Effects of Social and Ideological Cleavages
on Vote Choice in the Korean Presidential
Election of December 19, 2007*
Park, Chan Wook │Seoul National University
│ABSTRACT│
This paper investigates the effects of voters’socioeconomic characteristics and
ideological orientations on their candidate choice in the 2007 Korean presidential
election. The effect of a voter’
s residential region on candidate choice was demonstrated
by the regional cleavage of Honam versus the rest. Despite the lack of sharp generational
cleavage, older voters were more likely to support Lee Myung-bak. No discernible
pattern of class voting existed. However, a moderate degree of ideological cleavage in
electoral support was found. Progressives favored Chung Dong-young slightly more than
Lee, while centrists and conservatives overwhelmingly chose Lee over Chung. Voters’
ideological self-perception has substantive meaning in the sense that it has constrained
their stands on selected issues, such as the nation’
s most important problem and the
outgoing President Roh Moohyun’
s job performance. Voters who singled out economic
growth as the most important national objective indeed supported Lee much more
strongly than Chung. People who judged President Roh’s job performance more
negatively were more likely to vote for Lee of the main opposition party. Perceived
ideological distance from Lee monotonically decreased the likelihood of support for him.
With the regional, generational, and ideological cleavages becoming weaker than before,
Lee Myung-bak has launched his national leadership in a less divided political
environment. Still, depending on his leadership style and performance, the patterns of
social and ideological cleavages in electoral support will change. The Korean electorate
dynamically responds to what and how the government does.
* An earlier version of this paper was presented at the Annual Meeting of the Western Political
Science Association, San Diego, USA, March 22, 2008.
86 현대정치연구 | 2009년 봄호 (제2권 제1호)
I. Introduction
The main purpose of this paper is to investigate how voters’ socioeconomic
characteristics and ideological orientations have affected their candidate choice
in the 17th Korean presidential election held on December 19, 2007. Twelve
presidential candidates officially filed with the National Election Commission
on November 25 and 26, and later two of them withdrew from the race before
the polling day. In terms of the number of candidates, the election might have
appeared highly competitive. But Lee Myung-bak of the Grand National Party
(GNP), who was formerly Hyundai Construction CEO and also served as the
mayor of the capital Seoul, won a landslide victory. Lee actually kept himself
leading in the public opinion polls for about 15 months before the election day.
He received 48.7 percent of the vote, defeating Chung Dong-young of the
United New Democratic Party (UNDP) by a margin of 22.6 percentage points
(or 5,317,708 votes).1)
In the previous presidential election in 2002, the structure of competition
was totally different. Roh Moo-hyun of the New Millennium Democratic Party
waged a neck-and-neck election war against Lee Hoi-chang of the GNP, and
obtained 48.9 percent of the vote compared to Lee’s 46.6 percent. The winning
margin was much narrower: 2.3 percentage points (or 570,980 votes). In this
election, voters’ choice revealed serious social and ideological cleavages
prevailing at the time. As will be described later in this paper, electoral
alignment was clearly marked by the East-West regional divide, generational
gap, and progressive-conservative ideological split. President Roh Moo-hyun
1) About one and a half months before Election Day, Lee Hoi-chang who had run in the
previous presidential election under the GNP banner defected from the party and declared his
candidacy as an independent. This Lee came third with 15.1% of the vote.
Effects of Social and Ideological Cleavages on Vote Choice ~ 87
ascended to power in such heat of divided democracy.2)
To what extent has Lee Myung-bak’s overwhelming victory in 2007 lessened
the extent of regional, generational, and ideological cleavages in voters’
candidate support? How did these cleavage structures affect the vote choice in
the most recent Korean presidential election? This study seeks to answer these
questions using the data from a post-election survey and other supplementary
sources. First of all, the study briefly describes the conceptual framework for
analysis as well as the data. Second, it sketches and compares the voting results
of the 2002 and 2007 presidential elections with regards to the effects of social
and ideological cleavages. Third, it provides bi- and multivariate analyses to
show how voters’ socioeconomic characteristics and ideological orientations
have influenced their candidate choice in the 2007 presidential election.
Finally, the paper discusses the implications of the findings for cleavage
politics and the politics of integration on Korea’s path toward its democratic
consolidation.
II. Conceptual Framework and Data
The socioeconomic characteristic of an individual describes his or her
position in the society or a social network where the person interacts with
others. An individual’s ideological orientation is also defined in the social
context, which indicates the person’s belief and way of thinking about how the
society does and should work. Individual socioeconomic characteristics and
2) In the 2002 presidential election, Kwon Young-ghil of the Democratic Labor Party finished
far third with 3.9 percent of the vote.
88 현대정치연구 | 2009년 봄호 (제2권 제1호)
ideological orientations constitute the criteria or factors for cleavage patterns in
a society. People are divided along these attributes, and develop their group
identities or interests accordingly. These cleavages generate political conflict
when people in different groups are organized and mobilized in an attempt to
seek for their identities or interests. Such cleavage politics may bring a society
into ruin, and hence makes politics necessary as a means to moderate or bridge
the cleavages. In brief, cleavage and integration are two faces of politics.
In a democratic system, one can hardly think of cleavage politics and the
politics of integration without the working of political parties and elections.
Cleavage structures in a society shape the party system—political parties seek
electoral support based on social and ideological cleavages. Lipset and Rokkan
(1967) provided an in-depth sociological study of the consequences of
cleavage structures on voter alignment and the party system in modern western
societies. Particularly, in the voting behavior research, Lazarsfeld, Berelson,
and their associates (Lazarsfeld et al. 1944; Berelson et al. 1954) produced
pioneering results by focusing on social cleavage factors. They attempted to
explain voters’ behavior in terms of their socioeconomic status indicated by
education, income, and occupation; urban-rural residence type; and religion. In
western societies, studies on class voting are most prevalent and have given a
rise to much controversy on the continuing or declining effect of class
variables on vote choice (for example, Dalton 2002).
Existing studies on voting behavior in Korean presidential and National
Assembly elections since the country’s democratization have emphasized a
voter’s region of birth or residence as the most influential cleavage factor for
the vote choice (see Park 2000). A voter has region-based preference for a
candidate or a party. One supports the salient candidate who hails from the
voter’s native region (coinciding with one’s residential region for most of the
people living outside the Capital region), or whose party has a strong base in
the voter’s native region. Next to region, the age variable has recently become
Effects of Social and Ideological Cleavages on Vote Choice ~ 89
significant for explaining Korean voters’ choice behavior. One may find a lifecycle effect. As a voter grows older, he or she is more likely to vote for a
conservative candidate than a progressive one. When the electorate is grouped
into age cohorts, the younger is more likely to choose a progressive candidate
or party. Beyond a life-cycle effect, this may suggest a generation effect—the
members of one age cohort sharing common socialization experience during
their formative years have developed political attitudes similar among
themselves but different from those of another age cohort (Chung 1995; Kang
2003; Cho 2003).
In delving into the effects of ideological orientations, this study is eclectic in
drawing on both socio-psychological and economic approaches to the study of
voting behavior. The Michigan group of voting studies (Campbell et al. 1954;
Campbell et al. 1960) posited political attitudes as determinants of vote choice.
Among the attitudinal variables, the scholars of the school focused on party
identification, not much on ideology and issue positions. A typical American
voter in the 1950s and 1960s was depicted as an unsophisticated one with a
low level of ideological constraint on issue stands. However, later studies such
as The Changing American Voter (Nie, Verba, and Petrocik 1976) suggested
that the level of ideological thinking among the Americans rose during the
1970s. Notwithstanding the on-going controversy about the level of ideological
conceptualization among the American mass public, there is no denying that in
many advanced democracies voters’ liberal/conservative or left/right
ideological constraint on major issues provides an important clue for
explicating their voting decision.
