Daily Jakarta Composite Index Data (2004–2024)

The dataset employed in this study consisted of daily closing prices of the Jakarta Composite Index, which served as a comprehensive proxy for the overall performance of the Indonesian equity market. The observation period spanned from 6 January 2004 to 30 December 2024, yielding a total of 5,099 trading-day observations after excluding non-trading days. This long-horizon sample covered multiple market regimes, including periods of rapid economic expansion, global financial crises, domestic policy shifts, and post-pandemic normalization phases, thereby providing a rich empirical setting for volatility analysis. Daily log returns were computed from closing prices to obtain a scale-free and time-additive measure suitable for conditional variance modeling. The resulting return series exhibited substantial variability across time, reflecting heterogeneous market conditions and episodic stress events. Preliminary inspection revealed that the dataset was free from missing observations and structural inconsistencies, allowing for direct application of econometric volatility models without interpolation or data smoothing. To support both model estimation and forecasting evaluation, the dataset was partitioned into a training sample and a testing sample following an 80:20 split. The training period covered 6 January 2004 to 12 October 2020 and was used for parameter estimation, while the testing period from 13 October 2020 to 30 December 2024 was reserved for out-of-sample forecasting and risk backtesting. This design enabled robust assessment of model performance under recent high-uncertainty conditions while preserving long-term information for capturing persistent volatility dynamics.