초록
본 연구에서는 금융시계열 변동성 측정을 위한 다양한 방법들을 소개하고 비교분석 하였다. 최근 들어 활발한 연구가 이루어지고 있는 고빈도(high frequency) 자료에 기초한 변동성 측정방법을 국내 주가에 적용시켜 1분 단위 고빈도 주가로부터 일별 변동성을 계산하였다. 또한, 모형 기반 방법인 GARCH와 자료 기반 방법인 역사적 변동성(historical volatility)을 융합하여 새로운 변동성 측정법을 제안하였다.
Various computational methods for obtaining volatilities for financial time series are reviewed and compared with each other. We reviewed model based GARCH approach as well as the data based method which can essentially be regarded as a smoothing technique applied to the squared data. The method for high frequency data is focused to obtain the realized volatility. A hybrid method is suggested by combining the model based GARCH and the historical volatility which is a data based method. Korea stock prices are analysed to illustrate various computational methods for volatilities.