Carpenter Analytix

Carpenter Market Remarks
              Notes on Recent Statistics and Inference

                   
www.CarpenterAnalytix.com
                                

Carpenter Market Remarks--2006 

Annual archives see links at foot of page.

December 26, 2006. Holidays wreak havoc with data analysis. Friday's data offer a typical example. Cross-sectional dispersion was minuscule; broadly followed indexes were down somewhat while our Nasdaq Core up; our daily "XUB" exuberance index showed positive both on NYSE and Nasdaq (with the positive Nasdaq pulse nearly a record-setting).

Last week we remarked the Nasdaq Core "deviations" plot (Core Index vs EMA of the Core), saying "A new negative deviation below the recent negatives would be objective evidence the correction has actually arrived." Alas, you can take your pick as to whether that has happened. The deviations path has moved lower than its prior December low, but not lower than its early November low.

On the Hedge Fund Analytix front, December's 4th Friday reported last week that although the Managed Futures equity exposure has still not penetrated the 80th percentile of proven over-optimism, even the present level (69th percentile) has plenty of inference for below-average 1-month and 3-month returns. (Query regarding 4th Friday to Info@CarpenterAnalytix.com.)

I have attached here a graphic you will recognize from various earlier messages. It shows the 3-year path of Nasdaq's "Downside Tail" of cross-sectional returns distribution. It stands at lowest level since inception (1999), with the sole exception of Early July 2005 (indicated by intersection with light blue trace-back with vertical drops). By my count, there are 11 cases of "lows" in this path over the three years. (The count is subjective, in that "normal" level is not stationary.) In nine of the 11 cases, market decline ensued shortly; typically for at least a few weeks, or a couple of months. In two instances of the 11 (indicated by "Fail" on the graphic), renewed upside bursts ensued instead. (One of those two fails was just this past October.)

RLC
Hanover NH

December 18, 2006.  Nothing is new this week...except we're that much closer to the end of the trend. The downside "Tails" that we've been showing continue lower yet, with the Nasdaq downside Tail now at it's "shortest" (smallest extension, lowest reading) since July 1995. At that time, the Nasdaq Composite still had a couple of weeks of upside ahead of it before the 2 1/2 month correction, but that couple of weeks is the longest delay we've seen from these levels. I've not attached an updated Tails graphic here, as it looks just like last week's (but with the Downside Tail even lower), and I don't wish to be clogging mailboxes unnecessarily. If you want to see the update, just ask (RobinC@CarpenterAnalytix.com).

What I have attached here are two graphics of the Nasdaq Core Index. The first just is a picture of the Core Index and its EMA, such as you have seen here often. The second is a plot of just the Core Index deviations from the EMA. The point of the deviations plot is that the positive deviations have shrunken notably from the initial upthrust, and the path is now essentially centered on zero...and that this change looks quite like the condition leading into this year's May-June-July correction. The red rectangles highlight the parallels between today and last April-May. A new negative deviation below the recent negatives would be objective evidence the correction has actually arrived.

Finally, as an aside, we note that Bill Miller is in the news for NOT beating the S&P for a 16th year (absent miraculous recovery in two weeks). Miller is hailed as a stock-picker, but we find he also ranks as a successful timer. He's one of the five whose net effective equity exposure we track weekly at the Prime Timers page of the www.CarpenterAnalytix.com web site. We find him currently reducing exposure rapidly (although not drastically) over the last three months, and a graphic of his exposure path is posted at the Prime Timers page. (For counterpoint, we find Steve Leuthold building equity; graphic also posted.)

RLC
Hanover NH

December 11, 2006. The trend is still up, so far as our evidence can discern...but hanging on by its teeth in the face of ongoing dissipation. Two attachments here illustrate the dichotomy. The Nasdaq Core Index is not only still rising and still above its EMA, but is actually posting new highs while the Composite index lags. The Nasdaq Tails, on the other hand, are showing the energy of the uptrend is almost all wrung out. On the Tails graphic we see the downside (negative return) tail is clearly down into the range at which intermediate downturns occur.

The Core Indexes and the Tails analyses are conceptually complementary. The Cores track mean cumulative returns excluding extraordinaries and outliers (and excluding minimal "noise" level returns). The Tails track the behavior of the outliers, measuring how far out the tails of cross-sectional distribution extend (upward and downward). Short tails show lack of trading urgency across the market generally , and short downside tails show lack of urgency to sell in particular...which is to say, complacency.

The stage was similarly set over a month ago, only to be interrupted by a new surge of energy (extension surge in the Tails). That's a very improbable occurrence, but it occurred. It's also improbable now.

RLC
Hanover NH

December 04, 2006. The duration of this uncorrected uptrend is amazing every week. This past week showed the published averages down, but our NYSE Core Index rose half a percent, while Nasdaq Core gave up only about the same. We do have a putative crossover of the Nasdaq Core Index below its EMA on Friday. But (a) the degree is negligible at 16/100 of 1%, and (b) in the past year or so the Cores and their EMAs have sometimes traced lengthy snake dances over-and-under their EMAs--sometimes for months--before gaining traction for a reversal. Note (in attached graphic) how such a dance carried on for three months (yellow diagonal line) prior to the market break in May of this year. That behavior makes for frustrated timing...but the generally risky characteristics are unmistakable.

The most recent energy charge in this rally, represented by the bulge in our cross-sectional Tails, is now three weeks behind us, That bulge wasn't as great as the prior Tails bulge (July-August), and the downside Tails (NYSE and Nasdaq) have shrunken considerably. They're again approaching (almost) the all-played-out zone....adding to the risk of downturn.

Most mutual fund sectors have been at or near peak equity exposure for several months. The one exception is LargeCap/Value funds, which are at notably low exposures. In general, having Value funds showing notably more caution than Growth funds is a risk factor, too, but for now it's only the LargeCap/Value sector; the MidCap and SmallCap Value sectors are still heavily committed in stocks along with their Growth fund counterparts. Our Hedge Fund Analytix (HFA) update shows CTA funds finally increasing equity more vigorously this week...yet not in excessive zone yet. Email RobinC@CarpenterAnalytix.com for HFA info.

RLC
Hanover NH

November 27, 2006. Most market data from last week is little different from the week before. The NYSE and Nasdaq Tails continue to recede from their secondary bulge, indicating a dissipation of energy from the uptrend. But there's no present evidence that the uptrend has finished. (Well, there was some such evidence Friday, but nothing definitive...and data from truncated trading days are suspect in any case.)

One thing notable that did happen Friday (and Wednesday) was the hit on the Dollar. The Fed's Major Markets dollar index dropped 0.80% Wednesday and 0.85% Friday. That may not sound big, but the current daily standard deviation for this index is 0.33%, so those two days represent 2.42 sigma and 2.58 sigma declines. Moreover, the dollar is in a quite vulnerable position of late, based on our Trend Quality Index (TQI). The TQI is based on analysis of sequential trendiness, and ranges on a +1.00 to -1.00 scale. It stands at -.542 and falling as of Friday's close. The attached graphic shows five year history of the dollar index and its TQI. The blue vertical "drop lines" indicate prior occasions of TQI passing downward at Friday's level. Draw your own conclusion.

