WEEKLY ECONOMIC NEWS DIFFUSION INDEX (WENDI)
WENDI belted out a show tune this past week, sustaining more strength over 2 weeks than at any point in the Recovery (just a bit more than in August). Most notable this week were bullish contributions from 4 different manufacturing surveys: Kansas City, California, Empire State (NY), and Philly. The Richmond survey was a negative. As a number of one-off economic stimuli (e.g. tax rebates & the REFI boom) recede in the rearview mirror it is key that the Production end of the economy pick up and stimulate hiring. Thats what the whole Greenspan (Keynesian) gambit is designed to do. And this week's data suggest that there is reason to be optimistic about it working, at least in the near-term.
The Weekly WENDI came in at 34% for October 17, within 7 points of its alltime high. That pulled the 4-week Weighted Moving Average up 5 points to 17%. When the 4-wk Avg. is rising that indicates positive acceleration in the news flow. The Cumulative WENDI rose to a new high of 68% and has been steadily rising since July, albeit with a brief dip in September.
The news on the economy is decidedly and sustainedly bullish. Indeed there are projections for 3Q03 GDP reaching as high as 7%, about twice as high as originally expected. The bearish macro-economists I read are now in clear backpedaling mode, acknowledging 3Q's strength but dismissing it as fleeting, and some are now guiding 4Q's guesstimates lower, back to a 3% handle.
Merrill Lynch used to have a chief economist named Bruce Steinberg who would have been bullish straight through the apocalypse. When they fired him and hired a much more bearish Kahuna named David Rosenberg some of you may recall my mentioning that this was a sign that we were much closer to a bottom than a top. Well, when Rosenberg gets fired I'll be sure to let you know. 'Cause that'll be time to get all "beared up."
Rosenberg's weekly note from Friday is chock full of rationalizations for why he missed predicting 3Q's strength. It's also loaded with reasons that 4Q will disappoint. We don't want to dismiss that possibility, but in my view we want to keep an eye on the jobs market as the fulcrum on which the Recovery will hinge. If we see accelerating job creation (so far we have one month of very modest improvement) then Rosenberg's thesis is in deep doo-doo. If job creation languishes (courtesy of the productivity boom in concert with the export of manufacturing jobs) then Rosenberg may get to keep his corner office for a while.
(And by the way, if you're interested in an excellent piece on the subject of how it is precisely the exporting of jobs that may be a prime contributor to the soaring productivity statistics, which I've been thinking is the case for quite some time, read Stephen Roach's piece at Global: Imported Productivity. Roach is another of the backpedaling bears at the moment. But this report is well worth the read.)
And if that doesn't slake your unquenchable thirst for knowledge, here are the ingredients of this week's witches' brew (WENDI score).
1. Kansas City Fed Manufacturing Survey: Rose to a new all-time high in September (though "all-time" only goes back about 2 years). Shipments, orders, number of employees, and workweek all rose smartly. The 6-month outlook remained strong. Bullish (1).
2. California Purchasing Manger's Index: This is just the 5th report out of the CA, so there's little history on it. Immaturity notwithstanding, in 2Q03 the index rose to 63.3 (50 is neutral), its highest reading to date. All the numbers were to the good with the High-Tech Composite shooting up from 43.4 in 2Q03 to 71.3 in 3Q03. Our qualified score on this survey is only due to its lack of history. Qualified Bullish (0.5).
3. Richmond Fed Manufacturing Survey: Dropped from 0 (neutral) to -7 in September. New orders and shipments both deteriorated. Backlog of orders decelerated their decline but did not improve. Bearish (-1). (Bad time to be looking for work in Virginia.)
4. BT-M Chain Store Sales: Fell (-0.5%) for Oct. 11. However Y/Y growth was 5.3%. After a surprisingly strong September BT-M took this opportunity to revise down their October projection to about 3.5% from their previously outlook of +4-5%. That's two negatives and a positive just cited. Qualified Bearish (-0.5).
5. ABC News/Money Mag Consumer Comfort Index: Rose 1 point to - 19. This report has languished, oscillating between -19 and -20 since September 14. The consumer's psyche is lagging, but not worsening. Arguably Neutral (0).
6. MBA Mortgage Applications Survey: Dropped a whopping 20.5% for October 10 with both REFI and Purchase Applications taking it on the chin. The housing market is still in the process of passing the baton off to other sectors of the economy, it having run its leg of the Recovery Relay with admirable persistence and determination. I don't think we're in trouble here, but Housing's legs are definitely burnt. Neutral (0).
