Information-theoretic approach to quantifying currency risk

Abstract

Purpose – This paper aims to present a framework enriching currency risk analyses based on information theory. Design/methodology/approach – Information-theoretic measures of predictability (entropy rate) and co-dependence (mutual information) are used to enhance existing methods of analysing and measuring currency risk. Findings – The currency exchange rates have varying degrees of predictability, which should be accounted for in currency risk analyses. In case of baskets of currencies, a network approach rooted in portfolio theory may be useful. Research limitations/implications – The currency exchange rate time series must be discretised for the information-theoretic analysis (although the results are robust). An agent-based simulation may be a necessary further study to show what the impact of accounting for predictability in managing currency risk is. Practical implications – Practical analyses measuring currency risk should take predictability of currency rate changes into account wherever the currency exposure is actively managed. Originality/value – The paper introduces predictability into measuring currency risk, which has previously been ignored, despite the nature of the risk being inherently tied to uncertainty of the currency rate changes. The paper also introduces a portfolio theory-based approach to quantifying currency risk, which accounts for non-linear co-dependence in the currency markets.