Drops After a High
Tuesday, December 7th, 2004 8:15pm EST
Drops After a High
Bottom Line: Relatively large one-day declines so soon after the NDX made a new yearly high have not necessarily been a negative indication going forward, especially (curiously enough) if bullish sentiment is high.
In the last comment, I noted the tendency of the Nasdaq 100 (NDX) to decline in the short-term after large gap opens in Intel stock after the NDX had just made a new yearly high. So far, we’re seeing a fairly typical scenario play out in that there has been weakness after the semi-euphoric INTC news. The logical question is “what next?”.
Looking at instances where the NDX made a new yearly high within 5 days and then sold off at least 1.5%, we can get an idea as to whether today’s rather dramatic sell-off so soon after the index hit a new high is a buying opportunity or a harbinger of things to come.
Since tech shares bottomed in October 2002, there have been 8 instances of the NDX dipping at least 1.5% in a single day soon after making a new yearly high. Only twice was the NDX higher five days later. However, the weakness was short-lived, as every time but once it was higher within 10 days, and 30 days later it was higher every time and sported an average gain of 4.1%. Eight occurrences is a very small sample, so if we go back to 1985 and look for the same thing, we come up with a better sample of 90 occurrences. A still-healthy 71% of those 90 were higher after 30 days and the average return stayed relatively stable at 3.8%.
There are two instances in there that would make a die-hard bear salivate. This type of pattern occurred before the two most devastating drops in the history of the NDX. On October 6, 1987 and March 28, 2000, the NDX dropped significantly soon after making a new yearly high, and 30 days later it was an average of 30% lower. Even better (if you’re a bear), in some respects bullish opinion is even higher now than it was in either 1987 or 2000. Taking one simple measure, the Investor’s Intelligence sentiment survey, the bull ratio now is over 71%. This compares to 65% in 1987 and 65% as well in 2000.
While the logical conclusion from that is that this is a very bearish omen indeed, historically it turned out to be quite the opposite. There were 13 occurrences where the I.I. bull ratio was 70% or greater when the NDX dropped at least 1.5% within 5 days after making a new yearly high. After 30 days, every occurrence but one was positive, and the average return was 3.6%. It also performed well if sentiment was at the other extreme - looking at those times when this pattern set up and sentiment was at its lowest, the NDX was higher after 30 days more than 80% of the time and the average gain was a whopping 9.8%.
Curiously, the times the index performed the worst were when there were no extremes to speak of. When sentiment was in the middle of the pack, neither excessively bullish nor excessively bearish, the index was higher 30 days later “only” 65% of the time and the average gain was just over 2%. It seems as though if investors are not already optimistic enough to be in “buy the dip” mode or pessimistic enough to be in the “it can’t get any worse” mode, then this type of market behavior proves confusing enough to leave traders with their hands in their pockets.
If we look at this setup one other way, and that is those occurrences which happened to show up in December, we are limited to only two samples. In 1996 and 1998, the NDX dipped more than 1.5% after making a new high in the days prior several times each year. In both cases, it proved to be a phenomenal buying opportunity. After 30 days, the index was an average of 18% higher (mostly due to 1998) as the market climbed the rest of December and into the beginning of the following year. I’ve made mention before of the market’s consistent reluctance to sell off significantly heading into a new year, and while a sample of two is statistically meaningless, I think it highlights something important. Oversold conditions, however you define them, approaching the last two weeks of the year tend to get bought.
A Switch by Small Traders
Bottom Line: Odd lotters have switched their opinion pretty quickly. After scrambling for long exposure in November, they are now shorting to a very heavy degree.
There is a fair share of indicators we can point to which indicate investors are far too optimistic, even given the breakout in two out of the three major indexes. But there are also a couple that are worth mentioning as holdouts. Rydex traders, after going way overboard in early November, have settled back down and are actually showing a bit of skepticism towards the gains since then. And odd lot traders, after they too had become quite enamored with the long side a few weeks ago, have recently shown a desire to bet against the market.
I have included a chart of odd lot short sales below, which is the same chart we post to the site. Recall that odd lot trades are those executed for fewer than 100 shares, and a short sale is typically done as a bet on falling prices. So taken together, the odd lot short sales indicator should theoretically show us how much small, unsophisticated traders are betting against rising prices.
Over the past 5 trading days, these traders have averaged nearly 1.4 million short sales. This is an extraordinarily high amount, particularly for a market that has been rising or chopping around at worst. It is only the eighth time these short sales have risen to around 1.25 million shares or more, with the other times shown by green arrows on the chart below.
The chart clearly shows that those other times were good spots to be considering taking the other side of these trades. The recent data is unusual in that this is really the only case where these traders were not shorting into a down market. Either they know something now that they didn’t before, or this is one positive data point for the market going forward.
Conclusion
In recent comments I’ve been hanging on the opinion that if the market was going to decline this month, it was most likely going to occur in the first half. Given many of the signs we were seeing, that appeared to be a likely scenario. Now that we’re finally getting a little downside, my preference is shifting to find a buying opportunity for the balance of the month. A rally into year-end is very much a popular outlook, but as we saw with the Thanksgiving bias, once in a while a seasonal pattern emerges that it just doesn’t pay to fight. Obviously anything can happen in any given year, but if looking at the odds of a significant rise versus that of a significant decline, they overwhelmingly favor the long side in the latter half of December.
Our shortest-term model is now entering oversold territory, but it will take another couple of days for more readings to register that type of extreme. If we can get more downside over the next few days, or even just more of this type of choppy action that we’ve seen for a couple of weeks, it should set up nicely for a long trade into the beginning of 2005.
Jason Goepfert
President and CEO
Sundial Capital Research, Inc.
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.
© 2004 Sundial Capital Research, Inc. All Rights Reserved.