Skip to content

Do gold market returns have long memory? 1993

by on October 30, 2013

This paper offers a test for long memory in gold returns that is robust to short term dependence and heteroscedasticity, common problems in precious metals research.

Over the full sample gold returns are found to exhibit long memory but the authors show that this is mostly due to a small number of observations relating to Middle Eastern political tension and the activities of the Hunt Brothers in 1979. Once these are removed the long memory feature disappears. But the findings do highlight the sensitivity of the gold price due to it’s role as a safe haven, which can have the effect of reducing its informational efficiency over short periods.

Data: Mid-week London PM spot price, July 1973 – December 1987

Methodology: Rescaled Test for Long Memory

Citation: Cheung, Yin‐Wong, and Kon S. Lai. “Do gold market returns have long memory?.” Financial Review 28.2 (1993): 181-202.

Abstract: This study examines the long memory behavior in gold returns during the post-Bretton Woods period using a new rescaled range technique. Unlike the conventional rescaled range analysis, the new rescaled range analysis is robust to short-term dependence and conditional heteroscedasticity found in the gold data. Statistical results suggest that the long memory behavior in gold returns is rather unstable. When only few observations corresponding to major political events in the Middle East, together with the Hunts event, in late 1979 are omitted, little evidence of long memory can be found.



From → Empirical, Gold

Leave a Comment

Leave a Reply

Please log in using one of these methods to post your comment: Logo

You are commenting using your account. Log Out / Change )

Twitter picture

You are commenting using your Twitter account. Log Out / Change )

Facebook photo

You are commenting using your Facebook account. Log Out / Change )

Google+ photo

You are commenting using your Google+ account. Log Out / Change )

Connecting to %s

%d bloggers like this: