Downside Should be Minimal

Bottom Line: Most of our short-term measures are now very oversold, equating to times that have historically meant at least a short-term low was at hand.

As of today’s close, we’re getting some incredibly overdone readings from our shortest-term composite measures, including our indicator “score” and the STEM.MR model. Considering that the S&P 500 is only 40 points and a few days from recovery highs, these readings are remarkable.

The indicator score is computed by observing how stretched a series of 14 short-term indicators are from their typical values. The maximum score is 56, and as of today’s close we reached 43, or 77% of maximum. This is a rare occurrence, as the score has approached these heights only a few times since 2000. We hit 75% on 4/14/00, after which we saw an immediate rally. We hit 75% again on 10/12/00 which also marked the low before a nice rally over the next month. It was nearly another year before we saw such a confluence again, this time caused by the terrorist attacks of 9/11, and we know what happened after that. On 02/07/02, the score hit 73% and once again it marked an intermediate-term low.

The next (and last) two signals weren’t nearly as beneficial, as the score hit 75% on 6/11/02 and 77% on 9/19/02. The market continued to decline for another couple of days after those readings were achieved before seeing stiff rallies that took the market back to around the area it was at when the readings were given. Both rallies failed, of course, and eventually took us to lower lows.

The current reading of 77% is eclipsed only by those readings seen immediately after 9/11. Since the March 2003 bottom, the highest score we had seen was 68% near the low on August 5^th. Every time the score hit even 55%, it proved to be an impeccable buying opportunity. There are two ways to look at the current score – either it’s the mother of all short-term buying opportunities, or it’s a troubling sign that we are seeing extraordinary pressure that won’t be alleviated over just a few days. My opinion is a mix of the two – I think this type of reading shifts the risk/reward firmly to the long side over the next few days, but what happens after that is questionable. The fact that we’ve seen such heavy pressure so soon after new highs sparks some concern in my eyes, and how the market rallies after this will be especially telling.

I mentioned the STEM.MR model, since that too is notable. From the description on the site, you can see that the model is computed by observing how stretched each of several short-term indicators is from their 10-day moving averages, and those deviations are then ranked against others over the past six months and added together to come up with the final model reading. As of today’s close, the model stood at 71%. This is worth mentioning because it’s the highest reading in the model’s history. I changed the way the model was computed last year, but reconstructed the readings going back to July 2001 so we have enough history to encompass 9/11, and the July and October 2002 market lows.

At the height of the 9/11 panic, the STEM.MR model reached a peak of 65%, at the July 2002 low a peak of 67%, at the October 2002 low a peak of 64%, and at the March 2003 low a peak of 63%. Other than those readings, the model approached the current level a few times over the past couple of months (as you can see from the chart on the site), and it also reached 69% on 01/27/03. That last one, in January 2003, is the only one that didn’t result in a tradable low. It did halt the decline in its tracks, at least for a few days, but after some choppy action over the next week wore off the extreme model reading, the decline into February and March continued.

Looking at two measures – the indicator “score” and STEM.MR model - that encompass several of our shortest-term indicators, the prospect of much further downside over the coming days appears limited, while there is a good chance of decent upside potential. Obviously, these readings don’t mean the market HAS to do anything, but in weighing probabilities and risk/reward, it appears as though the bears will be risking much to make potentially little over the coming days.

Bottom Line: Recent TRIN readings, while possibly deceptive, suggest the narrow selling pressure may be overdone.

One of my market-related pet peeves is when traders or reporters talk about a high TRIN reading as being a measure of broad-based selling pressure (click here to see how it is calculated). I think it is important to know that it is most certainly not a good measure of broad selling pressure, but it is a decent measure of the relative relationship between how many stocks are up or down on the day and the volume flowing into each. Today is an excellent example of that. On the Nasdaq, I show that there were 716 advancing issues compared to 2449 declining issues – certainly a display of broad-based selling pressure - yet the TRIN for that index closed at 0.78. If you gave traders that TRIN figure alone and asked them if the Nasdaq had closed up or down on the day, probably 99% of them would say “up” – I suspect not even 1 in 1000 would suggest the Nasdaq was down 1.5% with a reading like that. The reason it was so low is because there was relatively less volume flowing into issues down on the day than there was flowing into issues up on the day. It has NOTHING to do with heavy, broad-based selling pressure.

