Approach to Lower End of Range Should Set Up Longs

Bottom Line: As stated before, program trading has wrecked several previously effective market indicators, but it appears as though it may also be affecting price action itself.

Last week I talked about program trading and its impact on total volume. The NYSE has now come out and addressed the figure of 70.5% that I noted:

“NYSE Program Trading announced today for the period June 21-25 reached a record 70.5%.

This reflects the impact of investment managers' portfolio rebalancing as a result of the annual Russell index changes last Friday. One of the impacts of this rebalancing is increased program trading activity.

Traders utilize program trading to help buy and sell baskets of stocks more effectively. This applies to Russell index funds and all funds tracked to Russell indices.

The Exchange's Crossing Session II, a type of program trading, peaked last Friday to more than 620 million shares (compared to a year-to-date average of 25.6 million shares a day), accounting for most of the increase in program trading last week.

It is also important to note that because the Exchange counts each program buy and sell order individually, and divides that number by single-counted NYSE volume (completed, two-sided transactions), the 70.5% figure effectively represents double the amount of actual program trading when compared to the single-counted NYSE daily volume.”

So, it looks like the figures from late June were, indeed, a fluke (though a fluke that emphasizes the importance of index funds in the current market). One thing the NYSE does not address, however, is the increase in program trading over the past five years and its impact on various aspect of market behavior. I’ve already discussed its impact on things like the Specialist Short Ratio, and the egregious figure from late June prompted me to look at price action itself.

It’s hard to talk to anyone who has traded the past few years without them noting the tight ranges and lack of volatility compared to years prior. To see if there was any link to program trading, I checked each week’s range in the S&P 500 from 1999 through the present. By “range”, I mean the highest high during the week minus the lowest low. The table below shows the percentage of weekly NYSE volume that was accounted for by program trading, along with the S&P’s pure point range and its point range expressed as a percentage of price (because point range differs when the S&P is at 1500 as opposed to when it is at 1000).

Year	Program Trading as % of NYSE Volume	S&P Point Range	S&P Point Range as % of Price
1999	19%	51	3.8%
2000	21%	67	4.8%
2001	28%	53	4.5%
2002	32%	46	4.7%
2003	40%	31	3.3%
2004	46%*	26	2.3%

From the table, we can see that in the year 2000, when programs accounted for barely more than 20% of volume, the S&P’s range was an average of 67 points. Put in other terms, that year the S&P gyrated an average of 4.8% from high to low during an average week. Now look at the current year – with programs accounting for 46% of total NYSE volume during an average week, the S&P’s range has declined to only 26 points, or 2.3%. So while program trading has doubled, the S&P’s range has been halved.

Still, comparing 1999 to 2003, program trading doubled but point range (as a percent of price, anyway) was tighter but not enormously so. Looking at it a few different ways, we can note that the 20 weeks with the tightest ranges (as a percent of price) showed program trading to be an average of 37% of NYSE volume. The 20 weeks with the widest ranges, however, showed program trading at only 28% of volume. Overall, the correlation between program trading and that week’s range in the S&P 500 cash index has been -.42, which is quite strong given the relatively large sample size.

Since we only have weekly data for program trading going back to 1999, there is a chance that the increasingly small range in the S&P is due to something else, and the correlation we’re seeing here is bogus. But since programs are now the elephant in the room, and since index arbitrage (buying stocks and selling futures, or vice versa) is the largest component of program trading, it makes fundamental sense that indexes like the S&P 500 should be seeing tighter and tighter ranges as programs become more prevalent. As traders and investors, to us it may mean getting used to relatively tight ranges in many of the broad market indexes, and adjusting our expectations accordingly.

Bottom Line: The TRading INdex for both the NYSE and Nasdaq is giving unusual readings. The impression is that these readings are a positive for the market going forward, but caution should still be exercised.

Back on March 14^th, I noted that at the time, the NYSE TRIN had closed above 2.0 for three consecutive days only twice since 1940, both of which proved to be good spots to be buyers. The most recent occurrence, which ended on March 11^th, also was a decent buying opportunity, as the S&P chopped for a few days, and ultimately bottomed 1.7% below the close on the 11^th before rallying for a 4% gain. I bring this up because once again, the TRIN has now closed above 2.0 for three consecutive days.

At the time, I showed that since we had so few instances of the TRIN closing above 2.0 three days in a row, it made more sense to look at those times the TRIN closed above that level for three out of five days, of which there were 12 prior occurrences since 1940. The results were positive going forward (though not spectacularly so) – for example, the Dow Jones was higher after 60 days after 9 of the 12 instances, for an average return of over 6%. These high readings are pushing up some of the moving averages that we look at as well. The 10-day TRIN, at a current reading of 1.5, is the highest since the March 2004 lows, and April 2003 prior to that.

Also of note is that the TRIN statistic for the Nasdaq closed just a tad under 4.0. This is the highest reading since April 9^th, 2003, which marked the end of the consolidation after the initial thrust coming off the major March low last year, after which we all know what happened. A Nasdaq TRIN reading over 4.0 had been extremely unusual up until 2001. In fact, I don’t show ANY instances from 1985 – 2000, but I show that there were 18 combined in 2001 and 2002. It’s no surprise that that data shows future returns to be quite dismal after high readings, as it was during the teeth of the bear market. We don’t really have any bull-market readings to compare the current one to, so I don’t believe looking at past occurrences will lend much help to us here.

As I stated in March, I’m leery of reading too much into these TRIN figures due to decimalization. I noted at the time that it was likely we would see more such extremes, and here we are not even two months later talking about the same thing. So my takeaway from this is the same as it was in March – the TRIN readings are a positive, but trading decisions should not be based on them alone.

We’re seeing just a few signs of a reversal of the excessive speculation we saw up to last week. Put/call ratios were high today, even with QQQ options removed. Volume in the QQQ exchange-traded fund was the highest in a month and a half, a welcome change from the excessively low volume I pointed out last week. I suppose that’s to be expected when the components of the Nasdaq 100 gain only a total of only 19 points over a 3-day period. This activity has pushed our STEM.MR Nasdaq model over the 70% level, something that has lead to short-term relief rallies consistently over the past year. The chart below shows the model since the beginning of the year, with prior peaks over 70% highlighted in green.

Still, the types of negatives I talked about last week don’t simply disappear after three days of downside. With short-term readings like we’re seeing, particularly on Nasdaq-type issues, we should see a bounce soon. However, it appears as though that bounce would most likely fail, and today’s lows taken out, leading to a better longer-term opportunity sometime in the next few weeks.

As I have been saying at the end of every comment, I believe that May saw an important intermediate-term low, declines that put us into an oversold situation should be viewed as opportunities to add or initiate long exposure, we should see modest gains for the year, but the best strategy will most likely take advantage of oscillating indicators (selling overbought conditions and buying oversold ones) rather than breakout strategies. The closer we are to the top of the trading range, the less inclined I am to be holding long positions. Vice-versa for approaches to the lower end of the range.

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.