How to Do xtabond2

1
How to Do xtabond2
David Roodman Research Fellow Center for Global
Development
2
xtabond2 in a nutshell

• First ado version in 11/03, Mata version in
  11/05.
• Extends built-in xtabond, to do system GMM,
  Windmeijer correction, revamped syntax
• Estimators designed for
• Small-T, large-N panels
• One dependent variable
• Dynamic
• Linear
• Regressors endogenous and predetermined
• Fixed individual effects
• Arbitrary autocorrelation and het. within panels
• General application

3
Outline of paper

• Introduction to linear GMM
• Motivation and design of difference and system
  GMM
• xtabond2 syntax

4
Black box problem

• Canned sophisticated procedure
• Dangers in hidden sophistication
• finite sample ? asymptotic
• Users should understand motivation and limits of
  estimator

5
Linear GMM in one slide
6
Linear GMM and 2SLS
7
Linear GMM in another slide(Holtz-Eakin, Newey,
and Rosen 1988)
8
Difference and system GMM
9
Problem Dynamic Panel Bias (Nickell 1981)
10
Partial solution OLS in differences
11
Problem Other endogeneity
12
Solution Instrument with lags (2SLS)(Anderson
and Hsiao 1981)
13
Problem Inefficiency
14
Solution GMM GMM-style instruments(Holtz-Eakin
, Newey, and Rosen 1988)
15
Problem Autocorrelation
16
Solution Restrict to deeper lags
17
Arellano-Bond AR() test
18
Problem Weak instruments
19
Solution Instead of purging fixed effects, find
instruments orthogonal to them(Arellano and
Bover 1995)
20
Relationship among moments(Tue Gorgens)
21
Problem Two-step errors too small
22
Problem, contd
23
Solution finite-sample correction(Windmeijer
2005)
24
(No Transcript)
25
Problem too many instruments
26
Solution consider limiting instruments
27
xtabond2 syntax
28
Examples
29
Examples, contd
30
Run times for bbest (seconds)700 MHz PC
xtabond2 ado 57
xtabond2 Mata, favoring space 14
xtabond2 Mata, favoring speed 11
DPD for Ox 3