Integrating High-Frequency Data In Volatility Prediction: A DCC-GARCH-MIDAS Approach

Traditional econometric models are restricted in their capacity to examine same-frequency data, resulting in the loss of valuable information from high-frequency data. To address this problem, We propose the DCC-GARCH-MIDAS model by combining dynamic conditional correlation modeling with information from high-frequency data. For our empirical study, we utilised historical data from 2000 to 2019 as in-sample data and trained a model for predicting volatility. We applied the trained model to forecast data from 2020 to 2022, calculating the discrepancy between predicted volatility and actual observations, and comparing differences between the predicted and actual values. The research findings not only enhance comprehension of the correlation between macroeconomics and financial market instability but also propose a distinct strategy for resolving the problem of incongruent data frequencies.