Data-Based Parametrization for Affine GARCH Models Across Multiple Time Scales—Roughness Implications

This paper revisits the topic of time-scale parameterizations of the Heston–Nandi GARCH (1,1) model to create a new, theoretically valid setting compatible with real financial data. We first estimate parameters using three US market indices and six frequencies to let data reveal the correct, data-im...

Full description

Saved in:
Bibliographic Details
Main Authors: Marcos Escobar-Anel, Sebastian Ferrando, Fuyu Li, Ke Xu
Format: Article
Language:English
Published: MDPI AG 2025-02-01
Series:Econometrics
Subjects:
Online Access:https://www.mdpi.com/2225-1146/13/1/6
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1850203782879641600
author Marcos Escobar-Anel
Sebastian Ferrando
Fuyu Li
Ke Xu
author_facet Marcos Escobar-Anel
Sebastian Ferrando
Fuyu Li
Ke Xu
author_sort Marcos Escobar-Anel
collection DOAJ
description This paper revisits the topic of time-scale parameterizations of the Heston–Nandi GARCH (1,1) model to create a new, theoretically valid setting compatible with real financial data. We first estimate parameters using three US market indices and six frequencies to let data reveal the correct, data-implied, time-scale parameterizations. We compared the data-implied parametrization to two popular candidates in the literature, demonstrating structurally different continuous-time limits, i.e., the data favor fractional Brownian motion (fBM)—instead of the standard Brownian motion (BM)-based parametrization. We then propose a theoretically flexible time-scale parameterization compatible with this fBM behavior. In this context, a fractional derivative analysis of our empirically based parametrization is performed, confirming an anomalous diffusion in the continuous-time limit. Such a finding is yet another endorsement of the recent and popular stylized fact known as rough volatility.
format Article
id doaj-art-204c5dd4a5ea4cb280986ff34660eebf
institution OA Journals
issn 2225-1146
language English
publishDate 2025-02-01
publisher MDPI AG
record_format Article
series Econometrics
spelling doaj-art-204c5dd4a5ea4cb280986ff34660eebf2025-08-20T02:11:25ZengMDPI AGEconometrics2225-11462025-02-01131610.3390/econometrics13010006Data-Based Parametrization for Affine GARCH Models Across Multiple Time Scales—Roughness ImplicationsMarcos Escobar-Anel0Sebastian Ferrando1Fuyu Li2Ke Xu3Department of Statistical and Actuarial Sciences, University of Western Ontario, London, ON N6A 5B7, CanadaDepartment of Mathematics, Toronto Metropolitan University, Toronto, ON M5B 2K3, CanadaDepartment of Economics, University of Victoria, Victoria, BC V8P 5C2, CanadaDepartment of Economics, University of Victoria, Victoria, BC V8P 5C2, CanadaThis paper revisits the topic of time-scale parameterizations of the Heston–Nandi GARCH (1,1) model to create a new, theoretically valid setting compatible with real financial data. We first estimate parameters using three US market indices and six frequencies to let data reveal the correct, data-implied, time-scale parameterizations. We compared the data-implied parametrization to two popular candidates in the literature, demonstrating structurally different continuous-time limits, i.e., the data favor fractional Brownian motion (fBM)—instead of the standard Brownian motion (BM)-based parametrization. We then propose a theoretically flexible time-scale parameterization compatible with this fBM behavior. In this context, a fractional derivative analysis of our empirically based parametrization is performed, confirming an anomalous diffusion in the continuous-time limit. Such a finding is yet another endorsement of the recent and popular stylized fact known as rough volatility.https://www.mdpi.com/2225-1146/13/1/6Affine GARCHmaximum likelihood estimationtime-scale parameterizationrough volatility
spellingShingle Marcos Escobar-Anel
Sebastian Ferrando
Fuyu Li
Ke Xu
Data-Based Parametrization for Affine GARCH Models Across Multiple Time Scales—Roughness Implications
Econometrics
Affine GARCH
maximum likelihood estimation
time-scale parameterization
rough volatility
title Data-Based Parametrization for Affine GARCH Models Across Multiple Time Scales—Roughness Implications
title_full Data-Based Parametrization for Affine GARCH Models Across Multiple Time Scales—Roughness Implications
title_fullStr Data-Based Parametrization for Affine GARCH Models Across Multiple Time Scales—Roughness Implications
title_full_unstemmed Data-Based Parametrization for Affine GARCH Models Across Multiple Time Scales—Roughness Implications
title_short Data-Based Parametrization for Affine GARCH Models Across Multiple Time Scales—Roughness Implications
title_sort data based parametrization for affine garch models across multiple time scales roughness implications
topic Affine GARCH
maximum likelihood estimation
time-scale parameterization
rough volatility
url https://www.mdpi.com/2225-1146/13/1/6
work_keys_str_mv AT marcosescobaranel databasedparametrizationforaffinegarchmodelsacrossmultipletimescalesroughnessimplications
AT sebastianferrando databasedparametrizationforaffinegarchmodelsacrossmultipletimescalesroughnessimplications
AT fuyuli databasedparametrizationforaffinegarchmodelsacrossmultipletimescalesroughnessimplications
AT kexu databasedparametrizationforaffinegarchmodelsacrossmultipletimescalesroughnessimplications