May 05, 2003

http://www.sentimentrader.com

Wednesday, May 7, 2003 8:06 PM EST

I've talked several times over the past couple of weeks about the 21-day "breadth combination". This is simply a combination of the 21-day moving average of the up issues ratio (a.k.a. advance/decline line) and the 21-day average of the up volume ratio. When I was discussing it a couple of weeks ago, it had reached 113% or so, which certainly was overbought according to recent history. Currently, the ratio is hovering around 118%, which is the highest reading since November 1998. Anything much over 110% had been a precursor to a substantial correction during this bear market.

I went back to the 1960's to see how the market reacted after such high readings as we're seeing now. The two charts below show the entire time span from 1966-present, with the S&P 500 in the top pane and the 21-day breadth combination in the lower pane. The vertical blue lines on the chart show you instances where the 21-day breadth combo reached 118% or higher.

One other note...you may notice a different look to these charts from what I have been using up until now. I have been searching for quite some time now for a charting program with the features I needed, and I believe I've found one that has many of them. If I don't receive too many protestations to the contrary, I will be converting ALL charts on the site to this new format. I believe this will make the data easier to observe, as well as allow me to quickly build a much-needed chart archive, so that you can see historical readings of all the charts on the site. I think you'll agree that will be a very welcome addition.

A couple of things stand out:

1. During the past 10 years, we have only seen two other instances of the breadth combo reaching 118%. The first was after the crash in 1997 and the second was after the crash in 1998. As with most oscillator-type indicators, this "overbought" condition didn't really matter since we were in a defined uptrend already, and were in the immediate aftermath of a severe selloff. The market continued higher unabated after the overbought readings.

2. The reaction in the market in the late 1980's demonstrates why it is extremely difficult to trade overbought readings in an uptrend (or oversold readings in a downtrend). More often than not, these types of extremes are notices that pressure is there, and will continue more often than not. You would have gone broke trying to short the market with these overbought readings during an uptrend.

3. It's impossible to find a comparable period to the current in this data, as we have not seen another three-year bear market like we have now. If we look at what could perhaps be the closest - the early 1970's - then these overbought readings did, in fact, often lead to at least a short-term correction or trading range at best. Typically, the only time when it did not lead to a correction was in the immediate recovery from a severe selloff.

It depends on how you define "trend", but currently the weekly trend is either neutral or down. It would be a severe stretch to suggest we're in a weekly uptrend. Also, the rather anemic decline into March and the subsequent 16% rally make it seem more like we are seeing just another bear market rally than it does that we are in the midst of a recovery from an intermediate-term low. Therefore, overbought readings DO matter, and should be looked at as opportunities to either lock in gains on long positions, or establish shorts in anticipation of a correction.

On Monday, I said that the action to end last week would likely move a significant number of sentiment survey respondents to the bullish camp. That appears to be the case, as the latest Investor's Intelligence survey showed a large jump in bullishness, finally moving this survey into extreme territory. The chart below outlines those instances since 1991 where the II bull ratio (bulls / (bulls + bears)) reached its current level of nearly 70%. Out of the 9 instances, only 1 (in 1999) lead to a steadily rising market.

If we play "what-if" and plug conservative increases in bullishness in the other surveys, then our AIM model has a very good chance at reaching 40% or just above this weekend. If so, I would likely become even more defensive than I am now (if possible).

Most of our shorter-term measures are back to a neutral condition. This has been all that's needed to allow the market to continue its upward advance over the past month. If we don't recover quickly from here, then this will be the first chink in the symmetry (in a sense) of this latest rally, and will provide the first evidence that we could be in for a deeper correction than we've seen. Longer-term, I remain of the opinion that this is a bear market (or trading range market at best) until "Step 2" is accomplished, which would be a weekly close above the 955 level at least. And in a bear market or trading range, selling an overbought market is a higher-odds strategy than jumping on the bandwagon with most of the others.

Disclosure: no positions

This disclosure is not intended as trading advice in any form. It is meant as a note to subscribers that the author may have a position directly affected by the market outlook reflected in the commentary. Although the author takes great pains to remain objective in any commentaries, it is only fair that readers should know that the author may have taken positions in accordance with his market outlook. Positions can and do change at any time, without notice to the reader.

