Scholarlink

<p>In this paper, we develop a quasi-maximum likelihood estimation (QMLE) procedure using the information filter, a diffuse Kalman filtering version, to estimate the parameters of a non-stationary stochastic volatility (SV) model. In addition to modeling the high persistence volatility as a random walk, we incorporate the leverage effect and the heavy-tailed nature of asset price return distribution into our proposed method to enhance the SV model’s ability to accurately capture these stylized facts. With this aim, we expand upon the asymmetric specification of the SV model that was introduced in Chirico’s most recent work (2024), using both normal and Student’s <i>t</i>-distribution for the investigation of heavy-tailedness. We provide simulation and empirical validations to demonstrate our method’s effectiveness in handling the stylised facts compared with existing methods.</p>

错误：搜索内容不能为空，请输入英文关键词
错误：关键词超出字数限制，请精简
高级检索

Quasi-maximum likelihood estimation for non-stationary stochastic volatility models: diffuse Kalman filtering approach

  • 期刊论文

  • 发表日期：

    2025年9月29日

查看全文
  • Abdeljalil Settar,
  • Ghassane Benrhmach,
  • Sara Chegdal

摘要

In this paper, we develop a quasi-maximum likelihood estimation (QMLE) procedure using the information filter, a diffuse Kalman filtering version, to estimate the parameters of a non-stationary stochastic volatility (SV) model. In addition to modeling the high persistence volatility as a random walk, we incorporate the leverage effect and the heavy-tailed nature of asset price return distribution into our proposed method to enhance the SV model’s ability to accurately capture these stylized facts. With this aim, we expand upon the asymmetric specification of the SV model that was introduced in Chirico’s most recent work (2024), using both normal and Student’s t-distribution for the investigation of heavy-tailedness. We provide simulation and empirical validations to demonstrate our method’s effectiveness in handling the stylised facts compared with existing methods.