In this study, voters’ ideological orientations mainly refer to their selfplacement and perception of candidates’positions on the progressive
/conservative continuum. The concept of ideological orientations also
encompasses voters’ issue positions constrained by and highly relevant to their
progressive/conservative self-identification.
90 현대정치연구 | 2009년 봄호 (제2권 제1호)
Downs (1957) opened a refined avenue for economic approach to the study
of voting decision based on voters’ ideological positions as well as those of
candidates or parties. Predicated on the axiom of voters’ rationality for
maximizing expected utility, he suggested a spatial theory of voting. According
to him, ideology is “a verbal image of the good society and of the chief means
of constructing such a society.” In a world of uncertainty, a voter uses
ideologies as a short cut for saving the cost of being informed on a wide range
of issues (96-100). From an economic perspective, the benefit of voting is the
actor’s expected utility differential between the candidates or parties (38-40).
The benefit is calculated based on the voter’s and candidate or party’s positions
on the ideological space. A Downsian rational voter chooses the candidate or
party nearest to him or her in terms of ideological proximity (e.g., Riker and
Ordeshook 1968; Enelow and Hinich 1984). In this study, this proposition is to
be empirically tested.
The main data for this study came from a post-election survey designed and
conducted by the East Asia Institute in collaboration with the Seoul
Broadcasting System, JoongAng Daily, and Hankook Research. The survey
was administered to a nationwide sample of eligible voters aged 19 or more
through the computer-aided telephone interviews on December 20 and 21,
2007. Telephone numbers and households were randomly selected, and within
the household the respondent was chosen so that the quota based on the region,
age, and sex could be fulfilled. The final size of a weighted sample is 2,111.
The sample error is ± 2.1% at the confidence level of 95 percent. The survey
includes questionnaire items necessary to investigate the effects of voters’
socioeconomic characteristics and ideological orientations on their presidential
candidate choice. In addition to this dataset, the study partly utilizes the official
election results publicized by the National Election Commission, and the data
collected from a couple of public opinion polls. The methods of statistical
analysis include simple descriptive analysis, bivariate correlation analysis,
Effects of Social and Ideological Cleavages on Vote Choice ~ 91
Ordinary Least Square regression analysis, and binary logistic regression
analysis. The manner of coding the variables considered in the analysis is
appropriately mentioned later in this paper.
III. Social and Ideological Cleavages Compared:
An Overview of 2002 and 2007 Presidential Elections
Between the 2002 and 2007 presidential elections, a sharp contrast is shown
in terms of regional, generational, and ideological cleavages. Let us first have a
look at voters’ regional alignment. Since the 1987 presidential election, voters’
region-based choice behavior and the consequential regional cleavage have
persisted in all of the subsequent presidential elections. Still, the results of the
2002 and 2007 presidential elections demonstrate significant differences in the
pattern and extent of regional cleavage. In this democratic era, most of the
major presidential candidates as well as all presidential winners have come
from the Southeast (or Yeongnam, including Daegu, North Gyeongsang,
Busan, Ulsan, and South Gyeongsang) and Southwest (or Honam, including
Gwangju, North Jeolla, and South Jeolla) regions. In this context, regional
voting particularly applies to the voters who are natives or residents of these
two regions.
Regional voting was conspicuous in the 2002 election. Lee Hoi-chang of the
GNP dominated in the Southeast region. As seen in
, he received
75.5 percent of the votes in Daegu and its adjacent North Gyeongsang, and
also 65.3 percent in Busan, Ulsan, and their surrounding South Gyeongsang.
Lee was not actually a native son of the region, but the region continued to be a
strong support base for the GNP whose predecessor parties had been led by the
92 현대정치연구 | 2009년 봄호 (제2권 제1호)
former Presidents Roh Tae-woo and Kim Young-sam. On the other hand, Roh
Moo-hyun of the NMDP got an overwhelming majority of the votes in the
Southwest region (95.2 percent in Gwangju, 91.6 percent in North Jeolla, and
93.4 percent in South Jeolla). Roh Moo-hyun was in fact a Southeast native,
but he ran as the candidate of the NMDP founded by the former President Kim
Dae-jung who had enlisted fervent support as a favorite son of the region since
the middle of the authoritarian era. In this election, Lee also won the most
votes in the Gangwon province. Meanwhile, Roh also finished first in the
Capital region (encompassing the capital city of Seoul, the metropolitan city of
Incheon, and their surrounding Gyeonggi province), in the Central region
(Daejon, North Chungcheong, and South Chungcheong), and in the Jeju island.
These election results clearly showed the East-West regional divide (see Figure
1; for detailed analysis, refer to Kang 2003, Kim 2003; Back et al. 2003).
In the 2007 election, Lee Myung-bak of the GNP garnered 48.7 percent of the
vote nationwide, and won an emphatic victory over the runner-up Chung Dongyoung who obtained 26.1 percent. As vividly demonstrated in , Lee
Myung-bak came up first in all the regions but the Southwest. His victory was
sweeping especially in the upper Southeast region, both Lee’s home ground
and the bastion of his party. He scored 71.1 percent of the votes in Daegu and
North Gyeongsang altogether, and 56.2 percent in Busan, Ulsan, and South
Gyeongsang. These percentage figures are somewhat smaller than 75.5 percent
and 65.3 percent recorded by Lee Hoi-chang of the GNP five years earlier in
the two parts of the Southeast region, respectively. But it should be noted that
Lee Hoi-chang declared his candidacy as an independent just one month before
the election day and made inroads into Lee Myung-bak’s support base. Chung
Dong-young of the UNDP only secured 6.4 percent in the upper Southeast and
13.0 percent in the lower Southeast. The UNDP was reorganized in August
2007 from the Open Uri [Our Open] Party, pro-President Roh Moo-hyun party,
which had split from the NMDP in November 2003. Chung, a native of the
Effects of Social and Ideological Cleavages on Vote Choice ~ 93
Official Results of 2002 and 2007 Korean Presidential Elections, by Region
(%)
Year
Candidate
2002
2007
Lee Hoi-chang
Roh Moo-hyun
Lee Myung-bak Chung Dong-young
Party
GNP
NMDP
GNP
UNDP
Nation
46.6
48.9
48.7
26.1
Seoul
45.0
51.3
53.2
24.5
Incheon/
Gyeonggi
44.3
50.5
51.4
23.6
41.3
52.5
37.1
22.6
Capital region
Central
Southwest
4.9
93.2
9.0
80.0
Upper Southeast
75.5
20.2
71.1
6.4
Lower Southeast
65.3
29.4
56.2
13.0
Gangwon
52.5
41.5
52.0
18.9
Jeju
39.9
56.1
38.7
32.7
Others
Note: The table shows the percentages received by the winner and the runner-up only, the
row total for one election being less than 100 percent; GNP = Grand National Party;
NMDP = New Millennium Democratic Party; UNDP = United New Democratic
Party; Central = Daejeon/Chungcheon; Southwest = Gwangju/Jeolla; Upper
Southeast = Daegu/North Gyeongsang; Lower Southeast = Busan/Ulsan/South
Gyeongsang
Source: National Election Commission; http://www.nec.go.kr
Southwest, received 80 percent of the votes in the whole region (79.8 percent
in Gwangju, 81.6 percent in North Jeolla, and 78.7 percent in South Jeolla).
These percent figures show vehement electoral support by the Southwestern
residents for their favorite son, but they are still short of the 93.2 percent
obtained by Roh Moo-hyun there in the 2002 election. In this region, Lee
Myung-bak of the GNP received 9.0 percent of the votes, somewhat greater
than the 3.6 percent obtained by Lee Hoi-chang, the GNP candidate in the
previous election. As far as the Southwest is concerned, the strength of
regional voting has weakened in the 2007 election.
