Panel and interpretation of baseline

Hi,

hope to find you well.
Basically, I would like to ask you a twofold question. First it's just a confirmation. I read the paper "Orthogonal ML for Demand Estimation: High Dimensional Causal Inference in Dynamic Panels" (Chernozhukov et al. 2018) and learned about the possibility of performing panel analysis. I had a discussion with one of you few months ago and understood that the panel structure is not yet available. I would like just to have a confirmation about that, please.
My aim is to estimate own and cross price elasticities of a set of products. Following what you suggested to me (i.e. creating a treatment, T, with prices of product "I" and average price of all other products; X containing time and fixed effects), I obtained the following CATE result for 9 random products:

array([[-1.4766871 , -1.93004447], [-2.41711183, -2.54188471], [ 3.90244306, 3.83095042], [-1.32182939, 0.8783108 ], [-0.83265341, -2.14551779], [ 1.24155941, -9.64598452], [-0.52252569, 18.90605457], [ 5.99370699, 2.06460821]])

So, please, my question is, how can I interpret the result with respect to baseline? Can I basically say that the results are with respect to the own and cross elasticity of the baseline product? So for instance, if baseline is product 0 that does not appear, product 3 has an own elasticity that is 4 times (3.90) bigger than the own elasticity of the baseline product and across elasticity that is 3.80 times bigger? (hence the baseline product has an elasticity normalized to 0 as, for instance, with multicollinearity?)

Thanks a lot,

Federico

No file, test, or entry point is identified in the issue. Start by checking EconML's support for panel structure and CATE interpretation for the described treatment setup; done means confirming whether panel analysis is available and explaining how the estimates relate to the baseline.