Jun 23, 2002

sentimenTrader.com

Sunday, June 23, 2002

I said on Thursday evening that I would not take any more short positions, as I believed the odds of a rally far outweighed those of a decline. That does NOT mean, of course, that the market cannot continue falling, just that the odds, as I calculate them, no longer give a favorable risk/reward expectancy on the short side. I actually took a long position early Friday morning and was stopped out for a small gain, and didn't see any more opportunities the remainder of the day. Like I said last week, the S&P could drop another 60 points, but it's going to do so without me aboard. The risks of a reversal are just too great to make it worthwhile. That continues to be my stance.

This is why...

BULLISH

NYSE MEMBERS REPORT - As the market has dropped, the public interest in shorting has become more robust, at least on the NYSE. The Specialist Short Ratio dropped again, this time 6% to a level of 32%. This means that the specialists on the NYSE accounted for only 32% of the total short sales on that exchange. Since 1966, this level or lower has only been seen 65 times (this is out of 1903 weekly readings). In the Research section, you can see the results of what kind of market performance we had after such a reading. After 16 weeks (4 months), the market was higher 80% of the time, with an average gain of 9.4% compared to an average loss of 3.7%. The maximum return was 29.0% while the minimum return was -8.7%. The short-term results were also generally positive, although not nearly as impressively so. I've said repeatedly that I don't like to use this report as a short-term timing devise, but rather to get a higher-level perspective. The data is two weeks old by the time we receive it, so it's hard to trade off of this information. It will be extremely interesting to see the next two weeks' reports to see what kind of shorting activities the public has done during this decline. The most recent report encompasses the week after the S&P broke the important 1050 level, which obviously gave more confidence to those trying to short. Next week's report will show the activity after the low established the prior week also fell, so we may have a good chance at seeing another drop in this indicator. The all-time low reading was 20% in September 2001.

COMPOSITE MODEL - We have almost the same story this week as we did last week, as the model never even made it out of bullish territory, finally closing the week at 74%. During the September 2001 decline, this model spent 15 straight days over the 64% threshold, which was of course the greatest stretch in the past two years - until now. As of Friday, we've now had 18 straight days over 64%. Since 95% of readings are contained under 64%, this stretch is extremely rare. This most recent decline, while possibly not showing the extreme panic of September, certainly has been more persistent. I keep hearing that sentiment is not negative enough for a rally. That is simply not true.

SENTIMENT SURVEYS - I mentioned on Wednesday that the II and Consensus surveys shed another batch of bullishness, and it was finally getting some casual sentiment watchers a bit excited. I also said that if the other surveys remained at their bearish levels, the AIM model would have a decent chance at hitting 90. We didn't get there, as the other surveys actually showed an increase in bullishness, the AAII survey drastically so. I have mentioned repeatedly that the AAII survey is the most noisy survey I follow, and thus should be viewed in the context of some kind of moving average. The 4-week moving average of the bullish ratio that's posted to the site is currently at 50%, even with the most recent data. This is a level that was last seen at the September 2001, April 2001 and October 1998 market bottoms. The German Neuer Market survey is now showing the lowest sentiment levels of the year, and the 2nd lowest in the history of the survey (granted, it's a short history). Notably, the institutional sentiment is markedly more positive than the small private investor. Both are suggestive of positive conditions going forward, as there's a high correlation between the Neuer market and our own Nasdaq 100.

AIM - I said last week that there was a good possibility we would see another drop in bullishness from the sentiment surveys, and thus another rise in this model, delaying a signal for another week. Well, we got a drop in two of the surveys, but an increase in the other two majors. The increase in bullishness of the AAII survey was not only enough to offset the decrease in the II survey, but it was even enough to turn the model down a bit. Therefore, I will be issuing a BUY signal on Monday, regardless of market direction. That means that all four models will now officially be on three star-rated BUY signals.

STEM - I mentioned on Thursday that the signal line of this model will be dropping a string of low readings, so there was a possibility that we could see a new all-time high. We didn't quite get there, although the line made a fairly large move upward. We still have a few more low readings dropping off on Monday, so it's likely the 50-period moving average will continue higher in the morning. Regardless of what happens with the signal line, this model is screaming oversold, equal to what happened in September. The model only goes back four years, but we are now seeing a level of negativity here seen only once before.

STEM.MR - Our shortest-term model has now spent 25 straight 1/2 hourly periods above 95%. This is the longest string of high readings in the past two years, longer even than September. 'Nuff said.

