1
$\begingroup$

I have a monthly time series and I'm trying to determine if such set of data is stationary or not; the dataset is about composed by 160 record.

Specifically, I'm running 2 test found in literature:

  1. KPSS: if $H_0$ has been rejected then one cannot assume the time series is stationary;
  2. Phillips-Perron test: if $H_0$ has been rejected then one cannot assume that the time series has a unit-root (then it is stationary);

I preferred to implement the Phillips-Perron test in place of the most common Augmented Dickey-Fuller test since the Phillips-Perron test adjusts for the heteroschedasticity and serial correlation.

Here below, one can find the output of such analysis.

The KPSS test returns not significant p-values both for single-mean, implying that you cannot infer that the time series is not stationary; likewise, the Phillips-Perron test returns significant p-values for the single-mean and trend component, but not for the zero-mean case.

How should I consider or interpret such result?

I wonder if one can evaluate the importance and the strength of such unit-root; for instance, in the question the user @ferdi deals with the variance ratio test to argue the framework to evaluate the importance of a unit root in a time series.

Could you suggest some reference about?

I'm currently running the analysis in SAS, but any programming language would be nice.

Thank all in advance!!

$\endgroup$
1
+100
$\begingroup$

Phillips-Perron test: if $H_0$ has been rejected then one cannot assume that the time series has a unit-root (then it is stationary);

This is not entirely correct. Lack of a unit root does not imply stationarity. It only implies lack of a very special kind of nonstationarity, namely, the unit-root kind. Other kinds of nonstationarity are, for example, level shifts, shifts of variance, and other. The PP test might not be sensitive to them.

How should I consider or interpret such result?

On the face value, and provided the test results are correct, you have a time series that does not have a unit root but is not stationary either. Perhaps it can be characterized by one of the kinds of nonstationarity mentioned above, or yet another kind.

I wonder if one can evaluate the importance and the strength of such unit-root; for instance, in the question the user @ferdi deals with the variance ratio test to argue the framework to evaluate the importance of a unit root in a time series.

The answer by Ferdi you quote seems to use some nonstandard/nonstatistical terminology, or at least one that is not widespread. I am not exactly sure what is meant there, but it reminds me of something related I heard in a lecture by prof. Ruey S. Tsay. He noted that a time series $x_t$ could be a sum of a stationary process $y_t$ and a unit-root process $z_t$: $$ x_t:=y_t+z_t. $$ If the variance of the stationary process $y_t$ is large (say, $100$), while the variance of the error term in the unit-root process $z_t$ is small (say, $0.01$), then $x_t$ might behave essentially as a stationary process in a finite sample. It would be very hard to detect the presence of a unit root unless the sample size is very large. Surely, asymptotically $y_t$ would dominate $z_t$, and $x_t$ would show its unit root. But for a fixed period of time that we care about, $y_t$ dominates. Hence, in practice we might treat $x_t$ as stationary. (And in the long run we are all dead.) Perhaps this is similar to what Ferdi meant in his answer.

$\endgroup$
2
  • $\begingroup$ Thanks for the answer Richard! Clear and very useful! $\endgroup$ – Quantopik May 18 '20 at 9:04
  • $\begingroup$ @Quantopik, I am very glad you found it helpful! Thanks for the bounty! $\endgroup$ – Richard Hardy May 18 '20 at 9:37

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.