A Parsimonious and Interpretable Factor Model of the Implied Volatility SurfaceÂ
with Wei Wei and Dan Zhu
[Download Most Recent Version] [Previous Version on SSRN]
The implied volatility surface (IVS) is a key gauge of market uncertainty, with implications for derivative pricing, risk management, and macro-financial analysis. This paper introduces a parsimonious factor-based model that decomposes the IVS into five components: level, tail asymmetry, term structure, curvature, and deviation. By incorporating the asymptotic properties of the IVS, the proposed framework offers a structurally grounded representation with enhanced interpretability. Empirical application to S&P 500 European options demonstrates that the model (i) recovers the asymptotic properties under specific underlying tail conditions; (ii) robustly captures the global geometry of the IVS, leading to consistent outperformance over traditional polynomial-based benchmarks in both surface fitting and forecasting; and (iii) provides predictive power for both realized variance and macroeconomic conditions. The factors prove particularly informative during episodes of heightened uncertainty, including the 2008 Global Financial Crisis and the 2020 COVID-19 shock.
Keywords: implied volatility surface, factor models, volatility forecasting