October 31, 2003

A Piece of Bad Economic News for the Fourth Quarter

A piece of disappointing news for growth in the fourth quarter:

Consumer spending dips 0.3% in September - Oct. 31, 2003: The Commerce Department said personal income rose 0.3 percent after rising a revised 0.3 percent in August. Economists, on average, expected it to rise 0.2 percent, according to Briefing.com. Spending by consumers, which accounts for about 70 percent of the nation's economic activity, fell 0.3 percent after rising a revised 1.1 percent in August. Economists, on average, expected spending to fall 0.1 percent...

That's a big enough piece of bad news to cause me to take a full percentage point off my personal estimate of the fourth quarter GDP growth rate...

Posted by DeLong at October 31, 2003 06:20 AM | TrackBack


Isn't this where you're supposed to say to Patrick R. Sullivan, "I told you so"


Posted by: Not Patrick Sullivan on October 31, 2003 06:32 AM


This also seems to raise the question of what the Q3 GDP revision will look like (noting Krugman's column this morning). The advance GDP report indicated 6.6% (annualized) real personal consumption growth. In today's report, the chained dollar personal consumption growth rates -- which were actually revised up for July and August, but down 0.6% for September -- only indicate a 1% rise for the quarter which, absent advanced mathematics, doesn't seem to annualize to 6.6%. To the macroeconomists out there: is there a good reason for these reports not to reconcile?

Posted by: Tom Bozzo on October 31, 2003 06:58 AM



I don't have an answer about the math (haven't tried), but I have a very strong impression that the PCE data for Q3 GDP and monthly and quarterly Q3 PCE reports released today contain the same survey data. The Q3 PCE report shows a 6.6% rise in real spending, same as GDP, so I assume the monthly data will work out. So no revision, based on what was reported today.

Now, here is the tricky part. Estmates for October vehicle sales are runnng around 13.3 mln units (annualized), vs a 14.3 mln pace in Q3. Chain store sales in October have been softish, too. If you combine those facts with a slowdown in consumption late n Q3 - a weak base for Q4 PCE growth relative to Q3 - then you start to get the picture of a much cooler showing for consumption in Q4. That is just the way the math works out. This is not a surprise, not unexpected. Many 3.5%-4.0% GDP forecasts for Q4 assume that sort of slowdown. You just need inventory and capital spending to pick up to keep overall growth strong.

Posted by: K Harris on October 31, 2003 08:40 AM


One only-slightly-less-than-fully-serious method of comparing international currencies to one another is via the price of a McDonald's "Big Mac". A commodity that's a buck in Omaha and 8 yuan in Guangzhou implies a "real" exchange ratio of 1:8, right?

Similarly, a so-far-fully-less-than-fully-serious hypothesis about trends in national employment has, by personal observation, proven to be fairly effective as well. To wit: when McDonalds and other fast-food vendors start hanging up banners advertising job vacancies; when fast food service begins to become noticably other-than-fast; and when the orders delivered become noticably other than the orders requested -- then that industry is having trouble getting good workers. The "good" workers are, I presume, moving away from the food service industry toward better (better paying, better benefit, better working condition, better Maslovian self-esteem providing) jobs. This is a mobility that hardly raises a blip in formal statistics. After all, a food service job provides very flexible opportunity for a job-seeker to take time to go on interviews. A current-McJob holder will likely give no notice at all before taking the new job, or may work thru a notice period (on an alternative shift) ever after starting the new job, while accruing enough hours to start the new paycheck coming in. So, initially, the uptake of McJob workers into the "new, better" jobs makes no impact on the unemployment rate. But soon the McBoss starts recruiting less good workers, or out-of-practice workers, to fill vacancies. And those folks move slow and screw up the orders ... at least during the first few months of employment while they learn their jobs.

ANYHOW, I've had FOUR of SIX orders this month
screwed up at fast food establishments around Dallas, Texas. (and my picky-eater kids get pretty upset about it.) Banner "Help Wanted" ads are up at Arby, KFC, McD, Wendy's, and Sonic. (The Wafflehouse and WattaBurger seem to be either well-staffed, or recruiting via other means.) And the pizza delivery guys I see running around in my neighborhood seem to be, once again, teenagers -- not the portly and balding delivery men of a year ago.

I expect the formal employment statistics to look better shortly.

But I wonder if they is any way to formalize my heuristic into a more presentable leading indicator?

Posted by: Pouncer on October 31, 2003 08:55 AM



The 'help wanted' banners indicate demand for new workers. Poor fast food service results from solar flairs.

Posted by: K Harris on October 31, 2003 09:04 AM


Pouncer: I was thinking the same thing. My local movie theatre had a "Help Wanted" sign in the window (for the first time in years), which made me think the economy is picking up.

Posted by: Walt Pohl on October 31, 2003 09:37 AM


More voodoo economics:

At a White House briefing, N. Gregory Mankiw, chairman of the Council of Economic Advisers, said that even though the number of payroll jobs has declined this year, the administration's tax cuts have kept employment higher than it otherwise would have been.

