Do Not Choose To Lose
Sometimes, losing is a choice. And if the Republican Party nominates Hillary Clinton donor and Clinton Foundation donor Donald Trump to be its presidential candidate, the choice to lose to Hillary will already have been made, and the only question will be how to manage the scale and collateral consequences of the defeat. Nobody should go to the polls Tuesday under the delusional impression that a vote for Trump is anything other than a choice to ensure that Hillary Clinton is the 45th President of the United States.
Let us count the ways. For simplicity, I’ll just compare Trump vs. Hillary Clinton to Marco Rubio vs. Hillary Clinton to illustrate why Trump would fail in a great many situations where Rubio would succeed. But a similar analysis would show, if for slightly different reasons, why Rubio would run better against Bernie Sanders than Trump and why Ted Cruz or one of the two pretend GOP “candidates” would run better than Trump against either.
This analysis will look at three factors. First, I’ll walk through the numbers, the polling we have so far. Second, I’ll offer some analysis of why we should not discount what the polling is telling us. And third, I’ll look at the structural differences between a primary and general election that will make it prohibitively difficult for Trump to translate a primary win into a plausible general election campaign.
Head-to-Head National Polls
Head-to-head general election polling during primary season is not the most reliable of indicators, but since we obviously can’t get post-primary polls until the primaries are over, it’s all we have in terms of polling evidence, so it’s as good a place as any to start. Here’s where the national general election polling averages stood as of this weekend:
You will notice not only that Trump trails Hillary by almost 3 points while Rubio leads her by almost 5, but also that Rubio has been steadily gaining ground on Hillary – he leads 12 of the last 14 polls (Hillary leads by 1 in the other two) and 15 of the last 20. Trump, by contrast, is flat-out lying when he claims to lead Hillary in every poll – out of 45 polls in RCP’s database, Trump leads Hillary in only 5, is tied in 2 others, and trails in 38 of them. Of 18 polls taken since late November, 9 show Hillary at 48 or higher; only two show Trump at 47% and one at 46. The only real variation is that some polls show more undecideds than others.
HuffPost Pollster collects a slightly different but substantially overlapping sample of polls – we can debate which is a better resource, as the HuffPost people let in a wider range of fringe-y partisan and online polls, but you should not just automatically disregard their data just because HuffPo is a left-wing website; it’s the same general world of polling. Anyway, it reaches similar conclusions to RCP throughout this article, only with different numbers. HuffPost sees Trump as decisively behind Hillary right now:
Whereas HuffPo estimates Rubio would start the general election slightly ahead of Hillary, and it is worth noting that HuffPo has Rubio ahead of Hillary in 6 of the last 7 polls, but Hillary up 10 points with 20% undecided in the one Morning Consult poll going the other way:
Besides head-to-head matchups, an early polling indicator to watch is favorable vs unfavorable ratings with voters, or “fav/unfav”. The more unpopular a candidate is, the harder it is to come back with undecided voters.
RCP has Trump’s fav/unfav at 34.4 favorable, 57.8 unfavorable (-23.4), so far underwater that it would be nearly impossible to win. How bad are Trump’s numbers? Barack Obama at this point in 2008 was at 69.0 fav/23.0 unfav (+46). Mitt Romney was actually in positive territory when he lost in 2012 (49.4 fav/44.6 unfav, +4.8), and never saw his unfavorables go above 51% in any poll the entire 2012 cycle; Romney grew from +7.6 in late January to +11 by the end of the primaries before the general election onslaught ground him down. John McCain was at 52.3 fav/41.5 unfav (+10.8) on Election Day 2008, and was around +20 at this juncture. Trump would undoubtedly see his favorables driven down during the general election campaign, just as Romney’s and McCain’s were, but he’s starting in a hole far deeper than the worst day of either of their campaigns. By this measure, Trump makes McCain and Romney look like Ronald Reagan on steroids.
In fact, Trump’s backsliding has been going on a while already. Trump’s favorables were getting better for a few months after he got in the race, as a faction of Republicans warmed to him, but they have been deteriorating in polls taken in December and January (RCP doesn’t have more recent ones than that). HuffPost likewise has Trump 20 points underwater, and steadily so for months:
Gallup, in a January poll, found that Trump – if nominated – would have the most unfavorable rating of any nominee from either party in the quarter century since it started polling the question, unfavorables on par with George W. Bush at the lowest point of his presidency. 60% viewed Trump unfavorably in Gallup’s poll, worse than George H.W. Bush a month before he got 38% of the popular vote in 1992. If you are keeping score at home, 60% of the people disliking you makes it hard to get 50% of them to vote for you.
