Geopolitical Intelligence Through Causal Inference
Correlation is not causation. We use econometric methods to discover which geopolitical risk signals actually drive commodity and financial markets — and by how much.
Why This Matters
Geopolitical risk is the largest unpriced variable in commodity markets. Most tools measure sentiment. We measure causation.
Causal structure discovery — not just correlations, but directed relationships between risk signals and markets
Impulse response functions — quantify how shocks propagate across variables over time
Variance decomposition — attribute market movements to their geopolitical risk sources
How It Works
Monitor
Continuously ingest geopolitical news from real-time sources. Classify messages using a multi-category keyword taxonomy.
Analyze
Apply causal inference methods (PC Algorithm, Structural VAR) to discover which signals drive markets and quantify their impact.
Visualize
Track risk signals against market data in real time. See correlations, causal structure, and shock propagation at a glance.
Backed by Causal Inference
Unlike sentiment dashboards that report correlations, CausalAlpha uses the PC Algorithm to discover directed causal relationships between 5 normalized risk categories (Conflict, Political, Energy, Financial, Trade) and market variables (VIX, Brent Oil, Gold).
Price series are first-differenced for stationarity. Risk indicators are normalized as share of daily messages. The result: a DAG showing what actually drives what.
Explore the full methodology
Causal DAG — PC Algorithm (Fisher Z, alpha=0.10) | First-differenced prices