Brad DeLong's Semi-Daily Journal (2004)

One can only wonder when this increase in jobs will be "adjusted downward" as they have just about every previous month. Don't these people ever under report job creation? Reminds me of the electronic voting systems where the flaws always seem to benefit the Republican candidate.

With the unemployment rate down a notch (and given that even the seasonally adjusted figures fall short of making up for population growth), this means the ranks of the discouraged workers are still growing. Work is slavery and joblessness is freedom! The four year plan to promote alchoolism, depression and crime is working! Congratulations to all patriotic comerades who have worked hard for this!

Brian Bock: table B suggests that every month in 2003 has been adjusted downward. While BLS could be fudging the numbers, we live in truly anomolous times and their models might need tweaking.

Overall, a better report than I expected, but obviously we are not even treading water at 116,000 jobs (break even is 150,000, correct?). It is amazing how 3 years can "downwardly adjust" one's expectations.

The administration will claim victory anyway since the unemployment rate ticked down. I bet in a few hours Bush will give a speech in front of a background with the word "Jobs" all over it and claim that the "tex relief is working."

Household survey shows 496,000 increase in employment - but could you explain the adjustment line that follows the data?

Hey Weisman got a job at the Wall Street Journal! The spin is on:

NEWS ALERT
from The Wall Street Journal

Feb. 6, 2004

The unemployment rate fell to 5.6% in January, the lowest rate since
President Bush was inaugurated in January 2001. Payrolls grew at the
fastest
clip in three years, adding 112,000 jobs, but fell short of
expectations.

So the Bush tax cuts were, by now, to have created north of 2.1M jobs since July 1, 2003.

The next member of the so-called liberal media to report this will be the first member of the so-called liberal media to report this.

I posted the table frowning clown refers to at the URL above.

Its incredibly consistent: The initially reported numbers overstate the number of employed.

If you really want a better look at the employment situation, check out the Augmented Employment data: http://bigpicture.typepad.com/comments/2004/01/augmented_unemp.html

Thanks for the pointer to the "Augmented Unemployment Rate", Barry. Definitely a useful measure these days...

I don't understand why you find the jump in retail employment disappointing. Over the 3 month period retail employment rose 9,000 with the seasonal pattern doing what was expected.

But this means that out of the 112,000 employmnet gain 76,000, or almost three quarters of the total employment gain was due to the seasonal adjustment factor for retail employees.
Excluding the impact of seasonal adjustment on retail, only 36,000 jobs were created.

Interestingly hours worked jumped 0.8%,compared to 1.4% increases in Nov. & Jan.. But average weekly earnings rose 0.7% because averge hourly earnings only rose 0.1%. I know the numbers do not add up. Probably because of rounding errors.

We are still facing the severe problem of extremely weak income growth.

Harold.-- the adjusment factor you asked about reflect the revised estimate of population, employment and unemployment made by Census to reflect newer census data.

The adjusment revised down the population by 560,000, labor force by 437,000 and employment by 409,000. They have not gone back and revised the historic data, but it shows that employment is now some 409,000 less than previously estimated because of the updated popultion controls. They are suppose to publish an article on the adjusment in the monthly BLS journal.

I am terrible at finding economic data, so I was hoping someone here could point me to data on real income growth going back at least a decade, broken down by quintet. My assumption is that thses numbers would be the best single statistic for seeing how the economy has been for particular groups.
Also, remember the parable about teaching a man to fish...thanks.

Actually, the raw payroll numbers show an increase of 443,000 jobs, but the annual adjustment of "net death of firms" reduces the raw number by 321,000 jobs. Given how fast the economy grew in the last two quarters, this is probably an over-adjustment, but may be correct.

Harold, the household survey did show and increase in jobs of 496,000 and an increase in the labor force of 422,000 (which severely weakens the "oh my god, there are so many discouraged workers" case). All these numbers are heavily seasonally adjusted. In this case, the total level of the population was adjusted making the change in jobs look wrong on the BLS site. It is hard to argue that this is a fudge factor, given that they do it every year.

