Party polarization is endogenous
Centrist wonks lament party polarization, but rarely point out that it’s not something that just happened. In the context of heterogeneous political geography and malleable district boundaries, a two-party system doesn’t yield the centrism it is often credited with, but a superficial and artificial polarization that demands an eventual populist response. Party polarization is the endogenous and predictable result of incentives created by a first-past-the-post voting system susceptible to gerrymandering.
For gerrymandering to work, the voting behavior of citizens must be identifiable and stable. The way gerrymandering works is to decrease the number of votes that are “wasted” in safe localities by redistricting a party’s voters into less safe or marginally unfavorable localities where they might tilt the balance. The expected outcome is more, closer, wins for the gerrymandering party. However, those closer wins are necessarily more fragile. Even small shifts in party identification or voting behavior, any deviation from the roles assigned to each of us by the gerrymanderers, can undo all their good work. Worse still, deviations are rarely idiosyncratic to a single district. Gerrymandering leaves a party more vulnerable to systematic deviations, which make “wave elections”. Thus gerrymandering creates incentives for incumbents and party operatives to try to cement party identification in stone. Candidates in both parties emphasize issues on which their members are unusually unified and the other party’s members are unusually opposed over issues that might be more important to their constituents, but less clearly drawn across party lines. They then work very hard to make positions on these issues, and only these issues, tribal markers, essential to voters’ identities. We end up in a world with extreme polarization of party platform and identification, despite not so much actual polarization among voters when a broader spectrum of issues than those emphasized by the parties is considered.
A puzzle in this account is why the party disadvantaged on net by gerrymandering should play along. The answer to this is simple. Both parties gerrymander when they get to draw the lines, although the Republicans gerrymandered unusually aggressively and with unusual success in 2010. Further, even absent gerrymandering, both parties “naturally” have districts where they have some advantage, but where elections remain close and competitive. The safest strategy for an incumbent member of Congress in a closely divided district is to cement in place the coalition that brung her the last time, rather than roll the dice on poaching her opponent’s voters (which may open up her own voters to poaching). Incumbent members of Congress in both parties tend to prefer a stably polarized electorate to one in which common ground increases the uncertainty surrounding voter behavior. Even when this disadvantages the national party, in the social science cliché, it’s a case of concentrated benefits and widely dispersed costs. Post 2010, nationally, the Democrats might have been better off de-emphasizing party-polarized issues and embracing causes with cross-party popular appeal. But neither members of Congress individually, nor party activists who’ve made their careers and raised their funds as fighters across well-worn political battle lines, would assent to that strategy. Further, issues with broad popular appeal that have been neglected by both political parties are often issues to which both parties’ “donor classes” are allergic.