[22nd General Election Research Series] Voters' Evaluation of District Representatives and District Vote Choice

1. Introduction

The 22nd general election, held this year, elected 254 out of 300 parliamentary seats through the single-member district plurality system. This means approximately 85% of the National Assembly members are elected in district constituencies. The single-member district plurality system is easy to understand and facilitates local representation and accountability of representatives at the district level. However, it has drawbacks, including a high disproportionality between a party's vote share and its seat allocation, and a weakening of political diversity. Critiques have persistently argued that the low proportionality between party vote share and seat acquisition fails to adequately reflect the will of the citizens in the legislature.

Despite these criticisms, the concentration of district seats in South Korea has gradually increased since democratization in 1987. The proportion of district seats in the 13th general election, the first after democratization, was 75% (224 out of 299 seats). This election saw this proportion rise to approximately 85%, an increase of about 10%. The primary reason for the high proportion of district seats in South Korean parliamentary elections is likely the political interests of stakeholders. However, a more significant reason is probably that more than half of the South Korean public prefers the system of directly electing a district representative for their residential area over proportional representation. The preference for district representatives over proportional representatives can be understood as a desire to secure local representation by directly electing a representative for their area and to hold that elected representative directly accountable. Theoretically, despite the weaknesses of reduced minority representation and increased disproportionality, the aspect of political accountability of individual National Assembly members is what makes the single-member district plurality system more highly regarded than the proportional representation system.

Therefore, for the district representative system to leverage its strengths, it is predicated that district voters must objectively evaluate the activities of their incumbent district representatives during their term and make their voting choices based on this evaluation. In other words, the principle of reward and punishment, where well-performing district representatives are re-elected and poorly performing ones are punished through votes, should be evident in parliamentary elections. However, in South Korean parliamentary elections, there have been many arguments that the elections often function as a mid-term evaluation of the president and the ruling party, rather than an individual evaluation of district representatives, and that voters' choices are determined by party preference or regionalism.

In this context, this study aims to analyze the impact of voters' evaluations of district representatives' activities during their term on their voting choices in the 22nd general election for district constituencies. If voters' evaluations of district representatives do not significantly influence their voting choices for district representatives, the legitimacy of the current district-centered electoral system will be further weakened. Conversely, if evaluations of district representatives significantly influence voters' district voting choices, the rationale for the current district-centered electoral system can be supported. This study will first examine the overall results of the 22nd general election district constituency voting, then explore how voters evaluate their district representatives, and finally analyze the impact of voters' evaluations of district representatives on their voting choices in district constituency elections.

2. District Vote Choice in the 22nd General Election

South Korean parliamentary elections can be characterized as reflecting voters' party preferences and evaluations of the incumbent government, particularly the president, more strongly than evaluations of individual National Assembly members. Furthermore, strong regionalist tendencies are evident, with candidates from the same party consistently winning in most districts in the Yeongnam and Honam regions. However, individual candidate factors are not entirely without influence on election outcomes. Notably, the incumbent effect is highly significant.

The results of the 22nd general election can be interpreted within this context. Amidst a backdrop of very low approval ratings for the president's job performance, the election saw a strong 'judgment on the administration' narrative. Consequently, the main opposition party, the Democratic Party of Korea, secured 161 district seats out of a total of 254, while the ruling People Power Party won only 90. Compared to the 21st general election, the Democratic Party of Korea won 163 district seats and the United Future Party (predecessor to the People Power Party) won 84. The Justice Party secured one seat, and independents won five. Therefore, overall, the district election results of the 22nd general election show little difference from the 21st general election, despite the People Power Party gaining six more seats. Looking at the total vote share for district seats by party, the Democratic Party of Korea garnered 52.33%, and the People Power Party received 46.29%. Compared to the 21st general election, where the Democratic Party of Korea received 49.91% and the United Future Party received 41.46%, both parties saw an increase in their vote share in district elections. Judging solely by district elections, the two-party system appears to have become more entrenched.

