Forecasting Value-at-Risk Based on Variant Smooth Transition Heteroskedastic Models
碩士 === 逢甲大學 === 統計學系統計與精算碩士班 === 102 === Value-at-Risk (VaR) is a popular instrument for financial risk management. This paper seeks to evaluate performance in VaR measures in a class of smooth transition (ST) heteroskedastic models. Three distinct ST functions with generalized autoregressive conditi...
| Main Author: | |
|---|---|
| Other Authors: | |
| Format: | Others |
| Language: | en_US |
| Published: |
2014
|
| Online Access: | http://ndltd.ncl.edu.tw/handle/51881023695477942734 |
| Summary: | 碩士 === 逢甲大學 === 統計學系統計與精算碩士班 === 102 === Value-at-Risk (VaR) is a popular instrument for financial risk management. This paper seeks to evaluate performance in VaR measures in a class of smooth transition (ST) heteroskedastic models. Three distinct ST functions with generalized autoregressive conditional heteroskedasticity (GARCH) models are employed: the first-order, the second-order logistic functions, and the exponential function. We investigate the properties of the second-order logistic ST function which introduces two smooth transitions among three regimes defined by two thresholds. The Bayesian solution is adapted and designed for parameter estimations through the Markov chain Monte Carlo scheme. We conduct an out-of-sample forecast of the proposed three variant ST-GARCH models with some existing and competing models, for nineteen stock market returns. Three backtests are used to measure and assert the forecast performances of a variety of risk models. The performance of a variety of risk models is examined by out-of-sample forecasts and the forecast accuracy for all models is diagnosed by three volatility proxies. Results reveal that the three ST-GARCH models were clearly favoured at the 1% level.
|
|---|