Demographic Tightrope or Broad Path for Democrats?

Shutterstock 534094786

For the Democratic Party, voters of color did everything right in 2016. No one turned out for Democrats like they did. And no one is suffering more due to the Democratic Party’s losses than people of color.

Their total share of the electorate was bigger in 2016 than 2012, and Democrats continued to win voters of color in astounding numbers, with Clinton getting essentially 90% of the Black vote and two-thirds each of the Latino and Asian-American votes. Many post-mortems tried to pin Democratic losses on a decreased turnout of African-Americans—and Black turnout was down some overall and in certain places in particular. But that is being judged against elections with the first African-American President on the ballot. Moreover, between 2008 and 2012, Black participation increases outpaced that of Whites in some key swing states (e.g., OH, NC, and PA), only to see Obama’s margin of victory decline between the two cycles.1

Democrats will no doubt ask these voters to deliver again in 2018 and beyond. And indeed, if Democrats are to have even a chance of victory in either the midterm or presidential cycles, voters of color will continue to do much of the heavy lifting.

But even if they come to the polls in record margins, this won’t be enough on its own. The reason? Math.

Whites, who remain the largest segment of voters in America, have become more Republican in their voting preferences over the last few cycles. And swings in their preferences can swamp fluctuations in Non-White turnout.

Voters of Color Weren’t the Problem but the Bright Spot in 2016

Voters of color have been delivering for Democrats for decades. According to Pew, in 1992, they made up 24% of the Democratic coalition, despite constituting just 16% of all voters.2 By 2016, their share of the total electorate had risen to 29%, but they were an even bigger part of the Democratic coalition, skyrocketing to 41%. And according to exit polls, Non-White voters made up 45% of Clinton’s total vote, which means they were certainly punching above their weight for her and the Party in 2016.

Black turnout in 2016 fell from the highs of 2008 and 2012, but mirrored the level of 2004 (which was higher than 1996 or 2000).3 And the truth is that Democrats must be able to win with that level of turnout from Black voters. From 2016 exit polls, 29% of the 2016 electorate was Non-White.4 That’s just two points shy of the 31% of voting age citizens that the census estimates are people of color. A steady turnout with Asian and Latino voters kept Non-White share of the electorate basically even between 2012 and 2016, despite a slight drop in Black turnout.

Given these facts, there are some who say Democrats must do more to attract Non-White voters to the polls. And certainly they should. But maybe what Democrats ought to say to Non-White voters is “thank you,” because no one did more to give Democrats a shot at winning last cycle.

Gains with Voters of Color in 2016 Weren’t Enough to Offset Massive Losses with Whites

That brings us to white voters.

If you compare 2004 and 2016 (the two most recent presidential elections in which Obama was not on the ballot), Hillary Clinton won four states that George W. Bush won in 2004 (Colorado, Nevada, New Mexico, and Virginia), while losing three states that John Kerry won (Michigan, Pennsylvania, and Wisconsin). These Bush/Clinton and Kerry/Trump states are indicative of the demographic-driven successes and—equally importantly—failures Democrats have experienced in the past decade.

All seven of those states saw their share of the White electorate go down over that time period. While the Bush/Clinton states often show a bigger drop (NV down 15%, CO & NM down 8%, and VA down 5%), some of the Kerry/Trump states came fairly close (MI down 7%, WI down 4%, and PA down 1%).

The Bush/Clinton states have two factors that contributed to their flip and to their continued Democratic shift overall. The first is education. Colorado and Virginia are two of the most educated states in the country, coming in at 2nd and 6th respectively in residents with at least a bachelor’s degree. In fact, of the 19 states plus DC with the highest proportion of college educated residents, Clinton won 18 of them. She won only 3 of the remaining 31 states. Nationally, according to exit polls, White college graduates voted for Trump by 3 points and White non-college graduates who voted for Trump by 37 points.

The second factor is race, though it’s far less correlative when it comes to state outcomes. Of the 19 states plus DC with the largest proportion of non-white residents, Clinton won 11 of them, or just over half. Nevada and New Mexico were the only states with low education levels that Clinton won. But these states, ranking 3rd and 5th respectively, have among the highest Non-White populations in the country.

In 2004, a majority of Kerry voters in Nevada were White, but by 2016, the majority of Clinton voters were people of color. New Mexico went from a coalition with a slight majority of people of color to one where nearly 60% of Clinton voters were people of color.

That increase with voters of color helped down-ballot as well in Nevada and New Mexico—which were two rare bright spots in a night of disappointment for Democrats. In Nevada, Democrats won back the State Senate and House after losing both in 2014, picked up two U.S. House seats, and held a U.S. Senate seat in a competitive, open-seat election. In New Mexico, Democrats won back the State House after losing it in 2014, increased their majority in the State Senate, and picked up the open office of Secretary of State in a special election. Metropolitan Albuquerque, Las Cruces, and Las Vegas, all places with large Latino populations, helped drive these gains.

Yet in the Kerry/Trump states of Michigan, Pennsylvania, and Wisconsin, Clinton’s share of the White vote cratered. Comparing 2004 and 2016, 5 to 8% fewer White voters in each of those states voted Democratic for President. What these Kerry/Trump states have in common is that they all fell well outside the top 20 in educational attainment (PA ranked 25th, Wisconsin 28th, and Michigan 35th).

Small increases in the share of voters of color or the margin of victory Democrats won in those communities simply couldn’t make up for such large losses with Whites. Of the four Bush/Clinton states, the only one in which Democrats did substantially worse with White voters this time around was Nevada. But the fact that Nevada went from 23% voters of color to 38% made up for that shift. Not many states are expected to diversify that quickly.

