http://www.sentimentrader.com

Sunday, July 6, 2003

Many of you noticed the very high level of the total put/call ratio on the CBOE (Chicago Board Options Exchange) on Thursday, with it closing above 1.0 for the first time since May 22^nd.  That last “signal” was an excellent tip-off to go long, but the makeup of Thursday’s reading was quite a bit different.  Instead of equity options traders concentrating on puts, like they had in mid-May, to a relatively large degree it was QQQ and OEX traders concentrating on puts on Thursday.  In mid-May, the equity-only put/call ratio (minus QQQ options) averaged over 0.70, while it was only 0.60 on Thursday.  A 0.10 difference in the ratio may not seem like much, but it is nearly 1 standard deviation. 

I’ve been asked many times why I keep mentioning the equity p/c ratio minus QQQ options.  The chart below shows the QQQ put/call ratio by itself, plotted against its underlying index, the QQQ.  For those of you not familiar with the QQQ, it is a hugely successful Exchange Traded Fund that tracks the Nasdaq 100.  A couple of months ago, I showed a chart showing the increasing number of times QQQ option volume spiked up to account for more than 15% of total equity option volume (which happened again on Thursday).  For that reason, and because of the chart below, I have been keeping it separate. 

It is clear from the chart that the 10-day average of the QQQ put/call ratio is about the highest it has been in nearly two years.  But is this a good thing or a bad thing?  Previously, very high readings have coincided with tops AND with bottoms, while low p/c ratios have also coincided with tops AND with bottoms.  If I had to give an opinion, I would say that higher ratios tend to lead to better market performance with a bit more consistency than low ratios do (particularly recently), but it would be a stretch.  To me, the QQQ put/call ratio has been not much better than random. 

Let’s take a look at the other group that accounted for the high put/call reading on Thursday – OEX option traders.  These traders also are not terribly consistent, but observing the put/call ratio here is certainly better than watching that of their QQQ counterparts, but on a non-contrarian basis (meaning high OEX put/call ratios tend to be bearish for the market while low OEX put/call ratios tend to be bullish).  The OEX put/call ratio on Thursday closed at an extremely high reading of 2.83, meaning OEX traders traded nearly 3 puts for every call.  This type of reading is extremely unusual, as Thursday’s was the 10^th highest reading in the last 20 years.  This was no doubt partially affected by the early close and thin volume, as total OEX option volume at the CBOE ran about ½ its yearly average.  However, even though total NYSE volume was the 5^th lowest in the past year on Thursday, OEX option volume was 42^nd.  This means that even though the total number of shares traded on the NYSE was extremely low, OEX option volume was heavier than it “should” have been. 

The table below shows how the OEX (S&P 100) performed after the other nine instances in the top 10 highest OEX put/call ratios since 1982: 

KEY

ROWS

·          Avg:  The average of each of the columns.

·          % Pos:  The percentage of time the S&P 100 was positive the given number of days later.

·          Max:  The maximum return seen within the given number of days.  This could be considered the “drawdown” if one sold short the OEX on the day of the high put/call ratio.

·          Min:  The largest decline within the given number of days.  This could be considered the maximum gain if one sold short the OEX on the day of the high put/call ratio.

COLUMNS

·          Close:  The change in the OEX from the close of the day with the high put/call ratio to the close the given number of days later.

·          Rise:  The highest high seen within the given number of days.

·          Drop:  The lowest low seen within the given number of days.

We can see from the table that such high OEX p/c ratios have been negative for the market.  After the other nine highest ratios, the market consistently had difficulty making significant headway up to 10 days later.  We can see that the market was usually negative after these readings (at least 6 out of the 9 times across each time frame), and the average (and maximum) decline was about twice as great as the average (and maximum) rise.  The greatest risk to the market appears to have occurred within the first few days after the reading. 

With a sample size of 9, it’s difficult to read too much into the results.  I went back and looked at all instances of an OEX put/call ratio of 2.50 and above, which almost tripled the sample size (even though it was still relatively small).  Those results confirmed the ones above, though the edge was a bit less.  The market was still usually negative up to 10 days later (about 60% of the time), and the average (and maximum) rise was less than the average (and maximum) decline, though it was less than the 2-to-1 from the table above. 

So out of Thursday’s high total put/call reading, we have equity options traders who did not appear to be any more or less bearish than usual, QQQ traders who had a huge bias towards puts (which really doesn’t tell us anything at all), and OEX traders who also had a bias towards puts (which appears to be bearish for the market).  From what I can determine, that high put/call reading looks more bearish than bullish. 

