sentimenTrader.com

All rights reserved.

Sunday, December 29, 2002

Last week, I said that the week just past would likely be as meaningless as they come, and it certainly felt that way.  NYSE volume ran about 55% of the yearly average, which is even lower than the 70% I mentioned was likely a couple of weeks ago.  The technical picture looks precarious, but short-term sentiment suggests some upside relief may be forthcoming, though short-lived.

____________________________________

Indicator:  SEASONALITY

Status:  BULLISH

Comment:  Well, if those who oppose using seasonality in their trading need another arrow in their quiver, it sure looks like they just got one.  Some of the most traditionally positive days of the year have failed, and that has been getting a lot of press.  Usually, I've seen statements about how seasonality no longer works, which is always very funny to me.  Hopefully, you understand that if something works 8 times out of 10, then there are 2 times where you should expect it to fail.  When you get one of those 2 times, it doesn't mean the other 8 will never happen again.  Since market logic tends to be perverse, something tells me that since the Santa Claus effect got so much press and it hasn't materialized, just when the national press will jump over its failure will we rally.  Seriously, watch for an increasing amount of press about the failure of the year-end rally.  When it seems you can't read a financial story that doesn't include something about it, then you should be on the lookout for an upside reversal.  From what I've been reading, it should be soon.  If December ended on Friday, the S&P 500 would have given its worst December performance since its inception in 1950.  The logical question now is if we cannot rally during historically the most positive month of the year, doesn't that bode ill for January?  Actually, no.  Since 1950, there have been 12 down Decembers.  After those negative Decembers, the following January was up 6 times, with an average return of just over 3.0%.  The three best Januarys ALL followed negative Decembers.  When the S&P was down over 2.0% during December, the following January was up an average of 4.1%, and was positive 4 out of the 7 times.  The best January after such dismal Decembers showed a return of 13.2% while the worst showed a return of -4.6%.  So, obviously, it would be incorrect to assume that January will be down just because December is down (did I mention that market logic is perverse?).  In fact, if anything, I would propose that we have a greater likelihood of a very positive January than we do a very negative one.  For what it's worth, the next week continues the string of relatively positive days historically (especially this coming Friday).

Bottom Line:  The much-hyped positive seasonality has failed so far this year, so the next week, while generally positive, has to be considered suspect.  However, if historical precedence holds true, January has a decent chance of giving a good performance.

____________________________________

Indicator:  STEM.MR MODEL

Status:  BULLISH

Comment:  Our shortest-term model is now the most positive it has been in over a month, with a current reading well over its upper standard deviation band.  While I do NOT use this information as a trading signal (and I suggest you don't either), it does serve as a useful guide of the probability of an upside reversal over the coming day(s).  Over the past year, when this model has been in strong positive territory (by exceeding its upper band), and the S&P closed above a prior 30-minute close, the average return a day later was 0.5% and the market was positive 62% of the time.  Nothing to get too excited about, but I have found that when this model is positive and we begin to see signs of a price recovery (such as closing above a prior 30-minute close or high), then the chances of upside follow-through are very high.  This is applicable to very short-term traders only.

Bottom Line:  This model is suggesting that if we begin to see signs an upside reversal, then there is a high probability of it continuing at least intraday.

____________________________________

Indicator:  COMMITMENTS OF TRADERS

Status:  BEARISH

Comment:  During the latest reporting period, commercial traders liquidated approximately 4,800 more short contracts than long while the small speculators reduced their net long position by about 23,400 contracts.  On the surface, this is quite bullish behavior, certainly the most bullish we've seen in a couple of months.  However, this reporting period includes the December futures expiration.  In September I went over some statistics regarding futures expiration and its impact on commercial and small speculator positions.  Over the past decade, since the CFTC has released the numbers weekly, the average expiration-week change (using absolute values) for the commercials has been 73% greater than on non-expiration weeks, and over 100% greater for the small specs. If we don't use absolute values and instead look at straight changes, the average commercial change during an expiration week is +3,039 contracts but during a non-expiration week it is -320 contracts. For the small specs, the average expiration week change is -5,446 contracts and the average non-expiration change is +594 contracts.  For the commercials, 36% of the expiration weeks had changes of greater than 10,000 contracts as opposed to only 8% of non-expiration weeks. For the small specs, 26% of expiration weeks had changes of greater than 10,000 contracts as opposed to only 5% of non-expiration weeks.  So we can see that expiration likely had a large impact on this week's reading, but even so the implications are positive.  It is perfectly normal for the commercials to become less net short and the small specs less net long during a generally down market, and expiration only exacerbates those changes.

