[EAI Opinion Review] Predicting the 2012 Presidential Election Through Changes in Regional Voting Behavior

Will Regional Voting Weaken? Cross-Pressures as a Variable

f2e34dd1a320d063

f2e34dd1a320d063

f2e34dd1a320d063

1. Regional Voting in the Democratic Era Since Democratization

During the democratization process, regionalism was a dominant factor in voter behavior. In the 13th presidential election, regional sentiments of Honam vs. non-Honam were confirmed in all regions except Seoul and Honam, resulting in a disadvantage for candidate Kim Dae-jung in most areas. However, through the 14th and 15th presidential elections, support increased in all regions except Yeongnam, indicating a weakening of the traditional Honam vs. non-Honam regional structure and a shift towards a partial regional structure between Honam and Yeongnam.

Since the 2000s, changes have begun to emerge even in this Honam vs. Yeongnam regional structure. In the 2002 16th presidential election, candidate Roh Moo-hyun, from the Yeongnam region, received overwhelming support in Honam, leading to the widespread diagnosis that 'regionalism-based voting behavior, which dominated Korean voter behavior during democratization, is weakening' (Kang Won-taek 2010; Lee Gap-yoon 2011; Lee Nae-young·Jeong Han-ul 2007). The fact that the Millennium Democratic Party, based in Honam, received around 10% support in Yeongnam, but achieved 20.0% in Daegu-Gyeongbuk and 29.1% in Busan-Gyeongnam, also supported the argument that regional voting behavior has weakened. In particular, subsequent general and local elections have provided increasing evidence of the weakening of regional voting behavior.

In the 17th general election in 2004, held after the impeachment, candidates from the current Saenuri Party's strongholds such as Daegu and Busan, running for the then United Democratic Party and progressive parties, increased in number and approached the winning margin. Cho Kye-tae, a candidate from the United Democratic Party, won in Saha-gu, Busan, and in the 18th general election, candidates Kwon Young-gil and Kang Ki-gap of the then Democratic Labor Party won in Gyeongnam. This proved that at least the Gyeongnam region is no longer a graveyard for the opposition. In the 2010 local elections, Kim Jung-gil, the Democratic Party candidate for Busan, received over 40% of the vote, and Kim Doo-kwan, considered a prominent pro-Roh figure despite running as an independent, was elected governor of Gyeongnam. Conversely, in Honam, although they did not win, Saenuri Party candidates performed well, especially in Jeonbuk, gaining double-digit support in general and local elections, underscoring the weakening of regionalism. Interest in generational voting and ideological voting is now filling that void.

However, there are significant counterarguments. Above all, recent presidential elections do not show any discernible weakening of regional voting. In particular, looking at the regional voting results, Honam still shows a strong tendency for regional voting. As shown in [Figure 1], even in the 16th presidential election, the support rate for candidate Lee Hoi-chang of the then Grand National Party in the Honam region was only 4.9%. Even in the 17th presidential election, where candidate Lee Myung-bak overwhelmingly defeated the opposition candidate, Lee Myung-bak's support rate in Honam was only 8.9%. Furthermore, if regional voting is defined not as support for a candidate from one's own region but as exclusive support for a regional party representing one's region, then Roh Moo-hyun's support in Honam is difficult to interpret as a weakening of regional voting, given that he was a candidate for the Millennium Democratic Party, which represented Honam, even though he was from the Yeongnam region. Conversely, Roh Moo-hyun's strong performance in Yeongnam is likely attributable to the 'hometown identity' of being from the region. Moreover, in the 16th presidential election, unlike previous elections, there were no strong third-party candidates like Chung Ju-yung or Lee In-je, making it essentially a 1:1 race between the ruling and opposition parties. Therefore, the increase in support in the Yeongnam region may also be influenced by the shift from a multi-candidate race to a two-party race. In the 17th presidential election, candidate Chung Dong-young of the United Democratic Party received an overwhelming 79.5% support in Honam, and the vote share gap between candidates Lee Myung-bak and Chung Dong-young in the Daegu-Gyeongbuk and Busan-Gyeongnam regions widened compared to the gap between Roh Moo-hyun and Lee Hoi-chang in the 16th presidential election.