Dollar trends do affect stocks, although the relationship is not stable over long periods. For the last two years we find dollar fluctuations leading U.S. equities by a week or two (positive correlation; same direction). Prior to that there were some years of negative relation. So the association varies and we can't know for sure what is the active function "today," but last seen it was positive with short lags.

RLC
Hanover NH

November 20, 2006. A week ago we wrote that because our "Tails" paths are elevated once again ... "If the downturn fails to materialize soon (as it did in May), the newly extended Tails could suddenly represent new energy for renewed advance (as in August)." Here we are another week later, with no downturn evident. Both NYSE and Nasdaq Core Indexes are above their EMAs.

I've attached a graphic. It is of the NYSE Tails paths. As with the picture of Nasdaq Tails paths (attached last week), both upside and downside tails are elevated, and (now) diverging with the up-tail stronger. It kind of looks like a new upward thrust developing without having had any correction. That's an extremely odd occurrence, and probably not a worthy betting proposition. But it is what the situation looks like. Sometimes the best bet isn't a good enough bet.

RLC
Hanover NH

November 13, 2006. Our long-anticipated downturn is still just "anticipated." If it's going to emerge at all, it had better emerge pretty soon. The attached graphic shows our Nasdaq "Tails" once again. The Tails measure "length" of the upper and lower limbs of daily cross-sectional returns distributions, and the lines plotted here are smoothed paths of the daily values. Downturn is often in order when the paths are very low (i.e., when the tails are short).

We have already had that low-energy condition, and we have now seen the tails shoot upward...which should be very closely followed or accompanied by downturn in prices. If the downturn fails to materialize soon (as it did in May), the newly extended tails could suddenly represent new energy for renewed advance (as in August). For now, the May downturn scenario remains the far greater likelihood, as with the similarity between the two yellow frames indicated on the graph.

(Other metrics seem to be in a kind of sloppy suspension of progress...much as with the poplar market indexes. Our Core Indexes are split, with NYSE above its EMA, but Nasdaq is below. VIX/DVAX ratio is low, but not as extreme as a week ago. The XUBerance measures have pulled back to neutral and slightly negative.)

RLC
Hanover NH

November 6, 2006. Both Core Indexes, NYSE and Nasdaq, have crossed negative...although not by much. The NYSE Core is a mere 0.04% below its EMA. The Nasdaq Core is a more noticeable 1.13% below its EMA. In the recent context of late-upturning "Tails" series we've noted, and the very low dispersion-adjusted VIX, we're inclined to believe the downturn has finally begun.

RLC
Hanover NH

October 30, 2006. After weeks of repeating the "no change" message (uptrend weakening, but still intact), this week we finally have evidence that change is afoot.

I have attached here a graphic showing the current state of the NYSE "Tails" analysis. As noted here so often, high-energy trends are evidenced by long "tails" of cross-sectional returns distribution. The Tails shrink as the energy dissipates. There is no steady boundary for how high or low the Tails can be, but we have noted for several weeks that the Tails had shrunk to a smallish size in the zone of soon change. Now the Tails have begun to expand (rising on the graph) with some vigor. (Nasdaq Tails are expanding similarly.) Because a whole new burst of upside energy seems quite improbable here, following the recent relentless advance, we have to conclude the change indication is for new direction down.

Tails-indicated trend turns are often not quite synchronous with new Tails bulge, as evident in the attachment. Generally turns seem to arrive within about 0-to-3 weeks. The downside Tail is now 10 days from its recent low at Fridays close, and upside Tail is 7 days from it's recent low. A downside crossover of the Core Indexes below their EMAs will pretty much make the case.

RLC
Hanover NH

October 23, 2006. Market uptrend continues to ripen, and we find increasing signs of change...but still no evidence of actual change. Our 23 daily "state scans" finally come up with no significant positives and one single significant negative. So that's "evidence" of a kind, but pretty slim. Our Core Indexes, both NYSE and Nasdaq are still above their EMAs. The cross-sectional "tails" paths, which we've shown here several times in recent weeks, have stopped declining and begun heading up (at least a little). Newly expanding tails of returns distribution (both positive tail and negative tail) is condition indicative of change.

Historically low VIX levels (option-implied volatility) are getting some attention now. I've attached here two graphics one shows five years of CBOE VIX (along with the S&P for context); the other shows five years of VIX adjusted for dispersion (ratio of VIX/DVAX). Both charts also show prior occasions (verticals) when each series crossed down through the Friday levels. Today's levels are very low either by the "raw" or dispersion-adjusted measure, but the adjusted measure has more stable range over time...providing more prior comparables, and more confidence of knowing when "low" is actually low (not just a new lower-range era). Today the ratio is indeed low. Friday''s ratio (.372) is at the 4th percentile of all daily readings over five years. On average, when VIX/DVAX is down in its bottom decile, one-month forward returns have been negative.

RLC
Hanover NH

October 16, 2006.  We hate to keep singing the same old song...but they keep playing the same old music. Here's part of what we wrote last week

"I have attached an update graphic of the cross-sectional "Tails" series (NYSE). The downside tail of a returns distribution represents a measure of investors selling with great urgency. As evident in the attachment, uptrends seem to thrive in and after a burst of such negativity, and downtrends tend to emerge when the downside tail is minimal. Today the negative tail series remains quite low (although not at lowest lows), and is seems resistant to further decline. When evidence of actual downtrend does emerge, it should be believed."

All that remains true today, with the slight exception that the downside Tail's seeming resistance to further decline turned out to be quite transient. That blue path resumed it's decline almost immediately.

This week I've attached another picture of cross-sectional Tails--this one based on Nasdaq (just for variety). The same condition prevails on Nasdaq as on NYSE. Energy of the uptrend is dissipating (occasional one-day bursts notwithstanding); the uptrend seems nearly exhausted; but a trend is expected to continue until evidence to contrary, and such evidence has not yet arrived.

We've marked the attachment to note the similarity between today's Tails condition and last May's. No two situations play out the same, but the similarly shrinking Tails (both upside and downside) in midst of price advance continues to show similar generality of dynamics...and now at similar magnitudes, too.

RLC
Hanover NH

October 9, 2006. The basic market situation is not much changed; the uptrend is still in place, buy the dynamics get weaker and weaker. One soon day will find the direction has turned. We may now have the beginning of such a turn, as our 23 daily state scans now showing one indicator in significant positive condition and two in significant negative. Our dispersion-adjusted VIX (VIX/DVAX ratio) is now at .377, which is its 15th percentile over ten years. That's low enough to imply negative "expected" returns over a month, but not low enough to be compelling. Both the NYSE and Nasdaq Core Indexes are showing relative weakness (as recently noted here), but remain above their EMAs.

I have attached an update graphic of the cross-sectional "Tails" series (NYSE). The downside tail of a returns distribution represents a measure of investors selling with great urgency. As evident in the attachment, uptrends seem to thrive in and after a burst of such negativity, and downtrends tend to emerge when the the downside tail is minimal. Today the negative tail series remains quite low (although not at lowest lows), and is seems resistant to further decline. When evidence of actual downtrend does emerge, it should be believed.

