Older Voters and the 2012 U.S. Presidential Election

By Lisa Rill and Larry Polivka (2018)

Abstract

In the 2008 presidential election, the majority of older persons failed to vote for the winner, which was only the second time in the past ten presidential elections. This trend continued in the 2012 election. In Robert Binstock’s studies of the 2008 presidential election and the 2010 midterm election, he examined age-group voting behaviors and possible electoral factors that persuade an individual’s vote. He concluded that although age may have an effect on voting behavior, race seemed to be a greater predictor of voting choices and regional figures suggest a racial period effect among older White voters in the South.

This paper is an extension of Binstock’s past studies. The first section discusses partisan leanings as a possible factor for why the older voters had a preference for Republican candidates. We then present the results of the three elections, focusing on age-groups and race/ethnicity. The next section explains the views regarding health care by the different age-groups. Finally, we discuss the findings of voting behaviors by region.


Older Voters and the 2012 U.S. Presidential Election

In the 2008 presidential election, the majority of older persons failed to vote for the winner, which was only the second time in the past 10 presidential elections. This trend continued in the 2012 election. Binstock examined age-group voting behaviors – specifically, partisan leanings, and the candidates’ contrasting ages and racial identities – to illustrate why older voters gave their majority vote to the Republicans (Binstock 2009). This report is an extension of Binstock’s past studies of the 2008 presidential election and the 2010 midterm election in which we examine the voting behaviors of various age-groups by partisan leanings, race/ethnicity, views regarding health care, and region in the 2012 presidential election.

The first section discusses partisan leanings as a possible factor for why the older voters had a preference for Republican candidates. We then present the results of the past three elections, focusing on age-groups and race/ethnicity. The next section explains the views regarding health care by the different age-groups. The final section discusses voting behaviors by region from the 2012 election.

Votes for U.S. President by Age-Groups: Partisan Leanings

Binstock examined the data from the 2008 national Election Day exit poll and found that individuals aged 60 years and older gave 51 percent of their vote to McCain and 47 percent to Obama (Binstock, 2009). He considered voters’ partisan leanings as one of the possible explanations for why McCain received distinctive majorities from older voter cohorts.

Binstock (2009) suggests that the partisan leanings of people in their late 60s are “definitely a factor in explaining the preference of a majority of current older voters for Republican presidential candidates” (p. 698). He explains that the political socialization of this cohort occurred during the eight years of Eisenhower’s presidency – the first Republican president in 20 years – and therefore, this cohort was more attracted to the Republican Party. Support for Republican presidential candidate among older voters was also evident in the 2004 election, which was a notable change from the 2000 election where the majority of the 60 and over age-group voted for the Democrat, Al Gore.

In the 2012 presidential election, Obama won over Romney by a margin of 51% -48% among all voters (NBC News, 2012). However, 50 percent of the people age 60-64 and 56 percent of those aged 65 and over voted for Romney (Exit poll, 2012). The older voters were not the only ones to give a majority vote to Romney, as voters ages 45-59 gave the Republican 52 percent of their votes. The younger age-groups, however, favored Obama with 52 percent of the votes from the 30-44 age-group and 60 percent from the 18-29 age-group. A possible explanation for these differences among age-groups was that older voters were more attached to the Republican Party than the preceding birth cohorts (Binstock, 2009). (Tables 1-3 provide a comparison of votes by age-groups for the 2004, 2008, and 2012 elections).

The first of the baby boom generation began turning sixty-five in January 2011, and over the next 17 years, around 10,000 people a day will enter the 65+ age-group (Pruchno, 2012; Stone and Barbarotta, 2011). Contrary to their preceding cohort, who were socialized into politics during Eisenhower’s presidency, the oldest of the baby boomers came of age during the Kennedy/Johnson era and tend to vote in favor of the Democrats. However, it is unlikely that this cohort had much of an effect on the percent of votes for those 65 and older in the 2012 election because the boomers were just beginning to enter this age-group. Although the boomers were a larger share of the older voting group in the 2012 election, they were in all likelihood, substantially outnumbered by voters from the older cohort, who voted for Romney. The results of the next election in 2016 could show a decrease in the majority vote for the Republicans as more boomers enter the 65+ age-group.

If Binstock’s (2009) argument of partisan leanings is valid, this trend may not continue for the following election in 2020, reflecting the increasing diversity of the baby boom cohort. For instance, the oldest boomers were raised during the time of much social change, including the Civil Rights movement, the sexual and drug revolutions, and the feminist and gay movements (Pruchno, 2012). The youngest boomers, on the other hand, reached the age of majority during the more conservative Reagan years. These younger boomers could have more conservative views than the older boomers, leading to a stronger affinity for the Republican candidate.

