Analyzing Cambodia Securities Exchange Index Returns using the Markov-Switching Autoregressive Model

Abstract

The examination of the weekly return behavior of the Cambodia Securities Exchange (CSX) index, spanning from 2012 to 2024, was categorized into two distinct states or regimes using the Markov-Switching Autoregressive model. The research findings indicated that the MS(2)-AR(1) model, which includes two states or regimes and a first-order autoregressive component, was the most suitable model. The empirical results showed that both the first-order lag of the dependent variable and the intercept term had a significant positive effect on the return of the CSX index at a 1% significance level, applicable to both Regime 1 and Regime 2 models. In contrast, the first-order autoregressive variable in the Regime 1 model demonstrated a significant negative effect on the return of the CSX index at the same 1% significance level, a relationship not observed in the Regime 2 model. The empirical results indicated a 38.35% likelihood of the CSX index transitioning from Regime 2 to Regime 1, while the probability of exiting Regime 1 was notably lower at 17.39%, as shown by the probability transition matrix. Additionally, the volatility of the CSX index returns in Regime 2 was found to be greater than that observed in Regime 1.