http://www.sentimentrader.com

Sunday, March 16, 2003

I ended the last weekly commentary two weeks ago stating that I didn't seen any signs that would indicate an imminent rebound, and in fact many of our shorter-term measures had relieved their oversold condition, so my preference was to short any additional upside with the understanding that trades should be kept short-term with tight reigns.  The S&P promptly sold off 50 points, only to recover most of it in the past three days.  Unfortunately, we are now at a point where both sides appear equally compelling, leaving us without a solid edge once again.

____________________________________

Indicator:  NYSE MEMBERS REPORT

Status:  BULLISH

Comment:  Not much to report here, as the Specialist Short Ratio inched up just a bit to 34%.  This is still rather low on a historical basis, as it is more than a standard deviation from its bear market mean.  The four-week moving average of the ratio remains at 33%, which is the lowest since the Spring of 1997.  This low level of specialist shorting (and by extension, high amount of public shorting) is what helped to launch the last legs of the bull market.  However, at that time we had seen a very extended bout of shorting pressure from the public, as the 52-week moving average was around 35% - a level that was seen at several of the major lows in the past 20 years.  Currently, that long-term moving average isn't so extreme, as it is sitting around 38%.  From a different historical perspective, the de-trended ratio is also more than a standard deviation from its long-term mean.  As is discussed in the explanation of the de-trended indicator on the site, this is simply a way to look at the data without the misleading effects of the secular rise in public shorting over time.  The fact that we are seeing the public short at a clip that is somewhat extreme on both a recent and long-term historical basis lends more credence to the bullish nature of the data.  At some point, many of these shorts will decide to cover, creating buying pressure in the underlying names.

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

Comment:  Once again this week, there was not much change in the futures positions in the large S&P 500 futures contract.  Both commercial traders and small speculators became a bit more net long, but the changes were minimal.  Overall, their current positions (or rather, the momentum of their positions) are moderately positive, as the Futures Balance Matrix is elevated on both a one-year and three-month time frame.  The Futures Balance Matrix gives us a way to see in one chart how aggressive the commercial traders have been versus the small speculators in accumulating long positions.  While neither time frame is extreme, they are both on the upper end of neutral, which is a slight positive for the market.

In the e-mini contract, there were relatively large changes in the commercial and small spec positions, and both of them were negative for the market.  Normally, during a down week such as we saw during the reporting period, what we would want to see is commercials adding to long positions and small speculators adding to shorts.  The opposite happened this week, as the commercials liquidated longs and added to shorts, for a reduction to their net long position of over 71,500 contracts.  Conversely, the small specs initiated a large number of longs and only a few shorts, for a reduction in their net short position of just over 33,500 contracts.  While this is not a trend that is friendly to equities, overall the positions for each group remain somewhat constructive, as the commercials are still heavily net long and the small specs remain net short.  As always when I mention the COT information for the e-mini, you should know that the forecasting record for this contract has been suspect until only recently (the last six months or so), and even then it has been only moderately successful (at least in every way I have analyzed it), so that should factor into any decisions where this data is taken into account.

Bottom Line:  Overall, I would view this data as more bullish than neutral, but only slightly.  The momentum of the positions in the large contract is moderately positive, as are the net positions in the e-mini.  Nothing to get excited about, but it could lend some support to a rally.   

____________________________________

Indicator:  SENTIMENT SURVEYS

Status:  BULLISH

Comment:  Continuing the theme of the previous indicators, there was not much change here this week either.  I presented a table in the March 2nd weekly commentary which looked at Chartcraft's Investor's Intelligence survey results at each major low since 1996.  As of two weeks ago, the survey was showing more bullishness than at almost any other major low, and certainly more bullishness than the average seen at the lows.  Even though overall bullishness in the survey has ticked down each of the last two weeks, we are still seeing more bulls and fewer bears than any other comparable low point.  While it would be ridiculous to suggest that the survey HAS to drop some bullishness before the market could find a low in here, it would be comforting to see.  Two weeks ago, I gave a potential reason why the survey is showing so much residual bullishness, so I suggest you check the archives for that commentary if you're interested.

The other surveys didn't show an appreciable change from last week, so we continue to be in a similar situation as we have had for the past few weeks.  While overall the level of bullishness in the surveys is low enough that a good low is certainly possible in here, we have not seen a confluence of true extremes that would suggest it may be imminent.

