[22nd General Election Research Series] Asset Voting in the 2024 General Election: Focusing on Metropolitan Area Voters

1. Introduction

The 2024 general election, held in the third year of the Yoon Suk-yeol administration, concluded with a landslide victory for the opposition Democratic Party of Korea. However, despite the election results ending in a victory for the 'regime critique' narrative, a notable phenomenon was observed in some areas of the metropolitan region: 'real estate' continued to be a significant variable influencing voters' choices.

If so, was the 'real estate class voting' phenomenon, which became clearly evident starting with the 2021 Seoul mayoral by-election, merely a temporary expression of dissatisfaction with the Moon Jae-in government's real estate policies? A post-election analysis by the Kyunghyang Shinmun, correlating apartment transaction prices in Seoul with the vote share difference between the two major parties in the recent election, reported a correlation coefficient of 0.76, a slight increase from 0.74 during the last presidential election (Kyunghyang Shinmun, 4/15/24). Specifically, in the 'Han River Belt,' among the nine constituencies, 'real estate voting' was evident in Dongjak-B, Mapo-A, and Yongsan-gu, where the People Power Party won, and in Yeongdeungpo-B and Yangcheon-A, which were close races. In Dongjak-B, the People Power Party garnered higher support in Heukseok-dong, Sangdo 1-dong, and Sadang 1-dong, where the average transaction price per 3.3㎡ is the highest. In Mapo-A, Ahyeon-dong and Yonggang-dong, where large apartment complexes are located, supported the People Power Party, and in Yongsan-gu, support for the People Power Party was strong in areas along the Han River such as Hannam-dong, Ichon-dong, and Seobinggo-dong. This implies that the higher the housing price in a neighborhood (dong), the higher the vote share for the People Power Party.

This phenomenon aligns with previous research that found a high correlation between 'housing prices' in the Seoul metropolitan area and party support in three previous elections: the 2018 local elections, the 2020 general election, and the April 7, 2021 Seoul mayoral by-election (Kim Soo-in, Kang Won-taek 2022). In the metropolitan area, neighborhoods with higher average apartment sale prices per square meter showed a higher probability of voting for the conservative party, a trend that was confirmed again in this election. While academic discussions continue regarding whether the voting behavior based on individuals' economic situations, observed recently, is a phenomenon related to social cleavages or a retrospective evaluation of real estate policies, correlation analyses between average real estate transaction prices by polling station or neighborhood and party vote share only show the proportion of parties voted for based on 'housing prices' at an aggregate level and have limitations in understanding voters' intentions at an individual level. Therefore, this study aims to examine the influence of assets, represented by 'real estate,' on the voting choices of metropolitan area voters in the 22nd general election.

2. The 2024 General Election and the Metropolitan Area[1]

In the 2024 general election, there was significant interest in the voting choices of voters in the Seoul metropolitan area. This is not new, given that more than half of South Korea's total population resides there and voters in this region are relatively free from regionalism.[2]

However, the metropolitan area attracted particular attention in this general election because many of the 'swing vote' regions, where election outcomes have been reversed in recent elections, are located there. Voters in the metropolitan area voted for the Democratic Party of Korea in the 2020 general election, and then for the People Power Party in the presidential and local elections held in March and June 2022, reversing their voting choices within two years. Leading up to the general election, domestic media reported that the sentiment of metropolitan area voters would determine the outcome of the election (Jungbu Ilbo 4/7/24; Financial News 4/8/24). Figure 1 is a word cloud created using the BigKinds news analysis system from the Korea Press Foundation, which visualizes frequently occurring words from articles related to the general election published in the month leading up to the election. The analysis of related words based on the search term ('general election') and their frequency and weight shows the order of related words as 'Democratic Party' – 'Han Dong-hoon' – 'Metropolitan Area' – 'Yoon Suk-yeol' – 'People Power Party.' Excluding major party and personal names, 'Metropolitan Area' was the word that appeared most frequently in media reports in the context of this general election. Media articles reflect the major issues of a given period and, especially when a 'decisive event' occurs, a flood of diverse articles are published, thus well illustrating the salient issues at the time.

