Friday, July 6, 2012

A tale of two jobs reports

The establishment survey of payrolls has been disappointing of late, finding an average monthly gain of only 105K new private sector jobs over the past four months. (I focus on private sector jobs because that's where the action is; the public sector workforce has been shrinking for the past three years due to budget cutbacks, and I view that as a good thing.) In contrast to the establishment survey, the household survey has registered a relatively strong average gain of 235K jobs over the same period. Since the jobs recovery began in early 2010, the household survey has registered about 5.1 million new private sector jobs, while the establishment survey has recorded 4.3 million. This discrepancy between the two surveys is large, and somewhat troubling. One survey says jobs growth is weak, the other says it is decent. Which one to believe?

The household survey is typically better at picking up new jobs in the early years of a recovery, since it is based on a sampling of the entire population, whereas the establishment survey only samples known businesses. So if there are lots of new businesses being formed, especially small ones, the household survey will find those but the establishment survey won't find them until a year or so from now when it is recalibrated to match IRS data. I'm inclined to trust the household survey more than the establishment survey right now, and so I conclude that the economy is still growing at a modest/moderate pace—nothing to get excited about, but nothing to get upset about either.

Another encouraging aspect of the June jobs report is that the labor force (all those with jobs plus those who are looking for jobs) is once again growing. Over the past year, the labor force has grown 1.1%, which is roughly its long-term average. Over the past six months, however, it's up at an annualized rate of 1.7%, and it has now reached a post-recession high. The overall growth of the labor force since the recession started is of course still miserable—the worst performance in modern times—but the change on the margin is now quite positive.


Anonymous said...

The household survey had been outperforming the establishment survey non-stop since 2003. One has to wonder if something else is going on in the economy that the BLS isn't picking up.

McKibbinUSA said...

The good news is that the US employment to population ratio marked a three-month rise in June 2012 -- more at:

brodero said...

Both the temporary help services and
the truck transportation components are growing faster than nonfarm payrolls....

Junkyard_hawg1985 said...

I think IBD has the most insightful piece on the jobs report. More people signed up for disability in June than were added to payrolls. This was not just a June phenomenon. More people have been added to the disability roll since the recession ended in June 2009 than have gained employment.

Squire said...

That the household firms not being picked up by the establishment data are growing but the establishment firms less so doesn’t make sense to me from my experience in the field. I think there is a problem in the procedures for collecting and analyzing the data but then again I am but just an anecdote.

The establishment non-seasonally adjusted non-farm private jobs figure, table B-1, has slowed just a bit. Not so bad considering we don’t have a dot com or real estate bubble going on.

The headline figure is manipulated low so they can do a catch up with vengeance for September and October before the election.
What me cynical?

Benjamin Cole said...

Junkyard's comments are worth noting.

We have created a dole class, through an extended recession and social welfare policies.

The "disabled" will join the public pensioners, farmers, military, and other groups dependent on federal cash to survive.

Far better we have a growth economy,and a limited public sector.

Charles said...

Does the Household survey pick up jobs in the off-the-books economy? My impression is that there has been considerable growth in the cash based economy - especially in remodeling and small-scale construction. Given what has happened to residential construction it would not surprise me if more people were collecting benefits and working for cash.

marmico said...

More people have been added to the disability roll since the recession ended in June 2009 than have gained employment.

What a crock. There are additions and terminations and the net is "in current payment status".

When Bush was in charge the current payment status increased by 5%+ for four consecutive years.

If the the second half of 2012 is similar to the first half, Obama is only charged with 2010 for a 5%+ increase.

If you need a laugh get IBD to chart Bush. It was certainly easier to move from applicant to award in his era.

Disabled worker stats.

marmico said...

I'm inclined to trust the household survey more than the establishment survey right now

Why? Is it confirmation bias? You are aware of the CES Net Birth/Death Model, no?

Why don't you look at the “adjusted” household survey employment series?

marmico said...

This discrepancy between the two surveys is large, and somewhat troubling.

I disagree.

Since February 2010, the household survey employment level has risen by 4.44% in total. The "jammed in" January 2012 positive adjustment of 216,000 jobs on account of the Decennial Census Survey should be subtracted which gives a rise of 4.26% in total.

The payroll survey employment level has risen by 4.09% in total.

The raw number differential of 800,000 jobs suffers from both the census effect and a baseline effect as the household level is larger than payroll level. You should make a second chart normalizing both series.

Correcting for both, the difference is 176,000 jobs in total or 6,300 jobs per month in the 28 month time frame.

I don't find that troubling at all.

Scott Grannis said...

The fact remains that over the past four months there has been a huge difference in the pace of job gains between these two surveys that cannot be easily explained away.

Bill said...

Who cares? It's still the worst job recovery since WWII.

McKibbinUSA said...

Employment statistics have always been "fuzzy" numbers to some extent -- measurements of "unemployment" have been subject to changes in recent years, and are especially suspicious these days -- although not perfect, the employment to population ratio provides a better view of the employment situation according to many economists -- more at:

For the record, the US employment to population ratio has improved each of the last three months on both a monthly and annual basis -- said another way, the US employment to population ratio is an optimistic indicator for labor markets.

marmico said...

Sure the EMRATIO paints the broad brush. Theoretically, it's a dependency ratio and it looks pretty flat to me. Maybe you could compare it to the early 1960s.

Bottom line, you can't simultaneously blow and suck about the payroll and household surveys ever month, like Scottie does.

Unknown said...
This comment has been removed by a blog administrator.