Bottom Line:  The most recent changes in the S&P 500 futures appears constructive, but there is an excellent chance that expiration played a large role in the position changes.  Therefore, until we can see the next report, I would still consider this data bearish in the intermediate-term.

____________________________________

Indicator:  SENTIMENT SURVEYS

Status:  NEUTRAL

Comment:  In Thursday's daily commentary, I pointed out the bearish implications of the stubborn bullishness apparent in the Investor's Intelligence survey, so I won't go over that again (the commentary is available in the Daily Commentary archives on the site).  The other major surveys showed no meaningful change during the last reporting period.  The lowrisk.com survey, which I find to be a leading indicator of the four major surveys (II, Consensus, Market Vane and AAII), showed a large jump in bullishness during the survey period ended last Sunday.  The bullish ratio in that survey rose from 31% to 54%, pushing the 4-week moving average to 46%.  This is an extremely high level of bullishness on a moving average basis considering the market environment we've been in during that time.  An unabashed bullishness in the face of steady or falling prices is not at all conducive to a meaningful upside reversal, especially in the context of a bear market.

Bottom Line:  Same as last week - the negative developments in the surveys (i.e. divergence between price and optimism, and high absolute levels of optimism) will take many weeks to work off.

____________________________________

Indicator:  AIM MODEL

Status:  NEUTRAL

Comment:  I've been pointing out the divergence between price and optimism (i.e. price makes a lower high but the sentiment surveys show more optimism than the prior peak) for several weeks, and amazingly this model is still above the reading it gave at the peak in August.  So despite falling prices and an apparent failure to exceed a previous high, the survey respondents were more bullish than when prices were higher a couple of months ago.  I pointed out historical instances of these divergences before, both positive and negative, and they usually take many months to wear off as investors gradually become disillusioned with prices going the "wrong" way.

Bottom Line:  The negative divergence between price and the model, and the low relative level of the model itself, will take at least several weeks to wear off.

____________________________________

Indicator:  TRIN

Status:  NEUTRAL

Comment:  The 10-day TRIN reading for the NYSE, at a current level of 1.54, is close to its one-year upper standard deviation band at 1.57.  I would consider a breach of this level to be significant in the short-term, as those bands should contain 80% of all readings.  We will be dropping two relatively low readings on Monday and Tuesday so a continuation of Friday's selling would almost certainly push us over the upper band.  This would be a positive indication for a 2-3 day relief bounce at least.  On the Nasdaq, the current 10-day reading of 1.49 is well below what I would consider to be an oversold reading of over 2.0, so there's still some work to do there.  One interesting thing about the TRIN is that since 1966, December accounted for only 4% of instances of the 10-day TRIN poking above its upper standard deviation band.  One would expect a percentage closer to 8% (1/12) if these readings were spread evenly.  Here is the breakdown by month:

We can see that January and October had more than their fair share of high sustained TRIN readings, while December had fewer than any other month.  So, I think it's safe to say that seeing such an elevated TRIN during this time of year is unusual.  But does it mean anything?  The following table gives the results for several days after the 10-day TRIN exceeds its upper band.  The study period was for the S&P 500 from March 1966 - December 2002.

We can see here that when a high 10-day TRIN occurred during the month of December, the following days were considerably more positive than if the reading had occurred during any other month.  If we look at just the readings during bear markets, there were really only two Decembers had readings that would count.  The one in 1974 preceded a further decline while the one in 2000 preceded a small rally.  Obviously, there's not much we can do with that.

