http://www.sentimentrader.com

Sunday, October 19, 2003

I’ve been mentioning lately the speculation displayed by traders of the Rydex funds, and how it had become extreme.  Our shorter-term measurements such as the Beta Chase Index and RSI Spread had reached areas that consistently corresponded with other short-term peaks in the broader market. 

Friday’s negative market performance switched these traders’ activities in a major way.  On Friday alone, $118 million left the long-side Rydex funds and $89 million went into the short-side funds.  Looking at other days in the past three years where the S&P declined between 0.9% and 1.1% (54 occurrences), this is one of the larger shifts in assets.  In fact, it is the largest one-day shift since late November 2000, when looking at market declines similar to Friday.  I don’t want to make too much out of one day, but this type of activity needs to be monitored closely.  If it becomes habitual for these traders to shifts assets to a bearish posture more than they “should” given the underlying market performance, it can give a good signal that at least a short-term bounce is due. 

Some time ago I showed a chart of what I called the Rydex Enthusiasm Index.  This index is created by measuring how much the assets shift around the various Rydex funds, and compares it to how they “should” move given how the market performs that day.  For instance, say the S&P 500 rises 2% one day.  Usually when the S&P rises that much, $170 million flows into the bullish Rydex funds and $30 million flows out of the bearish Rydex funds, on average.  But on this (example) day, $340 million poured into the bullish funds and $60 million fled the bearish funds.  Thus, we can say that these traders were roughly twice as optimistic as they usually are when the market puts in a comparable performance.  The Rydex Enthusiasm Index attempts to measure that type of confidence or fear.  Below, I’ve included a chart of a 3-day moving average of this index.  When the Index reached its upper extremes, showing that Rydex traders were more optimistic than they “should” have been, I drew in a red dotted line.  When the Index reached its lower extremes, highlighting those times when the traders were more pessimistic than normal, I drew in a green dotted line. 

This is one of the reasons the Rydex data can hold value as a reflection of sentiment.  Overall, I believe those who trade the Rydex funds are a representative sample of market participants in general, and they show the same tendencies to overreact in both directions.  By observing this type of behavior, when the traders overshoot in one direction or the other, we can place a bet with a reasonable chance of success that the market will gravitate toward the other extreme. 

Towards the end of September, these traders had become quite pessimistic indeed, as the Rydex Enthusiasm Index reached its lowest level in months.  This is one of the reasons I said at the time that the pessimism was palpable, and if we were about to see a situation that was the most confusing to the most number of players, then a sharp rally to the recent highs would be it.  By this past Tuesday, those worries had worn off and we were at the opposite extreme – by this time, these traders were shifting into the bullish funds and out of the bearish funds to a much greater degree than they should have been.  This was another tip-off that there was a lot of money counting on a breakout and continuation of the rally. 

Friday’s action went a long way towards working off the optimistic extreme this Index had reached last week.  In fact, if we get another day or two like Friday, it’s safe to say that this Index, as well as the other short-term Rydex indicators mentioned above, will enter oversold territory.  Should that occur, it’s a pretty safe bet that we’ll see at least a moderate bounce of 15-20 S&P points, even if the larger trend has changed. 

BREADTH FOLLOW-UP 

As I mentioned on Tuesday, truly extreme overbought breadth readings such as we saw this week (even after accounting for the distorting effects of decimalization) do not necessarily translate into a weak market.  In fact, such readings usually lead to a POSITIVE market more than a negative one.  The exception to this case is that when the market begins to lose steam right away, such as in the first 5 days.  In those cases, weakness was much more common later on than when the market continued to power higher immediately after becoming overbought.  At this point, we’ve declined just about 1% from the point where we became “maximum” overbought, which was the cutoff I used to determine weakness in the first 5 days.  So, we’re right on the cusp here, but I would like to see more weakness early next week before coming to the conclusion that the market is reacting to an overbought condition by going down instead of just sideways. 

ViXeN 

On Friday, the VXN declined even though the NDX lost more than 2% of its value.  This doubtless had something to do with expiration, and is not necessarily just a reflection of complacency among traders, so I’m not certain how much can be read into it.  Typically, when the NDX declines, the VXN rises as traders demand put protection and the market makers are more than happy to accommodate them (for ever-increasing prices, of course).  If we look at the period from 2/2/01 through Friday (the period that the CBOE has provided re-constructed VXN values with their new formula), then when the NDX declined, the VXN rose 74% of the time.  When the NDX dropped more than 2% in a day, then the VXN rose 83% of the time.  It could be that this is simply a “Friday phenomenon” due to how the new VXN calculation treats calendar days (a complicated detail that is unimportant for the most part), but that does not seem to be entirely the case.  Of the 22 days when the NDX declined 2% or more and the VXN declined as well, 5% of the days occurred on a Monday, 9% on a Tuesday, 41% on a Wednesday, 14% on a Thursday and 32% on a Friday.  There certainly appears to be a bias towards later in the week, as 86% of the occurrences happened on Wednesday or after, but it’s not skewed enough towards Fridays to conclude that what happened this past week was par for the course. 

