Breadth Says Up, Sentiment Says Down
Sunday, March 28th, 2004 9:30am EST
Thrust & Parry
Bottom Line: Monday’s downside thrust and Thursday’s upside parry carry with them some possible historical significance – more so than either of them separately.
In last Monday’s comment, I noted that the lopsided selling pressure we saw that day was not necessarily meaningful. Sometimes it marked a short-term low, sometimes not - historically a 10-to-1 ratio of downside volume to upside volume on the NYSE just didn’t lead to any kind of consistent future market performance.
Thursday’s rally, however, may have changed the historical meaning of Monday’s severe selling. On Thursday, we saw almost the opposite of what we saw on Monday. While we didn’t quite see 10 times as much advancing volume as declining volume, we came close, and these types of reversals over the past 64 years have lead to a consistent edge over the intermediate-term.
Looking at breadth data from 1940 through the present, I first identified all days that showed at least a 10-to-1 bias of down volume over up volume. I then marked any day that showed at least an 8-to-1 bias of up volume over down volume within 5 trading days, thus nearly reversing the severe selling seen previously. Comparing future performance in the Dow Jones Industrials Average over several time periods, there was a definite positive edge over the intermediate-term during those times when there was an upside reversal as opposed to those times where no upside reversal took place.
Looking at the past few years, this pattern has been extraordinarily positive. The last occurrence marked the low in March 2003, the time before that was the low in April 2001, then the low in September 1998, and finally the low in October 1997.
It didn’t always work out so well, of course. The chart below shows the performance in the DJIA the given number of days after these types of very lopsided selling days (the red bars and line), and it also shows the performance after the lopsided selling days were followed by lopsided buying days within 5 trading days later (the green bars and line).
There were a few cases where we saw multiple 8-to-1 upside reversals on successive days. Also, there were numerous instances of these downside thrusts and reversals occurring within just weeks or months of each other. In order to prevent “double-counting” of future returns, I used only the first day if there were consecutive upside reversals, and I also eliminated any instance that occurred within 6 months of another occurrence. These steps left us with 60 samples over the past 64 years.
It is very clear from the chart that the green bars and line are higher than the red bars and line over every time frame – meaning those times that saw upside reversals within 5 days performed better than did those times where we did not see a reversal. In the short-term, meaning 10 days later, the Dow actually showed a negative return. By the time 30 trading days had elapsed, however, the positive impact of the upside reversals was being felt. We can see that after 60 days, after those times where we saw an upside reversal, the Dow was higher nearly 80% of the time, with an average return of over 3%. If we compare that to those times where there was no upside reversal (and the Dow was higher 61% of the time with an average return of barely 1%), it becomes clear that the upside reversal lead to a more positive market.
As I mention nearly every time I talk about breadth figures, I do believe that decimalization makes it easier for us to achieve these kinds of outsized swings in advance/decline and up/down volume figures. While there have only been two of these occurrences since decimalization took effect, it is something that should continue to be monitored.
Fill ‘er Up
Bottom Line: Large opening gaps in the S&P futures tend to get filled sooner rather than later – particularly in a downtrend. This coming week should prove important if Thursday’s gap is not filled.
In an intraday comment on Thursday, I posted some figures on gap opens in the S&P 500 futures (the full contract, not the e-mini). I want to repeat myself a little here, because I think the ramifications could be telling.
What I looked at where those times when the futures gapped open more than 0.5% above their previous close, while the futures were trading below their declining 50-day moving average, which is a scenario similar to what happened on Thursday. Since 1998, the gap open was closed 56% of the time on the same day. Going back to 1986, the gap was closed the same day only 46% of the time. Just to clarify, by “closed” I mean that the futures declined enough to trade at or below the price of the previous close. For example, if the futures closed at 1110 one day, then gapped open at 1115 the next day, they would have to decline enough to trade at or below 1110 at some point in order to consider the gap closed.
What is more interesting to me, however, is how long it took before the gap was closed most of the time. Since 1998, there were 16 occurrences of the type of gap we experienced on Thursday; 15 times (94% of the time) the gap was closed within 5 days. Looking at all occurrences since 1986, there were 68 such gaps and 53 of them (78%) were closed within 6 trading days. Now this is what I find perhaps the most telling – if the gap was not closed within 6 days, then it also was not closed within 16 days. Meaning, if the futures didn’t fall back enough to close the gap in the first week, then they also didn’t fall back enough to close it up to three weeks later. 15 out of 15 times the futures continued higher, or at least didn’t decline enough to fill the gap. After 22 days, an additional 5 gaps had been closed, but after that, none of the remaining 10 instances were closed up to 30 days later.
