sentimenTrader.com

All rights reserved.

Sunday, August 11, 2002

I said last week that I would begin to look for upside reversals beginning last Monday, albeit with a greater requirement for confirmation.  That proved to be generally the right attitude as the S&P and NDX both rallied 5% week over week.  For the coming week, things are less clear, as three out of the four models are now neutral and our various indicators are spread out all over the map.  A 17% bounce off the lows in the S&P will do that to us.

Indicator Summary:

BULLISH

NYSE MEMBERS REPORT - The most recent reporting period was rather uneventful, and the members report reflects that ambiguity.  Although the absolute level of public shorting recorded another record, the relative level compared to NYSE specialists and members was nothing extraordinary, as the specialist short ratio inched up to 38% from 35% during the previous period.  Again, I refer you back to the July 28th weekly commentary (available in the archives) which presented a table of one-year returns after such a ratio.  This continues to be a long-term positive for the market.

SENTIMENT SURVEYS - Last week, I said that if the increased bullishness that the sentiment surveys showed were reflections of actual positions taken, then a level we may want to watch on the S&P was 832.  This week's low was 833.44.  I won't even begin to pretend that the 832 level held any real meaning, but I think when there are fairly dramatic moves in these sentiment surveys, it benefits us to pay attention to what general level many of those investors may have taken positions near.  I said on Monday with the "sneak peak" release of the lowrisk.com survey that we may see an increased level of bearishness from the more established surveys, and that was indeed the case in all but one of them.  The Investor's Intelligence survey continues to be an intermediate-term positive for the market as it regains its inverted status of more bears than bulls (although at some of the better lows this survey has remained inverted for months on end).  The Market Vane survey dropped a few percent to 24% and the noisy AAII survey dropped just a touch of bullishness.  The only survey to show an increase was the Consensus survey, which increased to 32%, right on the cusp of historically bullish readings.  This survey in particular has held up from its historical lows more than the other surveys, never really reaching an historical extreme.  Coming off of the truly unprecedented readings over the past few weeks, I don't expect these surveys to tell us anything new for quite some time.

AIM - With the increase in bearishness during the most recent survey period, this model has turned up once again, with a current reading of 90.5 as opposed to 86.4 last week.  This is our 10th straight week of readings at or above 80, surpassed only by the 22 weeks in the summer of 1994.  As readings over 90 have historically been extremely positive for the intermediate- to long-term prospects for the market, this model has changed stages from POSITIVE to STRONG POSITIVE with this week's reading.

NEUTRAL

VIX and VXN - The extremely high levels of implied volatility reflected in these indicators will not last.  That much is a given.  The only question is how high they will go before they begin that steady retreat back to their mean.  It's quite possible (even probable) that we have already seen the highs and these indicators are currently beginning to stair-step back down to more reasonable levels while the markets work off their historic level of oversold.  The action in these indicators has been very unusual, with some of the largest daily swings in their history.  In fact, the past few days have seen two of the three largest one-day high-low drops in the existence of VXN (which is not at all extensive).  These huge drops were exceed only by the drop on September 24th.  That volatility drop lead to an incredible Nasdaq rally.  Typically, these violent swings when implied volatility is elevated is a very positive development for the intermediate-term health of the market, and I believe that's also the case this time.  However, these daily swings also make me very uncomfortable and put my subjective "something fishy is going on" radar on full alert.  Please take this with a large grain of salt, but it just feels to me like these volatility measures want to spike higher one last time (corresponding of course with a large market drop).

STEM - Like the Composite model below, this short- to intermediate-term model has come off its historical extremes and is sitting on the upper end of neutral.  This is certainly not a negative for the market, it just tells me that the extremes we have seen have served their purpose and are now awaiting the next swing.

STEM.MR  - We've now had three straight days with no extremes in this shortest-term model, and that definitely won't last much longer.  Three days is about as long as this model ever goes without reaching one extreme or the other, so I would expect a large move early next week that tips the scales in either direction. 

COMPOSITE MODEL - The market turn that this model has been telling us was increasingly likely has happened, and so the model is rapidly losing its bullish extreme.  This by no means suggests that the recent rally is over, only that the sentiment extremes seen at bottoms are no longer there (or at least as extreme).  With a current reading of 54%, the model has slipped out of positive territory and is neutral at the moment.

