Hi Florian,
Good point - ultimately I do want to do the PSID stuff in Julia, as I'm solving my numerical model in Julia and the estimation I'll do will require hand-coding an estimator in any case. As I want to start from a correct reference point though I thought I'd create a panel using your R package first :) Your ideas sound good, once I've gotten psidR to give me a good panel to replicate I'll try to implement this!
