Macroeconomic Shocks and Jumps in the Long Memory Models of Daily KRW-USD and KRW-JPY Foreign Exchange Rates

Abstract

This paper focuses on providing proper models for the daily

Korean exchange rate dynamics which is subject to macroeconomic

shocks. By investigating the daily KRW-USD and KRW-JPY exchange

returns, this paper presents that the usual assumption of normal

distribution is not appropriate in representing the daily Korean

exchange returns due to the jumps which are related to the

macroeconomic shocks of Korea, Japan and the U.S. Thus, this

paper relies on the normal mixture distribution that allows for the

jumps in the process of the daily Korean exchange returns. The

normal mixture model with the Bernoulli distribution is found to

perform quite well and to be important for the estimation of the

long memory persistence in the daily Korean exchange return

volatility. In particular, using the time-varying jump probability

associated with the macroeconomic shocks of Korea, Japan and

the U.S., this paper finds that the macroeconomic shocks induce

jumps in the process of the daily exchange returns and appear to

increase the long memory persistence in the daily Korean exchange

return volatility.