Figure 1. Word Cloud – Search Term “General Election”

Within the metropolitan area, the 'Han River Belt,' comprising nine constituencies adjacent to the Han River in Seoul (including Mapo, Yongsan, Seongdong, Dongjak, Gwangjin, etc.), was identified as the most fiercely contested battleground, with considerable media coverage dedicated to opinion polls targeting voters in this region. Figure 2 is a graph illustrating the number of articles containing the keyword 'Han River Belt,' using BigKinds. It shows an increase in the number of articles with 'Han River Belt' in their titles as the election date of April 10th approached.

Figure 2. Keyword Trend Analysis for “Han River Belt” (Search Period: January 1, 2024 – April 10, 2024)

Meanwhile, as numerous opinion polls showed close races within the margin of error, each party presented various pledges to win over voters in this region. Notable examples include the People Power Party's pledge to connect the GTX-C line to Wangsimni Station and relocate the National Assembly to Sejong, and the Democratic Party of Korea's pledge to 'subterraneanize the entire Olympic Expressway in Seoul.' As the election day neared, campaign efforts also concentrated on the metropolitan area. During the official campaign period, Han Dong-hoon, interim leader of the People Power Party, visited metropolitan area regions more than 80 times, and Lee Jae-myung, leader of the Democratic Party of Korea, visited more than 40 times to support candidates (Yonhap News 4/11/24). Notably, Lee Jae-myung not only spent over 70% of his official schedule in the metropolitan area but also focused on highlighting the narrative of criticizing the Yoon Suk-yeol administration through remote campaigning via YouTube for areas he could not visit in person (MBN, 4/9/24).

3. Theoretical Background

Previous studies have largely explained the impact of real estate on elections through two main frameworks: retrospective voting and asset voting. First, from the perspective of retrospective voting, real estate is related to government policies and the subsequent rise or fall in housing prices. This can lead to two opposing outcomes in voters' choices: homeowners may vote in a way that rewards them for the increase in their home's value (Park Won-ho 2009), while those without homes may vote to punish the incumbent party for the difficulty in purchasing a home (Kim Ji-hye, Kwon Hyuk-yong 2020). Ansell (2014) found that in the US and UK, homeowners were more likely to support conservative parties, especially if their home values had increased significantly, and they showed weaker preferences for welfare policies like social insurance. Larsen et al. (2019), analyzing Danish elections, also found that as housing prices in a region increased, support for the ruling party also rose. However, in South Korea's 2022 presidential election, which occurred during a period of explosive housing price increases nationwide, both homeowners and non-homeowners voted to 'punish' the Moon Jae-in administration due to the real estate policy failures and related tax issues. According to previous research, the failure of President Moon Jae-in's real estate policies and associated tax issues had a statistically significant impact on votes for Yoon Suk-yeol, particularly among voters who had supported Moon in the 2017 presidential election but switched to Yoon in 2022 (Kang Won-taek 2022).

Second, from the perspective of asset voting, real estate is related to the objective scale of assets owned. The asset voting hypothesis posits that individuals with higher levels of owned assets tend to support right-wing parties (Foucault, Nadeau, and Lewis-Beck 2013). This argument follows the classical class voting theory proposed by Lipset et al. (1954), but focuses on assets rather than income and occupation, which they emphasized. Asset voting theory has gained attention in recent years through studies analyzing numerous domestic and international elections that have confirmed the effects of homeownership or asset size (Kim Do-gyun, Choi Jong-ho 2023; Kim Soo-in, Kang Won-taek 2022; Shin Jeong-seop 2022; Persson and Martinsson 2018).

