Disclaimer. Don't rely on these old notes in lieu of reading the literature, but they can jog your memory. As a grad student long ago, my peers and I collaborated to write and exchange summaries of political science research. I posted them to a wiki-style website. "Wikisum" is now dead but archived here. I cannot vouch for these notes' accuracy, nor can I say who wrote them.
Alesina and Rosenthal. 1989. Partisan cycles in Congressional elections and the macroeconomy. APSR 83:373-398.
Voters use the midterm elections to moderate the president's excessive shift in economic policy. With a new president, the pendulum swings too far. With the midterm, we pull policy back towards moderation. Thus, this model contrasts with the retrospective voting models, in which voters reward the incumbent if the economy is doing well before the election. Instead, voters (in this model) vote based on rational expectations: how will policy shift if I vote for the opposition party?
Lots of it, but not fully satsifactory. If you throw in a couple of controls, their results fall apart.
The authors overlook an important coordination problem: If we all vote against the president's party, the pendulum swings too far again. How do we coordinate how many of us will vote against the president's party?
Research on similar subjects