Intellectual honesty should suggest that both positive and negative addjustments be examined, not just those that fit one's political viewpoint.

correction: I accidently quoted smoothed hours worked data above.
the unsmoothed data was up 0.3 in Nov., down 1.0 in Dec and up 0.8 in Jan.

the 0.8 gain is the largest monthly gain since Feb 1997. Roughly hours worked times productivity equals real GDP. Hours worked is generally a leading indicator of employment.

Just to point the effect of seasonal adjustment, here are the adjusted and unadjusted numbers for "total nonfarm" Dec 2003p, Jan 2004p (i.e. both constituents are still preliminary), in thousands, and from:
http://www.bls.gov/news.release/empsit.t14.htm

130,043 130,155 delta +112 (adjusted)
130,878 128,254 delta -2,624 (unadjusted)

That is, there were roughly 2.6 million _fewer_ actual jobs in Jan (and even that is an estimate, as Pete Harrigan points out). It is only the adjustment for known patterns that evens it out.

Bad Brad, good spencer. The November to January retail sales wobble pretty much takes care of itself. If one had tallied up seasonally adjusted retail job losses in November and December (pre-revision), and added them back in for January, one would come within about 10k of the actual January retail sales employment gain. My personal "disappointment" was of the mangnitude of 18k jobs (I was in for 130k). The 250k-300k estimates seemed off the wall to me (as did a couple of 10k estimates - must have missed the retail employment stuff).

Spencer - thanks for the thorough explanation. After I posed the question but before I saw your reply - I noticed where BLS's reports the unadusted number for employment per the household survey at a figure equal to an 87,000 increase over the Dec. 2003 figure. Now that reported change ignores the footnote (which you gave a nice explanation of). Since the Bush spinmeisters have been ignoring the footnote in the past - I was if they will finally pay attention to it now that it's in their spin advantage.

Pete

Nice add-on to Spencer's comment. I especially liked your call for intelletual honesty (which I read only after my somewhat snide remark about the Bush spin lack of intellectual honesty).

http://www.jobwatch.org/

Nice up to date charts on the labor market....

Notice the the yield on the 10 year treasury is at 4.09%. No rate increase till this difficult labor market turns for real, and there is no reason to believe the turn is near. Darn.

http://www.epinet.org/content.cfm/webfeatures_econindicators_jobspict

February 6, 2004

Payrolls up, but growth not sufficient to boost wages

The nation's payrolls expanded by 112,000 in January, the best month for job gains since payrolls reversed their long slide last September, according to today's report from the Bureau of Labor Statistics. Unemployment fell slightly, to 5.6%, the lowest jobless rate since January 2002. Wage growth remains subdued, and weekly earnings—at 1.7% over the past year—are falling in inflation-adjusted terms.

Despite recent growth in payrolls, revised numbers released today reveal that the job losses over the recession and ensuing jobless recovery have left us with a significant job deficit. Payrolls remain 2.4 million down from when the recession began in March 2001 and 716,000 lower than when the recovery officially began in November of that year. In every other recovery on record, the recessionary losses had already been recouped by this point in time.

http://www.epinet.org/content.cfm/webfeatures_econindicators_jobspict

On the plus side, weekly hours grew by two-tenths of an hour last month. In tandem with employment growth, this increase in weekly hours led to a 0.8% gain in the total hours worked in the private sector, the largest monthly gain since February 1997, and the first hard data suggesting positive overall growth in the first quarter of 2004.

http://www.epinet.org/content.cfm/webfeatures_econindicators_jobspict

The household survey showed very large employment growth last month (496,000), and has consistently outpaced the payroll survey over the recovery, leading some analysts to discount the jobless recovery. However, in a detailed statement released with today's report, the BLS commissioner once again stressed the Bureau's position that the payroll survey offers a more accurate picture of job growth than the household survey. Importantly, as the figure below reveals, both surveys show this recovery to be particularly weak in terms of job growth; in fact, regardless of survey choice, this is the worst jobs recovery on record.

http://www.epinet.org/content.cfm/webfeatures_econindicators_jobspict

On a yearly basis, hourly earnings are growing at 2%, just about the rate of inflation. More recently—and despite faster overall economic growth—hourly wage growth has slowed further, and is up 1.5% at an annual rate over the past three months. Weekly earnings are also falling behind inflation, up 1.7% over the past year.

