Systemic Risk Contagion MCP Server

Financial system cascade failure simulation using DebtRank, Eisenberg-Noe clearing, multivariate Hawkes processes, and supra-adjacency tensor decomposition. This MCP server orchestrates 16 data sources spanning corporate registries, SEC filings, FDIC bank data, stock markets, cryptocurrency, macroeconomic indicators, and sanctions lists to construct 4-layer multiplex financial networks and run rigorous systemic risk analysis. Identifies G-SIBs, simulates cascade failures, detects stress clustering, and generates financial stability reports.

What data can you access?

Data Point	Source
Global corporate registrations and ownership	OpenCorporates
Legal Entity Identifiers and parent-child chains	GLEIF LEI
SEC annual/quarterly filings (10-K, 10-Q)	EDGAR Filings
SEC regulatory analysis and enforcement	SEC EDGAR
Executive insider stock transactions (Form 4)	SEC Insider Trading
US bank financial statements and condition data	FDIC Banks
Stock prices, quotes, and market data	Finnhub
US and global economic indicators	FRED
Employment and inflation statistics	BLS
IMF world economic data	IMF
Development and governance indicators	World Bank
Cryptocurrency prices and market capitalization	CoinGecko
Foreign exchange rates	Exchange Rates
Congressional stock trading disclosures	Congress Stock Trading
OFAC sanctions and blocked persons	OFAC
Consumer financial complaints	CFPB

MCP Tools

Tool	Price	Description
build_systemic_network	$0.04	Construct a 4-layer multiplex financial network from 16 data sources
compute_debtrank	$0.04	DebtRank (Battiston 2012): fraction of economic value affected by each node's distress
simulate_cascade_failure	$0.04	Eisenberg-Noe (2001) clearing payment cascade with fixed-point default propagation
identify_critical_nodes	$0.04	Supra-Laplacian spectral analysis for systemically critical entity identification
detect_stress_clustering	$0.04	Multivariate Hawkes process for stress event clustering and criticality detection
model_sanctions_shock	$0.04	Sanctions-induced shock propagation through counterparty exposure network
compute_contagion_channels	$0.04	Supra-adjacency tensor decomposition of cross-layer contagion pathways
generate_stability_report	$0.04	Comprehensive financial stability report with all analyses and recommendations

Data Sources

• OpenCorporates -- Corporate registrations and ownership structures for network construction
• GLEIF LEI -- Legal Entity Identifiers with parent-child ownership relationships
• EDGAR Filings -- SEC 10-K, 10-Q, and 8-K filings for financial exposure data
• SEC EDGAR -- Regulatory filings, enforcement actions, and analysis
• SEC Insider Trading -- Form 4 executive stock transactions for insider behavior signals
• FDIC Banks -- Bank financial statements, condition reports, and interbank exposure indicators
• Finnhub -- Stock market data for return correlation analysis
• FRED -- Federal Reserve economic data for macroeconomic context
• BLS -- Bureau of Labor Statistics employment and inflation data
• IMF -- World Economic Outlook data for global macro conditions
• World Bank -- Development and governance indicators
• CoinGecko -- Cryptocurrency prices and market data for crypto exposure layer
• Exchange Rates -- Foreign exchange rates for cross-border exposure analysis
• Congress Stock Trading -- Congressional stock trading disclosures for political risk layer
• OFAC -- Sanctions data for sanctions shock modeling
• CFPB -- Consumer complaints for retail financial stress signals

How the scoring works

The MCP constructs a 4-layer multiplex financial network and applies five analysis algorithms:

DebtRank (Battiston et al. 2012) measures the fraction of total economic value affected by each node's distress. Each entity is shocked individually, and stress propagates through weighted liability connections with no double-counting. Identifies G-SIBs (globally systemically important banks) and D-SIBs (domestically systemically important banks).

Eisenberg-Noe Clearing (2001) computes the fixed-point clearing vector via the Fictitious Default Algorithm. Each node pays min(obligations, available assets). An initial shock propagates through the network tracking defaults, cumulative losses, and cascade depth per round.

Supra-Laplacian Spectral Analysis builds the NL x NL supra-adjacency matrix (intra-layer blocks plus inter-layer coupling) and computes its largest eigenvalue via power iteration. If the eigenvalue exceeds the criticality threshold, the system is in supercritical regime where cross-layer cascades amplify.

Multivariate Hawkes Process estimates stress event intensity as self-exciting: past events increase future event probability. The branching ratio (spectral radius of alpha/beta) approaching 1 signals critical instability.

Supra-Adjacency Tensor Decomposition identifies dominant cross-layer contagion pathways. CP tensor rank estimation reveals which layer combinations (e.g., ownership to financial_exposure) are most dangerous for cascade amplification.

