sentimenTrader.com

All rights reserved.

Sunday, November 3, 2002

For the past two weeks, I've emphasized that when all of our models and a majority of indicators are either neutral or conflicting with each other, it typically precedes a period of volatility, which is what we have been seeing.  Not a whole lot has changed this week.

Indicator Summary:

BULLISH

SEASONALITY - We have positive seasonality working for us once again this week.  November has been historically positive (using the S&P 500 since 1950), with the month ending higher than it began 65% of the time, and an average return of 1.65%.  This makes it one of the most positive months of the year.  13 out of the past 20 Novembers have ended positively (with an average return of 1.32%), 7 out of the past 10 have been positive (average return of 2.10%), and 4 out of the past 5 have ended higher (average return of 2.36%).  The most positive times of November are the beginning of the month and the area around Thanksgiving, which I'll talk more about when we get there, as it's a phenomenon of which you must be aware.  Part of the reason the beginning of November is positive is due to the beginning-of-month positive bias which has existed over the past 50 years.  As you can see from the Seasonality section of the site, the first three trading days of a new month are some of the most positive days of the month, although the effect is muted somewhat when the 50-day moving average is sloping down (as it is now).  Seasonality can be a very powerful force, and override other fundamental and technical factors, so please be aware that the beginning of this week and the days surrounding Thanksgiving will not be completely "natural". 

BEARISH

RYDEX RATIOS - This week, the intermediate-term picture being displayed by the Rydex fund flows is the most bearish in over a year and a half.  The bullish fund flows (as defined by the 10-day moving average / 50 day moving average deviation) have set a new record high while the bearish fund flows have set a new record low.  Remember, the ratios as I compute them are weighted by the amount of leverage the funds employ, so even if the total amount of money flowing into the bullish funds is not terribly great, if it is flowing into the leveraged funds (called the Dynamic funds) then that will make a substantial difference in the ratios.  The charts below show the fund flows for the S&P 500 funds, both leveraged and un-leveraged.  The thick black lines are trendlines, so you can basically ignore the actual flows (gray lines).

BULLISH FUNDS

BEARISH FUNDS

We can see from these charts that there has been explosive growth in the leveraged funds, to the detriment of the funds that employ no (or less) leverage.  To ignore the dynamics the Dynamic funds have on this fund complex is not recommended.  In any event, the action we've seen in these funds over the intermediate-term, as I've said, is the most troubling it has been since these funds came into existence.  Not only are the 10/50 deviations incredibly bearish, but the three-month stochastic remains very high, the 3-day RSI is on the high end of neutral and the money market assets are fairly low (as a percentage of total assets).  All of this means that there has been a consistent shift in sentiment from very bearish near the October low to extremely bullish now.  Although there is some room in the short-term measures to reach higher levels, the longer-term indicators suggest that the enthusiasm recorded from this trader set has reached very troubling levels.

TRIN - The 10-day TRIN on both the NYSE and Nasdaq remain overbought, particularly on the Nasdaq which has reached its lowest level since last November.  In fact, our recent situation looks very similar to that time period, when the S&P floundered in a 54-point range for over two months.  Although there were spikes up and spikes down, we never really went anywhere until the market broke down going into late January.  We've now been stuck in an approximately 40-point range for about a quarter of that time.  As for the Nasdaq, the chart below shows what happened each of the times the 10-day TRIN has reached these low levels since the bear market began.

We can see that each instance lead to an extended period of consolidation (an average of about two months) or outright decline except for the period immediately following the terrorist attacks last September.  So if 5 out of 6 of past occurrences lead to consolidation or decline (and the one exception I think is safe to call an "outlier"), then it is a stretch to assume that this time will be any different.

VIX and VXN - The continued erosion in the volatility measures is not a complete surprise, as they have been elevated for an extended period of time, and it is only natural that they revert to their long-term means.  In the process, however, they are reaching an extreme - not necessarily in price, but in time.  You'll recall that in that past I've discussed times when the VIX is "elevated" or "depressed", which I consider to be 5% away from the 10-day moving average.  I then look at extremes in time, not price.  I think the last time I talked about this was in mid-June, when the VIX had recorded 12 straight days of being at an elevated level.  That resulted in a very quick 50-point (low to high) jump in the S&P.  We have the flip-side occurring now, as the VIX has been depressed for 6 straight sessions, and 13 out of the past 15 sessions.  Let's take a look at how each occurrence has played out in the past:

As usual, the percentages directly above or below each bar represent the probability of the S&P 500 being higher the given number of days later.  The top chart shows how the market performed 1,3, 5 and 10 days later when the VIX had been depressed for six straight days.  The lookback period is during this bear market only (3/00 - current).  The bottom chart shows how the S&P performed after the VIX had been depressed for 13 out of the past 15 days.  Since there was only one other instance of this occurring during the bear market (late November 2001 which lead to a 2.5% drop in the S&P within a week), the entire life of the VIX was used as the lookback period.  We can see clearly that both sets of circumstances lead to a market that consistently underperformed over the subsequent two weeks, with a low probability of the market being positive.

