Gmm

[Bin Dohry]

I the GMM model with a country level dependent variable is it ok the regress for firms level independent level?

DV = number of firms listed in the country
IVs= firms stock liquidity, volatility, ROE and ROA

if not ok what are the model can be suggested in such data

thanks

[Fei Wang]

Technically you can do that. But a more reasonable research question to me would be how firm characteristics affect its probability of being listed. In other words, the DV should be a dummy indicating whether a firm is listed or not.

[Bin Dohry]

Technically you can do that. But a more reasonable research question to me would be how firm characteristics affect its probability of being listed. In other words, the DV should be a dummy indicating whether a firm is listed or not.

first if dummy variable can not use the GMM, then we need to use logit
the answer for your question is the characteristics of firms like those with small size or high/low liquidity will encourage firms to pursue listing

so small firms can list for growth purposes and so on
the GMM will enable us also to examine wither the firms decision on the prior year will encourage firms in the current year to list