7. NY Empire State Manufacturing Survey: Like the Kansas City Fed this one rose to a new all-time (albeit juvenile) high. All the important internals were strong including Shipments, New Orders, Unfilled Orders, Prices Paid, Employment, Average Workweek, and 6-month Outlook. Inventories declined sharply, suggesting that pickups in final demand will rush through to production increases (which would further spur on the jobs market). The one negative may be Prices Received, which continued to deteriorate. That's indicative of a surfeit of capacity and supply. All things considered, though, this is report is quite Bullish (1).
8. Retail Sales: For September the headline number showed a drop of 0.2%. However the "core, core" index (Ex Autos Ex Gas) rose 0.2%. July's headline number was revised up from +0.6% to +1.2% but most of that revision was in the Autos component. Y/Y Retail Sales are up 5%, the strongest growth in 3 years. This is report was so mixed and multi-edged that I can only call it a Neutral (0).
9. Jobless Claims: Initial Claims fell more than expected to 384K for Oct. 11. (The prior week had reported 382K but that was revised up a greater-than-average 6K to 388K....so (apples-to-apples) the initial Initial Claims number rose 2K, but (apples-to-oranges) the Initial- Claims to Revised-Initial-Claims number fell 6K. What was it someone said about there being lies, damn lies, and statistics?) Continuing Claims remained near bear-market highs at 3.674M. There does appear to be a sluggish downtrend in Initial Claims but with those Continuing Claims remaining high the best we can call this report is a Qualified Bullish (0.5).
10. Consumer Price Index (CPI): The headline number for Septmember was 0.3% but the core CPI (ex food and energy) showed up at an immensely tame 0.1%, which knocked the Y/Y rate to a new 38-yr low of 1.2%. It continues to be my belief that this very modest inflation is the most bullish result at this point in the economicrecovery cycle. Obviously we don't want to see DEflation, but we want to see Growth LEAD inflation, not lag. If Growth lags and inflation leads then STAGflation becomes more likely. So this tame inflation is Goldilocks Bullish (1).
11. Industrial Production: More modest improvement was the story here. Production grew 0.4% in September with Capacity Utilzation edging up .03% to 74.8%. Qualified Bullish (0.5)
12. Business Inventories: Declined by more than expected in August, dropping 0.4%. Sales declined a like amount. The Inventory/Sales Ratio remained at its low of 1.36, which on a stand-alone basis might look good...but when it's because of contraction of both the numerator and the denominator that's a dicey conclusion to draw. Qualifed Bearish (-0.5).
13. NAHB Housing Market Index: Rebounded in October to 72, its highest reading since December '99. The 6-month Outlookis is soaring even as Mortgage Applications back down. A very strong report on the psychology of builders. (Let's hope it's not a last burst before the fall.) Bullish (1).
14. Philly Fed Survey: Trounced expectations (of 16), coming in at 28 (the highest level since 1996) and echoed KC, CA, and NY surveys' bullish tones. Here too all major internals were strong: Shipments, New Orders, Unfilled Orders, Number of Employees, Average Workweek, Number of Employees. And once again Prices Paid were stronger than Prices Received, which could have a marginsqueezing effect and speaks to the lack of urgency in final demand. The 6-month Outlook also backed off a bit, though it remains essentially robust. These last two caveats are not enough to qualify the Bullish implications. (1).
15. University of Michigan Consumer Sentiment Survey: Showed up at 89.4 for October, reversing September's decline. However it did not move above 90. The stock market tends to be supported by readings of 90 or above and/or acceleration in this report. It's been flirting with 90, but plateaued since May. The Job Market continues to weigh on the consumer. Nuetral (0).
16. New Residential Construction: Easily beat expectations for September, growing 3.4% M/M to a 1.89M seasonally adjusted annual rate. Permits declined just a smidge, but remain very strong. Bullish (1).
As of October 10 the Forward 12-month (FTM) consensus estimate for Operating Earnings on the SPX is at $59.53. The all-time high was just shy of $63, back in '00. Trailing Operating Earnings through the September quarter (according to Standard & Poors, as of 10/16) are at $51.27. Trailing Reported Earnings are at $37.02. But that gap of $14.25 between Reported and Operating Earnings will close at the end of 4Q03 by at least $7 when 4Q02 ($3 in reported earnings) falls off the look-back period. The current run rate for Reported Earnings is about $44.
