How to fix errors in unit root test panel data

[Isabel Hermann]

Dear All,

I am trying to run a unit root test in order to find the degree of autocorrelation among my y variable ( which is closing prices of various indices). Since my data is unbalanced, I used the fisher xtunitroot test but, still, Stata only produces this error r(2000):

. xtunitroot fisher closing_px_0, dfuller lags(10)
performing unit-root test on first panel using the syntax
dfuller closing_px_0, lags(10)
returned error code 2000

I checked my data, and there are is no missing data and no string variable so I don"t know what is wrong. I read about the fact that my data on closing prices could be a singleton but how do I solve this problem so that I can perform this test?

In case you need it: Overall, I am trying to assess the impact of terrorism on stock markets using daily data on the main stock index of the respective country (closing_px_0).

Any help would be much appreciated.

Best,
Isabel

[William Lisowski]

How does this problem differ from the one resolved on the earlier topic to which you posted http://www.statalist.org/forums/foru...66#post1308266 ?

Specifically, you presumably have too few consecutive observations in your first panel to calculate enough observations with 10 lagged differences to perform the test. The suggestion on that thread at post #4 was to use the user-written xtfisher command (see the output of search xtfisher).

[Scott Merryman]

I would second William Lisowski's reply.

One way to identify singleton would be use egen's count() and then exclude those panels with less than 10 observations.

For example:

. webuse grunfeld, clear

. replace invest = . in 1/10
(10 real changes made, 10 to missing)

. xtunitroot fisher invest, lag(10) dfuller
(10 missing values generated)
performing unit-root test on first panel using the syntax
dfuller invest, lags(10) 
returned error code 2000
r(2000);

.  egen count =count(invest), by(com)

. table com, c(mean count)

-----------------------
  company | mean(count)
----------+------------
        1 |          10
        2 |          20
        3 |          20
        4 |          20
        5 |          20
        6 |          20
        7 |          20
        8 |          20
        9 |          20
       10 |          20
-----------------------

. xtunitroot fisher invest if count>10, lag(10) dfuller
(10 missing values generated)

Fisher-type unit-root test for invest
Based on augmented Dickey-Fuller tests
--------------------------------------
Ho: All panels contain unit roots           Number of panels  =      9
Ha: At least one panel is stationary        Number of periods =     20

AR parameter: Panel-specific                Asymptotics: T -> Infinity
Panel means:  Included
Time trend:   Not included
Drift term:   Not included                  ADF regressions: 10 lags
------------------------------------------------------------------------------
                                  Statistic      p-value
------------------------------------------------------------------------------
 Inverse chi-squared(18)   P         0.0000       1.0000
 Inverse normal            Z              .            .
 Inverse logit t(4)        L*             .            .
 Modified inv. chi-squared Pm       -3.0000       0.9987
------------------------------------------------------------------------------
 P statistic requires number of panels to be finite.
 Other statistics are suitable for finite or infinite number of panels.
------------------------------------------------------------------------------

. xtfisher invest, lag(10)


Fisher Test for panel unit root using an augmented Dickey-Fuller test (10 lags)

Ho: unit root

         chi2(18)     =    0.0000
         Prob > chi2  =      1.0000