An example of this is the response to the exit poll question, “Do you approve or disapprove of the way Barack Obama is handling his job as president?” As indicated in Table/Figure 4, the age-group responses varied. With the 60-64 age-group, where most of older boomers fit, the majority (55%) approved of the way Obama was handling his job as president. In contrast, the majority (52%) of the 45-59 age-group, which includes many of the younger boomers, disapproved of Obama’s performance as president. This greater approval of Obama by the older baby boomers gives some support to the perception that the Democratic lean of the Kennedy cohort is responsible for this modest difference.

Votes for U.S. President by Age-Groups and Race/Ethnicity

Another factor Binstock (2009) examined in the 2008 and 2010 elections was race. Due to the birth cohort and period effects from an era when overt racial discrimination was stronger than succeeding eras, it was expected that “the contrasting racial identities of the two candidates might influence the voters” (p. 700). Although 80 percent of voters in the 2008 exit poll stated that race was not a factor in their choice, when looking at the breakdown of votes by age and race, McCain received the majority of the White votes. However, based on the data from the 2008 election, Binstock could not conclude that race was an important factor in older persons’ voting choices. He did briefly mention that there might be a racial period effect when looking at regional figures.

In his study of the 2010 midterm congressional election, Binstock (2011) found that only examining the voters by age-groups, “hides even greater variations in the voting patterns of racial and ethnic groups” (p. 414). Similar to the 2008 election, there were substantial differences between the White, Hispanic, and Black votes in the 2010 midterm election. The majority of Whites voted for Republican congressional candidates, while most of the Blacks and Hispanics voted for Democrats.

When examining the voting patterns of racial and ethnic groups in the recent 2012 election, the findings are comparable to the previous elections, as substantial differences appear among Whites, Blacks, and Hispanics. In the 2008 election, McCain received the majority (55%) of the White votes, compared to only 31 percent from Hispanic and 4 percent from Blacks (Binstock 2009). In the 2012 election, Romney received an even greater majority (59%) of the White votes, compared to just 27 percent from Hispanics and 6 percent from Blacks.

Broken down by race and age-groups, the results show that there is not much change between the two elections, except with the 18-29 age-group. In the 2008 election, the majority (54%) of the White 18-29 age-group favored Obama. However, in the 2012 election this same age-group favored Romney with 51 percent of their votes. As Table/Figure 5 illustrates, there was a slight increase in the Whites and Blacks support for Romney in all age-groups (apart from the Blacks age 65+, which stayed the same). For the Hispanics, the oldest and the youngest age-groups had a small increase in their support for the republican candidate in the 2012 election, and the two middle aged groups decreased their support.

Overall, it seems that race, rather than age, is a stronger predictor of voting choices. As Binstock (2009) pointed out, “regional figures suggest the possibility of a racial period effect among older White voters in the South” (p. 700). The final section of this report discusses the findings of voting behaviors by region from the 2012 election.

Age-Groups and Views about Health Care

In the 2010 midterm elections, when Republican candidates effectively linked possible threats to the Medicare program to Democratic policy initiatives, 59 percent of the 65 and older age-group gave their votes to the Republican candidates (Binstock, 2011). The one question on the Election Day exit poll that had some relevance to the threats of Medicare was, “What should Congress do with the new health care law? Expand it? Leave it as it is? Repeal it?” (Binstock, 2011, p. 412). The 65 and older age-group were most (55%) in favor of repealing the Patient Protection and Affordable Care Act (PPACA).

The same question was on the 2012 exit poll asking, “What should happen to the 2010 health care law? Expand it? Repeal it?” As Table/Figure 6 illustrates, 51 percent of voters aged 65 and older were in favor of repealing the health care law, compared with 42 percent of voters aged 60-64. This is less than the 55 percent of votes from the same age-group in the 2010 midterm congressional election to repeal the law. In addition, in the 2012 election voters aged 45-59 were the age-group most in favor of repeal (56%), which is contrary to the 2010 midterm congressional election where the 65 and over age-group had the highest percent supporting repeal.

In line with Binstock’s (2009) suggestion to focus on partisan leanings, these findings reflect differences in views by cohort. Voters in the 65 and older age group and the younger boomers (45-59 age-group) both supported repeal of the health care law. As mentioned earlier, these two groups reached the age of majority during the Republican years. On the other hand, the 60-64 age-group, which consists of the older boomers, supported expanding the health care law. This group reached majority under a Democratic president.

