[EAI Opinion Review] Predicting the 2012 Presidential Election Through Analysis of Generational Voter Turnout

The 50s and 60s age groups, which have increased by 5.7 million votes in a decade, will determine the presidential election.

f2e34dd1a320d063

f2e34dd1a320d063

f2e34dd1a320d063

1. The Proliferation of Voter Turnout Formulas

"If the voter turnout exceeds 55%, the opposition camp will win." This was the election prediction formula that spread through various election camps and major media outlets leading up to the April 11 general election. In the October 26, 2011 Seoul mayoral by-election, the prevailing theory was that if turnout exceeded 45%, candidate Park Won-soon would win, and if it fell short, candidate Na Kyung-won would win. In fact, Park Won-soon won with a 48.6% turnout. However, looking back, these formulas were not absolute laws. In the July 28, Eunpyeong-eul by-election held shortly after the 2010 local elections, despite a relatively high turnout of 40.5%, which contradicted the prediction that 'the opposition party would win if turnout exceeded 35%,' candidate Lee Jae-oh of the Grand National Party won by a significant margin.

These voter turnout formulas are circulated based on the authority of experts without rigorous models or empirical verification, and when these formulas are broken, they trigger baseless accusations and debates. As a recent example, when the opposition camp lost the April 11 general election despite a turnout of 54.2%, which was close to the election victory formula, rumors spread online that the voter turnout for those in their 20s was 27%. This led to intense reactions such as "Damn 20-somethings" and accusations from opposition party supporters. Shortly thereafter, when some reports indirectly estimated the turnout for those in their 20s in the Seoul metropolitan area to be 64% based on the exit poll results from the three major broadcasters, there was a reversal on social media, with the 20s being hailed as "heroes of the metropolitan area's victory."

However, according to the official announcement by the National Election Commission after the election, the voter turnout for those aged 19 was 47.2%, for those in their 20s was 41.5%, and for those in their 30s was 45.5%. For those in their 40s, it was 52.6%, for those in their 50s, it was 62.4%, and for those aged 60 and above, it was 68.6%, confirming once again the voting pattern where older generations have higher turnout rates than younger generations. Both the '27% damn-them theory' and the '64% hero theory' turned out to be mere rumors. If the voter turnout for those in their 20s was an abnormally low 27% with an overall turnout of 54.2%, then the turnout for those in their 50s and 60s must have been abnormally high. Conversely, if the turnout for those in their 20s was around 64%, the turnout for other generations would have to be lower for the average turnout to be 54.2%. Considering that younger generations have lower turnout rates than older generations, these claims were unrealistic from the outset.

As such, various controversies surrounding voter turnout continue during election periods. Among these, attention and discussion are concentrated on a type of "turnout engineering" that predicts election outcomes based on changes in voter turnout. In particular, current turnout engineering formulas contain concerning aspects due to several flawed premises and approaches. The public's fickle sentiments, which have swung between extremes since the April 11 general election, have shown signs of these concerns. This article examines the logic of such engineering-based turnout predictions and argues for the importance of analyzing generational voter turnout from a perspective of politics itself, rather than political engineering.

2. Will Voter Turnout Increase in the 18th Presidential Election?

High Competitiveness and Heightened Interest in the Presidential Election

One of the main arguments for the opposition's advantage in the December presidential election, considering the expected voter turnout, is the prevailing perception that presidential election turnout will exceed that of general elections, and higher turnout favors the opposition. The institutional characteristic of presidential elections, where voter interest and participation are concentrated in a presidential system, along with the high electoral competitiveness between the ruling and opposition parties, are expected to lead to higher turnout compared to the 17th presidential election or the recent general election.

First, as can be seen in [Figure 1], comparing presidential election turnout with the turnout in the general election closest to it, it is generally confirmed that presidential election turnout is relatively higher than that of local and general elections. The gap between the turnout for the 14th general election (71.9%) and the 14th presidential election (81.9%) in 1992, when the general and presidential elections were held in the same year, was the smallest at 10.0 percentage points. The gap between the 2007 presidential election turnout (63.0%) and the 18th general election turnout (46.0%) in 2008 was the largest at 17.0 percentage points.

