Trader’s Narrative

Here’s this week’s sentiment summary:

NASDAQ:NYSE Volume
I touched on the relative lack of volume and how this has historically marked tops, rather than spurred on rallies. Another troubling volume development is the ratio of volume on the Nasdaq compared to the NYSE. We are seeing a spike in this ratio, meaning that Nasdaq volume is significantly more than NYSE volume. Since the Nasdaq represents the riskier side of the market, this has usually meant that there is too much froth in the market.

Sentiment Surveys
Last week I pointed out the danger of having the AAII remain at 53% bullish when the market had gone down slightly. Although this week the AAII respondents were slightly less bullish (45%) it is good to see their excitement abate in the face of a week that saw the market go up.

This reversal however only slightly dilutes the contrarian bearish meaning of this indicator. We are still at a very high level of bullishness. Simply by receding from the extreme level of bullishness that we saw last week does not eliminate the topping signal that the AAII is giving us at this point. After all, it was in October 2007 when we last saw this sentiment measure at such heights.

Investor’s Intelligence, the measure of newsletter sentiment compiled by ChartCraft, showed little change going from 44.4% bullish to 46% bullish.

Hulbert Newsletter Sentiment Survey
In contrast to the sentiment surveys mentioned above, the Hulbert Stock Newsletter Sentiment Index is continuing to suggest that newsletter editor are still either skeptical of the market’s recovery or not really excited by it.

At the start of the week, the HSNSI was 16.2%, meaning that the average market timing newsletter was suggesting to their clients being long just 16.2% of their portfolio. This is much lower than the 27.5% which they were recommending in late April, even though at that time the market was lower than it is now.

Slicing and dicing the newsletters, Hulber finds that the stock market newsletters with the best track record of timing the market are continuing to be bullish, while those that have lagged buy and hold are much less so. This gap in sentiment and performance has, however, significantly diminished from mid-March - when the market hit its inflection point.

Is LowRisk Dead?
A reader already asked about the lack of updates from LowRisk. I emailed Jeff Walker, the keeper of the data and when or if I receive a reply I’ll write a follow up. The last update on their site is for March 23rd 2008. I haven’t received any email updates either - I’m subscribed. In any case, LowRisk was never one of the major sentiment surveys that I relied on. It was a bit too volatile and no one except Walker knew the size of the sample size.

Magazine Cover Indicator
Here’s an interesting magazine cover for analysis (below): “Barbarians at the vault”. It is the latest Economist cover showing a bank being besieged by a horde of angry “barbarians” carrying banners with such slogans as :

• Skin the fat cats!
• Just say no to CDOs
• Regulate now
• Salary limits

The building is on fire, there is a column being pulled and there are sledgehammers being taken to its pillars. Pretty powerful imagery. I wonder how it will play out with the Philadelphia Banking Index being where it is:

Check out the Economist’s “prowling bear” cover in 2006. As disclosure, I’m long the AMEX Financial Select Sector ETF (XLF).

On Wednesday - February 13th - after 3 consecutive up days, I mentioned the peculiar way that option traders were in denial of the rally and of the likelihood of seeing a pause:

It would be very normal for the market to pause and digest this short term move up but the negative sentiment is undeniable.

We got the “pause and digest” the following day on Thursday and today. So to dive into all that negative sentiment, here is the stock market sentiment recap for this past week:

LowRisk.com
I mention this sentiment survey sparingly because it is very jittery and much less famous than its peers. But this week’s reading of 64% bears and 24% bulls reminds me of the last time that bearish sentiment was 60% (June 2007).

Unlike then, this week’s bullish ratio (bulls divided by the sum of bulls and bears) is quite high at 27.27%. But there is no denying that the respondents are very gloomy about the Dow’s prospects. Their median guess of the Dow (closing value February 22nd) was 11852 - well below the Dow’s January swing low of 11,970.

Investor’s Intelligence
If you’ve been keeping up to date with these sentiment overviews then you know that the II survey has been insistently and stubbornly stuck with a clear bullish consensus. For contrarians, that has been disconcerting not only because II is a major sentiment survey but because it contradicted all the other surveys.

This week it seems “reason” has finally prevailed in newsletter land. According to ChartCraft, the keeper of this indicator, the bears now account for 35.6%, and the bulls 36.7%. While that may seemingly put them neck and neck, the historical data for this survey gives us a decidedly more bullish interpretation.

Newsletter editors are naturally bullish by nature, after all, optimism sells. So it is almost impossible to find less than a third of them bullish at any point in time, no matter what the market condition. The current percentage of bulls is as low as it was in the summer of 2006 and 2002 (and no other time since). So I can comfortably say that the II is officially flashing a contrarian buy signal - finally!!.

AAII
Meanwhile, the AAII (retail investors) sentiment has finally decided to come up for air from the depths of despair it had sunk to in January 2008. The AAII sentiment survey spent 5 consecutive weeks (December 21st, 2007 to January 18th, 2008) being 50% or more bearish. The bears are now “only” 42% (with the bulls at 33%).

We’ll have to wait a few more months to see if the stock market follows the previous script or if we stray. According to the above chart, we could find the S&P 500 at 1558 by June 2008.

