[EAI Opinion Review] How Much Did Generational Voting Affect Election Results?

EAI OPINION Review No. 201006-04

Focusing on the Seoul Mayoral Election

Turnout Increase in the June 2nd Local Elections

The voter turnout for the June 2nd local elections was 54.5%, a 2.9 percentage point increase from the 51.6% in the 2006 local elections. Some media outlets have analyzed that this increase in turnout was due to increased participation by voters in their 20s and 30s, who tend to support opposition parties, and that this rise in turnout was one of the reasons for the Grand National Party's crushing defeat and the Democratic Party's significant gains. However, with accurate data on generational voting rates not yet released, it is impossible to definitively confirm whether the increase in turnout was indeed due to increased participation by younger demographics, or whether the increase in turnout contributed to the Democratic Party's gains.

To accurately estimate the impact of age on voting, it is necessary to consider not only generational voting rates but also generational voting tendencies and the proportion of voters in each generation. While data on generational voting tendencies and voter proportions exist, in the absence of generational voting rate data, this analysis aims to estimate generational voting rates using data from the 2006 local elections and generational voting tendencies from election prediction surveys, and to analyze the generational influence on the Seoul mayoral election based on these estimates.

Generational Voting Tendencies

Exit polls revealed distinct generational voting tendencies that are crucial for interpreting the results of the 5th local elections. While it is a general trend for younger demographics to support progressive parties and older demographics to support conservative parties in elections, the extent of this tendency varies by generation and election. Furthermore, the differences in turnout rates among generations also vary by election. Therefore, a comprehensive understanding of the election results requires simultaneously examining the differences in choices between the Grand National Party and the Democratic Party within each generation, as well as the differences in turnout rates among generations. Figure 1 below shows the generational candidate support rates in the Seoul region based on exit poll data.

[Figure 1] Generational Candidate Support: Seoul Mayoral Election (%)

* Source: Exit poll data from three major broadcasters (June 2, 2010)

Support for Oh Se-hoon of the Grand National Party among voters in their 20s was 34.0%, less than half of the 71.8% support for Oh Se-hoon among voters aged 60 and above. Conversely, voting for Han Myeong-sook of the Democratic Party reached 56.7%, more than double. The most pronounced difference in support was observed among voters in their 30s, who supported Han Myeong-sook at a rate 2.3 times higher than Oh Se-hoon. Conversely, among those aged 60 and above, support for Oh Se-hoon of the Grand National Party exceeded that for Han Myeong-sook of the Democratic Party by more than 2.7 times. Overall, support for the Democratic Party candidate was significantly higher among voters up to their 40s, while support for the Grand National Party candidate led among older age groups.

Proportion of Voters by Generation

To ascertain the impact of generational voting on election outcomes, it is essential to consider not only voting tendencies but also the proportion of voters in each generation and their respective turnout rates. Based on the 2008 general election data, the voter proportions in Seoul are presented in the table below. Voters in their 30s and younger, who predominantly supported the Democratic Party, constituted 46.4% of the electorate, while those aged 50 and above, who largely supported the Grand National Party, accounted for 32.3%. Although younger demographics (20s to 40s) represent a larger proportion of voters, their turnout rates are significantly lower than those of voters aged 50 and above. Therefore, judging generational influence solely by voter proportion would be misleading.

[Table 1] Proportion of Eligible Voters by Household Type in the Seoul Metropolitan Area (%)

* Source: National Election Commission (2008) "The 18th National Assembly Election Compendium."

Estimated Generational Turnout Rates

Characteristics of Changes in Generational Turnout Rates

One phenomenon observed in past elections regarding generational turnout rates is that the difference in turnout rates between generations varies with the overall turnout rate. Specifically, the gap between the turnout rates of groups with lower and higher participation tends to widen as the overall turnout rate decreases.

[Table 2] Difference in Generational Turnout Rates in Past Elections

* Source: National Election Commission (2008) "Analysis of Turnout Rates for the 18th National Assembly Election."

The table above shows that in the 15th presidential election in 1997, the overall turnout rate was 80.7%, with turnout rates for those in their 50s and 20s at 89.9% and 68.2%, respectively, resulting in a turnout difference of 21.7 percentage points. However, as the overall turnout rate decreased to 70.8% in the 16th presidential election in 2002, the turnout difference between these two groups increased to 27.2 percentage points. Furthermore, with the turnout rate dropping to 51.6% in the 2006 local elections, the turnout difference between these two groups widened to 34.3 percentage points, a gap larger than the turnout rate of the 20s group itself.

