What drives US voters?

The 2022 US midterm elections presented a complex political landscape. This report uses data from the AP VoteCast survey to delve into the various factors that shaped voter behavior and electoral outcomes, focusing on three key areas: the impact of economic anxieties on voting behavior, the changes in policies on abortion and immigration, and the demographics of American voters. The 2022 midterms occurred against a backdrop of significant economic uncertainty, with inflation emerging as a dominant concern among voters. As inflation hit a 40-year high of 9.1% in June 2022, Americans expressed growing anxiety about their financial well-being, despite a low unemployment rate of 3.6%. By November 2022, inflation had started to fall, but was still very high at 7.1%.[1]

[1] Although on Election Day the most recently published CPI inflation figure was the September print of 8.2%, voters could already compare the prices of items they purchased in early November with those from a year before. More importantly, inflation was still very high, regardless of which measure voters relied on.

The 2022 midterms also demonstrated the rapid evolution of voter priorities. Just two years earlier, during the height of the Covid pandemic, healthcare and public health concerns dominated the political discourse. By 2022, these issues had taken a backseat to inflation. However, despite the changes in priorities compared to the 2020 elections, voters were still concerned with non-economic issues. Figure 2 and 3 show the responses of voters to the question, “Which one of the following would you say is the single most important issue for you?” The economy was the main concern for both Democrats and Republicans, with 33.2% and 63.2% of respondents choosing it, respectively. However, for Democrats climate change, abortion, and healthcare followed closely. For Republicans, immigration and crime were mentioned most often, although at some distance from the economy.

Regarding abortion, The Supreme Court’s decision to overturn Roe v. Wade in Dobbs v. Jackson Women’s Health Organization had a significant impact on the 2022 midterm elections. For about a quarter of voters, the Court’s decision was the single most important factor in their midterm vote. This figure increased to more than three in ten among groups traditionally aligned with pro-choice positions: Democratic voters, younger women, and first-time voters. These voters predominantly supported Democratic candidates. The impact was particularly pronounced among women of color, with majorities of Black and Hispanic women reporting that the Supreme Court decision influenced their voting behavior. Finally, a key issue for Republican voters in 2022 was immigration, because of a surge in the number of migrant apprehensions at the southern border. Democrats, however, ranked immigration as their last concern.

The data set from the 2022 midterms offers valuable insights into US voter behavior. It reveals that while economic concerns were at the forefront of voter priorities, these were not the only factors driving electoral decisions. The economic anxieties that dominated the political discourse led to a Republican majority in the House. However, the data also underscores the continued significance of non-economic issues, such as abortion and immigration, in shaping voter behavior. The Supreme Court's decision to overturn Roe v. Wade emerged as a pivotal factor, underscoring the powerful role of social issues in mobilizing the electorate.

Our analysis is grounded in a comprehensive dataset primarily sourced from AP VoteCast, a nationwide survey conducted after midterm and general elections in the US, comprising more than 100,000 respondents. This data provides direct insights into the demographics, sentiments, and perceptions on various economic and non-economic topics of individual voters. Initiated in 2018, AP VoteCast combines interviews with randomly sampled registered voters from state voter files and self-identified registered voters from NORC's AmeriSpeak® panel and nonprobability online panels.

To examine the relationships between voters' attitudes and their candidate preferences, we employ a logistic regression model. The regression analysis of the data at the individual level helps us understand how various factors interact and influence voting behavior. We can analyze the impact of economic factors and social policy issues, given the demographic characteristics, and vice versa. This approach allows us to isolate the pure effects of economic factors, social policy issues, and demographic characteristics on voter behavior at the micro level, rather than effects at the macro level that are distorted by the composition of the sample.

Our regression results can be described as follows. In 2022, voters who judged economic conditions as “poor” were more likely to vote for a Republican candidate in the Senate or the House of Representatives. This is plausible since the Democrats (the Biden-Harris administration) were responsible for economic policy. In contrast, those who considered economic conditions to be “good” were more likely to vote Democrat. Interestingly, voters who judged economic conditions as “not so good” also favored a Democrat. This suggests that some voters were willing to forgive Biden or did not (fully) blame him for the economic conditions. We should not forget that high inflation in 2021 and 2022 was not restricted to the US, although the empirical evidence suggests that excessive fiscal policy has made it worse. Additionally, some voters might have been making a trade-off between the economy and other issues, such as abortion rights.

Now let’s turn to these social policy issues. Voters who believed that abortion should be illegal in all or most cases were more likely to vote Republican. In contrast, those who thought that abortion should be legal in most cases were more likely to support a Democratic candidate. Voters who were not too concerned or not concerned at all about Covid were more likely to vote Republican, while those who were somewhat or very concerned favored the Democrats. Compared to voters who strongly favored increased border security, those who only somewhat favored it or opposed it were more likely to vote Democrat. Similarly, those who thought the racism issue was not too serious or not serious at all were more likely to vote Republican, compared to those who believed it was somewhat or very serious.

