Hello,

My data is at the product-level. However, many of the covariates in my specification of interest are at the firm-level. Multiple products could have been introduced by the same firm in the data set. I have 431 observations at the product-level introduced by 229 firms. My dependent variable is a count variable; therefore, I chose to run a Poisson model with standard errors clustered on firm ids. Due to the different units of analysis, I also ran a multilevel model with the command mepoisson with firm ids as the random intercept. However, the results for the two models are exactly identical. I'm confused as to why this is the case. Does clustering on the higher unit of analysis in the cross-section Poisson model account solve the problem of different units of analysis for the covariates?

Thank you in advance.
Austin