The relationship between changes in overall turnout and changes in intergenerational turnout reveals that when turnout decreases, it does not decline at the same rate across all groups. The rate of decrease in turnout is relatively lower for groups with above-average turnout, while the rate of decrease is significantly higher for groups with below-average turnout.

Estimating Generational Turnout Rates

Currently, neither the National Election Commission nor exit polls release generational turnout rates. The National Election Commission surveys turnout based on a 10% sample of the total electorate after each election, but the data collection and release process takes a considerable amount of time. Exit polls, on the other hand, survey only those who have voted, making it impossible to estimate overall turnout rates. Therefore, for the time being, it is necessary to use estimated figures for generational turnout rates.

[Table 3] Estimated Generational Turnout Rates, Voter Composition, Voting Share, and Voting Tendencies for the Seoul Mayoral Election

The table above presents estimated turnout rates based on voter proportions and generational voting tendencies revealed in exit polls. The estimated turnout rates were derived from generational turnout rates in the 4th local elections and data on voting intentions surveyed by YTN on May 25, 2010, prior to the 5th local elections. If voting intentions from YTN were directly reflected, the projected turnout rate would be 68.2%. Given that the actual turnout rate in Seoul was 53.9%, 'Estimated Turnout Rate 1', which is a calibrated rate obtained by multiplying each generation's projected turnout rate by the ratio (Actual Turnout Rate 53.9% / Projected Turnout Rate 68.2%), shows figures such as 40.8% for the 20s and 69.6% for the 60s. Calculating the vote share for each candidate based on these calibrated results, reflecting generational voting tendencies and voter proportions, yields a result of 46.4% for Oh Se-hoon and 47.8% for Han Myeong-sook, as shown in Table 4. This result differs significantly from the actual election outcome.

The basis for calibrating these results lies in the observation that the difference from the average turnout rate is larger for younger demographics, as discussed earlier. Therefore, 'Estimated Turnout Rate 2', presented in the second column, adjusts for this by setting the estimated turnout rates for the 20s and 30s lower than their stated voting intentions, while assuming a slightly higher turnout than the estimated rate for those aged 50 and above. This adjustment was determined through simulations, modifying generational turnout rates in 0.1% increments to align with each candidate's vote share. Furthermore, considering the increase in overall turnout compared to the 4th local elections, it was assumed that the turnout rates for younger demographics would have increased more significantly.

This estimation is not focused on precise accuracy but rather on understanding the general trend, aiming to provide a rough idea of generational voting patterns. More accurate information can be obtained once the actual turnout rates are released. Nevertheless, even with the estimated turnout rates in 'Estimated Turnout Rate 2' from the table above, Oh Se-hoon's vote share is calculated at 47.4%, and Han Myeong-sook's at 47.1%, slightly higher than their actual vote shares. This suggests that the turnout rate for younger demographics was calculated slightly higher than in reality.

[Table 4] Estimated Generational Turnout Rates / Voter Composition / Voting Tendencies for the Seoul Mayoral Election

According to these estimations, compared to the 2006 local elections in [Table 2], the turnout rate for the 20s increased by 2.0 percentage points, and for the 30s by 2.1 percentage points. The turnout for the 40s increased by 3.5 percentage points, for the 50s by 2.8 percentage points, and finally for those aged 60 and above, it is estimated to have increased by 1.6 percentage points. Analysis of past turnout trends indicates that when the overall turnout rate increases, there are rarely decreases in turnout rates across all generations. An increase in turnout is observed across all demographic groups, albeit to varying degrees. The difference lies solely in the magnitude of the increase.

Some media outlets have reported estimated figures suggesting that the turnout for the 20s was 34.6%, and for the 30s, it increased by 5.7% compared to the previous election (Chosun Ilbo, June 9, "Focus on the 30s Vote" by Byun Hee-jae). However, if these claims are accepted, and assuming that turnout rates for other age groups did not decrease compared to the previous local elections, it would be impossible to calculate the vote shares for Oh Se-hoon and Han Myeong-sook to match the actual election results. Therefore, it is unlikely that the turnout rate for the 30s significantly increased compared to the previous local elections.

Impact of Generational Voting Behavior on the Election

Generational Electoral Influence

If the turnout rates estimated above are valid, which age groups had the greatest influence on the Seoul mayoral candidates? Calculating the voter proportions and the bias towards candidates Oh and Han, voters aged 60 and above contributed more than twice as much to Oh Se-hoon's support compared to those in their 50s. Conversely, voters in their 30s contributed the most to Han Myeong-sook's support, contributing more than twice as much compared to those in their 40s.

