http://www.sentimentrader.com

Sunday, March 30, 2003

Last weekend, I outlined several factors that lead me to believe that we saw reckless speculation near the end of the prior week, and that short setups would be the preferred strategy going forward until some type of retracement was seen.  The nearly 4% decline in the S&P from last Friday's close did serve to alleviate much of that speculative fever as determined by some of our shorter-term measures, but it has not been enough to make much of an impact on some of the longer-term (intermediate) indicators.  This suggests that additional upside attempts from here, without either more downside or more elapsed time first, may struggle to get very far, particularly with the overhead resistance levels that we have.  If we do get that further retracement, and it's enough to work off the overbought nature of many of our intermediate-indicators, then we may be presented with an opportunity on the long side, provided we remain above the "line in the sand" at S&P 830.

____________________________________

Indicator:  SHORT SALES

Status:  BULLISH

Comment:  Public pressure on the short side picked up a bit in the most recent reporting period, to bring the Specialist Short Ratio down to 31%.  The four-week moving average of the ratio is now at the lowest point in seven years.  This is really the same story as the past few weeks, as the public had been shorting heavily since late January.  The average return after spikes lower in this indicator have averaged closer to 20% from low to high, so we may have room for another spurt higher.  This data is released with a two-week lag, so it does not encompass most of the recent move off the early March low.  It should be instructive to see how the public has treated the move - if they do not let up significantly in their shorting activity, it should bode well for continued upside momentum as they are scared out of their short positions.

Bottom Line:  From an intermediate-term perspective, the amount of public shorting seen since late January is constructive for building a base on which a sizable advance could spring.  However, if we look very long-term, we still have a ways to go before we will have seen the type of extended public shorting seen at other long-term lows during the past two decades. 

____________________________________

Indicator:  COMMITMENTS OF TRADERS

Status:  BULLISH and BEARISH (depends on time frame)

Comment:  Last week, I noted the large change in positions in the large S&P 500 futures contract.  At the time, I said that since the market had already risen over 3% from the time the report was generated, and the fact that expiration likely had an impact on the changes, that we should temper the bullishness of the readings.  There was also a report being circulated among institutional trading desks that a big reason for the large change was scrambling by traders who sold call options short, to buy futures in an attempt to mitigate their risk.  I was expecting to see another large change in positions this week (mainly due to expiration), and that happened, but in the opposite direction of how I was looking.  I thought we would see the commercials return to their large net short position, and the small specs add to their net longs.  Instead, we saw the commercials liquidate (or let expire) significantly more shorts than longs, returning them to a NET LONG position for the first time in three years.  May of 2000 (right before the last kick higher in the bull market) was the last time commercial traders showed a net long position, and they haven't been close since.  At the same time the commercials were becoming longer, the small specs were (of course) doing the opposite and becoming less net long.  Once again, they are the least net long since May 2000.  These two groups of traders are very close to becoming "inverted" for the first time in years, where commercial traders are more long than small speculators.  This was the typical relationship from 1986-1997, until small speculators caught on to the bull market  and began getting more long, more consistently.  Since 1997, it has been a rare occurrence for commercials to be more net long than small specs, and each time it happened, it lead to significant upside over the intermediate-term.  Of course, it hasn't happened yet this time around, and expiration undoubtedly had a large impact on these numbers, but this bears watching.

I also mentioned the troubling changes in the e-mini contract last week.  At the time, the small specs were the longest they had ever been in the contract, while the commercials were close to being the most net short.  That situation has been exacerbated this week, to a huge degree, as commercial traders stepped away from the long side in this contract.  The chart below looks at the small spec net position as subtracted from the commercial net position.  If the bar is high, then the commercials are largely net long and/or the small specs are net short.  This should be a positive for the market.  If the bar is low and below the zero line, then the opposite is true - the commercials are net short and/or the small specs are net long.  That should be negative for the market going forward.    For example, the current reading of -478,081 is derived from a commercial short position of 236,054 contracts and a small spec long position of 242,027 contracts (i.e. -236,054 - 242,027 = -478,081).  This differential was negligible in terms of the number of contracts prior to 2001.

