With Greg Caldeira

Legislators often rely on cues from colleagues to inform their actions. Several studies identify the boardinghouse effect, cue-taking among U.S. legislators who lived together in the nineteenth century. Nevertheless, there remains reason for skepticism, as legislators likely selected residences for reasons including political similarity. We analyze U.S. House members’ residences from 1801 to 1861, decades more than previously studied, and show not only that legislators tended to live with similar colleagues but also that coresidents with divergent politics were more likely to move apart. Therefore, we deploy improved identification strategies. First, using weighting, we estimate that coresidence increased voting agreement, but at only half of previously reported levels. Consistent with theoretical expectations, we find larger effects for weaker ties and those involving new members. Second, we study legislators who died in office, estimating that deaths increased ideological distance between survivors and deceased coresidents.



Replication materials