These questions were all about political issues, but we also found that demographic characteristics – even after correcting for economic and social policy issues – determined voter behavior. In particular, women and non-binary people were more likely to vote Democrat, as were Black, Latino, and Asian people. Age also matters, as people above 30 are more likely to vote Republican. This probability increases with age, peaking in the 50-64 segment. People aged 65 and older are also more likely to vote Republican, but not as much as those in the 50-64 cohort. If we look at education, we find that college graduates and people with postgraduate degrees are more likely to vote Democrat. In contrast, people with higher incomes are more likely to vote Republican. This may seem contradictory because higher education often correlates with higher income. However, our regression allows us to isolate the effects of education and income. For example, if two persons who have the same level of education differ only in their income level, then the one with the higher income is more likely to vote Republican. Conversely, if two individuals with the same income level differ only in their level of education, then the higher educated person is more likely to vote Democrat.

To summarize, the typical Republican voter in 2022 could be characterized as an older white male with a higher income but no college degree. This voter judged economic conditions as poor, believed that abortion should be illegal, strongly favored increased border security, was not too concerned or not concerned at all by Covid, and thought that the racism issue was not too serious or not serious at all. In contrast, the typical Democratic voter was generally a younger college-educated woman of color with a lower income. This voter judged economic conditions as good or not so good, believed that abortion should be legal, did not strongly favor increased border security, and was concerned about Covid and racism.

While table 1 may confirm several stereotypes, the regression results show some nuances. For example, the tendency to vote Republican rises with age but falls back a little for people of 65 and older. We have also taken a closer look at ethnicity and found that not all Hispanic groups exhibit the same voting behavior. In particular Cuban Americans are more likely to vote Republican than Democrat. This can likely be attributed to the Republican Party’s tougher foreign policy stance on Cuba. In contrast, Hispanics from all other countries of origin are more likely to vote Democrat. Finally, since the regression isolates the effect of being Black from being male, the results explain why Kamala Harris is currently having trouble getting support from Black men. Note that these regression results are based on the 2022 midterms, well before Kamala Harris rose to the top of the Democratic ticket. It shows that the trend of Black men being less likely to vote Democrat than Black women is not specific to presidential elections. While former president Obama recently suggested that Black men “just aren’t feeling the idea of having a woman as president,” this trend of Black men drifting toward the Republican Party has been noted for some time now. The Democrats’ “Opportunity Agenda for Black Men,” which they are now rolling out may be too little too late for this year’s election.

The 2022 midterms also highlight an intriguing connection between pop culture and political preferences. In 2020, Taylor Swift released the song “Only the young” – a political anthem aimed to encourage young adults to “speak up and stand for what is right.” The song references the surprise victory of Donald Trump in 2016 and how young voters were outnumbered. Swift’s influence reflects broader trends in youth political engagement. Her public support for Democratic candidates, particularly her advocacy for LGBTQ+ rights, gender equality, and opposition to systemic racism, aligns with issues that resonate strongly with younger voters. This alignment is significant given that Swift’s fan base largely consists of Millennials and Gen Z, demographics that statistically lean more liberal and are generally more receptive to social justice movements. The “Taylor Swift effect” thus serves as a microcosm of larger cultural shifts influencing political engagement among younger voters.

This raises two interesting questions: Are Taylor Swift fans more likely to vote Democrat, and if so, is this simply a reflection of their age cohort being more liberal, or are “Swifties” (Taylor Swift fans) actually more likely to vote Democrat than other people with the same demographic characteristics? We find that people with a “very favorable” or “somewhat favorable” view of Taylor Swift are more likely to vote Democrat. In contrast, those with a “somewhat unfavorable” or “very unfavorable” view of her are more likely to vote Republican. Notably, the regression takes age into account, so even within the subset of Millennials and Gen Z, those with a favorable view of Taylor Swift are more likely to vote Democrat. In this sense, the “Taylor Swift effect” is real. After the debate between Trump and Harris on September 10, Taylor Swift endorsed Kamala Harris, stating, “because she fights for the rights and causes I believe need a warrior to champion them.” Swift’s endorsement could help raise voter turnout among a demographic that is culturally close to the values of the Democrats.

Our findings reveal that while economic anxieties drove many voters toward Republican candidates in 2022, social issues and demographics provided countervailing forces, contributing to the Democrats' performance in the midterms. In our midterm preview in 2022, we explained that based on economic performance, Biden’s approval rating, and the usual midterm loss for the party occupying the White House, the Republicans should have been heading for a landslide victory. However, more modest polling results suggested that other factors, such as abortion, could be leading to Republican underperformance, as was evident on Election Day.

For the 2024 elections, our results suggest that the economy is likely a drag on the Harris campaign, but there are opportunities to offset this through social policy issues such as abortion. What’s more, while Kamala Harris may have been chosen as Vice-President to balance the 2020 Democratic ticket, our regression results show that she is not just another demographic of the Democratic Party, but she actually comes close to representing the typical Democratic voter. This could help explain the recent enthusiasm among Democratic voters that was absent when Biden was on top of the presidential ticket. This also meant that selecting Tim Walz as the VP candidate was necessary to balance the ticket and attract swing voters. However, our results also show that Black men have been less likely to vote Democrat than Black women well before Kamala Harris rose to the top of the Democratic ticket. Taking this key demographic for granted could hurt the Democratic Party well beyond the 2024 presidential election.