively smoothly. Measuring tops using this indicator is more difficult as it often produces lengthy negative divergences relative to the parent index - although a downward break of the zero level is usually a good trading signal to 'sell'. Because bottoms are quicker to form that tops the indicator has an advantage. For August of this year the indicator dropped to a low of -941, similar to that of August 2004 when it reached -950.

The 2004 bottom is a good example of one such early bottom signal. During May of that year the Nasdaq Summation Index made its first oversold bottom at -814 - itself a good 'buy' signal - but this was not the bottom for the parent index, and after a nice 5% gain from the -841 bottom the index rolled over for another 10% loss to mark its actual low at the eventual -950. Was there any warning for the first 'failed' bottom? The MACD histogram gives the best clue as the May low was well shy of a confirmation 'buy' as measured by the Summation index, but instead shaped a solid bullish divergence to the indicator - strengthening the second reaction low made by the Summation Index as a buying opportunity for Tech and the Nasdaq

How does this compare to today? The MACD suggests we are still someway from a bottom, although it is nicely primed for a bullish divergence on the next reaction low the NASDAQ Summation Index makes - the next low will provide another opportunity to buy or add (it should be noted, both of the 2004 lows proved to be good buying opportunities - just stocks were 'cheaper' on the second low):

Another possibility to watch for is for a positive divergence in the Summation Index relative to the Nasdaq (not necessarily in its supporting technicals - but this is hard to confirm based on available data) which occurred in 1998. 1998 is interesting as it marked the debut of the Nasdaq Summation Index at -647. The Summation Index did mark a bottom at -927, before it rolled into its lowest historical low on record at -1,648 (so, yes - things could get a lot worse from current 2007 levels). The Nasdaq lost over 500 points (about 25%) from the reaction high made in July of that year. Will this situation repeat? Hard to tell as there are no supporting indicators like a MACD to tell its own story. If there is something to take from this it would be buyers of the August 2007 low may have a month long rally to make some money before things turn really sour (the June-July 1998 rally added some 15% to its lows - not a bad effort from the markets).

I suppose time will reveal all!