Dems Diverse Coalition Graph

Note: Graph shows voting result as a percentage of the demographic's share of the electorate. The numbers from exit polls are rounded to the nearest whole number, and may not add up to the full result.

A Striking Data Point about Race and Education

Given the math outlined above, the way the 50 states and DC voted are correlated to their educational attainment more than to their racial composition (take a look at where the Bush/Clinton and Kerry/Trump states from above fall on the two charts below). While neither is a particularly precise measure, this is an important reminder that Whites are the majority of voters in most states, and a huge Republican advantage with non-college Whites can completely swamp a Democratic advantage with voters of color.

Dems Diverse Coalition Education-v-Race

Senate and House Maps are Rough if Democrats Can’t Do Better with White Voters

Without winning back some of these White voters, the Senate map in 2018 is particularly foreboding. If they repeat their 2016 performance among White voters, Democrats could potentially pick up the Senate seats in Arizona and Nevada, but they would face strong headwinds in at least half a dozen Senate seats in Indiana, Missouri, Montana, North Dakota, Ohio, and West Virginia. Trump won each of these states with more than 50% of the vote, all six are 80% White or more, and five of the six have college attainment rates in the bottom half of the country. Increased turnout among voters of color compared to White voters could potentially save Democratic Senators in other Trump states, like Michigan, Pennsylvania, and Wisconsin, but those three states are also Whiter than the national average and have low-to-middling college attainment rates.

Democrats need 24 seats to win back the House, but many of these swing districts are also overwhelmingly White. The average swing district, as determined by the Cook Political Report, is 74.4% White, as compared to the safe blue districts, which are 47.9% White.5

So it’s clear that increased turnout by voters of color could be enough to eke out an Electoral College win in certain scenarios, but it won’t fix Democratic deficits in the Senate or House.

2017’s First Test Case: Decent Non-White Turnout Wasn’t Enough in GA06

Let’s take the high-profile special election in Georgia as a test case. GA06 is a largely White district. The New York Times has shown that it went from 79% White in the 2014 election to 71.4% White in 2016. That makes sense given that more people turnout in presidential cycles, and Whites are historically more likely to vote in midterms.6 For the GA06 special election, the White share of the electorate was 75.6% in the first round while in the second round, it was 74%. Given that the White citizen voting age population (CVAP) of GA06 is 72.9%, a 74% White share of the electorate indicates that voters of color turned out in decent numbers compared to a typical special election.

The New York Times even noted that Democrats “brought a surprising number of irregular young and nonwhite voters to the polls.” It might shock some to note that despite the White share of the electorate dropping 1.6% from round one to round two of the special, the Republican share of the vote slightly increased from 51% to 52%. This was very much a case of voters of color doing their part but getting outvoted by a White majority that simply could not be persuaded to vote Democratic.

Democrats Must Appeal to Voters of Color, but it Might be in Ways You Wouldn’t Imagine

If Democrats want a good midterm election night, they certainly must invest in a concerted effort to get their most loyal voters—people of color—to the polls. A Salon report found that 67% of people who voted in both the 2012 and 2014 elections reported being contacted by a campaign in 2014, while only 39% of 2012 voters who didn’t vote in the 2014 midterms said that they had been contacted in 2014. Get out the vote efforts matter, and they will again in 2018.7

But Pew found that only 20% of non-voters from the 2014 midterms cited not liking candidates, not knowing enough, or not caring as the reason they didn’t vote.8 Compare that to 67% that said they didn’t have the time and 10% who couldn’t vote due to a registration problem or being physically unable. So the answer might be less about asking people to vote and more about ensuring that they can, by investing in things like registration drives, vote by mail programs, and driving early voters to the cast their ballots. This all requires long lead-times; Democrats can’t wait until the last minute to target voters of color to ensure they are able to cast their ballots.

So while Democrats shouldn’t rely on voters of color to restore their majorities, they must invest early and big in registering and turning them out. To accomplish that goal, organizations like the National Association of Diverse Consultants are urging the Party to hire consultants and staff that can better help understand the unique needs of these key communities.9 This is smart: reaching voters of color requires understanding that their lived experiences are unique and differ not only from the dominant White culture, but also from each other. Voters of color are not a bloc and should not be treated that way.

This point became especially clear during Third Way’s recent district visit in Miami. One leader in the Caribbean community told us that in 2016, a Democratic campaign staffer told him they planned on reaching the Haitian immigrant community through ads on traditional African-American media sources like BET—rather than investing in local radio stations that are popular with the Haitian community. He told us that elected officials don’t “have a true sense of the black community in Miami… They think BET will reach all the black people in the country. It’s not true.”10 Likewise, Latinos vary greatly on whether they own landlines or cell phones, and whether they are English or Spanish dominant. That has huge implications on phone banking efforts—and public opinion research. If they want to continue to run-up big margins with voters of color, Democrats must put their money where their mouth is in figuring out how to communicate these communities. The DCCC, under the leadership of Chairman Ben Ray Lujan (the first Latino to ever serve as Chair), has begun such efforts, but others in the Party hierarchy must follow suit quickly to prepare for the coming midterms.


Voters of color are an integral part of the Democratic coalition, but to be a winning one, it must also cut a broad path. That means continuing to earn the vote of people of color and winning back some of the White voters lost in 2016. Democrats would be wise to find a unifying message focused around economic opportunity, without ignoring the unique economic and societal struggles faced by people of color. That is what’s required if Democrats are to build a winning coalition up and down the ballot—and to sustain it over the long term.

End Notes