sentimenTrader.com

All rights reserved.

Sunday, October 27, 2002

I left off last week stating that due to the mix of models and indicators, it may be best for longer-term traders to stay cautious and not become too aggressive in either direction.  The violent whipsaws we saw this week are common when we have such varying perceptions in our indicators.  This week, we have a similar situation, although not as pronounced.

Indicator Summary:

BULLISH

SEASONALITY - Last week, I mentioned that the following week contained the two most historically negative days in October.  Although the 22nd followed through, the 25th did not.  That's why they're odds and not certainties.  This week has the seasonally strong end-of-month pattern working for it, as well as the October fiscal year end for many funds.  The tables below show the expectancies for the last week in October over the past 50 years, as well as the individual trading days leading up to the end of the month:

This is not overwhelmingly positive, but it is worth noting.  When you compare the results above to the end of month seasonality table on the site, you can see that October does have a more positive month-end bias than the average month, particularly in the days leading up to month-end.

BEARISH

RYDEX RATIOS - This week's readings are almost the exact flip side of two weeks ago.  At that time, the intermediate-term indicators derived from this information was still quite positive, while the shorter-term indicators were becoming bearish.  This week, the intermediate-term outlook is decidedly bearish while the shorter-term picture is a little more healthy.  The chart below is the composite chart of the bullish flows superimposed on the bearish flows.  The dashed lines are the current levels of the bullish flow (upper line) and bearish flow (lower line).

It quickly becomes obvious from looking at the chart that these traders have shown sustained optimism at a pace unlike any other over the past year and a half.  Considering that the rally off the 10/8 low is only 12 days old, it seems awfully early for these traders to become so optimistic (or considerable less pessimistic).  This shift in fund flows has also kept our three-month stochastic at an elevated level (currently .74, one of the highest readings over the past year and a half).  One other development that we haven't seen in several months is a shift out of the money market, which is currently at the lowest level (as a percent of assets) since May.  One feature of these traders is that they seem to shift into the money market when they are unsure of market direction, and this has a close correlation to market lows.  Alternatively, when they are more sure of where the market is going, they take money out of the money market and commit it to funds dependent on a market rise.  Although the flow of assets out of the money market is not large enough to be considered bearish, the shift we are beginning to see is troublesome.  The one positive sign, in the short-term, is that the three-day RSI of the bull/bear spread, which was so overbought last week, has come back down to a much more healthy level.  Not oversold by any means, but certainly better than last week. 

TRIN - I showed the nature of overbought TRIN readings in the context of a bear market last week, and that comparison still holds.  While both the NYSE and Nasdaq indicators remain significantly overbought, they have come off the extreme readings of last week.  I continue to believe that the readings registered last week are an important indication that market breadth has reached an unsustainably positive condition, and some rest would be the best thing this market could do.  If we continue higher unabated, it will not bode well going forward.

VIX and VXN - Both the VIX and VXN made one of their largest one-day moves in weeks on Friday, dropping close to 10% each.  The VIX has now lost over 15% of its "value" over the past 10 days.  While by no means an extreme drop historically, it has had some bearish connotations during this bear market.  Historically, when the VIX drops 15% within two weeks, the average OEX performance 10 days later is as follows:

We can see from the table that although such a drop has not lead to anything particularly memorable over the life of the VIX, during this bear market it has indicated "too much, too fast" and was remedied with lower equities prices a great majority of the time.

NEUTRAL

COMPOSITE MODEL - I said last week that there was a battle between time frames in this model, and the positive longer-term indicators would balance the bearish shorter-term indicators which would likely keep the model muted.  That turned out to be the case, as the model didn't move more than 5% all week.  Due to the large increase in bullishness in the sentiment surveys and big move in the commitments of traders data, the model will lose two supports for this coming week.  In fact, if we substitute the current weekly data for Friday's readings, the model would have dropped 8%.  With the wild swings in the market over the past year, the standard deviation bands around this model - which are used to determine extremes - have widened substantially.  This will make it very difficult for the model to actually pierce the lower band and give a negative reading.  However, should we approach that band over the coming days (or weeks), I would consider that significant and bearish.  If we have upside continuation next week, look for this model to approach that lower band.

SENTIMENT SURVEYS - To help with the interpretation of the weekly readings, I have added the bullish ratio (bulls / (bulls + bears)) for both the Investor's Intelligence and AAII surveys.  Along with these two charts, the Market Vane and Consensus surveys are now posted with bands that are 1.5 standard deviations from the 2-year mean.  Like all of the other indicators that I have switched to this format, the goal is to allow us to view current readings within the context of recent history and not get caught up in absolute levels, which is dangerous for sentiment analysis.  You can see from the charts on the site that we have reversed from the extreme negative sentiment we had been seeing.  This week saw a large shift to the bullish camp in the II, AAII and Consensus surveys.  This shift has caused the AAII bullish ratio and Consensus survey to approach their upper standard deviation bands, while the II bullish ratio and Market Vane survey are currently neutral.  The negativity is gone, and two of the surveys are suggesting that bullishness is approaching excess.  Below is the Consensus Inc survey going back to the beginning of 2000, with the circles corresponding to instances where the survey peaked at over 49% bulls:

We can see that although the survey has gotten significantly more bullish than it is now, when bullishness peaks above this level it has lead to a market drop within a few weeks without exception.  We can also see that the respondents are quite a bit more bullish than they were at the August top, even though the S&P has not rallied as much now as it did then (low to high).  The NDX has exceeded the July- August move, so undoubtedly that is influencing the futures traders who are the subject of this survey.

