[R] Tobit Regression with unbalanced Panel Data

[peter dalgaard]

On 06 Jan 2017, at 15:08 , Vanessa Romero <vanrome54 at gmail.com> wrote:

> BHHH maximisation, 150 iterations
> Return code 4: Iteration limit exceeded.
> Log-likelihood: -66915.77 on 10 Df
> 
> How can I calculate McFadden's adjusted  R2 in R?

Google gets you there soon enough (e.g., "mcfadden r2 in r tobit"). One of the hits point to a Stata FAQ, explaining why McF's R^2 is nonsensical for tobit models....

> How could I reduce iteration?

Better starting values? In the absences of that, I think you want to _increase_ the limit, so that you are more sure that the procedure has converged. Also, a logSigma of -5.4 suggests that you are working with small numbers -- it sometimes helps to scale things by a factor of 100 or 1000.

-pd

-- 
Peter Dalgaard, Professor,
Center for Statistics, Copenhagen Business School
Solbjerg Plads 3, 2000 Frederiksberg, Denmark
Phone: (+45)38153501
Office: A 4.23
Email: pd.mes at cbs.dk  Priv: PDalgd at gmail.com