Bottom Line:  I could repeat verbatim what I said two weeks ago, so I'm going to - overall, the sentiment surveys are showing a level of bearishness which could support a rally from these levels.  However, only the AAII survey is truly extreme, and the others have shown only a minimal amount of pessimism.  The II survey, which is certainly the most widely followed, has been the most reluctant to become bearish and has not yet reached a point that has appeared at almost every other major low.

____________________________________

Indicator:  AIM MODEL

Status:  BULLISH

Comment:  Obviously, since this model is based solely on the momentum in the sentiment surveys, and the surveys haven't really changed for a couple of weeks, this model remains pretty much where it was.  The Consensus bullish percentage was not available by the time I wrote this commentary, so the model cannot be updated yet, but unless there was a dramatic change in the Consensus survey (highly unlikely), then the model will not change much at all this week.  While I would prefer to see one final spike higher in the model (suggesting a sudden collapse in bullish opinion), it certainly is not necessary to consider this model positive. 

Bottom Line:  We have reached a peak here that is comparable to other intermediate-term lows.  There has been enough negative momentum in the surveys to suggest that it has reached an extreme, and that has been a positive influence on the market over the coming weeks/months with few exceptions.

____________________________________

Indicator:  SEASONALITY

Status:  NEUTRAL

Comment:  The end of this coming week will bring with it triple-witching, or the simultaneous expiration of options and futures contracts.  For such a mystical name and all the hype surrounding such events, I find that the evidence behind the headlines is not nearly as exciting.  Over the past 8 years, the week leading up to expiration (meaning this coming week), has generally showed a positive overall bias, with strength more likely to be seen earlier in the week.  Expiration Fridays have had a consistent negative bias.  This pattern has remained relatively consistent over the past year, as Monday and Tuesday prior to expiration are the only two days with a positive expectancy.  Recall that expectancy is calculated as follows:

Expectancy = (% positive * avg gain) - (% negative * avg loss)

This gives us one figure which measures not only whether a given day has historically been positive, but also HOW MUCH it has been positive.  The higher the expectancy, the more positive the bias.  Over the past year, Monday has had the highest expectancy, while Wednesday and Friday have had the worst.  8 of the last 12 expiration Mondays have closed higher on the day (and 5 of the last 6).  Only 3 of the last 12 Wednesdays have closed higher (and only 1 of the last 6).  While 12 samples is too small to generate anything that could be considered significant, I think it's an interesting factoid and something we may want to file away for the coming week.

Going back to the "non-mystical" nature of expiration days, I have actually found these weeks to be no more volatile than any other time.  To test that, I checked each expiration over the past three years to see if they did have some type of impact on volatility.  I tested all expirations, as well as just triple-witchings, and didn't see a discernable difference between the two.  I looked at volume on each of the days leading up to expiration compared to its 10-day moving average, and did the same for the average daily range in the S&P 500.  Interestingly, even though Monday has the most positive bias (discussed above), it had the lowest average volume, by far.  Only 26% of the Mondays prior to expiration had volume greater than its 10-day moving average, and 46% of those Mondays showed a range that was greater than average.  The least volatile day was Thursday, which showed a range greater than average only 34% of the time.  Wednesdays were the most volatile days, as they showed both an average volume and average range that were greater than average.  The "volatile" (according to the media) expiration Fridays showed an average range that was only 93% of average, with 60% of the days showing a range less than average.  Volume was just a bit heavier than average.  Expiration Fridays occupied 4 of the top 10 "least volatile days", more than twice any other day.  Once again, the media, in an attempt to find something sensational to say, will attribute any volatility this week to the looming expiration.  Please ignore it, as it is no more than hogwash.

Bottom Line:  This week will have a very slight positive bias to it, particularly early, while the bias turns negative later in the week.  Historically, we might expect a bit more volatility on Wednesday, but overall we shouldn't expect anything out of the ordinary volatility-wise.

____________________________________

Indicator:  BREADTH RATIOS

Status:  NEUTRAL

Comment:  The 10-day advance/decline line, my preferred measurement for short- to intermediate-term breadth, has been spending the past couple of weeks working off a mild overbought reading reached at the end of February.  Currently it is neutral, along with the longer-term 21-day moving average which is in the exact middle of its range.  The coming week will bring with it alternating days of positive and negative numbers dropping from the averages, so it'll be fairly difficult for us to reach one extreme or the other here unless we have a very dramatic move in the market.