I continue to see more and more talking heads (Lou Dobbs, don't say "Oh God"--average people listen to him) and politicians, including Democrats, jumping on the protectionist bandwagon by blaming India, China, etc. for our employment problems. This is a real concern especially in an election year. Does anyone else have an opinion on this?

All this talk from Fed officials about signs of labour market improvements, like when "Mr. Bernanke said he was 'pretty confident' the nation would soon see big numbers on employment.", was it yesterday?

So much for Bernanke's "pretty confidence"!

David--The problem you are addressing is that protectionism rears its ugly head when the economy is weak. The best protection against this type of thinking is a strong economy and strong job gains so the losers from free trade can quickly get new jobs.

"The best protection against this type of thinking is a strong economy and strong job gains so the losers from free trade can quickly get new jobs."

Yes, but that is where fiscal policy can help wonderfully. We have needed a jobs policy as a stimulus, but we never got close. China and India bashing is worrisome and self-defeating for workers.

http://www.epinet.org/printer.cfm?id=975&content_type=1

Outside of retail and accounting for returning strikers, net hiring amounted to 27,000, far below the 150,000 benchmark of that is needed to keep up with the growth of the working-age population. The lack of sustained job creation at a level commensurate with this stage of the economic expansion is generating long-term joblessness and stagnant wage growth.

I heard on the news this morning that the president acknowledges that the growth in employment is not keeping pace with the growth in the economy. According to the report, his solution is to make his tax cuts permanent. Gee, since his temporary tax cuts were so effective in creating jobs, keeping them in place is almost assured to keep us on this very impressive trend of growing employment, isn't it?

Thanks to everyone, especially anne, for the stats and comments. The comparison of both the household and payroll stats across post-war recoveries does put the debate on which survey is correct in perspective. The discrepancy does seem large than on average in previous recoveries, but it is plain that this is not a good recovery for jobs by any standard.

But the Pete Harrigan comment has me wondering. I thought I heard on the news this morning that there was an increase in discouraged workers this last month, or maybe that was some analyst giving opinion as fact. I looked at the BLS internet site but I am not familiar enough with employment stats to work through the numbers with a critical eye.

Does anyone know of site or an analysis were several of these time series can be compared across business cycles? I'd like to see several series (unemployment rate, employment from the two surveys, participation rate, discouraged workers, etc.) and see the time paths laid out as function of time since the last trough. Bar charts would be OK, but I would like to see monthly or quarterly time series.

Preferably adjusted and unadjusted.

Could anyone direct me? (I've just looked through the links on this thread but may have missed something)

Try economagic. If you don't see what you want in the index go to the bottom where you will see Labor Force and Non-Farm employment.

http://www.economagic.com/

Try here TheCoach:

http://www.bls.gov/ncs/ocs/home.htm

spencer: Can you please let us know what formulas you are using for your smoothing etc.? Are those plain moving averages? And how do you get your raw numbers? There is probably some interface on the websites where the data are available in CSV or plain text forms.
Thanks.

Pacta sunt servanda - Agreements are to be kept. (Cicero)

Nam et ipsa scientia potestas es - Knowledge is power. (Sir Francis Bacon)

Mens agitat molem - The mind moves the matter. (Vergil)

Vincit omnia veritas - Truth conquers all

Poeta fit, non nascitur - A poet is made, he is not born