>Look, I don't want to play a layman on these issues, since I'm not, but
>most people sort of expect when they hear "90,000 jobs were created last
>month" that this means that the Commerce Department
Bureau of Labor Statistics, actually.
> went out and found that
>many new jobs, period. All adjustments add to the uncertainty that the
>numbers reported reflect reality rather than problems in the statistical
>adjustments.
>
>Of course, the raw numbers can be deceptive as well in how we interpret
>them due to what is not sampled in the jobs report.
>
>But the statistical adjustments do lead a lot of folks to wonder how much
>change in the economy is due to underlying reality and how much is due to
>odd statistical decisions made by the government statisticians.
Crudele plays to a dumbass suspicion of pointy-heads and bean-counters which is completely wrong. The adjustments and imputations produce a more accurate picture of what's going on than the raw numbers would, and to stoke suspicions like Crudele does is a real disservice.
>Doug may feel that Crudele is accusing the BLS of political complicity with
>Bush, which I am not, but I have far less confidence that they've got the
>stats "right."
They're better than they ever were. Recent benchmark revisions, which are based on near=100% coverage of the employment universe (through unemployment insurance records) have been very small, much smaller compared to those of 10 years ago, when the BLS used a different method of estimating business startups and failures. They've done preliminary comparisons of the estimates for the fourth quarter of 2003 with the unemployment insurance records and found very little error. Also, their relatively new seasonal adjustment technique has proved to be much better at isolating underlying trends than the older method. So you're wrong here - they're much closer to right than they've ever been.
> I attribute this far more to poor funding of the
>statistical agencies, so they depend on poor data and have to extrapolate
>from there, but it's a problem that we are making large economic decisions
>based on pretty poor data.
No, the data is not poor at all. You claim not to be a layman on these issues, but you're talking lik one.
Doug