The 10-day up volume indicator is in a similar position, although it is a bit more oversold.  This is mainly due to the skewed downside volume I talked about earlier in the week which caused the abnormally high TRIN reading.  Once again, the 21-day average is in the exact middle of its range.

We haven't seen much progress on the new high / new low ratio lately, as the greatest number of new lows I recorded this week was 336.  I would much prefer to see something closer to 600 at a minimum for us to be able to begin making comparisons to recent intermediate-term lows.  I've been asked if perhaps we are seeing some sort of positive divergence, where the market is holding firm and so are the number of new lows.  In order for me to see a divergence, I would like either price to make a lower low while the number of new lows stays relatively muted, or price hold above the October lows while the number of new lows explodes higher.  Neither one has happened yet, so I'm not sure where a divergence would come in.

For only the third time since July 2002, we saw a one-day Down Pressure reading of 0% in the Nasdaq 100 on Thursday, which can only happen when all 100 stocks finish higher on the day.  The two other instances were July 5th and January 2nd.  This has pushed the 3-day average (which is what is posted to the site) down to an overbought 27%, a reading last seen on February 18th.  The reading for the S&P 500 on Thursday was 12% which helped to push the 3-day average down to 34%, which is modestly overbought.  Since the indicators on the site use a 3-day moving average of the daily readings, and since Friday gave us neutral readings, these indicators likely will not become significantly more overbought early this coming week unless we have another day or two like Thursday.

Our cumulative TICK readings are neutral across the board, after the intraday indicators worked off their overbought condition by Friday afternoon.

The 10-day TRIN on the NYSE is still showing an extremely oversold reading over 1.70.  However, if we (just for fun) adjust Monday's extraordinarily high reading over 5.0 and flip FNM and CE over to positive that day, the TRIN just about gets cut in half and our 10-day reading would be under 1.50.  That's still somewhat oversold, but it puts into perspective just how skewed the indicator is because of those two securities.  Even so, I'm not particularly impressed with oversold TRIN readings while in the context of a downtrend, as it has a bad track record so far in this bear market.  The table below shows how 10-day TRIN readings over 1.50 faired at calling the major lows:

We can see that only once did the TRIN peak the exact day the market formed a low.  The average delay between the peak in the TRIN and the market low was over 11 days.  What this shows us is that there is typically heavy, sustained selling pressure, often concentrated in a few issues, leading up to major low points.  When the selling becomes indiscriminate and broad-based, then the TRIN actually flattens out a bit, which is why it is unusual to see exceptionally high 10-day TRIN readings at major low points.  This suggests that we may currently be seeing only the initial stages of the formation of a major low.

Bottom Line:  We don't have any real extremes here in either oversold or overbought territory, other than the NYSE TRIN (and even that may not be telling us much).  We never saw the breadth extremes seen at most intermediate-term lows (refer to Wednesday's daily commentary for a comparison table), so these indicators are giving us no reason to believe that the recent rally was anything other than an oversold bounce.

____________________________________

Indicator:  RYDEX RATIOS

Status:  NEUTRAL

Comment:  Three out of the four short stock index funds hit one-year record inflows in the few days prior to Wednesday.  This show of pessimism pushed our RSI spread into bullish territory on Tuesday and was a good heads-up that these timers had become too bearish - a stance that was unlikely to be rewarded.  Not surprisingly, they have had a tremendous change of heart in the past two days.

From Wednesday to Friday, these traders shifted more than $425 million out of the short-oriented funds and into the long-oriented ones.  This two-day shift is nearly 10% of total assets in the stock index funds and money market combined.  Looking out over the past three years, this is one of the largest shifts of funds that we have seen, both on an absolute and a relative basis.  When we have seen a shift of greater than 9%, the S&P 500 was higher the next day only 23% of the time, with an average return of negative 0.33%.  Three days later, it was higher 46% of the time, with a return of negative 0.22%.  The NDX was higher the next day only 31% of the time, with a return of negative 1.0% and three days later it was higher 46% of the time, with a return of negative 1.5%.  If we take out the instances where we were recovering from a panic low, the figures above become considerably more negative.