RLC
Hanover NH

October 2, 2006. The uptrend is presumed intact until we find evidence to the contrary. We see lots of weaknesses, but don't find evidence of turn yet. Prices are kind of "floating" upward, with energy still ebbing. That ebbing energy is evident in the attached update graphic of positive and negative "tails" series of daily Nasdaq cross-sectional returns distributions. The key series is the blue "Ntail" path. It is clearly in low range from which downtrends emerge, as indicated by yellow circles. Perhaps the trend will continue all the way until this Ntail path is at the lowest lows (.040)...but it need not. In any case, the thrill is gone, and the next energetic move is highly likely down.

The second attachment here is the Nasdaq Core index with its EMA. The point in this picture is (again) how much weaker it is in this rally...both relative to the rally in the Nasdaq Composite (lower frame), and relative to the Core's own prior history.

RLC
Hanover NH

September 25, 2006. Market tenor is unchanged; a seriously weakening uptrend is approaching its end. We just don't know exactly where the end will arrive (if indeed it isn't already just arrived).

Two graphics are attached here. The first is an update of the time series of cross-sectional "tails" of daily returns distribution. The outside "reach" of the upside and (especially) the downside of returns distribution provides a measure of the internal energy of a trend. The red path shows the size of the upside (positive) tail, and the blue path shows the negative (downside) tail. You can see in the several intermediate cycles that a typical uptrend has an early surge in both upside and downside energy. Then the energy dissipates as the uptrend continues...eventually just running on fumes as the downside tail shrivels. The blue path has no absolute minimum, but is now clearly low enough to warrant great caution and minimal upside expectation. Although not now as low as it has reached in prior cycles, note that it also started from a higher level this time. (Note, too, that even the red path of upside tails is falling, even as prices advance.)

The second attachment shows our Nasdaq Core Index and its EMA. It is clearly weaker in this uptrend than the broad indexes like the Composite plotted with it here. Moreover, it is very close (0.63%) to crossing its EMA.

RLC
Hanover NH

September 18, 2006. The market is playing out its uptrend pretty much as scripted. Upside energy continues to dissipate, but the trend keeps inching along. Our 23 daily state scans are just in balance, with two significant positives and two significant negatives. Our Core Indexes (both NYSE and Nasdaq) are both positive (above their EMAs) but falling behind the broadly followed averages, which offers good possibility the Cores will turn negative pretty early in the next downturn. We don't expect this will take long now.

We find equity exposure among Managed Futures funds has been mostly unresponsive to the uptrend...as noted this past Friday by Walter Deemer in his Market Strategies And Insights ("It's Not Supposed to Be That Way," Sept 15). Well, change that "has been" unresponsive into "had been" unresponsive. In this week's data update (computed after Deemer's citation), we finally do see the equity exposure series pop upward. It's still not nearly "extreme" in our view, being at 58th percentile of its daily history. Historically, 80th (and 20th) percentile is a useful threshold of extremes. But the source data for these exposure metrics always arrive with a few days lag, so the little spurt at the end of the posted series may (may) have spurted still further in the few latest days. You can see the exposure path and it's latest posting in the attached graphic.

RLC
Hanover NH

September 11, 2006. The energy of the uptrend continues to dissipate, but is probably not yet spent. Our 23 daily state scans find two in significant positive state, and one in significant negative state. That kind of mixed condition doesn't give much likelihood one way or the other. But continuation is usually the better trend bet.

Friday's supposed rally did not provide much encouraging data. For example, our NYSE Core Index was actually down for the day (-0.34%). Also, it also crossed down through its EMA...just barely. [Graphic attached.] The more critical Nasdaq Core Index was up along with the conventional indexes, however, and remains 1.94% above it EMA

RLC
Hanover NH

September 05, 2006. The same conditions of a week ago still prevail today. We said the present trend is up, and it remains so; we have said the primary energy of the trend has peaked and is dissipating, and that dissipation continues.

What we don't know, of course, is just when the uptrend energy will finally be depleted. One useful measure of "energy" in a trend is the size of the "tails" of cross-sectional returns distribution. High-energy markets generate long tails. The attached graphic shows how upside and downside "tails" of Nasdaq cross-sectional distribution have varied over the past two years. (The NYSE tails we sent a couple of weeks ago are directly analogous...both in construction and in current condition.) Note how both the upside tail and downside tail have been contracting (falling) after the initial August thrust.

When the downside tail in particular gets too low (downside energy, or selling urgency, is just all wrung-out), it's time for a turn. Several such occasions are apparent over the two years shown. So the question for this metric is this How low is too low? In July 2005 the low was at .040 (18% below current). In June 2006, the low was at ,044 (10% below current). At the current rate of descent from the August high (a19.8% decline in 16 trading days) Those respective levels would be reached in 8 or 15 days. Alternatively, if the recent "higher low" leads to a even "higher low" this time, we might be there in a day or two.

That's a wide range (maybe 2 days, maybe three weeks). We can't predict such turns ahead of time. But sometime in soon weeks the tails will make their lows, and if we find actual direction is changing (for instance, the Core Indexes turning downward) as the cross-sectional tails expand, the stage will be set for notable weakness. But for right now, as best we can see, the trend is still up.

RLC
Hanover NH

August 28, 2006. Last week we noted the upside and downside "tails" of cross-sectional returns had peaked, observing that for prior occurrences, "...once the paths of the two tails peaked, the major energetic thrust was over, but the advance itself had further to play out." That observation still holds, but the "play out" so far is even less energetic than one might expect. With much less energy than usual, one might expect the play-out will also be briefer than usual.

Meanwhile, The Core Indexes are suddenly showing relative weakness. A picture of the NYSE Core is attached, and you will see that the Core has not nearly progressed with the S&P in the past month's rally. (The Nasdaq Core is similar.) Note, too, how different is the Core behavior today compared with the advance that started back in October 2005. That advance took the NYSE Core up 38% while the S&P rose 13% (October-May). Some of the gap reflects different volatilities (1.5:1), but most of the gap caused by the Core Index's greater "regularity" of progression through the advance. In the past couple of months, however, that former dogged upward bias is not apparent. From the June lows, the NYSE Core is up +7.7% compared with +5.9% for the S&P. Adjusted for volatility, the Core is actually falling behind.

Both NYSE and Nasdaq Cores are still in uptrend, however...even if weak uptrend. And we generally find it pays to not anticipate a new direction before it begins. So the outlook here is great caution, while still "expecting" further advance (referring more to mathematical expectation than actual anticipation).

RLC
Hanover NH

August 21, 2006. Commentators are noting a lack of volume in the past week's otherwise frisky advance. That's a valid concern, but probably premature. There are other ways to gauge the energy embedded in a move, and one of ours is to measure the length of the "tails" in cross-sectional distribution of daily returns. (This is obviously related to our DVAX dispersion metric, but in this case we just focus on the tails themselves, not the overall distribution).

