What question did this study set out to answer?

The aim is to examine biases in ML estimates of the GARCH(1,1) model and the required sample size for reliable estimates.

Small sample properties of GARCH estimates and persistence

View Full Paper

Ask AI

Bookmark

Authors

Research investigates bias and persistence in GARCH(1,1) estimates, suggesting sample size improvements.

The aim is to examine biases in ML estimates of the GARCH(1,1) model and the required sample size for reliable estimates.
Analyzed the performance of ML estimates in small samples for GARCH(1,1) models.
Proposed a measure for GARCH conditional volatility's effectiveness in explaining squared returns.
Recommended at least 250 observations for ARCH(1) and 500 for GARCH(1,1) models.
ML estimates in small samples show significant negative biases.
High persistence in large sample GARCH(1,1) models results in lower autocorrelations than those suggested by small sample estimates.
The proposed measure reveals that conditional volatility explains squared returns poorly when the ARCH parameter is very small.

Hwang et al. (2006) studied this question.

View Full Paper

Ask AI

Bookmark