http://www.sentimentrader.com/subscriber/daily_comment.php

September 2, 2006

9:30 AM EST

An Indicator Reveiw

Over the past five years, I've done thousands of studies.

From commonly cited technical indicators to the most obscure phenomenon you could imagine, we've looked under an untold number of rocks - hey, you never know where you might find an edge.

But in that time, I can't recall going through a period like the past month, where essentially nothing I've looked at, especially with an intermediate-term time frame, lead to anything consistent.  I've mentioned before that I have a "spreadsheet graveyard" on one of my PC's, where discarded studies get stashed until I bother to clear them out.  Over the past few weeks, that folder has gotten stuffed as I've closed the books on one study after another.

It's still difficult finding something that has lead to consistent results going forward, but I thought with the long weekend it would be a good time to do a general overview of our broad indicator groups.

VOLATILITY

Other than the last week of 2004, this week has been the least volatile of any other in the past decade, using a measure called Average True Range.

The same can be said of the implied volatility indicator, VIX.  If we look at the "volatility of volatility", we again see that we are experiencing a truly extreme level of non-movement, even by holiday trading standards.

In the chart below, what we see is the 10-day high - low range of the VIX, expressed as a percent of the current closing reading (so we can more accurately compare our recent conditions to those prior).

For example, the highest value the VIX reached over the past two weeks was 12.92, while its lowest level was 11.91, which it hit on Friday.  So the difference between the highest high and the lowest low was 1.01.  We then divide that by Friday's close of 11.96 and we get a "volatility of volatility" reading of 8.4%.

What's remarkable about this number is that it's the lowest in the 16 1/2 years for which we have data - never before have we seen the VIX contract into such a tight coil.

Going forward, anytime the "volatility of volatility" dropped under 10%, one week later the S&P 500 was positive 23% of the time (5 out of 22 days), with an average return of -0.9%.

When looking out between two weeks and one month later, there wasn't much difference than a random return, suggesting this kind of extreme coil in volatility was a short-term phenomenon in terms of its possible affect on the market in general. 

Bottom line:  BEARISH

PUT/CALL RATIOS

Most of the put/call ratios that we follow have been relatively muted lately, with a few standouts.

The ISE Sentiment Index, a ratio that tracks only options that were bought to open, has been exceedingly low lately, showing that traders have been busy buying put options.  The 10-day average of the ratio was recently at its lowest point in its three-year history, tied with mid-October 2002.

This ratio isn't skewed by many of the options-market strategies such as call or put writing, which theoretically makes it a more reliable measure of actual trader sentiment than the data released by the more widely-followed Chicago Board Options Exchange.

Confirming that is our ROBO put/call ratio.  Recall that this indicator only looks at the very smallest of options traders, those trading 10 contracts or less at a time, and restricts the volume to only those options that were bought to open.  So it's about as pure a sentiment gauge as we're going to get, since we know precisely who we're monitoring, and have a very good idea about what they're trying to do.

For the week ended August 11th, that gauge spiked to its highest level of the past few years, suggesting that these small-time traders were at their most-pessimistic since early 2003.  It has since dropped off a bit.

The one troubling development in put/call land comes from the OEX put/call ratio.  While volume has been very light this week, no doubt a contributing factor to the extreme readings we've seen, the ratio has spiked to an exceedingly high level.  Unlike most other put/call ratios, this is not a contrary indicator, rather it should be included in the "smart money" category, as the chart below shows.

The 10-day average of the OEX p/c ratio is now more than 25% above its yearly average (the red dotted line).  Over the past two years, every other time this has occurred the S&P 500 was lower one month later every single time, by an average of nearly 2%.

So we have conflicting signals in this data - the bullish being the ISE Sentiment Index and the bearish being the OEX p/c ratio.  There were a couple of times in the past few years, those being June 2004 and September 2005, when we saw a similar situation of a very high OEX p/c ratio and relatively low ISE number.  Equities declined going forward both times despite the bullish suggestion from the ISE index, but the spread wasn't as large as the extremes we're seeing now.

Bottom line:  NEUTRAL to BEARISH (?)

BREADTH

On July 17th, I posted a Chart in Focus Video regarding the percentage of Nasdaq 100 stocks above various moving averages.  The crux of the video was that the components of the index were washed out to a degree that lead to intermediate-term rallies consistently over the past decade.

The NDX bottomed the next day, and has subsequently tacked on over 100 points.  It should be no surprise that the breadth stats have changed markedly from that time, so that close to 80% of the component stocks are now above their 10-day and 50-day moving averages - the highest percentage since early January.

A curious development is that while nearly 80% of components are above their 50-day averages, less than 40% are above their 200-day.  This is an unusually wide spread - over the past decade, only three times have we seen a similar thing - those being early May 2001, mid-August 2002 and late October 2002.  All of those, of course, were classic bear-market rallies.  Hmmm...

There are a few other disturbing signs, like the fact that the S&P 500 is mere points from a new high, yet the cumulative a/d line for the S&P that we post to the site is below where it was at the market lows this spring.  And the TRIN on the Nasdaq, which has become so overbought on a 21-day basis that it's challenging the most extreme readings over the past five years.

