Long memory in real exchange rate changes.

Item

Title
Long memory in real exchange rate changes.
Identifier
AAI9820519
identifier
9820519
Creator
Christodoulou, Christodoulos Nicou.
Contributor
Adviser: Salih Neftci
Date
1998
Language
English
Publisher
City University of New York.
Subject
Economics, General | Economics, Finance
Abstract
This study examines the time-series dynamics of real exchange rate changes in the fractionally integrated autoregressive moving average (ARFIMA) framework, which allows for long-memory in the data and is a generalization of the standard ARIMA model. Specifically, the differencing parameter of an ARFIMA model is not restricted to the integer domain and can assume real values. This generalization makes an ARFIMA model a parsimonious and flexible model to study long-memory, which induces persistence, and short-run dynamics simultaneously. Fractional integration is a more general way to describe dependence between observations some distance apart than the unit-root specification and provides an alternative perspective to examine the unit-root hypothesis. This empirical study provides supportive evidence of long-memory in real exchange rate changes, meaning that purchasing power parity may hold as a long-run relationship, based on a statistical procedure that is asymptotically robust to short-run memory and has a well defined distribution. In addition, when ARFIMA models are fitted to the data the estimates of the AR and MA parameters imply that there is long-memory in all real exchange rate data. To better understand the long-memory characteristics of real exchange rates, estimates of the impulse response parameters are obtained based on the fitted models. Impulse-responses provide more convincing evidence of long-range dependence in the real exchange rate data and also give information regarding the pattern and speed with which shocks to purchasing power parity are propagated.
Type
dissertation
Source
PQT Legacy CUNY.xlsx
degree
Ph.D.
Item sets
CUNY Legacy ETDs