Markov Transition Matrices in Stata

[Guido Franco]

Dear Statalists,

I am performing a research in which I am interested to check how my variable of interest moves across states over time.
More specifically, I have panel data and, for each year, I divide my panel-id into 5 groups according to some thresholds. I want to see whether (and the probability) they fall in the same group next years and which is the probability they change to another group. Basically, I want to estimate a transition matrix in a panel framework.

I read Nichols (2014), however my panel has no gaps:

1) Is then xttrans (or xttrans2) a safe procedure to estimate my probability transition matrix?

2) Is there a way in Stata to impose the transition matrix to respect Markov properties? Indeed, in a second moment I would like to check if an ergodic distribution exists.

Thank you in advance for your comments!

[Oliver lee]

hey I am trying to use STATA to carry out one-step Markov transition probability results but i have no idea how to do it even i have read something about it but i still dont have a clue. Would you mind to share something that you know already with me and that will really help me get out of it. Thanks!

[Maarten Buis]

Oliver: that is very little to go on. Read your question again and imagine you are someone who wants to help, knows something about Markov chains in Stata, but knows nothing about your project. What would you write as an answer? Either a (multi volume) book or nothing. You need make a more specific question. Often it helps if you include some example data, See the Statalist FAQ (link is at the top of this page) for how to include that. Then you can tell us what exactly you want to do to that data. The example data often helps force you to be more concrete, and pinpoint the exact question you want answered.