For econometric discussions not necessarily related to EViews.

Moderators: EViews Gareth, EViews Moderator

willforbes1
Posts: 3
Joined: Fri Apr 07, 2017 3:35 am

Hi, i am new to eviews and i am conducting a few statistical tests for my dissertation, i am unsure whether i should reject or accept the null hypothesis of stationary according the critical values. Any advice regarding this would be appreciated. Thanks in advance for your help.

KPSS test results:
t-statistic= 2.4160
1% value= 0.7390
5% value= 0.4630
10% value= 0.3470

prob.= 0.9716
t-statistic= 0.1863
1% value= -3.4426
5% value= -2.8668
10% value= -2.5697

EViews Mirza
Posts: 68
Joined: Sat Apr 22, 2017 8:23 pm

Given your prob. value of the ADF test, you will most certainly NOT reject the null of a unit root for any reasonable significance levels. On the other hand, the null of the KPSS test is stationarity, but your statistic 2.416 is in the rejection region, so you will reject the null of stationarity and accept that your series has a unit root.

willforbes1
Posts: 3
Joined: Fri Apr 07, 2017 3:35 am

Hi, thanks for the quick response, this has really helped me. I also conducted a variance ratio test please could you take a look at the results from this test also if you dont mind? Should i accept the presence of a random walk or is there a random walk in the data? Thanks again in advance.

value= 0.621041
df= 523
probability= 0.9531

individual tests: stats for each period:
z statistics for period 2=-0.2306 prob=0.8176 variance ratio=0.9815
z statistics for period 4=0.3253 prob=0.7450 variance ratio=1.0461
z statistics for period 8=0.6210 prob=0.5346 variance ratio=1.1346
z statistics for period16=0.4766 prob=0.6337 variance ratio=1.1424

pitsi
Posts: 19
Joined: Wed May 28, 2014 2:43 pm

Dear all,

Could you please help me to understand if the variable has a unit root or not? The fact that the ADF t-stat is smaller than the critical value at 5% indicates rejection of the null but what about the critical value at 1% level of significance? In this case it indicates that null cannot be rejected and first differences should be tested, right? So which one should I account? Although I ve searched the forum for answers I couldnt find something relevant, so apologies if a post has skipped.

Null Hypothesis: L_IT_EV has a unit root
Exogenous: Constant, Linear Trend
Lag Length: 1 (Automatic - based on SIC, maxlag=15)

t-Statistic Prob.
Augmented Dickey-Fuller test statistic -3.655053 0.0272
Test critical values: 1% level -3.992933
5% level -3.426809
10% level -3.136666

*MacKinnon (1996) one-sided p-values.