[R] KPSS test

[Pfaff, Bernhard Dr.]

Hello Sachin,

a sequential testing procedure is described in the useR! book:

  @Book{,
    title = {Analysis of Integrated and Cointegrated Time Series with R},
    author = {B. Pfaff},
    publisher = {Springer},
    edition = {First},
    address = {New York},
    year = {2006},
    note = {ISBN 0-387-27960-1},
  }

Best,
Bernhard


Dr. Bernhard Pfaff
Global Structured Products Group
(Europe)

Invesco Asset Management Deutschland GmbH
Bleichstrasse 60-62
D-60313 Frankfurt am Main

Tel: +49(0)69 29807 230
Fax: +49(0)69 29807 178
Email: bernhard_pfaff at fra.invesco.com  

>-----Ursprüngliche Nachricht-----
>Von: r-help-bounces at stat.math.ethz.ch 
>[mailto:r-help-bounces at stat.math.ethz.ch] Im Auftrag von Sachin J
>Gesendet: Donnerstag, 6. Juli 2006 21:49
>An: markleeds at verizon.net
>Cc: r-help at stat.math.ethz.ch
>Betreff: Re: [R] KPSS test
>
>Hi Mark,
>   
>  Thanx for the help. I will verify my results with PP and DF 
>test. Also as suggested I will take a look at the references 
>pointed out. One small doubt: How do I decide what terms ( 
>trend, constant, seasonality ) to include while using these 
>stationarity tests. Any references would be of great help. 
>   
>  Thanx,
>  Sachin
>   
>  
>
>markleeds at verizon.net wrote:
>  >From: 
>>Date: Thu Jul 06 14:17:25 CDT 2006
>>To: Sachin J 
>>Subject: Re: [R] KPSS test
>
>sachin : i think your interpretations are right given the data
>but kpss is quite a different test than the usual tests
>because it assumes that the null is stationarity while dickey 
>fuller ( DF ) and phillips perron ( PP ) ) assume that the 
>null is a unit root. therefore, you should check whetheer
>the conclusions you get from kpss are consistent with what you 
>would get from DF or PP. the results often are not consistent.
>
>also, DF depends on what terms ( trend, constant ) 
>you used in your estimation of the model. i'm not sure if kpss 
>does also. people generally report Dickey fuller results but they
>are a little biased towards acepting unit root ( lower
>power ) so maybe that's why
>you are using KPSS ? Eric Zivot has a nice explanation
>of a lot of the of the stationarity tests in his S+Finmetrics 
>book.
>
>testing for cyclical variation is pretty complex because
>that's basically the same as testing for seasonality.
>check ord's or ender's book for relatively simple ways of doing that.
>
>
>
>
>
>
>
>
>
>
>
>
>>
>>>From: Sachin J 
>>>Date: Thu Jul 06 14:17:25 CDT 2006
>>>To: R-help at stat.math.ethz.ch
>>>Subject: [R] KPSS test
>>
>>>Hi,
>>> 
>>> Am I interpreting the results properly? Are my conclusions correct?
>>> 
>>> > KPSS.test(df)
>>> ---- ----
>>> KPSS test
>>> ---- ----
>>> Null hypotheses: Level stationarity and stationarity around 
>a linear trend.
>>> Alternative hypothesis: Unit root.
>>>----
>>> Statistic for the null hypothesis of 
>>> level stationarity: 1.089 
>>> Critical values:
>>> 0.10 0.05 0.025 0.01
>>> 0.347 0.463 0.574 0.739
>>>----
>>> Statistic for the null hypothesis of 
>>> trend stationarity: 0.13 
>>> Critical values:
>>> 0.10 0.05 0.025 0.01
>>> 0.119 0.146 0.176 0.216
>>>----
>>> Lag truncation parameter: 1 
>>> 
>>>CONCLUSION: Reject Ho at 0.05 sig level - Level Stationary
>>> Fail to reject Ho at 0.05 sig level - Trend Stationary 
>>> 
>>>> kpss.test(df,null = c("Trend"))
>>> KPSS Test for Trend Stationarity
>>> data: tsdata[, 6] 
>>>KPSS Trend = 0.1298, Truncation lag parameter = 1, p-value = 0.07999
>>> 
>>> CONCLUSION: Fail to reject Ho - Trend Stationary as p-value 
>< sig. level (0.05)
>>> 
>>>> kpss.test(df,null = c("Level"))
>>> KPSS Test for Level Stationarity
>>> data: tsdata[, 6] 
>>>KPSS Level = 1.0891, Truncation lag parameter = 1, p-value = 0.01
>>> Warning message:
>>>p-value smaller than printed p-value in: kpss.test(tsdata[, 
>6], null = c("Level")) 
>>> 
>>> CONCLUSION: Reject Ho - Level Stationary as p-value > sig. 
>level (0.05)
>>> 
>>> Following is my data set
>>> 
>>> structure(c(11.08, 7.08, 7.08, 6.08, 6.08, 6.08, 23.08, 32.08, 
>>>8.08, 11.08, 6.08, 13.08, 13.83, 16.83, 19.83, 8.83, 20.83, 17.83, 
>>>9.83, 20.83, 10.83, 12.83, 15.83, 11.83), .Tsp = c(2004, 
>2005.91666666667, 
>>>12), class = "ts")
>>>
>>> Also how do I test this time series for cyclical varitions? 
>>> 
>>> Thanks in advance.
>>> 
>>> Sachin
>>>
>>> 
>>>---------------------------------
>>>
>>> [[alternative HTML version deleted]]
>>>
>>>______________________________________________
>>>R-help at stat.math.ethz.ch mailing list
>>>https://stat.ethz.ch/mailman/listinfo/r-help
>>>PLEASE do read the posting guide! 
>http://www.R-project.org/posting-guide.html
>
>
>
> 		
>---------------------------------
>
>
>	[[alternative HTML version deleted]]
>
>______________________________________________
>R-help at stat.math.ethz.ch mailing list
>https://stat.ethz.ch/mailman/listinfo/r-help
>PLEASE do read the posting guide! 
>http://www.R-project.org/posting-guide.html
>
*****************************************************************
Confidentiality Note: The information contained in this mess...{{dropped}}