Reply to a comment on sequential elections and overlapping terms: voting for US senate

The phenomena of divided senate delegations, whereby several states are represented by senators from opposing political parties and differing political ideology, may at first blush seem at odds with the traditional median voter model. In a recent paper (Heckelman, 2000), I have modified the median voter framework by introducing two key institutional aspects, primary elections and staggered terms, which lead to predictions that the two parties alternate victories in successive elections, and the winning politicians hold extremely divergent positions. After a careful reading of the paper, Swank (2001) has called into question two of the underlying assumptions concerning voter motivation which drive the results. First, Swank believes voters in the primaries would not vote sincerely, rather they would behave strategically by supporting candidates most likely to capture the median voter in the general election instead of the candidate they most prefer themselves. It is then argued both parties would fully converge to the median. Second, Swank considers a different form of voter balancing than I utilized, but argues that even balancing within the state will result in full convergence among the two candidates in the general election (but not among the two winning senators). I consider each of these issues in turn and show the incorporation of strategic voting into