Purchasing power parity and long memory

Abstract

This article examines the validity of purchasing power parity by estimating long memory parameters with recently suggested exact local Whittle estimators of Shimotsu and Phillips (2005). Little evidence is found for stationarity in the real exchange rates spanning more than 100 years from 16 advanced countries. However, most of the estimates of long memory parameters are less than 1, so that they indicate mean-reversion to parity. Hence, purchasing power parity holds for most real exchange rates studied here.

Acknowledgement

This work was supported by research programme 2006 of Kookmin University in Korea. I would like to thank an anonymous referee for helpful comments and suggestions.

Notes

¹ See e.g. Abuaf and Jorion (1990), Lothian and Taylor (1996), Taylor (2002) and Lopez et al. (2005a).

² After allowing for broken trends, Culver and Papell (1995) find evidence for stationarity in the same data set.

³ All calculations in this article are carried out with GAUSS.

⁴ Recall that in Table 1, Canada, Denmark, Japan, Portugal and Switzerland are found to be I(1) by various unit root tests.