Hi all,

I am using stata to estimate demand for automobiles. I have panel data of 4 years 2014-2017. To have my estimates i am using Berry, Levinsohn and Pakes (or BLP) model, but without demographic data. i am trying to apply the first part of the model (mean utility part: ln(sj)-ln(s0)= -a*price+b*product characteristics+error term). As you can see my dependent variable is log difference of market share and outside good.
To get my results, i am using ivregress command as follows

ivregress 2sls d_share (price_t= localization) fe size weight engine1 origin chassys_2wd chassys_both suv hatchback minivan b1-b50 i.year, first

results seems ok, but now i need to calculate cross and own price elasticities. i tried to use -margins- command but i failed, i think i am using it in wrong way.

Could you please advise how can i use margins command to have elasticities.

Thanks,