In-Sample or Out-of-Sample Tests of Predictability: Which One Should We Use?


Speaker:Lutz Kilian, University of Michigan
Date: Friday 10 May 2002
Time: 16:15
Location: Room 106 Streatam Court

Further details

(with Atsushi Inoue)

It has become a standard empirical Þnding that significant in-sample evidence of predictability does not guarantee significant out-of-sample predictability. This is often interpreted as an indication that in-sample evidence is likely to be spurious and should be discounted. In this paper we question this conventional wisdom. We show that in-sample and out-of-sample tests of predictability are equally reliable under the null hypothesis of no predictability, provided that appropriate critical values are used. We then compare the local asymptotic power of these tests. We show that the ranking of in-sample and out-of-sample tests is ambiguous, but that in many cases in-sample tests have higher power. Our results provide an alternative explanation of the comparatively weak out-of-sample evidence of predictability. We conclude that results of in-sample tests of predictability will typically be more credible than results of out-of-sample tests.