Skip to main content

The Atlanta Fed GDPNow Model forecast for the 4th quarter fell a half a point this week whereas the New York Fed Nowcast Model rose a tenth of a point .

GDPNow Tracking

Image placeholder title

The yellow highlight shows the half-point drop on January 16.

Analysts, including me, generally concluded sales were solid, rising 0.3% in December, yet the GDPNow forecast fell.

I commented Retail Sales Mostly Solid But Department Stores Hit Again.

Q. Why the forecast decline?

A. The model expected better.

It's not the report that matters to these forecasts, but rather how well the forecast dose vs. what the model expected.

Housing and Industrial Production Balance Out

On January 17, I noted housing starts rose to a 13-year high.


I would have expected the GDPNow model to surge but nothing happened because Industrial Production balanced things out.

I commented Industrial Production Numbers Confirm Bias of Housing Report.

Weather Impact

"The drop for utilities resulted from a large decrease in demand for heating, as unseasonably warm weather in December followed unseasonably cold weather in November," according to the report.

It was that unseasonably warm weather in December that totally skewed the Housing report.

For example:

In the Midwest region, not-adjusted single-family housing starts rose from 9,200 to 9,600. This was reported at a seasonally-adjusted annualized rate of 180,000 and a whopping rise of 56.5%from November.

For more details, please see Housing Starts Surge to a 13-Year High Thanks to Massive Seasonal Adjustments

Word About Models

Models do not take weather or any other one-time issues into consideration.

Looking ahead, the GDPNow model is likely going to make some housing forecast assumptions that will not happen. Thus, from where things stand now, the model is likely on the high side.

Mike "Mish" Shedlock