Simulation study for change-point of AR-GARCH models and rank tests based trading strategies

Abstract

It is very important to estimate the location of the change-point in statistical models. This thesis first gives simulation studies for the performance of the estimating coefficients and change-point in the AR-GARCH model. As an application, the structural change AR-GARCH models are used to analyze the Hang Seng Index. This thesis also proposes a new indicator, called Moving Signed-Rank Statistic, to detect the buy and sell signals in price series. We use the bootstrap approach to estimate the quantiles of our test statistic and then use them as trading signals. The performance of the proposed trading strategies is given.

Date of Award	2012
Original language	English
Awarding Institution	The Hong Kong University of Science and Technology