Bookmark and Share

Notice: On March 31, it was announced that Statalist is moving from an email list to a forum. The old list will shut down on April 23, and its replacement, is already up and running.

[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

st: Xtabond or dummy variables

From   Ayman Farahat <>
To   "" <>
Subject   st: Xtabond or dummy variables
Date   Mon, 24 Sep 2012 14:55:31 -0700 (PDT)

I have synthetic data set generated by  (fixed effects)

y_{i,t} = a_i + b_1 * y_{i,t-1} + e
where e is iid zero mean and unit variance

I have a 100 panel members and for each i have 10 observations. 

I tried to estimate ``b_1" using two methods

1) Using the Arellano bond 

2) Using OLS with dummy variables for panel ID. 

If i understand it correctly, the reason we use  the AB instruments is that when we difference , the lagged difference are endogenous and we need to use instruments. 
However, if i use dummy variables, i can in theory use dummy variables for the individual effects. 

I did a simulation and the AB approach always give a better estimate of the true value of "b_1" (lower bias). 

Any explanation /pointers as why that might be the case?


*   For searches and help try:

© Copyright 1996–2016 StataCorp LP   |   Terms of use   |   Privacy   |   Contact us   |   Site index