"About 1.5 million more people are working than would have been had the president never changed the tax code," Mankiw said. "It's hard to . . . determine the exact cause and effect of any particular economic change, but certainly there's lots of reasons to believe that what we saw in [the third] quarter is attributable to the president's jobs and growth package. The president's package put money in the hands of consumers, and you saw very strong consumer spending. The jobs and growth package was aimed at lowering the cost of capital, and you saw very rapid investment spending."


Posted by: Kosh on October 31, 2003 09:49 AM


K: Thanks for the discussion.

I looked at the report closely, and figured out what's going on with the math -- it amounts to a start- and endpoint issue. The June-September change annualizes to ~4%. Whereas the quarterly PCE growth compares averages of the monthly PCE for the respective quarters, and annualizes to 6.6%. Which is just to say that the PCE growth was somewhat front-loaded in Q3 and back-loaded in Q2 (no surprise).

Posted by: Tom Bozzo on October 31, 2003 10:04 AM


Recall also that the leading indicators index was down in the last report. The 3rd quarter of 2002 was about 3.5%, but Bush then killed things off for several quarters with Iraq anxiety and higher energy prices. Perhaps this last quarter was just a snap-back reflecting pent-up demand.

Posted by: BobNJ on October 31, 2003 10:11 AM


The compand growth rate of 3 mth moving avg of real PCE in sept vs june was 6.6%. The Sept real PCE vs the 3rd Q avg is 0.76% (annualized).
ie, if 4th Q real PCE = the sept number the growth rate would be 0.76%.

The real funny thing in the data is recent trade data. In aug both real exports & imports fell sharply. In June real exports had had heir first significant jump in this cycle. Part of import weakness is oil. Imports are a lagged function of real final demand and change in inventories. This implies that imports will explode in the 4th Q.
and will be a major negative for 4th Q growth.The coefficient on real domestic final demand lagged 2 quarters is about 0.3 -- over past decade imports filled one-third of growth in real domestic final demand. Thus, really funny thing about 3rd Q real GDP is that trade added to growth.

Posted by: spencer on October 31, 2003 10:31 AM


A couple of points: I believe that the 6.6% increase in consumption in the GDP report was real, while the figure cited in Brad's post is nominal. Real PCE fell by 0.6% in Sept. Are we definitely comparing apples with apples here?

Also, since this confirms that consumption boom in Q3 really happened just in July and August, we can infer that the more late data that comes in (i.e. Sept. data), the lower will be the final figure for Q3 GDP. Hence I still expect a substantial downward revision.

See the post on Angry Bear for more.

Posted by: Kash on October 31, 2003 11:34 AM


>>More voodoo economics>>

Mankiw might be right, in some technical sense - the tax cut may have made things better than no cut, at least in the short term. But there are much much better ways to stimulate the economy - the real sin is not having done something better.

Posted by: richard on October 31, 2003 12:22 PM


I have no doubt that the tax cuts did, in fact, have some stimulative effect.

What I question is Mankiw's statement that:

"About 1.5 million more people are working than would have been had the president never changed the tax code."

How on earth did he arrive at this estimate? Or did he just pull it out of ............. the air?

Posted by: Kosh on October 31, 2003 12:34 PM



Check out this Mankiw speech for a hint of the source:

Mankiw refers to "simulations of a conventional macroeconomic model" showing that GDP would have been 2 percent lower mid-year absent the tax cuts, which somehow translates into 1.5 million jobs. Where you consider that to have been pulled presumably depends on the macro model in question ;-).

Is it an outrageous figure? Not necessarily -- if you consider a cost per job calculation (a la the one that got the Krugman stalkers fired up), you have (per CBPP) some $205 billion of the FY03 deficit attributable to the tax cut, which works out to over $135,000 per job (this year alone!). In short, there seems to have been plenty of stimulus, not well-targeted at all.

I'd also ote that Snow used this line in Chicago last week, and in fact Bush seems to have introduced it back in his August 2 radio address (per a Google search).

Posted by: Tom Bozzo on October 31, 2003 01:07 PM


Focusing on monthly data is an exercise in futility. GDP growth was good and may herald a return to normal growth.

On the other hand, it may mean nothing at all.

I wouldn't bet anything on the apparent slowdown of consumer spending in September.

Posted by: Charles on October 31, 2003 01:46 PM


>>Mankiw refers to "simulations of a conventional macroeconomic model" showing that GDP would have been 2 percent lower mid-year absent the tax cuts, which somehow translates into 1.5 million jobs. Where you consider that to have been pulled presumably depends on the macro model in question<<

It assumes that the Federal Reserve is a potted plant, and keeps interest rates on the same path no matter what the deficit. That's not an assumption anyone who knows anything about the Federal Reserve should make.

Posted by: Brad DeLong on October 31, 2003 02:26 PM


Post a comment