However, analyzing the impact of real estate on South Korean voters' choices using the retrospective voting theory has limitations because it fails to consider the unique characteristics of real estate. For instance, even if the price of a house owned by an individual increases, it is difficult to experience economic benefits through buying or moving if housing prices are rising simultaneously in adjacent areas or nationwide. In other words, an increase in housing prices does not necessarily equate to an increase in wealth. Meanwhile, in South Korea, it is almost accurate to say 'assets = real estate' as the majority of household assets are dependent on real estate. According to the Bank of Korea and Statistics Korea, real estate, including housing, accounted for 75% of domestic household net worth (2023 National Balance Sheet). Furthermore, previous research indicates that South Korean voters perceive themselves as middle or upper class based on the scale of their assets, rather than income, occupation, education, or number of homes owned (Kim Soo-in, Kang Won-taek 2022). Piketty (2020) argued that assets accumulated over the long term, compared to income which is temporary and volatile, have a more significant and lasting impact on shaping class attitudes. Therefore, this report will utilize the asset voting hypothesis for analysis, focusing on the class voting tendency based on the magnitude of assets.

4. Analysis of Asset Voting in the 2024 General Election

As discussed in the introduction, correlation analyses between average real estate transaction prices by polling station or neighborhood and party vote share are methodologically prone to the ecological fallacy. In other words, they do not necessarily imply that individual assets had a significant impact on individual voters' choices. Considering this, this study conducted an empirical analysis using data from a post-election survey of voters conducted by the East Asia Institute (EAI) after the 2024 general election.

The operationalization of the main variables is as follows. First, the asset size, the variable of interest in this paper, was operationalized into five dummy variables for each of the following groups: lower (under 100 million KRW), lower-middle (100-300 million KRW), middle (300-500 million KRW), upper-middle (500-900 million KRW), and upper (over 900 million KRW). Since asset distribution typically has a long right tail, this study classified the five groups based on the median asset value, not the average. According to the 'Survey of Household Finances and Living Conditions' published by Statistics Korea in 2023, the median asset value in the Seoul metropolitan area was 386.2 million KRW in Seoul, 315.0 million KRW in Incheon, and 415.9 million KRW in Gyeonggi Province. The categories for each group were also based on the same data. The dependent variable is the party voted for by the voter in their constituency. Since the purpose of this study is to determine whether the asset level of metropolitan area voters influenced their voting choices, voting for the People Power Party was coded as '1' and voting for the Democratic Party of Korea as '0'. Additionally, gender, education level, age, household income, region of origin, and occupation, which are known to influence political attitudes, were included as control variables in a binary logistic regression model. The operationalization of each variable is summarized in Table 1.

Table 1. Operationalization of Variables

f2e34dd1a320d063

(1) Assets and Political Orientation

Before proceeding to verify the asset voting hypothesis in the 2024 general election, we examined the political orientation of metropolitan area voters by dividing them into five groups based on asset size. Table 2 shows the differences in subjective ideology based on voters' assets. The upper asset group (over 900 million KRW) appeared to be the most conservative, followed by the upper-middle asset group (500-900 million KRW), and these group differences were statistically significant. Previous research analyzing the 2020 general election also found a pattern of increasing conservatism in higher asset groups, but without statistically significant differences between groups (Kim Soo-in, Kang Won-taek 2022).

Table 2. Subjective Ideology by Asset Group

f2e34dd1a320d063

As shown in <Table 3>, there were statistically significant differences in party support among the five asset groups. The conservative People Power Party garnered the highest support at 61.26% among the top asset group, while support for the Democratic Party of Korea was relatively lower. The lowest asset group showed a slight preference for the progressive Democratic Party of Korea. Notably, the third party, Cho Kuk Innovation Party, received 14.04% support among the upper-middle asset group, a significantly higher proportion compared to other asset groups. This can be seen as similar to the so-called 'Brahmin Left' in the West, who tend to be economically affluent but politically support progressive values such as gender equality, race, and welfare policies (Piketty 2020). In Korea, a similar group is referred to as 'Gangnam Left,' characterized by high education levels, high income, and substantial assets, yet being sensitive to progressive values and supporting progressive parties (Kang Jun-man 2011).