I've attached a graphic tracing the paths of the extreme upside and extreme downside tails of NYSE returns distribution. Both the upside and downside values have just recently posted highs, and begun to retreat. Two very similar prior occasions (and one "sort of similar" occasion) are marked. In each of these cases, once the paths of the two tails peaked, the major energetic thrust was over, but the advance itself had further to play out.

RLC
Hanover NH

August 14, 2006. Market state remains unclear. Our state scans show two metrics significantly positive, and two significantly negative; volatility measures are middling (with the recently remarked VIX/DVAX no longer historically low); cross-sectional skew is mostly negative, but not remarkably so.

We've noted a few times that the declines from May highs has pushed several data series down to levels where rally would be "normal." The attached graphic shows an example, based on the Russell 2000 "Momentum" series (a double exponential of log returns). Four occasions of prior crossings of current Momentum level are shown with "drop lines" pointing to the R2K index itself. Three out of four times, current levels occurred just at the end of the downtrend. The fourth time, however (July 9, 2002), the index would shed 15% in the following two weeks...and 24% to the eventual bottom (The situation is not limited to the Russell small caps, although this index shows a cleaner example than most.) The inference here is that we may be quite close to readiness for upturn...but that such upturn needs to give some evidence of actual emergence before it permits confidence.

RLC
Hanover NH

August 07, 2006. Not much new was revealed last week in terms of market state and market tone. In fact, our daily state scans (which have maintained a distinctly downbeat balance of 3-or-4 significant negatives through the saw-tooth rally of the past two weeks) have finally "relaxed" to showing only single significant negative state. That is, of 23 market state metrics, almost all are at levels (or rates of change) that are not statistically associated with notable market direction in past cases. Perhaps that's just a confirmation that "everyone" is waiting for Fed rate policy on Tuesday.

Barron's readers may have noted Mike Santoli's mention (The Trader column) that our equity exposures analysis finds hedge funds at their "lower third of their range." Just to clarify here, that's not for hedge fund universe, but for funds in the Managed Futures hedge funds index. These funds are demonstrably more aggressive and more labile than most. In fact, the "lower third" exposure is not merely more negative than average, but in fact net short in our calibration. I've attached here a graphic showing the equity exposure path so you can see for yourself.

The three boxed areas show periods of equity exposure (2004, 2005, 2006) declining into net short territory. The two "red" boxes are occasions (including now) selling into decline. The yellow box shows selling into a rally. In our far longer experience with tracking mutual fund exposures (which are almost never short, yet sometimes vary notably), we find the "contrarian" interpretation is generally valid in the former (red box) case. In the latter (yellow box) case, contrarian inference is often wrong. This has been borne out in the Managed Futures cases shown.

Thus, the current situation would seem to be building toward an upturn. We say "building" because the net short reading seems not yet very deep.

RLC
Hanover NH

July 31. Directional metrics here remain split. The Nasdaq Core Index crossed positive over its EMA on Friday (joining the NYSE Core which crossed on Tuesday), and Momentum paths remain at depressed levels from which renewed advance is common (although not compelling). But cross-sectional skew has been net negative for many weeks, and dispersion-adjusted VIX is very low.

Speaking of dispersion, one metric that is not ambiguous is our DVAX index of cross-sectional dispersion. DVAX is quite high...now at levels not seen since the spring of 2003. The clear implication is for increased volatility. In fact, actual volatility is already up a good bit (although not yet as much as the DVAX suggests is coming), even while option-implied volatility (VIX) is only middling so far (hence the low "adjusted VIX" cited above). Options volatility will typically follow shortly (and higher actual volatility, too). A five-year plot of the DVAX series is attached here, showing the recent runup.

RLC
Hanover NH

July 24. Apparent trend is still down, last Wednesday's Bernanke testimony and ensuing one-day rally notwithstanding. Our NYSE Core index did cross above its EMA for that one day--then back below--but our Nasdaq Core Index (which has far the better long term consistency of the two) never did cross positive, and at Friday's close is more deeply negative than before Bernanke's remarks.

A number of momentum and distension measures are now down into "normally" reversible range. So the question, which we can't answer, is whether this down trend will prove to be "normal" or not. Barron's has a notably upbeat cover story in today's issue, arguing that market value in general is pretty attractive here. Meanwhile, other articles in the same issue point out potentially dire housing and structural financial strains. So it seems that, like the market's internal stats, the external economic background is also saying this is a plausible point for renewed advance...if events will just proceed "normally."

This dichotomy is plainly evident in comparing mutual fund and hedge fund equity exposures. I have attached two graphics here one shows the statistically inferred equity exposure path in our mutual Growth Funds index; the other shows comparable exposure path among Managed Futures Hedge Funds. The red rectangles highlight the diametrically opposing tactics in the down trend since May. The mutual funds are plainly betting this is a buying opportunity (highest exposure in years), while the hedge funds have gone short. (In truth, most hedge fund categories are not net short. The Managed Futures funds are plainly more aggressive and more directionally active than most.)

For our part, a trend is presumed to continue until evidence shows otherwise.

RLC
Hanover NH

July 03.  Last week we noted "it's not hard to find various market measures holding up manfully," but also "current trend is presumed down." That downward presumption dissolved in last Thursday's Fed-relief rally, with our NYSE Core and Nasdaq Core Indexes both moving up through their EMAs.

Also notable Thursday and Friday is that the ratio of actual to implied volatility suddenly jumped. Most of the time, actual S&P volatility is less than VIX (ratio runs about 0.60 to 0.80), but shot up to1.24 and 1.18 on Thursday and Friday respectively. Readings above unity are often associated with upturns lasting for more than a few weeks. Notable exceptions occurred in 2001-2002 bear market. (We intended to include a graphic illustration on this volatility ratio, but technical problems prevent transmittal for now. Please advise if you want a follow-up transmittal when available.)

Not that everything's rosy. Our daily state scans still show three significant negatives as of Friday's close. That's down from six a week ago, however, and may well just be slow responding.

In last week's remarks about high equity exposure among mutual funds, we noted the appropriate inference depends on largely on the market's next ensuing impulse. If down, the high exposure can readily produce a rapid (and damaging) exit; but if up, the exposure can get worked off in rising prices. With the most current trend now presumed up (relying on the Core Indexes), the potential high-exposure vulnerability seems likely deferred.

RLC
Hanover NH

June 26. The key question--as to whether the market highs of mid-May marked the end of that bull market--is yet unanswered. We've observed longstanding weaknesses for months, so it would be easy to infer the recent declines are "obviously" the manifestation of those accumulated weaknesses...and destined to play out further down. On the other hand, it's not hard to find various market measures holding up manfully. (Among our own metrics, we have the Cumulative Skew still near its highs, and both NYSE and Nasdaq data show daily upside outliers still running ahead of downside.)

