Scholarlink

<p>This paper investigates the daily exchange rate volatility of the Swedish krona (SEK) against the USD, EUR, GBP, and NOK over the period 2010–2023. Using asymmetric power ARCH (APARCH) models, the analysis uncovers significant differences in volatility dynamics across currency pairs and subperiods. A negative asymmetric return–volatility relationship is identified for SEK/EUR, indicating stronger reactions to negative shocks, while an inverted asymmetry is found for SEK/NOK—a pattern rarely documented in prior studies. No significant asymmetry is detected for SEK/USD or SEK/GBP. To address the limitations of conventional parametric models in small, open economies, a distribution-free support vector regression (SVR) approach with a wavelet kernel is integrated within the APARCH framework. In this study, the SVR-APARCH model demonstrates superior forecasting performance compared with standard APARCH models estimated via maximum likelihood estimation. Furthermore, it is shown to be generally more suitable for volatility forecasting than the hybrid ANN-APARCH and boosting-APARCH models. The results underscore the model’s enhanced capacity to capture the complex and nonlinear features of exchange rate volatility. The study contributes to the literature by providing robust evidence of asymmetric volatility patterns in SEK exchange rates and by introducing an effective hybrid modeling approach for improved volatility forecasting.</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

Investigation of Swedish Krona exchange rate volatility using APARCH-Support Vector Regression

  • 期刊论文

  • 出版模式：

    开放获取

  • 发表日期：

    2026年2月11日

查看全文
  • Hyunjoo Kim Karlsson,
  • Yushu Li

摘要

This paper investigates the daily exchange rate volatility of the Swedish krona (SEK) against the USD, EUR, GBP, and NOK over the period 2010–2023. Using asymmetric power ARCH (APARCH) models, the analysis uncovers significant differences in volatility dynamics across currency pairs and subperiods. A negative asymmetric return–volatility relationship is identified for SEK/EUR, indicating stronger reactions to negative shocks, while an inverted asymmetry is found for SEK/NOK—a pattern rarely documented in prior studies. No significant asymmetry is detected for SEK/USD or SEK/GBP. To address the limitations of conventional parametric models in small, open economies, a distribution-free support vector regression (SVR) approach with a wavelet kernel is integrated within the APARCH framework. In this study, the SVR-APARCH model demonstrates superior forecasting performance compared with standard APARCH models estimated via maximum likelihood estimation. Furthermore, it is shown to be generally more suitable for volatility forecasting than the hybrid ANN-APARCH and boosting-APARCH models. The results underscore the model’s enhanced capacity to capture the complex and nonlinear features of exchange rate volatility. The study contributes to the literature by providing robust evidence of asymmetric volatility patterns in SEK exchange rates and by introducing an effective hybrid modeling approach for improved volatility forecasting.