Rubio, in RCP, has seen much more volatile favorables with the general electorate, and is presently at 35.2 fav, 38.0 unfav (-2.8), a little underwater but with a whole lot of people not sure yet. HuffPost likewise shows a lot of Americans still not settled in the kind of negative views they already have of Trump:
On the whole, everybody’s favorables (except Bernie Sanders) are lower now than everyone’s in 2008, but Rubio still looks a lot healthier than Trump. And there are a lot of subsets of voters at the margins where this gives Rubio room to win votes in the general election that are already closed off to Trump – as I noted a few days ago, Rubio actually has positive favorables with Hispanics nationally, while Trump is hated and seen as offensive by over 70% of them. College-educated voters would likewise be closed off to Trump; a mid-February Quinnipiac poll had Rubio at 46-40 (+6) with college-educated voters nationwide, Trump at 29-66 (-37), and trailing Hillary 52-37 with college-educated voters, who even in the GOP primaries so far have been a majority of the electorate.
Nate Silver did a deeper dive on this in mid-January, noting that Gallup found Trump at +27 with Republicans (compared to Rubio at +46), -27 with independents (compared to +4 for Rubio) and -70 (!!!) with Democrats, compared to -27 for Rubio. Silver, using the HuffPost Pollster database of polls, generated an average over the whole November-mid-January period and found Trump the most unpopular GOP candidate with the general electorate:
In fact, the most recent Gallup poll released February 29, 2016 shows Trump’s net favorables with Republicans down to +15, half what it was a month ago and his lowest since Gallup started tracking in early August. The grind of the primary campaign had also dropped Ted Cruz’s favorables – but not Rubio’s, which stood at +34, a bit lower than his campaign-long average but the same as it was when the year began.
Compared to Rubio, that’s the picture of an extraordinarily vulnerable candidate, even before you consider the historic trends working against the Democrats. But replace Rubio with a candidate whose negatives are as deeply underwater as Trump, and suddenly Hillary looks pretty safe.
Head-to-Head State-By-State Polls
Ideally, since Presidential elections are conducted at the state rather than national level, we could supplement our national polling with robust state-by-state poll averages. That’s what we’ll have to work with as the general election warms up, but for now, only a handful of states have been polled for general election matchups, and most of those only once or twice, in some cases not since November or earlier (Missouri was last polled in August) and in one case (Pennsylvania) the most recent Rubio vs. Clinton matchup is from February, while the most recent Trump vs. Clinton matchup is from October. So take all these with a grain of salt.
That said, the evidence we have from the state-level polling is consistent with the national polling: Rubio runs well ahead of Trump, winning states Trump loses, keeping close in states where Trump gets blown out. Of the 13 ‘battleground’ states where we have polling for both candidates, Rubio runs better than Trump in 11 of them, eight of those by 5 points or better, three by double digits (including New Hampshire, where Trump beat Rubio by 24 in the primary – a reminder that primary pluralities and general election majorities are two radically different things). Oddly enough, one of the two states where Trump polls better than Rubio is Rubio’s home state of Florida, but then only by a single point; Georgia is the only state where Trump would run significantly better, although both would start at 48% of the vote. Trump would enter multiple battleground states polling below 40%, consistent with his massive unfavorables.
(Yes, I realize it’s exceptionally unlikely that the Q Colorado poll is correct that Hillary would lose Colorado by double digits to either Rubio or Trump. The point here is how they stack up in an apples-to-apples comparison in the same polling environment).
Trump claims he would put his home state of New York in play. We have one poll of New York, a Siena College poll from early February that shows Hillary up 57-32 on Trump, with Trump’s fav/unfav at 25/71 unfav (-46). Rubio would likewise lose New York to Hillary, at 54-37, and his unfavorables are 34/45 (-11), but Rubio isn’t the one running around claiming he would be competitive in a home state where 71% of the voters disapprove of him. You can safely repeat that for most any other blue state you like.
