When estimating the GARCH model, an intermediate step is to predict tomorrow's return.

Question

BlackTom AI · Accepted Answer

In GARCH modeling, the primary focus is on modeling the conditional variance process, not directly predicting tomorrow's return as a primary objective.
Option 1: 'Make sense' — This would imply that forecasting tomorrow's return is a central intermediate step in estimating a GARCH model. While some models include a mean equation that could estimate an expected return (mu), the core purpose of GARCH is to capture how volatility evolves over time, i.e., the conditional variance h_{t+1|t}. Therefore, predicting tomorrow's return is not the main intermediate step and makes sense only in a limited sense if you separately specify a mean model; it is not the standard core procedure.
Option 2: 'Make no sense' — This aligns with the standard view that GARCH estimation emphasizes the conditional variance dynamics, not the point forecast of tomorrow's return. The usual workflow involves estimating parameters of the volatility process and then producing a one-step-ahead variance forecast, rather than treating tomorrow's return as the central intermediate target. If a mean model is present, its forecast could be mu, but that is ancillary to the volatility dynamics, not the essence of the GARCH estimation step.
Overall, the common interpretation is that predicting tomorrow's return is not a fundamental intermediate step in GARCH estimation, whereas focusing on the conditional variance is. The contrast highlights why the second option is more appropriate in typical GARCH applications.

When estimating the GARCH model, an intermediate step is to predict tomorrow's return.单项选择题

类似问题

On Tuesday, you calculated the volatility of Wednesday as 5% using the GARCH model, which information will make the Thursday volatility become even higher?