VOLATILITY DIFFUSION AND UNCERTAINTY IN THE CURRENCY MARKETS: CASE OF USD/JPY MARKET

Özet

This study investigates the symmetry and asymmetry of volatility dynamics in the USD/JPY foreign-exchange market using high-frequency data and advanced conditional variance models. The analysis compares the performance of Unified GARCH-Itô, GARCH-Itô-Jump, and EGARCH specifications to identify how volatility evolves under continuous and abrupt uncertainty. The continuoustime GARCH-Itô framework is particularly suited to this setting, as uncertainty in foreign exchange markets propagates almost continuously rather than in discrete steps. Empirical results show that USD/JPY volatility is highly persistent, predominantly symmetric, and diffusion-driven, with jump components contributing negligible additional explanatory power. The Exponential GARCH reveals temporary asymmetry during specific periods. In particular, positive return shocks, during yen depreciation episodes, amplify volatility more than negative ones, without altering long-run equilibrium behavior. Value-at-Risk backtesting indicates mild risk overestimation during tranquil periods, suggesting that parametric models require state-dependent calibration across different market environments.

Abstract

This study investigates the symmetry and asymmetry of volatility dynamics in the USD/JPY foreign-exchange market using high-frequency data and advanced conditional variance models. The analysis compares the performance of Unified GARCH-Itô, GARCH-Itô-Jump, and EGARCH specifications to identify how volatility evolves under continuous and abrupt uncertainty. The continuoustime GARCH-Itô framework is particularly suited to this setting, as uncertainty in foreign exchange markets propagates almost continuously rather than in discrete steps. Empirical results show that USD/JPY volatility is highly persistent, predominantly symmetric, and diffusion-driven, with jump components contributing negligible additional explanatory power. The Exponential GARCH reveals temporary asymmetry during specific periods. In particular, positive return shocks, during yen depreciation episodes, amplify volatility more than negative ones, without altering long-run equilibrium behavior. Value-at-Risk backtesting indicates mild risk overestimation during tranquil periods, suggesting that parametric models require state-dependent calibration across different market environments.

Yazarlar

Erdem KILIÇ

Anahtar Kelimeler

Symmetry, uncertainty, volatility symmetry, financial markets, high-frequency data, foreign exchange markets.

JEL Codes

C40, C50

Yayın Bilgileri

Cilt 6, Sayı 1, 2026 · Sayfa 29-42

DOI: 10.52898/ijif.2026.2

Dosyalar