June 01, 2007
Don't worry, be happy
Yesterday the Bureau of Economic Analysis told us that first-quarter real GDP grew not at the anemic 1.3% annual growth rate as was originally reported in the "advance" estimate given to us at the end of April, but instead was a barely-positive 0.6% as now claimed in the "preliminary" 2007:Q1 estimates. So what's worse than we thought?
Not consumption spending, which was even stronger in the new "preliminary" estimates than was reported in the initial "advance" estimates. BEA told us last month that, if all other components of GDP had remained constant, the increase in consumption spending alone would have resulted in a 2.66% annual growth rate for 2007:Q1 GDP. Now they say that number is even higher at 3%. Whatever problems the economy had in the first quarter, they weren't a lack of consumer spending.
The revisions to fixed investment spending were also favorable. Nonresidential fixed investment grew more than originally reported, and residential fixed investment fell by less. Together, revisions to these two components would have added 0.2% to the reported first-quarter GDP growth.
Wait a minute. If the core components of demand would have added a half percent to GDP growth, how come the revision resulted in a half percent slower growth? Part of the answer is that, although consumer and fixed investment spending were higher than originally reported, essentially all of that newly discovered spending was attributed to imports. As Menzie noted yesterday, the upward revision to imports subtracted a little over half a percent from 2007:Q1 GDP growth.
But if the revision to consumption plus fixed investment would have increased GDP growth by half a percent, and the revision to imports would have decreased GDP growth by half a percent, shouldn't reported GDP still be exactly where it was initially?
Yes, it would be, except that there was also a big revision to another component of GDP-- inventory investment. Inventories are now regarded as significantly lower than was reported last month. GDP counts production, not sales. Since more of the sales came out of inventories than we thought, production must have been lower than we thought. This downward inventory revision by itself subtracted 0.7% from the 2007:Q1 growth rate, and could be viewed as the primary reason that number fell from the initial 1.3% to the current 0.6%.
In an environment like the present, where my number one concern is the possibility of a demand-led recession, the fact that inventories were drawn down by more than we originally thought is not necessarily bad news. It does mean that production and income weren't as high as we thought. But it should not be viewed as an indicator that the downturn is upon us.
If you were nonetheless depressed by the new GDP figures, you could take some cheer in this week's employment numbers. According to the Bureau of Labor Statistics survey of establishments, the number of people employed in the U.S. grew by a healthy 157,000 workers in May (seasonally adjusted) relative to April, aided by the fact that construction employment mysteriously failed to fall for yet another month in May. For a second opinion, I always turn to the separate BLS household survey, which collects the numbers from workers rather than firms. It also (miraculously) came up with the identical number as the establishment survey-- 157,000 more people working in May compared to April. A third number I like to consult comes from the data directly collected by Automatic Data Processing, Inc. for the 22 million workers whose payrolls the company processes. ADP inferred from their own numbers that seasonally adjusted May private sector employment grew by 97,000 workers. When you add to that the 22,000 new government workers (as counted by the BLS establishment survey), you get 119,000 new jobs. So everybody agrees, for once-- employment growth in May was pretty solid. I had regarded the earlier April household employment numbers as one of the most bearish indicators we had. The new numbers therefore go a long way toward allaying those concerns.
Calculated Risk thinks we won't get the real story on construction employment until October, and maybe he's right. But the data available so far have been coming in better than he and I had been expecting.
a hard landing can take two forms: a growth recession (i.e. a period when growth is well below potential and ranging in the 0% to 1% range) or an outright recession (i.e. negative growth). Thus, in Q1 of this year the US entered in a growth recession, and a pretty serious one indeed.
I find that perspective interesting, since slow growth for the start of 2007 is exactly what people like Dave and I had been expecting all along, and indeed is precisely what the Fed has been aiming for. Bernanke's program from day one, in my opinion, has been to slow growth down sufficiently so as to contain inflation, but not go so far as to produce a recession. The latter phenomenon has a very clear and unambiguous data signature, including such features as a spike in the unemployment rate and significant declines in real GDP.
That may yet come, but let's be clear-- you can't declare it to have started yet on the basis of the data currently in hand. If anybody wants to crow "I told you so," in my opinion the one person who could legitimately claim that right would be Ben Bernanke.
But, being a cautious guy, I expect even he is still not sure just how this story is going to end.
Posted by James Hamilton at June 1, 2007 09:38 AMdigg this | reddit
Prof Hamilton, I don't know if this is true but I heard economist Dianne Swonk say that plenty of residential construction workers have been able to find at least some work on the commercial side (and of course there are immigrant workers who are not counted in the payrolls).
