Friday, July 22, 2011

Duration models with measurement error

The issue of how to think about duration models when there is measurement error in the dependent variable came up in a discussion the other day. If anyone out there in reader-land has a pointer to a paper or two on this, please post a comment or send an email.

3 comments:

Dave Giles said...

Hi - have you checked Daniel Gutknecht's recent working paper?

John said...

There's a paper by Abrevaya and Hausman (http://econ-www.mit.edu/files/1041) where measurement error in the duration is treated as unobserved heterogeneity and they propose a consistent semiparametric estimator.

econjeff said...

Thanks for the helpful suggestions!