Forecasting security's volatility using low-frequency historical data, high-frequency historical data and option-implied volatility
Pith reviewed 2026-05-25 02:08 UTC · model grok-4.3
The pith
Two GARCH-Itô models that integrate low-frequency, high-frequency and option-implied volatility data outperform benchmarks for security volatility forecasts at five-minute sampling intervals.
A machine-rendered reading of the paper's core claim, the machinery that carries it, and where it could break.
Core claim
The GARCH-Itô-OI model treats option-implied volatility as an observable exogenous variable influencing the security's future volatility, while the GARCH-Itô-IV model constructs a relationship between option-implied volatility and the security's volatility to extract information; both models integrate low- and high-frequency historical data and exhibit superior out-of-sample forecasting performance compared with existing models when high-frequency sampling occurs at five-minute intervals.
What carries the argument
GARCH-Itô-OI and GARCH-Itô-IV models that extend the GARCH-Itô framework by incorporating option-implied volatility either as an exogenous input or through a constructed link to extract information.
If this is right
- Volatility forecasts improve when low-frequency data, high-frequency data and option-implied volatility are used jointly in the integrated models.
- The quasi-maximum likelihood estimators for model parameters are consistent and asymptotically normal.
- The forecasting gains are observed specifically when high-frequency data is sampled at five-minute intervals.
- Simulation results support the theoretical properties and the empirical advantages of the two models.
Where Pith is reading between the lines
- The models could be tested on additional asset classes to check whether the five-minute advantage generalizes beyond the studied securities.
- Comparing the direct-influence version against the constructed-relationship version on different markets might reveal when one specification is preferable.
- Extending the analysis to other sampling intervals could identify whether an optimal frequency exists for each security or market.
Load-bearing premise
Option-implied volatility either directly influences future volatility or supplies extractable information through the constructed relationship without model misspecification that would bias the forecasts.
What would settle it
An out-of-sample test on the same or comparable securities showing that the GARCH-Itô-OI and GARCH-Itô-IV models do not produce lower mean squared forecast errors than standard GARCH or HAR models at the five-minute high-frequency sampling interval.
read the original abstract
Low-frequency historical data, high-frequency historical data and option data are three major sources, which can be used to forecast the underlying security's volatility. In this paper, we propose two econometric models, which integrate three information sources. In GARCH-It\^{o}-OI model, we assume that the option-implied volatility can influence the security's future volatility, and the option-implied volatility is treated as an observable exogenous variable. In GARCH-It\^{o}-IV model, we assume that the option-implied volatility can not influence the security's volatility directly, and the relationship between the option-implied volatility and the security's volatility is constructed to extract useful information of the underlying security. After providing the quasi-maximum likelihood estimators for the parameters and establishing their asymptotic properties, we also conduct a series of simulation analysis and empirical analysis to compare the proposed models with other popular models in the literature. We find that when the sampling interval of the high-frequency data is 5 minutes, the GARCH-It\^{o}-OI model and GARCH-It\^{o}-IV model has better forecasting performance than other models.
Editorial analysis
A structured set of objections, weighed in public.
Referee Report
Summary. The paper proposes the GARCH-Itô-OI and GARCH-Itô-IV models to integrate low-frequency historical data, high-frequency historical data, and option-implied volatility for forecasting security volatility. It derives quasi-MLE estimators with asymptotic properties and, through simulation and empirical analyses, claims that these models have better forecasting performance than other models when the high-frequency data sampling interval is 5 minutes.
Significance. If the reported out-of-sample forecasting superiority is robust, the models could offer a valuable advancement in volatility forecasting by combining multiple data sources, with potential applications in financial risk management.
major comments (1)
- Abstract: the assertion of superior forecasting performance lacks any supporting metrics, baselines, sample sizes, or robustness checks, making the central empirical claim unverifiable from the provided text.
Simulated Author's Rebuttal
We thank the referee for the detailed review and constructive suggestion. We address the single major comment below and will revise the manuscript accordingly.
read point-by-point responses
-
Referee: Abstract: the assertion of superior forecasting performance lacks any supporting metrics, baselines, sample sizes, or robustness checks, making the central empirical claim unverifiable from the provided text.
Authors: We agree that the abstract would benefit from greater specificity to allow readers to assess the central claim immediately. The body of the manuscript (Sections 4 and 5) already reports the simulation design, the empirical sample (including the 5-minute high-frequency sampling interval), the competing models used as baselines, and the out-of-sample forecasting metrics (MSE and MAE). In the revised version we will condense the key quantitative results—such as the reported percentage improvements over the benchmark models and the data period—into the abstract while preserving its length constraints. revision: yes
Circularity Check
No significant circularity
full rationale
The paper derives quasi-MLE estimators for its GARCH-Itô-OI and GARCH-Itô-IV models, establishes their asymptotic properties, and reports simulation plus empirical out-of-sample forecasting comparisons against other models. These performance claims rest on external benchmark comparisons rather than any equation or parameter that reduces by construction to the fitted inputs themselves. No self-citation is invoked as a load-bearing uniqueness theorem, no ansatz is smuggled via prior work, and no prediction is statistically forced by the estimation procedure. The derivation chain is therefore self-contained against external benchmarks.
Axiom & Free-Parameter Ledger
free parameters (1)
- GARCH-Itô model parameters
axioms (1)
- domain assumption Option-implied volatility contains information about future volatility that can be incorporated either as an exogenous variable or via a constructed relationship
discussion (0)
Sign in with ORCID, Apple, or X to comment. Anyone can read and Pith papers without signing in.