A key problem Trump faces is the threat that a lot of Republicans (myself included) simply won’t vote for him in November. This, too, is a hard thing to measure at this juncture – when emotions are running high in a primary, it’s not that unusual to see people swearing they won’t vote for the nominee, then coming home in the fall. On the other hand, as Silver reminds us, the high levels of party loyalty we have seen in the last four elections are something of a historical anomaly – it was a frequent occurrence from 1952 to 1996 to see nominees of both parties who lost 20% or more of their own party’s voters in November, either through aisle-crossing or to third party candidates (Ross Perot, George Wallace, John Anderson). Of course, a major reason for that is an ideologically polarized electorate, which Trump would alter given his long record as a big-government social liberal. But I’m getting ahead of myself. First the numbers – since we don’t have poll averages, we have to go poll by poll.
Bloomberg Super Tuesday Poll: A Feb. 22-24 online poll of GOP primary voters in the Southern Super Tuesday states (Alabama, Arkansas, Georgia, Oklahoma, Tennessee, Texas, and Virginia) only two of which (Virginia and Georgia) are typically competitive in a normal general election these days. Trump has a 60/37 (+23) fav/unfav in this poll, compared to 65/28 (+37) for Rubio, an overall 37-20-20 lead on Rubio and Cruz, and a 48-44 lead in a hypothetical two-way matchup with Rubio. But in a vaguely worded question, 20% say they would “never” vote for Trump, and another 14% aren’t sure (only 10% would never vote Rubio, although 20 are unsure, numbers much more in line with soft disapproval from voters who would mostly come home in November). When given a choice between Trump and Hillary, 27% would not vote for Trump (9 for Hillary, 14 for a third party, 4 would stay home) and another 6% were not sure. That means only two-thirds of GOP likely primary voters in a set of mostly deep-red states would prefer Trump to Hillary. That’s an enormous red flag.
Democracy Corps poll: A poll by a Democratic firm hunting for opportunities, so these are results after some push-polling to see what Democratic messages worked with a “likely voter survey of 800 Republican base voters was conducted online using a voter file sample. The results were weighted to match Democracy Corps’ national likely voter data set for self-identifying Republicans and Republican-leaning independents who will vote in the Republican primaries or caucuses.” The poll shows that 20% of the Republican base could be induced to desert Trump in the general election, and indeed that self-identified moderates (even though they’re the group containing the most Trump supporters) would give just 60% of their votes to Trump against Hillary compared to 83% for Rubio:
The survey found that Evangelicals were not likely to desert Trump, but Tea Partiers, Moderates and Observant Catholics (the other three segments in the poll) were, depending on the anti-Trump messages tested – “[t]he strongest attacks on Trump charge that he is an ego-maniac who cares more about himself than the country, that he is very disrespectful towards women, and that he is a threat to national security and should not have control of our nuclear weapons.”
CNN/ORC national GOP primary voter poll: Conducted Feb. 24-27 – of 1,001 adults including 920 registered voters interviewed, the poll found 306 Republicans and 121 independents who lean Republican, and it was absurdly favorable to Trump, showing him with 49% of the vote and nobody else over 20. And yet, even in a spectacularly good poll for Trump, 48% of those not supporting him in the primary said they were likely to do so in the general election, including 35% certain they would not; only 25% of his non-supporters were sure they’d vote Trump in November. Rubio, by contrast, would find only 29% of his primary non-voters likely bailing on him in November, only 12% sure to do so. While the “won’t vote for them” contingents in this poll are pretty similar when you adjust for the baked-in support of their backers, this is a pretty alarming sign if you think Trump’s actual support is a lot less than 49%.
Fox News national poll, Feb. 15-18: 40% across both parties said the candidate they would most dread watching the next four years was Trump, 31% Hillary, 2% Rubio. Trump gets 79% of Republicans and 39% of independents, and 9% of Democrats against Hillary; Rubio gets 88% of Republicans, 48% of independents, and 12% of Democrats. 26% of Republicans and 57% of independents would be “not at all satisfied” if Trump won the presidency; we don’t have results for Rubio, but they asked the same question about Jeb! Bush and he did better than Trump, 22% of Republicans and 49% of independents. 67% of all voters in the poll said Trump lacked the temperament to be President, compared to 55% who said Hillary lacked the integrity for the job and 46% who said Rubio lacked the toughness.
Elon Poll of North Carolina: Feb. 15-19. Trump gets 80% of Republican votes and 48% of independents in head-to-head matchups with Hillary; Rubio gets 92% of Republicans and 55% of independents. Among voters of all parties, 38.7% said Trump was the worst candidate in either party running; 28.8% said Hillary; 1.3% said Rubio.
