Date of Award

January 2013

Document Type

Thesis

Degree Name

Master of Science (MS)

Department

Economics & Finance

First Advisor

David T. Flynn

Abstract

Financial time series are often characterized by nonlinearity and volatility bunching. Standard regression analysis models cannot capture changing volatilities, potentially leading to erroneous results. The need to more completely model the characteristic volatilities inherent to financial time series eventually led to the creation of the GARCH model. Typical GARCH parameters are (1,1) incorporating a 1-period lag of the regression residual as well as a 1-period lag of the regression volatility. The primary question investigated in this paper is whether the typical GARCH(1,1) parameters are in fact optimal over all time periods and attempts to improve on the typical parameters by minimizing a modified AIC value using a genetic algorithm.

Download

Share

COinS