Forward EPS estimates have risen at an 18% annualized rate over the past 3 months, at a 15% rate over the past 6 months, and at an 8% rate over the past year. Trailing Operating EPS have risen 23% in the past year. Reported EPS have also risen 23% in the past year.
All these earnings lines are rising. The FTM EPS (blue) line is within about $3.50 of its all-time high. The TTM Op. EPS (yellow) line is about $5.50 below it's all-time high. The Reported EPS line (magenta) is set to close at least half the gap up to the yellow line at the end of the year.
In this context it's tough to predict a market crash. We do, though, want to keep an eye peeled for how the market responds when the acceleration in earnings tapers off, which it is bound to do sometime next year.
RISK PREMIUM & FAIR VALUE
In Closing Bell, Oct. 5, 2003 we looked at a couple of measures of risk in the stock market. I'd like to return here to the Fed model of Fair Value.
In essence the Fed model suggests that the Forward 12-month (FTM) Operating Earnings Yield on the SPX should be about the same as the dividend yield on the 10-yr Treasury. (The implicit algebra is that Risk in stocks is offset by Growth in earnings.) So we can calculate Fair Value in this model by dividing the FTM consensus estimate by the 10-yr Note's yield. Right now the FTM EPS estimate is $59.53. The 10-yr Note's yield is 4.39%. So, 59.53/0.0439=1356. That's the Fed's Fair Value for the SPX, about 317 points higher than current levels.
And what accounts for this 317 point disparity? Risk Premium.
The Risk Premium we're looking at is what the denominator on the left side of the equation (10-yr yield) would have to be in order for the market's current level to be "fair" in the Fed's model. The simple algebra looks like this.
Solving for X...
...which is to say that the 10-yr's yield would have to be 5.73% to make SPX 1039 Fair Value. With the 10-yr yield now at 4.39% the difference between the Fair yield and the Actual yield is +1.34%. And that' how we'll frame it for current purposes.
Now in the October 5 edition we noted that the average Risk Premium on this reckoning since '94 has been 0.11%, which makes the current perceived risk in the market (at 1.34%) high by comparison. But some astute readers pointed out that this data might be skewed owing to the prolonged existence of an apparently overvalued market during the latter half of the '90s. So I decided to do some further research.
I only have Earnings Estimates in my data base going back to '94. But I do have actual earnings data going back much farther (to 1927 in fact). So, what I've done is take a look at the last 43 years, going back to 1960 (probably not coincidentally, the year of my birth.) using ACTUAL yearforward earnings as a proxy for the Consensus Estimate in the Fed's model. (And I think this proxy is actually better than using the estimates because we don't have to worry about corrupt or stupid analysts and their opinions, just about how the market itself discounted what earnings indeed turned out to be. But of course the drawback is that our calculation of Fair Value on this reckoning has to stop a year ago.) And here's what I found out.
Our look back encompasses the yellow area. Why? Modernity and pertinence. If we go back further, then we start dealing with the aftermaths of the Great Depression, WWII, the Korean War, and the immense structural economic transitions that were wrought in the first half of the 20th Century. If one wanted to argue that the markets should function as they did prior to 1960, then that's just a "whole 'nother ballawax." For our purposes here, we'll suggest that 43 years of modern history are worth considering.
And guess what. Since 1960 the AVERAGE SPX FTM OPERATING EARNINGS YIELD HAS BEEN IDENTICAL TO THE AVERAGE YIELD ON THE 10-YR NOTE. In other words, the AVERAGE RISK PREMIUM from 1960 to the present (using year-forward actual operating earnings in solving for X) is ZERO. The Fed's model is just about SPOT ON.
This look-back period includes the sexy '60s, the inflation-ridden '70s, the deficit-burdened '80s, the Goldilocks '90s, and the anti-bubble of the past 3 years. It's an interesting cross-section of economic history. And it suggests that the Fed's model ain't a load of horse**** either. Not by a long-shot. (And by the way, one standard deviation from the mean is 2.2% over this period, so it looks like 68% of the time the Risk Premium is between -2.2% and +2.2%.
Of further interest is that since 1980 the average risk premium is -1.46% with one standard deviation being 1.3%. So, over the past 23 years risk premium has been between -.16% and -2.76% about 68% of the time.
