Marginal VaR Analysis — Agentic Doc-Driven Orchestrator

Marginal VaR decomposition calculates each child node's risk contribution to the parent VaR, computing relative shares and diversification metrics, then presenting the results with ASCII bar charts and donut charts for visual risk distribution.

Marginal VaR analysis supports FX, IR, EQ, CR, and CO risk factors using Historical Simulation VaR at 99% confidence with a 500-day lookback period to compute and visualize child node risk contributions.

Provide a parent node and an ISO date to start marginal VaR analysis, which queries VaR data for the parent and its immediate one-level children, then runs a self-check to ensure data completeness before computing risk contributions.

Marginal VaR analysis applies to a parent node with one-level immediate child nodes only, focusing on single-depth risk decomposition rather than multi-level portfolio hierarchy aggregation.

Visualizing risk contributions is best handled by generating ASCII bar charts and donut charts that illustrate each child's share of VaR and the overall risk distribution across FX, IR, EQ, CR, and CO risk factors.

Historical Simulation VaR limitations include reliance on a 500-day lookback period and 99% confidence level, restricting analysis to one-level child nodes and five specific risk factor types without multi-level hierarchy aggregation.