11 Comments
Sep 24, 2023Liked by Claudia Sahm

Such a great headline to this article. So many of us get caught up in talking about GDP for which the real number last time I checked is expected to be around 4.9% for the 3Q. One would think that is as solid as it gets. I guess in hindsight if the fed had focused more on GDI and GDO that showed much weaker growth that maybe the fed wouldn't be in such a rush with its rate hikes. Of course all this important econ data this week to me could be overshadowed by a looming government shutdown in DC for which there is no deal in sight.

Expand full comment

Now this is “the scoop.” Thanks for catching it early and sharing your take — truly great reading and setup for the revisions!

Expand full comment
Sep 24, 2023Liked by Claudia Sahm

So very very important - great article.

Frequent and often substantial revisions undermine the case for simple rule-based approaches, and reinforce the case for risk balancing and humility.

Our inflation target of 2% has produced 2 deflation scares this century, which is too many in my view. Errors in contemporaneous data imply that we need a greater cushion on top of that. There's a case for a 3% inflation target.

I'm a fan of nominal GDP targeting, but in light of this article, we really need to backtest such an approach, even if this targeting is done informally.

Revisions are a familiar problem in this field but we nonetheless need to think even harder. Perhaps analysts should estimate measurement errors - do these errors increase during economic turmoil?

Expand full comment

Great work on this one Claudia!

Expand full comment
author

thank you!

Expand full comment

Whoa - looks like the revisions came in the other way! GDI was revised up to match GDP. Just goes to show how hard it is to calibrate these takes! 2023 Q1 GDI growth revised from -1.8% to 0.5%...

Expand full comment

If, as it should, the Fed is running its inflation machine with an idea of allowing relative prices to adjust to shocks so as to minimize failed transactions( = un/under employment of resources) and maximize successful transactions( = real GDP), then learning that there had been more failed transactions/not as many successful transactions than one though ought imply that the inflation machine should have been running a little faster than it was and that the model linking the settings of the Fed's policy instruments needs to be recalibrated. The new information from the data revision would not itself lead to a resetting of policy instruments, but it would make the Fed more likely to react to new information with a pro-inflation direction than would have the un-recalibrated model.

Expand full comment

If the Fed is running its inflation machine with an idea of allowing relative prices to adjust to shocks so as to minimize failed transactions( = un/under employment of resources) and maximize successful transactions( = real GDP), then learning that there had been more failed transactions/not as many successful transactions than one though ought imply that the inflation machine should have been running a little faster than it was and that the model linking the settings of the Fed's policy instruments needs to be recalibrated. The new information from the data revision would not itself lead to a resetting of policy instruments, but would make the Fed more likely to react to new information with a pro-inflation direction than would have the un-recalibrated model.

Expand full comment
author

It would depend on the revision.

Expand full comment

There was an article in the Financial Times as couple of weeks ago about how some countries in average revise their GDP numbers up and others down. US revised down.

Expand full comment
author

Interesting. I will look for that.

Expand full comment