<Table 3> Party Support by Asset Group

f2e34dd1a320d063

<Table 4> and <Table 5> present the results of the analysis of asset ownership and voting choices. <Table 4> summarizes the constituency vote choices for the five groups categorized by asset ownership, showing that support for the conservative People Power Party was highest among the top asset group, while the Democratic Party of Korea received a higher proportion of votes among the middle and lower-middle asset groups.

Similarly, <Table 5>, which examines proportional representation votes, confirms that support for the conservative People Power Party was highest among the upper-middle and top asset groups, indicating a trend of asset-based voting. Meanwhile, support for the Cho Kuk Innovation Party was lowest among the lowest asset group, suggesting, as discussed earlier, that a significant portion of its support base consists of individuals with high education and middle-class status. The cross-tabulation results in <Table 3> and <Table 4>, which show statistically significant differences (p=0.000) in support for major parties across asset groups, indicate a clear trend of patrimonial voting among voters in the Seoul metropolitan area.

[Table 4] 2024 District Vote Choice of Seoul Metropolitan Area Voters by Asset Size (%)

f2e34dd1a320d063

[Table 5] 2024 Proportional Representation Vote Choice of Seoul Metropolitan Area Voters by Asset Size (%)

f2e34dd1a320d063

(2) Analysis of Asset Voting

Table 6 presents the results of a binary logistic regression analysis conducted to determine whether assets have an independent influence even after controlling for factors known to significantly affect voting behavior. In 'Model 1', which included only control variables, male, Honam origin, and individuals aged 60 and above were found to be statistically significant. The significant influence of the region of origin variable implies that voters born and raised in regions with strong partisanship are influenced by regionalism even after migrating to other regions. This finding aligns with previous research that analyzed the voting behavior of voters from the Honam region, who were highly likely to support candidates from the Democratic Party (Kim Young-tae, 2009). In 'Model 2', which added income variables, income did not significantly affect the voting choices of voters in the Seoul Metropolitan Area. This result is similar to the findings of numerous domestic studies that report no income-based voting in South Korea (Han Gwi-young, 2013; Moon Woo-jin, 2019). Moon Woo-jin (2020) explained that in comparison to Western countries, South Korean voters historically have not formed an income-based identity, and this is because redistribution policies of progressive parties are not clearly defined, leading to no support for progressive parties based on redistribution preferences. Finally, in 'Model 3', which added the asset variable to 'Model 2', assets in the upper asset group had a statistically significant impact on voting for the People Power Party, supporting the asset voting hypothesis, which is the focus of this study. The confirmation of the independent influence of assets, even after controlling for variables known to affect individual voting behavior, implies that the 'haves' support the conservative People Power Party.

Table 6. Binary Logistic Model: District Vote – Democratic Party (0), People Power Party (1)

f2e34dd1a320d063

<Table 7> summarizes the odds ratios and marginal effects calculated by converting the probability of voting for conservative party candidates into odds ratios for each group divided by asset level, compared to progressive parties. The marginal effect measures how the probability of voting for a conservative party changes as the asset variable changes. When other independent variables (income, gender, education level, region of origin, age, occupation) are fixed at their mean values, the probability of high-asset voters voting for the conservative People Power Party increases by 24% compared to low-asset voters, and this difference is statistically significant. In other words, the wealthier the asset group, the higher the probability of voting for a conservative party.

<Table 7> Odds Ratios and Marginal Effects on the Probability of Voting for Conservative Parties

f2e34dd1a320d063

<Figure 3> Marginal Effects of the Probability of Voting for Conservative Parties by Seoul Metropolitan Area Voters According to Asset Size

Meanwhile, since the upper asset group generally has a higher distribution among older age groups than younger ones, the results in <Table 6> could be a '। illusion' due to age factors. Considering this, in <Table 8>, after conducting regression analysis in the same format excluding voters aged 50 and above, not only was conservative party voting confirmed in the upper asset group, but it also showed statistically stronger significance (p<0.01). Interestingly, in the analysis targeting only voters aged 20-40, the gender and region of origin variables did not show statistical significance. While the Yeongnam and Honam origin variables maintain their positive and negative directions, respectively, their influence is weaker than that of assets, suggesting that other cleavage factors such as issues and class may emerge in future Korean elections.