This sudden shift has pushed the RSI spread between the bullish and bearish funds to +51, more than a standard deviation from the bear market mean.  While I would prefer to see something closer to +80 before becoming too aggressive on the short side, even peaks in the indicator at this level have portended negative things for the market on a very consistent basis over the past couple of years.  This indicator is extremely effective at identifying contra-trend extremes (i.e. overbought in a downtrend or oversold in an uptrend), and I think it is important to watch the extremes here.  The fact that we are already overbought suggests that short-term upside will be limited and it makes the most sense for short-term traders to sell rather than buy if higher prices are seen over the next day or two.

Something interesting also happened in the Bond complex.  Apparently, the fact that bonds got whacked on Thursday was the signal so many were looking for that bonds are about to collapse.  I've heard explanations for why bonds can't go any higher for months on end now, yet they keep setting new highs.  Such negative sentiment is one reason they have kept chugging higher.  Obviously, they can't rise forever since rates will not drop below zero, but they have already gone considerably higher than many thought probable.  It's quite amazing how negative bond sentiment is compared to its price performance.  If the S&P was hitting new all-time highs nearly every day, yet traders were doubting it would continue any longer and sentiment was as bearish as it is on bonds, how many people do you know who would be salivating at adding to their long positions on any pullback?  I'm guessing quite a few. 

In any event, Friday's slight recovery in bonds was the opportunity many were looking for to establish short positions.  The money flow out of the long Bond fund and into the bearish Juno fund was the greatest one-day move in at least three years.  In fact, it was 50% greater than the next biggest move.  Going over the other large one-day asset shifts, a great many of them came near lows in bonds.  And if you've checked out the primer I've posted to the site about why it's good for stock traders to pay attention to bonds, then you'll know that lows in bonds have often coincided closely with highs in equities over the past few years.  As I say in the primer, at some point the bond/stock relationship will revert back to a positive correlation, but for now I don't think it makes sense to think that trend has changed.  This spurt of bearish bets on bonds is a positive for bonds and a negative for stocks. 

Bottom Line:  Historical precedent suggests that any additional upside from this point forward would likely serve as a good shorting opportunity for short-term traders.  The market timers who use the Rydex funds have committed heavily to the most recent rally, and these types of long-side commitments in the context of a downtrend are almost never rewarded handsomely.  It has paid over and over again to take the other side of these trades, and likely will going forward as well unless we have seen a major turn in the trend.  Our longer-term measures in this fund family remain neutral to slightly oversold.

____________________________________

Indicator:  PUT/CALL RATIOS

Status:  NEUTRAL

Comment:  There isn't much to say here, as both the equity and OEX put/call ratios remain relatively muted.  We have seen a slight drift down in the OEX put/call ratio over the past few days, which is bullish, at the same time the equity put/call ratio has rose slightly, which is also bullish.  This has caused the spread between the 10-day averages of the two ratios to come in significantly, to a level that has coincided with other good lows. 

After being relatively bullish up through Wednesday, our SPX put/call bid/ask bias ratio has dropped below 1.0, to currently read .65.  This is not positive in the short-term as it shows institutions have been selling calls and/or buying puts on the S&P 500 during the late-week rally.  This low of a level has often portended a noticeable drop in the market sometime in the following two-three days.

Bottom Line:  We have slightly conflicting ratios here, although they work on different time frames.  The bias ratio is suggesting strength should be sold or shorted early in the week, while the OEX-equity p/c ratio spread is suggesting that savvy OEX traders are more optimistic than the wrong-way equity options traders, and therefore any weakness seen should be bought. 

____________________________________

Indicator:  VOLATILITY MEASUREMENTS

Status:  NEUTRAL

Comment:  The combination of envelopes and RSI on the VIX which is outlined on the site did a masterful job at identifying a low-risk buying opportunity on Wednesday, and I'm still kicking myself for not giving it more weight.  This combination, while not suitable for trading on a mechanical basis, often serves as a very good tipoff that a trend may be about to change, and it suggested Wednesday that the VIX was overbought and thus equities were oversold.  I did mention this in Wednesdays' intraday note, so I hope at least a few of you were more aware than I.  The late-week rally has served to alleviate that overbought condition, and we are currently neutral here.

The VIX Fear Premium remains elevated, and I suggest you read Wednesday's daily comment for an overview of just how positive this indicator appears to be for the intermediate-term, as well as a brief overview of what happened during Desert Storm.