Similar to the previous election, the 22nd general election exhibited strong regional voting patterns. The Democratic Party of Korea swept all 38 seats in the Honam region (North Jeolla, South Jeolla, Gwangju). Meanwhile, the People Power Party secured 59 out of the 65 seats in the Yeongnam region (North Gyeongsang, South Gyeongsang, Busan, Ulsan). The remaining six seats were won by the Democratic Party of Korea in Gyeongsangnam-do (3 seats), Busan (1 seat), and Ulsan (1 seat), with the Progressive Party winning one seat in Ulsan.

Regarding the incumbent effect, out of the total 254 district constituencies, 152 saw incumbent National Assembly members seeking re-election, representing approximately 60%. Of these, 138 incumbent members were re-elected in their districts, while 14 incumbents were defeated. This means approximately 90% of incumbents were re-elected in the same district. Additionally, considering districts where incumbents did not run for re-election, the re-election rate for candidates from the incumbent's party was as follows: Excluding districts that were newly created or merged due to electoral boundary redrawing, where the incumbent's party affiliation was unclear, candidates from the incumbent's party were re-elected in 219 out of 246 districts. In 27 districts, candidates from parties other than the incumbent's party were elected. This indicates that candidates from the incumbent's party were re-elected in approximately 90% of the districts.

As such, the district constituency landscape of the 21st National Assembly appears to have largely remained unchanged and repeated in many aspects in the current election. What about individual voters' choices? <Table 1> presents a cross-comparison of voters' district voting choices in the 21st general election and the 22nd general election, based on the EAI 22nd General Election Survey results. <Table 1> shows that the district voting choices for the 21st and 22nd general elections are very similar. In particular, voters who voted for the Democratic Party of Korea and the People Power Party (then the United Future Party) in the last general election tended to vote for the same parties again in this election. Of the voters who voted for the Democratic Party of Korea in the district elections of the 21st general election, 73.5% voted for the Democratic Party of Korea again. In contrast, 84.7% of voters who voted for the United Future Party in the district elections of the last general election voted for the People Power Party again. In terms of district voting, the voting choices of People Power Party supporters were more consistent compared to the last general election, and there seemed to be a slightly higher rate of defection among Democratic Party of Korea supporters. However, the difference was not large enough to change the election outcome. Meanwhile, voters who did not vote in the last general election tended to not vote in this election as well.

Based solely on these results, one might expect that voters in South Korea do not significantly alter their district voting choices, implying that their choices are not greatly influenced by their evaluations of district representatives. However, to truly examine the impact of voters' evaluations of district representatives on their district voting choices, it is necessary to analyze these evaluations and explore the correlation between district representative evaluations and voting choices.

<Table 1>. Comparison of Voting Choices in the 21st and 22nd General Elections for District Constituencies

3. Voters' Evaluation of District Representatives

How do voters evaluate their district representatives? In an ideal representative democracy, voters in a district should periodically monitor and evaluate the activities of their district representative to ensure they are being well-represented. However, realistically and from the perspective of rational choice, it is not easy for district voters to closely monitor and evaluate the activities of their district representative. <Table 2> serves as an example to understand how much attention voters pay to their district representatives before examining their evaluations of these representatives.

<Table 2>. Comparison of Respondent Answers on Incumbent District Representative's Re-election Bid and Actual Re-election Status

Note: 'Uncertain' refers to districts where it was difficult to confirm the incumbent's re-election bid due to electoral boundary redrawing.

In the EAI 22nd General Election survey, respondents were asked about their residential area and whether the incumbent representative for that district was running for re-election in the 22nd general election. Despite this being a very simple question, which should naturally be known for making a voting decision in the district election, nearly 20% of all respondents answered 'Don't know.' Additionally, about 5.5% of respondents provided answers different from the actual re-election status of the incumbent. This means approximately 25% of all respondents were unaware of who was running in their district constituency. One in four respondents who voted in a district election did not consider the candidates at all. This poses a serious threat to South Korean democracy.

Of course, not all voters are interested in politics, so voters who are unaware of who is running in their district may not have voted. If they did not vote in the district election, there would be no need to know who the candidates were.

