f*******i 发帖数: 8492 | 1 看到一个评论,蛮有意思的
-----------------
Funny thing about that increase in unemployment: the household survey
actually showed an increase in employment of 400,000 (which would be
ridiculously good if it was true), but an even larger increase in labor
force participation.
Of course, the household survey is an inaccurate piece of shit at high
frequencies, so we shouldn't take this number very seriously. But to the
extent that we care about that number, it's actually good news.
Note also that the crappy establishment survey number is all coming from the
seasonal adjustment; removing the seasonal adjustment, there was an
increase in employment of 800,000. Of course, seasonal adjustment is done
for good reasons, and the SA numbers are surely more accurate than the NSA
numbers; but I find it hard to believe that seasonal adjustment is accurate
enough that structural errors of, say, 75,000 might not crop up by accident.
(Everyone's talking about the mild winter; that's certainly a possibility.)
Add that to the fact that the 95% confidence band on the establishment
survey is already +/- 100,000, and you get that these month-by-month
fluctuations are pretty much meaningless. It's sad how obsessively certain
pundits follow each twist and turn of such error-ridden data. |
|