A1674
Title: Closing the gap between state-space and score-driven models
Authors: Xia Zou - Vrije Universiteit Amsterdam (Netherlands) [presenting]
Andre Lucas - VU University Amsterdam (Netherlands)
Yicong Lin - Vrije Universiteit Amsterdam (Netherlands)
Abstract: State-space and score-driven models are compared for option implied volatility surface dynamics. Point forecasts of both models behave similarly, but density forecasts of plain-vanilla score-driven models are substantially worse. This phenomenon is explained, and it shows how a simple adjustment of the measurement density of the score-driven model can put both models back on an equal footing. The score-driven models can subsequently be easily extended with non-Gaussian features to better fit the data without complicating parameter estimation. The findings are illustrated using S\&P500 index options implied volatility surfaces.