NEUTRAL

COMPOSITE MODEL - The churning action in the market this week didn't allow for a large change in the model, which dropped 6% from its level a week ago.  As I said last week, it will be very difficult for the model to actually pierce the lower standard deviation band (which would be incredibly bearish), but an approach to that band should be noted.  We're still quite a ways from that area, although the model is quite close to where it troughed in late August, corresponding to the market high.  This model isn't suggesting caution yet, but it should be watched closely.

SENTIMENT SURVEYS - I talked about the Investor's Intelligence survey earlier this week, so I'm not going to go over that again.  All of the others showed a slight dip in bullishness.  The AAII poll came in with a reading of 52% bulls, which is the highest since March, but the bears also increased so the bullish ratio (bulls / (bulls + bears)) actually dropped a couple of percent.  One troubling aspect that I'm finding is that although the broader market has so far not approached the August highs, and in fact it would take another 8% rally for the NYSE Composite to match its August peak, the one-year stochastics of the surveys are showing more bullishness now than they did then.

We can see from this chart that what we're seeing now is very similar to January - March of 2001.  During that time, the market rallied a bit before taking a 15% dip, accompanied by a large drop in bullishness from the surveys.  We then had a rally that recovered about 80% of that decline, but with a rise in bullishness that surpassed the previous peak.  That rally quickly failed and took us down to the low in September.  Of course, we're further along in the bear market now, and the S&P has lost over 30% of its value from that peak in May 2001, so the dynamics are different.  But when we see a higher high in optimism but a lower high in prices, it usually does not end well.  Conversely, when we see positive divergences, with prices making a higher low but pessimism making a lower low (suggesting survey respondents are looking for the lows to break), it has typically lead to a nice rebound off of those lows.  This is an example from early 2000:

These divergences between price and sentiment are fairly rare, and can be quite useful when they happen.  I suggest you check out these stochastics each week on the site.

AIM - Since this model is made up of the stochastics from the surveys mentioned above, it should be no surprise that the model has made a lower low while price made a lower high.  The divergence is not particularly strong, and I wouldn't read a whole lot into it, but as I mentioned above it is something to note.

COMMITMENTS OF TRADERS - We saw basically no change in the S&P 500 information this week, as the total net change in the commercial and small spec net positions was the smallest in 12 years.  The commercials decreased their net short position by 60 contracts while the small specs increased their net longs by 193 contracts, both of which are meaningless.  This is not a surprise considering the market was essentially flat during the reporting period.  The change in the Nasdaq futures was larger, with the commercials adding a couple of thousand contracts to their net shorts while the small specs decreased their net longs by a couple of hundred contracts.  Not much to read into this week's numbers.

STEM - Talk about flat.  This model has been stuck in a 3-point range for the past 15 days, which is unprecedented.  This model has not seen such a period of little movement for such an extended period of time in the past four years, but again, considering the market action during these 15 days it is not entirely surprising.  The only thing this tells me is that we're seeing a matched battle between those looking for higher prices and those looking for lower, and once it breaks in either direction we should see some give-up by the other side.  This consolidation has allowed the standard deviation bands to come down, so it is becoming more difficult for the model to become overbought and easier to it to become oversold, which is a positive going forward.

STEM.MR - Like the STEM model above, the past two weeks have trapped the model in a range, although not quite as severely as that model.  Each time this model was about to hit an extreme, the market would reverse (usually with a gap open in the opposite direction) and alleviate whatever conditions were building up.  As of Friday's close, we're once again approaching the lower (bearish) band, and a push above the recent highs will likely give us the "oomph" needed to record an extreme here.  However, that breakout will likely bring in new buying and I'm not convinced it would be wise to fade such a move.

