Analysis of Volatility of the Return of Composite Stock Price Index Using ARCH/GARCH Model, January 2015 - September 2024

Abstrak

 The objectives of this paper is to identify and measure the volatility of the return of composite stock price index in the time period January, 2015 – September, 2024 using model ARCH/GARCH. It has been identified that the best model in explaining the volatility of the return in the time period was GARCH (1,1). The interesting findings, among others,  firstly, the average return of the index is 0.4548 or 45.48 percent monthly in the time period. Secondly, the volatility of return of index at the certain month affected by squared residual of previous months of 27.63 percent. Thirdly, 53.58 percent of the volatility of the return of the index at the certain month affected by the volatility of the return of the index at the previous month larger than the effect of squared residuals. Finally, fourthly, the fluctuation of conditional variance in pandemic period higher than nonpandemic period. The volatility of the return of index are useful  for either individual investor or institutional investor in the decision making at the stock exchange especially in their asset allocation. If the volatility increases, investor probably to choose the safely assets such as obligation or defensive stock. Instead, as the volatility is low, the investors tend to taking risk by choosing the aggressive stocks. Remains a consideration in using this research findings. Because of model using in this research is univariate GARCH model of course it will be also consideration in using multivariate GARCH. Both internally including inflation, monetary policy, politics, or domestic economic event and externally including exchange rate, commodity price (world oil price, gold), global interest rate or international market sentiment should be considered into model to obtain comprehensive return volatility analysis. Among others, using model Generalized Autoregressive Conditional Heteroskedasticity-in-Mean (GARCH-M) or Exponential Generalized Autoregressive Conditional Heteroskedasticity (EGARCH) can be used to observates the factors already mentioned.