Predicting daily highs and lows of exchange rates : A cointegration analysis

Research output: Journal Publications and ReviewsRGC 21 - Publication in refereed journalpeer-review

17 Scopus Citations
View graph of relations

Author(s)

Detail(s)

Original languageEnglish
Pages (from-to)1191-1204
Journal / PublicationJournal of Applied Statistics
Volume36
Issue number11
Publication statusPublished - Nov 2009

Abstract

This article presents empirical evidence that links the daily highs and lows of exchange rates of the US dollar against two other major currencies over a 15 year period. We find that the log high and log low of an exchange rate are cointegrated, and the error correction term is well-approximated by the range, which is defined as the difference between the log high and log low. We further assess the empirical relevance of jointly analyzing the highs, lows and the ranges by comparing the range forecasts generated from the cointegration framework with those from random walk and autoregressive integrated moving average (ARIMA) specifications. The ability of range forecasts as predictors of implied volatility for a European style currency option is also evaluated. Our results show that aside from a limited set of exceptions, the cointegration framework generally outperforms the random walk and ARIMA models in an out-of-sample forecast contest. © 2009 Taylor & Francis.

Research Area(s)

  • Daily high, Daily low, Direction of change, Implied volatility, VECM