One interesting and notable development in the recent market weakness is the vigorous buying evident among mutual funds. Interesting and notable...but with implications also still ambiguous for now. The graphic attached here shows the equity exposure path among LargeCap Growth funds. Exposure increased dramatically as prices fell, and now stands at levels not seen since 2004 (briefly) and 2000 (massively). Those two situations turned out so differently, in our view, based on whether the optimism was promptly rewarded. That is, a simple "contrarian" interpretation doesn't work. Overexposure is not always a negative (see 2004), but turns very negative when the excess exposure starts losing (as in 2000). Current high exposure isn't confined to the Large/Growth sector.

Meanwhile, current trend is still presumed down, with both NYSE and Nasdaq Core Indexes below their EMAs, and six of our daily state scans showing significant negative condition at Friday's close.

RLC
Hanover NH

 June 19. Almost anything to be said about current market would repeat things you already know. Volatility is clearly up, and not likely to go back down. (In fact, fairly likely to go up further.) With increasing volatility, after the relentless May-June sinking spell, there's a reasonably good chance of fierce rally (as we saw last Thursday). Nevertheless, the presumed present trend is down, and trying to anticipate and bet on sharp counter-trend bounce is difficult and dangerous.

Two attachments illustrate the conflicted situation. Our Nasdaq Core Index can only be interpreted as in a downtrend state. The Index is below its EMA (our standard directional criterion), and it has also made a recent high lower than the previous high and a low lower than previous low. The Core Indexes are not "predictive," except that they are highly trendy...meaning that they are always more likely to continue than to reverse. (The NYSE Core Index is similarly in downtrend condition.)

The second attachment (NQ-d1d2) traces out prior occasions over five years when our Nasdaq "Momentum" and "Trend" series were at same levels (and moving in same direction) as today. There were seven occasions in the past five years when one or both of these measures were where they are today. (The color coding of the 2001 occasion is obscured because both red and blue conditions occurred at once...just before the World Trade Center was hit.) For six of the seven occasions, the outcome was an imminent rally, either for a couple of weeks, or several months. In the seventh occasion, however, prices were only half way down a horrific slide that gave no safe exits. And that is why it's so risky to bet on counter-trend rallies.

RLC
Hanover NH

June 12. Market state shows not much new. Our expected rally continuation did not continue, as you may have noticed. So we see we should have been paying more attention to the four state scans in significant negative condition last week. Every once in a while we get into overlaying a bit more judgment on top of the actual evidence.

At this Friday's close, the daily state scans now show one positive and three negatives. So that's more mixed than a week ago, but negative enough. The Core Indexes are of course still below their EMAs, and so presumptively negative. We note, however, that the Cores are showing a slight relative strength. A graph of the Nasdaq Core and Nasdaq Composite is attached, and you can see that while the Composite has fallen below its May 24 low, the Core has not. This preliminary strength is not actionable, but may build into an opportunity.

RLC
Hanover NH

June 05. Last week we wrote "Some further rally seems likey," This week we repeat that view, although the likelihood has become pretty marginal. As of Friday's close, our 23 daily state scans produced four scans in significant negative condition; none in significant positive condition. Unless that situation is soon relieved, this zig-zaggy rally will end up as nothing more than a choppy selling opportunity.

Our Core Indexes (NYSE and Nasdaq) both remain below their EMAs, but only by 0.08% and 0.37% respectively. Both indexes rose on Friday by 0.57% while most broad indexes were essentially flat. So there's not only a relative strength in the Core Indexes, but a fair chance they'll carry forth for a positive crossover. Crossovers themselves are not "signals," because there is no magic. But statistical history shows that being on he wrong side of the Cores is a bad position on average. A graphic is attached, showing one year of the Nasdaq Core and its EMA.

Also attached is a picture of our Fixed Income exposure path for Directional/Tactical hedge funds. You may recall that these exposure paths emerge from statistical inference of daily fluctuation. The graph shows these hedge funds to be deeply short in bonds (or other fixed income instruments), suggesting potential fuel for further bond market gains.

RLC
Hanover NH

May 30. Last week we were expecting some prompt rebound, based largely on the Core Indexes' deep deviations from their EMAs. Some rally materialized Thursday (and Friday), just slightly less promptly than envisioned. Some further rally seems likely, although prices are down about a percent at this writing Tuesday morning.

Various statistics here are still negative, however. Both NYSE and Nasdaq Core Indexes remain below their EMAs, and our 23 daily state scans show three in significant negative condition at Friday's close.

One interesting metric that seems utterly oblivious of the May declines is our index of cumulative skew (CSQ) in the S&P 100. It posted a new high on Friday, as shown in the attached graphic. This index is not a predictive device, and in fact it sometimes turns late. But the continuing positive skew (upside deviations from mean bigger than downside deviations), in the face of several weeks of pull-back, indicates there has not been a shift from urgency-to-buy toward urgency-to-sell.

RLC
Hanover NH

May 22. After all the up-and-down crossovers of our Core Indexes, it's a relief to finally have a new direction "hold" for a bit. At Friday's close, the S&P 500 is off 4.4% from its high (and off 3.0% from the NYSE Core crossover on May 10). Nasdaq Composite is off 6.4% from its high (and off 5.4% from Nasdaq Core crossover May 10). Not only is the direction clear, but volatility has increased notably, just as we expected.

Two questions emerge from any sharp drop; (a) Is this a major downturn? and (b) When to expect an oversold rally?

Our answer to the first is We don't know. Aside from the fact that we never know when a direction will be enduring, even uncertain estimation must await the next rally. It had better show some good vigor if it's going to undo the current weakness.

So are conditions ready for at least an oversold rally? Yes, we think so. Attached graphics show the NYSE Core Index deviation, and the Nasdaq Core Index deviation, from their EMAs. Each graph has a red horizontal drawn back from the latest posting, with verticals dropped from each prior occasion of deviation upticking at same level as today. Three facts are apparent. First, that current negative deviations are about that the normal extremes. Next, that the market most usually rallies at these extremes. Finally, that the only occasions when current levels of negative deviation did NOT precipitate prompt rally were occasions of profound and extended weakness...that is, in the midst of heavy duty bear market. We conclude that prompt rally is odds-on bet. (Conversely, of course, in case such rally does not materialize here, it may imply some much more profound weakness.)

RLC
Hanover NH
P.S. Due to technical issues, the Prime Timers page at web site will not be updated until Monday.

 

May 15. The Nasdaq Core turned negative (below its EMA) Wednesday, with NYSE Core following Thursday. We've seen lots of whipsaw reversals in these indices recently, but other metrics, aimed at gauging the "thrust" of the Thursday-Friday declines, now suggest it's quite unlikely last week's drops would simply end here (although at least some soon bounce would be in order).

A picture of the Nasdaq Core and its EMA is attached. As you can see, last week's drop (5.2% from Monday's high) is still within fairly normal range of fluctuation. And cross-sectional skew is still neutral-to-positive. So we do not have current evidence of a substantial down turn here...only that it's a downturn with probably more to go.

[Up or down, it still looks like our expected increase in volatility has now begun. Until a week ago, daily standard deviation for the S&P 500 had been running about .0050 (5/10%), with median change about .0040 (4/10%). At Friday's close we have standard deviation 1/3 higher, at .0067. Volatility has persistence, so look for larger daily changes than we've been used to of late--both up and down.]

