sentimenTrader.com

All rights reserved.

Sunday, September 1, 2002

From last week's commentary:

"There are many bearish signs entering the new week, but they are almost exclusively short-term and could be corrected quickly.  The intermediate-term and longer picture continue to look fine from a sentiment perspective, and another rally attempt after a short-term decline is very likely.  For now, I believe that focusing on short setups will present the better risk/reward relationship for short-term traders."

The weakness we saw this week alleviated many of the short-term negatives I mentioned last week, so we will enter the new month with a better chance to see some short-term upside.  We will also likely see a pickup in volume - since 1965, the week after Labor Day (compared to the week before) has seen higher average daily volume 78% of the time, with an average increase of around 10%.

Indicator Summary:

BULLISH

STEM - The rather sharp decline this week has brought this model back to oversold levels.  Although we're not at levels seen in late July, we're close.  On Thursday, the 20-period moving average hit 94%, which is almost always followed by some short-term upside relief.  Although the 20-period average has now begun to back off a bit, the 50-period average (the signal line) has jumped from the 40's last week to the 80's this week, finally closing at 82%.  Readings above 80% have historically been positive for the market, for at least a short-term pop.

COT - Let me first clarify time frames here.  I am considering this information bullish in the short- to intermediate-term only.  Long-term, I still believe the relative levels of long-side commitment both groups are showing is a negative indication for market potential.  With that said, we did see a rather positive surprise in the latest reporting period (the data is available on the site).  After a rise such as we've had off of the July lows, we would normally see the commercials increasing their short positions and the small specs increasing their longs.  Not this time, as you can see from the chart series below.

Commercials are not increasing their shorting...

While the small specs are liquidating longs and increasing shorts...

Results in a positive futures environment...

It's obvious from the charts above that this time, so far, is different from recent large rallies off of intermediate-term low points.  We will have to watch closely over the next few weeks to see if the situation changes, as it can do so very quickly.

VIX and VXN - Last week I showed a chart of a model that I keep which tracks VIX reversals.  I said that the model was suggesting that over the short-term, the VIX should head higher and the market lower, and indeed the VIX tacked on 7 points by Thursday while the market sold off.  The picture this week has changed considerably.  Although the model isn't oversold, it's close, and further weakness this week will likely trigger a positive signal.  There's no need to consult a model to find out what the VIX is doing, as the charts of the VIX and VXN on the site will give you all of the same information I'm looking at.  The RSI is overbought and the indicators are poking out of their 10% bands, while the VIX and VXN themselves have made new short-term highs and begun to reverse.  These are all classic signs of overbought volatility and oversold equities, and hint at a short-term market rebound ahead.

BEARISH

BREADTH RATIOS - Last week, I showed the chart of the one-year stochastic reading of the 10-day advance/decline line (which simply compares the current reading to all of those over the past year), and said that if we reached the 90% level, that it would be wise to at least tighten stops on any intermediate-term long positions, as it is extremely rare to see significant market advances after such readings.  We actually hit 90% on Monday, and that proved to be the high of the week, and perhaps for several weeks.  It will take some time for the breadth indicators to work off their overbought condition, and a sustained upside surge is unlikely.  Three of the next four days will see the 10-day a/d average drop very large positive readings, which will go a long way towards moving this indicator back to an oversold condition, provided we don't replace them with more high readings.  The cumulative TICK indicators are all neutral, with the exception of the intraday NYSE indicator.  That is still overbought, as we actually saw quite a few high TICK readings on Friday before the late-day decline.  If we see a drop early next week, this should quickly cycle down to oversold.

NEUTRAL

SENTIMENT SURVEYS - The biggest story this week in the sentiment surveys was the large increase in bullishness in the Investor's Intelligence poll.  This was one of the largest increases in the bullish ratio (bulls / (bulls + bears) in the past 15 years, at 11%.  Such an increase has been matched or exceeded only twice, in September of 1989 and February of 1991.  In 1989, it lead to a steep decline within two weeks, and the market was still lower 8 weeks from the increase.  In 1991, however, it lead to an even sharper RALLY in equities, as the market never looked back over the ensuing two months on its way to a 10% gain from the time the increase was recorded.  Looking back to lesser increases in the bullish ratio (anything above 8%), the picture is mixed.

