Poisson Jumps and Long Memory Volatility Process in High Frequency European Exchange Rates

Abstract

This paper analyzes the intriguing features of 30-minute

European foreign exchange rates during the year 1996; jumps

and long memory volatility process. The FIGARCH model with

the Poisson distribution is applied in order to consider both the

jumps in the conditional mean process and the long memory

property in the conditional variance process of the high

frequency foreign exchange returns series. The general results

show that the Poisson distribution accounts for the jumps in

the high frequency foreign exchange rates returns quite well and

that the jumps seem to spuriously induce higher long memory

property in the high frequency foreign exchange returns. Hence,

the FIGARCH model with the Poisson distribution appears to be

quite appropriate for the specification of the high frequency

returns.