MARKOV SWITCHING MODELS FOR THE EXCHANGE RATE AND THE STOCK RETURN BASED ON A CHARTIST-FUNDAMENTALIST EXPECTATION APPROACH

Abstract

This dissertation consists of two essays to develop empirical models to describe and to forecast daily exchange rate and stock return dynamics by using Markov switching (MS) models based on a chartist-fundamentalist expectation approach. The first essay is “a Markov switching model of the exchange rate based on a chartist-fundamentalist approach.” This study develops a Markov switching model by allowing each group of foreign exchange traders to use both chartist and fundamentalist expectations in the two unobservable states. This study makes use of the Hodrick-Prescott filter to separate the market participants into two independent groups: short-term speculators and longer-term investors. These two types of participants have different expectations and impacts on daily exchange rate movements. In addition, we assume that the two unobservable states affecting the expectations of these two groups are not the same. This study examines empirical evidences of the five most traded currency pairs. The second essay is “an application of the chartist-fundamentalist approach to a model of speculative behavior for Asian stock markets.” This study examines empirical evidences of the daily data of four stock indices in four Asian countries for comparison purpose, i.e., Hong Kong, India, Korea, and Thailand. The market participants are divided into two types: chartists and fundamentalists. The empirical results exhibit that the proposed models can reasonably and significantly explain the dynamics of daily exchange rates and stock returns.