Identification for difference in differences with cross-section and panel data

被引:53
作者
Lee, Myoung-jae
Kang, Changhui
机构
[1] Korea Univ, Dept Econ, Seoul 136701, South Korea
[2] Singapore Management Univ, Sch Econ & Social Sci, Singapore 178903, Singapore
[3] Natl Univ Singapore, Dept Econ, Singapore 117570, Singapore
关键词
difference in differences; identification; panel data; independent cross-sections;
D O I
10.1016/j.econlet.2006.03.007
中图分类号
F [经济];
学科分类号
02 ;
摘要
For difference-in-differences (DD), the identification question in using three types of progressively more informative data is addressed: independent cross-sections, 'mover' panels, and "no-mover' panels. Although DD identifies 'the average effect on the treated', its meaning and the identification conditions differ across the data types. (c) 2006 Elsevier B.V All rights reserved.
引用
收藏
页码:270 / 276
页数:7
相关论文
共 5 条
[1]   How much should we trust differences-in-differences estimates? [J].
Bertrand, M ;
Duflo, E ;
Mullainathan, S .
QUARTERLY JOURNAL OF ECONOMICS, 2004, 119 (01) :249-275
[2]   Unnatural experiments? Estimating the incidence of endogenous policies [J].
Besley, T ;
Case, A .
ECONOMIC JOURNAL, 2000, 110 (467) :F672-F694
[3]  
Krueger AB., 1999, HDB LABOR EC
[4]  
Lee Myoung-Jae., 2005, MICROECONOMETRICS PO
[5]  
Rosenbaum P.R., 2002, Springer Series in Statistics