Oddly, our daily state scans are balanced, with three significant positives and three significant negatives.

RLC
Hanover NH

May 01.  No clear directional indications yet. Our 23 daily state scans continue fluttering back and forth with one or two significant positives...or one or two negatives. The Core Indexes turned just barely negative (below their EMAs) on Thursday, only to turn around and cross back to positive--by an even teensier margin--on Friday. Subjectively, the Core Indexes look relatively strong, as shown in the attached graphic for the Nasdaq Core. Each of the recent minor pull-backs in the Composite has been scarcely even detectable in the Core.

This week's Barron's (page M2) notes our finding this week that dispersion-adjusted VIX has dropped to levels usually associated with market weakness. The dispersion adjustment is just a ratio of the VIX to the DVAX dispersion index. Longtime readers will recall that our DVAX simply measures how widely daily returns are spread out. Very large dispersion obviously means that stocks are changing price in large amounts. That is, a high dispersion day or era is a high volatility day or era...even if the individual large price changes may happen to cancel each other out in the averages. So the VIX/DVAX ratio is just a way of scaling implied volatility in terms of the market's actual internal volatility. The scaling matters because raw VIX is notoriously non-stationary for judging what values are high or low.

Anyway, at Friday's close the VIX/DVAX ratio stood at .294. This is the lowest ratio since March 10, 2000. This comparison does not warrant an inference that April 2006 will play out "like" March 2000. It does warrant an inference that one dimension of market tone is showing similar weakness.

RLC
Hanover NH

 

April 24. Another week of data in hand, and still we see great uncertainty as to when the uptrend will end. Our daily state scans say it may already be done, with two metrics showing significantly negatives. But we've found a small handful of significant negatives (intermittently) for weeks. And both NYSE and Nasdaq Core Indexes are now back above their EMAs, for presumed uptrend. Various other metrics are also at sixes and sevens; our aggregate Cumulative Skew has been posting highs (confirming uptrend), but a good part of that progress is due to extremely atrophied downside limb of cross-sectional distribution (implying end of trend).

So our outlook remains guarded...but probably, sort of, maybe more likely to progress further. With luck.

One manager who seems to have resolved his own uncertainty about emergent direction is Bill Miller of Legg Mason Value Trust. This fund is one of our five gurus whose equity exposure is tracked weekly at our Prime Timers page of the web site. This past week the LMVTX fund's inferred exposure dropped from its 71st to 27th percentile. (It's still measured at 1.00x market, but for this fund, that's a defensive posture.) Miller's retreat shifts the balance of the 5 funds to three below-median and two above-median. A graphic of the LMVTX exposure path is posted at the site.

RLC
Hanover NH

April 17. Stocks still seem to be in midst of a stop-and-start slow-motion roll-over. Indeed, the slow-motion seems to get slower. Although prices were down on the week just past, very little downside energy was apparent in the data, and our XUB indexes of NYSE and Nasdaq exuberance remained positive in three out of four days. Our daily scans still show on-balance negatives, and our NYSE and Nasdaq Core Indexes turned marginally negative on the 10th (below their EMAs).

Portfolio managers have begun to lighten up on stocks. Among Long-Short hedge funds we find equity exposure remains high (net long), but now a bit off from their highest exposures (in February). Among Growth mutual funds, equity exposures also remains high, but a bit further off from their high exposures (in March). Among our Prime Timers posted at the web site, outlook is still split (three with high exposure and two low).

Speaking of change in exposures, you may have seen mentioned in this week's Barron's ("Up and Down Wall Street") that we find Managed Futures hedge funds are suddenly re-entering long positions in crude. The increase sharp, but observable exposure is nowhere near the highs back in September, so there's lots more room if there's another run. I've attached a graphic of the Crude Oil exposure path..

RLC
Hanover NH

April 10.  Market direction remains uncertain, even after Friday's sharp drop. In fact, our count of daily state scans, which in recent weeks has been consistently showing a negative balance as the market wiggled higher, on Friday's decline has finally clammed up entirely; no significant positives, no significant negatives. Neither NYSE nor Nasdaq Core Index has crossed down through it's EMA, but now are now close enough to do so readily.

We've written often about "underlying weakness," and one new indication of weakness has emerged this week. The ratio of CBOE VIX to our DVAX index of cross-sectional dispersion has now dropped to a level associated with imminent downturn. Everyone knows about VIX, of course. It's been quite low for months, which is taken as a weakness itself. Trouble is, VIX has lengthy eras of prevailing high or low values. Because we never know what is the current prevailing norm, we never quite know whether a given reading is high or low relative to that norm.

VIX norms are generally stable, however relative to dispersion. The ratio still fluctuates widely, but the range of normalcy is stable. (The ratio's stability is entirely expectable in terms of statistical mechanics. When the market is "drawing from" a wide-returns distribution, sample means can deviate widely and volatility will be high; when "drawing from" a tight distribution, sample means will deviate little and volatility will be low.)

In Friday's sell-off, VIX jumped 8.8%, as it does when prices drop. But DVAX dispersion jumped 12.6%, so the VIX/DVAX ratio declined. And it declined to a very low level (.3712)....its 3rd percentile of daily values over five years. I've attached a graphic showing the daily series, along with the S&P for market context. The vertical lines identify prior occasions when the ratio fell to the present level. (The vertical at January 2004 is "disconnected" at the ends to recognize that the ratio did not quite reach Friday's actual level, but held immaterially above it.) Note that although these levels are clearly associated with intermediate tops, specific timing easily varies by several weeks. Hence we call it an "underlying" weakness...even though it's sometimes pretty timely.

RLC
Hanover NH

April 3. A week ago, we said first "Just as things seemed finally to be rolling over toward a comfortably confident downturn, our stats have started to suggest all is flux, and another upward try may be developing." And we said last "Evidence still shows vulnerability, but the actual turn still deferred." Those inconclusive observations remain valid, but perhaps with greater near-term possibilities today. Our daily state scans are "improving," in the sense that those significant negatives have been slowly dissolving...from 5 negs two weeks ago, to four last week, and only one at last Friday's close. And both NYSE and Nasdaq Core Indexes are clearly rising. The Nasdaq Composite has indeed exceeded its January highs (albeit without the attention we posited). None of this obviates the broader vulnerability, but does seem to indicate some further advance likely.

We're adding Vanguard Asset Allocation Fund to the Prime Timers page at the web site. This fund has an excellent exposure history, as evident in the attached graphic. Note the dramatic exposure reduction in 2001-02, and the almost precision re-entry in 2002. The reduced commitment in 2004 was not as productive, because the correction was not severe, but the timing of the reduction was well justified. The fund is fully committed today, as you can see.

RLC
Hanover NH

March 27. Just as things seemed finally to be rolling over toward a comfortably confident downturn, our stats have started to suggest all is flux, and another upward try may be developing. The Nasdaq Core Index has now followed the NYSE Core Index, crossing upward through it's EMA. I've attached a picture of the Nasdaq Core, showing how it has simply refused to do anything but float higher and higher, even as the Composite has been consolidating for the entire quarter. If the Composite ends up complying with the Core, it would likely attract notable activity as it finally posts new rally highs (past 2331).