PUT/CALL RATIOS - The rather unusual situation of the equity put/call ratio being in agreement with the OEX put/call ratio continues to play out.  By this, I mean that the equity put/call ratio is rather low, and is close to hitting its lower standard deviation band (bearish).  At the same time, the OEX put/call ratio is also dropping and is approaching its lower band.  Over the past year, this has happened only one other time:

The areas highlighted in yellow are the instances where the 10-day equity p/c ratio (red line) touched or came close to its lower band.  The blue line is the 10-day OEX put/call ratio.  As you can see, most of the time when the equity ratio came close to touching the lower band (suggesting that small-capitalized equity options traders were betting heavily on further upside), the OEX p/c ratio was peaking at a relatively high level (suggesting the more savvy index option traders were putting money on downside protection).  The only time we had a "divergence" like we have now, where both the equity and index options traders were placing money relatively heavily on further upside was in late November of 2001.  Remember from above that this was also the time when the VIX was 5% below its 10-day average for 13 out of the past 15 days, and lead to a decline of 2.5% over the next week.  Due to the small sample size of divergences such as we're seeing now between these ratios, I cannot state confidently that this augurs anything in particular.  In the S&P 500 options, after being bullish for most of the week, the put/call bid/ask bias ratio closed Friday at the most bearish level since October 21st.  Friday's reading of 0.20 is the lowest since 0.18 on the 21st, which lead to a sharp decline the next day.  The reason for the low ratio is that in the back months, 51% of puts went off at or above ask while only 2% went off at or below bid, and 53% of the calls went off at or below bid while only 11% were traded at or above ask.  What this means is that in the expiration months after November, there was a clear bias towards buying puts and selling calls by institutional investors.  Although there is no way for me to tell what strategy was being employed by these traders, and in fact it could actually be part of a bullish overall position, when we see such a lopsided series of trades it has typically lead to tradable market weakness within 1-3 days.

BREADTH RATIOS - Over the next couple of weeks, the 10-day advance/decline ratio will be dropping a hodge-podge of readings, which will make it difficult for the indicator to go much of anywhere.  Again, that should not be surprising considering the market action during the past 10 days.  The same thing will be occurring with the 10-day up volume ratio, so I would expect these two indicators to continue to chop around near the overbought levels they are at now unless we get a sustained trend in either direction this week.  The daily cumulative TICK readings remain where they were near the August peak, which is just under where I would consider them to be overbought.  The intraday NYSE TICK is now fully in overbought territory.  As you can see from the site, when this indicator became overbought over the past couple of weeks, there was weakness (or at least no strength) the following day.  I would be a hesitant to become too aggressive based on this indicator alone, however, since I think it's safe to say we are in a 30-minute uptrend.  As I continually say, these indicators are not as effective when giving a signal in the direction of a trend (i.e. overbought in an uptrend and oversold in a downtrend).  In any event, it urges caution from the long side in the short-term.  Overall, I would rate this complex of indicators as neutral, but with a bearishly overbought tint.

NYSE MEMBERS REPORT - During the first week of the rally off the October low, the public decreased their shorting activity while the specialists on the NYSE picked theirs up a bit.  This caused the Specialist Short Ratio to rise a couple of percent to 40%, which is firmly in neutral territory after being bullish the past couple of weeks.  This is still below the August peak of 43% and well below the 5-year upper standard deviation band of 50%, so I don't see any cause for concern here.  Not until the public curtails their shorting activity greatly, driving this ratio closer to 45%, will I consider this information to be troublesome.

Like last week, all of our models remain neutral.  As I said then, I do not believe in being aggressive when there is no clear edge, so I continue to suggest keeping trades light this week, or shortening your time frame and taking advantage of smaller price moves.  That isn't to say there could not be a 100-point move in the S&P this week, but it's a matter of weighing the odds and placing yourself in a position that has a positive risk/reward profile.  If you cannot define the risk or the reward, then you will be best served by staying cautious.  We still have mix of indicators, and several are suggesting that this chop will continue.  There will be spikes higher and spikes lower, but net/net we could be relatively unchanged several weeks from now.  During times like this, with no clear trend, it is beneficial for short-term traders to concentrate on oscillator-type indicators.  You can see proof of this by the tremendous job the S&P and Nasdaq price oscillators have done over the past couple of weeks (the charts are posted to the site daily and are included in the intraday notes I send out).  Of course, at one point the market will break out of the range and these oscillators will give a terribly false signal, but your risk control would minimize any losses.  Because the range we have been in is so well-defined, it's likely that a break in either direction will become self-fulfilling, and cause at least a short-term trend in that direction.  The NDX broke out near the close Friday, so it will be interesting to see how that is treated come Monday.  If we fall back into the range, then I expect the lower end to be tested once again.  But if we hold above this level, then I expect some heavy buying to carry us higher.

 - Jason Goepfert