Does anyone understand today's employment report?

[BlueCheese]

The two commonly used figures are wildly contradictory.

First, the standard unemployment rate, as measured by the household survey, fell by 0.4% (good news). Following a similar 0.4% drop for December 2010, this represents a stunning 0.8% drop in two months.

On the other hand, the payroll report was barely even anemic. The establishment survey says there were 36,000 jobs added last month--way below what is supposedly needed to support natural growth of the work force.

The usual explanation when the two reports diverge is that the unemployment rate doesn't count people who are so discouraged that they stop looking for work. But the measure that I think counts these folks, U-6, also shows a big drop, from 16.7% to 16.1%. (see http://data.bls.gov/cgi-bin/surveymost and check Alternative measure of labor underutilization U-6).

One news article suggests that the difference might be because of very small businesses, or self-employed, who don't tend to show up in the establishment survey. Did that many people really start their own businesses in January?

Can anyone help? Thanks.

[Statistical]

1. There are two different reports.

One report (which all U-3 to U-6 numbers are based on) is the employment survey. They call PEOPLE to determine their status.

A second completely different report is the business survey where they call businesses to find out how much payroll has changed. That is where the "net jobs" number comes from.

Now both reports are released at the same time but understand they are COMPLETELY seperate. The issue with the business survey is the BLS is often unaware of new and small businesses. Those people are employed but BLS doesn't count them in net jobs.

[BlueCheese]

3. I wrote something about that in my OP.

Do you think new and small businesses can explain the entire discrepancy, which seems quite a bit larger than usual?

[JDPriestly]

7. Businesses may be hiring people as independent contractors

to avoid having to pay the business's share of various types of employment taxes. That's one possibility.

[BlueCheese]

2. Just as I write that...

This article seems interesting. Warning: WSJ. Apparently January is the time that the BLS updates its population numbers, adding to baseline confusion.

http://blogs.wsj.com/economics/2011/02/04/messy-new-estimates-complicate-explanation-for-unemployment-rate-drop/

[Hillprop]

4. Would take too long..

..to explain completely, but the short answer is the drop in unemployment was essentially because of book keeping. A new household survey estimate is being used, as the Census showed the old BLS estimate was overstated by approx. 500k.

The old figures were not revised based on the new figures, hence the discrepancy and what appears to be a drop by 0.8%, when in reality it probably ticked up slightly.

[2Design]

5. purely fictional for centuries - not based on real peopple unemployed

does not count those not collecting unemployment - it is a fictional government smoke screen

[Statistical]

6. False. Unemployment has nothing to do with benefits.