Interpreting House Midterm Elections: Toward a Measurement of the In-Party's “Expected” Loss of Seats

Abstract

Midterm congressional elections have been subjected to relatively little analysis. This is surprising because these elections exhibit three quite striking features which, when taken together, call for further explanation. First, every midterm House election since the Civil War, with the exception of 1934, has brought a net loss of seats to the President's party. Second, in the large majority of elections the net loss has approximated the gross loss. The in-party (i.e., the President's party) seldom has captured seats from the other party to offset its own loss. And third, although the in-party's loss has been persistent, the number of seats lost has varied widely.Attempts to incorporate midterm elections into a broader interpretive framework of American election studies usually stress one of the first two features outlined above. The fact that only the in-party loses—and that its losses are mainly in marginal districts—has led commentators such as V. O. Key Jr. and the authors of The American Voter to interpret these midterm elections as part of the stable, long-term trends in voters' party allegiance.