<Table 8> Binary Logistic Model – Excluding Voters Aged 50 and Above

f2e34dd1a320d063

3. Conclusion

The analysis in this study demonstrates that asset variables, observed in recent South Korean elections, also had a significant impact on the 2024 general election. Their influence became particularly pronounced and evident in some areas of the Seoul metropolitan region. The concept of 'real estate voting' is not entirely new, having emerged in the early 2000s with the surge in redevelopment issues in Seoul and having been discussed since the Grand National Party's landslide victories in the 2007 presidential and 2008 general elections. This report's analysis is also limited to voters in the Seoul metropolitan area, thus it is difficult to generalize the possibility of asset-based class voting in South Korean elections.

Nevertheless, this study shows that class-based factors, such as assets, which were identified in recent South Korean elections, had a similarly significant impact on the 2024 general election. Furthermore, younger voters, excluding those aged 50 and above, were found to make voting decisions based on a complex interplay of various factors, moving beyond the regional and generational divides that have long explained South Korean elections. These changes are noteworthy as they could emerge as significant cleavages in future South Korean elections.

Moreover, the findings of this study offer implications for the future of South Korean conservative parties. Ultimately, the phenomenon of asset-based voting signifies that the People Power Party is increasingly becoming a 'party of the wealthy.' The image of the conservative party is solidifying as a party of the 'haves,' akin to the upper asset group. The disparity in seats won by the People Power Party and the Democratic Party in the Seoul metropolitan area in the past two general elections was 16 seats to 103 in 2020, and 19 seats to 102 in the 2024 general election. In a society where economic polarization is a serious issue, the People Power Party becoming a 'party of the wealthy' can pose a significant constraint on expanding its political reach and increasing its support base. As this 'class cleavage' solidifies amidst deepening economic inequality in South Korean society, the People Power Party's prospects for electoral victory or governing will become increasingly difficult with its image as a 'party of the haves.' This study indicates that the People Power Party is not only geographically confined to the Yeongnam region but also generationally tied to the elderly, and is now also exclusively limited to the 'haves' in terms of class. The People Power Party now needs to draw on examples of how conservative parties worldwide have overcome crises. It must expand its reach by approaching the centrist and metropolitan voters with new agendas.

References

Kang, Won-taek. 2022. “Issues in the 2022 Presidential Election: Focusing on the Evaluation of the Moon Jae-in Administration's Real Estate Policy.” *EAI Working Paper*, 1-21.

Kang, Jun-man. 2017. *The Gangnam Left: Elitism After Democratization*. Inmulgwasasangsa.

Kim, Do-kyun, and Choe, Jong-ho. 2023. “Analysis of the Effects of Asset-Based Voting and Mediating Variables: Focusing on the 20th Presidential Election.” *Journal of Social Science Studies* 54(2): 1-27.

Kim, Su-in, and Kang, Won-taek. 2022. “Assets and Voting Choices: Focusing on Voters in the Seoul Metropolitan Area.” *Korean Political Science Review* 56(1):187-215.

Kim, Young-tae. 2009. “Elections and Regional Background in Korea.” *Korean Journal of Political Studies* 2(2): 61-85.

Kim, Ji-hye, and Kwon, Hyuk-yong. 2020. “Apartment Price Increases and Support for the Ruling Party: Analysis of South Korean Local Elections, 2006-2018.” *Journal of Northeast Asian Studies (formerly Journal of Unification Studies)* 35(1): 305-338.

Moon, Woo-jin. 2020. “Why Do Koreans Not Vote Based on Income?: A Comparative Analysis of Income-Based Voting in the UK, US, and South Korea.” *Korean Political Science Review* 54(2): 5-30.

Moon, Woo-jin. 2019. “Causes of the Inactivation of Income-Based Voting: Analysis of the 7th National Simultaneous Local Elections.” *Peace Studies* 27(1): 131-168.