Furthermore, voter turnout is significantly influenced by institutional factors such as mandatory voting or voter promotion systems, as well as the electoral competitiveness between candidates and parties. In Korea, excluding controversial systems like mandatory voting or electronic voting, voter promotion systems are relatively well-developed compared to other countries (Lee Gap-yoon et al., 2008; Cho Sung-dae, 2009). Moreover, since significant changes in the overall voting system are unlikely due to the election schedule, institutional factors are unlikely to be major variables. However, electoral competitiveness is expected to be the biggest variable influencing presidential election turnout. The higher the competitiveness, meaning the smaller the gap between competing candidates and parties, the greater the public's interest in the election and the greater the perceived value of a single vote in determining the outcome, thus increasing the sense of voting efficacy (Blais, 2006; Lee Nae-young et al., 2010).

[Figure 1] Changes in Voter Turnout in Past Elections Since Democratization

Source: Reconstructed by the author based on the results of the National Election Commission's "Analysis of Voter Turnout in the 19th National Assembly Election" (2012).

Looking solely at the vote count results of the April 11 general election, when the votes obtained in the district constituencies nationwide are summed up, the Saenuri Party received approximately 9.32 million votes (42.3%), the Democratic United Party received 8.15 million votes (37.4%), and the Unified Progressive Party received 1.29 million votes (5.9%). Combining the votes received by candidates from the Democratic United Party and the Unified Progressive Party, they slightly surpassed the Saenuri Party's vote count by about 1% point (120,000 votes). Party support rates also show a similar trend: in 2007, the Grand National Party's support rate was 45-50%, while the combined support for opposition parties was only around 25%, indicating an overwhelming advantage of more than double. However, as of 2012, the Saenuri Party's support fluctuates between 35-45%, and the combined support for the Democratic Party and the Unified Progressive Party ranges between 25-35%. This structure inherently leads to high competitiveness in an election held with a balanced power between the ruling and opposition parties.

[Figure 2] Changes in Interest in the 17th and 18th Presidential Elections Before the Election

Source: EAI, JoongAng Ilbo, SBS, Korea Research Panel Survey for the 17th Presidential Election (April 2007), EAI, Korea Research Regular Opinion Survey (June 2012).

Source: EAI·JoongAng Ilbo·SBS· 한국리서치 17th Presidential Election Panel Survey (2007), EAI·한국리서치 Regular Public Opinion Survey (2012.6)

3. The Resurgence of Generational Voting: Generational Voting and Generational Turnout

With the expected increase in voter turnout, interest is growing in generational voting and the gap in turnout rates among different generations. The focus on generational voting is due to the restoration of distinct generational voting patterns, with strong support for the opposition among those in their 20s and 30s, and strong support for the ruling party among those in their 50s and 60s. In fact, the 16th presidential election, where Roh Moo-hyun won with overwhelming support from younger voters despite the dominance of regionalism in Korean elections, sparked significant interest in generational voting, not only in the political sphere but also academically. However, in the 17th presidential election, Lee Myung-bak also received majority support from those in their 20s and 30s, raising doubts as to whether generational voting was merely a temporary phenomenon. The 2010 local elections, with the emergence of new media like social networking services (SNS), saw increased voter participation among younger demographics and were credited as a major factor in the opposition's victory, reigniting interest in generational voting ([Table 1]).

First, the phenomenon of generational voting, where support for candidates differs by generation, is being restored. The power of generational voting is easily confirmed in the April 11 general election. As shown in [Table 1], when examining the support rates for Saenuri Party candidates versus opposition candidates (Democratic Party + Unified Progressive Party), the opposition had overwhelming support among those in their 20s (25.4% vs. 64.9% [Democratic 58.5%, Progressive 6.4%]), 30s (33.8% vs. 56.9% [Democratic 49.4%, Progressive 7.5%]), and 40s (30.4% vs. 59.3% [Democratic 52.6%, Progressive 6.7%]). Conversely, among those in their 50s, the support was 52.3% vs. 35.5% [Democratic 33.5%, Progressive 2.0%], and among those aged 60 and above, it was 67.9% vs. 23.0% [Democratic 21.7%, Progressive 1.3%], clearly showing generational differences.

As the influence of generational voting grows, attention is shifting to which generations will see a significant increase in turnout, i.e., generational voter turnout. Looking at the April 11 general election alone in [Figure 3], the overall voter turnout was 54.2%, but it was below average for those in their 20s (41.5%) and 30s (45.5%). For those in their 40s, it was 52.6%, close to the overall turnout. For those in their 50s, it was 62.4%, and for those aged 60 and above, it was 68.6%, significantly exceeding the average. Therefore, if the overall voter turnout increases in the presidential election, there is a premise that the potential for increased turnout is greater among those in their 20s and 30s compared to the 50s and 60s age groups, who already have high turnout rates.

