The Hidden Structure of Policy Networks
How lobbying coalitions self-organize — and what graph theory reveals about their durability.
Network analysis has long been applied to social systems, but rarely to the formal structures of policy influence. What happens when you map the full graph of organizational co-signatories across a decade of federal comment letters?
The Data
The Federal Register’s comment database contains millions of submissions, each tagged with submitting organization. Co-occurrence in the same proceeding creates a weighted edge. Run this across 10 years of EPA rulemaking and you get a graph of roughly 40,000 nodes.
What Emerges
The structure is not random. High-betweenness nodes — organizations that bridge otherwise disconnected clusters — turn out to be neither the largest trade associations nor the most politically active, but mid-sized technical consultancies with deep domain credibility.
This has real implications for how we understand durable policy coalitions. The most connected node is not always the most influential. The most critical node is often invisible until it disappears.
Implications
Policy influence maps built from declared positions systematically miss the connective tissue. The organizations that actually hold coalitions together rarely announce it.