If I want to do research in causal inference/causal ML, which PhD is better?

[statisticsphd]

I am broadly interested in causal inference and causal machine learning (e.g. double/debiased ML or meta-learners), and I am trying to decide between applying to statistics PhD programs and economics PhD programs (and specialize in econometric theory). I definitely can apply to the mix of both, but due to the extremely competitive nature of PhD admissions these days, it seems like a better strategy to focus on a single program and cast a wide net. I have done research in both fields and I think my profile is equally good for both programs. I don't have clear post-PhD plans yet, but I am slightly leaning toward industry than staying in academia (something like industry 6 : academia 4). My rough thought process is

Economics PhD

Pros: Excellent training in causal inference, get to/expected to work with interesting yet messy observational data in economics, better academic job market

Cons: mandatory first-year coursework in micro and macroeconomics and second-year field courses, seem to be obsessed with classical/old/outdated methodologies, will take a year longer potentially (median 6 years to complete), more stressful and longer publication process, worse culture (hierarchy, bureaucracy ...), might not be able to focus on computational side of research (where all the novel and interesting works with actual practical value are happening these days), expected to submit papers to journals rather than (ML) conferences, might not be perceived as good/smart as a statistician in the industry job market.

I think statistics PhD has roughly the opposite pros/cons to economics PhD. What do you think?

Edited September 19 by statisticsphd
.

[bayessays]

If you don't want to actually study economics, I think an economics PhD sounds like a painful journey.  I personally think even going to a biostatistics programs can potentially be a bad choice for someone who isn't interested in doing medical stuff, so this sounds like a more extreme version of that.

One thing I would consider is looking into how many econometricians are actually doing cutting-edge research in the topics you're talking about.  My suspicion is they're mostly only at the top econ programs which are extremely competitive, so I bet you'd have way more choices of good-fit programs if you choose statistics.

There is almost no chance you will get better causal inference training in an econ program than going to a stats PhD program with a good causal inference group.  You can work with interesting data in any stats program.  Are you sure the econ job market is that much better?

Personally, I could only imagine doing the econ PhD if I was excited about being an economist.  It really sounds like you want to be a statistician.