Volatility clustering

Volatility clustering refers to the phenomenon in financial time series where periods of high variability are followed by periods of low variability, and vice versa. This effect can be observed in various macroeconomic time series, such as manufacturing activities, consumer price indices, and housing and asset prices. It is a well-documented phenomenon that can be captured by statistical models such as S t (Hamilton 1988; Kim, Shephard, and Chib 1998; Ang and Timmermann 2012).From: Next Generation of Data Mining [2019], Time Series [2019], Penalized Estimation of Sparse Markov Regime-Switching Vector Auto-Regressive Models [2023]

Nonlinear Time Series Analysis

View Chapter

Purchase Book

Published in Tucker S. McElroy, Dimitris N. Politis, Time Series, 2019

Tucker S. McElroy, Dimitris N. Politis

Remark 11.3.1. Volatility Clustering Financial time series exhibit volatility clustering, i.e., periods of higher variability followed by periods of low variability, and so forth. For example, with intra-day trading data, the beginning and end of the trading day are characterized by heightened activity, yielding a higher variance in stock price during these times. Also, at the onset of a recession, expansion, or noteworthy event there can be increased volatility; see Figure 1.4, which exhibits volatility clustering in Dow Jones log returns.

A hybrid machine learning method for procurement risk assessment of non-ferrous metals for manufacturing firms

View Article

Journal Information

Published in International Journal of Computer Integrated Manufacturing, 2022

Jian Ni, Yan Hu, Ray Y. Zhong

The GARCH model is originally developed to deal with the commonly observed heteroscedasticity in financial time series, which cannot be handled by traditional regression analysis. Engle (1982) firstly proposed a p-order autoregressive conditional heteroscedasticity (i.e. ARCH (p)) model to analyze time series with volatility clustering characteristics from the perspective of conditional distribution. The disadvantage of the ARCH (p) model is that a number of parameters are needed if the p-order is too large. Bollerslev (1986) then introduced the GARCH model to solve this problem. GARCH models suppose that the variance of the current term is a function of both the square terms of the past error term and previous conditional variance terms. To formally introduce the GARCH model, several required math symbols are summarized in Table 1

Volatility clustering

Explore chapters and articles related to this topic

Nonlinear Time Series Analysis

A hybrid machine learning method for procurement risk assessment of non-ferrous metals for manufacturing firms