Bottom Line:  The 10-day TRIN is close enough to its upper trading band to consider the market oversold, although a couple more days of selling would truly get us to a point where an upside reversal lasting several days would be likely.  Over the past 36 years, very high TRIN readings have lead to market out-performance over the short-term, particularly so during the month of December.

____________________________________

Indicator:  VOLATILITY MEASUREMENTS

Status:  NEUTRAL

Comment:  The combination of moving average envelopes and RSI that I post to the site for the VIX and VXN proved to do a pretty decent job over the past couple of weeks at identifying the short-term swings during the trading range we've been in.  Currently, we're closer to the upper end of the range on the VIX (no coincidence that coincides with the lower end of the range on the S&P), which would suggest a bounce is forthcoming if the recent pattern holds.  The VXN hasn't moved much at all over the past few days, and is still closer to the bottom part of its range.  If we spike below the 1000 level on the NDX, I suspect the VXN would show more movement than it has been, but as long as we remain above that level, the VXN could continue its descent.  A couple of weeks ago, I presented three charts which show different measures of volatility.  The first was 10-day NYSE volume as a percentage of the 200-day average; the second was the 10-day daily average range of the S&P, expressed as a percentage of price; and the third was the 10-day average daily range of the NYSE TICK, with standard deviation bands around the readings to put them into the context of recent action.  The three charts are reprised below:

We can see that each of the measures are dangerously close to levels that have coincided with previous intermediate-term peaks in the S&P 500 during the bear market.  The suggestion here is that low volatility is a sign of complacency, and in the context of generally falling prices, is not an encouraging sign.

Bottom Line:  While not extreme by any measure, the VIX is close to a point that has coincided with recent short-term lows in the S&P.  If the pattern holds, we should see something of a bounce in equities early next week.  However, from the looks of our other volatility measures, any bounce should be short-lived.

____________________________________

Indicator:  RYDEX RATIOS

Status:  NEUTRAL

Comment:  After reaching two-year bullish (bearish to us) extremes in November, the intermediate-term Rydex ratios have almost completed a 180 degree turn.  We've seen a gradual shift of assets from the bullish funds to the bearish, pushing our longer-term indicators very close to levels that in the past have signaled extremes of bearish sentiment (and short- to intermediate-term upside in the market).  While that is constructive, we haven't seen the sharp swing in asset flows that has accompanied most of the significant swings of 50+ points in the S&P over the past year.  If we would see a 2-3 day exodus from the bullish funds, I would be more comfortable in expecting a tradable bounce sometime soon, but that hasn't happened yet.

Bottom Line:  The asset flow in this mutual fund family is the most constructive it has been in a couple of months, although we have not quite seen the short- and intermediate-term swings we have seen at past tradable low points.

____________________________________

Indicator:  COMPOSITE MODEL

Status:  NEUTRAL

Comment:  YAWN.

Bottom Line:  Since the October low, this model has gone nowhere, which has been a good representation of the market conditions.  Until some kind of extreme is reached, we shouldn't read too much into this model's readings.

____________________________________

Indicator:  STEM MODEL

Status:  NEUTRAL

Comment:  The positive indication this model gave in early December lead to not much more than a three-week trading range, and in fact it was closer to its bearish band than upper just a couple of days ago.  Currently, even with a rather heavy two-day decline, we're still neutral here and quite a ways from a positive reading.

Bottom Line:  This model is suggesting that there's more work on the downside before a meaningful multi-day reversal is likely.

____________________________________

Indicator:  PUT/CALL RATIOS

Status:  NEUTRAL

Comment:  Tuesday and Thursday we saw a combination of low equity put/call ratios and high OEX put/call ratios.  Those who have been reading this site for a while know that we view the equity p/c ratio in a contrarian light (i.e. high equity p/c ratios are bullish and low equity p/c ratios are bearish) and the OEX p/c ratio in a non-contrarian way (i.e. high OEX p/c values are bearish and low OEX p/c values are bullish).  So the combination we saw mid-week was bearish, and that continued - although to a lesser extent - on Friday.  This has served to push the 5-day spread between the two ratios above its upper standard deviation band for the first time since mid-September, meaning that the usually wrong-way equity options traders have likely been betting significantly more money on an upside reversal than their more savvy index-trading cousins.  That type of activity certainly didn't bode well in September.  Other than that, the 10-day and 21-day moving averages of all of the  ratios are stuck in the middle of their trading bands and will most likely stay there for the coming week.

