In Edward Tufte’s classic 1975 _APSR_ article “Determinants of the Outcomes of Midterm Congressional Elections” (“gated”:http://www.jstor.org/pss/1958391), he notes that:
bq. ..the translation of votes into seats has changed considerably over the period covered in our study: comparable shifts in the midterm partisan division of the vote are now worth less than half as much in terms of congressional seats compared to 35 years ago [i.e., 1940], thus significantly muting the impact of midterm elections on party alignment in the House.
Has anyone followed up on this line of research over the subsequent 35 years since Tufte published the piece? Has the rate at which vote shifts in midterm elections translate into seat shifts continue to decline since 1975? Or has it stabilized? Or reversed course? It’s a kind of an interesting question in its own right, but I imagine this would also be germane to our “ongoing discussion of forecasting models”:https://themonkeycage.org/2010/09/doug_hibbs_on_the_fundamentals.html.