That’s only a sampling of polls, none of them all that conclusive, but every piece of data we have keeps pointing in the same direction.
Could The Polls Be Wrong?
So those are the numbers, such as we have. Now for the reasons why they may be right or wrong.
As I noted, polls this early do have some drawbacks, such as the fact that the candidates are engaged in divisive primary battles within their own parties (following which there’s sometimes a bit of a coming-together effect) and, on the other hand, that the candidates have not started to take fire from the other side. Trump fans, Trump apologists and Trump defeatists point to the fact that Trump was able to pull off an unprecedented reversal of his stratospheric prior unfavorables among Republicans after he joined the race. But of course, he did that in large part by (1) joining the party and (2) demonizing a bunch of targets disliked by many Republicans. Trump didn’t become more likable, he was viewed more favorably by rallying people on one side to tribalist dislike of people on the other. But that’s a trick you can’t pull in the general election; the very nature of “us against them” tribalism is that there has to be a “them,” and they vote too.
Will Trump’s unfavorables improve? Some will rally to the nominee, but that’s also true of Rubio and true of Hillary. It’s probably most true of Hillary, as some of her unfavorability right now is Bernie Sanders voters who think she’s not liberal enough, but have voted for plenty of non-socialists in the past and will come around when the alternative is a Republican. The same would largely happen for Rubio – some hardliners on immigration won’t be reconciled, but his overall profile is pretty conventional in terms of a Republican drawing contrasts to the Democrat (in fact, by most any measure he’d be the most conservative nominee since Reagan), and outside of immigration, it’s hard to see any issue where Rubio would lose a lot of people who vote for Republicans on a regular basis. It’s debatable how big the unreconcilables are on immigration (they hated McCain, but he only got 1.5% less of the vote than immigration-hardliner Romney despite running in much more unfavorable year; on the other hand, the issue is obviously more polarizing right now). But as I’ll discuss more below, Trump has far deeper problems that go beyond people liking some other candidate better.
And remember: Hillary’s negatives are very well known, although she has not really faced a sustained negative ad barrage of late. So, her opponent can put a bit of hurt on her, but Hillary’s standing with the voters is mostly baked into the cake already. Her chief problem is re-creating the supercharged turnout of non-white voters that propelled Obama to victory in 2012, while maintaining Obama’s extraordinary levels of support among those voters. Against Rubio, that kind of enthusiasm will be hard to come by, and at least some marginal deterioration of the Democrats’ advantage with Hispanic voters and young voters would be expected against a young, Hispanic nominee. Against Trump, however, his identification with white supremacists, hatred of Mexicans and Muslims, and all sorts of other racially polarized stances – plus the fact that he’s even older than Hillary – make him the absolute perfect candidate to drive up Democratic turnout while depressing Republican turnout.
And bear in mind: Rubio has been softened up so far by tens of millions of dollars of negative ad spending from Right to Rise. The attacks are not the same ones that would be used by the Democrats, but they mean that his polling right now already includes a substantial amount of the hits over things like his Senate attendance record and finances. Yet Rubio is also unknown enough with enough people that he could still win over more converts (he has repeatedly proven himself a strong closer in the primaries, winning late deciders). Trump, by contrast, while he’s a very known quantity who would have difficulty fixing his existing negatives, has been almost totally untouched by negative ads until the past week, so his standing in the general election polls right now – abysmal as it is – is likely to be closer to his ceiling than his real floor.
Trump has also had the advantage of running against an all-male field that is less directly able to exploit his notoriously disrespectful and boorish treatment of women. Hillary Clinton, who is running mainly on her gender and whose greatest asset is the sympathy she engenders for putting up with Bill’s infidelities, couldn’t ask for a more perfect foil than a thrice-married serial adulterer.
And that is before we get into Trump’s notoriously brittle temperament; we’ve seen him rattle badly at setbacks (losing Iowa) and provocations large (Rubio and Cruz going hard and personal at him in debates) and small (Megyn Kelly asking him questions). The one thing that keeps him out of full meltdown is his ability to recite his strong standing in the polls and his wins in primaries. But in a general election where he starts out behind, he will have that thrown constantly in his face instead.
It is not at all difficult to see Trump as the kind of candidate that nobody would admit having supported in the primaries by Labor Day. Barring some totally unforeseen scandal – a built-in risk with every candidate – it is all but impossible to see that happen to Rubio.