From this, you can see why I have been relatively hesitant to use the e-mini as a forecasting tool.  Its record until recently has been spotty at best.  Since the last quarter of last year, however, it has improved significantly, and as contract volume grows it may become more so.  In any event, the differential has gone from the most positive in history just prior to the low a few weeks ago to by far the most negative in its history.

How do we reconcile the fact that the large contract appears very positive while the small contract is showing just the opposite?  Not an easy question, for sure, and you're bound to get different answers depending on who you talk to.  Personally, I consider the two contracts completely separate.  I think it is a mistake to try to create one figure out of both contracts, by adjusting the small contract for the large.  What I mean by this is that a one point move in the large contract is worth $250.  In the e-mini, that same point move would be worth $50.  So some analysts simply divide the number of e-mini contracts in the COT information by 5, and add it to the large contract.  However, I think that can be terribly misleading because the groups of traders who trade each contract can be vastly different.  The CFTC reporting limit for the large contract is 1,000 contracts.  With a current margin requirement of $17,813, that translates to a position margin of at least $17,813,000 before a trader would have to be reported.  In the e-mini, by contrast, the reporting limit is only 300 contracts.  With a current initial margin requirement of $3,563, it would only take a position of $1,068,900 to be reported, a difference of nearly 17:1 between the two contracts.  Since initial margin requirements are reduced by around 20% for those who designate themselves commercial traders, there may be an incentive for some funds to do so, which would be much easier to accomplish in the e-mini contract.  My point here is that there are most likely very different groups of traders making up the commercials and small speculators in the full contract vs. the e-mini, and I do not think the two should be mixed.

From what I have seen of this data so far, the e-mini has appeared to "work" on a shorter time frame than the large contract.  The positions change with much greater frequency and magnitude from week to week, so I don't think those positions should be extrapolated out too far.  The large contract, in contrast, has been most effective in the intermediate-term, which I consider to be at least several weeks.  So, taken as a whole, I believe the changes in these contracts, while likely influenced to a great degree by expiration, indicate that we may be more likely to see short-term weakness in the days ahead, but strength thereafter.

I want to touch on the COT information for 30 year bonds as well this week.  I calculate the same type of differential between commercial and small spec positions here as I do in the S&P 500 e-mini as discussed above.  This week, the variance went positive, which has been a rare occurrence in the past few years.  EVERY time the variance has gone positive since 1999 (and almost every time since 1995), meaning the commercials were relatively long and the small specs relatively short, bonds have rallied shortly thereafter.  The green arrows on the chart depict peaks in the variance above zero.  With the variance now positive, it would argue that bonds may be due for a rebound sometime soon (especially considering the state of the put/call and Rydex ratios for bonds). 

Since equities and bonds have shown a very strong negative correlation for the past couple of years (see the "Why Bonds" link on the bond indicator page), this would suggest that a rise in bonds would coincide with a decline in equities.  Let's take that same variance chart above and superimpose it on the S&P 500 since 2000:

Pretty good negative correlation here.  When commercial traders became positive on bonds, it correlated nicely with weakness in equities soon after.  We've now reached that same point this week.

So now we have three pieces of conflicting information within this complex.  The large S&P 500 contract argues for higher equities prices, while bonds and the e-mini suggest lower prices are more likely.  I prefer to weight them in order of record and relevance.  The large S&P contract has the most relevance, and the best long-term record.  Therefore, I would give the highest odds to a market rise in the intermediate-term.  But the recent record of the e-mini, and the high negative correlation with bonds, tempers that outlook considerably.

Bottom Line:  The position changes in both the full and e-mini contracts are notable, though they give opposite conclusions.  It may make more sense when we factor in the time frame in which each appears to be most effective.  In that case, we may expect further weakness in the short-term (since the e-mini information is so negative), but more strength in the intermediate-term (due to the positive changes seen in the large contract).  However, the accuracy of the COT information in bonds, combined with the high negative correlation between bonds and stocks, should temper the bullish outlook for equities.

