Divided governments are becoming the norm, rather than the exception in countries where the executive and the legislature are both directly elected. In the United States, for instance, the Republican Party appears to have a solid grip on the Presidency, while the Democratic Party appears to have an advantage in Congress. In this project the investigators explain this phenomenon as the result of a process where moderate voters bring about a balance of power that insures a moderate policy outcome. The investigators start from the presumption that the two political parties have different objectives. That is, when in office, the two parties follow different policies because they pursue the interests of different constituencies. However, the final policy outcome is not only a function of which party is holding the Presidency, but also of the composition of Congress. Moderate voters, that is voters who prefer policies in between the policies most preferred by the two parties, have two opportunities to"moderate" the President. In Presidential election years some voters may vote, Republican for President and then "balance" by voting for a Democratic candidate for Congress. This explains the common observation of split.ticket voting. Second, in mid.term Congressional elections some additional moderation can take place. In these elections there is no uncertainty about the identity of the President. This makes "moderation" easier relative to Presidential election years when there is uncertainty about the outcome of the Presidential race. Thus, some additional moderation can take place in mid-term. This explains a second widely observed phenomenon: the mid-term cycle. That is, the party holding the White House almost always loses votes and seats in mid-term elections relative to the preceding Congressional elections held in Presidential election years. The researchers have developed a model of this process and consider a specific policy issue: the inflation/unemployment tradeoff. The researchers generate several testable propositions of the correlations between economic conditions and electoral outcomes. Using aggregate voting data for Congressional elections for the period 1950-1984 the researchers assess the extent to which the model predicts aggregate Congressional election outcomes in terms of incumbency effects and state of the economy.
