Sep 28, 2003

Last week, the CBOE total put/call ratio averaged 0.86, which is near the upper end of its range. The equity-only p/c ratio averaged 0.73, which is also near the upper end of its range. Contrasting with that data somewhat, was the equity-only p/c ratio with QQQ options removed, which averaged 0.62 and is only about average. Each of those ratios gets us successively closer to a “pure” ratio which should be an effective contrary indicator. Like I’ve discussed several times over the past few weeks, however, each of those ratios is trumped by the ROBO put/call ratio^TM, which allows us to know precisely who is trading the options and what their strategy is. That’s what makes this week’s reading, at 0.38, so troubling, as it is the lowest since the week ended January 26^th, 2001. It is especially bothersome since the market was so negative during this time. There doesn’t appear to be any particular post-expiration bias to this data, so I don’t think that had a very large impact. If we look at the strategies these smallest of traders (trading 10 contracts or less) employed during last week, bullish strategies, including call buying and put selling, accounted for 54% of the volume, which is the greatest in over two years.

Put selling accounted for 21% of total volume for these traders. Interestingly, this was the greatest percentage since the week ended January 17^th, 2003. The reason I think that is interesting is because that was the week that equities topped preceding the decline into February and March. It could be that these traders felt at the time (and perhaps do this time as well), that it was just a normal, garden variety short-term correction and so they began selling puts in earnest. Selling puts in an opening transaction is a moderately bullish strategy, as it tends to profit the seller as long as the underlying security does not decline below the strike price (and stay there) prior to expiration.

Not only was put selling as a percentage of the total the highest since January 2003, on an absolute basis these traders sold a whopping 523,500 put contracts to open. These traders had sold more than 500,000 puts to open during only eight other weeks since 2000. Seven of the weeks were in 2000 itself, and the other was, not surprisingly, the week of January 16^th, 2001 (the same week as the extremely low ROBO p/c ratio from two paragraphs above).

With the traditional put/call data, we don’t know who is trading the contracts, how many contracts they are trading at a time, or what their strategy is. With this data, we can be relatively confident that we know each of those important points. With that in mind, we can say that the smallest of options traders - the ones usually most wrong on market direction and most likely to lose their money - are extremely optimistic about the market and believe last week’s decline was just another blip that will be quickly recovered.

On Wednesday, I talked about the high level of the Down Pressure indicators for the S&P 500 and Nasdaq 100. With the additional downside we experienced late in the week, these indicators are now in truly extreme territory – in fact, the Down Pressure indicator for the Nasdaq 100 just set a new record, going back to July 2002, at 93%. This means that over the past three days, of all the point changes in the 100 component stocks of the NDX, over 90% of them can be accounted for by points lost (as opposed to points gained), and more than 90% of the volume has gone into those that closed down on the day. That is true selling pressure. While Friday’s reading was a new record, there were three other days which showed a Down Pressure indicator reading of 90% or above, and the table below illustrates how the NDX performed the given number of days later:

	1 DAY LATER	3 DAYS LATER	5 DAYS LATER
07/10/02	+4.1%	+6.2%	+7.3%
08/05/02	+5.2%	+10.4%	+9.5%
11/11/02	+3.0%	+8.8%	+7.5%

Obviously, we can’t read too much into three past occurrences, but it does at least suggest that the rubber band has been stretched a bit too far, and is likely to snap back.

Since this only goes back to 2002, I wanted to approximate these conditions and go back further. So I looked at all instances where the NDX closed down more than 1% for three consecutive days, with a cumulative loss of at least 5%, for all days since 1985. There were 73 such occurrences, and the results were quite positive. The NDX was up greater than 65% of the time 1, 3, 5 and 10 days later, with an average return that far outpaced that of a random return. However, while this type of selling pressure preceded some fantastic gains over the next few days, it also preceded a couple of sharp selloffs, most notably in mid-October 1987.

Also notable are our intraday cumulative TICK indicators for the Nasdaq. The one that’s posted to the site is currently at -3556, a reading not seen since March 11^th. I also keep a longer-term intraday cumulative TICK indicator, and that one is the most oversold since October 9^th, 2002. As I’ve said many times, these types of indicators are most effective when giving contra-trend extremes (i.e. oversold in an uptrend or overbought in a downtrend). The most obvious reason for this is because it provides for high-probability, low-risk entry points for a resumption of the larger trend. However, they can also be valuable by just letting us know that things are going along as they should be.

An excellent example of this was seen during the strong uptrend this Spring. These cumulative TICK indicators gave overbought reading after overbought reading throughout April and May. Other than perhaps a day or two of rest, the NDX kept powering higher. That told us (or should have told us, anyway), that there was something different about this rally, as it was not being stopped by overbought readings. We can put the market to the same test now, only in reverse. If the uptrend is intact, then we should not continue to decline here. We should see a rally of at least a few percent off of these oversold readings. If we do not, then we have a good piece of evidence that the intermediate-term momentum has deteriorated, and the larger uptrend is likely exhausted for now. That would have me wanting to look for short sale possibilities on any further rallies.

Last week, I mentioned the large changes in the Commitments of Traders data for the full and e-mini S&P 500 futures contracts. I noted that while the changes were positive (for the market), the week prior to and the week of expiration tend to show the largest position changes of all the weeks, and was wary of reading too much into them. As expected, we saw another large change this week, and again it was in a direction that would normally be positive for the market.