____________________________________

Indicator:  RYDEX RATIOS

Status:  BEARISH

Comment:  The persistent and extreme asset flows I mentioned last weekend have had a large impact on our longer-term indicators here.  The bullish and bearish flows are now in truly nose-bleed territory, and this is a bearish development.  In fact, both set new all-time records on Friday.  Since the history of these indicators is relatively limited, it's hard to read too much into their meaning, but the fact that they have reached the degree of overbought that they have while we're still in the context of a longer-term downtrend is certainly troubling.  Interestingly, the previous records were set after the move in October 2002, which is also the closest comparison we saw in changes in the sentiment surveys, as outlined below.

The Beta Chase Index is only just beginning to roll over from its all-time record reading on Wednesday.  As I stated last week, such a high reading shows just how much speculation was seen in the run-up since the low in early March.  With a reading over 7, the index is telling us that high-beta funds have a relative strength more than 7 times that of the "safe", low-beta funds.  Such a huge shift in asset momentum to the most speculative funds that Rydex offers cannot be considered a positive sign, although it should settle back considerably this week, as the excesses prior to last week are wrung out.

Some have noted that the RSI spread indicator, our shortest-term one in this complex, is now at an oversold level.  This indicator is based on the relative momentum of funds shifting between the bullish and bearish index funds, and it does indeed show that the tide this past week shifted.  However, this is more because of a slowdown in asset movements into the bullish funds, as opposed to a pick-up in betting on the short side by these market timers.  This is not especially bullish, and it should be expected as we come off several days of the indicator reaching the unprecedented +100 level.

Bottom Line:  These timers have taken something of a breather this week, after their frenzied panic into the long side the week before last.  As I stated last weekend, it will take time for this hangover to work itself off.  Our longer-term indicators have just entered extreme overbought territory, and it often takes several weeks for these types of readings to be alleviated.  We saw a tremendous amount of speculation that the move off the March low was going to continue, and the result of that speculation is usually a flat (trading-range) or declining market.  The least likely scenario would be enough upside to reward these traders for their panic.

____________________________________

Indicator:  SENTIMENT SURVEYS

Status:  NEUTRAL

Comment:  The negative momentum the major surveys had displayed going into March has dissipated, as they are all back to readings that have been historically neutral.  With a 10% rise off the low, this is not surprising.  I went back over the past decade and looked at periods where we declined at least 10% from high to low (weekly closes) over the previous 60 days, then rallied at least 5% off the low.  I wanted to see how our current readings compared to those other instances after such a move, to see if investors were buying into this rally more or less than previous ones.  The table below shows the change in each survey from the time of the low to the time where the market rallied at least 5%.  For example, an II reading of 9.1 below means that the bull ratio (bulls / (bulls + bears)) in Chartcraft's Investor's Intelligence survey increased by 9.1% percentage points between the low and the time the market rallied at least 5%, based on weekly closes.

* estimate

We can see that investor attitudes towards recent move are quite optimistic, based on previous similar moves.  While the percentage move off the low is about average, the percentage change in the sentiment surveys is above average for each one.  While none of the current changes are the largest we have seen, overall it is one of the greatest confluence of changes of any of the other lows, perhaps eclipsed only by October 2002.  I don't see any particular pattern between increased optimism and the market return 2 weeks later, although there does seem to be a slight relationship between the move off the low and the changes in the surveys - the bigger the move, the more optimism as displayed by the surveys.  Nothing earth-shattering or especially useful about that, though.  Unfortunately, I don't think this table tells us too much about what to expect going forward, except perhaps that the returns 2 weeks subsequent to these moves are not particularly dramatic.  This goes back to what I was saying earlier last week - that one of the most likely scenarios for the coming week(s) is a trading-range environment.

Bottom Line:  Investor attitudes may be a bit more optimistic that this recent low will hold than they have been about most others.  While we have seen more "buy-in" of this recent move than any other low, except perhaps for the move in October 2002, that doesn't necessarily mean we have seen too much and are about to roll over.  If anything, I think this tells us that we may see a continuation of a trading-range environment.

