October 14, 2005: Today's Events:
If you read this commentary to get a different spin on the data from the typical Wall Street regurgitation, today is your lucky day...because when I look at the clouds, I don't see any of the funny shapes that economists seem to be seeing.
The data as I read them today were weak, bordering on ugly. While some of this was clearly due to Katrina and Rita, some of it is difficult to shrug off.
Headline Retail Sales was near expectations at +0.2% although the trajectory was worse (upward revision to Aug; Sep below expectations). This will continue to be the case next month, because October car sales are shaping up as weak. Core sales were a solid +1.1%, but a lot of that came from gasoline sales. Even stripping out gasoline sales, retail sales appear to remain reasonably strong (if you ignore the auto weakness).
Industrial Production, though, was awful. The -1.3% decline was the worst since 1982, and the 78.6% Capacity Utilization erases a year of gains. This was much worse than was expected, even accounting for the hurricanes, and what is interesting is that almost every sector of manufacturing lost ground. Only high tech, and cars & car parts, were strengths in the report, so this is more than just the 9%-10% fall in mining activity (and a smaller decline related to a Boeing strike). And of those two sectors, the strength in autos is a head-scratcher. Building inventory when you have poor sales? Gee, and we wonder why the auto companies are having troubles. Get ready for more incentives! But anyway, the point is that the weakness in IP, while exaggerated because of the hurricanes, is real.
The Michigan Confidence figure was supposed to bounce! Why didn't it bounce? It in fact declined further, to 75.4. That's the lowest level since 1992, and the three-month decline is the third largest in the history of the series, behind only January-March 1991 and June- August 1980. You might recognize those dates as the dates that more or less signaled the beginning of two of the three recessions of the last quartercentury. Yes, I'm quite aware that confidence is low for "one-off" reasons...but it's one-off reasons that tend to trigger recessions! In this case, the one-off reason is high energy prices. In those cases, the one-off reason was... high energy prices. This is the sort of fancy econometric analysis you can get from Ashton Analytics. I really don't see what the confusion is.
Now of course, our work here focuses on inflation, so I want to say just a bit about that release! The headline was, as everyone suspected, scary, at +1.2%. But the core rate was only +0.1%, which is what our model sees as the basic run-rate right now.1
Some observers dismissed that print, saying it came from the treatment of housing costs. Because rents are somewhat inflexible, and because the amount of rent that accrues to the cost of housing must be the difference between the overall rent and the amount of it that covers the utility bill, increases in energy prices tend to (in the short-run) decrease the cost-of-shelter component of CPI. In my view, that's the exact correct treatment of rents: if my rent hasn't risen, then in what way is the rise in the cost of heating my building inflation to me? In the future, the rent might go up...but that will be captured in due time, when the rent does go up. This is the part of the crack econometric analysis where my lack of a Ph.D. in mathematics plays a key (and positive) role.
You can't just shrug off the soft core CPI that easily. The Cleveland Fed's "Median CPI" indicator also printed +0.1%, the first time it has done so this year. The design of the Median CPI - that it tracks the median, and not the average - means that outlier inflation events have less effect on the figure. The Y/Y rise in that CPI is 2.3%; it has been 2.2%, 2.3%, or 2.4% every month for over a year, and the high print for the '04-'05 expansion is 2.5%. Note that when the Fed was easing aggressively in 2001, the Median CPI was rising between 3.2% and 4%.
But what's more, look: the 6-month rise in core CPI is only +0.5%, which is (even I can do this math) an annualized rate of +1.0%. Meanwhile, Crude prices have been above $30 and rising since late 2003. At what point does the economics establishment start to question whether the hypothesis that energy inflation is eventually passed into core inflation might be flawed, or at least that there may be other effects that are larger than any pass-through that may be occurring? Gee, we'd throw out for consideration the income effect, which is what the Fed said for years until recently (high energy prices tend to cause lower core inflation because disposable income evaporates, meaning less cost-push inflation on the rest of the market basket); our favorite hypothesis, though, is that the debt effect restrains inflation very well when private debt levels grow onerous, as they are now.
If the pass-through does not begin soon, then when energy prices eventually oscillate lower - I'm not saying $30, I'm saying $50 - we'll be in full-fledged deflation alarm again. I know it seems outrageous to think that way right now, and I'm not saying that's a trading view. But our forecast continues to be for a recession beginning in 2006Q1, and I suspect we'll be worried about deflation right around then. Headline and core inflation will cross over, and typically that happens mainly because headline inflation regresses to core rather than the other way around.
Somehow this is all bullish for stocks...I don't know how; I presume today's +0.83% gain is some sort of "relief rally" that happens in equities but apparently not in bonds, where yields rose 1.6bps for the 10y note (4.487%) after reaching 4.52% earlier in the day. Of course, the CPI print is terrific for TIPS, especially the short ones. The Jan- 07 breakeven widened by about 9bps, the Jan-08s widened 7bps, and the Jan- 09s widened 5bps. After those issues, the tepid core reading had more effect and the rest of the breakeven curve rose only 1-3bps. Commodities were mixed; the dollar closed lower.
Over the weekend, the Iraqi constitutional referendum will be held, and while there is likely to be violence the real key is whether the referendum passes. The rules of the vote result in passage unless two-thirds of the voters in at least three of the eighteen provinces reject the referendum, which seems unlikely at this point (but who knows?). Passage, along with the recent transfer of policing duties in much of the country to Iraqi forces, means the clock is clearly ticking on the eventual decision to begin withdrawing troops from the country. This should be regarded as bullish for bonds, whenever it happens, but that will not be this weekend.
Also over the weekend, keep one ear tuned to listen for any big-name bankruptcies; as noted yesterday, the deadline for filing under the old rules is here.
On Monday, the first look at October data on manufacturing will arrive in the form of the Empire Manufacturing Index (Consensus: 19.0 from 17.0). I think the consensus is loopy on this one - Empire is already well above the level of other manufacturing surveys, and people are calling for it to rise?
Question of the Day: What is TASER an acronym for?
Answer to Prior Question: The question was, "How many Oscars did Frank Capra's "It's a Wonderful Life' win?" Trick question: the movie won no Oscars, and only became popular when RKO failed to renew the copyright in 1973, which allowed television to pick up the program and air it for free.
1 Actually, the fall in the Y/Y core rate pushes the model's 2006 forecast for core CPI to +0.96% because of trend effects.