For the week ended this past Tuesday, commercial traders reduced their net short position by nearly 24,000 contracts, and are now barely net short. Small speculators reduced their net long position by over 14,000 contracts but are still carrying a relatively hefty net long position. The current net position by commercial traders is one of the most positive seen since the bubble burst in 2000, with the only period exceeding the current one occurring after the March 2003 expiration.

After going over the data extensively this weekend, I am becoming increasingly cautious about giving any of it too much consideration. The changes we are seeing around expirations are extremely violent. I don’t think it’s a coincidence that this odd behavior has coincided with total position sizes (commercial long and short positions plus small speculator long and short positions) in the e-mini approaching or exceeding 1,000,000 contracts beginning with the March expiration.

Over the past year, the correlation between commercial net positions in the full contract versus the e-mini is -0.92. For small speculators, the correlation is -0.94. This is out of a scale of -1 to +1, and with the number of samples we have, it’s fairly safe to say that the chances that this has occurred by chance is essentially zero. Statistically, this means that 85% of the movement in one contract can be explained by the movement in the other contract. This is a very significant correlation, and I don’t believe it can be ignored. I’ve stated before that I don’t know why this distinct correlation exists, whether it be hedging or arbitrage between the contracts or whatever, but it does exist and it is significant. Until we have more data, I no longer feel comfortable relying on the COT data for hints as to market direction. If we continue to see commercial traders building up a net long position in the full contract over the coming weeks, then I think that will be a positive development, but the activity in the e-mini needs to be factored in there somehow.

For the month ended September 10^th, short interest on the NYSE declined 1.5% from August. This is after a 4.2% decline in August and a 3.2% decline in July, and marks only the second time in a decade that we’ve seen a three-month decline of this magnitude (the only other time being January 1999). The total number of shares of NYSE issues sold short and not yet covered has now declined 10.5% since the peak in October 2002. Our modified short interest ratio, which adjusts short interest for the seasonality inherent in volume and short interest itself, now stands at 4.10. This is lower than what was seen in November 2000, which was the previous low in recent history. In fact, we have to go back over a decade to see a value lower than what we’re seeing now. The situation isn’t nearly as egregious in the Nasdaq, as the short interest ratio is only coming back to neutral after setting a new record high in March.

I’ve received quite a few inquiries about margin debt carried by NASD firms. This figure has been talked about frequently lately because of a chart similar to the one below, most notably shown by perma-bear Alan Abelson in Barron’s, which shows total debt in customers’ cash and margin accounts at NASD-designated clearing firms only.

These figures are current through July only, but we can clearly see that margin debt has exploded, more than tripling from May to July, and is now higher than what was seen at the height of the bubble.

Curiously, they do not show the related chart, which is free credit balances in customers’ accounts. Credit balances are essentially cash sitting in a customer’s account. It is unfortunate that I have not seen this chart anywhere else, since it is needed to give a more objective view of the chart above.

This paints quite a different picture. Credit balances have traditionally been thought of as a small-trader contrarian indicator, since most savvy traders would not leave a credit balance in their account, earning little or no interest. As credit balances rise, that shows these traders are selling stock and leaving the cash sitting idle in their accounts. We can see that credit balances have nearly doubled in the past two months.

That leads us to the logical next step, which Alan Abelson most certainly will not show you. For this chart, I simply took the difference between the credit balances and the margin balances, essentially showing how much cash was in customers’ accounts over and above the debt that they owed.

While cash went NEGATIVE to the tune of $4 Billion at the height of the bubble, it was a POSITIVE $12 Billion in April of this year. While this has dropped by nearly half over the past two months, we can see that this is still positive by over $6 Billion.

Lastly, I want to show a chart I first presented several months ago, and that is margin debt in NYSE firms’ customers’ accounts. This is a very long-term, de-trended look at margin balances, and as we can see it is still the lowest it has been in four decades.

The indicator is computed by comparing the average margin balance over the past year to that over the past five years. Currently, the one-year average is more than 20% below the five-year average, which is more severe than it was at the major buying points in 1971, 1975, 1982 and 1990. It should be understood that the increasing use of margin can be a very GOOD thing for the market, as it provides a consistent and powerful fuel to propel stocks higher. It is only when this type of activity reaches speculative extremes that trouble often sets in.

We are extremely oversold in the short-term, especially on the NDX, and I expect to see at least a small bounce early this coming week. If we are able to exceed Friday’s highs, that should set us up for some short-term follow-through to the upside. How that bounce is treated (should it come) will tell the tale longer-term. I’ve been talking about how I wanted to see three distinct failures before becoming too anticipatory on the short side (i.e. failure at a previous high, a break of a simple downtrend line from August 6^th, and an inability to rally off short-term oversold conditions), and if the market does not rally pretty much immediately, then I think it’s safe to say that we will have seen all three failures, and my preference would shift to viewing rallies as potential shorting opportunities.

From a longer-term perspective, we are not yet what I would call oversold, and it would likely take another week or so of a steady downtrend to get us there. If we are able to get into an intermediate-term oversold condition while remaining above 960 on the S&P 500, then I think that could set us up nicely for a trade to the long side.

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.