We can see that large increases in bullishness in this particular survey usually lead to short-term outperformance in equities, but over longer time spans the optimism seemed to weigh on the market a little.  This does not seem to be a significantly bearish development, especially since the increase came from a bullish ratio that was quite low to begin with.  In fact, the current bullish ratio doesn't even make it into the top 40% of readings over the past 15 years - hardly excessive bullishness.  The only other survey to show much of a change was the AAII survey, whose constituents also seem to feel more optimistic about the immediate future.  The bullish ratio there jumped from 52% to 63%, but again that didn't even make the top 30% of bullish ratios over the past 15 years.  I mentioned several weeks ago about a quick method I use to determine approximate support and resistance levels when there is a large change in the sentiment surveys.  I use the average price for the week (determined by the formula:  (LOW + HIGH + CLOSE)/3) to approximate where buyers or sellers may have come into the market.  Since there was a large jump in bullishness, and we can make the assumption that the opinions expressed in the surveys may correspond to actual positions taken in the market, we will want to watch the average price during the survey week as a very general level of resistance.  This is because those bulls who bought during that week and haven't already sold are now most likely down on their investments, and may sell if prices rebound back to breakeven.  The average price during the survey week was 944 on the S&P and 1019 on the NDX.  Again, these are very general levels that I will be watching, but certainly wouldn't use them as specific trading levels.

AIM - The streak of 12 straight readings over 80 has ended, in violent fashion.  This model shed 29 points, moving from 80.2% last week to 51.2% this week.  Although not bearish by any measure, the model is now at the lowest point since April, due to the large increases in bullishness from the II and AAII surveys.  Although it may seem alarming that the model dropped nearly 30 points in just one week, this drop doesn't even make the top 10 over the past decade.  We've actually seen larger drops - and from lower levels - before, so this isn't unprecedented by any means.  What it means is that we now have another model that has switched from being extremely positive during the historic bearishness we saw in late July to neutral after the large move from the lows.  That doesn't tell us anything specifically, other than that the intermediate-term (and longer) sentiment situation isn't as positive as it was when the S&P was 100 points lower, and those who were astute enough to buy into the panic of others should now be looking at protecting those profits.

SEASONALITY - Hopefully, you've had some time to explore the new Seasonality and Daily Market Bias sections of the site.  For those of you that haven't, these sections currently present biases that have shown themselves to be present over the past 50+ years in the S&P 500.  More will be added, and plans are to add the data for the Nasdaq and Dow Jones Average as well.  We can see from the site that entering September, the average return for the month is -0.52%, one of only three months to show a negative average return.  September also shows the worst probability of being positive, at 40%.  This is 11% lower than the next-lowest month.  If you look at the Daily Market Bias page, you can see that the first three days of September are historically the best days of the month, showing both a higher-than-average return and probability of being higher on the day.  However, that bias is skewed by the holiday effect of Labor Day.  Since 1950, the S&P has been up the day before Labor Day 77% of the time, with an average return of 0.43% - quite significant.  The day before Labor Day very often falls on September 1st, 2nd or 3rd.  The day AFTER Labor Day doesn't show much of a bias at all, with an average return of 0.01% and a probability of being positive of 56%.  When the market is weak the day before Labor Day and it doesn't show it's usual pep (i.e. it's down on the day), the day following the holiday is up only 40% of the time.  It's very easy to get lost in all of these statistics, and they are secondary to market action itself, of course.  The purpose is to give us a better feel for inherent market biases that exist for whatever reason.  What we can glean for this week is that usually the days before Labor Day are positive, and they are often negative after traders return from their holiday vacations - especially when the market is weak heading into the holiday.  This will cause me to have a bit more of a negative bias going into the week than I normally would.