Another health care question in the 2012 exit poll data asked, “Who would better handle Medicare? Obama or Romney?” Table/Figure 7 reveals a pattern similar to the results from the health care law question. The majority (52%) of the 65 and older age-group believed that Romney would handle Medicare better, compared to only 44 percent of the 60-64 age- group. Likewise, there was a slightly higher percent of people in the 45-59 age-group who felt that Romney would handle Medicare better; however, it was only a one percent difference. The response regarding Medicare from the 45-59 age-group is not a strong indicator that cohort effects are a factor. Yet, taken together with the data on the other health care questions, the argument becomes a bit more robust.

Overall, Binstock suggests that although race seems to be a greater predictor of voting behavior than age, the voting behaviors of the 65 and older age-group from the 2010 midterm congressional election were “affected by their perception that President Obama’s health care reform legislation would have negative consequences for health care paid for through the Medicare program” (Binstock, 2011, p. 415). Therefore, the older voters had the chance influence the outcome of the 2012 election to a greater extent than they have historically.

Age-Groups and Race/Ethnicity by Region

The political history of the White South has been the focus of much debate over the past four decades. There are a variety of perspectives discussed in the literature regarding the partisan realignment in this region, which changed from being mostly Democratic in the pre- civil rights period, to being a majority Republican over the last several years. The most common debate today is whether the South remains politically exceptional. Southern exceptionalism has been defined as a unique social environment that is the “result of some underlying cultural framework, nurtured by the geographic setting, history, and socioeconomic relations” (Aistrup, 2010, p. 909), which generates different patterns of behavior among southerners when compared to other regions.

Some studies argue that the South is no longer unique (Aistrup, 2010; Shafer and Johnston, 2006), while other studies suggest that the political attitudes of the South remain distinctive (Hayes and McKee, 2008; Lewis-Beck, Jacoby, Norpoth, and Weisberg, 2008; Osborne, Sears, and Valentino, 2010). In this section, we use the regional results from the 2012 election to support the argument that the South continues to be politically exceptional.

Two general factors seem to lead to political realignments; one being change among party elites, and the other is shifting attitudes of the mass public (Valentino and Sears, 2005). In this case, the party elites changed their views on racial and other social issues. More specifically, “Democratic elites began to move to more liberal positions on noneconomic issues such as national defense or abortion in the 1970s, and the Reagan era heightened the distinctive economic conservatism of the Republican Party” (Valentino and Sears, 2005, p. 673).

In the early 1950s, the Democratic Party dominated over three-quarters of the Southern electorate; however, by 2002, it had declined to a meager twenty-six percent (Osborne et al., 2010). Osborne and colleagues (2010) suggest that political realignment occurred because, “White southerners have long been more conservative than Whites elsewhere in the nation, and they finally came to see that the Republican Party better represents their distinctive preferences…[and] partisanship has become aligned in a more consistent fashion with underlying ideological preferences” (p. 84-85).

Partisanship is gained through socialization, specifically the transmission of party identification from parent to child (Hayes and McKee, 2008; Valentino and Sears, 2005). Over the last three decades, the number of southern White voters who identify with the Republican Party has increased dramatically and a generation of southern Republican parents is now transferring their support of the Republican Party to the next generation of voters. This generational exchange, along with the decline of older southern Democrats, has led to an increase in southern White Republicans.

Examples of this can be found in the 2012 election. As Table/Figure 8 illustrates, the majority vote favored Romney in the South by 10 percent. In the Midwest, the vote was even at 49 percent for each candidate; the remaining 2 percent went to other candidates. In the East and the West, Obama was favored by approximately 15 percent. Overall, the South was the only region with a majority vote for Romney.

A closer look at each of the southern states in Table/Figure 9 shows that, except for Florida and Virginia, all of them voted for Romney. However, the percent of difference was very small, with a one percent difference in Florida and a two percent difference in Virginia. Of the states that voted for Romney, they all had a much higher percent disparity (between 8% and 34%). The only state with a small percent difference in favor of Romney was North Carolina (3%).

The results can be broken down further by age-groups and race for each region. In the 2008 election, voters in the South of all races who were aged 60 and older favored McCain at 62 percent, compared to 45 percent in the East, 46 percent in the Midwest, and 47 percent in the West (Binstock, 2009). The results from the 2012 election in the South illustrate racial differences in votes for Romney. Whites, in all age-groups, were the only ones to favor Romney (68% for those 18-29, 69% of the 30-44 year olds, 70% of the 45-64 age-group, and 72% of the 65+ age-group). In the other three regions, all of the White 18-29 year olds had a majority vote for Obama, and Whites in all of the other age-groups (except the 30-44 age- group in the West) favored Romney. However, the margin of difference in all of the groups in the three other regions was much smaller than in the South. In all four regions, Blacks and Hispanics of all ages (with data available) favored Obama.