[Figure 1] Regional Distribution (%) of Current Ruling and Opposition Party Candidates in Past Presidential Elections

Source: Historical Election Information, National Election Commission

In fact, when examining voter sentiment in hypothetical 1:1 matchups, regional differences are as pronounced as in past elections. The results of a panel survey conducted by the East Asia Institute, JoongAng Ilbo, SBS, and Korea Research immediately after the April general election, shown in [Figure 2], indicate that in a hypothetical contest between candidate Park Geun-hye and Ahn Cheol-soo, Park Geun-hye held a significant lead in Daegu-Gyeongbuk (68.4% vs. 26.3%) and Busan-Ulsan-Gyeongnam (58.1% vs. 39.1%), while Ahn Cheol-soo held a lead in Honam (79.6% vs. 17.9%). In Seoul, Incheon, and Gyeonggi, where no single party held a dominant advantage, Ahn Cheol-soo led in the metropolitan areas, while the two candidates were in a close contest in the Chungcheong region. Based on these regional support disparities, the view that regional voting factors will intensify in the 2012 presidential election is growing. Candidate Park Geun-hye is leading in the Yeongnam and Gangwon/Jeju regions, while candidate Ahn Cheol-soo is competing on equal footing by leveraging his support in Honam and the metropolitan areas. The undecided voters in the Chungcheong region are currently playing the role of kingmakers.

Notably, as shown in [Table 1], there has been little variation in voter turnout by region in past presidential elections. Up to the 16th presidential election, voter turnout in the Honam region was relatively higher than in other regions. However, in the 17th presidential election, voter turnout increased in the Daegu-Gyeongbuk region, albeit not significantly. In the July regular survey conducted by the East Asia Institute and Korea Research, the Daegu-Gyeongbuk region showed the highest intention to vote ('I will definitely vote'), but overall turnout intentions are evenly distributed. If this pattern continues, regional turnout itself is unlikely to be a significant variable. In this case, it seems difficult to predict the outcome of the election without a change in the current balance of power between the ruling and opposition parties.

[Figure 2] Park Geun-hye vs. Ahn Cheol-soo, Park Geun-hye vs. Moon Jae-in 1:1 Matchups

Source: East Asia Institute·SBS·JoongAng Ilbo·Korea Research Joint <2012 General Election Presidential Panel Survey> 2nd Survey (April 12-15, 1666 respondents)

[Table 1] Regional Voter Turnout Since Democratization and Voter Turnout Intention in the 18th Presidential Election (July)

Note: Historical turnout data from the National Election Commission; 18th presidential election voter turnout intention from East Asia Institute·Korea Research Joint Monthly Regular Survey (July 28, 800 respondents).

2. The Buffer Zone for Regional Voting: Swing Voters in the Metropolitan Area

Currently, regarding regional voting, arguments are being made to focus on the shifting sentiments in the traditional regional fault lines of Honam and Yeongnam, particularly in the Busan-Gyeongnam region where opposition candidate support is higher than in past elections, and on the Chungcheong region, which has played a casting vote role. This is likely because the metropolitan area is considered to be free from regionalism.

However, the regions that warrant the most attention in terms of regional voting tendencies, along with the Chungcheong region, are the swing voter areas in the metropolitan region. [Figure 3] shows the trend of regional changes in 1:1 hypothetical matchups from September of last year, when the Ahn Cheol-soo phenomenon began to emerge, to July 2012. In the Honam, TK, and PK regions, where regional voting tendencies are concentrated, except for the sharp rise in Ahn Cheol-soo's support from September until before this general election, candidate Park Geun-hye is generally leading significantly in Yeongnam, while Ahn Cheol-soo is leading in Honam.