PUT/CALL RATIOS -  Last week I mentioned that I was troubled by the relatively large amount of public call buying that was coupled with a large amount of index put buying.  That was widening the spread between the two p/c ratios, which was not a positive sign.  With the decline from the highs this week, we are beginning to see a shift in sentiment.  This week has seen a persistently high total p/c ratio, enough to push the 3-day moving average of the one-year rank back up to bullish territory.  This indicator had touched a low of 17% on the 19th, and now it is back up to 84%.  The 10-day and 21-day moving averages have also begun to cycle back up from their close-to-overbought levels last week, although they have a long ways to go before they could be considered bullish.  In the S&P 500 options, the action in the September expiration on Friday was quite bullish, with a bid/ask bias ratio of 5.76 (for the September expiration only).  32% of puts traded at or below bid compared to 16% at ask.  13% of calls traded at bid while 38% traded at ask.  As you'll recall, one typically sells at bid and buys at ask, so if the institutions (which account for the bulk of SPX trading activity) are selling puts and buying calls, then we can conclude that they have a bullish view in the short-term.  As should be expected, volume was very low.  There was no particular bias in the later expirations.

NYSE MEMBERS REPORT - Even though the NYSE Composite Index was up around 2% during the most recently released reporting period (ending 8/16), every group tracked in this report curtailed their shorting activities.  In fact, shorting activity dropped about 8% across the board, so the Specialist Short Ratio and Public/Specialist Short Ratio remained unchanged.  We're still hanging right around the 20-year median at this point, so I don't believe there's too much to read into this report at the moment.

STEM.MR  - Most components of this model are neutral as of Friday's close, with the exception of the VIX.  The somewhat elevated nature of that indicator is keeping this model on the higher end of neutral, although that could change within 1/2 hour on Tuesday.  The moving average which is used to calculate the deviation for the VIX has dropped dramatically over the past two weeks, so it will be much easier for this indicator (and the model in general) to achieve oversold status.

COMPOSITE MODEL - After dropping to the mid-30's on August 22nd, this intermediate-term model has clawed its way back up close to oversold levels, at 61%.  Obviously, it's nowhere close to the upper 80's we saw at the July low, but it certainly a positive change from the low readings only one week ago.  Next week, the increased bullishness from the sentiment surveys will begin to weigh heavily on the model, as those readings make up 15% of the model's weight.  Countering that somewhat should be the better situation we will likely be in with the 10-day moving average of the advance/decline line, which should continue to move lower from the extremes we saw last week.  With those two components in conflict with each other, I don't see a big movement coming out of this model either way, unless we see a major move up or down.  Any further selling, however, could push the Composite Model into positive (though probably not strong positive) territory.

TRIN - From the charts on the site, you can plainly see that both the NYSE and Nasdaq TRINs were reached critical mass, touching levels that normally coincide with market high points.  The decline last week went a long way to relieving that overbought pressure, as both 10-day averages are now back to neutral territory.  Three of the next four days will see the 10-day moving averages drop off extremely low readings, so any further market drop will likely move these averages back up to oversold.

Most of our indicators are now neutral, and on balance we certainly have a better situation than we did last week.  There are some troubling negatives - seasonality, the breadth ratios and the fact that we no longer have the support of extremely negative investor sentiment - but there are some positives as well.  Other than the failed signal in early July, the futures markets are the most positive since the April 2001 and September 2001 lows, even though we are more than 15% off the lows already.  We are beginning to see fear creep back into the market, even though the selloff from the highs hasn't been particularly severe.  The STEM and Composite models are pushing positive territory, and the VIX and put/call ratios are perking up again.  I expect volatility to increase over the coming week(s) as traders return from vacation and volume begins to pour back into equities.  I suspect we will see several triple-point days in the DOW, in both directions, as bullish and bearish perceptions battle for dominance.  I don't see any compelling reasons to be long on a short-term basis at this point, but if we sell off enough early in the week to push the STEM.MR model into positive territory, I believe the chances for a tradable rally to the upside will be great.

 - Jason Goepfert