94 현대정치연구 | 2009년 봄호 (제2권 제1호)
The pattern of regional cleavage revealed in the 2007 election is quite
different from that in its preceding 2002 election. The presidential winner Lee
Myung-bak came past the post first everywhere outside the Southwest, while
Chung Dong-young dominated only in the Southwest. In brief, the notorious
East-West regional divide has transformed into that of the Southwest versus the
rest. This way the regional cleavage has lessened in the 2007 election.
The generational cleavage or gap in voting drew much attention from the
analysts of the 2002 presidential election. Roh Moo-hyun was popular among
the young voters in their twenties and thirties. By a sharp contrast, Lee Hoichang was highly favored by the old voters in their fifties and over. The exit
poll conducted by the Media Research for the Korean Broadcasting System
showed that roughly six out of every ten voters in their twenties or thirties cast
their ballots for Roh, while about six out of every ten voters in their fifties and
older supported Lee (see
). The voters in their forties were more or
Top Vote-Getter by Metropolitan City and Province
Effects of Social and Ideological Cleavages on Vote Choice ~ 95
less evenly divided in their choice of Roh and Lee. Interestingly, the younger
half (ages 40 to 44) favored Roh more than Lee, while the older half (ages 45
to 49) supported Lee over Roh. Such stark generational differences in the vote
choice never happened before the 2002 presidential election. In this sense, this
presidential election could be dubbed “war of generation” (Song 2003).
To some extent, this cleavage may be explained based on the notion of lifecycle effect. Roh Moo-hyun had a comparative advantage in candidate image
so that he could be a favorite choice for the young generation less oriented
toward the establishment than the old generation. As of 2002, Lee Hoi-chang
was 66 years old, and Roh was 55. Lee came from an eminent family,
graduated from Seoul National University College of Law, served as a
Supreme Court judge and also later as a prime minister, and hence could be
perceived as a conservative champion of law and order. Moreover, Lee’s two
sons were exempted from mandatory military service for physical and health
reasons, which gave rise to scandal and made him a symbol of the privileged.
In the contrary, Roh was born into a poor farming family, received formal
education up to vocational high school only, passed the state bar exam to
become a judge, built his reputation as a righteous politician, and was a
progressive with populist appeal.
However, the generational differences in electoral support can be fully
understood by considering different socialization experiences among the age
cohorts and also significant events occurring in election periods. The voters in
their twenties at the time of the presidential election grew up in the atmosphere
of political democratization, economic prosperity, and social diversity. Their
sense of national pride and self-expression was strong. The cohort of ages 30 to
39 was represented by the “386 forces” which included those in their thirties
who attended college in the 1980s after being born in the 1960s. They engaged
in pro-democracy protests against the military authoritarian regime, and
espoused anti-U.S. and pro-North Korea nationalism. The voters in their forties
96 현대정치연구 | 2009년 봄호 (제2권 제1호)
were brought up under economically improving but still straitened conditions.
They had zeal for democracy, but lacked actual democratic experiences. The
voters in their fifties and over were socialized under the strong influence of
anti-communism, lived through economic hardship and social unrest, and thus
were highly likely to favor economic growth, national security, and social
order. All this suggested the generational cleavage in electoral support greatly
overlapped the ideological divide.
Besides different socialization experiences and value orientations, the events
and campaigns in the election year galvanized contrasting electoral support
between the young in their twenties and thirties, and the old in their fifties and
over. Both the World Cup soccer games and the accident of a U.S. military
vehicle in June 2002 deserve mention. The quadrennial World Cup games were
co-hosted by Japan and Korea in 2002. The Korean soccer team won a series
of games and finished fourth. The Korean “red devils” rooting enthusiastically
for their national team were predominantly in their twenties or thirties. Several
hundreds of thousands of them filled the City Hall Square and its adjacent high
streets in Seoul every time the home team took the field. This event boosted
national pride of the young generation. In the midst of the World Cup games,
two middle school girls were accidentally run over by an armored vehicle near
a U.S. military base. This accident attracted nationwide attention when the two
U.S. soldiers were acquitted at a U.S. court martial in November 2002, a
month before the polling day. The court decision enraged many ordinary
Koreans and provoked a sudden upsurge in anti-American sentiment. Mostly
young Koreans took part in countless candlelight vigils and violent
demonstrations in front of the U.S. Embassy. The Korea-U.S. and North-South
Korean relations had become major campaign issues. The conservative
candidate, Lee Hoi-chang, campaigned for stronger Korea-U.S. alliance and
hardliner policy toward North Korea going nuclear. On the other hand, Roh
Moo-hyun advocated continued engagement with North Korea to prevent
Effects of Social and Ideological Cleavages on Vote Choice ~ 97
heightened tension on the Korean peninsula. Roh expressed his objection to
Bush administration’s tough stance toward North Korea. In his campaign, Roh
put a stress on his determination to establish an equal relationship with the U.S.
(for further discussion, see Kang 2003, Song 2003, Cho 2003, Park 2005,
Chung 2002, and Jaung 2003).
The exit poll results in
hardly confirm the clear existence of
generational cleavage in the 2007 presidential election. Lee Myung-bak, the
conservative GNP candidate earned a greater rate of support among the old
generation aged 50 or more than among the young generation aged 39 or less.
About six out of every ten voters in the oldest cohort favored Lee, while some
four out of every ten in the youngest cohort did so. This may correspond with
the expectation based on a life-cycle effect. A more significant fact, however,
Presidential Vote by Age: Exit Polls in 2002 and 2007
(%)
Year
Candidate
2002
2007
Lee Hoi-chang
Roh Moo-hyun
Lee Myung-bak Chung Dong-young
Party
GNP
NMDP
GNP
UNDP
All
46.8
49.1
51.3
25.0
20-29/19-29 of age
31.7
62.1
42.5
20.7
30-39
33.9
59.3
40.4
28.3
40-49
48.7
47.4
50.6
27.1
58.5
23.5
58.8
24.8
(40-44)
(45-49)
50 or more/50-59
60 or more
(46.1)
(52.7)
58.3
(49.4)
(44.4)
39.8
Note: The table shows the percentages received by the winner and the runner-up only; and
the row total for one election is less than 100 percent; GNP =.Grand National Party,
NMDP = New Millennium Democratic Party, UNDP = United New Democratic
Party
Source: KBS-Media Research exit poll on December 19, 2002(N = 261,223); SBS-TNS
Media Korea exit poll on December 19, 2007 (N = 100,000 odd)
98 현대정치연구 | 2009년 봄호 (제2권 제1호)
is that Lee led Chung Dong-young of the progressive UNDP across all age
cohorts. In Chung’s case, the greatest level of support, which is three out of
every ten within the group at most, came from the voters in their thirties. The
lowest level of support for Chung was shown by those in their twenties: about
two supporters out of every ten voters. There is no linear tendency of Chung’s
electoral support changing with age. In short, the Korean electorate has
become far less divided along the generational lines in the 2007 presidential
election than in 2002.
Ideological cleavage needs to be discussed on its own, though it may be
closely interrelated with generational cleavage. In this study, a voter’s ideology
is his or her own perceived position on the progressive/conservative
continuum, measured by the relevant post-election survey item. Immediately
after the 2002 presidential election, the Korean Social Science Data Center
conducted a survey that contained a questionnaire item measuring the
respondent’s ideological self-placement based on a five-point scale. On this
scale, 7.1 percent of those respondents answered “(1) very progressive”; 34.0
percent, “(2) somewhat progressive”; 32.3 percent, “(3) middle of the road”;
23.5 percent, “(4) somewhat conservative”; and 3.2 percent “(5) very
conservative.” In the wake of a progressive candidate’s victory, progressive
identifiers outnumbered conservative identifiers, 41.1 percent versus 26.7
percent with 32.3 percent being in the middle.
