Surface Classification for Direction Predictions of the Foreign Exchange Implied Volatility Surface

Skip Navigation
menu
COVID-19 Update: Trusted Partner Information

click here

0 of 0 Search Results
 |   Displaying

Sorry, we couldn't find any results matching

Please:

MACRO STRATEGY

Surface Classification for Direction Predictions of the Foreign Exchange Implied Volatility Surface

Jun. 08, 2020

Predictions of the value of the FX implied volatility of an asset are commonly used as inputs to probabilistic models of profitability before expiration. However, depending on trading strategy, a metric that quantifies magnitudes of changes in implied volatility can be of supplemental value or of more interest than level prediction.

  • Through transformation of the distribution of the observed implied volatilities, a predictable target representing the direction of change in the implied volatility of an asset can be generated.
  • Correlation analysis of this target leads to significant dimensionality reduction with little loss of information.
  • Cluster analysis of implied volatility surface data shows clear separation by overall change in the directions of surfaces.
  • The clustering of surfaces shows evidence of Markov properties, leading to predictability of future values of cluster classification.
  • Surface cluster classification shows predictability through historical, differenced, implied volatilities, interest rate and economic indicator data.
  • The results of our directional forecasting models (i.e., Nested Markov Chain, Random Forest and Ensemble) show lower out-of-sample error when compared to deriving forecasts of the target using traditional value forecasts.