I find the TRIN useful not for spotting those times of broad-based rallies or declines, but quite the opposite. Extreme TRIN readings are made only when there is an unusual relationship between breadth and volume. If we have an extremely negative day, say with 3 declining stocks for every advancing one, and 3 times as much down volume as up volume, then the TRIN is still only going to be 1.0. Only when volume doesn’t match breadth will there be a high or low reading. The situation over the past couple of days has been unusual in that we have seen the TRIN on the NYSE reach at least 3.0 on two consecutive days. This very rarely happens, except in the first few minutes of trading which should be ignored. Looking at the past few years, since decimalization was introduced in 2001, there have only been four other instances of the TRIN hitting 3 or greater on two consecutive days. Again, this is ignoring the first few minutes of trading, and some of you may have different data depending on your quote vendor. Those four occurrences were all in 2002, and were on 4/11 - 4/12, 6/6 - 6/7, 8/28 - 8/29 and 9/3 - 9/4. Two of them lead to smallish bounces over the next few days, while the other two just lead to further declines, which doesn’t tell us a whole lot.

When the TRIN ended up over 3.0 on a closing basis, however, the results over the next few days were quite positive, and extremely so over the past six years. While we technically didn’t close above 3 either today or yesterday, we came close enough to at least consider those other instances. Each of the past 12 times the TRIN closed above 3.0, the S&P 500 was higher 3 days later, with an average return of over 2%. This is more applicable to yesterday than today, so we’re already starting in quite a hole if this occurrence is going to match its recent predecessors.

Bottom Line: This volatility index is now at a point that has signaled good short-term lows in the past.

Another sign of extreme oversold conditions is reflected by the fact that the VIX (CBOE Volatility Index) is more than 20% above its 10-day moving average. A rising VIX shows that S&P 500 put options are becoming relatively more expensive relative to calls as demand for them increases, and observing it in relation to a short-term moving average has been an effective way to pinpoint short- to intermediate-term extremes in equities.

A VIX reading of 20% above its 10-day moving average is a rare event. Since 1986, there have only been 90 days on which that has happened (on a closing basis). Even rarer is such an occurrence when the market is in an uptrend, broadly defined as the S&P 100 index (OEX) being above its 200-day moving average. When those two conditions (VIX more than 20% above its 10-day moving average, and the OEX being above its 200-day moving average) are met, the number of days shrinks to 36. As an example of how rare this type of thing is, the last occurrence was nearly five years ago.

S&P 100 Performance When VIX > 20% Above 10-day Moving Average and OEX > 200-day Moving Average 1986 - 2004
	1 Day Later	3 Days Later	5 Days Later	10 Days Later	20 Days Later
Avg Return	0.6%	1.2%	1.1%	1.3%	1.6%
% Positive	77%	86%	83%	74%	60%

Notably, three days after such instances, the OEX was higher 30 out of 35 times, with an average gain of just over 1%. Up to 10 days later, the market was higher considerably more often than a random period.

Once again, the broader market is at a testing point. Each of the other occurrences of our current level of short-term oversold conditions has resulted in minimal additional downside, and more often than not a stiff rally. This time is already quite different simply because of the severity of the selling, according to many of our measures either outlined here or shown elsewhere on the site. But these are all short-term in nature and suggest nothing beyond what may occur over the next week or so. Most of our intermediate- to long-term indicators continue to be either neutral or, in many cases, still overbought. Since January I have been looking for a lengthy trading range to develop, with new highs and new lows both likely to fail. The new high failed last week, and now we are close to testing new lows (for the trading range). The downside prospects over the coming days should be very limited if the market holds to form. If not, stand aside and look out below, as that type of failure would likely lead to a more severe and time-consuming decline.

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.