____________________________________

Indicator:  AIM MODEL

Status:  NEUTRAL

Comment:  With the move in the surveys back to neutral, this model has obviously followed.  The model is now about 10% of its highs, after reaching an extreme of 66.5% two weeks ago.  While the current reading is still somewhat high, it is certainly no longer extreme and is moving in a pattern that is typical of most lows such as we have seen.

Bottom Line:  Not much to glean here, other than the fact that we're not seeing anything unusual from past lows.

____________________________________

Indicator:  SEASONALITY

Status:  NEUTRAL

Comment:  Last week, I outlined a study which lead me to believe that any upside we saw in the broader market going into the last few days of the month was NOT related to quarter-end manipulation by mutual funds and hedge funds, even though it would surely be the reason given by almost every media outlet.  I received quite a bit of feedback on that comment, so let me make a clarification - I have no doubt that this type of price manipulation occurs.  I'm sure that it does, particularly in some of the smaller capitalization issues with volume patterns more susceptible to that type activity.  However, I am not convinced that it is widespread enough that it is able to lift the entire market for days on end.  One subscriber suggested that with the proliferation of hedge funds in the past decade, perhaps just looking at the last 10 years or so would lead to a different conclusion.  I narrowed the study to just that time frame, and indeed the results changed somewhat.  If the market was up for the month going into the last few days of the quarter, then the average return those last days was 0.24%, with a 67% probability of being higher.  This is quite a bit higher than what the other study indicated.  However, if the market was DOWN going into quarter-end, then the last few days gave an average return of negative 0.46%, with a 41% chance of closing higher.  This is worse than the other study.  On the surface, I think this may show that the increased presence of hedge funds has exacerbated the volatility going into quarter-end, and on both sides of the market.  With so many funds now concentrated on the short side, they certainly have an incentive to press the market down and increase their returns during a down month.  However, the sample size is small enough, and the results not variable enough, that any conclusions drawn may be misleading.

This coming week, we will have the end-of-month, beginning-of-next-month phenomenon working for us.  As you can see from the Seasonality section of the site, this pattern has been relatively strong and consistent for many years.  One factor which may dampen the effects this time around, however, is the fact that the 50-day moving average on the S&P is still declining.  Over the time period studied, this has served to significantly weaken the effects of this bias. Also, the first trading day in April has been negative the past two years, the second trading day has been negative the past three years, and the third trading day negative the past four years.  Reversing that trend, the fourth trading day in April (Friday in this case) has been positive 8 of the past 10 years, and the two negative instances were just barely down.  Overall, April has been a fairly strong month, and is the end of the traditionally strong November-April period.

Bottom Line:  I don't think there's anything about this week that would give us a tradable edge.  While the end of month pattern is relatively strong, the fact that we are in a down market, and recent history has been negative, serves to reverse a lot of that positive bias.

____________________________________

Indicator:  BREADTH RATIOS

Status:  NEUTRAL

Comment:  I mentioned the overbought nature of our breadth measurements last week, and they haven't changed to a great degree yet.  The 10-day advance/decline and up volume ratios would have been "maximum" overbought had we gotten an up day last Monday, but they came close enough to warrant our attention.  At this point, it will be difficult for either of them to become more overbought than they were last week, as we'll be dropping a string of positive days all next week.  As positive days are dropped, it requires additional positive days just to keep the averages where they were.  The longer-term 21-day averages of both indicators are now on the upper end of neutral, but nothing particularly notable there.

Our Down Pressure indicators have worked off the overbought conditions they entered last week, and are now neutral on the S&P and NDX.

The daily cumulative TICK readings on the NYSE and Nasdaq remain in overbought territory and at the most extreme they have been in over a year.  Persistently high TICK readings on a daily basis in the context of a longer-term downtrend are usually a sign of a market proceeding at an unsustainable pace, so I think this should be watched carefully.

After approaching overbought early last week (and achieving that status on the Nasdaq), the 10-day TRINs have worked their way higher and relieved some of that overbought condition.  Two weeks ago, I showed a table comparing peaks in the TRIN at oversold extremes, and how it typically lead actual market lows by a couple of weeks.  That suggested we were in the initial stages of a selloff, and a market low was not imminent.  Obviously, that was completely wrong as the market reversed immediately, but the fact remains that overbought TRIN readings have worked much better during this bear market than oversold readings. 