Regarding the candidate choice item, the survey got 56.4 percent for Roh
Moo-hyun, roughly 7 percentage points higher than the vote rate received by
Roh, while it got 38.7 percent for Lee Hoi-chang, 8 percentage points lower
than Lee’s actual rate. Regardless of these discrepancies between real official
and survey results, a more important question is whether or not one can find a
lucid ideological cleavage, that is, a contrasted voting pattern in which a
majority of progressive voters are aligned with the progressive candidate and a
majority of conservatives come along with the conservative candidate. Among
Effects of Social and Ideological Cleavages on Vote Choice ~ 99
every ten progressive voters, seven tended to support Roh, and three did Lee.
In contrast, among every ten conservatives, six voted for Lee, and four did so
for Roh. Among the voters on the center, there were somewhat more
supporters for Roh, the winner than for Lee. The point is that clear evidence of
ideological cleavage could be seen in the 2002 presidential election.
Existing studies on voters’ ideology in the 2002 presidential election have
found that voters’ ideological self-placement was significantly associated with
their positions on an array of issues raised during the election. Economic
issues, including that of growth versus distribution, turned out to be less
correlated with the perceived ideology than were political or social issues, such
as North-South Korean relations, Korea-U.S. relations, national security law,
domestic order, and traditional values. Among these, voter’s ideological selfplacement showed the strongest constraint on his or her stance toward the U.S.:
Presidential Vote by Ideological Self-placement:
Post-Election Survey in 2002 and 2007
(%)
Year
Candidate
2002
(N)
Party
2007
Lee Hoichang
Roh Moohyun
(N)
Lee Myungbak
Chung Dongyoung
GNP
NMDP
GNP
UNDP
(1,314)
38.7
56.4
(1,931)
53.2
23.8
Progressive
(552)
20.8
71.6
(432)
34.3
37.3
Centrist
(403)
41.9
54.8
(657)
53.5
27.4
Conservative
(359)
62.7
34.8
(842)
62.6
14.0
All
Cramer’s V
.252***
.179***
Note: The table shows the percentages received by the winner and the runner-up only; and
the row total for one election is less than 100 percent; *** p< .01; GNP =.Grand
National Party, NMDP = New Millennium Democratic Party, UNDP = United New
Democratic Party
Source: The 2002 post-election survey conducted by the Korean Social Science Data
Center, and the 2007 post-election done by the East Asia Institute in collaboration
with the SBS, JoongAng Daily, and Hankook Research
100 현대정치연구 | 2009년 봄호 (제2권 제1호)
Progressives were more likely to be anti-American than were conservatives
(for detailed analysis, see Ka 2007; Kim et al. 2003; Park 2005; Back et al.
2003; Lee 2003).
In the 2007 post-election survey which provides the main data for this study,
the pattern of ideological distribution among the voters has gone reverse. A
respondent’s ideology was measured by an 11-point scale ranging from 0 (most
progressive) to 5 (middle of the road) to 10 (most conservative). When this
measure was condensed into a 3-point scale, the scores of 0 through 4 were
recoded into 1 (progressive); 5 into 2 (centrist); and 6 through 10 into
3(conservative). Conservative identifiers overwhelmed progressive identifiers
by 43.3 percent to 22.5 percent. And 34.2 percent located themselves on the
centrist position. A new political landscape under the conservative hegemony
has arisen along with Lee Myung-bak’s triumph.
Has the ideological cleavage vanished this time at all? The post-election
survey data in
indicates that a moderate degree of ideological
cleavage still exists. Among the progressives, 37.3 percent voted for Chung
Dong-young and 34.3 percent did for Lee Myung-bak. Chung led Lee by a
small margin of 3 percentage points in this group. On the other hand, Lee
dominated Chung among the centrists and conservatives as well. Within the
centrist group, Lee’s electoral support was twice as strong as Chung’s.
Furthermore, within the conservative electorate, Lee enlisted support four
times or so more than Chung did. The Cramer’s V statistic, an indicator of the
strength of relationship between voter ideology and candidate choice, is .252
for 2002 and .179 for 2007, both significant at the level of .01. One cannot
conclude that the meaningful ideological cleavage has not been played out in
2007. There still exists the ideological divide in electoral support, but it has
become perceptibly weaker than it was five years before.
Effects of Social and Ideological Cleavages on Vote Choice ~ 101
IV. Voters’ Candidate Choice in the 2007
Presidential Election
1. Voters’ Socioeconomic Characteristics and Candidate Choice:
Bivariate Analysis
In the post-election survey providing this study with the main data, 53.2
percent of the respondents who participated in the 2007 presidential election
said that they voted for Lee Myung-bak, and 23.8 percent for Chung Dongyoung. Compared with the actual official returns, the rate of support reported
by the survey was 4.5 percentage points over for Lee and it was 2.3 percentage
points below for Chung. Considering the sampling error, the winner was
overrated slightly, but the survey results did not seriously deviate from the
actual aggregate ones.
In examining the effects of voters’ socioeconomic characteristics and
ideological orientations on their candidate choice, the study looks first at the
simple relationship between the determinants and the vote choice. As
mentioned in the previous section, in the 2007 presidential election, electoral
alignment has not sharply taken a shape of the regional divide between the East
and the West. Still, there exists the divide between the Southwest and the rest
or particularly the upper Southeast region (see
). One needs to check
if the survey data will also bear this out.
shows that Lee enjoyed the
highest rate of support for him—75.4 percent among the residents of the upper
Southeast region, while he got the lowest 15.6 percent among the residents of
the Southwest. On the other hand, Chung Dong-young upended Lee only in the
Southwest, his party’s stronghold, but received the poorest 7.1 percent in the
upper Southeast. The survey data in
exactly agree with the
aggregate data in
, showing the divide between the Southwest and
102 현대정치연구 | 2009년 봄호 (제2권 제1호)
the rest.
The analysis based on the exit poll data in the previous section has shown
the lack of generational cleavage in the 2007 presidential election, remarkably
contrasting with that in 2002. The data in
again point to this. In
every age cohort, Lee came ahead of Chung in electoral support.
Even within the youngest cohort, including 19-year olds eligible for voting
since 2006, Lee’s support rate (46.1 percent) was more than twice as high as
Chung’s (21.6 percent). The age variable was recoded by taking into account
only the voters who had been eligible to vote in the 2002 presidential election.
Remember that in 2002 the two young cohorts (20 to 29 and 30 to 39 groups at
that time, that is, 25 to 34 and 35 to 44 groups five years later) predominantly
voted for Roh Moo-hyun, the progressive candidate of the MDP as opposed to
the older, especially in their fifties and more, who rallied behind the
conservative Lee Hoi-chang of the GNP. In the 2007 presidential election,
however, the young voters followed behind the old ones in choosing Lee well
above Chung. Most young voters have shifted their support from a major
progressive to conservative candidate in five years, probably because they have
become discontented with President Roh’s ill-performance, such as growing
youth unemployment and have developed pragmatism in their political
thinking. Generational cleavage has become inconspicuous in the most recent
presidential election.
Although clear generational cleavage lacked in the 2007 presidential
election, one cannot still totally deny the effect of voters’ age on candidate
choice. More than 40 percent but less than half of the voters below the age of
40 supported Lee Myung-bak. His support rate monotonically increased from
53.1 percent within the cohort of 40 to 49 years old, to 57.7 percent within the
50 to 59 cohort, and up to 70.4 percent within the cohort of 60 and over. In a
bivariate context, age is significantly related to the vote choice for Lee Myungbak.
