May 19, 2002

sentimenTrader.com

Sunday, May 19, 2002

I said last week that sentiment was in a place where a decent rally could form, and I guess I would have to say that's what we got this week. It depends on what you mean by "decent", but a 5% S&P rally and an 11% NDX move counts in my book. Unfortunately, the only BUY signal we had was a STEM.MR indication issued last Friday that went into Monday. Maybe that says something about the potential for this move.

I want to take a quick second to talk about what our signals mean. By the nature of the models, they almost always go against the current trend. That's the whole point. They are designed to give us a clue as to when sentiment becomes overheated on three different time frames. I would never recommend someone take a trade against the predominant trend in whichever time frame you're trading. If contra-trend moves are part of your methodology, then fine, but if they are not, then don't force it. I neither expect nor recommend anyone take a trade based on these model signals alone, as they usually fight the current trend. They are meant to be nothing more than an indicator or a heads-up that a given trend may soon exhaust itself based on probabilities derived from past occurrences of similar conditions. I use "BUY" and "SELL" nomenclature to make it absolutely clear where the model stands, not to imply that any trades are being taken based on this information alone.

Let's look at our indicators one by one...

BULLISH

SPECIALIST SHORT RATIO - At 41%, this indicator can still be considered bullish. However, the last reading available is for the period ended Friday, May 3rd (there is a large lag between when the data is gathered and when it is released). The next two readings could and should stay low, but it will be interesting to see the figures a few weeks from now on this current week with the rally we've had. One of the research items on my list is to try to see what effect market movement has on this ratio. For example, if the market drops by 5% in a given week, then with what confidence can we say this ratio will drop as well? And by how much? That may give us a clue as to future readings of this ratio before the numbers are actually released.

BEARISH

STEM - With both the fast and slow line hovering in the .30 area, this model is suggesting sentiment is intermediate-term overbought. Although there's lots of room for this model to drop further and become more overbought, the current reading is consistent with recent market tops. The reason Friday's signal was given a one-star rating was the relatively weak reading (just under the SELL threshold) and the longer-term sentiment picture. Now that we have the new sentiment survey and COT numbers, that longer-term picture is not looking quite so oversold and I will be much quicker in upgrading the signal than I probably otherwise would have been. However, unless we get another day or two of upside, I most likely will not get a chance to upgrade it. The slow line will be dropping some extremely low numbers on Monday, which will force a rise in the signal line.

STEM.MR - The BUY signal last Friday quickly evaporated with Monday's jump and for the rest of the week this model popped into and out of SELL territory. It was able to indicate a couple of minor (very minor) contra-trend reversals, but again it was fighting the upward movement like it always will. We left off Friday with another relatively minor SELL signal, but I believe if we get any more upside, we may see the first three star-rated signal from this model for quite some time. All components (except the put/call ratio) are hovering in the lower portion of their ranges, particularly the VIX.

VIX and VXN - Please go to the Research area to see a current study of what the VIX is telling us. The VIX is currently 12.2% below its 10-day moving average. In the past, this has lead to significant short-term (<10 days) underperformance in the broader market. You can see from the study that of the 16 times the VIX was around this low level (on a relative, not absolute basis), 14 of those times the S&P 500 was lower 3 days and 10 days later. Across all time periods studied, the S&P was lower a majority of the time. If one would take out the times where the VIX continued lower after such a low reading, the results would become much more negative. This tells us that if the VIX does NOT continue lower, but instead reverses up from here, probability would suggest that the S&P may be 40 - 50 points lower several days from now. Of course, this is all academic and doesn't mean squat in the practical world of trading, but I believe it really does pay to be aware of these tendencies. It suggest that there is complacency and that leads to a period of market underperformance. On a related note, the VXN has dropped 15% during the 11% rise in the NDX, which also suggests that options traders are embracing this rally. There is not enough data for this indicator to draw meaningful conclusions, but it is still not good sign.

COT - Back to bearish here, as the commercials and small specs both made an about-face and returned to their longer-term trend of being bearish and bullish, respectively. The commercials decreased their long positions by over 4,000 contracts while at the same time they increased their short positions by nearly 2,000 contracts (for a net increase in the net short position of 6,000 contracts). The small specs covered 1,000 short contracts while opening 8,400 long contracts (for a net increase in the net long position of 9,400 contracts). As you can see from the stochastic charts, that's not a positive trend in the least. During the survey period the S&P rose about 4%, and it has continued to rise about another 1% since the survey period ended, so one would logically conclude that the numbers have not changed much since then, except to become more bearish to us. Although it may be difficult to see on the Futures Balance Matrix chart, the one-year reading is now 18% commercials and 82% small spec (meaning the small specs are accounting for 82% of long commitment). This is a decline from 29% commercials and 71% small specs last week. Drops such as this are typically the start of a sustained trend and not a one-reading wonder, so looking at the history of these two groups, it appears as though the commercials are now on their way to building up their short positions once again while the smalls continue to get longer. That is most definitely a negative for longer-term market prospects.

NEUTRAL

NEWS ITEMS - Once again, a fairly neutral news environment, as economic numbers were greeted with "logical" market reaction. Company-specific news was generally positive (except of course for the now-typical bankruptcy hunt). One question mark is the DELL-inspired futures rally that pretty much petered out. If the market was looking for an excuse to rally and found one but couldn't hold it, that would be a negative. We'll have to withhold judgment on that one for now. One rather amusing anecdotal aside is an article in Friday's Wall Street Journal about fund families taking steps to no longer release the names of their portfolio managers. It wasn't too long ago that every time you turned on TV, a portfolio manager was angling to get their mug on the tube and bring in assets, but have they now become so reviled that they want to remain anonymous? That's gotta be a positive.