Bottom Line:  Often, after we reached the level of overbought that we have, and it comes after a large move from an oversold condition, the immediate market reaction is not a steady decline.  It is more of a choppy, trading range environment which is what we saw last week.  This can persist for several weeks, and would be the most healthy thing the market could do.  If we instead rally from here, without giving the market a chance to rest, any rally would likely be doomed to failure, and we would probably fall back harder than we rose. 

____________________________________

Indicator:  PUT/CALL RATIOS

Status:  NEUTRAL

Comment:  The equity put/call ratios were once again skewed by a large trader in the QQQ options on Friday, much like what happened in February.  According to my quote vendor, someone traded over 200,000 contracts each of the April and May 22 put options on the QQQ.  My vendor shows that the April contracts went out at bid (suggesting they were sold), while the May contracts went out at ask (suggesting they were bought).  This is a debit calendar put spread, and as a pure option trade, is neutral to bearish, depending on how aggressive the trader wants to be.  However, and this is my problem with trying to decipher this information with no particular inside scoop, there is no telling if this is part of a larger position, which it almost certainly is.  This trader is most likely long or short the underlying security, which changes the complexion of the trade.  On the surface, it is neutral or bearish (but NOT bullish), but with no knowledge of the underlying strategy, it is impossible to say for sure what the trader is trying to accomplish.

Needless to say, such a large trade greatly skewed the put/call numbers, and we'll have to live with that until they drop off the averages.  I assume these trades will be closed out sometime in the next few weeks, so we can look forward to another bulge in the ratios sometime down the pike.  While Friday's reading helped push the 10-day and 21-day averages off the overbought levels they were flirting with, we still are only just beginning to work off some of the excess we've seen recently, as I mentioned last weekend.  It will take time for the averages to work back to neutral, so in the meantime the suggestion is that any upside attempts from here would likely fail fairly quickly, as we've seen too much speculation recently to support another solid push higher.

Bottom Line:  The large QQQ trade will "artificially" inflate the equity put/call ratio for the foreseeable future.  We're just beginning to work off the excess speculation seen during the last leg higher, and that suggests that any rallies from here would likely fail quickly. 

____________________________________

Indicator:  VOLATILITY MEASUREMENTS

Status:  NEUTRAL

Comment:  The VIX and VXN are on their way to reaching the other side of the extreme they hit on March 12th.  While we haven't seen yet the spike lower in these indices that would suggest there was a sudden lifting of uncertainty, they have slowly drifted down to a slightly oversold condition (due to the contrary nature of the indicators, this translates to an overbought reading in equities).  The way I look at them, both indicators are now the most oversold they have been since early January.

On March 12th, I mentioned the extreme nature of the VIX Fear Premium, and that it had reached a level rarely seen at anything other than intermediate-term lows.  I wished I had paid more attention to the reading, as it was an excellent notification that we were seeing an unsustainable amount of uncertainty being priced into the market.  The indicator posted to the site is a 10-day moving average, and it began to decline quite dramatically this week, as historical volatility has risen while implied (i.e. estimated future) volatility has dropped.  Should we continue a choppy trading environment, the indicator will likely decline even further this week, as several high readings will be dropping off the 10-day average.  This is typical after moves such as we've seen, and should only be considered troublesome if we drop back close to a 0 reading.

My biggest problem with expecting a rally in early March was that we were not seeing the type of volume and volatility almost always seen prior to an intermediate-term low.  The beginning of the rally was more reminiscent of the quick pop off the December low, which also saw a low amount of volume and volatility.  Obviously, the reaction of this low has been quite a bit different than December, so I don't think that comparison holds, but we still have not seen volume come into the market in a sustained way.  The action now, as you can see from the Volume Differential indicator posted to the site, is much more typical of declining markets than rising ones.

Bottom Line:  The unwinding of market uncertainty has not yet reached a stage that would suggest a top is imminent.  However, another spike higher in equities would likely accomplish that, which suggests that it would probably fail pretty quickly.  We never saw the volume and volatility pattern that have lead to sustainable lows in the past, so that makes this move suspect anyway.