However, we still (or rather, again) have four of the daily state scans in negative condition, and none showing significant positives at end of week. But positives popped in and out on Tuesday and on Thursday during the week, so the 4-to-0 end-of-week count may be a lot less "settled" than the count appears.

As noted in today's Barron's, our hedge funds exposure tracking finds Directional/Tactical funds in a "high-neutral range" of equity commitment, which is down and declining from a couple of months ago. A separate hedge fund set (U.S. Long/Short) is still just plainly high, at it's 86th percentile of exposure (but on a briefer historical base). We also still find mutual funds at well above-normal exposures. So there's room for possibly a good deal of selling whenever that mode finally does emerge.

So the bottom line is that there's not yet a clear bottom line. Evidence still shows vulnerability, but the actual turn still deferred.

RLC
Hanover NH

March 20. Overall market state is virtually unchanged from last week. NYSE Core Index remains positive, and Nasdaq Core remains negative (respectively above and below their EMAs). Daily significance scans show five negatives (versus 6 a week ago). The up-week did undo two fixed-criteria indicators (PDQ and Curve Shift) from negative to neutral.

Although it's possible these small changes (from 6 to 5 negative scans, PDQ and Curve Shift going neutral) are indicative of a firming preparation for renewed uptrend...it's a stretch. More likely we're just in midst of a slow-motion expiration of the uptrend. The attached graphic shows five years of the Russell 2000 along with a "Momentum" series (which is a double EMA of daily log relatives). The horizontal blue line traces backward the current level of the Momentum series, showing vertical drops at each prior occasion of passing upward through the current value. (At the 2002 and 2005 verticals, the Mom line didn't quite make it, but are so close as to be indistinguishable.) Four of the five occasions are either right at an intermediate top, or (in 2004) within a few weeks. The only exception is at May 2003, which was at the vigorous upthrust from the bear market lows of 2002...a plainly different situation from today.

Nasdaq Composite is alone among the major market indexes not posting new highs. At Friday's close it's 1.08% shy of its January high. That's less than two average daily changes (recent average running about 0.6%). If Nasdaq does join in with new recovery highs, that could generate popular excitement. Other than that, the plain likelihood is continuing sloppy rollover.

RLC
Hanover NH

March 13. Friday's runup looked pretty good. Taken by itself, it would suggest likely follow-through...which would new highs and new headlines (and all that). But most of our metrics look not so good. Both NYSE and Nasdaq Core indexes are below their EMAs, and our daily state scans turn up six metrics in significant negative condition, and none in significant positive conditions. (For a given metric, "significant negative condition" means that on prior occasions at similar values, average five-day outcomes differed from outcomes of "all other" occasions by a margin that is statistically significant.) Lopsided significance like 0-to-6 can sometimes be counterindicative (so negative it's a positive), but it's not a good bet.

Hedge funds and mutual funds are pretty heavily committed long in equity. We do have one index of U.S. Long/Short hedge funds that is only about at median exposure, but another U.S. Long/Short sample is about at its highest exposure on record (last three years). Non-U.S. Long/Shorts are also at high equity commitment. As for mutual funds, I've attached a five-year exposure path of Large/Growth funds. They're at 1.20x market at Thursday's close, which is at its 98th percentile of the past five years. High exposure isn't bad when it's built on a correction; but when it's built on upward trend-chasing, it's dangerous. Other style sets are not quite so extended as the L/G funds, but they're all showing above-median exposure.

RLC
Hanover NH

March 6. We must be getting somewhere near the end of this flaccid, trendless, market hiatus, because the lack of tone is now becoming widely remarked. Interpreting current market state is more than daunting when the predominant state is pretty much...stateless!

Take a look at the attached graphic of our Nasdaq Core Index; it's pretty clear the drift is still upward, but also pretty clear that drift is lacking any effective punch. The index just can't seem to get enough energy to deviate from its EMA. And it's deviations that represent an energetic market.

RLC
Hanover NH

February 27. Probabilities for immediate direction remain almost impenetrable. Our Core Indexes are back to positive, while the daily state scans at Friday's close find a single metric in significant positive and three negatives.  So the short-term outlook is mixed...which is to say, there isn't much short-term outlook.

Last week we noted the above-normal equity exposure in Long-Short hedge funds and growth-oriented mutual funds, with negative (but not necessarily imminent) implications. This week we point to the TXP indexes. Longtime readers will recall that our TXP model not only tracks average Tech exposures, but also measures cross-sectionally whether the better Tech exposure timers or the lesser Tech timers are the more aggressively positioned. That balance is captured in a series called the TXP Slope. It happens that the TXP Slope has now crossed from positive to negative, meaning the lesser timers now have the greater Tech exposure. A graphic is attached, which also indicates two prior occasions of such crossover. (Crossovers notwithstanding, this is not a specific "signaling" device; it's a Who's doing what? metric.)

You may have noticed that Barron's this week remarks in The Trader column our recent study of changing market behavior The Death of Momentum. Bottom line is that price momentum is getting scarce, and strategies that rely on momentum need to be selective for vehicles that actually demonstrate trendiness. The study is posted at the web site, with links both on the home page (near center) at in the Analytica section of market essays.

RLC
Hanover NH

February 20. The week ago negative crossing of the Nasdaq Core Index (Feb 10) turned out short-lived and unproductive, re-crossing positive (Feb 15) with the Composite having advanced 1.5% in the interim. The NYSE Core It's now still positive (above its EMA) at Friday's close, by a slim 0.34%. Meanwhile, our NYSE Core Index remains below its EMA, by an even slimmer 0.12%. Our daily state scans currently show one metric in significant positive state, and two in significant negative state. Conditions are mixed.

A number of more general indicators seem to be pointing to danger. Average equity exposure among Long-Short hedge funds is high, as is average exposure among growth-oriented mutual funds. Although high and rising equity commitment is not always a "contrary" indicator (it sometimes indicates healthy shared confidence), it usually spells danger when it emerges in midst of rising prices. With the DJIA within 2% of its all time highs, and the Value Line (geometric) within 0.7%, we have to believe the supra-normal fund exposures present danger here. Although immediate conditions are indeed mixed (and changing somewhat more than usual), so immediate trend is quite uncertain, the next consequential trend seems likely to be down.

RLC
Hanover NH

February 13. Market direction may be finally at hand. The NYSE Core Index crossed negative on Tuesday, and the Nasdaq Core followed on Friday. Meanwhile, our daily state scans have turned up a growing number of significant negatives, from none on Monday to five by end of week. There's also one significant positive (just to guard us from complacency), but from here the odds have to favor downside.

A graphic of the Nasdaq Core is attached.

RLC
Hanover NH

February 6. Alas, we have yet another week with no definitive directional state. At Friday's close we find no significant positives and no significant negatives across the 23 daily state scans. Our NYSE and Nasdaq Core Indexes both remain above their EMAs, however, so we'll continue to presume the uptrend remains in place.