Effects of Social and Ideological Cleavages on Vote Choice ~ 103
indicates no significant influence of sex on candidate choice,
though Lee Myung-bak received a higher percentage of support among female
voters than among their male counterparts. This study further considers voters’
education, income, and occupation as indicators of their socioeconomic status
or class. Lee Myung-bak led Chung Dong-young in any subgroup with regards
to these socioeconomic status variables, indicating no clear effect of class
Voters’Socioeconomic Characteristics and Candidate Choice, 2007
(%)
Variables (N)
All(1,930)
Lee Myung-bak Chung Dong-young Cramer’s V
53.2
23.8
Region of residence
.333***
Seoul (411)
58.4
21.4
Incheon/Gyeonggi (520)
54.2
19.8
Central (Daejeon/Chungcheon) (201)
48.3
16.9
Southwest (Gwangju/Jeolla) (205)
15.6
78.0
Upper Southeast (Daegu/N. Gyeongsang) (211) 75.4
7.1
Lower Southeast (Busan/Ulsan/
57.2
13.2
54.4
25.3
Male (950)
50.8
25.2
Female (980)
55.4
22.4
S.Gyeongsang) (304)
Others (Gangwon/Jeju) (79)
Sex
.046
Age (19 or older)
.160***
19-29 (371)
46.1
21.6
30-39 (441)
42.0
28.6
40-49 (450)
53.1
22.7
50-59 (307)
57.7
26.1
60 or more(361)
70.4
19.7
Age (25 or older)
.140***
25-34 (440)
47.7
23.2
35-44 (471)
44.2
28.2
45-54 (395)
55.2
24.1
55-64 (331)
65.3
22.1
65 or more (182)
68.7
19.8
104 현대정치연구 | 2009년 봄호 (제2권 제1호)
Variables (N)
All(1,930)
Lee Myung-bak Chung Dong-young Cramer’s V
53.2
23.8
Education
.077***
Middle school or less (467)
56.5
27.0
High school (717)
55.4
22.0
College or more (748)
48.8
23.5
Household income (monthly, in million won)
.073**
Less than one (209)
60.3
21.1
1-1.99 (297)
49.5
30.6
2-2.99 (456)
52.0
21.1
3-3.99 (395)
48.9
27.3
4 or more (528)
53.8
22.7
Occupation
.146***
Agriculture, forestry, and fishery (76)
60.5
26.3
Self-employed (396)
58.3
26.0
Blue-collar worker (228)
43.4
32.5
White-collar worker (460)
43.9
24.3
Homemaker (541)
58.6
21.3
Student (102)
45.1
16.7
Unemployed and others (125)
68.0
12.6
Note: The table shows the percentages received by the winner and the runner-up only, and
the row total for one election is less than 100 percent; **p<.05, ***p<.01
Source: The 2007 post-election conducted by the East Asia Institute in collaboration with
the SBS, JoongAng Daily, and Hankook Research
cleavage on voting. If the pattern of class voting existed among the Korean
electorate, the voters in high socioeconomic status would be expected to
support the conservative Lee Myung-bak more than those in low
socioeconomic status. Contrary to this class voting hypothesis, the rate of
support for Lee among the best educated (college or more) was the lowest. The
pattern of relationship between voters’ household income and their candidate
choice also betrays the hypothesis. The highest rate of support for Lee came
from the poorest income bracket. Some occupational groups were significantly
Effects of Social and Ideological Cleavages on Vote Choice ~ 105
more supportive of Lee than others. But this does not clearly agree with the
class voting hypothesis, either. Lee was more supported among the farmers,
self-employed, housewives, and unemployed than among the blue collars,
white collars, and students. By all standards, no discernible nationwide pattern
of class voting can be found in the Korean electorate.
2. Voters’ Ideological Perceptions and Candidate Choice:
Bivariate Analysis
The preceding analysis has suggested the significant effect of voters’
ideology, measured by their self-placement on the progressive/conservative
continuum, on their candidate choice. Now the study further analyzes how the
proximity or distance between a voter’s and the candidate’s perceived ideology
affects the vote choice. In so doing, the study first examines whether and to
what extent a voter’s ideological self-identification constrains his or her
positions on the issues raised during the campaign.
The post-election survey asked the respondents what they thought was the
nation’s most important task faced by the incoming administration. The two
most frequently mentioned responses were: first, moderating economic
polarization by 42.0 percent of the respondents and second, accelerating
economic growth by 35.8 percent (see
). The other problems, such as
quality of life, integration of the people, political reform, international
competitiveness, inter-Korean relations, or national security were mentioned
by at most 7 percent of the respondents. Based on these response categories,
eight dummy variables were created. Accelerating economic growth turned out
to be the single dummy showing a statistically significant correlation with the
voter’s self-placement on the ideological continuum. The more conservative a
voter is, the more likely he or she would single out economic growth as the
106 현대정치연구 | 2009년 봄호 (제2권 제1호)
Relationship of Voters’Ideological Self-placement with Issue Positions
(Pearson’s simple correlation coefficient)
Issues
Pearson’s r
The nation’s most important task for the next administration (N = 2,081: 100%)
a
Moderate economic polarization (42.0%)
-.03
Accelerate economic growth (35.8%)
.05**
Improve quality of life (6.9%)
-.03
Integrate the people (5.5%)
.01
Pursue political reform (3.5%)
-.01
Strengthen international competitiveness (2.2%)
-.01
Better North-South Korean relations (2.1%)
.00
Tighten national security (2.3%)
.04
Raise welfare spendingb (N = 2,080)
-.03
Oppose to foreign trade expansionb (N = 2,080)
-.10***
Promote university autonomyb (N = 1,881)
.19***
Negative retrospective evaluation of President Roh’s job performance (N = 2,109)
b
.20***
Note: The progressive/conservative continuum is an 11-point scale; a Dummies; b 4-point
scale; ** p<.05 *** p<.01
Source: The 2007 post-election conducted by the East Asia Institute in collaboration with
the SBS, JoongAng Daily, and Hankook Research
nation’s most important problem.
A voter’s ideological orientations broadly defined in this study encompass
not only the voter’s perceived ideological position but also his or her stands on
these ideologically relevant issues at the election time. The survey solicited the
respondents’ opinions on welfare spending, foreign trade, and university
autonomy from government regulation. The analysis has found that a more
conservative voter has significantly more favorable attitude toward the
expansion of foreign trade or university autonomy. In the survey, the
respondents were also asked to pass retrospective judgment on the job
performance of President Roh who is regarded as ideologically progressive.
The analysis has proved: the more conservative a voter is, the more likely he or
Effects of Social and Ideological Cleavages on Vote Choice ~ 107
she will become negative toward President Roh’s job performance. Voters’
ideological self-placement on the progressive/conservative continuum has
substantive meaning by constraining their attitudes toward selected issues, such
as foreign trade, university autonomy, and President Roh’s job performance.
displays how three kinds of voters’ ideological perception are
distributed along the progressive/conservative continuum. One curve with the
single, highest peak at the center shows the distribution of voters’ selfplacement on the ideological continuum. This curve is not exactly symmetric
in shape, but more symmetric than the other two curves. About one third (34.2
percent) of the respondents locate themselves at the center, 22.5 percent to the
left, and 43.3 percent to the right. The average score of voters’ own perceived
position is 5.5, meaning a largely centrist position leaning a bit toward the
conservative side. The other two curves show the distribution of voters’
perception about Lee Myung-bak’s and Chung Dong-young’s ideological
positions, respectively. Roughly speaking, Lee’s curve has a bimodal shape
having two peaks on the right. About one fifth (21.7 percent) of the
respondents place Lee at the exact center, and another fifth (19.6 percent) does
so at 7 or a moderately conservation position. The average of the distribution
concerning Lee’s ideological position is 6.2, a moderately conservative one.
The final curve portraying the voters’ perception of Chung’s ideological
position peaks at the center: 30.5 percent of the respondents put him in the
middle. This curve is denser on the left than on the right, and the average of
Chung’s perceived position is 4.5 or a moderately progressive one.
