Quantitative analysis of Bitcoin derivatives market microstructure. Implements implied volatility surface modeling, risk-neutral density extraction, and econometric tests of price discovery mechanisms.
This project investigates three empirical questions about Bitcoin options markets:
- Does the options market lead spot price movements? (Granger causality, VAR models)
- What is the magnitude and dynamics of the variance risk premium? (IV vs realized volatility)
- How do monetary policy shocks propagate through the volatility term structure? (Event study methodology)
The analysis uses daily options data from Deribit (2023-2024) and spot prices from Binance, combined with FOMC announcement dates and macroeconomic indicators.
Data Sources
- Options: Deribit API (BTC perpetual options, all strikes and maturities)
- Spot: Binance WebSocket (tick-by-tick, aggregated to 1-minute OHLCV)
- Macro: FRED API (Federal Funds Rate, CPI, DXY index)
Derivatives Pricing
- Black-Scholes model for European options
- Implied volatility extraction via Newton-Raphson
- Greeks calculation (Delta, Gamma, Vega, Theta, Rho)
- Risk-neutral density via Breeden-Litzenberger (1978)
Econometric Models
- GARCH(1,1) for realized volatility
- VAR(p) for lead-lag relationships (AIC selection)
- Granger causality tests (F-statistic, bootstrap)
- Event study with cumulative abnormal returns
Statistical Tests
- Augmented Dickey-Fuller (stationarity)
- Ljung-Box (autocorrelation)
- White test (heteroskedasticity)
- Jarque-Bera (normality)
Full technical details in /docs/methodology.md.