An examination of foreign exchange risk under fixed and floating rate regimes

Abstract

This paper examines variability risk in foreign exchange markets during fixed and floating rate regimes. It is demonstrated that variance is an inadequate measure of variability, both theoretically and empirically. Although the normal model is commonly employed in foreign exchange research, the paper shows that spot and forward rate changes can be better described as if they were generated from nonnormal members of the Pareto-Levy class of probability distributions. Further, the paper discusses what statistical measures of variability are most appropriate. Several measures are developed to assess the characteristics of spot and forward rate changes during the period 1962–1975. It is found that floating rates are generally many times riskier, in the variability sense, than fixed rates.