A New Kind of Balancing Act: Electoral Certainty and Ticket-Splitting in the 1996 and 2000 Elections

• Author: Kyle L. Saunders,Jonathan Williamson,Alan I. Abramowitz
• Published date: 01 March 2005
• DOI: 10.1177/106591290505800106
• Date: 01 March 2005
• Subject Matter: Articles

The results of the 2000, 2002, and 2004 elections

notwithstanding, over the last fifty years, divided

government has been the norm in the United States.

Only 13 of 30 elections since World War II have resulted in

unified control of the executive and legislative branches of

the federal government. In order to explain split party con-

trol, Fiorina (1988, 1992, and 1996) argues that divided

government is a result of purposive balancing behavior by a

portion of the electorate seeking moderate policy outputs

not possible under unified party control.1However, thus far

empirical support for his spatial theory has generally been

weak (e.g., Alvarez and Schousen 1993; Born 1994; Burden

and Kimball 1998, 2002).

Despite these results, we would argue that balancing

theory should not be dismissed. In line with Alesina and

Rosenthal (1995), we argue that balancing behavior (1) is an

option only for a subset of politically sophisticated voters

and (2) can be expressed as split ticket voting in presiden-

tial election years only under specific circumstances: when

voters are relatively certain prior to voting who will win the

presidential election. This article tests these hypotheses with

data from the 1996 and 2000 presidential elections, exam-

ples of elections with a fairly certain outcome and a very

uncertain outcome, respectively.

BALANCING THEORY

In its simplest form, balancing theory suggests that voters

compensate for reservations about the character or policy

positions of the expected presidential winner by voting for a

congressional candidate from the opposing party. As such,

balancing is a strategic action in which voters deliberately

seek to bring about divided government. One form of

balancing behavior is expressed as split ticket voting. Ticket-

splitters can be divided into two groups: those who split

their ticket in response to local factors such as the relative

visibility of the congressional candidates, and those who split

their ticket as a means of balancing partisan control.2

The idea that some voters split their ticket due to local

factors is not a new one. Declining party loyalties allow

short-term forces, the issues, candidates, and conditions

peculiar to a given election, to play a larger role in vote

choice (Nie, Verba, and Petrocik 1976; Wattenberg 1991).3

Some voters split their ticket as a consequence of their sat-

isfaction with their congressional incumbent. Constituency

service, the franking privilege, gerrymandering, and cam-

paign finance practices all provide incumbents advantages

in their reelection efforts. For some voters, the appeal of

incumbency is strong enough to overcome national political

69

A New Kind of Balancing Act:

Electoral Certainty and Ticket-Splitting

in the 1996 and 2000 Elections

KYLE L. SAUNDERS, COLORADO STATE UNIVERSITY

ALAN I. ABRAMOWITZ, EMORY UNIVERSITY

JONATHAN WILLIAMSON, LYCOMING COLLEGE

Using data from the 1996 and 2000 American National Election Studies, this article analyzes the behavior of

voters who split their tickets, voting for one party’s presidential candidate and the opposing party’s House can-

didate, in presidential election years. We test the hypotheses that balancing behavior is likely to occur only

when the outcome of the presidential election is relatively certain, and that balancing is most likely to occur

among relatively sophisticated voters who have reservations about the policy positions of their preferred pres-

idential candidate. The results of the study support the presence of this type of balancing behavior in the 1996

election and suggest that balancing can play an important role in producing divided government.

1Erikson (1988), Ingberman and Villani (1993), and Alesina and Rosen-

thal (1995) offer similar balancing perspectives to explain divided gov-

ernment and split ticket voting.

Political Research Quarterly, Vol. 58, No. 1 (March 2005): pp. 69-78

2In some circumstances, however, balancing strategy results in straight

ticket voting. For instance, if a voter’s preferred presidential candidate is

likely to lose the election, then the rational choice for the voter is to vote

a straight ticket with the goal of achieving divided government. It is

nearly impossible, therefore, empirically to discern a balancing strategy

from straight ticket behavior because of the myriad other factors leading

to this pattern of voting. Keeping this possibility in mind, we follow the

lead of other researchers by focusing on the balancing strategy of voters

that, in turn, results in split ticket voting.

3However, Kimball (2004) points out that the size of this group of influ-

ential moderate balancers appears to be shrinking and that ticket-split-

ting has been in decline since the 1980s. Kimball also argues the decline

is attributable to increased salience of party labels, which is in turn

caused by party polarization. In addition, Bartels (2000) demonstrates

that, compared with 25 years ago, a larger percentage of American voters

identify themselves as Republicans or Democrats and a larger percentage

of these partisans cast their ballots along party lines; therefore partisan-

ship is increasingly affecting the behavior of those who vote. Abramowitz

and Saunders (1998) explain these polarizing trends have occurred

because voters have been choosing their party identification more on the

basis of their policy preferences rather than maintaining the party alle-

giance that they inherited from their parents, which has resulted in an

ideological realignment of partisan identification.

PRQ_March05_II 3/24/05 9:18 AM Page 69