I looked at what happened after the NDX declined 2% or more in a single day, both for those days on which the VXN increased as it should have, and those days when the VXN declined as it shouldn’t have.  The table below outlines those results.  Again, this is for the period 2/2/01 - 10/17/03 only.  In the table, “Avg Ret” refers to the percentage change in the NDX the given number of days later; “% Pos” shows how often the NDX was higher the given number of days later; “Max” is the largest gain seen the given number of days later (close-to-close only); and “Min” is the largest loss seen the given number of days later.   

Overall, these results show that the NDX performed moderately more poorly after those days the VXN declined than those days when it rose.  For example, 10 days later, the NDX showed an average gain of 0.5% when the VXN rose (and it was higher 45% of the time), but it showed an average loss of 2.9% when it declined (and it was higher only 33% of the time).  Also, look at the maximum gain and maximum loss:  when the VXN rose, the max gain was 43% and max loss was -20%; when the VXN declined, the max gain was 10% and max loss was 14%.  Again, I don’t want to read too much into one day’s activity, particularly with the VXN just getting switched over to its new calculation, but I feel comfortable saying that how the VXN performed on Friday is a mild negative for the Nasdaq 100 going forward. 

SMALL TRADERS 

Small option traders continue to pile in.  For the latest week, the smallest of options traders – those trading 10 contracts or less – bought to open 792,000 calls as opposed to 317,000 puts.  Other than a reading seen a few weeks ago, this has pushed the ROBO put/call ratio^TM to its lowest level in years, at 0.40.  And, once again, they are paying handsomely for their upside exposure – they paid 27% more for their calls on average than they did for their puts.  This marks the 10^th straight week that we have seen a “demand premium” for calls that is 10% or greater, which is the longest streak since the Spring of 2000.  For the second week in a row, call buying made up 35% of their volume allocation, and bullish strategies in general made up 53% of the total.  This is the fifth week out of the past seven that bullish strategies have outweighed bearish ones, and it should be a surprise to nobody that this is the most concentrated amount since 2000. 

A subscriber (thanks Damien) pointed out to me that even though these traders may be paying up for their calls more than their puts, they may not necessarily be bullish, which is absolutely correct.  If a trader is buying a deep in-the-money call, it is going to cost far more than a far out-of-the-money call, and in fact this trader may be selling calls just above.  His suggestion was to compare not only the premiums paid for options that were bought to open, but also those that were sold to open, and to look at the volume flowing into each.  From his suggestion, I created the index below, which tracks the premiums for all calls and puts bought and sold to open, and weights them according to the volume flowing into each.  This gets somewhat complicated, but the bottom line is that the lower the index, the more bullish the traders.  In contrast, when the index is high, it shows downright panic among these smallest of traders.  This tracks the ROBO put/call ratio quite closely, but is more susceptible to extremes.  As you can see, the index is at its lowest value since the week ended July 14^th, 2000.

WEEKLY REPORTS 

I’m getting tired of saying this, but once again this week the weekly reports didn’t show any notable changes: the sentiment surveys continue to show excessive optimism across the board; no major position changes are evident in the S&P futures contracts; and the relationship of public short selling to NYSE specialist short selling remains near long-term historical (bullish for the market) extremes.

CONCLUSION 

We are now entering the seasonally weakest time of the month, as I pointed out last week.  Over the next six days, only Monday has a positive expectancy.  I don’t give much weight to specific days, but I certainly do think there is something to be said for the strong bias exhibited in the middle of the month as opposed to the beginning and the end.  Also, on Friday we closed just below the 1040 level on the S&P, which is what I had been looking for to indicate that a more serious correction could be setting in.  Last week, we reached a breadth and sentiment extreme, and closed below the low of the high bar.  That is a consistent tipoff to expect more weakness ahead, so my preference continues to be with the short side. 

Shorter-term, the extreme overbought condition from last week is wearing off quickly, and as we can see from the Rydex traders, it may not take much to swing sentiment to a short-term pessimistic extreme, similar to what was seen in late September.  Should that occur next week, I would be inclined to look for a bounce of at least 15-20 S&P points. 

 - 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.