Abstract

Members of the U.S. Congress have become increasingly extreme and polarized over the past 50 years, but explanations for this phenomenon remain hotly debated. This paper tests whether and to what extent Congressional polarization can be explained by the increasing partisanship of constituencies that resulted from geographic polarization and nationalized voting. I estimate that approximately 40% of the polarization in the House and 70% in the Senate can be explained by the increasing partisan alignment of constituencies. Some of this trend is explained by incentives—members cast more extreme votes when their constituency becomes more partisan, but most is explained by selection—more partisan constituencies elect more extreme representatives. Additional analyses assess the mechanisms behind this trend. Partisan alignment is mostly explained by geographic polarization, but nationalized voting has also played an important role.