Font Size: a A A

Empirical Research On Volatility Forecast Models Based On CSI300Index

Posted on:2014-03-27Degree:MasterType:Thesis
Country:ChinaCandidate:D JinFull Text:PDF
GTID:2269330422964553Subject:Applied Statistics
Abstract/Summary:PDF Full Text Request
Volatility, as a measure of market risk, its effective estimation has an direct impacton asset pricing, asset allocation and risk management. Whether we can understand andmeasure the market risk and pricing the risk efficiently or not related to asset allocationefficiency in economic activities and economic operation cost, which is the common concernproblem of the investors, business and government department. At the time of the CSI300stock index options, which is the first option in domestic coming soon, it’s play animportant role in the accurate pricing of stock index options to study the variation ruleof its volatility and model it. It also has important theoretical and practical significancein the management, avoidance and control financial risks, and will affect the long-termdevelopment of China’s options market.This paper provides through and detailed introductions of the characteristics ofthe market volatility and the volatility forecasting models. Realized volatility based onhigh-frequency data is introduced as a proxy of the real market volatility.We use statisticalmethods to determine the optimal sampling frequency of CSI300Index is15min. Basedon the CSI300Index daily yield data since2005when it had been released, I do apreliminary study on the characteristics of the volatility in Chinese stock market. Thensome of the major volatility models were used to forecast the volatility on the basis ofthe nearly five years intraday high frequency data of the CSI300Index, which is calledthe rolling time window out-of-sample forecast. Finally, we use6loss functions and theDW test to compare the prediction result between the realized volatility models based onintraday high frequency data and the GARCH models and SV models which are basedon the daily return data. The main contribution of this paper is to provide comprehensiveintroductions to newer volatility forecast models, and compare realized volatility modelsbased on the high-frequency data with the traditional ones. At last, taking into accountof the characteristics of China’s stock market, the paper provides some references andsuggestions to volatility forecast, derivatives pricing and risk management.
Keywords/Search Tags:Volatility forecast, Realized volatility, Intraday high frequency data, DWtest
PDF Full Text Request
Related items