Stata: unit-root test on unbalanced panel data

[John Michael Perdue]

This question is cross posted here.

I have the following data:

    import excel "\\dfs\wiwi\home\Perdue\Desktop\statafiles\unitroot.xls", sheet("Sheet1") firstrow clear
    
    encode cid, gen(panel_id)
    
    drop cid
    
    xtset panel_id date
           panel variable:  panel_id (unbalanced)
            time variable:  date, 2006 to 2014
                    delta:  1 unit
    
    xtunitroot fisher leg_totalbills, dfuller trend demean lags(1)
    performing unit-root test on first panel using the syntax
    dfuller leg_totalbills, trend  lags(1)
    returned error code 2000
    r(2000);
    
    count if mi(date, years_exp, leg_totalbills, panel_id)
        0

I simply need to do a unit-root test. I have tried using R but was unable to find a package that could handle unbalanced panels. I thought that Stata would be easier but I have encountered consecutive obstacles. This question had evolved from here and here.

[Carlo Lazzaro]

John Michael:
the error code r(2000) creeps up when your code does not match your observations (missing values is one possible cause, but not the only one).
I'm not familiar with unit-root test; however, I would take a look at -lags- as a possible source of problem.

[Isabel Cour]

Dear Carlo,

According to the theoretical description of unit root test and the examples, unit root test may be to do on the dependent variables. however I can see that on each independent variable separately is also applied. Do you have a theoretical article where I can read more about it. It is not still clear for me why I should do for independent variables. Then what happen if you have the situation where one of the variables show that one of the panel is not stationary? I really appreciate any comments. Regards