FX option data predicts large currency swings
A Bank of Japan working paper develops a model to predict large currency swings using transaction-level foreign exchange option data from trade repositories. The study finds that heterogeneous currency risk perceptions significantly improve these predictions.
Unlocking currency tail risks
The paper introduces a novel model for forecasting significant currency fluctuations, specifically in the USD/JPY rate, leveraging transaction-level data from foreign exchange options collected by trade repositories.
This granular data allows researchers to capture diverse currency risk perceptions among individual market participants, a key differentiator from models relying solely on macroeconomic variables.
By employing a quantile regression approach combined with machine learning for variable selection, the study empirically demonstrates that these extracted market participants' views substantially enhance the accuracy of predicting large currency swings.
This method focuses on the "tails" of the currency return distribution, addressing extreme appreciation and depreciation events.
Beyond macro fundamentals
Large currency swings, exemplified by the sharp JPY appreciation in August 2024, can significantly impact financial conditions and business activities.
Traditional models often rely on macroeconomic variables and assume homogeneous information among market participants.
This research, however, builds on market microstructure studies by extracting heterogeneous currency risk perceptions from transaction-level FX option data.
These investor views, reflecting expectations of future volatility and hedging, provide valuable information for forecasting tail risks, a distinct contribution to the literature.
A new lens on FX volatility
This study offers a crucial advancement for financial authorities seeking to better forecast extreme currency fluctuations.
By tapping into granular, transaction-level data, it moves beyond aggregate indicators to reveal the predictive power of individual market participants' nuanced risk perceptions.
While data limitations and time-varying market narratives remain challenges, the findings underscore the value of micro-level insights for macro-level stability monitoring.