Whoops

The webosphere is talking about badly reported statistics from the BEA,via a blog post by Michael Mandel.

He points out something we have noticed, the necessity of deep downward revisions after the first estimate of nominal GDP output. If we look at the crash, a lot of intermediate businesses exited the field, and the ratio of consumer prices to producer prices went way up; business inputs dropped in price faster than we expected. So during the post crash era, the BEA underestimated the price drop for intermediate goods and and assumed the gains after the crash came from productivity gains.

Calculated risk twice noticed the downward revisions to GDP. The economy changed state, and the BEA uses the older state model for the new economy. Keynesians pull the same stunt.

What happened was not a collapse of the retail consumer, but a collapse of intermediate producers, leaving the field with much fewer suppliers. With the contracted supply chain, gains along the chain appear impressive. But the rank order in the channel, being reduced, means that the sum(n*log(n)) is smaller because n has dropped, fewer steps in the production chain.

We do not have intermediate solutions when we drop rank, the BEA assumes we do.  The result, to make matters simple, is that the BEA assumes a smaller amount of volatility, and ascribes a temporary inventory recovery to real growth. Inevitably, the BEA has to issue a downward revision.

The economy has and is continuing to contract.  I keep pointing out that M1 Velocity continues to drop, and the gain from specialization (PPI/CPI) have gone back to previous, and terrible, values.   We are going to drop rank one more time, have a bit of a Double Dip.