Other bullish signs - last week retailers' sales were up 2.9%, consumer spending and confidence are up, jobless claims are lower, corporate spending, manufacturing and stocks are higher and construction spending actually rose in April (albeit by the minimal amount possible).
So it does look like we should thank Ben Bernanke for doing a good job.
Posted by: Charlie Stromeyer at June 1, 2007 10:45 AM
I think the construction story is most are independent contractors. They are neither employed or unemployed but only have work or don't.
Posted by: Lord at June 1, 2007 11:10 AM
Bloomberg added today that:
"There are also signs that the labor market may not have been as strong last year as earlier estimates suggested.
A report this month from the Labor Department, based on tax records from all businesses, showed the economy added 19,000 private-sector jobs in the third quarter. That contrasts with the government's monthly payroll figures, based on a smaller survey, which showed a gain of 498,000 jobs for the period.
The report showed declines in residential-construction jobs, and more losses may follow in coming months, economists said."
Does that make you any less sanguine about this week's employment numbers?
Posted by: Underlord at June 1, 2007 11:17 AM
So basically, Nouriel Roubini is redefining the term recession to include slower growth than he considers acceptable. Isn't that what most of the rest of us consider a soft landing? Talk about tailoring the facts to fit a preconceived notion, jeez!
Posted by: Hitchhiker at June 1, 2007 12:29 PM
Take a close look at real consumer spending in April compared to March (BEA Table 2.3.6U). Yes, overall, it is growing at an annualized rate of 3.0%.
But, back out the two suspect series -- housing (is owner's equivalent rent really believable, growing at 2.4% annualized in real terms, March to April, given the softness in home prices?) and medical expenses (third party financed outlays will be the last to turn south) -- and household operation (elec., gas, and other) (which, while it is seasonally adjusted, still bounces around a lot), and personal consumptions in March to April fall, at an annualized rate of 0.9%.
Posted by: jg at June 1, 2007 12:32 PM
Roubini is crowing about a "growth recession" now because his call for a recession in Q1 (Q2 "at the latest") hasn't materialized.
You're exactly right: what Roubini calls "growth recession" is basically what the Fed has been counting on to slow inflation.
Posted by: Steve at June 1, 2007 12:54 PM
I agree with your expectation of stagnation -- note I said stagnation not stagflation.
Consumer spending is OK, but really has been growing at a fairly steady growth rate for a couple of years. Capital spending is weak -- no surprise as the surprise was it actually turning negative for a short period. But if trend growth is 2% rather then 10% small fluctuations can give you a negative number. I see little reason to expect really strong capital spending. One reason for this is the very large share of profits going to the energy industry where they are increasing dividend payouts as much or more then capital spending. Given the large jump in oil profits the rebound in oil capital spending has been very weak by historic standards.
We had an inventory problem develop last year. However, in the new world of greater information firms do not let inventories get out of hand like they use and that is exactly what happened this time. They quickly cut back orders and have pretty much gotten inventories back in line. The consensus is that this will now generate a bounce in output. In the old days that would have been true. But now the cut back in orders show up as much or more as a drop in imports that has the perverse impact of increasing GDP. So now that inventories are back in line imports should rebound with the opposite impact of cutting gdp growth. What we have had for a year is the real trade balance being flat -- a big improvement from the long term trend. But the long term trend now appears to be reemerging. So far this cycle trend real import growth has been 7.5% and trend real export growth has been 6.5%. Recently imports were weaker and exports stronger, but this seems to be a temporary development and these trend growth rates appear to be reemerging.
Posted by: spencer at June 1, 2007 01:05 PM
What is your opinion on the discrepancy between real-time estimate by BLS, and lagged-time actual number by BED of jobs gain in Q3,06? The link is below. Thanks,
Posted by: Anonymous at June 1, 2007 03:47 PM
Is it true that the birth/death model, which accounts for most of the job growth - employment has to be one of the most mystical of all government data - is not available for review by economists?
Posted by: Footwedge at June 1, 2007 05:26 PM
Underlord and anonymous, the Bloomberg story is a little misleading. It seems that the 19,000 number they are quoting is based on taking the difference between two series that are first seasonally adjusted separately. In general that is not at all a sound thing to do, though I see the BLS has done the same thing in their press release. I need to look into the Business Economic Dynamics data a little more. I'd welcome any insights from other readers on this as I research it.