That’s not a prediction, by the way. I’m just extrapolating from the historic averages. But it could turn out to be prescient, so write it down somewhere or bookmark this so you can come back and mock me

Consumer Sentiment
The most recent Reuters/University of Michigan’s consumer sentiment survey was released today and it shows a stumble from 78.4 to 69.6 - the lowest since 1992!

As I’ve discussed before, consumer sentiment measures are a contrarian indicator. By the time they reflect doom and gloom, it is too late to sell, and in fact a better time to buy.

Whether that is because of the time lag built into this kind of survey or whether it is because of the forward discounting ability of the stock market (or both), the historical evidence shows that significant lows in consumer sentiment are buy signals for stocks.

I’ll going to write more about this indicator soon.

Everyone these days pays at least some attention to sentiment. Which means that the usual sentiment indicators run the risk of becoming saturated with attention, and therefore less effective.

Which is why I love finding novel ways of measuring sentiment. One that I’ve already written about is the “sheeple index” which is a measure of retail broker’s web traffic.

Another indicator is to measure the level of interest that people have in specific internet searches. For example, using Google Trends, we can see a history of people searching for “stock market crash”:

Since people would have to be really bearish to even think of typing in such a search string, it may have value from a contrarian point of view. As you can see each time there was a spike high in the number of searches for “stock market crash”, it was a great time to actually go long!

And here’s the chart for the keyword: “subprime”:

Again, note that the two spike highs correpond to intermediate bottoms in the market.

What about you? Do you have an unorthodox way of measuring market sentiment?

According to sentiment, it is time to sell bonds (and buy stocks).

Bond Market Sentiment
A few days ago the Hulbert Bond Newsletter Sentiment Index reached 47.4% — that’s four times the average sentiment over the past year and eerily enough, it is a repeat of what happened last year at this time. We had a slightly higher level of bullishness last November as bond prices carved out a major top.

Two other bond sentiment measures also show a dangerous level of bullishness. Market Vane sentiment survey shows 68% bulls - you’d have to go back to 2005 to find a higher level. And Consensus Sentiment shows 63% bulls, down from a recent peak of almost 80% but still quite high.

According to contrarian analysis, you want to fade the crowd and that would mean selling bonds.

Chart Request
In the comments section of my update on the timing the stock market using the rate of change of the 10 year T-Bonds, scood asked for a chart of the same covering the 2001- 2004 timeline:

It depends on how you want to interpret that signal. One could see it as a failure or success since price meandered before continuing to rise again. If you want to be really strict, it could be interpreted as a failure since price didn’t immediately go up.

In any case, nothing is 100%. This and other indicators are mere guides and should only be used as such.

Well, that was something. Wasn’t it?

Simply breathtaking.

I love it when things align like that. I had to take care of some things yesterday or you would have found me gloating about this much earlier

An old saying around Wall Street is that the market moves to exert maximum pain to the maximum number of participants. If this isn’t what just happened, I don’t know what is. Who in their right mind could have foreseen what happened this week? the breakout? the sheer intensity and unrelenting buying pressure?

Unfortunately for them, those that bear the brunt of the damage tend to be the outsiders, the unwashed masses, otherwise known as retail investors and traders. For the most part, insiders make out just fine.

Speaking of insiders, the latest Commitment of Traders report, hot off the presses, shows a continuation of the same lopsided buying from the commercials. Brace yourself, because since the last report, commercials have increased their aggregate net long position by 36%.

Eventhough I was bullish as I outlined this past week, I was still taken aback by the force of the bulls. They simply demolished the bears. I don’t care what thesis a bear trots out: inflation, the dollar, China, etc… it is all meaningless in the face of such a clear and powerful breakout.

What’s more fascinating is that according to the Hulbert sentiment measures, newsletter publishers are not at all excited about the market… even as it has gone up. In fact, in the face of a rising market, sentiment has actually become less bullish. To be specific, in late February 2007, when the market was going strong (before the correction), bullish sentiment as measured by Hulbert, was 62.4%. Since then, the market has recovered the correction and then some. But newsletter bullish sentiment is now 40.6%

And Thursday’s rocket ride did not make one iota of difference! It seems like the newsletter writers out there are scoffing at the price action as merely a mirage. Who knows, it may very well be. But with history as a guide, we know that the probability of it being real is high when there isn’t much excitement accompanying market gains.

But (there’s always a but) the Investor’s Intelligence report on newletter sentiment shows a completely different picture. According to II there are 21.3% bears and 49.5% bulls. That is a fairly toxic ratio of bulls to bears. But if we look back to 2003 we see that after the bear market low was made, and as the market rallied almost continuously for the whole year, the sentiment was equally bearish according to II. So maybe history will repeat. But why are the sentiment measures so at odds?

I believe it is because Michael Burke and his team over at II use a different method to categorize newsletters. Rather than going by what the newsletters are recommending (which Hulbert does), they go by their feel of the bullishness or bearishness (or neutralness) of the newsletter after a reading. This is a much more subjective method and may account for the variance between the two sentiment measures. Ideally, we would like to see them agree with each other. But we can’t have everything now, can we?