Ordered Logit Regression

[Andreas Baltin]

Dear Statalisters,

I have a quick question regarding ordered logits.

I have a panel dataset with individuals and t=2. I use the following xtset

xtset individual year

I then run a simple ordinal logit

1. xtologit y x1 x2 x3 x4, vce(cluster state)

Now my question: ALL individuals lived in the same state in period 1 and period 2. I.e. no one moved, but people are living in different states. I cluster for states, as can be seen above.
Does it make sense to include dummies for each individual state in the regression? I can not really wrap my head around it. It would then look.

2. xtologit y x1 x2 x3 x4 i.state, vce(cluster state)

Is 1. or 2. better? Or does it depend on other factors?, If yes - which ones?

Many thanks in advance,
Andreas

[Carlo Lazzaro]

Andreas:
I would go a bit different (or halfway between your codes):

xtset individual year
xtologit y x1 x2 x3 x4 i.state, vce(cluster individual)