Assessing the relationship between the all share price index and some selected economic indices in Sri Lanka

Abstract

The purpose of this research is to create a predictive model based on changes in macroeconomic variables for the Colombo Stock Exchange's (CSE) All Share Price Index (ASPI). The macroeconomic variables considered are interest rate, Colombo Consumer Price Index (CCPI), and Gross Domestic Product (GDP) growth. The dataset, which includes quarterly data from 2007 to 2023, was obtained from the CSE, Central Bank of Sri Lanka (CBSL), and Census of Statistics. From the descriptive analysis, it was found that the ASPI values vary from 1503 to 12226 within the time range. The cross-correlation coefficient was used to examine the correlation between ASPI and independent variables at different time lags. Interest rates and GDP growth have a marginally negative relationship with ASPI, while the relationship between CCPI and ASPI is negligible. A univariate time series analysis was conducted and the model was fitted using ARIMA to forecast the ASPI values in the future. The analysis involves testing for stationarity, checking normality, and residual diagnostics among independent variables. The stationarity of variables was checked using the Augmented Dickey-Fuller test. Further, the cointegration among variables was checked using the Johansen co-integration test as the data was not stationary. After identifying there is no cointegration, the VAR (Vector Auto-Regressive) model was fitted to find the short-term dynamics. Significance of the parameters and model adequacy measures were conducted for the VAR model. The autocorrelation among errors was checked using the Ljung-Box test while the white test was used to check the homoscedasticity of the errors. The normality of the errors was checked using Shapiro-Wilk tests and can be visualized from Q-Q plots.