Although candidate Park Geun-hye once received over 20% support in Honam, her support has stabilized in the mid-to-high teens as Honam's choice has tilted towards Ahn Cheol-soo. Conversely, in the TK region, Park Geun-hye has widened the support gap, recording support in the high 60% range, while Ahn Cheol-soo is in the high teens to low 20% range. In the PK region, despite significant fluctuations initially, Ahn Cheol-soo has now secured stable support of up to 35-40%, primarily from younger and progressive voters. Ahn Cheol-soo's strong performance in the PK region should now be understood as a constant rather than a variable.

In contrast, the swing voter regions of Seoul, Gyeonggi, and Chungcheong exhibit a fluctuating pattern depending on the verification process of candidate Park's leadership and the evolution of the 'Ahn wind.' Ultimately, the changes in the support rates of both candidates are influenced by the changes in support within these regions. Generally, the magnitude of these changes is within a 10 percentage point range, indicating a close contest, unlike the initial 'Ahn wind' which saw changes exceeding 15 percentage points.

Among these, the changes in the metropolitan area voters are particularly crucial. Although there are significant fluctuations in support rates in the Daejeon-Chungcheong region, considering the voter composition based on the 19th general election in [Table 2], it accounts for approximately 4.06 million voters, or 10.1% of the total electorate. The actual number of voters will inevitably be lower due to turnout, and the absolute number of votes resulting from support fluctuations within this group is small. However, the metropolitan area, with Seoul having 8.38 million voters and Gyeonggi/Incheon having 11.44 million voters, accounts for 49.4% of the total electorate, nearly half. Meanwhile, voters in Honam and Daegu-Gyeongbuk are similar in number to Chungcheong voters, at approximately 4.11 million and 4.15 million respectively, representing about 10% of the total electorate. However, the number of voters in Busan-Ulsan-Gyeongnam is 6.36 million, or about 15% of the total electorate. Combined with the Daegu-Gyeongbuk region, it represents about 25% of the total electorate, 2.5 times the number of voters in Honam. Ultimately, candidate Park and the Grand National Party are in a situation where they must compensate for their disadvantage in the metropolitan area through their lead in the Yeongnam region, while candidate Ahn must overcome his disadvantage in the Yeongnam region by widening his lead in the metropolitan area.

Assuming a voter turnout of approximately 75% in each region for this presidential election, similar to the 16th presidential election, the number of voters in the metropolitan area would be around 14.86 million. If, as seen in the recent changes in support rates between candidates Park and Ahn, one candidate gains 5 percentage points while the other loses 5 percentage points within a month or two, it translates to a shift of 1.48 million votes compared to before. Assuming 75% turnout among the 4.11 million voters in the Honam region, this amounts to 3.08 million voters. In other words, a ±5 percentage point change in support in the Seoul region can lead to a shift in voter sentiment equivalent to nearly half the voters in the Honam, Chungcheong, and Daegu-Gyeongbuk regions.

[Figure 3] Trend of Park Geun-hye vs. Ahn Cheol-soo 1:1 Support Rates

Source: East Asia Institute·Korea Research <Regular Survey>, 18th East Asia Institute·SBS·JoongAng Ilbo·Korea Research Joint <2012 General Election Presidential Panel Survey>

[Table 2] Number of Electorate and Voter Composition Ratio by Region in the 16th-18th Elections

Note: Reconstructed from historical election information from the National Election Commission. The voter numbers and composition ratios for the 18th general election may differ slightly from those of the 18th presidential election due to using the composition ratios from the 19th general election.

