Trader’s Narrative

Back at the beginning of August, we looked a study of momentum thrusts which showed that historically, when the 10 day ratio of NYSE advance decline was pushed to an extreme the market tended to enter a protracted rally.

There were only 10 instances in the past 4 decades, with 2 of them occurring this year. Since some time has passed, I wanted to update the table and look at where breadth stands now:

From March 23 2009, the market rallied about 12% in 3 months and almost 30% in 6 months. And forward 3 months from July 23rd, the market rose 18.6%. That’s in line with the previous returns historically (or slightly better).

When Wayne shared this historical study with me, my initial concern was that this is relying on NYSE breadth data. As I’ve stressed several times before, NYSE market internals are now skewed by the increasing number of non-operating company securities trading on the big board. While the NYSE is the most well known stock exchange around the world, increasingly ETFs, municipal bond funds and other CEFs and even bonds have started to take a larger and larger share of trading.

But, if you look closely at the Nasdaq data, it corroborates the last two momentum thrusts that are shown above. In late March and late July you can see two distinctive spikes:

So far, so good. Or more accurately, in line with historical norms. Lately though, the breadth has been horrible. We haven’t seen market internals this bad since the beginning of the year - just before the spring rally was launched.

This is a guest post by Wayne Whaley (CTA):

I have written over the last few months on the importance of the historic momentum thrust that we have experienced this year and how they could possibly push the market higher than most would anticipate.

Today, Sept 16th, 58% off the lows of 666 on the S&P, I am rolling my eyes, looking to the heavens in disbelief and sharing with you that today we had the third “Ten Day 2:1 Advance Decline” reading in the last 6 months. The previous two were on March 23rd and July 23rd. I haven’t had a great deal of time to study it yet, but it appears unprecedented with really very little to compare too. But I would caution against interpreting this as a sign of an overbought market. As I have shown in the past, single 2:1 advance decline thrust signals are very bullish. Two in a short period of time, even more so, and I am assuming until I find evidence to the contrary that a third is bullish as well. Since I have no tri-signal data, let’s take a look at the three other occasions where there were double signals in a short period of time.

The table below shows the three previous double signal dates, followed by the percentage change in the S&P 500 and the returns 3, 6 and 12 months later.

Double Ten Day 2:1 Advance Decline Signals in less than six months:

Note that these three previous double readings occurred in different decades and notice that although the second reading came well after the initial advance was launched, the S&P 500 gained an average of an additional 26.37% over the next year, with nary a less than 24.28% gain. The S&P 500 index is currently up 9.5% since the second thrust that occurred on July 23rd, 2009.

This market is similar to all three of these markets in some ways. But appears to correlate the closest with 1975, which similar to this rally followed a 48.4% sell-off in equities that lasted 20 months; was often postulated to precede the next great depression and was accompanied by massive federal stimulus.

For much of the last 6 months, analyst have compared the recent rally to the bear market rally that took place between November 13th 1929 and April 17th, 1930 - rising 48% over five months. One can not totally rule that possibility out, but a major difference was the fact that in 1929, the preceding sell-off in stocks occurred in only two months, while the current rally followed a 15 month long bear market. It seems that in just in the last few weeks such comparisons have dissipated.

I understand that it works against human instinct to buy the market this high off the lows, but I assure young readers that investors in 1975 and 1982 had the same dilemma. If you have been waiting and are tempted to reenter, I suggest dollar cost average into it.

I see a lot written on indicators that are at historic levels with many interpreting this market as overbought. But if you look back through history, anytime there is a 50% or more move in the major indexes, the oscillators, sentiment indicators, etc, that are bounded from x to y are going to be tested repeatedly. Divergence analysis between range bound indicators and a potentially rangeless market tend to mislead at such times.

The chart below is from James Montier. It shows the average holding period for NYSE stocks (expressed in years) from 1920 to today.