Park, Won-ho. 2009. “Real Estate Price Fluctuations and South Korean Elections in the 2000s: A Methodological Exploration of Economic Voting After Regionalism.” *Journal of Korean Politics* 18(3): 1-28.

Shin, Jeong-seop. 2022. “Housing Ownership and Voting Choice in the 20th Presidential Election: Retrospective Voting vs. Asset Voting.” *Korean Journal of Political Studies* 15(3): 5-33.

Han, Gwi-young. 2013. “The 2012 Presidential Election: Why Did the Poor Support the Conservative Party?” *Trends and Prospects* (89): 9-40.

Ansell, B. 2014. “The political economy of ownership: Housing markets and the welfare state.” American Political Science Review, 108(2), 383-402.

Larsen, M. V., Hjorth, F., Dinesen, P. T., & Sønderskov, K. M. 2019. “When do citizens respond politically to the local economy? Evidence from registry data on local housing markets.” American Political Science Review, 113(2), 499-516.

Piketty, Thomas, trans. Ahn, Jun-beom. 2020. *Capital and Ideology*. Munhakdongne.

Lipset, S. M., Lazarsfeld, P. F., Barton, A. H., & Linz, J. 1954. “The psychology of voting: An analysis of political behavior.” Handbook of social psychology, 2, 1124-1175.

Foucault, M., Nadeau, R., & Lewis-Beck, M. S. 2013. “Patrimonial voting: Refining the measures.” Electoral Studies, 32(3), 557-562.

Persson, M., & Martinsson, J. 2018. “Patrimonial economic voting and asset value–new evidence from taxation register data.” British Journal of Political Science, 48(3), 825-842.

<Newspaper Articles and Data>

Kim, Jae-deuk. 2024. “Ruling and Opposition Parties Go All Out to Win Over Swing Voters in the Seoul Metropolitan Area.” <Joongboo Ilbo> April 7. (https://www.joongboo.com/news/articleView.html?idxno=363646030)

Kwon, Jeong-hyuk. 2024. “The 'Real Estate Vote' That Divided the Presidential Election... What About Our Neighborhood in This General Election?” <Kyunghyang Shinmun> April 15. (https://www.khan.co.kr/politics/election/article/202404150600131)

Seo, Young-soo. 2024. “[April 10 General Election] Han Dong-hoon's 'Semiconductor Belt' vs. Lee Jae-myung's 'Han River Belt'... Most Frequent Words During Rallies?” April 9. (https://www.mbn.co.kr/news/politics/5018150)

Ryu, Mina, and Han, Ju-hong. 2024. “[April 10 General Election] Democratic Party Achieves 'Overwhelming Victory' in Seoul Metropolitan Area Including Han River Belt... Ruling Party 'Performs Well' in Nakdong River Belt.” <Yonhap News>, April 11. (https://www.yna.co.kr/view/AKR20240411024700001?input=1195m)

Statistics Korea. 2023. “Household Finance and Welfare Survey.” https://kostat.go.kr/board.es?mid=a10301040300&bid=215&act=view&list_no=428364

■ Author: Kim, Suin_ Ph.D. Candidate, Seoul National University.

■ Editor: Kim, Sunhee_ Senior Researcher, EAI.

Inquiries: 02-2277-1683 (ext. 209), shkim@eai.or.kr

[1] The Seoul Metropolitan Area refers to the regions of Seoul, Incheon, and Gyeonggi Province, as defined by Article 2 of the Seoul Metropolitan Area Readjustment Planning Act.

[2] As of March 2024, approximately 26.03 million people reside in the Seoul Metropolitan Area, accounting for more than half of South Korea's total population. The proportion of voters in the Seoul Metropolitan Area among all eligible voters in the 2024 general election is 50.78%. (Ministry of the Interior and Safety Resident Registration Population Statistics: https://jumin.mois.go.kr/index.jsp, National Election Commission Voter Roll: http://info.nec.go.kr/electioninfo/electionInfo_report.xhtml)