TRIN - With the high readings we received toward the end of last week, the 10-day moving average of the TRIN is now fast approaching 1.50 once again. We would need another reading above 2.0 to get there Monday. That doesn't mean anything in and of itself, but it is widely watched. The TRIN study posted to the Research section of the site suggested that at the time of a 1.65 peak reading in the 10-day moving average, the NYSE Composite was higher 10 days later 70% of the time (since 1966). Well, 10 days later we are now lower by about 3.0%, which is almost exactly the average loss. The average loss actually increases over 15, 20 and 30 days but that was because of one instance in 1987 that skews the numbers greatly. Without that occurrence, the average loss would be about 2% over those time periods. And remember, the market was only lower 3 times out of the 10 over the past 36 years after such a reading. The Nasdaq TRIN 10-day moving average is now the highest since April of 2000 (except for a brief period in early May of this year). Both extreme readings lead to tradable rallies.

CBOE RATIOS - Although there is probably no "maximum" reading when it comes to these ratios, we're now about as close as we're going to get. The 10-day moving average, at .98, is the second-highest reading on record (September's 1.03 is the highest). The 21-day moving average, at .91, is tied for first along with a .91 reading during the panic of September of 1998. Forget the nonsense that expiration is skewing these numbers - we have seen a great amount of downside betting over the past few weeks, and these traders are never right at the extremes. In the Research section, you can now see a study of the open interest put/call ratio since 1986 when it peaks at or below .70, where it is now. You can see from the study (and 5-year chart) that these peaks have often coincided with intermediate-term bottoms. Although these readings sometimes lead to short-term underperformance, for time periods greater than a few days it proves to be a pretty accurate buy signal. It's also fairly rare, since we've only encountered 19 unique occurrences in the past 16 years. This set of indicators is screamingly bullish.

BREADTH RATIOS - We're generally bullish here as well, although we're not at extremes by any measure. We've seen a steady erosion in the market, but only a couple of days of hugely lopsided selling. The new high/new low and volume ratios remain low, but not historically extreme. The advance/decline ratio is oversold, although again not historically extreme. A most unusual occurrence is the cumulative TICK indicators, which are actually closer to overbought than oversold on an intraday basis. As I mentioned in the daily commentary, it's not as though we're seeing persistently high ticks, but it just so happens that at the times I collect my data, the ticks have been rather high. Although I'm discounting these a bit because of that anomaly, it's not a positive sign overall - I would have preferred to see a gathering of high tick readings instead of the sporadic ones we saw late last week. The average tick range, which I posted in last week's commentary, is still the highest it has been in three years, with the 10-day average range now at 1700.

VIX and VXN - Last week, I said the spike we saw on Friday may be an exhaustion of the current trend, and it was - for a grand total of two days. We then turned around and headed higher, but not explosively so, which is somewhat disappointing. We're still more than 10% above the 10-day moving average, and the RSI is still above 70, so the indicator is technically overbought (the way I look at it). The action on Friday was troubling, with the VIX actually dropping while the market declined as much as it did. That may have been influenced by expiration, but trying to determine the exact impact of that event is next to impossible.

NEUTRAL

TOPM - Same story as last week - this model has been completely flat. After giving some very decent heads-up over the course of the past couple of months, this option pricing model has been relatively muted during this most recent decline. Normally we would see the model decline as put prices are being bid heavily, but that's just not happening. I've determined that it's not expiration-related, but I still don't have a good explanation for why it's not reflecting more fear.

COT - This one is iffy, as it's somewhat bullish shorter-term, but still bearish historically. The three-month stochastic is 100 for the commericals and 0 for the small specs, which is as bullish as it gets. However, the one-year stochastics are not quite where they "should" be, and on an absolute basis both are still very bearish. Last week saw the greatest one-week change in posture since September, with the commercials shedding nearly 18,000 of their net short contracts while the small specs decreased their net long posture by just over 12,000 contracts. The interesting thing is that this reporting period covered the time up to 6/18, when the S&P closed at 1037. The market was actually UP during the week of this big change (although volatility was extremely high), and it's down 4.6% since then. I will be very interested in seeing this week's numbers on Friday to see what kind of change this decline (so far) has brought.

I could repeat last week's closing paragraph verbatim and be comfortable with it. We are now seeing a tremendous confluence of negativity, equal to past major bottoms. There will always be indicators that are not in line (insider buying, margin levels, etc.), but I only follow those indicators that I have proven to have a high correlation to market direction when an extreme is reached. And those indicators are coming together as they rarely do, which tells me the odds of an intermediate-term rally are tremendous. We never know if this will be the time that sentiment doesn't work, but it never pays to think like that. What good would it do to go to a casino and place your bets on blackjack with the assumption the dealer is going to deal herself 21 every time?

There seems to be a lot of talk about a possible crash, or a one to two-day decline that generates enough fear to tell the herd that a bottom is in place. I can certainly see that happening, and would welcome such a buying opportunity. I know there are many of you that are nursing long positions during this ordeal, and I certainly do not wish you any more pain. But if that's what it's going to take to make some people commit funds, then I say bring it on and get it over with. I'm not going with the assumption that such an event is necessary, since we are already at a point that equals past bottoms. But it should be clear by now that my opinion is any cascading moves lower will be very buyable for an intermediate-term rally.

- Jason Goepfert