BREADTH RATIOS - I said a couple of weeks ago that it wouldn't take long for the advance/decline oscillator to cycle from extremely oversold to overbought, and it certainly didn't as we reached overbought this week.  Since we've seen rather large positive breadth readings this week, the large numbers that are dropping from the moving average are not having a major impact on the oscillator.  We have two more large positive readings to drop over the next two days, so it's likely this line will drop a bit before making another thrust upward as we begin to get rid of a string of three very large negative readings.  That thrust upward will most likely take us to an extremely overbought situation, unless we sell off from the get-go on Monday.  The cumulative TICK indicators, particularly on the NYSE, continue to hover around dangerous territory.  The intraday indicator didn't suffer much of a decline with Friday's morning dramatics.  As I said last week, however, these cumulative TICK indicators lose much of their effectiveness when giving overbought signals within an uptrend or oversold signals in a downtrend.  Since I think it's safe to say we're in a 30-minute uptrend, I have to assign this overbought reading a little less weight than I normally would.

TRIN - The 10-day average is still stuck in neutral, as it has been for several weeks, although we are slowly approaching the May lows.  The overbought signals in this indicator have certainly fared better than the oversold signals, and I'll be paying close attention to it if we continue to get consistently low readings over the coming week.

COT - This week's reports were pretty much a wash, as the commercials reduced their net short position by 4,800 contracts while the small specs increased their net long exposure by 5,700 contracts.  The end result is no change in the one-year Futures Balance Matrix and a slight decrease in the three-month Matrix.  The most extended moves tend to come when the one-year and three-month stochastics for the commercials are in direct opposition to the small spec stochastics (meaning the commercials are getting longer at the same time the small specs are getting shorter).  The fact that the small specs are continually trying to game the bottom of the market, after two straight years of being wrong, is still troublesome to me.  However, as of the most current reading ending this past Tuesday, the one-year commercial stochastic is the most positive it has been since late January and late September before that.  This is an encouraging sign for the intermediate-term.

PUT/CALL RATIOS -  After being generally bullish for the better part of a month, the various options indicators that we follow are beginning to show some cracks.  The total put/call ratio is beginning to have more frequent low readings, suggesting retail options traders are beginning to see enough upside possibility as to risk their funds on calls in greater proportion than puts.  The 3-day moving average of the one-year p/c rank shows that we are now seeing the most optimistic "dumb money" trading since the 7/30 high and the May high before that.  It's nowhere near a bearish extreme at this point, but it does show a clear shift in sentiment.  The 10-day and 21-day moving averages of the p/c ratio are well off their bullish extremes hit several weeks ago and have been slowly working themselves lower since.  They are both firmly in neutral territory at this point.  On the bullish front, the open interest ratio remains firmly seating in bullish territory.  Expiration usually has a profound impact on this indicator, so it will be interesting to see these numbers during the week of the 19th.  Most notably, and most troubling from a short-term perspective, is the put/call bid/ask bias ratio.  For those of you who are new, this proprietary indicator looks not simply at the volume flowing into the institutionally-dominated S&P 500 options, but also at the flow of orders going off at the bid versus the ask (one generally sells at bid and buys at ask).  This gives us a feel for whether institutions are buying or selling calls and buying or selling puts.  We then take that information and combine it into a signal ratio - if the ratio is high, then that means there is an institutional bias towards selling puts and/or buying calls (bullish for the market); when it is low, then that means there is a bias towards buying puts and/or selling calls (bearish for the market).  This ratio is usually effective at signaling short-term (< 5 days) market direction.  Friday's reading is the lowest I have ever recorded, at 0.09.  There was a very clear and quite aggressive bias towards selling calls AND buying puts, particularly in the September expiration.  In the most striking example to date, only 5% of September puts went off at or below bid, as opposed to 32% at ask.  Conversely, 35% of the September calls went off at bid as opposed to 8% at ask.  That is a crystal-clear bearish bias by the institutions who use these options.  As a point of reference, the July 30th high saw a reading of 0.24 while the explosive moves in late July were preceded by a reading of over 7.0.

The large gains off of the lows and the "retest" have left us in the most ambiguous state of affairs we've had for some time.  The trend on many timeframes is open to interpretation and the majority of our most useful sentiment indicators (and models) are sitting in neutral.  When we have situations where the sentiment situation is neutral, then the logical course is to simply follow the trend until another extreme, one way or the other, is reached.  The problem this time is trying to find and define a trend.  It depends on your time frame, of course, but on many it's ambiguous.  This will have me largely on the sidelines until a clearer picture emerges - when there's no edge to be had, it's usually best to step aside and wait.

 - Jason Goepfert