http://www.sentimentrader.com

Sunday, August 17, 2003

The e-mini contract of the S&P 500 has been getting increasing attention as a valid and useful trading tool for several years.  Part of the effect this has on our data is reflected in the Commitments of Traders information, which I have discussed many times over the past year.  My opinion has been that the e-mini data for the S&P 500 futures contract has been of limited use.  That contract has only seen “adequate” position sizes for about a year, and for information that is released weekly, as the COT data is, this just doesn’t give us enough data points to be comfortable, even if it seemed effective as a market forecasting tool. 

I’ve been hearing anecdotally that some commercial traders are using the e-mini contract as opposed to the full contract for a variety of reasons, such as better fills and opportunities to arbitrage the e-mini against the full contract.  I’ve also heard that some commercials have been using the e-mini as a hedge against their full contract positions.  This became even more pronounced after 9/11, where traders needed access to markets that traded nearly 24 hours a day, as the e-mini does.  I always take these anecdotes with a grain of salt, unless they can be proved with some degree of confidence by how the data plays out.  In this case, there does seem to be a possibility that those who trade the full contract also have a hand in the e-mini. 

The chart below shows the “variance” for the full contract versus that of the e-mini.  The variance in this case is defined simply as commercial net positions minus small speculator net positions.  The scale for the e-mini (on the right-hand side) is inverted to more clearly show this relationship. 

Since the scale for the e-mini is inverted, when the two lines move together, it actually shows an inverse relationship between the two contracts.  And as we can clearly see, the two lines move together a great majority of the time.  In fact, the correlation since August 2001 is -0.82 on a scale of +1 to -1.  This is a very high negative correlation, and from this we can say that nearly 70% of the movement in one contract can be explained by movement in the other contract.  It is very unlikely that this is due to chance alone. 

Let’s run with the theory that commercial traders are using the e-mini to hedge.  If we back out those hedges, then we may get a more accurate picture of what their true take on the market may be.  The following chart shows the traditional net position of the commercial traders (the tan line), which is computed simply by subtracting short positions from long positions – when it is greater than 0, commercials are in the rare position of being net long.  The red line is computed by backing out the e-mini positions, adjusting for the size difference between the two contracts.  If commercials are using the e-mini to hedge, then this red line should be more representative of their true market outlook. 

We can see that up until 2003, the two lines tracked rather closely.  Beginning with the market descent into March of this year, however, the paths began to diverge.  Near the March low, the traditional net position (again, the tan line) was relatively short, which was suggesting that commercials were not especially intent on buying into that decline.  The “corrected” position (the red line), however, was the least net short it had been in over a year, suggesting just the opposite – that commercials were in fact quite keen on the market and were expecting it to rise.  This is more in keeping with their historical accuracy in calling major market turns. 

The traditional measure stayed bullish throughout the entire rally, until the June expiration as I’ve discussed several times before.  The corrected measure began to become bearish as May approached, and has been chopping around since then.  This brings up an interesting juncture that we’re seeing now – the traditional measure continues to become more bearish, as commercials keep adding to their net short position.  However, in the e-mini, commercial net positions turned positive in July, and have been rising steadily since the June expiration.  This battle between the two contracts is what is making the “corrected” measure chop around.   

Frankly, I still don’t know if this argument carries any weight.  We simply do not have enough data to make any judgments about whether large contract positions should be adjusted in some manner for what is going on in the e-mini.  The striking negative correlation between the two does suggest that there is some link between the contracts, but adjusting for them somehow does not appear to consistently improve the forecasting ability of this data.  This juncture in the market should give us more good information, as the distance between the traditional measure and the “corrected” measure is quite large.  For now, I am sticking with the assumption that the relatively large short position for commercials in the large contract is moderately bearish for the market, despite their large (and growing) long position in the e-mini. 

The VIX is once again down to the lower end of its recent trading range.  Though a low VIX reading does not necessarily mean anything in and of itself (in fact, looking at the VIX Fear Premium, the VIX is right about where it should be, given how tame the market has been recently), I did a quick look at every other August 15^th since 1986.  Out of the twelve occurrences, the VIX was higher 21 days later 58% of the time, with an average gain of 12%.  More notably, however, is that after 42 days (roughly equating to mid-October), the VIX closed higher than it did in August 83% of the time, with an average gain of nearly 20%.  The average VIX reading for August 15^th historically has been 21, while one month later it has averaged 23.7 and two months later 24.6.  This is another sign that historically, there is a good chance that we will see volatility increase, not decrease, over the coming months.  I want to once again include a chart I first presented a while back, as I think it is important to keep in mind - the seasonal pattern of the VIX since 1986: 

As we can see, we are now moving out of the lowest-volatile period historically (the early Summer months), and are moving into the most-volatile (late Summer / early Fall).  This must be kept in mind, as increased volatility is usually seen during market declines and not rises. 

The equity put/call ratios have been quite low this week.  As I’ve laid out several times, however, these put/call ratios tend to trend at least 7% below normal when expiration week has a positive bias, as it did this week.  Still, the equity-only ratio (minus QQQ options) closed Friday at 0.37 which is the lowest figure seen since the June market high.  For as far back as the CBOE has kept QQQ option numbers separately (May 2001), such low put/call ratios have tended to precede market weakness consistently.   

If we look at equity-only put/call ratios (minus QQQ options) under 0.40 since 2001, the S&P 500 closed higher the next day only 3 out of 11 times, with an average return of -0.6%.  Going out five days, the bias becomes even more negative, as the S&P was higher only 1 time , and the average return dropped to -1.6%.  Most notably, if one had sold short the S&P 500 (cash index) at the close on the day with the low put/call ratio, the average drawdown (loss) over the next five days would have been 0.9% (with a maximum loss of 2.9%), while the average gain would have been 2.7% (with a maximum gain of 4.5%).  The table below outlines each occurrence of the equity-only put/call ratio (minus QQQ options) under 0.40, with the return of the S&P 500 1, 3 and 5 days later, along with the drawdown and maximum gain seen over the next 5 days. 

Obviously, the suggestion here is that market weakness is more likely than strength over the next week, though the relatively limited sample size and predominant downward market environment during the lookback period temper the bearishness of Friday’s reading somewhat. 

Over the past couple of years, there has been a consistent negative bias to days following option expiration (which was this past Friday).  In fact, so far in 2003, every day immediately following option expiration has closed in the red, with an average loss of 1.8%.  This negative bias has tended to persist for the rest of the week.  We have enough evidence from a short-term and long-term sentiment perspective that makes me want to look for entries to short this market.  While I continue to be cognizant of the bullish arguments, and in fact I’ve laid out several of them myself over the past couple of months, there is still too much enthusiasm in the air for my liking.  If we decline below 980 on the S&P 500, I would want to be quite aggressively short.  However, if we are able to overtake and hold 1005, I would back off the short side, and a break of 1015 would force me to cease looking at the short side for the time being, as it would validate market strength and my short-side preference would be wrong.

 - Jason Goepfert

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.

© 2003 Sundial Capital Research, Inc.  All Rights Reserved.