Replication-Consistent Liquidity Forecasting for Derivatives: Forward Funding Sensitivities and a Liquidity Valuation Adjustment for Settlement Lags

Battle Count: 50

The paper is primarily relevant to risk management, treasury operations, and derivative valuation rather than direct quantitative trading strategies. However, it has indirect relevance for: (1) understanding funding costs embedded in derivative pricing, (2) liquidity-aware hedging strategies, (3) XVA-adjusted trade pricing, and (4) operational risk in trading desks. The funding sensitivity framework could inform liquidity-aware execution and position management. The LVA concept is relevant for desks managing settlement timing. Overall, it is more relevant to risk management and valuation than to alpha-generating trading strategies.

Implementation requires: (1) a multi-curve risk-neutral valuation model with funding and collateral curves, (2) American Monte Carlo simulation (e.g., Longstaff-Schwartz) for pathwise valuation, (3) stochastic adjoint algorithmic differentiation (AAD) for computing hedge ratios and their rebalancing increments, (4) scenario-wise extraction of funding positions and net cash increments, (5) settlement-lag modeling with appropriate discount factors, and (6) portfolio-level netting conventions. The AAD framework alone is demanding in computational resources. The paper provides implementation hints but no code or detailed algorithmic specifications.

The paper is a 34-page theoretical/conceptual note with no figures, no tables, and no code repository. It provides mathematical proofs, worked examples, and implementation hints (American Monte Carlo with AAD), but no comprehensive numerical analysis or reproducible computational experiments. The scope is explicitly stated as 'conceptual clarification and methodological guidance' with 'a comprehensive numerical analysis left for future work.' Reproducibility would require implementing the described Monte Carlo and AAD framework independently.