Bottom line:  BEARISH

SENTIMENT SURVEYS

For the most part, we're not seeing a whole lot of people out there who are saying that they're bullish on equities.

The most egregious example is the lowrisk.com survey, which asks respondents if they think the Dow Jones Industrial Average will be higher or lower 30 days in the future.  The latest results showed only 16% voted for higher while 67% gave the the Dow their thumbs-down.

The four-week average of the opinion of the lowrisk participants is at one of its lowest points in the nearly ten-year history of the survey.  While the Dow was mostly higher going forward after other occurrences of such apathetic responses, its average maximum gain was not a whole lot higher than the average maximum loss, something that always puts me on guard about reading too much into the results.  I prefer to see at least a 2-to-1 max gain-to-loss ratio, and this one doesn't come close to that.

The Investor's Intelligence survey has also shown a low bull ratio, even after the market recovery over the past few weeks.  The Market Vane, Consensus and AAII polls, on the other hand, have either continued to show a high level of bullishness, or at least rebounded significantly from excessive bearish opinion seen in July.  The AAII survey in particular, has seen the number of respondents expecting a decline to dry up dramatically lately.

In March, I had gone over what happens when we see the unusual circumstance of traders and investors becoming more bearish while prices are rising.  Contrary to the "wall of worry" theory, we saw that such behavior was typically quite negative for the stock market going forward, so it's certainly not a lock to believe that high levels of bearish opinion now necessarily will lead to higher prices going forward.  However, given that we have been coming off a fairly deep correction, the hesitancy we've seen for investors to at least say they are significantly more bullish is more of a slight positive than anything.

Bottom line:  MILDLY BULLISH

COMMITMENTS OF TRADERS

In early August, the Commodity Futures Trading Commission (CFTC), the regulatory body for the futures markets, changed the way it classified several large traders in various commodities, including the stock index futures.

Because of those changes, we're not really comparing apples-to-apples when we look at the positioning of commercial hedgers and large speculators now versus those positions a couple of months ago.  That makes it very difficult to come to any solid conclusions about what those traders may be up to.

Fortunately, it didn't affect the "non-reportable", or small speculator, category.  In the latest release, covering positions active through this past Tuesday, we see that those small specs reduced their net long position in the index futures by nearly 40% over the past month.

Small specs are considered "dumb money" since they seem to be their most long at market peaks and least long near market troughs, so that reduction in bullish bets should be good for the market.

The chart below shows the nominal dollar value of small speculator holdings in the three major equity index futures, including both e-mini and full size contracts.  The green arrows highlight other times in the past five years that the dollar value dropped near or below the current figure of $15 billion.

Sure enough, the other occurrences were almost universally bullish for equities.  Out of the 28 weeks in that time that the net position was at or below this level, the S&P was higher four weeks later 21 times (for a 75% "win rate") and its average change was +2.7%.  The average maximum loss during those months was -3.8% compared to an average maximum gain of +5.3%, which is not at all spectacular, but good enough to pay attention.

If we look out over the next quarter, then the win rate jumps up to 89% and the average return rises to +6.9%.  The average drawdown didn't increase much at -4.6% while the average max gain doubled to +10.6%...suggesting that if we are going to see weakness after these kinds of numbers, then it would be more likely to occur during the first month.

Bottom line:  BULLISH

SHORT SALES

One of the big questions that gets asked on any rally is "is this just short-covering?".

The theory is that if a rally is mainly traders covering short positions by buying back shares, then the buying is not because "strong hands" are moving the market, and the rally attempt is more likely to fail.  I haven't seen any studies that would support or refute such a theory, but it makes sense to me and on a very short-term time frame, my experience would seem to support it.

It's difficult to know whether buying pressure is organic or due to a covering of shorts, but we can get something of a clue by looking at short interest figures released by the exchanges each month.  Short interest is the number of company shares sold short, betting on a market decline, that have not yet been bought back.

One that has stuck out to me over the past couple of months is the short interest in the Russell 2000 exchange-traded fund, IWM.  Below is a chart of the short interest in that ETF expressed as a percentage of the float (i.e. available shares):

In May 2005, the short interest exploded to nearly twice the available shares (due to how exchange-traded funds are structured, it's possible to see short interest greater than the actual number of shares outstanding).  That proved to be a particularly conspicuous buy signal.

In June of this year, we once again saw short interest shoot well above the number of shares outstanding.  While small caps didn't take off like they did in 2005, it coincided with the halt of the spring decline.  Over the past couple of months, the short interest as a percent of the float went from 145% in June to 100% in August - but that's not due to a covering of short positions, it is due to an increase in the number of units outstanding.

As for broader-market gauges, short interest on the NYSE last month rose to another new record high, but when adjusted for volume, the short interest ratio rose only slightly and is about in the middle of its range over the past decade.