SENTIMENT SURVEYS - The bullish attitude resumed in these surveys for the first time really since the March peak (although there was a small blip after the mid-April rally). On the whole, I'm not suggesting the surveys are showing too much bullishness since overall they are still quite bearish, but the trend towards increased bearish has paused for now. Interestingly, all the surveys showed increased bullishness except for the II survey, which went very slightly in the other direction. There has been a very unusual disconnect between that survey and the others, which I touched on a couple of weeks ago. I said then that there were three times in the past decade where there was such a divergence, and the only solid conclusion we could draw was that each occurrence preceded extreme volatility. So far that trend continues. The increase in bullishness in the surveys was enough to push the AIM model squarely into the neutral camp...

AIM - We never got a BUY indication from this model, as it reversed from just under that threshold in a big way. We are now back to a reading of 57 on this model from 77.3 last week. I showed the results last week of looking at past occurrences when the model was around 77, and the returns were encouraging, especially 5 weeks out and later. But if bullish sentiment returns too quickly, it could cancel out those indications.

CBOE RATIOS - Fairly neutral here as well, although with a slight bullish bias. We have not seen very low put/call readings during this recent rally, which is a positive, as it suggests that traders are not scrambling to buy calls to take advantage of further upside. That conflicts with the VIX, which is giving significantly bearish readings. I've said several times that during periods where there is a divergence between the two indicators, I like to refer to the TOPM model (see the April 28th Weekly Commentary in the Archives for a description of this model). And this model is now...perfectly neutral. It's not telling us anything at this juncture, as it is squarely in the middle of its range. The 10- and 21-Day moving averages of the put/call ratio (now posted daily in the CBOE Ratios group in the Indicators section of the site) are still quite high, corresponding to past intermediate-term market bottoms. Over the next 11 days, the 21-day moving average will be dropping 8 readings over .77. So should we get a drop in the put/call ratio going forward, this average could drop significantly. The past two times this average has been this high was in late September after 9/11 and early to mid February, at that short-term bottom. Both times, the 10- and 21-day moving averages reached readings just higher than we have now, but the VIX was also higher then. In September, it reached 20% over the 10-day average and in February it was 14% above its average. Although we had a quick spike in late April, now the VIX is 12% BELOW its average, as discussed above. All in all, it's a mixed picture and I don't think we can come to a conclusion one way or the other.

TOPM - We got some low readings in this model earlier in the Month during the most recent bottom, but it has stayed neutral since then. One would expect higher readings if enthusiasm for calls was overdone, but we don't have that. This could almost be considered bullish, but it is very marginal. If this index drops at all during a market rise next week, that would be considered very bulllish and I'll be sure to mention it in the daily commentary.

TRIN - The extremely oversold conditions I pointed out last week have now dissipated. We recorded several low readings this past week, which when taken in conjunction with the fact that we will drop four readings over 1.2 (and two readings over 2.0) in the next five days, suggests that the 10-day moving average could drop severely in the coming week, provided we don't have a market selloff. It is very possible that this could be going into the bearish column next week, however at this time it is still neutral.

BREADTH RATIOS - We are on the high (overbought) end of neutral here, but neutral nonetheless. The rally last week gave us some quite high breadth readings, but nothing extraordinary. Over the next five days, we will be dropping the following numbers from the 10-day moving average of the advance/decline line: -1110, -377, 816, -978 and -959. If we get mild numbers next week, this indicator could become extremely overbought. Over the past couple of years, the average daily a/d reading is 63 (63 more issues advanced than declined on a daily basis on average). So looking at a totally hypothetical situation, if we have an "average" breadth week next week, that would take the advance/decline average to a level that would place it in the top 30% of all readings over the past few years. And it wouldn't take very high positive readings to put it into the top 10% - very overbought. We'll have to watch this closely.

COMPOSITE MODEL - Our newest model, which you can now see daily in the Models section, came very close to giving a BUY indication at the recent bottom. However, there just was not enough of a confluence of bearishness to push it into BUY territory. We have now dropped significantly from that precipice and are neutral. It would take quite an effort to get any kind of BUY indication here, as the sentiment surveys and COT figures will have a negative impact on the model next week. That coupled with the breadth, TRIN and VIX situations make it seem almost impossible that we will see any indications from this model next week. In fact, it's very conceivable that it will tip towards overbought instead of oversold early next week. I'm very excited about the release of this new model and I think it will help us tremendously in the future.

Like most of the indicators, I am squarely neutral from a sentiment perspective this week. I can just as easily see us going up to challenge the recent highs (well, that would be a stretch) as I can see us going down to challenge the lows. However, as I have been saying, I have no idea how far this recent rally can take us, but I do believe it will ultimately fail. We did not see the confluence of bearish indicators and model signals that we almost always see at solid intermediate-term bottoms. Of course, the key in that sentence is "almost always". There is nothing written that says that the market HAS to drop just because we didn't get our BUY signals. However, if we want to look objectively at the odds of what could happen, probability suggests that this is not a sustainable bottom.

See you next week!

- Jason Goepfert