And where are we now? Well, obviously we don't have actual earnings for October '04 yet but we do have analysts' consensus estimate. (Risk Premium using FTM Estimates is charted in magenta above, and moves pretty much in sync with the FTM Actual line in blue). On FTM Estimates Risk Premium is now at about 1.34%, which is high but not excessive relative to the past 43 years (inside 1 standard deviation). Relative to the past 23 years, however, the Risk Premium is MORE THAN 2 Standard Deviations above the mean.
So, what can we conclude? On a relative valuation model (relative to interest rates) the market is still scared (cheap). It's either somewhat scared (cheap on a 43 year comparison) or extremely scared (cheap on a 23 year comparison).
What's the market so cheap about? Well, it's either scared that interest rates are going higher or that earnings growth is will slow down. Very likely some of each will happen. Over the intermediate term, however, and with the Fed explicitly on hold for some significant time to come, it would be unlikely to see the 10-yr yield go above 5-5.25%, as a high upper limit. With the 10-yr yield in that range (62-87 basis points higher than at present) the Fed's Fair Value model (dead on over the past 43 years and conservative over the past 23 years) would put the SPX in the 1130-1187 range (absent a collapse in earnings).
A LOOK DOWN THE MARKET'S THROAT
This past week almost all our dozen benchmark indices broke above their September highs. Some of them have tucked their heads back into their shells, however, and our 5-day Stochastic oscillators are all headed down.
The pink oscillators have a number of bearishly divergent double tops. That suggests that we're due for some short-term pullback. The SPX has closed virtually on the top of its yellow band (at the September high). The Dow is stronger, maintaining its loft above its September high. The Dow Transports are confirming the Dow's loft while the OEX is solidly below its September high and may be showing us a double top.
The S&P Midcap is maintaining its posture above its yellow band while the Russell 2000, like the SPX is sitting virtually ON the upper limit of its band. Both the Advance/Decline Line and Cumulative Volume are strong, though the A/D line is notably the stronger of the two.
Our look at the Tech indices is also ambivalent in its opinion about the breakout over Septembers highs.
The Nasdaq Composite is sitting on the yellow zone's upper limit. The Nasdaq 100 is back inside the yellow as is the Philly Semiconductor Index. The Nasdaq 100 A/D line is still poised above its yellow congestion band.
Six of our benchmarks are above their September highs. Three have penetrated back below their September highs. And three are virtually AT their September highs.
In all of these charts our short-term momentum oscillator is heading down. In almost all of them there is a lower oscillator peak associated with a higher price peak. That's a short-term bearish divergence.
What's it all mean? It means that the breakout over the September highs is enjoying only a limping kind of confirmation (half our indices). And that short-term momentum is pointing down. Absent some killer earnings blowouts the market looks too tired to rocket ahead.
We have a nice chart that uses the Put/Call 3-dma and 5-dma to illustrate the market's fatigue.
As you can see these low levels on the P/C moving averages have strongly correlated with local tops. If a slight pullback in the SPX brings on a strong bout of put buying, which pushes the moving averages sharply to their upper band, then we'll assume the downside will be limited. If, though, complacency continues to reign, and the P/C Ratio is sluggish to rise, then we'll look for a deeper pullback.
Short-term oscillations notwithstanding, the structure of this cyclical bull continues to be healthy.
There are a load of Support (S) lines beneath the market. Should we puncture through the S4-R3 range (1028-34) then we'll look for the R2-S5 range (1007-1015). Below that there are support targets at each S line on the chart.
We'll consider our short-term call for more downside to be obviated if we close above last week's high of 1054. In that event we're looking for R1 at 1070. And we'd expect that level to provide some significant resistance, just as R2 provided a ceiling over the summer.
Short-term: Expecting downside as the market realizes it has priced in a load of good news lately.
Mid-term: Continuing to expect higher prices by year-end, with the SPX at least hitting 1070 and possibly heading toward the 1130-1187 range as per our Fair Value discussion above.
Long-term: If we see job creation accelerate then we'll be looking for further upside into '04. In the absence of job creation we'll become more neutral to bearish as improvement in final demand will likely be unsustainable.
If you'd like to follow along during the week with our analyses, please join us at The Agile Trader for our daily Pre-Market Updates, Afternoon Notes, and intraday Trading Alerts.
Best regards and good trading!