Any explanations for the big upwards revisions to August and September? Revised August numbers are now over 200K.

Anecdotal evidence on the slowdown in our neck of the woods (Southeast Michigan):

In our year-old subdivision, it is now the exception for a house to sell to the original purchaser. i.e. people are walking away from their deposits, and houses that were marked sold are now marked available. What's worse is that our builder now only seems to be building houses when someone puts down a deposit.

Our sub has a "cheaper" set of models ranging from 170k for 1400 sq ft to 225k for 2700 sq ft. The builder just added a new model of 1000 sq ft for an as yet undisclosed price. My assumption from this is that even around relatively affluent Ann Arbor, people are walking away from 200k homes due to the cost.

So far we've only seen one house in our sub for sale, and that notice has been taken down without any change in owner. I know both our immediate neighbours are very worried about continuing to afford their

I've come with the same question as Tom and Kett82. Those sure don't look like sampling issues. Did somebody have a couple sheets of data fall off their desk last month or something?

And I could be wrong, but haven't most of these reports been adjusted downward, any explanation for that switch?

Does it have something to do with workforce composition (i.e. reading a particular sector wrong)?

Unemployment statistics have become mostly meaningless. Look instead at the employment data and you'll get a different picture. The employment rate for men age 25-64 is peaking/stagnating at around 83.5% -- which was the employment level nadir in the early 1990s.

Moreover, the last time the US added 200,000+ private sector jobs was November 2005. Thus 2006 hasn't had a single month over 200,000. That's TERRIBLE performance for an economy not in recession -- shades of 1992 and the jobless recovery, of course, although we're now four years past our (official) recession.

For the bottom 60-80% of the population, it's still hard to tell the difference between growth and recession.

GG

I expect the GOP to trumpet that 4.4 number as if it were the last puff on that horn that was supposed to let that charging grizzly know you were coming AND you were NOT his dinner.
(Actually I need to let you horn blowers out there know that the short blast from my tuba is that I can't hear too much with an estimated 8% of the work force not being all that well documented.) [But does this stop me and my underemployed tuba?]

4.4% looks good, but is it tasty?
This at first glance looks like a near balance: loss of jobs in the goods producing sector are more than made up by gains in the services sector. In the exodus from goods, construction (slightly more than half the size of manufacturing) lost slightly more than half the manufacturing sector. (Those auto industries are not to be upstaged by some fly-by-night construction companies.)
It looks like if you can swing a hammer, you can swing a mop and the place might be twice as clean as it was last month (in some areas, under some conditions that might not apply to your particularly messy circumstances). So fears about the mighty job dislocation look to be premature at this point and possibly unfounded.
But it is early IMSoPaitentO and that dislocation from generous (Ok, not you illegals out there who are too busy to read this.) housing related employment may still be on its way.
Last little puff: Do falling retail employment numbers represent technological changes (stores processing better sales/employee ratios) or declining sales? [If sales/employee is improving, does that register as an improvement in productivity (recently claimed to be zero) or just longer wait lines for customers?]

Silly me. No need to guess about the government contribution to revisions. It's right there in the data. Of the total revisions to August and September, 43.2% were accounted for by government jobs.

calmo: I'd say it's a "productivity" increase when sales per employee's hour worked (e.g. standing behind the register) rise. The purpose of any retail operation is to move product from the warehouse into the customers' hands, and the customers' money from their wallets into the corporate bank account (arguably the former is merely required for the latter and not the primary purpose of the business). Either or both divided by employee hours is productivity, I guess.

5:59 AM Pacific! And I thought this West Coast Angrybear had gotten up earlier with my 6:30 AM Pacific post.

That's strike two.

dryfly: On a pessimistic note, at some point you ask yourself what the point of obsessing about all the data crunching is.

From a cybernetic (process control) point of view, indicators that are not used purely for observation, but feed back into policy, tend to get decorrelated with what they are supposed to measure.

Another angle from which to look at this is psychological (and indeed in some cases psychiatric!) -- esp. with indicators that can be described as "performance metrics" (and this applies to most economic series), there are powerful incentives to game the numbers. The harder you try to manage through the numbers, the more they decorrelate/get gamed.

I'm not suggesting that improprieties are happening inside the BLS statistics dept. The feedback network is larger, including things like setting policies of what to measure/publish and how, how to model, what definitions to use, shaping of public image of the indicator, etc. It's more of a group-dynamic kind of thing than simple direct manipulation.

For example, aggregate (?) ideology determines how (un)employment is defined, measured, estimated, evaluated, controlled. As a consequence, you get things like excluding people who have not made "specific efforts" to find a job within 4 weeks, rhetoric about "jobs that (US?) Americans won't take", branding the unemployed as pampered slackers, policies of getting petty offenders (e.g. minor drug possession) "off the street", birth/death adjustments, contract workers who cannot secure a job with benefits reporting themselves as employed when "between contracts", people reluctant to admit they are unemployed because of the stigma, etc.

One can argue it's even larger and indeed systemic -- employers converting to contracting/outsourcing to dump cost of benefits and termination, etc.

The unemployment stigma reminds me of an anecdote -- I took my car to the shop and took the shop's shuttle driving people to their offices. When determining the route (by mostly stating corporate addresses), one of the passengers gave a residential address, and indicated he was "working from home". The driver, an elderly ex-military guy, promptly misunderstood this as the passenger being unemployed, and responded by saying, these days it is tough to keep a job, but it's good when one has makes an effort to keep up the skills and attitude. The other guy made a face, but did not clarify further.

Thanks fatbear. Wait til I get my hands on him...

fatbear: Indeed, thanks for pointing out.

Lindsey: Productivity is harder to measure, or even define, with activities that don't produce easily quantifiable output (and where quantitative measures are relevant). For example, consider consulting services. What is the unit of output there (and input)?

Or software development. Number of copies sold depends on paying demand, and marginally ability to press CD ROMs, or put up download server capacity. How does it relate back to software engineering productivity? The software engineer will do (roughly) the same amount of work whether 10 or 10,000 copies are shipped. You can go to revenue per employee, lines of code per employee, etc., but those are all wishy-washy. When revenue numbers are not met, some engineers will be laid off, based on some cost model. Does that mean the productivity of the firm will increase?

Back to retail, your "productivity" depends on how well your store is frequented, and how much (and what) customers buy. There is a fixed cost for running the store, and at least one checkout must be (semi) manned even when only 1 customer comes in per hour.

dryfly: You know your stuff. Unfortunately (?), this comes back to previously discussed points of industrial complexity. With somebody like your's experience of how the business works, it is apparent that what goes into making widgets and getting them into customers' products is a complex interaction of many people across several disciplines and levels of intermediaries, with a significant time component.

After Business Administration 101, one could think you buy a few machines, hire some accountants, managers, and guys who can show up for work, and that's it.