Posted by: JDH at June 1, 2007 05:26 PM
Does anyone know if the BLS or ADP surveys do an adequate sampling of small businesses? I ask because in this country 2 out of every 3 new jobs are generated by small businesses and about 50 million employees work for small businesses.
Sure Payroll looks at the data from their over 18,000 small business customers and concludes that hiring and wages are up this year and that there is a shortage of skilled workers:
Posted by: Charlie Stromeyer at June 1, 2007 06:14 PM
Footwedge, I will ask that question of BLS and let you know if I get a response.
Posted by: JDH at June 1, 2007 06:39 PM
Professor, on the construction employment issue.
It appears likely the BLS missed the downturn in employment; at least that is my take on the Business Economic Dynamics data. That would be good news for the general economy; perhaps we have already lost several hundred thousand residential construction jobs and the economy is still OK.
The bad news would be if the BLS is correct, and the job losses are still ahead of us. One thing we can sure of - the job losses will happen.
Posted by: CalculatedRisk at June 1, 2007 07:09 PM
"Other bullish signs - last week retailers' sales were up 2.9%, consumer spending and confidence are up, jobless claims are lower, corporate spending, manufacturing and stocks are higher and construction spending actually rose in April (albeit by the minimal amount possible)"
Please. Show some brain power. "Retailers" sales weren't that impressive. Claims weren't either. If anything, Claims are useless as a leading indicator. Manufacturing isn't higher, but expectations are. If those can't be met, a fallback, possibly major is coming.
Consumption has slowed this quarter and if it slows again next quarter, look out.
Posted by: Johnson at June 1, 2007 07:33 PM
Johnson, see this article and the related interviews to the right of the article:
In May, the ISM factory index unexpectedly rose (whereas European manufacturing unexpectedly slowed). However, last month we did lose 19,000 manufacturing jobs according to the payrolls data.
Posted by: Charlie Stromeyer at June 2, 2007 05:19 AM
JDH said: "...I need to look into the Business Economic Dynamics data a little more. I'd welcome any insights from other readers on this as I research it."
I spent an hour or so reading their methodology and putting together some of the basic data into a spreadsheet. It appears as though they exclude some significant employment sectors, like self-employed workers and government employees.
"...Major exclusions from UI coverage are self-employed workers, religious organizations, most agricultural workers on small farms, all members of the Armed Forces, elected officials in most States, most employees of railroads, some domestic workers, most student workers at schools, and employees of certain nonprofit organizations.
Gross job gains and gross job losses data do not include government employees, private households, and establishments with zero employment. Data from Puerto Rico and Virgin Islands also are excluded from the national data."
Also, it appears as though the data only becomes available *long* after the fact, since it's taken from a quarterly census. I'm not sure what this series does for us.
Posted by: Anonymous at June 2, 2007 11:01 AM
sorry i am a bit off topic.
my question is:
we are having a slowing economy (falling GDP) but increasing employment.....does this means a slow economy but higher inflation?
I have seen a trend in IT that after 2000 bust....there was not much hiring (just enough to maintain) and IT has started hiring big time...since 2006 because their systems are getting old.
can this be happening in other sectors too?
to me it looks like we are going to fulfill our own prophecy...what if businesses now take a clue from the bears and slow down hiring (in prep for slowing economy)....i think that will be the last nail in the coffin.....because consumers are already in debt, which means they are now at the peak of their consumption....i expect a fall in consumption (add to that resetting of ARM mortgages to higher rates).
personally i have not bet on a slowdown......but the probability is increasing (if we can trust the data available)
Posted by: TECHY2468 at June 3, 2007 10:29 AM
i would like to add that i see wages increasing in IT....since there is a shortage now since they were hiring like crazy since last 6 months....
Posted by: TECHY2468 at June 3, 2007 10:32 AM
Footwedge, details of the BLS birth-death model are publicly posted here.
Posted by: JDH at June 4, 2007 12:57 PM
Two things to remember: Population growth requires about 150k new jobs just to stay about even with the current employment rate.
Secondly, unlike others, I have no idea if we will definitely have a recession. But I suspect its possible, perhaps more likely than many believe. If pressed, I will say the odds have moved from 50 to 60% over the past few months.
Posted by: Barry Ritholtz at June 5, 2007 02:16 PM
Thanks, JDH. Still petty mystical but I think I understand generally now. It's easy to see how the model can miss the turn in employment (not that anyone really knows until years later!)
Posted by: Footwedge at June 7, 2007 06:31 PM