Hi
Do anyone read the paper of Galvao (2011) about quantile regression for dynamic panel data with fixed effects? I'm a little bit confused since the paper is too technique. At the begenning of the paper Galvao wrote that the individual fixed effects are restricted to be independant of the quantile order and the restriction can be implemented by estimating the model for several quantiles simultaneously. I think he is talking about restricted regression types (such as Lasso). However, When he explained later how to use instruental variables for the lagged dependant variable, in the process, the individual effects depend on the quantile and the estimation is sequentially for a grid of quantiles values!!
Do any one can help me about this?