5 Comments
User's avatar
BTJ's avatar

Which recession(s) exactly has the Sahm “Rule” predicted exactly?

Expand full comment
Alexander Campbell's avatar

Fair. "Predicted" is probably the wrong word. Indicated. Qualified. Guess that's part of the problem, is a recession just when you have two quarters of real GDP contraction or is it something broader/different. Does it even matter? All great questions!

Expand full comment
BTJ's avatar

The answer is none, never.

It is a backtested model, with an n of 11. It’s a joke. Ultimate correlation without causation, and overfitting. There are a billion other variables that could be backtested to correlate with 11 instances. You don’t have a rule that has NEVER forecasted a single time in advance. Not a serious tool for serious people.

Expand full comment
Alexander Campbell's avatar

Did you see the part where I said you could replace it with a 12m change?

Expand full comment
BTJ's avatar

Yeah. So it’s not a “rule.” It’s made up fungible BS that has no predictive value.

Expand full comment