Unit Root Test Error

[David Brown]

Hi again,

I was trying to run a unit root test on my unbalanced panel data and I received this error. From my knowledge, the Fisher tests should work on unbalanced panel data? I am unsure what is happening. I have also tried other variables and same result.

xtunitroot fisher roe, dfuller lags(0)
performing unit-root test on first panel using the syntax
dfuller roe, lags(0)
returned error code 2000
r(2000);

xtunitroot fisher roe, pperron lags(0)
performing unit-root test on first panel using the syntax
pperron roe, lags(0)
returned error code 2000
r(2000);


Thanks!

[Carlo Lazzaro]

David.
as always, the error description is a good first place to start from:

search for r(2000)                                                                                                                                                    (manual:  [R] search)
-------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------

Search of official help files, FAQs, Examples, SJs, and STBs

[P]     error . . . . . . . . . . . . . . . . . . . . . . . . Return code 2000
        no observations;
        You have requested some statistical calculation and there are
        no observations on which to perform it.  Perhaps you specified
        if or in and inadvertently filtered all the data.

A careful scrutiny of your dataset and/or commands seems the way to go.

[David Brown]

Hi Carlo,

Thanks for your help again. I looked through my dataset; however, it doesn't appear to have any missing observations. It is just unbalanced with different years. I have attached my dataset.

Thanks!

[Scott Merryman]

Because you have singletons, when xtunitroot fisher calls dfuller it returns an error code:

. use "C:\Users\Scott\Desktop\Data.dta", clear

. dfuller roe if companyid ==1 ,lag(0)
no observations
r(2000);

You could try the user written command xtfisher (ssc desc xtfisher) which will ignore problematic cases:

. xtfisher roe, lag(0)


Fisher Test for panel unit root using an augmented Dickey-Fuller test (0 lags)

Ho: unit root

         chi2(1070)   = 2125.7340
         Prob > chi2  =      0.0000

[David Brown]

Hi Scott,

Thanks so much. What does singleton mean? I'm not familiar with the jargon. I will try the xtfisher. Thanks!

[Carlo Lazzaro]

David:
- missing observations are not the only cause for r(2000); more often, the reason of that call rests on the fact that your dataset has no observations that fulfill the requirements of the Stata command that you've invoked;
- you can see what Scott meant by singleton (i.e., some of your companies have observations that refer to one year only; so lags are not allowed) via:

tab companyid

[David Brown]

Thanks for the explanation Carlo! I understand now. I have used the xtfisher user written command from Scott and it works now.

[David Brown]

I am using both the ADF and PP from xtfisher and am receiving identical chi2 for each. Is this because of the singletons as well?
The code I am using is xtfisher var, lag(0) and xtfisher var, lag(0) pp. I was also wondering if I can still implement lags in these tests or would that not make sense? To clarify, I am unsure of how many lags to run in my data. Thanks again.

[Scott Merryman]

The results from ADP and PP are the same because you are using zero lags. The sample sizes are the same and the interpolated critical values for Z_t are the same.

. use "C:\Users\Scott\Desktop\Data.dta", clear

. dfuller roe if companyid ==10 ,lag(0)

Dickey-Fuller test for unit root                   Number of obs   =         7

                               ---------- Interpolated Dickey-Fuller ---------
                  Test         1% Critical       5% Critical      10% Critical
               Statistic           Value             Value             Value
------------------------------------------------------------------------------
 Z(t)            -14.750            -3.750            -3.000            -2.630
------------------------------------------------------------------------------
MacKinnon approximate p-value for Z(t) = 0.0000

. pperron roe if companyid ==10 ,lag(0)

Phillips-Perron test for unit root                 Number of obs   =         7
                                                   Newey-West lags =         0

                               ---------- Interpolated Dickey-Fuller ---------
                  Test         1% Critical       5% Critical      10% Critical
               Statistic           Value             Value             Value
------------------------------------------------------------------------------
 Z(rho)           -7.822           -17.200           -12.500           -10.200
 Z(t)            -14.750            -3.750            -3.000            -2.630
------------------------------------------------------------------------------
MacKinnon approximate p-value for Z(t) = 0.0000

[Daisy Nguyen]

Hi all, I used xtfisher for my unbalanced panel after tsset. However, stata gives the message that "too many values", how can I solve this problem? Many thanks in advanced!!

[Isabel Hermann]

Hi there,

Like David, I am trying to run a unit root test on my unbalanced panel data in order to find the degree of autocorrelation among my y variable (closing prices various indices). Since my data in 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?

Any help would be much appreciated.

Best,
Isabel