Good day scholars,
In order to properly determine the true order of integration of X, Y and Z series, I subjected each series to Augmented Dickey-Fuller (ADF), Phillips-Perron (PP) and Kwiatkowski-Phillips-Schmidt-Shin (KPSS) tests respectively. Results reported by the ADF showed that X and Y series are stationary at levels {i.e they are I(0)s} wile Z is not. However, PP reported that only X series is I(0) whereas Y and Z are not. Finally, KPSS reported that all the X, Y and Z series are stationary series. Though ADF and PP tests are testing the hypothesis: X has a unit root etc. while KPSS is testing the hypothesis: X is stationary etc.
Can I rely on the results reported by KPSS tests and conclude that time series variables X, Y and Z are truly level stationary series?
Very many thanks for always being there!!!
Dear Saheed,
So, In your case you need to check if your series is stationary and integrated. You can use the correlogram test and unit root test here. The unit root test is the most popular stationarity test. For example, by calculating the first difference, your series will be integrated by the first-order I(1). In addition to the ADF (Augmented Dickey-Fuller test) there are other unit root tests, such as the Phillips-Perron test, the Kwiatkowski-Phillips-Schmidt-Shin test, and so, as you mentioned. More explanation of the advantages, disadvantages and criticism of these unit root tests, for example, can be found in Gujarati (2003) or you can consider Maddala as a good one source. Good luck! Goran.
See these links:
https://www.statisticshowto.com/order-of-integration/
https://www.statalist.org/forums/forum/general-stata-discussion/general/1291442-determining-the-order-of-integration
https://corporatefinanceinstitute.com/resources/knowledge/other/cointegration/
Check https://www.statisticshowto.com/order-of-integration/
Also check http://www.lexjansen.com/sesug/1997/SESUG97095.pdf
Hello Akanni,
Kindly look at this sir.
https://www.machinelearningplus.com/time-series/kpss-test-for-stationarity/
- Badges
- Science topic
- Similar topics
- Statistics
- Time Series