Ideological Congruence: Illusion or Imperfection?

• Date: 01 February 2018
• Published date: 01 February 2018
• DOI: http://doi.org/10.1111/lsq.12177
• Author: G. Bingham Powell

G. BINGHAM POWELL

University of Rochester

Ideological Congruence: Illusion

or Imperfection?

Paul Warwick (2016) argues that much of the research on ideological congruence

leaves the erroneous impression that a close match of median left-right voter opinions

and government ideological positions usually emerges from elections. I propose further

clarifications. I offer a “natural metric” based on the average distances from the median

voter of the most distant and the closest parties competing in all these countries’ elec-

tions. I suggest that by these standards average ideological congruence in the Western

liberal democracies in the last 20 years has been fairly successful, but not as successful

as it could be.

Paul Warwick (2016) provocatively titles his recent article “The

Ideological Congruence Illusion.” This thoughtful and analytically pow-

erful piece of analysis, building as well on several f‌ine pieces coauthored

with Maria Zakharova (2013, 2014), argues against the view that “a close

matching of government policy positions with median left-right voter

opinion generally prevails in liberal democracies.” Rather, “valence”

(nonideological) issues lead voters frequently to desert the ideologically

closest political party and shift their support systematically to the left or

right. These shifts lead parliamentary governments also to form system-

atically to right or left of the median voter, a pattern Warwick calls

“bilateralist, rather than center-concentrated.”

Warwick also argues that the frequent use of the approach to esti-

mating the position of the median voter devised by HeeMin Kim and

Richard Fording (1998) is partly responsible for the “illusion” of ideo-

logical congruence. It has always been clear that this measure, which I

shall henceforth call the “KF” measure, as Warwick does, assumes that

all voters vote for the political party that they perceive as closest and then

derives the estimated median voter position from the party positions and

the aggregate party votes. Even if not all voters vote ideologically, as

long as some deviate to the right and an equal number to the left, the

measure will still identify the median voter position well. But if the vot-

ers systematically deviate in one direction, the KF measure will misplace

LEGISLATIVE STUDIES QUARTERLY, 43, 1, February 2018 21

DOI: 10.1111/lsq.12177

V

C2017 Washington University in St. Louis