Weekend price discontinuities in cryptocurrency markets arise because liquidity and trading behavior change across the calendar. Exchange maintenance, fewer professional market makers, and concentrated retail activity during leisure hours combine with news arriving asynchronously to create gaps between Friday and Monday prices. These gaps matter for risk management because models that ignore discontinuities will understate extreme moves and misprice options and margin requirements.
Models that handle weekend gaps
Classic ARCH and GARCH frameworks capture volatility clustering and conditional heteroskedasticity and remain a baseline for forecasting. Robert Engle New York University introduced ARCH and Tim Bollerslev Duke University extended it to GARCH, both providing solid foundations for modeling time-varying variance. However, pure GARCH often misses abrupt jumps that characterize weekend gaps. Models that explicitly include jump components or use stochastic volatility with jumps better represent sudden discontinuities. Research on high-frequency and realized measures by Torben G. Andersen Northwestern University and collaborators shows that using intraday data to construct realized volatility improves forecasts of future variance, because realized measures can separate continuous volatility from jumps. Combining realized measures with conditional variance models, or employing stochastic volatility models augmented with jumps, tends to outperform single-equation GARCH in the presence of weekend discontinuities. No single model is universally best; performance depends on data quality and the prediction horizon.
Relevance, causes, consequences and nuances
Practical implications include under-estimated tail risk, inappropriate hedge sizing, and stress on liquidity during early-week trading. Human and cultural patterns influence weekend behavior: markets with predominantly Western traders show different weekend volume profiles than regional exchanges where weekends differ, and regulatory or operational practices in certain jurisdictions can restrict withdrawals or order types over local holidays, amplifying gaps. Environmentally, automated trading and 24/7 infrastructure have reduced some barriers but concentrated centralization on a few major exchanges still creates territorial vulnerabilities. For practitioners, combining realized volatility estimators with jump-aware stochastic models and using heavy-tailed error distributions improves robustness to weekend gaps. Calibrating models on exchange-specific data and including time-of-week adjustments captures local market microstructure and reduces forecast bias. Model choice should be guided by empirical validation on the specific exchange and instruments being traded.