Papers That Changed How Markets See Volatility
- Connect models to live data streams and observe regime shifts in real time
- See a clear plan from theoretical framework to practical implementation
- Adjust parameters as market conditions evolve without rebuilding from scratch
- Equity
- Rates
- FX
- Commodities
-
Regime Papers 87%
-
Clustering Studies 64%
-
Risk Applications 91%
Featured Research
The working paper series covers latent variable methods for volatility clustering, regime-switching detection in high-frequency data, and nonlinear dynamics in derivatives markets. Recent publications examine particle filtering for real-time state inference, Hawkes process extensions for self-exciting volatility, and robust validation frameworks that prevent overfitting in regime models. Each paper includes replication code and datasets where licensing permits. The team maintains a public repository of estimation algorithms and backtesting utilities used in the research.
Foundational Regime Paper
The first publication introduced a Markov-switching framework for equity volatility that outperformed fixed-parameter GARCH by separating crisis and calm regimes. The model correctly identified the 2020 market dislocation three weeks before standard risk tools flagged elevated volatility. Replication code attracted over eight thousand downloads in the first year.
High-Frequency Extension
The team extended regime detection to tick-level data using particle filtering methods. This breakthrough enabled real-time state inference rather than end-of-day summaries. The paper demonstrated that intraday volatility clustering contains regime information invisible at daily frequencies. Implementation required solving computational bottlenecks that had prevented practical deployment.
Hawkes Process Integration
A landmark paper combined self-exciting point processes with latent regime variables to model volatility clustering and jump dynamics simultaneously. The framework captured contagion effects across asset classes that pure diffusion models missed entirely. The methodology was adopted by two central banks for stress-testing applications within eighteen months of publication.
Validation Framework
Recognizing that regime models are prone to overfitting, the team published a comprehensive validation protocol. The framework requires separate testing within each detected regime, walk-forward analysis, and comparison against random-walk benchmarks. This paper established standards that subsequent researchers in the field now follow routinely.
Derivatives Pricing Applications
The most recent major publication applies latent regime detection to implied volatility surfaces. The model identifies regime-dependent skew and term structure patterns that static surface models cannot represent. Practitioners use the framework for timing hedging adjustments and identifying mispriced options during regime transitions.
Cross-Asset Clustering
Current research examines how volatility regimes propagate across equity, fixed income, foreign exchange, and commodity markets. Early results suggest that latent state variables estimated from one asset class improve forecasting accuracy in others. The team is developing a unified framework for global regime detection with applications to macro hedging and systemic risk monitoring.
Research in Practice
Two case studies demonstrate how latent volatility clustering detection moves from academic publication to institutional application. Each example shows the model, the implementation challenge, and the outcome for the practitioner.
Crisis Detection for a Regional Bank
Commodity Hedging for an Exporter
A dairy exporter faced volatile milk powder prices that standard futures hedging could not address. The latent regime model identified two distinct volatility states in dairy derivatives. During high-volatility regimes, the firm increased hedge ratios and extended contract maturities. During calm regimes, it reduced hedging costs. Over three years, the approach lowered effective hedging expenses by a measurable margin without increasing unhedged exposure.
Collaborate on Volatility Clustering Research
Working paper series with peer feedback from quantitative finance researchers and econometricians
Replication datasets and estimation code provided for all published methodology papers
Conference presentation invitations accepted for topics related to latent volatility clustering
Institutional dataset access available for qualified researchers with verified academic affiliations