Interestingly, approximately 80% of respondents who answered 'Don't know' or provided incorrect information ('Mismatch') about the district representative candidates actually voted in the district election. <Table 3> shows the correlation between knowledge of the incumbent district representative's candidacy and voting behavior. Approximately 80% of voters who responded that they did not know whether an incumbent was running in their district reported voting, and about 85% of voters who responded incorrectly about the incumbent's candidacy reported voting. These results suggest that a considerable portion of voters do not consider the candidates important in district parliamentary elections. In other words, about 20% of South Korean voters participated in the district election without knowing who the candidates were.

<Table 3>. Knowledge of Incumbent District Representative's Candidacy and Voting Behavior

However, the fact that approximately 20% of voters participated in the election without interest in the district candidates does not mean all voters lack interest. Among the remaining 80% of voters, there may be citizens who view the district election not merely as a proxy for central politics but as a means to ensure the political accountability of their district representative.

Before investigating whether voters make their voting choices based on their evaluation of district representatives in district elections, this study examined the distribution of voters' evaluations of district representatives and their independence from partisan alignment.

The following figures <Figure 1>, <Figure 2>, and <Figure 3> show the distribution of evaluations of incumbent district representatives by district in Seoul, Daegu/Gyeongbuk, and Gwangju/Jeonbuk/Jeonnam, respectively. Due to limitations in the survey sample, a sufficient number of samples could not be secured for each of the 254 districts. Of the 254 districts, 250 had at least one respondent, with the highest number of respondents (15 each) in 'Gwanak-gu Gap' and 'Yongin-si Eul'. The four districts with no respondents were 'Seocho-gu Gap,' 'Donghae-si, Taebaek-si, Gangneung-si, Jeongseon-gun,' 'Boryeong-si, Seocheon-gun,' and 'Haenam-gun, Wando-gun, Jindo-gun.' The average number of respondents per district was six. Given the small sample size per district, directly comparing evaluation scores for incumbent representatives by district may not be appropriate. Therefore, this study does not report which district's incumbent received the highest score or the lowest score. However, it is important to examine the extent of variation in evaluations of incumbent representatives across districts and within districts.

Although the number of samples per constituency is small, to ensure maximum representativeness, the figures below (Figures 1, 2, and 3) include only constituencies where the number of respondents is at least 8 and where respondent party preference is not concentrated in a single party. First, examining the evaluation of parliamentary activities by constituency in Seoul, where the Democratic Party of Korea maintains a recent strong presence but which is relatively neutral and exhibits less regional voting behavior, is shown in Figure 1.

<Figure 1>. Voters' Evaluation of District Representatives' Activities Over the Past 4 Years (Seoul)

<Figure 1> includes a total of 14 districts. The navy blue represents the distribution of evaluation scores for the district representatives' legislative activities, and the red represents the distribution of evaluation scores for their district activities. Each evaluation score is measured on a scale of 0 to 10. The rectangular boxes in the box plots show the distribution of the middle 50% of the data, excluding the bottom 25% and top 25% of the sample. The lower whisker indicates the minimum score, and the upper whisker indicates the maximum score. The line in the middle of the rectangle shows the median value. Specifically, in districts like 'Guro-gu Eul' or 'Seocho-gu Eul,' the evaluations of district representatives by respondents are relatively high compared to other districts, and the distribution of evaluations shows little variation. In contrast, in districts such as 'Gangnam-gu Byeong,' 'Gwanak-gu Gap,' 'Nowon-gu Eul,' and 'Yangcheon-gu Eul,' there are significant differences in the evaluations of incumbent representatives' activities among respondents.

<Figure 2> and <Figure 3> show the distribution of evaluations of incumbent representatives' activities by district in the strongly regionalist areas of Daegu/Gyeongbuk and Gwangju/Jeonbuk/Jeonnam. Interestingly, compared to Seoul, the differences in evaluations of incumbent representatives across districts are even larger in these two regions. Furthermore, within a single region, there are areas where evaluations vary significantly and areas where they vary little.