3. Factors Weakening Regional Voting in 2012: Cross-Pressures

Cross-pressures is a concept popularized by the Columbia school of election theory, pioneered by Lazarsfeld et al. (Lewis-Beck et al. 2008). When an individual belongs to multiple groups with different political characteristics and shares their identities, they may not exhibit a single political attitude, leading to psychological conflict and consequently preventing their political inclinations from being consistently expressed. For example, a person who is from Honam but belongs to a group with strong conservative ideological leanings is typically exposed to cross-pressures and is likely to exhibit attitudes different from those of people from Honam. This is useful for explaining the buffering factors of regional voting discussed in this study and warrants examination for predicting the 2012 presidential election. This paper examines how regional cross-pressures arising from the mismatch between residence and place of origin, as well as the cross-pressures between residence and ideological identity, and residence and generational identity, lead to specific voting tendencies among residents of the metropolitan area.

1) Regional Cross-Pressures: Defection of Honam/Yeongnam Natives Residing in the Metropolitan Area

Is the public sentiment in the metropolitan area unrelated to regional voting? Not necessarily. A significant portion of voters residing in the metropolitan area are from other regions, and many of them maintain their regional identity (hometown), which plays an important role in shaping their political attitudes. Recent studies suggest that the identity as a resident of the current region and the identity as a native of the place of origin are not always aligned, and that the identity of the region of residence has a stronger influence on voting behavior. However, a considerable number of individuals still maintain the political characteristics of their region of origin rather than those of their region of residence. Let's compare the voting tendencies of voters from Honam, Yeongnam, and Chungcheong, who constitute a large portion of the metropolitan area electorate, with residents of those respective regions.

First, for those from Honam, in the 16th presidential election, they overwhelmingly supported candidate Roh Moo-hyun regardless of whether they resided in Honam or the metropolitan area. However, after the 17th general election, a considerable difference emerged. Specifically, among those from Honam residing in Honam, 79.6% supported candidate Chung Dong-young and 13.8% supported candidate Lee Myung-bak, showing a strong regional voting tendency. In contrast, among Honam natives residing in the metropolitan area, 27.9% supported candidate Lee Myung-bak, and support for candidate Chung Dong-young was only around 50.0%. In the 18th presidential election, only 18.5% of residents in the Honam region stated they would vote for candidate Park Geun-hye, but 28.1% of Honam natives residing in the metropolitan area indicated they would support candidate Park Geun-hye, similar to the previous election.

Similarly, voters from the Yeongnam region show the same pattern. In the 16th presidential election, 61.3% of those residing in Yeongnam supported candidate Lee Hoi-chang, while only 48.5% of Yeongnam natives residing in the metropolitan area supported him. In the 17th presidential election, it was unusual that Yeongnam natives residing in the metropolitan area supported candidate Lee Myung-bak at a higher rate than Yeongnam natives residing in Yeongnam. However, in the 18th presidential election, 64.5% of Yeongnam residents stated they would vote for Park Geun-hye, and 32.5% for Ahn Cheol-soo. In contrast, among those residing in the metropolitan area, only 55.3% supported Park Geun-hye, and 40.7% supported Ahn Cheol-soo.

Of particular note are the voters from the Chungcheong region. As typical swing voters, residents of Chungcheong showed higher support for candidate Roh Moo-hyun in the 16th election and candidate Lee Myung-bak in the 17th election. In the 18th election, support for candidate Park Geun-hye was high at 53.4%. Residents of the metropolitan area also follow this pattern. The main difference is that Chungcheong natives residing in the metropolitan area tend to prefer ruling party candidates over opposition candidates compared to residents of Chungcheong. In the 16th election, support for candidate Lee Hoi-chang was 38.0%, higher than for residents of Chungcheong, while support for candidate Roh Moo-hyun was lower at 54.4%. In the 17th election as well, support for candidate Lee Myung-bak among Chungcheong natives residing in the metropolitan area was 62.8%, significantly higher than among residents of Chungcheong. Currently, support for candidate Park Geun-hye is 57.8%, 4.2 percentage points higher than among residents of Chungcheong, but support for Ahn Cheol-soo is only 35.3%, 6.7 percentage points lower. When considering strategies to expand support among metropolitan voters, political parties and candidates need to accurately understand these differences in characteristics between those based on place of origin and those residing in the metropolitan area.

