Paper: arXiv 2410.22471
Authors: Jihyun Park, Andrey Sarantsev
Abstract
We model time series of VIX (monthly average) and monthly stock index returns. We use log-Heston model: logarithm of VIX is modeled as an autoregression of order 1. Our main insight is that normalizing monthly stock index returns (dividing them by VIX) makes them much closer to independent identically distributed Gaussian. The resulting model is mean-reverting, and the innovations are non-Gaussian. The combined stochastic volatility model fits well, and captures Pareto-like tails of real-world stock market returns. This works for small and large stock indices, for both price and total returns.
Complexity vs Empirical Score
- Math Complexity: 6.5/10
- Empirical Rigor: 8.0/10
- Quadrant: Holy Grail — high math complexity, high empirical rigor
Why this score: The paper employs a sophisticated stochastic differential equation model (log-Heston) with AR(1) dynamics and non-Gaussian innovations, but is grounded in extensive empirical analysis of real VIX and index return data over 462 months, including statistical tests (skewness, kurtosis, ACF) and model fitting with public data and code.
Research Flowchart
flowchart TD
A["Research Goal: Model VIX & Stock Returns<br>to understand volatility dynamics"] --> B["Data Input<br>Monthly Avg VIX &<br>Stock Index Returns"]
B --> C["Methodology: Log-Heston Model<br>Log VIX ~ AR1 Process"]
C --> D["Key Insight: Normalization<br>Returns / VIX ≈ Gaussian I.I.D."]
D --> E["Computational Process<br>Fit Combined Stochastic<br>Volatility Model"]
E --> F{"Model Diagnostics"}
F -->|"Good Fit"| G["Key Findings<br>Mean-Reverting Process"]
F -->|"Tail Analysis"| H["Key Findings<br>Captures Pareto-like Tails"]
G & H --> I["Outcome<br>Validated for Small & Large Indices<br>(Price & Total Returns)"]
Evidence Card
What the authors claim, on what data, and who has checked it. Read from the paper's own text by a language model: every quote is verbatim, nothing was re-run. How to read this card
Proposes
The paper proposes a log‑Heston model for monthly average VIX and stock returns, using normalized returns (dividing by VIX) to achieve near‑Gaussian IID behavior.
Source sentence
We call it the log-Heston model.
Assumes 6
- 0 < β < 1 ensures mean‑reversion in the log‑Heston volatility equation. model
Source sentence
Here, β ∈(0, 1); this condition ensures mean-reversion.
- Wt are independent identically distributed mean‑zero innovations (non‑Gaussian). statistical
Source sentence
are independent identically distributed but not Gaussian.
- Zt are IID normal random variables (Gaussian marginal). statistical
Source sentence
Zt can be modeled as IID normal random variables: Zt ∼ N(µ, σ2).
- The volatility index VIX is observed directly, not inferred. data
Source sentence
Also, here we observe VIX directly, whereas in estimating GARCH‑type models we need to imply the volatility.
Show 2 more
- Monthly returns are more regular and suitable for IID Gaussian modeling. model
Source sentence
First, returns with larger time step exhibit more regularity. It is easier to write a model of them with IID Gaussian innovations.
- The initial condition V0 is independent of all Wt and has finite moments. model
Source sentence
the initial condition V0 is independent of all Wt, with E[V^u_0] < ∞.
Data
- VIXCLS / VXOCLS monthly average VIX series options public
Source sentence
The main data series in this article is Chicago Board of Options Exchange (CBOE) Volatiltiy Index (VIX). This is taken from the Federal Reserve Economic Data (FRED) web site.
- Monthly stock index returns (Top 30% and Bottom 30% portfolios) market-prices public
Source sentence
The data is taken from public financial data libraries: Federal Reserve Economic Data and Kenneth French’s Dartmouth College Financial Data.
Costs
Not applicable
Finds 5
- Normalizing monthly stock returns by VIX yields skewness and kurtosis much closer to zero, indicating near‑Gaussian IID behavior. statistical
Source sentence
... We see that the skewness and kurtosis for normalized retruns are much closer to zero.
- The log‑Heston model fits the monthly VIX series well: residuals are IID but exhibit exponential‑like tails, not Gaussian. statistical
Source sentence
This model fits well in the sense that residuals Wt are independent identically distributed. But they are not Gaussian.
- The combined model reproduces Pareto‑like tails of real‑world stock returns and guarantees finite moments up to a certain order. theoretical
Source sentence
It captures a well‑known property of real‑world stock
- Under the stated assumptions, the log‑Heston system possesses a unique stationary distribution and is V‑uniformly ergodic. theoretical
Source sentence
This process ((ln Vt, Zt), t ≥0) has a unique stationary distribution, and is V‑uniformly ergodic.
- Statistical analysis shows that ln Vt is well‑modeled by a mean‑reverting AR(1) process with finite exponential moments. statistical
Source sentence
Statistical analysis shows that ln Vt is well‑modeled by mean‑reverting autoregression of order 1.
Checked by others
No other paper in the archive has been found to support or contradict these findings (compared with 29 related papers). Coverage is partial, so this is neither replication nor refutation.
2 papers in the archive
Code
No repository by the authors found in the paper.
Extracted 2026-11-14 · same card as JSON · A language model chose these items: check the source sentence before relying on one.