A Semiparametric Approach to the Detection of Change-points in Volatility Dynamics of Financial Data

TL;DR

This paper presents a semiparametric change-point detection method for financial volatility data that improves accuracy over traditional parametric GARCH models by incorporating flexible error distributions.

Contribution

It introduces a novel semiparametric GARCH-based algorithm with penalized likelihood and binary segmentation for more reliable change-point detection in financial volatility.

Findings

Outperforms Quasi-MLE in change-point detection accuracy

Effective in diverse volatility scenarios

Retains GARCH structure with added flexibility

Abstract

One of the most important features of financial time series data is volatility. There are often structural changes in volatility over time, and an accurate estimation of the volatility of financial time series requires careful identification of change-points. A common approach to modeling the volatility of time series data is the well-known GARCH model. Although the problem of change-point estimation of volatility dynamics derived from the GARCH model has been considered in the literature, these approaches rely on parametric assumptions of the conditional error distribution, which are often violated in financial time series. This may lead to inaccuracies in change-point detection resulting in unreliable GARCH volatility estimates. This paper introduces a novel change-point detection algorithm based on a semiparametric GARCH model. The proposed method retains the structural advantages of the GARCH process while incorporating the flexibility of nonparametric conditional error distribution. The approach utilizes a penalized likelihood derived from a semiparametric GARCH model and an efficient binary segmentation algorithm. The results show that in terms of change-point estimation and detection accuracy, the semiparametric method outperforms the commonly used Quasi-MLE (QMLE) and other variations of GARCH models in wide-ranging scenarios.