#TITLE_ALTERNATIVE#
Most volatility models are generated based on the Normality assumption of their returns. This assumption is empirically inappropriate since asset returns have high kurtosis. In other words, returns data follow heavy-tailed distribution. In this project, Student-t and Burr Type II distributions are e...
Saved in:
| Main Author: | |
|---|---|
| Format: | Final Project |
| Language: | Indonesia |
| Online Access: | https://digilib.itb.ac.id/gdl/view/11596 |
| Tags: |
Add Tag
No Tags, Be the first to tag this record!
|
| Institution: | Institut Teknologi Bandung |
| Language: | Indonesia |
| Summary: | Most volatility models are generated based on the Normality assumption of their returns. This assumption is empirically inappropriate since asset returns have high kurtosis. In other words, returns data follow heavy-tailed distribution. In this project, Student-t and Burr Type II distributions are employed for ARCH/GARCH class of models and Stochastic Volatility (SV) model. For parameter estimation, we have used the Maximum Likelihood (ML) method and Quasi Maximum Likelihood (QML) method for ARCH/GARCH and SV models, respectively. Comparisons among these volatility models with Normal, Student-t, and Burr Type II distributions are provided. <br />
|
|---|