Huge Difference Between GDPNow and Nowcast Models


The Atlanta Fed GDPNow and the New York Fed Nowcast GDP models once again are wildly different.

Wide Discrepancy

  • In the past week, the GDPNow model forecast rose to 2.0% from 1.5%.
  • In the past week, the Nowcast model forecast fell to 0.6% from 0.8%.

Not only were the starting points very different, the models responded in opposite fashion to the same economic news.

GDPNow rose 0.5 percentage points while the Nowcast fell 0.2 Percentage points, a net difference of 0.7 percentage points.


The models do not react to the news itself, but rather to what happened vs what the model expected to happen.

The economic news was disappointing to Nowcast expectations, but hugely positive to the GDPNow expectation.

Blue Chip Forecast Via GDPNow

Image placeholder title

The range of the Blue Chip professional forecasts is 1.0% to +2.0%.

GDNow is at the very top end of Blue Chip whereas Nowcast, at 0.6% is 0.4 percentage points lower than the lowest Blue Chip estimate (but that Blue Chip estimate is a bit stale).

Jobs Data

For a discussion of today's jobs data please see:

  1. Jobs Surge in Strike-Ending and Seasonal Adjustment Rebound
  2. Trump Tweets "Manufacturing Blowout": What's the Real Story?

Mike "Mish" Shedlock

Comments (13)

How can anyone possibly take any economic numbers seriously coming out of this administration? (and its side business, the Fed). GDP, inflation and unemployment numbers are so squishy and massaged as to be useless. The stock market was a real, market-based number for awhile, but now even that is compromised.

Tony Bennett
Tony Bennett

"The models do not react to the news itself, but rather to what happened vs what the model expected to happen."


Yes. And Q4 can be boosted by downward revision to Q3 numbers.

Take yesterday's Monthly Wholesale report (October) on Sales/Inventories ... report showed a +0.1% gain in inventories. Bullz relish as "sign" business has confidence in economy so adding inventory.

Well, looking under the hood -

Inventories previous report ... September $676.710 billion

This report ... September $674.897 billion ... October $675.573 billion.

If not for downward revision to September, headline for Inventories would have been negative ... September sales revised downward, as well. Inventories / Sales creeping ever higher ... October 2018 1.30 ... October 2019 1.37 ... which is very close to cycle high of 1.38 last seen in near recession of early 2016. Any higher and we're talking 2009.

Tony Bennett
Tony Bennett

The consumer is still hanging in there. Courtesy of ever increasing load of debt, of course.

Yesterdays consumer credit report (October)

Consensus ... +$15.8 billion range of "experts" ... +$11.6 billion to +$17.6 billion

Actual ... +$18.9 billion.

The economy will continue its low quality growth (debt based) until the delinquency / default monster storms into the room. The first rumblings are present. I fully expect it to escalate quickly in the near / mid term.


Personally, I appreciate all the models, all the statistics, all the numbers; whether they come from government, economists, think tanks, etc. I also appreciate the work Mish performs, analyzing the available data. (Thanks Mish; you are why I keep coming back.) The main reason I appreciate this micro data is because I can’t be bothered to gather it myself.

All of the data comes with a caveat; none of it is acccurate. It is all an estimate; an approximation; a best guess.

So when people say “don’t trust the data“; I can understand. When they say it is fake; I simply laugh at them and tell them to go get their own data.

Personally, I take a macro approach. Use the available data; pay attention to the trends; understand the reality around the information. And importantly, ignore the politics; it blinds people. Like the moronic Trump supporters who post on this site, thinking Trump is some kind of genius. You can’t get more stupid than that. So sad.

I see no reason to change my opinion regarding economic growth yet. I expect slow growth to continue for the foreseeable future; no recession; no stock market crash. Slogflation; slow growth, low inflation.

Until someone significant drops a ball.

And I am realistic enough to know that I have no idea when that will happen.



@scoot the GDPNow model has generally been the more accurate of the two. Not always, and it is far more volatile


One model might be slightly less bad than the other, but both models are garbage.

Academics, in university and staffing the Fed's "research" divisions, are people who can't hack working in the real world.

Academic models are useful tools to help understand what already happened, but in that role the models are incomplete.

Used for forecasting, the models are garbage. TV personalities use them to fill air time.

Global Economics