Stock Daily Price Regime Model Detection using Markov Switching Model

Abstract

Changes in stock prices randomly occur due to market forces with reoccurrence
possibilities. This process, also known as the structural break model, is captured through
changes in the linear model parameters among periods with the Markov Switching Model
(MSwM) used for detection. Furthermore, using the smallest Akaike Information Criterion
(AIC) value on all feasible MSwM alternatives formed for a daily stock price, the complete
MSwM model with its Markov transition is determined. This method has been tested and
applied to daily stock price data in several sectors. The result showed that the number of
regime models coupled with its transition probability helped investors make investment
decisions.