<![CDATA[This study aims to model volatility and predict the stock returns of Bank Central Asia Tbk (BBCA) using the ARCH-GARCH model with a daily frequency time series approach in the Indonesian capital market during the 2020–2025 period. The data used consist of the daily closing prices of BBCA stock obtained from Yahoo Finance, comprising 1,447 observations of daily returns. The research methodology includes descriptive statistical analysis, stationarity testing (ADF test), ARIMA model identification through ACF and PACF plots, selection of the best model based on the AIC criterion, testing of classical residual assumptions (normality, autocorrelation, and heteroscedasticity), and estimation of the GARCH(1,1) model to model conditional variance. The results show that BBCA stock returns exhibit non-normal characteristics (Shapiro-Wilk p-value = 8.55×10⁻¹⁸), positive skewness (0.5787), and high kurtosis (11.64), indicating a leptokurtic (fat-tail) distribution, as well as containing ARCH effects (p-value = 1.82×10⁻⁹), confirming the presence of volatility clustering. The ARIMA(2,1,0) model was selected as the best mean equation model with an AIC value of 18032.73 and MAPE of 1.15% (categorized as excellent). Estimation of the GARCH(1,1) model produced ARCH (α) and GARCH (β) parameters of 0.1198 and 0.8163, respectively (both significant at p < 0.001), with volatility persistence (α + β) of 0.9361 (93.61%), indicating that shocks have a long-lasting impact on volatility with a half-life of approximately 10–11 trading days. Diagnostic tests proved that the GARCH(1,1) model is adequate because the standardized residuals contain neither autocorrelation nor remaining ARCH effects, and no leverage effect was found. Volatility forecasting for the next 20 days shows a gradual increase from 1.38% to 1.55%, indicating the potential for increased price fluctuations in the future. In conclusion, the GARCH(1,1) model is proven to be superior and appropriate for modeling volatility and predicting BBCA stock returns because it is able to capture the characteristics of volatility clustering, fat-tail distribution, and high volatility persistence in stock return data in the Indonesian capital market.]]>
