topic Re: TSA STATIONARITYTEST gives null pvalue in SAS Forecasting and Econometrics

TSA STATIONARITYTEST gives null pvalue

Edoedoedo — Thu, 17 Jun 2021 14:59:18 GMT

Hi, consider this code:

proc tsmodel data=MYCAS.ZZZ outscalar=MYCAS.AAA;
    id DATE interval=day;
    var VALORE;
    outscalars pvalue rc;
    require tsa;
    submit;
        declare object tsa(tsa);
        rc = tsa.stationaritytest(VALORE,,,,'SSM',pvalue);
        rc = rc;
        pvalue = pvalue;
    endsubmit;
run;

I want to calculate the stationarity test pvalue for the time series (attached as csv):

However in the results I see:

so that the pvalue is missing and the return code is 0, which according to the documentation (link) means Time series is stationary with the default significance level of 0.05 which clearly is false here.

So what's wrong? Can you help me to figure out what is happening? Moreover, what kind of test is stationaritytest performing? Is unclear from the documentation.

Note: I must use only procedures cas-enabled.

Thanks

Regards

Re: TSA STATIONARITYTEST gives null pvalue

imvash — Thu, 17 Jun 2021 17:37:16 GMT

Your data seems to be weekly, but you put interval=day. That might fix your problem.

Re: TSA STATIONARITYTEST gives null pvalue

imvash — Thu, 17 Jun 2021 17:52:13 GMT

Also, the augmented Dickey-Fuller test is used in STATIONARITYTEST method.

Re: TSA STATIONARITYTEST gives null pvalue

Edoedoedo — Thu, 17 Jun 2021 18:04:53 GMT

Thank you, I completely forgot to check the periodicity, That time series is part of a large set of time series on which perform the stationarity test and I didn't noticed that most of them were daily but some were not.
That explains the calculation failure, thank you very much!