Last week, I went over the extraordinarily high bullish sentiment displayed by the AAII survey.  This week, the bullishness dropped considerably, as the percentage bullish went from 71% last week to 41% this week, and bears rose from 9% last week to 35% this week.  I’ve been asked if this sudden change negates the import of the extreme seen last week.  The answer is no.  Looking back over the 20 weeks with the highest bull ratio (bulls / (bulls + bears)) in the history of the survey, the following week showed a decrease in bullishness every time except twice, with an average decline of 15%.   

If we switch focus just a bit and look at the 10 largest drops in bullishness from one week to the next (excluding the current one), the recent change is the 2^nd largest in history.  Looking at the other 10 occurrences, the market was higher the next week (this coming week for current purposes) 8 out of the 10 times, with an average gain of 1.2%.  But the following week (the week following next), the market reversed course again and closed down 7 out of the 10 times, with an average loss of 1.5%.  Three weeks after that, the S&P 500 was down an average of 2.6%, with 6 of the occurrences being negative. 

I showed in the June 26^th daily comment (see the archives for reference) how such extremes of bullishness in the AAII survey tend to portend market weakness in the following one- and two-month periods.  This week’s huge drop in bullish sentiment perhaps spells a little short-term relief, but it doesn’t appear to change the overall longer-term negative market action that often follows such displays of investor confidence. 

Over the past two months, the VIX has been stuck in under a 5-point range.  During the history of this indicator, such a small range over such a sustained period is quite unusual, as less than 5% of the days have shown such a small range over the prior 42 days.  Since volatility is supposed to be mean-reverting, meaning periods of low volatility precede periods of high volatility and vice-versa, I thought it would be instructive to see what usually happened over the NEXT 42 days.  For those days that showed the smallest range in the VIX over the previous 42 days (the bottom 5%), the next 42 days showed a higher average range more than 85% of the time.  For those days that showed the largest range in the VIX over the prior 42 days (the top 5%), the next 42 days showed a higher average range only 1.5% of the time.  So, it does appear that periods of low volatility are followed by periods of higher volatility.   

If we look at the VIX’s range over the past two months as a percentage of the VIX itself, then our current situation becomes even more interesting.  Viewed in that vein, the current situation has been the least volatile in history, except for one other period – mid-July 1997.  During that time, the VIX was also stuck in a very tight range heading into the summer months, with a slightly smaller range than the current one.  The result was a small rally before the market declined into August and September.  The market eventually recovered during the latter half of September and October before the crash in October of that year.  Interestingly, the next least volatile period was early June 1989.  The result of that was a market decline throughout June, then a recovery during the rest of the summer – then a mini-crash in October of THAT year as well.  It would be ridiculous to leap to the conclusion, based on these precedents, that we will decline for a month or so, recover, then crash in October.  But I do think it’s reasonable to expect an increase in volatility over the coming months – and as we all know, an increase in volatility (as measured by the VIX), is normally associated with declining markets, not rising ones. 

There is one positive that sticks out among our longer-term sentiment measures, and that is the very low level of specialist shorting on the NYSE.  Specialist shorting dropped again during the most recent reporting period (the week ended 06/20/03), even though it was an overall positive week for the broader market.  Also, public shorting as a percentage of the total rose for the week, and is up considerably since early June.  I’ve noted before that it is relatively unusual for specialists to decrease shorting – and for the public to increase it – when the market rises.  The usual relationship is exactly the opposite.  While the public generally has decreased its shorting activity while the market has rallied since March (until recently, as noted just above), specialists have not increased their shorting consistently during the entire rally.  In fact, specialist shorting as a percentage of the total is now below where it was in early April, and just above where it was at the low in mid-March.  This is a lagging indicator, and doesn’t reflect the most recent two weeks’ activity, but it still should be considered one bright spot among all the other sentiment negatives. 

For those of you wondering, the most eagerly anticipated report in quite some time – the Commitments of Traders – will not be released until Monday afternoon due to the July 4^th holiday.  If there is something notable, I will be sure to let you know that evening. 

Last week, I said that any rally we saw would likely only serve to set up a high-odds trade to the short side with a time frame of at least several weeks, and I continue to believe that.  As long as we remain under the recent highs, I will be viewing upside attempts as opportunities to short, as I don’t see many compelling reasons from a sentiment perspective to be long in any time frame.

 - Jason Goepfert

Disclosure:  long OEX puts

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.