Dave Backus writes:
We macroeconomists are thrilled with the Nobel prize for Sargent and Sims. But on causality: they spent more time showing how hard it was to identify causality than showing how to do it. And that’s a fair assessment of our field [economics]: causality is almost always in doubt.
More here.
If I were in a snarky mood, I’d say something like, Causality is always in doubt in economics . . . unless you’re talking about abortion and crime, in which case you can be absolutely certain.
But I’m in a good mood right now so I won’t say that. Instead I’ll just remark that, as a statistician, I’m positively thrilled that somebody named “Sims” received a major award.
Wow, this is passive aggressive.
Apophasis!
As a fellow Bayesian, you should be pleased that Sims was named. Not too many Bayesian econometricians out there!
Agreed!
Some lighter stuff with a statistical flavor:
Why Econometrics Should Always and Everywhere Be Bayesian (slides form a presentation)
Comments on Angrist and Pischke (critiques quite a lot)
On an example of larry wasserman (defending the bayesian viewpoint and fun)
Pingback: More reason to like Sims besides just his name « Statistical Modeling, Causal Inference, and Social Science
Isn’t causality always in doubt in every field?
Pingback: Through A Glass Darkly | Poison Your Mind
Have you previously weighed in on the abortion-crime argument?
This is all I’ve ever written on the topic.
I am wondering about a statement you made:”Causality is always in doubt in economics . . . unless you’re talking about abortion and crime, in which case you can be absolutely certain.” Why do you believe that the causality of crime (or abortion) is absolutely certain? If I do understand you right, you believe that the causes for criminal behavior are to 100% definable. As a social researcher/criminologist I doubt that thought. And as a PhD student I would be thrilled if you could elaborate on that.
Julian:
I was writing ironically. I do not actually think the abortion/crime link is certain. See here for more.