The Polls

Democrats Don’t Have to Come Home for Clinton to Win

But certain independents do.

Clinton Looks to Unify Party After Obama Endorsement

In the wake of Donald Trump’s horrible, very bad, no-good weekend and Hillary Clinton winning the California primary, capturing the title of presumptive Democratic nominee, and earning the endorsement of President Barack Obama, what should we be watching for in the polls? Was the “coming home” media mantra—Republicans having returned, some Democrats not having come back as of yet—that emerged after the last round of polls correct?

On average (and depending on which of the various averages you prefer), Clinton leads Trump by 3 to 5 percentage points. The closeness of the polls has surprised some pundits—but a deeper dive suggest that there’s a lot that makes sense in the current polling and that they still provide a good way to understand the structure and math of the 2016 race.

And, yes, even in a year when the rules supposedly don’t apply, and even in a week of unprecedented events, there is math and structure, and a range of predictable outcomes.

The first indication of this is that national and state polls to date have been logically consistent—they’ve been telling essentially the same story. Close polls in battleground states yield results in line with what we would expect when the national margin is around 4 percentage points. For example, according to the compilation of polling averages, Clinton is ahead in Florida, Ohio, and New Hampshire by about 4 percentage points and Trump leads in North Carolina by 2 percentage points.

The polls are also in line with the well-established fundamental factors that structure and explain presidential elections. The changing face of the American electorate (becoming less white), Obama’s improving job-approval rating, and an easier path to 270 in the Electoral College benefit the Democrats. On the other hand, economic growth remains tepid, the most recent jobs report was weak, optimism in the electorate remains low, most say things are headed off on the wrong track, and it is difficult for an incumbent party to win a third term.

So, without incorporating anything specific about this particular and peculiar year or this past particularly peculiar week—not to mention a certain particularly peculiar candidate—the fundamentals point to a competitive election with a slight advantage for the Democrats.

This happens to be a finding exactly in line with the current polls. Put another way, a slight under-performance by Clinton or over-performance by Trump over the next five months makes this race a toss-up. A slight over-performance by Clinton and under-performance by Trump puts this in landslide (at least under the modern definition) territory for the Democrats—something akin to Obama’s 2008 victory.

For Clinton to win, she needs Democrats to remain at a solid advantage over Republicans as a share of the electorate (this is what most current polls are showing); retain the loyalty of Democratic voters; and not lose independents by too large of a margin. For Trump to win, he must have the loyalty of Republicans while winning a larger share of independents than Mitt Romney did in 2012 (Romney won independents by 5 percentage points). Theoretically, his candidacy could also shape the electorate to be more Republican, but it is more likely that the partisan composition of the electorate in 2016 will be what the polls are showing now and be roughly what it was in 2012.

That’s the stark math for both candidates.

This business of coming home has been exaggerated, except among one specific group. In all the polls—at least the ones that release cross tabs by party—we see that Republicans have, in fact, come home to Trump. But, Democrats have also come home—that is if they ever left the house—to Clinton.

In the most recent Quinnipiac University poll, Democrats are even more at home and slightly less divided than Republicans with 85 percent of Democrats voting for Clinton and 83 percent of Republicans voting for Trump. In short, contra the widespread perception of a substantial contingent of core Democrats who are still reluctant to support Clinton, even before Obama’s endorsement of her, it does not appear that Democrats were particularly “divided.”

That said—and sparing you the long debate among political scientists—all partisans and all independents are not created equal. Some independents truly feel unattached to either party while others may identify themselves as independents, but tend to vote for one party or the other. In terms of the classic party-identification question asked in identical fashion in just about every survey, some voters, who initially answer independent, can be pushed into one party or the other when asked a follow up question on whether they “lean” toward one party or the other.

Looking at vote choice by these more precise slice of party attachments can show us who is at home and who is still out on the town. On the last four national polls conducted in the past 10 days by Morning Consult among a sample of about 8,000 registered voters, we subdivided voters in seven categories of party identification (i.e., strong Democrat, not strong Democrat, lean Democrat, pure independent, lean Republican, not strong Republican, and strong Republican).

Pure independents favor Trump over Clinton by a 34-to-24 margin, with fully four in 10 (43 percent) saying they are undecided (many of those may end up not voting). A nearly identical number of strong partisans support their party’s presumptive nominee—84 percent of strong Democrats support Clinton and 85 percent of strong Republicans support Trump. The only difference in loyalty is with the leaners.

Six in 10 (59 percent) lean-Democrats support Clinton, whereas a slightly higher proportion of lean-Republicans, about seven in 10 (67 percent) support Trump. These lean-Democrats currently have an unfavorable view of Clinton (46 percent favorable, 51 percent unfavorable), but elections are choices and they have much less love for the Donald (16 percent favorable, 81 percent unfavorable).

If these lean-Democrats, yes, we’ll say it, come home, Clinton should have a lead in the range of where Obama won in 2012 (4 percentage points). If Clinton can break even with pure independents, she will have a national lead in the range of where Obama won in 2008 (7 percentage points).  If GOP and lean GOP identifiers follow the lead of GOP elites and distance themselves from their party’s presumptive nominee, the lead for Clinton could be much greater. So, in the wake of Trump’s horrible, very bad, no-good weekend and Clinton’s capture of the title of presumptive nominee, it is those sorts of voters who we will be watching in the next round of polling and over the course of the entire campaign.

Ken Goldstein is professor of politics at the University of San Francisco and is Bloomberg Politics’ polling and political advertising analyst.

Before it's here, it's on the Bloomberg Terminal.