AIM - What a difference a week makes, as the rise in bullishness in the sentiment survey pushed this model from strong positive to neutral.  The model is now just above the level during the August top, although still far above the lower (bearish) standard deviation band.

COMMITMENTS OF TRADERS - The string of relatively muted changes in this data was broken during the latest reporting period, and not in a good way.  The S&P 500 commercial traders increased their shorting activity heavily, adding over 11,300 contracts to their net short position.  The rally off the 10/08 low didn't fail immediately, and that was apparently the sign the small specs were looking for to cover their short positions.  This has resulted in an increase in their net long position of just over 11,100 contracts.  This net change between the two groups is the most bearish development since just before the July waterfall selling and the March high before that.  Although the rate of change certainly adds a tone of caution to our outlook, the stochastic readings from the two camps are still on the bullish end of neutral and not a cause for concern at this point.

STEM - Not much to say here, as the model basically didn't move all week.  The one thing that did happen was that the standard deviation bands began to come down, which allows us to consider the model neutral instead of overbought.  If we sell off soon, this will let the model become bullish earlier than it would have otherwise.

STEM.MR - The choppy move higher over the past two weeks has frustrated this model, as it has not been able to reach an extreme.  When it gets close, price reverses (usually with a gap open) and alleviates much of the pent-up pressure that was beginning to build.  The current situation is more of the same, as the decline Thursday afternoon was offset by Friday's staggered advance.  Net net, the model is neutral and will remain so until something of a trend in either direction is evident.

PUT/CALL RATIOS - We didn't see any real extremes this week, and the readings were neutral to a little low in the total and equity ratios.  This has caused the 10-day and 21-day moving averages to drop considerably, and they are now approaching their lower standard deviation bands.  We continue to see muted activity in the OEX options, so those averages continue to hug the upper band.  These two factors are causing the 5-day and 10-day spreads between the two to remain close to the bearish deviation band, which lends a slightly bearish overtone to the complex, as it suggests that while the poorly capitalized traders are decreasing their put purchases, their more savvy index trading cousins are not betting aggressively on further upside potential.  In the S&P 500 options, the put/call bid/ask bias ratio was relatively high at Thursday's close (bullish) but Friday's rally caused it to drop to .69, which is neutral to very slightly bearish.

BREADTH RATIOS - Last week I said that it would be extremely difficult for the 10-day advance/decline line to become oversold, but quite easy for it to become overbought.  It did indeed become somewhat overbought on Wednesday, and is working off that condition now.  This coming week, we will be dropping a mix of readings, several of them largely positive, so it will become easier for the indicator to cycle back to oversold with any selloff.  If we rally, it will have to be on extremely positive breadth in order for us to become overbought again.  This either means that we are losing momentum and any rally will fail, or we have rested an adequate amount of time for the market to catch it's breath (I refuse to do the obvious pun).  I've seen both occurrences historically, and am not convinced of the validity of either - I prefer to stick with strict overbought and oversold readings and not try to read too much into the indicators.  The 10-day up volume indicator has come off its hyper-overbought reading and will spend the next week dropping a series of large readings, which will help push it down further.  The cumulative TICK indicators are now all at or very near overbought territory.  The NYSE intraday indicator closed Friday at +6160, which is extremely overbought.  It is very rare for the market to muster significant upside - even in the context of a 1/2 hourly uptrend - when this indicator is so overbought, without first working off the condition by either consolidating or staging something of a decline.

NYSE MEMBERS REPORT - Total shorting activity increased yet again in the latest reporting period, which ended on 10/11.  The chart below shows total shorting activity plotted against the NYSE Composite Index, with the red and green bands representing 1.5 standard deviations from the one-year average.

We can see that when the total amount of shorting picks up enough to exceed 1.5 standard deviations from the average - regardless of who's doing it - it has tended to indicate an excessive amount of pessimism and occurs near market lows.  We are at that point again, as the total shorting level has increased dramatically since August.  Of these total shorts, the NYSE Specialists accounted for 38%.  While not extreme (determined by deviation from a long-term average as you can see on the site), it is a relatively positive indication that the public is shorting these moves down rather heavily.

The positive sentiment indications for the health of this market are dwindling quickly.  While we had a large number of positives just a few weeks ago, we're down to seasonality as the lone positive this week.  Conversely, we also don't have a large number of negatives to point to - rather a majority of our indicators are neutral.  And for the first time that I can remember, every one of our models is neutral.  Anyone who has subscribed for a few months knows that I believe in aggressively pressing my bets when all of the models are skewed in one direction or the other, and that edge has performed well and consistently.  When all of the them are neutral, that tells me to stay cautious and either go with the trend or shorten my time frame.  Currently, the short-term trend is up and the long-term trend is down.  So, by default, that is going to have to be my stance as well until something either changes the trend or changes the models and indicators.  While I would be hesitant to hold long positions after a nearly 20% move off of a low in the context of a bear market (and the negatives mentioned above), I have to respect the fact that the market has not made a sizable correction in the face of some terrible news and overbought breadth.  This tells me there could be underlying strength that will force us higher, and I do not wish to fight that until such time as a confluence of our sentiment indicators and/or models becomes overbought.

On a side note, I want to remind everyone that the subscription rates will be increasing on Friday, November 1st.  I have been allowing current subscribers to re-subscribe (even if their term is not yet up) and extend their term at the current rate, so if you wish to lock in the current savings please re-subscribe before Friday.  After that date, the new rates will apply.

 - Jason Goepfert