xttobit with no observation

[Patrick Schulze]

I am investigating how certain factors affect hedging. For this I used xttobit. Now my exogenous variable "HEDGING" consists of a total of 450 observations and 50 of them are empty. I thought if I set xttobit's left censoring to zero, it works. But that doesn't work, stata says I have over 200 left censored observations (he also takes all zeros with him). How can I set the censoring so that it doesn't look at the empty lines?

My idea was to enter a -2 anywhere where it is empty and then enter -1 for the link censoring. This works, but I think the estimate is biased.

I hope you can help me,

many greetings,
Patrick

[Patrick Schulze]

no one?

[Phil Bromiley]

You didn't get a quick answer. You'll increase your chances of a useful answer by following the FAQ on asking questions - provide Stata code in code delimiters, readable Stata output, and sample data using dataex.

I'm having trouble following your explanation. I have no idea what it means when "he also takes all the zeros with him" means. You also refer to link censoring but I associate links with generalized estimators rather than xttobit..

It is odd that it treats missing as censored, but if the issue is missing treated as censored, then the easiest thing would be an if hedging <. to drop all missing observations.