To understand the significance of generational voting more concretely, it is helpful to compare it with the generational voting patterns during the 2002 presidential election, when Roh Moo-hyun won. This is because subsequent elections were won by conservative parties, requiring the control of numerous factors for comparative analysis of generational turnout. However, the 2002 election, won by progressive forces, allows for a direct comparison.

[Table 5] Comparison of Vote Shares: 2002 Presidential Election and 2010 Local Elections

* Source: Hong Young-rim, "U-turn of the 40s Generation," Chosun Ilbo (June 7, 2010)

The table above compares generations with a 10-year difference, given the nearly 8-year time gap between the 16th presidential election and the 5th local elections. Therefore, the voting tendencies of the 20s in 2002 are reasonably comparable to the 30s in the current local elections. Compared to the 2002 presidential election, the Grand National Party's vote share increased by 1.3 percentage points in the 2010 local elections, with the most significant increase in support observed among those in their 50s. Examining the changes in the Democratic Party's support, the current 30s, who were in their 20s in 2002, showed an increase of 3.6 percentage points. Considering the overall decrease of -1.6 percentage points in the Democratic Party's total vote share, this represents a substantial increase in support. As seen in the table, the 40s in the 16th presidential election, who are now in their 50s, showed the largest change, with a difference of 16.1 percentage points (11 + 5.1).

While it is true that the 30s contributed most significantly to the Democratic Party in the Seoul mayoral election, as stated earlier, a major reason for not achieving victory, unlike in the 2002 presidential election, was the decline in support for the Democratic Party among the current 40s by over 6 percentage points and the significant increase in support for the Grand National Party among the 50s. Considering the changes in turnout rates and voting tendencies, the 'rebellion' of the 20s or the high support for the Democratic Party among the 30s did not act as factors capable of altering the electoral landscape from previous elections. Furthermore, the newly eligible 20s showed reduced support for progressive parties compared to the previous cohort of 20s.

Differences in Political Consciousness by Generation

The 40s generation exhibits distinct differences from the 30s generation in terms of political consciousness. Post-election surveys of Seoul voters reveal that in foreign policy, 45.8% of voters in their 30s emphasize the ROK-US alliance, while this figure rises to 51.5% among voters in their 40s. Similarly, regarding the reliability of the joint investigation team's findings on the Cheonan incident, only 48.3% of voters in their 30s expressed confidence, whereas 62.6% of voters in their 40s showed high confidence. Moreover, the proportion of those who believe economic growth is more important than income distribution (46.6%) is higher than among the 20s and 30s. Additionally, the proportion agreeing with the statement that order is more important than freedom is 14.5% among the 30s, compared to 29.8% among the 40s, nearly a twofold difference. These results consistently indicate that while the 40s generation experienced democratization, as they age, they increasingly value a stable society in a pragmatic sense.

Compared to the 20s and 30s, the 50s generation showed a relatively lower proportion of agreement with the opinion that the opposition party should be supported to prevent the Grand National Party's unchecked dominance in this election. While over 75% of younger voters were conscious of the Grand National Party's monopoly on power, this figure was just over 50% among the 50s. This perception is shared with voters aged 60 and above. Consequently, voters aged 50 and above prioritize the efficiency of national policies and believe it is desirable to avoid political instability.

[Table 6] Preference for Key Political Issues by Generation (%) - Seoul 3rd Round

*Note: EAI, SBS, JoongAng Ilbo, and Korea Research Joint 2010 Local Elections Seoul Mayoral 3rd Panel Survey (June 3-5, 2010)

Overall Evaluation

The extent to which the results of this local election differ from the previous local election can be viewed differently depending on one's perspective. It is a consistent phenomenon, both internationally and in Korea, for the ruling party to lose seats in mid-term evaluation elections such as general elections or local elections following a presidential election. If the Grand National Party's sweep in the previous local elections was an exceptional political outcome, then the current local elections could be considered a more normal electoral result. Furthermore, voter turnout, with the exception of the 68.4% in the 1st local elections in 1995, has been relatively stable at 52.7% in the 2nd, 48.9% in the 3rd, 51.6% in the 4th, and 54.5% in the current election. Since mayoral elections are single-member district systems, there are differences in the winners and losers compared to previous elections, but the party vote shares show little variation. Moreover, there were no significant changes in generational turnout rates, and the characteristics of voting tendencies were not significantly different from the past. Ultimately, while there were differences from predictions, the overall trend in this local election was characterized more by continuity than by change. ■