Corporate Network Analysis

Corporate Network Analysis

Exploit slow information diffusion through supply chain relationships and director interlocks to identify predictable stock returns. Academic research shows customer momentum yields ~150bps monthly alpha; board centrality generates ~4.68% annual risk-adjusted returns.

Core Investment Theses

1. Supply Chain Momentum: When a customer gets a big order, upstream suppliers benefit with a lag (40% of shock diffuses over subsequent months)
2. Director Interlock Signals: Shared board members facilitate information flow, M&A activity, and strategic alignment
3. Cross-Holding Synergies: Executives/directors with positions in multiple companies may orchestrate beneficial strategic moves

Workflow

Phase 1: Identify Target Company Networks

1. Map supplier-customer relationships (>5% revenue materiality)
2. Identify board members and their other directorships
3. Track executive cross-company positions
4. Document insider holdings across connected firms

Phase 2: Data Collection

Use free data sources in this priority order:

| Relationship Type | Primary Source | How to Access | |-------------------|----------------|---------------| | Suppliers/Customers | SEC 10-K Item 1/7 | EDGAR full-text search | | Board Members | DEF 14A Proxy | EDGAR company filings | | Insider Transactions | Form 3/4/5 | EDGAR insider filings | | Beneficial Ownership | Schedule 13D/13G | EDGAR ownership filings | | Institutional Holdings | Form 13F | EDGAR quarterly filings |

See references/sec-data-guide.md for detailed extraction procedures.

Phase 3: Network Construction

Build network graphs connecting companies through:

• Customer-supplier links (weighted by revenue %)
• Shared directors (first-degree: direct; second-degree: shared third board)
• Executive overlaps
• Cross-insider holdings

Calculate centrality metrics:

• Degree: Number of direct connections
• Betweenness: Importance in information flow paths
• Closeness: Speed of information propagation

See references/network-metrics.md for calculation methods.

Phase 4: Signal Generation

Customer Momentum Signal:

1. Calculate weighted 12-month returns of major customers
2. Rank universe by customer momentum
3. Long top quintile, short bottom quintile
4. Optimal: 4th-layer relationships (less crowded)

Board Connection Signals:

• New shared director between firms → potential M&A signal
• Director clustered buying across positions → information signal
• High centrality + adverse event → watch for network contagion

See references/signal-generation.md for detailed procedures.

Phase 5: Due Diligence Checklist

Before acting on network signals, verify:

• [ ] Relationship confirmed via multiple sources
• [ ] Revenue materiality >5%
• [ ] Network position shows high centrality
• [ ] Relationship stable (duration >2 years)
• [ ] Information asymmetry present (limited analyst coverage)
• [ ] Sufficient liquidity for execution
• [ ] No MNPI exposure risk

Phase 6: Compliance Review

Permissible:

• All public SEC filings (10-K, DEF 14A, Form 4, 13D, 13F)
• Public earnings calls and press releases
• Mosaic theory: combining immaterial public info

Prohibited:

• Material non-public information from any source
• Shadow trading: using MNPI about Company A to trade Company B

See references/compliance-guide.md for legal boundaries.

Output Templates

When completing analysis, produce:

1. Network Map: Visual or tabular representation of company relationships
2. Signal Summary: Ranked opportunities with confidence scores
3. Due Diligence Notes: Verification status for each relationship
4. Risk Factors: Data quality issues, concentration risks, compliance flags

Key Academic References

• Cohen & Frazzini (2008): Customer-supplier momentum, 1.37-1.55% monthly alpha
• Ali & Hirshleifer (2020): Shared analyst coverage unifies spillover effects
• Larcker, So & Wang (2013): Board centrality predicts returns
• Cai & Sevilir (2012): Board connections in M&A transactions

See references/academic-papers.md for full bibliography.

Data Quality Considerations

• SEC disclosures only capture >10% revenue customers
• Companies may strategically withhold customer names
• Board data has survivorship bias (failed firms disappear)
• Post-publication alpha decay affects well-known signals
• Multi-source verification essential for high-confidence signals

Source & license

This open-source skill is cataloged on AgentStack and links to its original source — we do not rehost the code.

• Author: HHFinAi
• Source: HHFinAi/claude-equity-research-skills
• License: MIT

Install and usage instructions live in the source repository linked above.