From the results above, we can see that the evaluations of incumbent representatives' legislative and district activities vary significantly by region and by voter within a region. This is a crucial aspect for analyzing the correlation between voters' evaluations of district representatives and their voting choices. This is because there must be sufficient variation in voters' evaluations of district representatives across voters and regions for us to analyze whether voting choices differ based on these evaluations.

<Figure 2>. Voters' Evaluation of District Representatives' Activities Over the Past 4 Years (Daegu/Gyeongbuk)

<Figure 3>. Voters' Evaluation of District Representatives' Activities Over the Past 4 Years (Gwangju/Jeonnam/Jeonbuk)

However, before proceeding to analyze the correlation between voters' district evaluations and their district voting choices, it is necessary to examine the extent to which voters' evaluations of district representatives are independent. This is because existing research has consistently raised concerns that voters' evaluations of incumbents or economic conditions are heavily influenced by partisanship rather than being independent or objective.

<Table 4>. Evaluation of Incumbent District Representatives' Legislative Activities Over the Past 4 Years by Party Identification

<Table 5>. Evaluation of Incumbent District Representatives' District Activities Over the Past 4 Years by Party Identification

<Table 5> shows the correlation between voters' party identification and their evaluation of incumbent district representatives' legislative activities, and <Table 6> shows the correlation between party identification and their evaluation of incumbent district representatives' district activities. Voters' party identification was categorized as 'Support Incumbent Party' if they supported the party of the incumbent district representative, 'Support Other Party' if they supported a different party, and 'Unaffiliated' if they had no party preference. As some had feared, the evaluation of district representatives' activities showed a significant correlation with voters' party identification. That is, respondents who identified with the incumbent district representative's party were more likely to positively evaluate the incumbent's activities than those who did not. Conversely, unaffiliated voters and supporters of other parties were more likely to negatively evaluate the incumbent's activities than supporters of the incumbent's party, with unaffiliated voters showing a higher probability of negative evaluation than supporters of other parties.

In summary, it can be said that voters' evaluations of incumbent district representatives' activities are not entirely free from partisan preferences. However, the correlation coefficients (gamma) in <Table 4> and <Table 5> are not very high. Furthermore, considering that more than half of the respondents provided evaluations that contradicted their party identification, it is difficult to argue that voters' evaluations of district representatives' activities are completely contaminated by partisanship.

Meanwhile, this survey asked about the evaluation of district representatives' activities separately for legislative activities and district activities, as these are the two main areas of a National Assembly member's work.

<Figure 4>. Voters' Perception of the Proportion of District Representatives' Activities

<Figure 4> shows the results of asking respondents which aspect of a district representative's activities—legislative or district-based—should be prioritized. South Korean voters believe that district representatives' activities should be more focused on the district than on legislative work. Excluding 'Don't know' and 'No response,' only 19.30% of all respondents answered that legislative activities are more important, 21.92% held a neutral stance, and 58.78% responded that district activities are more important.

4. Impact of District Representative Evaluation on District Election Voting Choices

Finally, we will examine the impact of evaluations of district representatives' activities on voting choices in district elections. Specifically, we investigated whether positive evaluations of district representatives' activities increase the probability of voting for the incumbent or candidates from the incumbent's party. <Table 6> and <Table 7> show the correlation between evaluations of legislative activities and district activities, respectively, and voting for candidates from the incumbent's party.

<Table 6>. Correlation Between Evaluation of District Representatives' Legislative Activities and Voters' Voting Choices

<Table 7>. Correlation Between Evaluation of District Representatives' District Activities and Voters' Voting Choices

Both <Table 6> and <Table 7> show a significant correlation between voters' evaluations of district representatives' activities and their voting choices for candidates from the incumbent's party. However, as noted above, voters' evaluations of district representatives can be influenced by other factors such as party identification. Therefore, to determine whether evaluations of district representatives' activities significantly impact voters' voting choices, a multivariate analysis including factors that can influence voters' voting choices, such as party identification, ideological orientation, and presidential job performance evaluation, is necessary. <Table 8> presents the results of a binary logistic regression analysis including these variables. The legislative activity evaluation and district activity evaluation variables were not included in the same model simultaneously due to a high correlation between the two variables, leading to multicollinearity issues. The Pearson correlation coefficient between voters' evaluation of district activities and legislative activities for district representatives was a very high 0.848.

