Prev: Database Speed
Next: array assignment issue
From: jia li on 2 Jul 2010 10:22
Dear Matlab community:

I am new to the group and recently started using Matlab for an econometric analysis task I am working. I would like to get some suggestions on the best way to tackle it using Matlab.

My research involves modeling a system of simultaneous equations with both discrete and continuous endogenous variables. I have three nonlinear consumer demand equations that are assumed to have joint normal distribution, and two discrete choice equations that have logistical distributions. I want to use two strategy to estimate the system: limited information maximum likelihood (LIML) where the system of demand equations are analyzed and parameter constraints will be imposed to estimate the choice equations. I will use iteration to ensure convergence.

A second strategy will be a full information maximum likelihood analysis, where the entire system is analyzed, perhaps using the starting values from the LIML procedure.

I have codes for the discrete choice models, but need to write my own for the system of nonlinear equations. I have seen -fsolve- being used for nonlinear system equations, but not sure whether it uses maximum likelihood method. Would I be better off to write my own maximization code given the other steps involved?

If you have any suggestion on examples related to two-step MLE and FIML, I would highly appreciate receiving them!

Thank you so much and I look forward to hearing from you!

Jia
 | 
Pages: 1
Prev: Database Speed
Next: array assignment issue