Our hedge fund exposures paths show equity commitment is pretty heavy, with the equity Long-Short funds at or near their highs. (Histories are a bit limited there, however.) The broader "Directional/Tactical" category (which subsumes some of the Long-Short) also shows above-average equity commitment, but not nearly at historical extremes.

Equity exposure among mutual funds is also pretty high. Large/Growth funds in particular are at about 115% (reflecting a bias for volatility, obviously, not actual financial leverage above 100%). That is about at the 85th percentile of normal exposure range..."normal" meaning excluding the outlandish exposures taken on in the Tech bubble.

Two notable funds showing specific under-exposure are the Hussman Strategic Growth fund and Ned Davis Asset Allocation fund. Hussman is an explicit timer, and has been steadfast at zero equity exposure for a few months. The Davis fund is a bit more like a traditional allocator. It has also been at below-average exposure for months, but shows sharply reduced exposure just in the past two weeks, now at the lowest apparent equity position since inception (April '03). Graphic exposure paths for both of these funds are posted at the Prime Timers link at the CarpenterAnalytix web site.

For an alternative equity outlook, take a look at the exposure path for Vanguard Asset Allocation Fund. A five-year graphic is attached here. This fund also has a very strong record of equity timing (high exposure in market strength and low exposure in market weakness)....quite impressive for a firm that officially eschews timing strategies. And the fund is clearly carrying high exposure these days; an unblinking 100%.

RLC
Hanover NH

January 30. Last week's state scans found four significant positives and only one significant negative at Friday's close, and we noted both NYSE and Nasdaq Core Indexes remained positive (above their EMAs), so in spite of the market's drubbing of January 20, the numbers forced us to presume upside continuation. And upside continuation is exactly what unfolded, with broad indexes up 2% to 4% for the week (with smaller-caps getting the bigger push).

This week we're back down to a much thinner level of presumption. The Core Indexes are still positive, but at Friday's close the daily state scans are entirely neutral (no significant positives, no significant negatives).

One somewhat broader constructive observation is that the CSQ index of cumulative cross-sectional skew has posted new highs. The two attachments here show (a) the cross-sectional distribution of OEX component returns for this past Friday Jan 27 with notably positive skew, and (b) a plot of the OEX cumulative skew (CSQ) over the past five years. Distributional skew matters because it shows which way the more dramatic action is running. The accumulated CSQ path is very "trendy" and rarely changes direction unless there is a new trend afoot. This series gave cause for concern in October-November 2005 when it posted its largest correction since the 2003 lows, right in the face of market advance. But CSQ path did re-reverse this time, and for right now is posting new highs.

RLC
Hanover NH

January 23. With all our fretting over "underlying" weaknesses in the market, one might think that last week's decline (especially Friday's relentless drubbing) would tip the probabilities clearly to downside expectation. But no. In fact, we suddenly find more statistical positives than negatives. Near-term positives, but we live in the near-term.

Both NYSE and Nasdaq Core Indexes remain above their EMAs. The 23 daily state scans at Friday's close show four significant positives and one significant negative. Last week's DejaVu attachment of a "tight" scan on Nasdaq tails condition implying a negative and significant 2.63% expected value for the week (turned out -2.99%) has dissipated entirely.

Always trying hard just to paint-by-the-numbers, we have to say the current stats favor upside. If, on the other hand, last week's ominous finish turns out to have more import than is currently reflected in the stats, it will likely unfold with lots of downside to take advantage of in due course.

RLC
Hanover NH

January 17. Our presumed uptrend has been uptrending, but the stats are still not giving confident inference. Rolling along on mere presumption is uncomfortable, but that's still where we are. Both Core Indexes remain positive (above their EMAs) but our daily state scans just don't seem to be able to find much confirmation in the data histories; only a single scan of the 23 is showing significance (1-positive, 0-negative).

A strange contradiction shows up in current DejaVu screens of Nasdaq cross-sectional distributions. In general, trends are notably positive when the extreme positive tail of returns is "growing" and extreme negative tail is "shrinking." That's a well-established connection, and it makes sense. And, it describes the current situation. But we also find at this time that a very "tight" screen on current specific values for the same variables (rates of change of the tail growth and shrinkage, rather than merely growing or shrinking) gives a very different result; a statistically significant loss of -2.63% over ensuing week. The statistics are shown on the attached page of DejaVu output. Although the number of prior occasions (25) may seem quite limited when considered isolation, given the magnitudes (large) and the variance (small), the significance test says 25 cases is sufficient to be taken seriously.

So we continue with a weak presumption of continuing uptrend, but with conflicting evidence preventing any real confidence.

RLC
Hanover NH

January 9. Last week we noted the uptrend was tenuous, but still presumed to be in fact an up trend. The ensuing week has confirmed that presumption. Needless to say, market measures such as our Core Indexes are now comfortably ahead of their EMAs, and indeed at new highs. (Well, it's not quite "needless" to say, but of course we'd be vocally alarmed were it not so at this point.)

What's more interesting (at least potentially) is that there are now some signs of uptrend renewal. For instance the concern about the weakened Cumulative Skew index (CSQ) may turn out to be repairing itself. This series is an accumulation of daily cross-sectional skew across the 100 OEX components. Back in November it went into a more prominent decline than any time since the market bottom. But right now it's heading upward in the market's new year revival, and has made up much of the prior weakness. If it gets to post a new high, we'd probably infer the prior weakness was canceled. But that's getting just a bit ahead of the evidence. For right now, we simply have a current uptrend that is still an uptrend.

A graphic of the CSQ is attached, as well as a look at the one week OEX cross section of returns. Note the overall positive skew of the distribution, even if the outlier (Viacom) were held out.

RLC
Hanover NH

January 3 (2006). With a brand new year underway, it feels like there should be lots to say. But no. Indeed, there seems to be less than usual of note. One reason is that these holiday weeks actually generate less information. Not only are there fewer days of trading, but the days available are thinly traded, with some markets (bonds) closing early and others just lethargic. For example, last Friday's cross-sectional dispersion had one of the smallest standard deviations on record (0.70% for the S&P, 1.1% NY overall).

For what it's worth, the our Nasdaq Core Index remains above its EMA...barely. A graphic is attached, and you can see that the Core and its EMA are so tight that the balance would be indeterminate if the numerical values weren't also shown. The less reliable NYSE Core is slightly below it's EMA. The daily state scans have more positives (3) than negatives (1), but mixed indications aren't very encouraging. Things just seem to be in neutral, with everyone waiting for someone else to make the next move.

As an aside, we note that Bill Miller's Legg Mason Value Trust (One of our Prime Timers tracked at the web site) reduced equity exposure a bit last week. Miller just completed his 15th calendar year of beating the S&P 500, so angling to maintain that record may be a factor, or maybe there's a change of outlook afoot. A graph of the exposure path is posted at the Prime Timers link at the site.

RLC
Hanover NH

 Market Remarks Archives:

2006        2007      2008       2009