X hits on this document

106 views

0 shares

0 downloads

0 comments

28 / 31

would be used for ̂. Then

  • ̂ A T T

N

  • ̂ 1N1

i1

  • Wi p̂XiΔYi

  • 1 p̂Xi

.

is consistent and N -asymptotically normal. HIR discuss variance estimation. Imbens and Wooldridge (2007) provide a simple adjustment available in the case that p̂ is treated as a parametric model. Similar approach works for ATE Regression version: .

ΔYi on 1, Wi, p̂Xi, Wi ̂  p̂Xi, i 1, . . . , N.

The coefficient on Wi is the estimated ATE. Requires some functional form restrictions. Certainly preferred to running the regression Yit on 1, d1t, d1t Wi, p̂Xi. This latter regression

28

(40)

Document info
Document views106
Page views106
Page last viewedSun Dec 11 00:29:30 UTC 2016
Pages31
Paragraphs261
Words3934

Comments