My main issue with the decline that we've seen from the January high is that it has been accompanied by pathetic volume and volatility.  While much of that can probably be explained away by geopolitical uncertainties, the fact remains that weak holders of stock are normally shaken out when volatility becomes extreme.  Because we have not experienced any great volatility during the past few weeks (other than isolated occasions), we probably still have a greater-than-average number of hangers-on that will panic if/when volatility picks up.  This will exacerbate a decline if one should occur. 

On a positive note, volume picked up this week, and that is a trend that must continue in order to have more confidence in rising prices.

Bottom Line:  We're seeing the very beginnings of what could be a positive development, in that heightened volume and volatility appear to be making a comeback.  These components exist at nearly every intermediate-term low, especially during a bear market, and their absence is a major negative.

____________________________________

Indicator:  STEM.MR MODEL

Status:  NEUTRAL

Comment:  As I outlined early last week, our shortest-term model reached an oversold level that had only been seen a few other times in the past few years, each of which lead to at least a temporary halt in the selling pressure, and all but one preceded a major low.  That was our signal that additional selling pressure would likely lack the momentum needed to continue the trend, and I should have been more open to the possibility that we would see a reversal.  By early Friday morning, the model had swung to the other direction - although not quite overbought, it was at a point that had coincided with levels last seen near intraday market peaks over the past month.  Friday's choppy action was enough to work off some of that slightly overbought condition, but if recent action is a guide, then we may see a little more weakness in the market in order to begin the swing back to oversold.

Bottom Line:  Currently, the model isn't telling us much.  However, if we have another spike higher accompanied by a spike into overbought on the model, that should provide short-term traders with a decent short entry.  Conversely, a model move back to oversold may give a good long entry provided we hold above Wednesday's lows. 

____________________________________

Indicator:  COMPOSITE MODEL

Status:  NEUTRAL

Comment:  It is virtually impossible for this model to reach any kind of extreme as long as breadth remains neutral and our volatility measurements do not see a sustained degree of "stretch".  The model has risen quite a bit in the past week and a half, mainly due to the very high 10-day TRIN and the little spike in the VIX.  If we see continued volatility and a move towards an oversold extreme in some of our breadth measurements, the model has a good chance at reaching a true extreme like July and October. 

Bottom Line:  Unless and until we see more volume and volatility come into the market, accompanied by some oversold breadth readings, this model will be stuck in neutral.

____________________________________

Indicator:  STEM MODEL

Status:  NEUTRAL

Comment:  Two weeks ago, I said that the market was recovering from an oversold reading, so when the model reached its overbought band, we should assume that we are just seeing another oversold bounce and it should be treated as a selling opportunity.  The model did touch its lower (overbought) band on March 3rd, and that proved to be a very good selling opportunity for short-term traders.  Currently, the model is once again approaching its lower band but it will be more difficult this time since the bands are curling down.  We'll need to see more extremes - and more sustained buying pressure - in order to become overbought here.  Conversely, it would take a bit less selling pressure in order to turn around and become oversold.  As of now, however, we're neutral here with no clear direction going forward.

Bottom Line:  It will be relatively easier for this model to become oversold than overbought, suggesting that it may be easier for the market to make significant headway than it would be for the market to suffer a meaningful decline.  That conclusion is fairly tenuous, however, so I wouldn't read a whole lot into it.

I've been using the 830 level on the S&P as my "line in the sand", suggesting that there was at least the possibility we had seen an intermediate-term low on Wednesday as long as we stayed above this level.  If we drop below, then all bets are off and I wouldn't become bullish again unless we either rallied back above it, or reached a true pessimistic sentiment extreme.  We waffled above and below that level on Friday, proving once again how important it is.  I'm not suggesting that I buy every time we go above and short every time we drop below, I am only using it as a frame of reference.  I don't do this often, and I don't like it, but I feel these levels are so widely watched that they become pivotal.  I see no reason at this point to change my outlook, and I would continue to concentrate on long positions as long as we remain above 830 and until we become overbought.  If we lose 830 and continue down, then my preference switches to short positions unless and until we become oversold.  Sentiment-wise, we have a confluence of neutral readings, especially shorter-term, which doesn't give us a lot to work with.  Longer-term, we have the positive influence of the negative momentum in the surveys and heavy public shorting pressure, but that's about it.  I will continue to stress that we have not seen the type of sentiment normally seen at an intermediate-term low, but we have seen enough that we have to at least consider the possibility that a low was reached on Wednesday.  That's why, even though I don't like to do it, I have determined a line in the sand, and will use that as my reference for now. 

 - 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.