[R-SIG-Finance] half life are different when applying diff frequencies to mean reverting time series

[Thomas X]

Hi, All

I try to calculate the lengths of the half life of a mean reverting time
series by applying different frequencies (1 second, 1 minutes, 1 hour, etc)
to the same time series in its format (using to.period). I use the adfTest
to calculate the p-value. For different frequencies, the p-values of the
time series are all below 0.01.  However, the values of half-life confuse
me. For example.

In the frequency of 10 seconds, half-life is 0.25, which is equivalent to
2.5 seconds.
In the frequency of 1 hour, half-life is 0.15, which is equivalent to 9
minutes.

The values of the half life are different.

Any clue? Thanks.

-Thomas

--
View this message in context: http://r.789695.n4.nabble.com/half-life-are-different-when-applying-diff-frequencies-to-mean-reverting-time-series-tp4275915p4275915.html
Sent from the Rmetrics mailing list archive at Nabble.com.