The Causes and Effects of Preferences for Party Government: A New Test of Policy Balancing*

Divided party control of government has become a defining feature of American politics over the last two decades. A number of hypotheses–both structural and electoral–have been offered to explain split party control, but the most provocative has been Fiorina’s (1988, 1992, 1996) argument that divided government results from purposive behavior by some portion of the electorate. He presents a spatial theory of policy balancing to show that some voters–principally moderates who see the parties as ideologically polarized–prefer the policies produced by a government where control of Congress and the Presidency is split between the two parties to those produced under the unified control of either the Democrats or the Republicans (see Ingberman and Villani 1993; Alesina and Rosenthal 1995; and Erikson 1988 for similar arguments). These voters act, perhaps unconsciously, on these preferences by splitting their presidential and congressional t ickets on election day. Though compelling logically, the policy balancing thesis receives scant support in most empirical analyses. However, these efforts generally offer only indirect tests of the policy balancing theory or make use of flawed direct measures of the preference for divided government. Recent studies (Lacy 1998; Lacy and Paolino 1998; Smith et al. 1999) develop better measures of preferences for party control, but they do not examine whether these preferences depend on voters ’ ideological locations an