Stata The Stata listserver
[Date Prev][Date Next][Thread Prev][Thread Next][Date index][Thread index]

Re: st: long-term parameters in dynamic panel data

From   Giovanni Bruno <[email protected]>
To   [email protected]
Subject   Re: st: long-term parameters in dynamic panel data
Date   Wed, 6 Jul 2005 12:55:58 +0200

"L. Surdeanu" <[email protected]> wrote:

> Does anybody know how to estimate in Stata long-term parameters and their
> standard errors in dynamic panel data, using Arellano-Bond estimator?
> For short run I used "xtabond" and obtained short-run estimates of exogeneous
> variables (b)  and estimate of lagged dependent variable (a).
> I can calculate the long-term parameter estimates (by hand) using the formula
> b/(1-a), but I have no ideea about their respective standard errors, or if
> it's an easier way to do it.

That long run estimator is a continuous function of short run estimators 
a and b, as such it is consistent provided that a and b are consistent.
But this is a large sample property.

In small samples it can be severely biased for two reasons mainly: 
1) b and a are affected by small sample bias and 
2) E(b)/(1-E(a)) != E(b/(1-a)), 

This basically implies that in small samples even if you succeeded 
in bias-correcting a and b, then there would be still another source of 
bias to accommodate, which is due to non-linearity.  

A good reference on this is [Pesaran, M.H., Zhao Z., 1999. Bias 
reduction in estimating long-run relationships from dynamic 
heterogeneous panels.  In: Hsiao, C.,Lahiri, K., Lee, L.-F., 
Pesaran, M.H. (Eds.), Analysis of Panel Data and Limited Dependent 
Variables. Cambridge University Press,Cambridge.]

Hope this helps

> *
> *   For searches and help try:
> *
> *
> *

Giovanni S.F. Bruno
Istituto di Economia Politica, Universit´┐Ż Bocconi
Via U. Gobbi, 5, 20136 Milano
tel. + 02 5836 5411
fax. + 02 5836 5438
*   For searches and help try:

© Copyright 1996–2024 StataCorp LLC   |   Terms of use   |   Privacy   |   Contact us   |   What's new   |   Site index