Does it matter that we were trading below a declining 50-day moving average in the futures? Yes, it does. The graph below plots the percentage of time opening gaps of 0.5% or more were filled the given number of days later depending on whether we were above a rising 50-day m.a. at the time (the green line) or below a declining 50-day m.a. (the red line).
After 5 trading days, gaps that occurred when we were in a downtrend were closed 78% of the time. When we were in an uptrend, however, the gaps were filled only 60% of the time. Similar distinctions are apparent across all time frames up to 30 days later. This tells us that the market is more likely to experience a “gap and run” if we are trading in an uptrend, but more likely to “gap and crap” if in a downtrend. That may seem like common sense, but common sense when applied to the stock market is not very common at all.
This data makes the coming week notable in that if we can levitate long enough to keep the gap open, history suggests that we shouldn’t fall back to fill it within the next two weeks either.
Bearish Put Spread
Bottom Line: Last week, large traders bought puts to open to the greatest degree in three years compared to small traders. This is not a perfect indicator, but it is not encouraging for bulls.
One of my larger concerns for several months has been the overwhelming lack of fear on the part of those most likely to lose their money – very small options traders. I highlighted the R.O.B.O. put/call ratioTM last week, and for the week just past it remained at 0.43 – well below other points that marked “fearful” market bottoms.
This past week, traders of 10 contracts or less concentrated on selling calls to open. That particular strategy accounted for 42% of all of their volume, which is a relatively large percentage historically. Call selling as an opening option strategy generally reflects a bearish opinion, as the traders would lose if the stocks on which they sold calls rose in price. However, opening put purchases accounted for only 12% of total volume, which is on the low end of the historical range. Apparently these traders are assuming the market will not rise much, but not fall by much either. There remains no scramble for protection.
Larger traders buying 50 contracts or more at a clip showed a different set of priorities. For these traders, put buying accounted for 28% of total volume. Large traders normally buy more puts than small traders, most likely because they are hedging existing long positions in their portfolios. Still, last week’s reading is one of the largest percentages in the past few years. So we have large traders buying puts to a degree that’s on the high side, while small traders’ put buying is on the low side. In fact, the spread in put buying between the two is now the widest it has been since 2000. That doesn’t sound particularly encouraging to me, and the following chart reinforces that feeling.
In the chart above, when the tan line is high, it shows significantly more put buying by large traders than small traders; when the tan line is low, it shows large traders buying less puts than usual compared to small traders.
On average, large traders buy around 8% more puts to open than do small traders, and the spread between them tends to range between 4% and 12%. When it extends beyond those bounds, typically the market has formed some sort of turning point. Low readings, below 4%, have tended to coincide with market low points, while high readings above 12% have more likely lead to market weakness going forward. While this is certainly not a perfect predictor of market weakness going forward, the fact that this spread is now the widest in three years is not a positive indication to me.
Conclusion
I have pointed out several intermediate-term positives over the past week or so, mostly related to the extensive selling pressure we saw as reflected in the TRIN or other assorted breadth measurements like those in “Thrust & Parry” above. But many of our pure sentiment measures continue to show a lack of fear among those traders who should most likely be showing it. This stubborn bullishness is not a positive thing for the market, as we should prefer to see these types of traders washed out before looking for another leg higher (that may sound terribly crass, and it is, but that’s how the game is played).
There are a few indications that sentiment is shifting – the lowrisk.com and AAII surveys are both showing low levels of bullishness, but unfortunately those two surveys also have the most flawed methodologies out of all of them. Still, I consider their recent movements encouraging, and the AIM model is back to where it was in August of last year from its historic bearish (for the market) extreme in January. We’re also seeing Rydex traders trying to fade rallies (such as I pointed out with SMH on Wednesday) and pile out of long funds on declines which is a positive sign. Overall, we haven’t reached a point that would have me feeling comfortable on the long side - another push lower below last week’s low would likely do it, though. I think it helps to define an “uncle” point, which is some indication that I’m just plain wrong and we won’t see another push lower. A rally (and hold) above 1125ish on the S&P would do it for me, and would be enough to have me favoring longs as opposed to continue looking for another decline.
Jason Goepfert
President and CEO
Sundial Capital Research, Inc.
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.
© 2004 Sundial Capital Research, Inc. All Rights Reserved.