Montier is an economist and global strategist who uses behavioral finance to make sense of the financial markets. He started his career at Dresdner Kleinwort, moved to SocGen and just recently moved to the hedge fund world. Montier has written several books, among them, Behavioural Investing: A Practitioners Guide to Applying Behavioural Finance.

I’m not sure where exactly Montier got the raw data for this graph but considering the caliber of research he does, I’m assuming it is an accurate reflection of the underlying change in the structure of the market over time. The chart is remarkable in setting out what we all intuitively know to be true. Driven mostly by high frequency trading, we’ve seen an explosion in advance decline volatility.

It seems I was wrong, when I said this isn’t your grandfather’s stock market. This is exactly your grandfather’s stock market. Indeed, your grandparents would readily recognize the sort of stock market we’ve had recently. What I should have said is this is not your father’s stock market (1950’s - 1960’s) where people actually invested by holding stocks for years at a time. In comparison, what we do now is push buttons with the attention span of a housefly:

Source: James Montier formerly of Société Générale

It is absolutely remarkable to notice that the turnover in 1929 - a time where trading was done over telegraphed message or scribbled notes and hand gestures - is equivalent to recent times when trading is done by blazing fast computers interconnected directly via FIX to the exchange.

If you enjoyed this, don’t miss Montier’s brutal take down of EMH (via John Mauldin’s Investor Insights). Also, in a world where we are all traders, the least we can do is be better traders:

• Why do traders fail?
• How to Fail as a Trader in 10 Easy Steps
• Dennis Gartmen’s Rules of Trading
• The Definitive Guide to Trading Mastery & Success
• 5 Fatal Flaws of Trading

A recent mention of cumulative TICK on Trader Feed blog caught my attention and I looked at this indicator today. Now I know, Dr. Steenbarger’s chart of cumulative TICK is using the NYSE TICK data and it is very short term in contrast to my analysis.

But he does briefly mention that “…the Cumulative NYSE TICK has stayed well above May levels.” And then goes on to extrapolate:

Continued strength in Cumulative TICK would suggest to me that we’re experiencing a correction in a bull market, not the start of a renewed bear.

I’m not so sure we can draw that conclusion. For reasons that I’ve outlined many times before, I prefer to use the internal breadth data from the Nasdaq. So here is a look at a few years worth of cumulative TICK for the Nasdaq:

If a higher high is a sign of a correction within a bull market, then by that account according to cumulative Nasdaq TICK we’ve never even entered a bear market!

Now I know this is the Nasdaq data but the NYSE chart doesn’t look all that different. Which reminds me of the uselessness of cumulative breadth numbers (advance decline) as any type of indicator - NYSE Breadth Is Strong: Why It Doesn’t Matter.

Instead of looking at TICK data cumulatively, I prefer to smooth it using a simple short term moving average:

Although it has come down from the extreme in April 2009, it isn’t anywhere close to the range that has historically coincided with market lows (or lasting market bottoms).

A very old indicator that measures the magnitude of speculation on Wall St. is the volume ratio between the NYSE and the OTC market. This ratio hearkens from the early days when Nasdaq was the over the counter market where smaller and riskier securities traded.

This was before the time of Microsoft (MSFT) and Intel (INTC) - bellwethers of both the stock market and the US economy. Today it would be more apt to substitute the Pink Sheets OTC BB market instead.

Even so, for some reason the volume ratio still holds and has flagged important market tops. Including the start of the bear market in October 2007.

There is no absolute level however in this ratio that can give us oversold and overbought readings. As with all analysis of volume, we have to contend with not only a seasonal pattern, but also a continuous increase of trading over time. Since we are looking at a ratio, most of the upward creep in volume should be canceled out.

Right now, the volume ratio is extremely high and points to a lot of froth in the market. We’ve seen other signs of this in the sentiment data as well.

To play devil’s advocate, the ratio could be higher because of the Nasdaq’s leadership (it has rallied much more than the NYSE since March 2009).