Monday, May 5, 2003 9:54 PM EST

Last week, I outlined what it would take to move the odds in favor of the "new bull market" camp. There were four steps, and the market completed #1 last Friday. This does NOT mean the bear market is over. It means that we have gone one step in a long journey, and the market has a lot of work to do to prove itself. Until it completes all four steps, I believe rallies should be sold.

The latest Commitments of Traders data continued the trend it has been on for two months now. In the large contract, commercial traders remain net long while small speculators continue to increase their short activity. Those small specs are now actually net SHORT for the first time since October 1999 and to the largest degree since September 1998. This data has been spot-on so far during the latest rally, so it's hard (and not advisable) to ignore its bullish implications.

Another trend that is continuing is the divergence in the e-mini contract. In that contract, commercials keep adding to their short positions while the small specs add to their longs. I had been mentioning that while this contract had only been effective for about six or seven months now, its most effective time frame was in the short-term. If that is the case, then it has broken down (thus demonstrating the difficulty of placing too much weight on a "new" indicator). The positions held in this contract are expanding exponentially, so there is really no telling at this point what exactly could constitute an "extreme". I have received a lot of questions asking how these two contracts could be so diametrically opposed. Anecdotally, explanations I've heard include everything from arbitrage strategies between the contracts, to hedging, to an outright switch from the full contract to the e-mini due to its electronic platform. I don't know which is correct, possibly all of them, but until its trading pattern becomes more clear, I think the action in the e-mini should continue to be discounted.

The sentiment surveys regained some bullish momentum, which has been enough to drive our AIM model into negative (for the market) territory for the first time since November 2002. There is certainly more room for the model to become more extreme, as it is currently around 46% and its effective "minimum" is closer to 40%. I think there's a decent chance we could approach that level with the next reading, as last Friday's breakout in the S&P should have moved a significant number of poll respondents to the bullish camp. If we do approach that level in the model and the market continues to rise, it would be the first model failure sell signal since the bull runs in 1997 and 1998.

I've been asked if I am worried about the latest Barron's Big Money Poll. In case you're not familiar, Barron's magazine runs a semi-yearly poll of around 150 money managers with a large amount of money under management, typically at least several hundred million dollars. The latest poll was released this weekend, which means that the responses were probably sent in beginning about a month ago. The chart below shows the comparison of the big money's predictions at where the S&P would be six months from now (on average) to the actual outcome (much of this had to be estimated due to the methodology of the survey).

We can see that during 1999, the big money actually UNDER-estimated the power of the bull market, and their guesses were consistently below where the S&P ended up closing. Once the bear began, however, the big money could not lower their average guess low enough, as they did nothing but OVER-estimate the market potential, often by a very wide margin of 10% or more. Contrary to the protestations of Barron's editorial staff, this big money does seem to be something of a contrary indicator, at least recently. While their individual stock picks and pans have a much better record, their broad market calls leave something to be desired. So the fact that the percentage of respondents who are bearish is the lowest in over five years should not necessarily be encouraging to anyone who actually bothers to look at their record. And that only 0.8% of them are very bearish is downright scary - since there were 147 responses, this means that only 1 person was very bearish. This isn't a group that I would normally look to "fade", or take the other side of, but they have been resolutely bullish throughout this bear market, and continue to be to the greatest extent yet.

Many of our shortest-term measures are now back to a neutral stance. This has happened while a confluence of our longer-term measures have finally begun to reach something of an overbought condition, as evidenced by the intermediate-term indicator score chart. This reinforces what I said last week - short-term rallies are more likely going to turn into intermediate-term shorting opportunities rather than the continuation of a strong bull-market move. The break of the weekly downtrend line, and the persistent bullish implications from the full S&P 500 futures contract, has taken some of the wind out of the sails of potential shorts, but we would need to see a new high (i.e. at least 955) before serious damage to the short case would be done.

- Jason Goepfert

Disclosure: no positions