Shanghai Stock Exchange Composite Index Prediction Using Macro Variables Based on ARIMAX Model

Abstract

Stock market prediction has always been a prevailing topic among investors and researchers. Among numerous market index, the Shanghai Stock Exchange Composite Index (SSE Index) is recognized as one of the most indicative stock indexes in China's A-share market. As it is discovered that the fluctuations of SSE index and exchange rate (USD/CNY) displays highly similar pattern since the subprime mortgage crisis, this study aims to use macro variables (including exchange rate) to predict the SSE index based on ARIMAX model. The data are collected since 2006 and based on which ARIMA (14,1,4) model is generated. Distinct macro variables are added in this ARIMA model respectively and it concludes that the incorporation of exchange rate with certain lags can significantly increase the fitting degree. Cross validations on different lengths of validation sets are implemented, which shows that the model can make accurate forecast in relatively short term. The result manifests increasing accuracy when incorporating certain explanatory regressive factors, which might provide valuable and enlightening information in short term for researchers or investors.