Let us examine the effect of a voter’s ideological proximity to Lee Myungbak or Chung Dong-young on the candidate choice. A voter’s comparative
ideological distance from the two candidates was computed as follows. First,
the absolute value of ideological distance was obtained between a voter’s selfidentified position and Chung’s position perceived by the voter. Second, the
absolute value of ideological distance was similarly obtained between the
108 현대정치연구 | 2009년 봄호 (제2권 제1호)
Voters’ Perception of Their Own and Candidates’ Ideological Positions
Note: The progressive/conservative continuum is an 11-point scale; the number of cases
ranges from 2,101 to 2,111
Source: The 2007 post-election conducted by the East Asia Institute in collaboration with
the Seoul Broadcasting Service, JoongAng Daily, and Hankook Research
voter’s self-identified position and Lee’s position perceived by the voter. Third,
the voter’s comparative ideological distance from Lee and Chung was
calculated by subtracting the value at the first step from that at the second one.
A smaller negative value of comparative ideological distance means a greater
degree of the voter’s ideological proximity to Lee than to Chung. The value of
zero tells that the voter is ideologically equidistant from these two candidates.
A larger positive value indicates a greater degree of the voter’s ideological
proximity to Chung than to Lee.
provides the data on voters’ comparative ideological distance
from the two major candidates and their candidate choice. The respondents
who voted for either Lee Myung-bak or Chung Dong-young numbered 1,472.
Effects of Social and Ideological Cleavages on Vote Choice ~ 109
Those who are ideologically closer to Lee than to Chung constitute 45.3
percent of the respondents and outnumber those closer to Chung than to Lee
(34.9 percent of the respondents). The remaining 19.8 percent are equidistant
from the two candidates. Given this distribution of voters’ comparative
ideological distances from the candidates only, Lee is expected to obtain more
support than Chung. In actuality, factors other than the ideological variables
must have worked for Lee’s victory over Chung as suggested by the fact that
Lee was favored instead of Chung in any of the five groups categorized based
on voters’ ideological proximity to Lee or Chung. At any rate, evidence
corroborates the effect of a voter’s ideological proximity to the candidate on
the vote choice. The voter’s increasing ideological proximity to Lee
monotonically strengthens the likelihood of voter’s support for Lee. On the
other hand, the voter’s ideological proximity to Chung is positively related to
the voter’s choice of Chung over Lee.
Voters’ Comparative Ideological Distance from Two Candidates
(Lee Myung-bak minus Chung Dong-young), and Candidate Choice
(%)
Candidate Choice
Voter’s Comparative Ideological Distance from Two
Candidates Lee Myung-bak minus Chung Dong-young (N) Lee Myung-bak
All (1,472)
Chung Dong-young
69.0
31.0
Much closer to Lee:
-10 ~ -6 (59)
93.2
6.8
Closer to Lee:
-5 ~ -1 (608)
80.8
19.2
Identical:
0 (292)
67.8
32.2
Closer to Chung:
1 ~ 5 (484)
53.1
46.9
Much closer to Chung:
6 ~ 10 (29)
51.7
Cramer’s V
48.3
.282***
Note: The row totals 100 percent; the progressive/conservative continuum is an 11-point
scale; *** p<.01
Source: The 2007 post-election conducted by the East Asia Institute in collaboration with
the SBS, JoongAng Daily, and Hankook Research
110 현대정치연구 | 2009년 봄호 (제2권 제1호)
3. Effects of Voters’ Socioeconomic Characteristics and
Ideological Orientations on Candidate Choice:
Multivariate Analysis
Now the analysis further proceeds to ascertain the effects of voters’
socioeconomic characteristics and ideological orientations on the vote choice
in the multivariate context where all the independent variables are
simultaneously taken into account.3)
As a preliminary step, an Ordinary Least Square regression model was
estimated. The results are displayed in
. The dependent variable,
candidate choice, was coded 0 for Chung Dong-young and 1 for Lee Myungbak. The model explains 37.4 percent of the variance in the dependent variable.
Among the socioeconomic characteristics, sex, age, region, education, and
some occupational categories have a statistically significant influence on the
vote choice at the .10 or lower level. In terms of the variables’ relative
importance indicated by their beta coefficients, a set of regional variables stand
out. In the 2007 presidential election, male voters supported Lee less than
female voters. An older voter more likely endorsed Lee at the polling booth.
Gangwon or Jeju being considered for reference, a voter’s residence in Seoul,
Inchon/Gyeonggi, Chungcheong, or Yeongnam (the Southeast) was positively
related to choosing Lee over Chung, whereas residence in Honam (the
Southwest) was negatively associated with such candidate choice. The better
educated were more likely to vote for Lee instead of Chung. Relative to the
unemployed, the voters engaged in self-employed, blue collar and white collar
jobs together with homemakers were less likely to support Lee.
3) Two ideological orientation variables, a voter’s respective position on the issues of
university autonomy and foreign trade, were omitted from the multivariate analysis to
prevent the increasing number of missing cases.
Effects of Social and Ideological Cleavages on Vote Choice ~ 111
Effects of Voters’ Socioeconomic Characteristics and Ideological
Orientations on Candidate Choice, Lee Myung-bak vs. Chung Dong-young:
OLS and Binary Logistic Regressions
Independent
Variables
OLS
Coefficient B (S.E.)
Beta
Binary Logistic
Coefficient B (S.E.)
(Constant)
-.166* (.101)
Sex (reference: female)
-.043* (.025)
-.046
-4.349*** (.816)
-.189 (.189)
Age (years old)
.002** (.001)
.073
.018** (.008)
Region of residence (reference: others)
Seoul
.215*** (.052)
.191
.450 (.351)
Incheon/Gyeonggi
.216*** (.051)
.207
.518 (.345)
Central
.213*** (.056)
.138
.689* (.393)
-.346*** (.056)
-.231
-2.371*** (.386)
Upper Southeast
.386*** (.056)
.258
1.868*** (.430)
Lower Southeast
.290*** (.053)
.231
.956** (.371)
.031* (.017)
.052
.275** (.131)
.002 (.009)
.005
-.020 (.067)
-.075 (.064)
-.030
-.744 (.497)
Self-employed
-.179*** (.044)
-.155
-1.043*** (.400)
Blue-collar worker
-.231*** (.047)
-.164
-1.410*** (.422)
Southwest
Education
a
Household incomeb
Occupation(reference: unemployed)
Agriculture, forestry, and fishery
(Lee – Chung)
Most important task: economic growth
Negative retrospective evaluation of President
Roh’sjob performance
Adjusted R Square
.374
Number of cases
1,485
1,456
Chi-square
608.744***
-2Log Likelihood
1,179.851
Cox & Snell R2
.346
112 현대정치연구 | 2009년 봄호 (제2권 제1호)
Independent
Variables
OLS
Coefficient B (S.E.)
Nagelkerke R2
Percentage Correct
Beta
Binary Logistic
Coefficient B (S.E.)
.486
82.5%
Note: S.E. = standard error; the progressive/conservative continuum is an 11-point scale;
Central = Daejeon/Chungcheong, Southwest = Gwangju/Jeolla, Upper Southeast =
Daegu/N. Gyeongsang, Lower Southeast = Busan/Ulsan/S. Gyeongsang; a 3-point
scale; b 4-point scale; * p<.10, ** p<.05, *** p<.01
Source: The 2007 post-election conducted by the East Asia Institute in collaboration with
the SBS, JoongAng Daily, and Hankook Research
All of the independent variables relevant to voters’ ideological orientations
exert a significant impact on candidate choice at the .01 level. Interestingly,
among the ideology variables, the most important variable is voters’ negative
retrospective evaluation of President Roh Moo-hyun’s job performance, as
indicated by its beta value. This has proven the widely accepted notion of the
2007 presidential election as a negative verdict by the people on President Roh
and the UNDP. The more conservative a voter’s self-perception was, the more
likely the voter would choose Lee Myung-bak. A voter perceiving a larger
distance from Lee than from Chung was less likely to support Lee. Support for
Lee was more likely to come from the voters who thought of economic growth
as the single most important task to be carried out by the incoming
administration than from the voters who did otherwise.