____________________________________

Indicator:  STEM.MR MODEL

Status:  NEUTRAL

Comment:  This model has bounced around overbought a couple of times this week, and is sitting just above that territory now.  Another move higher in equities early this coming week should give us an extreme overbought reading fairly quickly.  That type of reading normally gives a good heads-up that we are about to enter a short-term trading-range or declining market environment, so short-term traders should pay attention should we see a rising market early next week.

Bottom Line:  This model isn't telling us much currently, but a rising market early next week would likely push the model to overbought fairly quickly. 

____________________________________

Indicator:  COMPOSITE MODEL

Status:  NEUTRAL

Comment:  Even with positive readings from longer-term indicators such as the sentiment surveys and Commitments of Traders data, this model declined sharply last week.  The main culprit for the decline in the model was overbought breadth readings and an unwinding of market uncertainty.  While the latter may likely continue this coming week, the former will probably not be a factor unless we get a large ramp in equities.  The fact that breadth will have a difficult time becoming more overbought than it already is will prove to have a dampening effect on the model, and will make it almost impossible for it to become overbought this week, especially considering the longer-term positives. 

Bottom Line:  This model has quite a ways to go before being considered overbought, and it will have an exceedingly difficult time becoming so barring a large rise in equities.

____________________________________

Indicator:  STEM MODEL

Status:  NEUTRAL

Comment:  We're still neutral here on both a relative and absolute basis.  Until we see more sustained optimism in traditional readings such as the put/call ratios, VIX and TRIN, the model will flop around not far from where it currently is.  Two weeks ago, I said that from looking at the model, it should be easier for the market to make significant headway than it would be to suffer a meaningful decline.  That situation has reversed now, as the model will become overbought much quicker with a rise in equities than it would become oversold with a selloff.

Bottom Line:  Although the model remains neutral, its current positioning suggests that a meaningful rise in equities should be met fairly quickly with an overbought reading in the model.  Conversely, a market decline would not be met with oversold readings for quite some time.  This suggests that the market has limited prospects on the upside compared to downside.

This week has seen an incredible contraction in volatility.  I compute a simple indicator which scores days based on their volatility relative to other days over the past week.  As of Friday, the indicator reached, by far, the lowest score it has given in three years.  Volatility tends to be mean-reverting, meaning that periods of low volatility often give way to periods of high volatility, and vice versa.  We've seen that play out in spades over the past couple of weeks, as the large moves off the low lead to the tight ranges of the past week.  While we may see a few more days of this type of choppy trading, it should lead to a large move sometime soon, most likely sometime this coming week. 

Like all of our models and most of our indicators, I am neutral at this point.  Previously, I said I was bullish as long as we were above 830 and not overbought.  Then we became overbought and I said last week that my preference was to the short side.  Now that the short-term overbought condition has dissipated, I don't see a high probability of a move in either direction.  I think we will see a large point move sometime in the next week, but unfortunately I cannot say with any degree of confidence whether it will be to the upside or down.  I do think that the most likely scenario at this point is an environment something like last October, where we had spikes higher and spikes lower, but remained relatively unchanged after several weeks.  In that type of trading range, time frames should be kept short (under 5 days), and a focus on oscillator-type indicators (e.g. stochastics, RSI, etc.) often work better than breakout or breakdown strategies.  It's very difficult to be aggressive in anything other than intraday trades in our current environment, as the market is tied to war developments to the exclusion of pretty much everything else.  The should dissipate somewhat as the war develops, and we will see more normal trading, other than after very major developments.

 - Jason Goepfert

Disclosure:  no positions

This disclosure is not intended as trading advice in any form.  It is meant as a note to subscribers that the author may have a position directly affected by the market outlook reflected in the commentary.  Although the author takes great pains to remain objective in any commentaries, it is only fair that readers should know that the author may have taken positions in accordance with his market outlook.  Positions can and do change at any time, without notice to the reader.

© 2003 Sundial Capital Research, Inc.  All Rights Reserved.