Overall, based on the data from the 2008 and 2012 presidential elections, it seems that race/ethnicity and region, rather than age, are the predominant factors that influence voting choices. Furthermore, the results from the 2012 election illustrate that the politics of the South continue to be exceptional as Whites vote overwhelmingly for Republican candidates, while the White vote in the rest of the country is split much more evenly between Democratic and Republican candidates.

Methodological Limitations

Edison Research is the exclusive provider of the National Election Exit Polls to the major US television news networks and the Associated Press. It was formed in 2003 in order to provide information on election night about the vote count, election analysis, and election projections. The voter surveys provide political, demographic, and geographic information detailing voters and whom they chose for President, Senator, and Governor, as well as the results for newsworthy state ballot issues.

The National Election Pool members (ABC, AP, CBS, CNN, FOX, NBC) prepared the questionnaire. The survey was conducted at 350 polling places among 26,565 Election Day voters. In addition, there were 4,408 absentee and/or early voters interviewed in a pre-election telephone poll. Results from the phone poll were combined with results from voters interviewed at the polling places.

The authors acquired the data for this analysis directly from Edison Research and from other media publication websites. One limitation to using this data is that it was aggregated by Edison Research and presented in crosstab format. With aggregated data, the researchers could not study the effects of variables at the individual subject level. Therefore, limited results could be presented. In addition, without access to the individual data, multivariate analysis could not be included in this report.

Conclusion

Binstock’s studies of previous elections focused on age-group voting behaviors and attempted to explain the electoral factors that are likely to influence an individual’s vote. In his study of the 2008 presidential election, he found voters partisan leanings, the candidates’ contrasting ages and racial identities, and how the voters viewed the candidates’ judgment and experience to be explanations for why older voter cohorts favored the Republican candidate.

Binstock’s article on the 2010 midterm congressional election examined whether alleged threats to the Medicare program had an effect on older age-group bloc voting (Binstock, 2011). Binstock found that for the first time in four decades, there was evidence of an emerging old-age voting bloc and suggested that the older voters could be the key to the outcome of the 2012 election. However, Binstock’s (2011) findings also revealed notable differences in the age-group data by gender and race/ethnicity. He anticipated that “the contrasting racial identities of the two candidates might influence the voters” (p. 700). Although Binstock could not conclude that race was a factor in older persons voting choices, he did suggest that there might be a racial period effect when looking at regional data.

Continuing along the same lines as Binstock’s past studies, we examined the findings from the 2012 exit poll data. The responses suggest that partisan leaning, the candidates’ contrasting racial identities, views of health care, and region are all factors that may influence an individual’s vote. Overall, it seems that race/ethnicity and region are greater predictors of voting choices than age-groups with southern and some western state white voters strongly supporting the Republican candidate. This does not mean, however, that these margins could not be reduced in favor of Democratic candidates in future elections.

The policy issue that could possibly be used to reduce the Republican advantage among older voters in the future is retirement security and the likelihood that Social Security, Medicare and Medicaid will play an increasingly important role in supporting the economic well-being of future retirees. The data reported here and elsewhere would seem to indicate that Republican candidates have benefitted most from any concerns older voters may have regarding the current and future status of these programs, especially the potential impact of health care reform (The Affordable Care Act, 2010) on Medicare.

The Obama Administration, however, has made it more complicated for Democratic candidates to take this tack as aggressively as many would like, by supporting the chained Consumer Price Index (CPI) for determining the size of the annual cost of living adjustments. Use of the chained CPI would reduce the Cost-of-Living Adjustment (COLA) increases by a modest, but significant amount over time compared to the regular CPI, which has been used to determine COLA increases since 1973. Furthermore, the Administration, as well as many Democratic congressional leaders, have been relatively quiet about what many analysts consider a steady erosion in retirement security caused by the decline of defined benefits private pensions and rising out-of-pocket medical costs for Medicare beneficiaries since the 1980s. These developments have diminished the Democrats ability to fully exploit their historical advantage on Social Security and Medicare during the most recent elections and helped the Republicans become dominant among older voters. The Republicans are likely to retain their large edge among these voters in the absence of a concerted Democratic effort to make retirement security and a strong, unambiguous defense of Social Security and Medicare central to their campaign strategy in future elections.

Although the Republicans may be at a long-term electoral disadvantage in the face of major demographic changes over the next 20 to 30 years, their currently solid advantage among many older voters could keep them competitive in state and national elections for some time to come. Younger and minority voters are increasingly important electoral groups, but the older voter population is also growing with the aging of the boomer population and their high level of participation in elections is not likely to decrease in the future.

Democrats could potentially offset this Republican advantage by increasing voter participation rates among younger and minority groups. This kind of initiative, however, could lead to even higher voting percentages among older white voters and make the contest between the parties over retirement security and the future of Social Security and Medicare an even more salient electoral issue in the future than it is today.