Predicting Realized Variance Out of Sample: Can Anything Beat The Benchmark?

ArXiv ID: 2506.07928 “View on arXiv”

Authors: Austin Pollok

Abstract

The discrepancy between realized volatility and the market’s view of volatility has been known to predict individual equity options at the monthly horizon. It is not clear how this predictability depends on a forecast’s ability to predict firm-level volatility. We consider this phenomenon at the daily frequency using high-dimensional machine learning models, as well as low-dimensional factor models. We find that marginal improvements to standard forecast error measurements can lead to economically significant gains in portfolio performance. This makes a case for re-imagining the way we train models that are used to construct portfolios.

Keywords: volatility forecasting, realized volatility, machine learning, portfolio construction, equity options

Complexity vs Empirical Score

• Math Complexity: 7.5/10
• Empirical Rigor: 8.0/10
• Quadrant: Holy Grail
• Why: The paper employs advanced statistical and machine learning methods for volatility forecasting, supported by formal derivations. It demonstrates strong empirical rigor through an extensive out-of-sample analysis on S&P 500 options data from 1993–2019, evaluating portfolio performance and economic significance.

  flowchart TD
    A["Research Goal: Determine if marginal improvements in volatility forecasting error lead to economically significant gains in portfolio construction for equity options."]
    
    subgraph B ["Data & Inputs"]
        B1["High-dimensional machine learning models"]
        B2["Low-dimensional factor models"]
        B3["Realized Variance (benchmark)"]
    end
    
    subgraph C ["Methodology"]
        C1["Daily frequency analysis"]
        C2["Compare model predictions vs. Realized Variance"]
        C3["Measure Forecast Errors"]
    end
    
    subgraph D ["Computational Process"]
        D1["Construct portfolios based on improved forecasts"]
        D2["Evaluate economic significance of gains"]
    end
    
    E["Key Finding: Marginal forecast improvements yield significant economic gains, advocating for re-imagined model training approaches."]

    A --> B
    B --> C
    C --> D
    D --> E