[Table 1] Generational Voting Divide: 19th General Election and 16th-17th Presidential Elections

Note: Figures in parentheses are actual vote shares announced by the National Election Commission.

[Figure 3] Changes in Generational Voter Turnout from the 16th Presidential Election to the 19th General Election (%)

Source: Reconstructed by the author based on the results of the National Election Commission's "Analysis of Voter Turnout in the 19th National Assembly Election" (2012).

To examine the differences in the potential for increased generational voter turnout, we looked at the difference between the highest and lowest voter turnout rates for each generation from the 16th presidential election to the 19th general election. A larger difference can be interpreted as a greater potential for turnout increase, indicating greater variability in turnout for that generation. The gap between the lowest and highest voter turnout rates over the past 10 years of national elections was largest for those in their 30s at 31.9 percentage points (35.5% to 67.4%). Those in their 20s (28.1% to 56.6%) and 40s (47.9% to 76.3%) showed variations of 28.5 and 28.4 percentage points, respectively. While those in their 50s had a significant variation of 23.4 percentage points (60.3% to 83.7%), those aged 60 and above had the smallest variation of 13.2 percentage points (65.5% to 78.7%), indicating the lowest potential for turnout increase (Figure 4). Furthermore, according to the regular June survey jointly conducted by EAI and Korea Research, the proportion of "definitely will vote" respondents for the 18th presidential election was similar to the 17th presidential election for those in their 30s and 40s, but it increased for those in their 20s and decreased for those aged 60 and above, creating a situation favorable to the opposition.

[Figure 4] Difference Between Maximum and Minimum Generational Voter Turnout (%p) [Figure 5] Voter Intention for 17th-18th Presidential Elections (June)

4. Is the Formula "70% Voter Turnout = Opposition Candidate Victory" Valid?

Simulation Results: Comparison of the 16th and 17th Presidential Election Models

With an expected increase in voter turnout, the 20s and 40s age groups, along with the 40s who show relatively stronger support for opposition candidates than the current ruling party candidate, have a greater potential for turnout increase compared to the 50s and 60s age groups. Projections suggest an increase of at least 10 to 15 percentage points compared to the 16th presidential election. Furthermore, recalling the result of the 16th presidential election, where Roh Moo-hyun won with a turnout of 70.8%, predictions are cautiously emerging that the opposition will win if the turnout in 2012 exceeds approximately 70%. However, this study posits that as long as the current public support structure remains unchanged, an increase in voter turnout alone will have limitations.

To verify this, let's conduct a simple simulation based on the generational voter turnout rates during the 16th and 17th presidential elections. The 17th presidential election, which had the lowest turnout since democratization and was characterized by the overwhelming advantage of the Grand National Party, is considered a representative type of election. The 16th presidential election, on the other hand, was an election where generational voting trends were most clearly observed and the opposition won, making it a favorable voting model type for the opposition.

Considering only the generational variable, the prediction of election outcomes is ultimately determined by the combination of 'generational voter turnout' and 'generational candidate support rates.' To maximize the opposition's vote-getting potential, we calculated the vote-getting power in a one-on-one competition between Representative Park Geun-hye and Ahn Cheol-soo, assuming Ahn Cheol-soo, the most popular opposition candidate, as the unified opposition candidate. For generational candidate support rates, we used the estimated generational support rates from June 30, when the support rates for Representative Park Geun-hye and Ahn Cheol-soo were closest (Park Geun-hye 46.4%, Ahn Cheol-soo 45.4%). For generational voter turnout, we used estimated turnout rates from the 17th presidential election, which was most favorable to the ruling party, and the 16th presidential election, which yielded the most favorable results for the opposition.

First, using the generational voter turnout rates from the 17th presidential election, we multiplied the number of voters in each generation by the generational turnout rate and then by the candidate support rate for each generation to calculate the vote counts for Representative Park Geun-hye and Ahn Cheol-soo, respectively. The estimated overall turnout was 64.5% [(12,575,613 + 11,102,402) / 40,181,623]. Representative Park Geun-hye held an advantage in the 50s and 60s age groups, while Ahn Cheol-soo led in the 20s and 40s. However, due to the relatively low turnout rates of 46.6% for those in their 20s and 55.1% for those in their 30s, combined with the high turnout rates of those in their 50s and 60s and Park's support, Park Geun-hye was estimated to have an advantage of approximately 1.47 million votes (Table 2).

