Forecasting Malaysian stock returns using GPR and EPU - an in-sample and out-of-sample analysis

Hong, Way Hong and Lai, Yong Qin and Lim, Chen Yen and Woo, Kuan Yee (2020) Forecasting Malaysian stock returns using GPR and EPU - an in-sample and out-of-sample analysis. Final Year Project, UTAR.

Preview

PDF Download (2262Kb) | Preview

Abstract

From earliest infancy, human being tend to classify the world into categories, forecast how things work, and assess those predictions. Such thinking, which is one of the human nature, is now being expanded into economics, business, management and finance contexts. That is why nowadays there are growing literature and empirical studies which examined the forecasting extreme by applying several types of predictors, since they wished to search for the most accurate forecasting method and predictors for different contexts. In finance context, especially on stock returns forecasting, prior studies focused on the assessment of the predictive power of macroeconomic determinants and financial ratios on stock returns. Thus, this paper focuses on the application of the latest risk indicators which comprises of economic policy uncertainty (EPU) and geopolitical risk (GPR) on forecasting Malaysian stock returns by using monthly data that covered from January 2005 to April 2019. This studies makes three contributions to the literature on forecasting Malaysian stock returns. First, we examine the predictive ability of EPU and GPR by undertaking in-sample tests on the FGLS model of both predictors. Through this, we find that both predictors have the ability to predict Malaysian stock returns. Second, by comparing the in-sample and out-of-sample results of both predictors, both tests provide a consistent outcome by showing that EPU and GPR are favourable in forecasting Malaysian stock returns. Third, through the comparison between the predictive ability of EPU and GPR by implementing the out-of-sample analysis, GPR is said to be performed even well in forecasting Malaysian stock return than EPU by referring to Theil U2 coefficient.

Actions (login required)