More importantly, binary logistic regression analysis was conducted where
the dependent variable was the probability of voting for Lee Myung-bak over
Chung Dong-young. The estimated model performs reasonably well, whose
goodness of empirical fit is indicated by the value of pseudo R2: Cox & Snell
R2 is .346 and Nagelkerke R2 is .486. Based on the model, 82.5% of the cases
under analysis are correctly predicted.
Among the variables which emerged statistically significant from the OLS
Effects of Social and Ideological Cleavages on Vote Choice ~ 113
regression analysis, some ceased to be so in the binary logistic regression
analysis. Such variables include sex, two region dummies (Seoul and
Incheon/Gyeonggi), and one occupational dummy (homemaker). Others
remain significant in their independent impact and also the same in its
direction.
Ceteris paribus, the older were more likely to vote for Lee Myung-bak.
Voters’ residence in Chungcheong and Yeongnam influenced the chances of
their choosing Lee over Chung Dong-young, but Honam residence worked the
other way around. In particular, the voters living in the northern part of
Yeongnam showed a very strong tendency of voting for Lee, and those in
Honam gave him a strikingly low level of support. Voters with higher level of
education had a greater probability of voting for Lee instead of Chung. In the
multivariate context, the effect of education on the vote has reversed in
direction as compared to the result of the bivariate analysis. The class voting
hypothesis holds in the multivariate context. However, when it comes to the
effect of occupation on the vote, self-employed people and those with blue or
white collar jobs were much less likely to vote for Lee than the unemployed. It
is suggested that voters with job security tended to be more lukewarm in
supporting Lee. This result is contradictory to the class voting hypothesis. But
it may tell that Lee’s campaign promise of economic growth and job creation
was less appealing to the voters in relatively secure jobs. Ideological
orientation variables altogether maintain their significant effects on vote choice
in the multivariate analysis. A more conservative voter showed a greater
likelihood of choosing Lee over Chung. A longer distance from Lee in the
voter’s ideological perception was a prohibitive factor for supporting him.
Voters who considered economic growth as the uttermost important goal for
the incoming administration were more likely to vote for Lee. Again, evidence
indicates that a voter’s negative assessment of President Roh’s job performance
was a significantly positive factor for increasing the probability of voting for
114 현대정치연구 | 2009년 봄호 (제2권 제1호)
Lee, the candidate of the main opposition GNP.
V. Conclusion
In the 2007 Korean presidential election, Lee Myung-bak of the GNP won a
landslide victory by garnering 48.7 percent of the vote. Lee Myung-bak led his
opponent Chung Dong-young in all the regions but Honam. Lee received the
highest rate of support in the northern part of Yeongnam, his native region and
the stronghold of his party. The East-West regional divide in the previous
presidential elections has transformed into the regional cleavage of Honam
versus the rest. This has been corroborated by a series of analyses using the
officially election returns and survey data.
This study hardly confirms the clear existence of generational cleavage in
voter alignment at the time of the 2007 presidential election. In any age cohort,
Lee Myung-bak was more supported than Chung Dong-young. The lack of
sharp generational cleavage notwithstanding, both the bi- and multivariate
analysis of the survey data show that age is still significantly related to the vote
choice of Lee Myung-bak—Older people were more likely to vote for him.
The study has sought to find the effect of class or socioeconomic status
cleavage on voting. The bivariate analysis has suggested that education has a
negative effect on the voter’s support for Lee when both its direct and indirect
effects are combined. On the other hand, the multivariate analysis shows that
other factors being controlled for, education directly exerts a pure positive
effect on choosing Lee in the presidential vote. In this multivariate context
only, the better educated were more likely to support the major conservative
candidate as suggested by the class voting hypothesis. Income suggests nothing
Effects of Social and Ideological Cleavages on Vote Choice ~ 115
significant about the class cleavage in electoral support. The analysis has
reached the conclusion that the self-employed and blue or white collar voters
were much less likely to vote for Lee than the unemployed. This may imply
that Lee’s emphasis on economic growth and job creation for his campaign
theme was less appealing to the voters with job security in a relative sense. But
this is not what the class voting hypothesis indicates. Overall, the study is led
to conclude that no nationwide discernible pattern of class voting could be
found in the 2007 presidential election.
For the 2007 presidential election, a moderate degree of ideological cleavage
in electoral support has been found. Ideologically progressive voters favored
Chung Dong-young who was seen by most people as moderately progressive,
slightly more than Lee Myung-bak, perceived by the bulk of the people as a
moderate conservative. Lee dominated Chung in receiving electoral support
among the centrist and conservative voters as well.
In this study, ideology was measured by the voters’ self-positioning or their
perception of candidates’ positions on the progressive/conservative continuum.
Voters’ ideological self-perception had substantive meaning in the sense that it
reasonably constrained their stands on selected issues during the 2007
presidential election, such as the nation’s most important problem, foreign
trade, university autonomy from the government, and the outgoing president’s
job performance. Voters who emphasized economic growth more than any
other national objective indeed supported Lee Myung-bak much more strongly
than Chung Dong-young. People who passed a more negative judgment on
President Roh’s job performance were more likely to vote for Lee who ran
under the opposition GNP banner. In explaining Korean voters’ candidate
choice, the study applied the Downsian notion of ideological proximity. In the
bi- and multivariate analysis as well, perceived ideological distance from Lee
Myung-bak monotonically decreased the likelihood of voter support for him.
The voter’s ideological proximity to a candidate was positively related to his or
116 현대정치연구 | 2009년 봄호 (제2권 제1호)
her choice of the candidate over the opponent.
Kim Dae-jung won the 1997 presidential election and succeeded in bringing
about the first-time peaceful transfer of presidential power from the ruling
party to the opposition since the country’s transition from authoritarianism in
June 1987. Lee Myung-bak’s victory in the 2007 presidential election has
marked the second inter-party transfer of presidential power by peaceful
means. Korea has now passed what Huntington calls the “two-turnover test”
(Huntington, 1991, p. 266), which signals the consolidation of Korean
democracy.
The results of the 2007 presidential election have provided Lee Myung-bak
with a good basis for launching his leadership on the political terrain much less
rough and divided than five years ago. He has set to engage in the politics of
integration with the regional, generational, and ideological cleavages having
become less salient and serious. But there is no guarantee that such favorable
conditions would last for long during the Lee Myung-bak administration. The
young generation, now seemingly more absorbed in jobs, livelihood matters,
real life goals, and pragmatic values than before, may give a vent to its pent-up
desire for change in no time. Ideological cleavage looks dormant. In the 2002
presidential election, progressive voters were triumphant. Five years
afterwards, the conservatives have hegemony over the progressives. Korean
voters’ ideological dispositions shift easily due to the influences arising from
turbulent events and persistent mobilization efforts. People dynamically
respond to what the government has done and how.
During the recent years under the progressive President Roh Moo-hyun’s
leadership, distribution-oriented rather than growth-oriented policies were
emphasized. Despite this, the country has faced an increasing degree of
economic polarization. Real household income is growing slowly, annually in
the neighborhood of two percent or so. The Gini coefficient, a measure of
income distribution, has recorded the highest value of .352 in 2007 since 2003.
Effects of Social and Ideological Cleavages on Vote Choice ~ 117
Over the same period, the income distribution ratio between households in the
top 20 percent and those in the bottom 20 percent has also increased to 7.66 in
2007 (“Countrys income disparity worsens,” Korea Herald, February 15,
2008). Under these conditions, why is one not able to find a nationwide pattern
of class voting in the 2007 presidential election? It was because low-income
voters supported the moderately conservative Lee Myung-bak more than any
progressive candidate. As a national leader, Lee has shouldered a burden of
achieving two often clashing goals almost simultaneously. He is a neoliberal
politician stressing economic growth based on market-oriented and businessfriendly policies. Under his policy scheme, economic growth will create more
jobs, which will in turn generate effects of easing economic polarization. If his
policy does not deliver much on the matter of economic polarization, it will be
raised as a major political issue and the class cleavage may emerge significant
on the political arena. The existence of any cleavage based on class or ideology
is not necessarily detrimental to democratic consolidation on Korean soil.
