Hypothesis: Median voters prefer divided government. People distrust unified government and want ambition to check ambition. Is this always true (during wars and crises too)? When does it occur (only when opposing party holds presidency)? Is it just on the national level or is it true on state level too?
At the state level, balanced budget amendments check the power of unified government; there is less danger that one party will get carried away with spending or tax cuts. At the federal level, legislators aren’t constrained to balance the budget; there is more danger of one party getting carried away with spending or lowering taxes.
How often do voter order preferences for candidates when preferences in one race depend on who wins other races? How stable are preferences? Do moderate voters act as a ballast, shifting their allegiances so no party gets too powerful? It may seem irrational to occasionally vote against the party you prefer, but it’s a rational way to carry out a preference for divided government. We see evidence for divided government preference in midterm elections where the President’s party typically loses seats in Congress.
Most models of voter behavior, whether it’s a model of retrospective voting, prospective voting, partisan voting, etc., don’t capture the various ways people think about voting. General models must make simplifying assumptions, but it’s also important to consider their scope and applicability. All models of voting behavior probably exist in the voting population but how prevalent is each approach to voting? If we think, for example, 60% of the population will always vote by party identification, how does the rest of the populations, the other 40%, divide up among other approaches to vote choice? I don’t think preference for divided government is a predominate style of voting, but I suspect it characterizes median voters who want to see the pendulum of party control swing back and forth.