[Table 3] Differences in Presidential Election Support (%) Based on Cross/Alignment of Hometown-Residence Identity

Source: 16th election data from the Korean Political Science Association (2003); 17th election joint survey by East Asia Institute·SBS·JoongAng Ilbo·Korea Research <2007 Presidential Panel Survey>; 18th election joint survey by East Asia Institute·SBS·JoongAng Ilbo·Korea Research <2012 General Election Presidential Panel Survey> 2nd Survey.

2) Cross-Pressures of Residence-Ideology Identity and Residence-Generational Identity

Another type of cross-pressure that can weaken regional voting behavior is the cross-pressure between residential identity and ideological identity. Voters with Honam origins and conservative leanings, and voters with Yeongnam origins and progressive leanings, are representative examples of individuals experiencing attitudinal conflict. In fact, among residents of Honam with progressive leanings, only 10.5% support candidate Park Geun-hye, while 30.8% of conservative residents of Honam defect. In the Yeongnam region, ideological cross-pressure is greater: 84.1% of conservative residents of Yeongnam support candidate Park Geun-hye, while only 38.3% of progressive residents of Yeongnam support Park Geun-hye, with 58.9% supporting Ahn Cheol-soo. This indicates that residents of Yeongnam completely deviate from regional voting tendencies based on their ideological leanings.

Looking at cross-pressures with generations, among residents of Yeongnam, 85.1% of those in their 50s and 60s support candidate Park Geun-hye, but this drops to 43.8% for those in their 20s and 30s, with support for Ahn Cheol-soo being higher at 53.7%. However, residents of Honam support candidate Ahn Cheol-soo over candidate Park Geun-hye regardless of age group (50s-60s vs. 20s-30s), indicating strong regional voting tendencies. Conversely, voters in the Yeongnam region deviate significantly from regional voting tendencies due to cross-pressures of ideology and generation. In Honam, while there were no generational differences, there was a relatively stronger tendency for conservatives to shift their support to candidate Park Geun-hye instead of Ahn Cheol-soo.

[Table 4] Cross-Pressures Between Residence Identity and Ideology, and Residence Identity and Generation

4. Conclusion: Regional Voting and the Ahn Cheol-soo Variable

The analysis above suggests that the metropolitan and Chungcheong regions will be key swing voter areas that will determine the outcome of the 2012 presidential election, and highlights the need to pay attention to factors that buffer traditional regional voting. Notably, there is a clear trend of deviating from regional voting, particularly in the Yeongnam region, due to cross-pressures of ideology and generation. This implies a message to focus on conflicting attitudinal factors between residence and origin, ideology and region, and generation and region, rather than relying on outdated regional mobilization strategies for victory. It calls for a direct confrontation by strengthening political activities targeting conflicted voters, moving beyond the old slogan of 'Are we not family?'

Regionalism in the Honam region is also entering a new phase, largely due to the Ahn Cheol-soo variable. As shown in [Figure 2], Honam now prefers external candidates over the Democratic Party candidates who have traditionally represented the region. In a hypothetical contest with Ahn Cheol-soo, candidate Park Geun-hye receives over 30% support in Honam, compared to remaining in the 10% range. This is a change comparable to Roh Moo-hyun receiving 30% support in the Busan-Gyeongnam region. While overwhelming support similar to past elections is likely if Ahn Cheol-soo runs, it demonstrates that the exclusive dominance of a specific party's candidate in Honam is already significantly shaken.

Although regional voting patterns remain strong, existing regionalism has conditions and buffering factors for its weakening. This study aimed to show that with bold efforts by parties and candidates to overcome regionalism, the possibility of an election free from regionalism is sufficiently present. The issue lies in the choice of the political establishment. Attention is focused on whether they will remain content with safe and easy regional mobilization or evolve towards establishing new political confrontation structures and competitive systems by further exploiting these weakening factors. ■