Meanwhile, the simulation results applying the 16th presidential election model show that the turnout rates for those in their 20s, 30s, and 40s increased by 15.5, 21.9, and 13.7 percentage points, respectively, compared to the 19th general election, reaching 56.5%, 67.4%, and 76.3%. For those in their 50s, the increase was 21.3%, and for those aged 60 and above, it was 10.1 percentage points. This model reflects a relatively higher increase for those in their 20s and 40s. The overall voter turnout was calculated to be 72.7%, an increase of more than 10 percentage points compared to the 16th presidential election, and slightly higher than the 70.8% turnout in the 16th presidential election. Comparing the vote counts of the two candidates according to the simulation, Representative Park Geun-hye received 13.97 million votes, and Ahn Cheol-soo received 12.8 million votes, resulting in an advantage of approximately 1.1 million votes for Park, which is about 400,000 votes less than in the simulation applying the 17th presidential election model (Table 3).

[Table 2] Simulation 1: Applying Generational Voter Turnout from the 17th Presidential Election

Note: Generational voter turnout is based on the "Analysis of Voter Turnout in the 17th Presidential Election" (2008) announced by the National Election Commission.

[Table 3] Simulation 2: Applying Generational Voter Turnout from the 16th Presidential Election

Note: Generational voter turnout is based on the press release for the "Analysis of Voter Turnout in the 16th Presidential Election" (2003) announced by the National Election Commission.

Overlooked Variables: Changes in Generational Support Rates and the Significant Increase in Voters in their 50s and 60s

However, contrary to expectations, the overall election results predicted a significant lead for Representative Park. In the 16th presidential election, Roh Moo-hyun won with the same generational voter turnout rates. The reason why Representative Park Geun-hye shows an advantage in the current one-on-one hypothetical matchup can be attributed to two factors.

First, there is a difference between the generational support structures of Representative Park Geun-hye and Ahn Cheol-soo today and those of Roh Moo-hyun and Lee Hoi-chang during the 16th presidential election. As seen in [Table 1], the generational support gap between Roh Moo-hyun and Lee Hoi-chang in the 20s and 40s age groups was actually smaller than the gap between Ahn Cheol-soo and Representative Park today. However, Roh Moo-hyun significantly narrowed the gap with Lee Hoi-chang by garnering 40.1% support from those in their 50s and 34.9% from those aged 60 and above. In contrast, while Representative Park Geun-hye's support in the 50s and 60s is similar to the support Lee Hoi-chang received in 2002, Ahn Cheol-soo's support is significantly lower than it was then, with 34.7% in the 50s and 22.9% in the 60s and above.

Furthermore, the current 50s generation was the 40s generation 10 years ago, when support for Lee Hoi-chang and Roh Moo-hyun was nearly equal at 47.9% and 48.1%, respectively. Today, the support gap between Park Geun-hye and Ahn Cheol-soo has widened to 18.8 percentage points. The current 60s and above generation was the 50s and above generation 10 years ago. At that time, the gap in the 50s was 17.8 percentage points (Lee Hoi-chang 57.9%, Roh Moo-hyun 40.1%), and in the 60s and above, it was 28.9 percentage points (Lee Hoi-chang 63.5%, Roh Moo-hyun 34.9%). Ten years later, the current 60s generation shows a gap of a staggering 43.6 percentage points. This indicates that the generation that was in their 40s and above 10 years ago has become overwhelmingly supportive of the Saenuri Party candidate today. Conversely, it implies that Roh Moo-hyun received considerable support from the 50s and 60s age groups compared to the current support for Ahn Cheol-soo or other opposition candidates. While the opposition's advantage remains consistent in the 20s and 30s, and the opposition shows an advantage in the 40s compared to 10 years ago, the significantly widened gap between the ruling and opposition parties among the 50s and 60s compared to a decade ago appears to be a key factor contributing to the overall simulation results favoring Representative Park. This suggests the need to shift focus from generational turnout to strategies for changing generational support rates.

[Table 4] Aging Effect: Reduction in the 2030 Generation and Increase in the 5060 Generation

Note: Figures in parentheses represent the proportion of each generation in the total voter population. The number of voters in each generation was calculated by multiplying the total number of voters by the proportion. Data sources are the National Election Commission's "Analysis of Voter Turnout in the 19th General Election" (2012), "Analysis of Voter Turnout in the 17th Presidential Election" (2008), and Table 2 from the "Analysis of Voter Turnout in the 16th Presidential Election" (press release).