On the Nasdaq, it's a different story.  Short interest exchange-wide rose modestly, but average daily volume dropped off about 7% (using Nasdaq-provided volume figures).  That resulted in a nearly 9% jump in the short interest ratio to a new record high.  The ratio has been an OK indicator at highlighting long-term extremes in tech, and this could be considered a bullish factor going forward.

Bottom line:  BULLISH

CASH

According to some brokers like Charles Schwab, investors are beginning to consider "cash" an asset class once again.

And it's no wonder why - money markets and the like were barely yielding 1% a couple of years ago, and now one can find these "risk-free" deals north of 5%...that's approaching the long-term historic return for a risky asset like stocks.

In the past, when there has been a marked increase in short-term rates, we have seen a coincident rise in cash levels at mutual funds.  Portfolio managers are no different than anyone else - if they see a risk-free 5% yield, they're going to park more of their available assets there.

But not now.

Three years ago, the yield on 90-day Treasury Bills was 0.96%, and at the time mutual funds were holding 4.5% of their assets in cash.

As of last month, the yield on those T-Bills was a whopping 4.97%, yet funds actually decreased their cash on hand to 4.1% of assets.

The Mutual Fund Cash Premium / Deficit that we post to the site takes these short-term rates into consideration, and computes how much cash mutual funds are holding compared to how much they "should" be holding.

The latest release showed that funds now have about 2.9% less cash on hand than they historically have given the current level of short-term rates.  That's a truly extreme figure that has been exceeded only twice in the past 50 years.

In the chart below, I have highlighted the five other times in history that the cash deficit either exceeded the current level, or got close enough to count.

Obviously the precedents aren't that encouraging for the long-term bulls out there.  The average decline in the S&P 500 (using monthly closing prices) before a sustained advance was -21% over about a year's time, though there was a wide variation among the occurrences.

What did not show a wide variation, however, was that any additional upside after the indicator had become this extreme was limited and temporary.

I've gone over several reasons in the past why we may be seeing a permanently lower range in this measure, but I'm not going to re-hash them here (that's what the archives are for...).

Bottom line:  EXTREMELY BEARISH

RYDEX ASSET FLOWS

The Rydex data has been compromised, folks, and it's time to adjust our thinking.

Sure, there are certain short-term discrepancies that continue to have some value (I still like our Beta Chase and RSI Spread indicators, and the sector assets), but the longer-term index fund bull / bear flows have changed.

Constantly changing market dynamics, and quite possibly the popularity of the Rydex Ratio itself, have turned a large number of these traders into trend-faders, not trend-chasers.  In addition, new trading vehicles like the Profunds leveraged and inverse ETFs have taken away a major advantage that Rydex had, and I wouldn't be surprised to see the bread-and-butter Rydex index fund assets just dwindle away into shadows of their former selves.

A case in point is how the fund flows have behaved since the July low.  The S&P has gained nearly 80 points in that time, yet assets in the leveraged S&P and NDX bull funds have actually dropped by $70 million and the leveraged bear funds haven't really lost any assets at all.

One could of course suggest that this is evidence of distrust in the rally and thus should be bullish.  I'm not so sure...one would have had that same mindset this spring, but as I showed then, this kind of behavior is not necessarily a bullish "wall of worry".

Bottom line:  NEUTRAL

Conclusion:  Short-term Looks Dicey, But Longer-term

Should See a Wide Trading Range at Worst

When we see a high level of neutrality in our indicators, and/or an approximately equal number of them giving bullish and bearish signals, then we typically see a wide, extended trading range form.

I've been harping on that concept in the intraday notes, as it's what experience has shown us is a common result when the overall market trend is questionable and a seemingly large number of traders have equal but opposite opinions.

In the short-term, the market this week was subject to numerous very tight ranges (some of the tightest in recent history) and relatively low volume.  We knew that would be the case coming into the week, but the lack of volatility at some points was remarkable.

Even considering that it was a holiday-influenced week, such lack of movement has been consistently negative looking ahead, as we saw above with the "volatility of volatility".  A few of my favorite shorter-term guides are quite stretched into overbought territory at this point as well, and it rarely pays to hop onto the bullish stagecoach when they're in the front pulling back on the reigns.

I expect the gains made this week to be given back in fairly short order, but the intermediate-term is sort of a muddled mess.  Our Smart Money Confidence is under 40% for only the third time since 2002, something that tells us that the bigger picture is beginning to become negative.  That leaves plenty of room for shorter-term wiggles, but if the Dumb Money climbs over 60%, then it's going to be time to give serious thought to hedging any existing long positions.

All the best,

Jason Goepfert

President and CEO

Sundial Capital Research, Inc.

Forwarding or otherwise distributing this copyrighted material is a breach of your subscriber agreement.  Violators are subject to termination of their subscription with any received subscription fees forfeited.  Any references to historical performance are based on data we deem to be reliable, but are based upon feeds from third parties.  We do not recommend subscribers take positions based on data presented here alone, but rather incorporate it into a comprehensive investment outlook.

© 2006 Sundial Capital Research, Inc.  All Rights Reserved.