Autoregressive Conditional Heteroskedastic Models推荐.pdfVIP

CHAPTER 8 Autoregressive Conditional Heteroskedastic Models n linear regression analysis, a standard assumption is that the variance of Iall squared error terms is the same. This assumption is called homoske- dasticity (constant variance). However, many time series data exhibit het- eroskedasticity, where the variances of the error terms are not equal, and in which the error terms may be expected to be larger for some observations or periods of the data than for others. The issue is then how to construct models that accommodate heteroskedasticity so that valid coefficient esti- mates and models are obtained for the variance of the error terms. Autore- gressive conditional heteroskedasticity (ARCH) models are the topic of this chapter. They have proven to be very useful in finance to model return variance or volatility of major asset classes including equity, fixed income, and foreign exchange. Understanding the behavior of the variance of the return process is important for forecasting as well as pricing option-type derivative instruments since the variance is a proxy for risk. Although asset returns, such as stock and exchange rate returns, appear to follow a martingale difference sequence, observation of the daily return plots shows that the amplitude of the returns varies across time. A widely observed phenomenon in finance confirming this fact is the so- called volatility clustering. This refers to the tendency of large changes in asset prices (either positive or negative) to be followed by large changes and small changes to be followed by small changes. Hence, there is tempo- ral dependence in asset returns. Typically, they are not even close to being independently and identically distributed (IID). This pattern in the volatil- ity of asset returns was first