Bottom Line:  While the ratios taken in isolation are currently neutral, the recent activity in the OEX vs. equity options is bearish.  Such high levels of institutional-quality put buying and retail call-buying is never a good sign, especially in the context of a defined downtrend.

____________________________________

Indicator:  BREADTH RATIOS

Status:  NEUTRAL

Comment:  The 10-day advance/decline line has been stuck in neutral for the past couple of weeks after working off its overbought condition from late November.  Chances are good that it will remain neutral this coming week, unless we get a substantial multi-day rally.  We'll be dropping relatively large negative readings four out of the next five days (thus pushing the 10-day average higher), so it's quite unlikely we'll approach oversold this week.  Volume has been more skewed to the downside over the past week, helping to push the 10-day up volume ratio down to 41%, just above the 40% threshold that serves as a good approximation of an oversold market, but well above the readings of 30% or so seen at recent intermediate-term lows.  The cumulative TICKS are all neutral, although the intraday NYSE indicator is approaching oversold at -344 (-1000 could be considered extreme).  As always, however, oversold in the context of a downtrend is not as effective as oversold in an uptrend.  Our proprietary Down Pressure indicator for the S&P is on the upper end of neutral with a reading of 73%.  The Down Pressure on the NDX is currently bullish (oversold) with a current reading of 82%, above the 80% threshold.  Over the past five months, such a high reading has been an effective buy signal in the NDX, with the average next-day return being 2.1%, with 7 out of 9 signals being positive.  After 3 days, the average return climbs to 2.8%, with 6 out of 9 signals resulting in a higher market.  As you'll recall, this indicator is constructed by measuring the volume and magnitude of point moves in the 100 component stocks of the NDX. 

Bottom Line:  A couple of our indicators, namely the intraday NYSE cumulative TICK and the NDX Down Pressure, are indicating that some short-term upside relief should come soon, most likely within a day or two.  However, there is nothing in the intermediate-term to suggest that the move (if we get one) should last for more than a couple of days.

____________________________________

Indicator:  NYSE MEMBERS REPORT

Status:  NEUTRAL

Comment:  Total shorting activity was relatively low during the latest reporting period (for the week ended 12/13/02), and the ratios between the various participants remained stable.  The Specialist Short Ratio was unchanged at 37%, which is a bit surprising.  I say it's surprising because over the past 40 years, when the NYSE Composite declined around 1.9% during a given week (as it did during this reporting period), then the SSR also declined 72% of the time, for an average of 3%.  The fact that it didn't suggests that the public didn't pick up their short activity during a down market, which is a sign of complacency and argues for lower prices.  However, I would need to see more of a trend before drawing too many conclusions here.

Bottom Line:  Although we should have seen a pickup in public shorting activity during this reporting period, the fact that we did not is not especially bearish considering the relatively heavy shorting they have engaged in over the past few months.

We have a few short-term signs that pessimism has reached a sufficient extreme to trigger something of an upside reversal, so I would look to play the long side if we start to see a price recovery.  However, this is with the understanding that it would be short-term only, within the context of a downtrend and overall negative sentiment condition.  I've been writing about the 885 level on the S&P and 1000 level on the NDX, and the fact that we broke that level on the S&P has me cautious on any longs.  The 875 level on the S&P (where we are currently) should also serve as support, so this level will be important to watch on Monday.  If the bulls are going to stage a reversal, it better happen here.  Any reversal which takes us up 5%-8%, however, should continue to serve as an opportunity to initiate intermediate-term short positions.

 - Jason Goepfert

Disclosure:  long QQQ calls, long QQQ puts

This disclosure is not intended as trading advice in any form.  It is meant as a note to subscribers that I have a position directly affected by my market outlook.  Although I take great pains to remain objective in my commentaries, I believe it is only fair that readers should know that I have taken positions in accordance with my market outlook.