No announcement yet.
  • Filter
  • Time
  • Show
Clear All
new posts

  • RCL random coefficient logit stata module, issue with elasticities() option

    My issue is regarding the rcl command < > for estimating the random coefficients logit model of BLP:
    Berry, Steven, James Levinsohn, and Ariel Pakes. "Automobile prices in market equilibrium." Econometrica: Journal of the Econometric Society (1995): 841-890.

    I am estimating this model with a random coefficient on price. My data set is an unbalanced panel in which products are not present in every market. This involves approximately 30 firms producing 300 total products. Utilizing the elasticities() option produces a matrix in which the elasticities associated with some firms or some products consist entirely of zeros. Dropping observations from these firms does not alleviate this issue, but instead results in a different firm experiencing elasticity rows and columns consisting entirely of zeros.

    I believe this may be an issue of averaging elasticities across markets. If I restrict my data set to only include a single market and run my random coefficient model, the elasticities() option runs correctly with elasticities estimated for every firm or product. However, these elasticities are calculated using separate estimates of my coefficients for every market and will not be appropriate for my research.

    Has anyone else experienced this issue? Any insight will be greatly appreciated.

  • #2
    Welcome to Statalist, Jacob.

    You should perhaps contact the author of the rcl command at the email address found at the link you provided, as well as in the output of help rcl and net describe rcl, from( The command has a 2016 release, so there's a good chance the author is still available and interested in helping.


    • #3
      I will do this, thank you.