<Table 8>. Retrospective Voting for Candidates from Incumbent District Representatives' Parties

Note: *P<0.05, **P<0.01, ***P<0.001

The results in <Table 8> show that both evaluations of legislative activities and district activities of district representatives have a significant impact on voting choices for candidates from the incumbent district representative's party. This finding remained significant even after controlling for variables that could influence voters' voting choices, such as party identification, ideological orientation, and presidential job performance evaluation. In other words, when party preference, ideological orientation, and presidential job performance evaluation are held constant, the more positively voters evaluate the incumbent district representative's legislative and district activities, the higher the probability of choosing the incumbent or a candidate from the incumbent's party in the district election.

Among the key control variables, party identification, ideological orientation, and presidential job performance evaluation also had a significant impact on voters' district voting choices. The fact that presidential job performance evaluation emerged as a significant variable, as shown in previous research, indicates that South Korean district elections are utilized by voters not simply as a means to elect district representatives but also as a tool for mid-term evaluation of the president and the ruling party.

5. Conclusion

Historically, South Korean parliamentary elections have been evaluated as being determined by assessments of the president and political parties, rather than by holding individual National Assembly members politically accountable. The 22nd general election can also be characterized as an election where voters' evaluations of central politics strongly manifested. As shown in the results of <Table 8>, party identification and presidential job performance evaluation were significant factors influencing voters' voting choices for district representatives. For these reasons, at a macro level, the overall election outcome resulted in a landslide victory for the opposition party based on the narrative of judgment on the administration.

However, the results of this study demonstrate that, contrary to the prevailing views in academia and the media, district parliamentary elections are not solely determined by being a proxy for central politics or a judgment on the incumbent government party. Although approximately 20% of all voters participated in district elections without accurate information about the candidates, statistical analysis revealed that a significant number of voters made their voting decisions by referring to the legislative and district activities of incumbent district representatives. In other words, the statistical test results showed that voters' evaluations of district representatives' activities had a significant impact on their voting choices in district elections. This impact remained significant even after controlling for strong variables influencing voters' choices, such as party identification, ideological orientation, and presidential job performance evaluation. That is, when other conditions were equal, the more positively voters evaluated the incumbent district representative's activities, the higher the probability of voting for the incumbent or a candidate from the incumbent's party in the next election.

The findings of this study indicate that a significant number of South Korean voters utilize district parliamentary elections as a political tool to ensure the political accountability of their district representatives. In other words, the current district parliamentary election system can be assessed as fulfilling its role to some extent as a mechanism for ensuring the political accountability of district representatives. However, further research is needed to determine under what conditions and which voters meaningfully utilize their evaluations of district representatives for their voting choices. ■

References

Gil Jeong-ah and Kang Won-taek (2020) "Retrospective Voting in the 21st General Election: Evaluation of the President's COVID-19 Response and Partisan Bias" *Korean Political Party Studies* 19(4): 101-140.

Hwang Ah-ran (2017) "Direct and Indirect Effects of National Assembly Members and Incumbency: Focusing on the 20th General Election" *Korean Political Science Review* 51(4): 123-146.

The National Assembly Archives of the Republic of Korea (https://www.rokps.or.kr/dataroom/data_05.asp)

Newsmin. 2023. “If the single-member district system is maintained, a runoff (preferential) voting system should be introduced: How to implement the ‘500-Person Conference Decision,’ Part 2.” June 7. (https://www.newsmin.co.kr/news/89606/)

■ Author: Shin Jeong-seop_Professor, Department of Political Science and International Relations, Soongsil University.

■ Editor: Kim Sun-hee_EAI Researcher

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