The four network layers are:

• Ownership -- Corporate registry and GLEIF parent-child relationships
• Financial Exposure -- SEC filings, FDIC interbank lending, derivative exposure
• Market Correlation -- Stock return correlations, crypto, and FX co-movement
• Supply Chain -- Sector co-occurrence and trade relationship data

How to connect this MCP server

Claude Desktop

Add to your claude_desktop_config.json:

{
  "mcpServers": {
    "systemic-risk-contagion": {
      "url": "https://systemic-risk-contagion-mcp.apify.actor/mcp"
    }
  }
}

Programmatic (HTTP)

curl -X POST https://systemic-risk-contagion-mcp.apify.actor/mcp \
  -H "Content-Type: application/json" \
  -H "Authorization: Bearer YOUR_APIFY_TOKEN" \
  -d '{"jsonrpc":"2.0","method":"tools/call","params":{"name":"compute_debtrank","arguments":{"query":"US banking sector"}},"id":1}'

This MCP also works with Cursor, Windsurf, Cline, and any other MCP-compatible client.

Use cases for systemic risk intelligence

G-SIB/D-SIB Identification

Run DebtRank analysis to identify which financial entities would cause the largest fraction of total economic value loss if they entered distress. Rank systemic importance quantitatively.

Stress Testing and Resolution Planning

Simulate cascade failures using Eisenberg-Noe clearing to estimate how an initial shock amplifies through counterparty networks. Evaluate resolution plan effectiveness by comparing cascade depth with and without intervention.

Criticality Regime Detection

Use supra-Laplacian spectral analysis to determine whether the financial system is in subcritical, critical, or supercritical regime. Supercritical regime means cross-layer cascades amplify rather than dampen.

Stress Event Clustering Analysis

Apply Hawkes process analysis to detect whether financial stress events are clustering and accelerating. A branching ratio approaching 1 provides early warning of systemic instability.

Sanctions Impact Stress Testing

Model how new sanctions would propagate through the financial network. Estimate direct exposure loss, indirect cascading impact, and propagation depth for each sanctioned entity.

Cross-Layer Contagion Channel Mapping

Identify which types of financial connections (ownership, exposure, correlation, supply chain) carry the most contagion risk. Tensor decomposition reveals the most dangerous cross-layer amplification pathways.

How much does it cost?

This MCP uses pay-per-event pricing. Each tool call costs $0.04.

The Apify Free plan includes $5 of monthly platform credits, which covers 125 tool calls.

Example Use	Approximate Cost
DebtRank analysis for a sector	$0.04
Cascade failure simulation	$0.04
Full stability report (all algorithms)	$0.04
Complete 8-tool analysis suite	$0.32

Note: Each tool runs all 16 actors in parallel, making the per-tool cost extremely efficient for the data volume and computational complexity involved.

How it works

1. You provide a financial entity, sector, or systemic risk topic
2. 16 Apify actors run in parallel fetching corporate registries, SEC filings, bank data, market data, macro indicators, crypto, FX, sanctions, and more
3. A 4-layer multiplex network is constructed with ownership, financial exposure, market correlation, and supply chain layers
4. The selected algorithm runs on the network -- DebtRank, Eisenberg-Noe, spectral analysis, Hawkes process, tensor decomposition, or comprehensive stability report
5. Structured results are returned with network statistics, algorithmic outputs, criticality assessments, and actionable signals

FAQ

Q: What is a multiplex network? A: A multiplex network has multiple layers of connections between the same set of nodes. In this case, financial entities are connected through ownership, financial exposure, market correlation, and supply chain channels simultaneously.

Q: How accurate are the cascade simulations? A: The models implement published mathematical frameworks (Battiston DebtRank, Eisenberg-Noe clearing). Results depend on the quality and completeness of publicly available exposure data. They should be interpreted as scenario analysis, not predictions.

Q: What is the branching ratio? A: In the Hawkes process, the branching ratio measures self-excitation intensity. Values near 0 indicate independent stress events; values approaching 1 indicate critical instability where each stress event triggers additional events.

Q: Does this include private exposure data? A: No. The network is constructed from publicly available data (SEC filings, FDIC reports, market data). Private bilateral exposure data is not available.

Q: Is it legal to use this data? A: All 16 data sources are publicly available government databases and open market data. See Apify's guide on web scraping legality.

Q: Can I combine this with other MCPs? A: Yes. Use alongside the Sovereign Debt Contagion MCP for sovereign-level stress analysis or the Sanctions Evasion Network MCP for sanctions compliance.

Related MCP servers

MCP Server	Description
ryanclinton/sovereign-debt-contagion-mcp	Sovereign fiscal stress and contagion modeling
ryanclinton/sanctions-evasion-network-mcp	Structural sanctions evasion detection
ryanclinton/investment-alternative-data-mcp	Alternative data for investment intelligence

Integrations

This MCP server is built on the Apify platform and supports:

• Apify API for programmatic systemic risk analysis pipelines
• Scheduled runs via Apify Scheduler for recurring stability monitoring
• Webhooks for triggering alerts when branching ratios or spectral radii exceed thresholds
• Integration with 200+ Apify actors for extending financial data coverage