Second, as Korean society rapidly enters an aging society, there have been significant changes in the composition of the voter population. This phenomenon, which has received little attention but holds great significance, is noteworthy (Table 4). In the 16th presidential election, when Roh Moo-hyun was elected, the proportion of voters in their 20s was 23.2% and in their 30s was 25.1%, meaning the 20s and 30s cohorts combined accounted for nearly half of the electorate. In contrast, those in their 50s constituted 12.9% and those aged 60 and above constituted 16.3%, totaling 29.2% for the 50s and 60s combined. However, 10 years later, the proportion of voters in their 20s is 18.2% and in their 30s is 20.4%, a combined decrease of nearly 10 percentage points to 38.6% compared to a decade ago. Meanwhile, the proportion of voters in their 50s has increased to 18.8%, a rise of 5.9 percentage points, and those aged 60 and above have increased by 4.4 percentage points. Consequently, the proportion of the 50s and 60s cohorts has increased as the proportion of the 20s and 30s cohorts has decreased. The increase in the 50s and 60s cohorts in the 19th general election may be partly due to the inclusion of overseas voters, but the significant increase in the 50s and 60s and the decrease in the 20s and 30s were already evident in the 17th presidential election, even before the overseas voting measure was implemented. Therefore, this is primarily attributable to the aging phenomenon in Korean society. This societal phenomenon of aging can be seen as one of the main reasons why, even when applying the 16th presidential election model, which reflects the relatively high turnout of the 20s and 30s, the gap in support between Representative Park and Ahn Cheol-soo was reduced but did not change the election outcome.

5. The Calculus of Generational Voting Needs to Change: From Generational Conflict to Generational Politics

Analysis of generational voter turnout is a crucial issue from the perspective of election strategy. Given the overlapping divisions of "20s-30s = progressive" and "50s-60s = conservative," and the clear divergence in generational voting preferences, it is inevitable to focus on generational voter turnout. However, voter turnout is merely an indicator reflecting voters' political interest and willingness to participate. Turnout is important because it signifies participation, not because it holds independent value or is an independent variable that determines election outcomes. An increase in voter turnout can be achieved through policies and visions that inspire public trust and confidence during the election competition between the ruling and opposition parties, and through a fair competition of leadership to realize them.

Excluding these political processes, which involve effectively gathering their demands, creating good policies to represent their interests, and communicating effectively during their implementation, there are few specific measures. While it is true that efforts to encourage voting through SNS and other means have contributed significantly, it is questionable whether they were sufficient to fundamentally alter the balance of power between the ruling and opposition parties. As confirmed by the simple simulations above, if the fundamental generational support structure between Representative Park and the opposition candidate does not change, there are limitations to relying solely on voter turnout.

When the inverted value system and misconceptions that "X% voter turnout determines the outcome" become excessive, they can lead to actions that undermine the value of democracy or shake its foundations. The notion of devaluing the votes of the elderly, expressed by the opposition party which is weak among the 50s and 60s, and the strong apprehension regarding the voter turnout of the 20s and 30s, which led to the "NEC DDoS" attack that shook the foundations of the state, can be seen as examples. Both the ruling and opposition parties need to shift their perspective on generational politics.

In elections, generational politics should primarily involve demonstrating greater responsibility to the generation that supports you, while simultaneously listening to the voices of those who do not support you and strengthening policy development and communication to secure their support. In other words, the fundamental approach is to expand support within a specific generation to change the overall support rate, which aligns with the essence of electoral politics. Nevertheless, clinging to voter turnout formulas rather than increasing support rates reduces generational election politics to election engineering. A dominant narrative that fails to gain the love of future generations lacks legitimacy and makes it difficult to guarantee electoral victory. This is why Representative Park Geun-hye cannot afford to be complacent in her current lead. Conversely, it will be difficult for the opposition to dream of governing while consistently ignoring the 50s and 60s age groups, which are concentrated with the aging population and socially vulnerable individuals. Both the ruling and opposition parties should reflect on whether generational politics has been narrowed down to politics focused on the 20s and 30s. As the presidential primaries intensify, we hope to see a rise in generational election politics rather than voter turnout engineering. ■