Monday, November 23, 2009

If You Must Choose Between The Data And Your Model, Go With The Model


Who's Lying About Personal Spending?

Here are your possibilities:

* The BEA is lying. In the third quarter they claim that PCE changed +0.2, +1.4, and -0.6% for July, August and September, respectively, leading to an aggregate change of +1.2%.

* Business that remit sales tax are lying. The overall sales tax collections in the 3rd quarter were down 8.2% from last year's levels, and this is the fourth quarter in a row that year-over-year declines were posted.

One of these two reports is a lie.

One is a count of actual monies remitted by businesses in satisfaction of taxes collected by them from real consumers processing real retail transactions (that's the spending that matters in the real economy, right?)

The other, if you read the BEA methodology papers, has the word estimate peppered liberally throughout.

Which do you believe?

Do you believe that retailers are intentionally under-reporting and under-paying sales taxes?

Or do you believe the BEA's "estimates" are complete horsecrap and that the government is intentionally overstating economic activity?

When you have two radically different claims of measurement of the same activity (in this case consumer spending) that are impossible to reconcile within reasonable "measurement error" the conclusion one is forced to reach is that one of the two reports is false.

No comments: