Luis Loret de Mola
Jan 8, 2018 · 1 min read

Awesome post !

I would try a VAR model here.

This would allow you to use lags on the “exogenous variables” that affect pricing. And hopefully improve prediction at time t since you have more relevant info at time t -1 .

Note : Under VAR, these variables will be treated as endogenous so your initial assumption changes a bit.

Anyway, I really enjoyed your post.

Amazing work.

    Luis Loret de Mola

    Written by