hetroskedasticity test for random errors model

[nikhil shukla]

i am working on 20 projects with data from 15 years. I have performed the Hausman test that recommended using the random effects model for my analysis. I have checked the random effect model for autocorrelation and found no autocorrelation in the model. Now I want to check for heteroskedasticity in the random model. How can I do it?

Also, suggest whether should i use random effect or any other dynamic model

[Manh Hoang Ba]

You can use the test proposed by Montes-Rojas & Sosa-Escudero (2011). Run the xttest4 command after estimating REM to perform this test.