Abstract
In the Engel and West (2005) exchange rate literature, some hold that commodity prices should be predicted by exchange rates and not vice-versa. This hypothesis runs counter to a vast literature that regards commodities
as financial assets, whose price changes are normally hard to predict. Prior empirical results have not lead to a consensus. Some commonly used indices have features that create biases in tests like ours, we show. Furthermore, we suggest a new index weighting scheme that should improve one’s chances of unearthing predictive power, if any. We also try higher frequencies, principal components indices, model averaging and trading rules — all to no avail: we find no asymmetric predictive ability either way, and the link is essentially contemporaneous.
as financial assets, whose price changes are normally hard to predict. Prior empirical results have not lead to a consensus. Some commonly used indices have features that create biases in tests like ours, we show. Furthermore, we suggest a new index weighting scheme that should improve one’s chances of unearthing predictive power, if any. We also try higher frequencies, principal components indices, model averaging and trading rules — all to no avail: we find no asymmetric predictive ability either way, and the link is essentially contemporaneous.
Originalsprog | Engelsk |
---|---|
Tidsskrift | Journal of Banking and Finance |
Sider (fra-til) | 1 |
Antal sider | 49 |
ISSN | 0378-4266 |
Status | Afsendt - 2019 |