0

I am refreshing my knowledge with respect to econometric modelling in general.

I came across page 201 of the book 'Enjoyable Econometrics by Philip Franses' and I had some difficulties interpreting the results of a censored regression model.

The estimation results of the censored regression model are depicted in the picture below.

enter image description here

The dependent variable is given as:

enter image description here

Then the interpretation of the censored regression model continues as follows:

enter image description here

Question: Why is the estimated variance of the errors $(\frac{0.964}{2})$ used for the estimation?

I believe this perhaps has to do with the fact that we try to estimate unobserved values $(y_i^*)$ instead of the observed and censored $(y_i)$ observations. Note that the errors are assumed to be normally distributed.

Useful I used to understandm, but unfortunately could not grasp fully to answer my question:

Documentation SAS censored regression

Related CV-post

Nadia Merquez
  • 140
  • 2
  • 9
  • See this [post](https://stats.stackexchange.com/a/93103/7071). and also [this](https://stats.stackexchange.com/a/49868/7071) one. This has to do with logging the outcome to make the assumptions of censored regression more realistic, but still caring about the outcome in levels. An alternative would be to use the Duan smearing approach. – dimitriy Dec 10 '19 at 19:25
  • Thank you, I will look into it. – Nadia Merquez Dec 10 '19 at 20:48

0 Answers0