Long RangeMemory Modeling and Estimation for Financial Time Series

This paper deals with several aspects in time series modeling concerning estimation and tests of long memory, fractional integration, and cointegration, as well as applications to financial data. The aim of the paper is to develop new and improved estimation and testing techniques, in particular to extend existing work concerning fractional processes and also to introduce new areas of application. The formulation allows the widely used fractional autoregressive integrated moving average ARFIMA models and our asymptotic results provide a theoretical justification of the findings in simulations that the local Whittle estimator is robust to deterministic polynomial trends. Finally, we explore the existence of long memory in some financial time series and conclude using a novel approach in their exploration.

long memory, fractionally integrated autoregression models, periodogram, Whittle method