Korean democracy can advance if the disintegrating effects of the cleavage can
be addressed through rational discussion and institutionalized policy
competition.
118 현대정치연구 | 2009년 봄호 (제2권 제1호)
REFERENCES
Back, Jun-Kee, Jung-kwan Cho, and Sungdai Cho. 2003.“Ideology, Regionalism, and the
2002 Korean Presidential Election.”National Strategy. 9:4, 139-168 [in Korean].
Berelson, B. R., P. F. Lazarsfeld, and William N. McPhee. 1954. Voting. Chicago:
University of Chicago.
Campbell, Angus, Gerald Gurin, and Warren E. Miller. 1954. The Voter Decides.
Evanston: Row, Peterson.
Campbell, Angus, Philip E. Converse, Warren E. Miller, and Donald E. Stokes. 1960. The
American Voter. New York: Wiley.
Cho, Choongbin. 2003.“The 2002 Presidential Election and Political Generation.”
Presented at the spring meeting of the Korean Political Science Association [in
Korean].
Chung, Jin Min. 1995.“A Generational Analysis of the 1992 Presidential Election in
Korea.”Korean Social Science Journal, 22, 145-164.
Dalton, Russell J. 2002.“Political Cleavages, Issues, and Electoral Change.”In Lawrence
LeDuc, Richard G. Niemi and Pippa Norris, eds. Comparing Democracies II: New
Challenges in the Study of Elections and Voting. Sage Publications.
Downs, Anthony. 1957. An Economic Theory of Democracy. New York: Harper and Row.
Enelow, James M., and Melvin J. Hinich. 1984. Spatial Theory of Voting. Cambridge:
Cambridge University Press.
Fiorina, Morris P. 1981. Retrospective Voting in American National Elections. New Haven:
Yale University Press.
Huntington, Samuel P. 1991. The Third Wave: Democratization in the Late Twentieth
Century. Norman, OK: University of Oklahoma Press.
Juang, Hoon. 2003. President Roh Moo-hyun and the New Politics of South Korea. New
York: Asia Society.
Ka, Sangjoon. 2007.“The 2007 Presidential Election and Ideology.”The Journal of NGO
Studies, 5:1, 27-58 [in Korean].
Effects of Social and Ideological Cleavages on Vote Choice ~ 119
Kang, Won-Taek. 2003. Electoral Politics in South Korea: Ideology, Region, Generation
and Media. Seoul: Pureungil [in Korean].
Key, V. O. 1966. The Responsible Electorate. Cambridge: Belknap Press of Harvard
University Press.
Kim, Ju-chan, and Seongyi Yun. 2003.“The Influence of Ideology in the 2002 Presidential
Election.”21st Century Political Science Review, 13:2, 1-17 [in Korean].
Kim, Man-heum. 2003.“The 2002 Presidential Election and Regionalism.”In Kim Sekyun, ed. The Electoral Process and Significance of the 2002 Presidential Election.
Seoul National University Press [in Korean].
Kim, Wook. 2006.“Influences of Generation, Ideology, and Values in the 2002
Presidential Election.”In Soo Young Auh, ed. Korean Election V: The 2002
Presidential Election and the 2004 General Election for the National Assembly.
Seoul: Oreum [in Korean].
Knutsen, O. 1995.“Value Orientations, Political Conflicts, and Left-Right Identification: A
Comparative Study.”European Journal of Political Research, 28:1, 63-93.
Lazasfeld, Paul, Bernard Berelson, and Helen Gaudet. 1944. The People’s Choice. New
York: Columbia University Press.
LeDuc, Lawrence, Richard G. Niemi, and Pippa Norris, eds. 2002. Comparing
Democracies II: New Challenges in the Study of Elections and Voting . Sage
Publications.
Lee, Junhan. 2003.“Major Issues and Voter’s Choice in the 2002 Presidential Electin.”In
Kim Se-kyun, ed. The Electoral Process and Significance of the 2002 Presidential
Election. Seoul National University Press [in Korean].
Lipset, Seymour M., and Stein Rokkan, eds. 1967. Party Systems and Voter Alignment.
New York: Macmillan.
Nie, Norman, Sidney Verba, and John R. Petrocik. 1976. The Changing American Voter.
Cambridge: Harvard University Press.
Park, Chan Wook. 2000.“Elections in Democratizing Korea.”In Fuh-sheng and David
Newman, eds. How Asia Votes. New York: Chatham House.
. 2005.“Social Cleavage and Voters’Candidate Choice in Korea: An Analysis of the
120 현대정치연구 | 2009년 봄호 (제2권 제1호)
2002 Presidential Election.”In Yoshiaki Kobayashi and Hyug Bag Im, eds.
Political Process in Civil Society [in Japanese].
Park, Myung-Ho. 2007.“A Presidential Election and Generation: focusing the 386 Political
Generation.”The Korean Regional Politics Review, 1:2, 53-70 [in Korean].
Riker, William H., and Peter C. Ordeshook. 1968.“A Theory of the Calculus of Voting.”
American Political Science Review, 62: 1, 25-42.
Song, Ho-keun. 2003. What is Happening in Korea: Generation and Aesthetics of Its
Conflict and Harmony. Samsung Economic Institute [in Korean].
Walker, Scott, and Kyung-Tae Kang. 2004.“The Presidential Election in South Korea,
December 2002.”Electoral Studies, 23:4, 840-845.
Yoon, Jong-Bin. 2007.“The Metropolitan Voting Behavior in the 2007 Presidential
Election; Regionalism, Ideology and Candidate Factors.”Korean Party Studies
Review, 6:2, 65-95 [in Korean].
Effects of Social and Ideological Cleavages on Vote Choice ~ 121
논문요약
2007 한국 대선에서 사회균열이 유권자의
투표선택에 미친 영향
박찬욱│서울대학교
이 논문은 2007년 한국 대통령 선거에서 유권자의 사회경제적 특성과 이데올로
기적 정향이 후보선택에 미친 효과를 분석한다. 유권자의 거주지역이 미친 영향은
호남 대 비호남의 지역균열에 의하여 분명하게 나타났다. 선명한 세대균열이 존재
하지는 않았으나 나이가 많은 투표자일수록 이명박 후보를 지지하였다. 계급투표의
형태가 분명히 드러나지는 않았지만 선거지지에 있어서 이데올로기적 균열은 어느
정도 발견되었다. 진보성향 유권자들은 이명박 후보보다는 정동영 후보에 대하여
호의적이었으며, 중도성향과 보수성향 유권자들은 정 후보보다는 이 후보를 압도적
으로 선택하였다. 유권자가 자신의 이데올로기에 대하여 갖는 주관적 인식은 국가
가 당면한 가장 중요한 문제나 퇴임하게 될 노무현 대통령의 직무수행과 같은 특정
한 쟁점에 대하여 그 유권자의 입장을 제약하였다. 노 대통령의 직무수행을 부정적
으로 평가하는 사람일수록 제1야당의 이 후보에 표를 던질 가능성이 높았다. 유권
자와 이 후보 간의 이념적 거리는 그에 대한 지지를 단조적으로 낮추었다. 전반적으
로, 지역, 세대 및 이념 균열은 이전 선거보다는 약화되었다. 이 후보는 덜 분열적인
정치적 환경에서 그의 국가적 리더십을 행사하게 되었던 것이다. 이명박 대통령의
리더십 양식과 성과에 따라 선거 지지에 반영되는 사회균열의 형태는 변화할 것이
다. 한국의 유권자들은 정부가 무엇을 어떻게 하는가에 동태적으로 반응한다.
