If I didn't know you better, Doug, I'd have guessed the "So let me see if
I've got this right" gambit is a holdover from your adolescent right-wing
Yale period. You can take the boy out of the Buckley, but you can't take the
Buckley out of the boy. Jez kiddin'.
No, of course you don't have it right. The BLS can produce stats based on
whatever technique it thinks is appropriate to whatever purpose. Smart folks
should understand that whatever stats comes out of the BLS will be used for
some purposes it's not suited for, often by people who should know better.
The briefness of turning points is not the issue. Someone could have a
perfect driving record for 35 years, take his eyes off the road for 5
seconds and plow through a crowd of school children in an intersection. The
casualties will be just as dead or maimed as if the guy was chronic violater.
Trust me on this, Doug. I'm not a reader/consumer/recycler of ready-made
statistics. It's my job to mold fully formed stats out of raw data. I'm good
at it, too (references on request). There's more than one way to skin a stat.
But -- so let me see if I've got this right -- "plenty of other numbers"?
"an accurate and almost real-time picture of what's going on"? There's a
saying, "do you want it now, or do you want it done right?" To glibly claim
that numbers can be simultaneously accurate and almost real-time is to
disparage the hard work that statisticians do and the complexity and
ambiguity of the materials they work with. How many great meals have you had
that were prepared in a microwave? Numbers that are BOTH accurate and
instantaneous (stock indexes, for example) are fundamentally trivial, which
is evidenced by their volatility.
Doug Henwood wrote,
>So let me see if I've got this right - the BLS shouldn't use a more
>accurate technique because there's a one in ten chance it will be
>briefly inaccurate? Turning points, after all, are even briefer than
>recessions themselves - we're talking about a few months out of many
>years. And they produce plenty of other numbers - e.g. the household
>survey and the unemployment claims figures - which do give an
>accurate and almost-real time picture of what's going on.
Tom Walker
Bowen Island, BC
604 947 2213
