Jump risk premia in the presence of clustered jumps

Battle Count: 77

Highly relevant for quantitative trading in cryptocurrency options markets. The paper provides actionable trading signals based on jump risk premia for delta-hedged option strategies (selling calls when positive jump premium is high, selling puts when negative jump premium is low). The predictive power for BTC futures cost-of-carry (R²=31% in full model) is directly applicable to basis trading strategies. The model's ability to capture regime-conditional skewness is valuable for options market-making and volatility trading strategies.

High complexity due to: (1) bivariate Hawkes process with self- and cross-excitation requiring recursive intensity computation; (2) non-linear system of ODEs for MGF computation solved via Runge-Kutta methods; (3) two-stage calibration procedure combining MLE for intensity parameters and MAPE minimization for risk premia; (4) Lewis-Lipton Fourier inversion for option pricing; (5) POT procedure for jump identification; (6) delta-hedging with hourly rebalancing for empirical validation. Python prototype code is available but production implementation requires significant engineering.

The paper provides Python code via a GitHub repository (StochVolModels) and mentions Quantlet for reproduction. The Deribit API provides access to BTC price and options data. The two-stage calibration procedure is well-documented with explicit equations. However, the full empirical pipeline including data preprocessing and specific parameter choices for the POT procedure may require additional implementation effort.