As they reel from the election of Donald Trump, Democrats are turning to the 2018 midterms as their next best hope. The party in control of the White House tends to lose seats down-ballot, and when the incumbent is unpopular the losses tend to be far greater. Already, Trump has sparked massive and popular protests, and has achieved a majority disapproval rate in a record eight days (it took Obama 936 days and George W. Bush 1,205).
The Democratic Congressional Campaign Committee recently released its list of 2018 targets. Democrats are also targeting gubernatorial and state legislative races, which will determine redistricting after the 2020 census. However, Democrats also face uphill battles. In 2018, the Democratic senators swept into office in Obama’s re-election will face battles on ground that has shifted toward Republicans over the intervening years. Democrats are defending 10 Senate seats in states Trump won, although it’s worth noting that an incumbent advantage and an unpopular president will likely help these candidates.
How can Democrats maximize their chances? First, they need to get the basics right. They should target widely because it’s impossible to know where the floor is for Trump. They don’t want to be in a situation where new terrain opens up and they’re unprepared. They need to start winning back state-level and county-level positions that feed into higher office. They’ll need money and an aggressive recruitment strategy to get good candidates to run. But, ultimately, the 2018 election, like all others, will be determined by who shows up. The Democratic Party must make a concerted effort to target the voters who have voted in presidential elections but stay home during the midterms.
For this analysis, I look at a specific pool of voters: people who voted in the 2012 presidential election, but failed to turn out to vote in 2014. These voters are younger, more diverse, lower income and more Democratic than the voting population. If Democrats can mobilize these voters in a midterm, they can dramatically increase their chances. Progressives should see these voters as a key way to push Democratic members of Congress to the left.
The Missing Marginal Voters
I analyzed only Cooperative Congressional Election Study (CCES) respondents who were matched to Catalist, a progressive voter file database, and reported voting in 2012. Among those respondents, 25 percent did not vote in 2014, and 75 percent did. For this analysis, I’ll compare those who voted in 2014 to those who did not. (For 2012 I am relying on the respondent’s recollection, whereas for 2014 I am relying on actual voter file data. CCES does not extract vote history data from Catalist.) To distinguish the group of 2012 voters and 2014 nonvoters from consistent nonvoters, I’ll refer to them as "marginal voters" (see here for a discussion of marginal voters).
First, I explored demographic differences between the 2014 voters and the marginal voters. For one thing, 2014 voters were 81 percent white, compared to 65 percent of the marginal voters. The marginal voters were lower income, with 31 percent earning less than $30,000 a year, compared with 21 percent of all voters. While 35 percent of marginal voters earned $60,000 or more a year, 47 percent of all voters did. Marginal voters were also younger, with an average age of 43, compared with an average age of 54 among all voters.
These marginal voters are more liberal on budget questions as well. I examined the generic budget-cutting question, which asks respondents whether they would prefer defense spending cuts, domestic spending cuts or tax hikes. Marginal voters prefer defense spending cuts (44 percent) over domestic cuts (36 percent) and tax hikes (20 percent). Voters, however, prefer domestic cuts (42 percent) over defense cuts (35 percent) and tax hikes (23 percent).
Among voters, 43 percent identified as Democrats, 10 percent as independents and 47 percent as Republicans. On the other hand, 51 percent of marginal voters were Democrats, 15 percent were independents and 34 percent were Republicans. How would this affect races? The CCES survey asks all respondents an automated question that generates the gubernatorial, Senate and House candidates for their district or state. I examined these questions, excluding those who answered "other," "not sure" or "no one," but including independent candidates (the state-level questions are only asked of respondents who have a race in their state).
Across the different races, marginal voters favored the Democratic Party. Marginal voters favored Democratic gubernatorial candidates (55 percent Democratic to 43 percent Republican) while voters favored Republicans (52 percent Republican to 46 percent Democratic). The gap was smaller in House races, where marginal voters favored Democrats (49 percent to 48 percent) and voters favored Republicans (51 percent to 46 percent). The chart below shows net Democratic preference (that is, the share in favor of Democrats minus the share in favor of Republicans). The marginal voters lean Democratic in all races, from a 12-point Democratic advantage in gubernatorial races to a two-point Democratic advantage in Senate races. This analysis aligns with academic research on the subject.
A Failure of Mobilization
The implication of this analysis is that Democrats could improve their fortunes by getting presidential voters to turn out in midterm elections. These marginal voters would be quite a bit easier to mobilize than persistent nonvoters (81 percent of marginal voters in 2014 were registered). I find that 39 percent of those who voted in 2012 but did not vote in 2014 reported campaign contact, compared to 67 percent of those who voted in both elections. There were class, race and age divides in contact. Among those who voted in 2012 and were matched into Catalist, 63 percent of whites, 52 percent of African-Americans and 46 percent of Asian-Americans reported contact. Only 36 percent of those under 30 reported being contacted by a campaign, compared to 78 percent of those 70 or older. Fewer than half of those who earn less than $30,000 a year reported contact, compared with 72 percent of those earning $100,000 or more.
CCES does not include a question about which party contacted an individual. However, using the Catalist match, I can examine only individuals who were registered Democrats (in states that require party registration). Among those who are matched into Catalist and registered as a Democrat, 34 percent reported that they received no campaign contact. Among those who did not vote in 2014, only 43 percent reported a contact. The share of Republican voters reporting contacts was slightly higher than Democrats, suggesting that the vaunted Democratic ground game is no more effective, and perhaps slightly less so, than the Republican ground game. In the end, 76 percent of registered Democrats voted in the 2014 election, compared to 84 percent of registered Republicans. If Democrats want to seize on Trump’s unpopularity, they need to find a way to get these presidential voters to turn out in the off-cycle election. Donald Trump will probably help.