Community Structure in Congressional Cosponsorship Networks

We study the United States Congress by constructing networks between Members of Congress based on the legislation that they cosponsor. Using the concept of modularity, we identify the community structure of Congressmen, as connected via sponsorship/cosponsorship of the same legislation, to investigate the collaborative communities of legislators in both chambers of Congress. This analysis yields an explicit and conceptually clear measure of political polarization, demonstrating a sharp increase in partisan polarization which preceded and then culminated in the 104th Congress (1995-1996), when Republicans took control of both chambers. Although polarization has since waned in the U.S. Senate, it remains at historically high levels in the House of Representatives.

💡 Research Summary

The paper applies network‑science methods to the United States Congress by constructing bipartite graphs that link legislators through the bills they co‑sponsor. Each member of the House or Senate is represented as a node, and an edge between two legislators is weighted by the number of bills they jointly sponsor. Using data from the Congressional Record spanning 1979‑2004, the authors build separate weighted, undirected networks for the House and the Senate for each two‑year Congress.
Community detection is performed with the Louvain algorithm, which seeks to maximize Newman‑Girvan modularity (Q). Modularity quantifies how strongly a network decomposes into dense sub‑communities relative to a random null model; higher Q indicates that intra‑community ties are much denser than inter‑community ties. By tracking Q over time, the authors obtain a single, interpretable metric of partisan polarization.
The results reveal a dramatic rise in modularity during the 104th Congress (1995‑1996). At that point the Republican Party seized control of both chambers, and the co‑sponsorship network reorganizes into two almost disjoint clusters—one dominated by Republicans, the other by Democrats. In the House, Q climbs above 0.6, signifying an exceptionally sharp partisan split. In the Senate, Q also increases but begins to decline in the early 2000s, suggesting a modest re‑integration of cross‑party collaboration, likely aided by longer terms and the committee‑centric structure of the upper chamber.
Further analysis of node centrality (betweenness, closeness, eigenvector) identifies key legislators—such as the Speaker, minority leaders, and senior committee chairs—who act as bridges between parties when polarization is low. As Q rises, the number and strength of these bridging ties drop sharply, confirming that the network’s structural cohesion erodes under intense partisan pressure.
The authors supplement quantitative findings with visualizations that show a transition from a tangled, multi‑party mesh in the early 1980s to a clean, bipartite coloration after the mid‑1990s. This visual shift mirrors the modularity trajectory, reinforcing the claim that the observed increase in Q reflects a genuine realignment rather than an artifact of the detection algorithm.
Methodological limitations are acknowledged. Co‑sponsorship captures only one dimension of legislative cooperation; other forms of interaction (e.g., committee work, informal negotiations) are omitted. The choice of edge weighting and the reliance on a single community‑detection heuristic may bias results, and alternative algorithms could yield different partitions. Nonetheless, the study demonstrates that modularity provides a robust, longitudinal indicator of political polarization.
In conclusion, the paper shows that congressional co‑sponsorship networks experienced a sharp surge in partisan modularity around the 104th Congress, marking the onset of the contemporary era of polarization. While the Senate’s modularity later receded, the House has remained at historically high levels, indicating persistent structural division. This network‑based approach offers a powerful complement to traditional political‑science analyses, enabling scholars to quantify and visualize the evolving architecture of legislative collaboration.