Arguments for and against the existence of an American cultural divide are frequently placed in a religious context. This paper seeks to establish that, all politics aside, the American religious divide is real, that modern religious polarization is not a uniquely American phenomenon, and that religious divides can be understood as naturally emergent within the club theory of religion. Analysis of the 2005 Baylor Religion reveals a bimodal distribution of religious commitment in the US. International survey data reveals bimodal distributions in twenty-eight of thirty surveyed countries. The club theory of religion, when applied in a multi-agent model, generates bimodal distributions of religious commitment whose emergence correlates to substitutability of club goods for standard goods and the mean population wage rate. Ramifications of religious bimodality include potential instability of majority rule electoral outcomes. Median estimators, such as majority rule democracy, are non-robust with bimodal distributions. When religion is politically salient and polarized, small errors can disproportionately shift the election result from the preferences of the median voter.
