How would a national primary campaign between Bernie Sanders and Hillary Clinton play out? Recently both David Wasserman and Nate Silver have attempted to sketch out what a longer-running electoral battle between Sanders and Clinton might look like.
Both Wasserman and Silver argue that because Sanders’s support stems from a base of white liberal Democrats, Bernie might be able to compete with Clinton in Iowa and New Hampshire, but he will struggle in states with different voter populations.
Let us leave aside the question of whether it’s reasonable to assume that Sanders — who has already defied conventional wisdom to push Clinton to the brink in the early states — cannot conceivably broaden his appeal to include non-white or non-liberal voters.
For now, I just want to consider the basic assumption about what Sanders’s base looks like right now: specifically, how “liberal” it really is. Both Wasserman and Silver assume, rather cavalierly, that a handful of national polls is sufficient to determine the character of the Sanders coalition. (Wasserman builds his entire model from a single NBC/Wall Street Journal survey.)
And in national polls, Sanders supporters do indeed tend to look the way that Wasserman and Silver describe: white, liberal, and often college-educated. But as I argued here a few days ago, the voter coalition that Sanders has built in Iowa and New Hamsphire — the states where he has campaigned intensively — looks somewhat different.
Sanders’s strength with voters making less than $50,000 a year — and his relative lack of appeal among voters making above $100,000 — sets him apart from Democratic primary challengers in years past like Bill Bradley, Howard Dean, and Barack Obama. All these “progressive” underdogs attracted their strongest support from wealthier voters, while struggling, in relative terms, to win lower-income support. (Nate Cohn notices the same trend in today’s New York Times.)
Compared to Dean or Obama, there’s also reason to believe Sanders’s early state backers are much less monolithically “liberal.” Public Policy Polling (PPP), which has tracked Iowa voter ideology most comprehensively, divides likely caucus-goers into five different groups, ranging from “very liberal” to “very conservative.” Their results, over three polls since December, do not at all conform to the Wasserman/Silver presumption that liberals are flocking to Sanders, while moderates support Clinton.
Very Liberal (23%–25%)
|PPP Jan 27–28||45.0%||47.0%|
|PPP Jan 8–10||47.0%||48.0%|
|PPP Dec 10-13||35.0%||59.0%|
Somewhat Liberal Liberal (31%–37%)
|PPP Jan 27–28||36.0%||53.0%|
|PPP Jan 8–10||37.0%||51.0%|
|PPP Dec 10-13||38.0%||54.0%|
|PPP Jan 27–28||45.0%||41.0%|
|PPP Jan 8–10||38.0%||44.0%|
|PPP Dec 10-13||30.0%||52.0%|
|PPP Jan 27–28||42.0%||43.0%|
|PPP Jan 8–10||37.0%||44.0%|
|PPP Dec 10-13||45.0%||28.0%|
Very Conservative (1%–3%)
|PPP Jan 27–28||21.0%||36.0%|
|PPP Jan 8–10||15.0%||44.0%|
|PPP Dec 10-13||9.0%||23.0%|
|PPP Jan 27–28||40.0%||48.0%|
|PPP Jan 8–10||40.0%||46.0%|
|PPP Dec 10-13||34.0%||52.0%|
In general, PPP’s Iowa polls have favored Clinton. But her advantage stems not from any disproportionate appeal with “moderate” or “conservative” voters — the most recent survey finds Sanders ahead among moderates, despite trailing by eight points overall. In fact Clinton’s most loyal base appears to be “somewhat liberal” Iowa voters. Sanders, meanwhile, might hold a slight relative edge among “very liberal” voters, but they are hardly the bedrock of his support.
Other Iowa polls also show uneven results. Two January surveys from Quinnipiac do show Sanders winning strong “very liberal” support — but Clinton again leads among “somewhat liberal” caucusers, while the “moderate”/”conservative” vote splits down the middle. Only CBS/YouGov polls have yielded a result that follows the linear ideological pattern in which “very liberal” voters break to Sanders, “somewhat liberal” voters are split, and “moderates” back Clinton.
In New Hampshire, meanwhile, PPP’s early January survey showed Sanders and Clinton deadlocked among “very liberal” voters, while Sanders leads “moderates” by 12 points. Four recent polls from Marist and Fox News have Sanders leading “liberal” New Hampshire voters at about the same rate — no more and no less — than the overall Democratic electorate.
In other words, the bulk of the New Hampshire and Iowa polls do not reveal Sanders winning over an ideologically consistent base. His early state supporters are certainly white — as is 94 percent of New Hampshire and 92 percent of Iowa — but they do not seem to be particularly “liberal.”
How can Sanders, whose campaign program obviously leans much farther to the left than Clinton, have won so much “moderate” and even “conservative” support? The simplest answer points to the fuzziness of the ideological descriptors so beloved by the national media. “Liberal,” “moderate,” and “conservative” may well seem like stable identities among the pundit and professional class, but for many Americans — and perhaps especially for lower-income Americans — they are much more arbitrary, at once too narrow and too ambiguous to be of real use.
In the case of Sanders, a democratic socialist who seldom describes himself as a “liberal,” the general unreliability of these conventional categories seems even more conspicuous. A large chunk of Iowa and New Hampshire voters, at any rate, seem to think that there’s nothing especially immoderate about a candidate who stands for universal health care, a living wage, and free tuition at public universities.
Certainly, to project what a Sanders-Clinton race might look like after the early states, we need a much more textured portrait of Bernie’s coalition than the experts have given us so far. In Iowa and in New Hampshire — as in Vermont, a Republican state when he first won statewide in 1990 — Bernie’s appeal does not seem limited to “liberals.”