Could Trump Rally The #NeverTrump Faction?
Moving to the specific problem of Trump’s issues with the GOP base, we can clearly see that he would face challenges vastly different from those that Rubio would face. How would he appeal to the lifelong party faithful who are desparing for the party and ready to search for the exits if he’s the nominee? It’s no answer just to say that people will suck it up and vote even if we nominated the proverbial ham sandwich, as we have plenty of experience losing elections when not enough people turned out to vote for the ham sandwich.
To start with, a significant part of the #NeverTrump vote – not just the online commentators but the ordinary rank-and-file Republicans – consists of people who just think the man is dangerously unfit to be President, and there is only so much a man with near-100% name recognition and off-the-charts media saturation can do to solve that.
Partisans will forgive a lot if they see major issues at stake. But Trump has nothing at all to offer Republicans on the issues. “But the Supreme Court!” is traditionally the strongest difference between the parties, and with Justice Scalia’s death it’s a vital issue, but Trump has not given Court-minded conservatives even the slightest reason to believe he would appoint better or different judges than Hillary. He certainly has a long record of calling himself pro-choice well into his sixties (even to the point of publicly endorsing partial-birth abortion in 1999), of touting his pro-abortion federal judge sister, of welcoming pro-abortion Scott Brown as a possible VP candidate, and of defending federal funding for Planned Parenthood repeatedly during these primaries. He supported the assault weapons ban and same-sex marriage.
His record on size-of-government issues is even worse than on social issues; Trump has talked repeatedly about favoring something like single-payer healthcare and the Obamacare mandate, he has an authoritarian streak a mile wide, he’s practically the living embodiment of crony capitalism, he loves bailouts and eminent domain, he has campaigned for big-governmnent protectionism and against entitlement reform, and he’s previously proposed colossal tax hikes. The man doesn’t have a limited-government bone in his body.
What this means is, Trump in the general election will have to spend a vast amount of his time and effort just trying to get Republican voter turnout up to McCain/Romney levels, or maybe even Bob Dole levels. He will have to rely on surrogates yelling at people to show party loyalty solely for the sake of party loyalty, on behalf of a guy who was bankrolling the Democrats well into the Obama Administration. And yes, he will depend on the pure tribalism of the GOP base disliking Hillary – but the GOP base hated Obama with the fire of a thousand suns, and that still got us 46-47% in the last two elections. Hatred of the Clintons got Dole even less than that. And if the election comes down to a referendum on who is hated more, that’s a battle Donald Trump can’t win with anyone, even Hillary.
I’ve discussed here the hard numbers and soft factors that go into evaluating Trump as a general election candidate. But it’s also worth remembering that he’s a completely untraditional candidate who would face structural obstacles unique to Trump.
First, even among elected officials who won’t go full #NeverTrump (as Ben Sasse has), a lot of the party is already planning to run away from him, either quietly or openly. The press and the Democrats can make a lot of hay out of the occasional defector (think of how hard John Kerry leaned on Eric Shinseki), but here you could have dozens of Republican Senate, House and Governor candidates running ads distancing themselves from Trump. And who will speak on his behalf as a surrogate? Chris Christie, within days of endorsing Trump, is already bailing out on media appearances for Trump, refusing to defend his stances on issues, and refusing to answer questions about him in press conferences. Sarah Palin was quickly dispatched after an incoherent attempt to talk up The Donald. His chief press spokesperson, Katrina Pierson, is a barely literate conspiracy theorist. And of course, large parts of the conservative message apparatus is firmly set against Trump and could be backing a third party or openly rooting for him to lose. Trump’s entire message right now is that his victory is inevitable, but in a general election he would actually need to persuade people, and he’d have to do it all by himself.
Second, where will Trump get money? General election campaigns are notoriously expensive billion-dollar affairs. But Trump is too cheap to self-finance more than a fraction of his campaign, and he has scarcely raised any money for his primary campaign and can’t have much of a mailing list built. Two of the groups he has alienated the most are the GOP’s large donor class, who find him and his agenda horrible, and its small grassroots donor base, which consists of the sort of people most likely to be found in the #NeverTrump faction because they actually believe in things. His best supporters are economically downscale people who think Trump is self-financing and aren’t going to shell out their meager disposable income to help out a billionaire. Even people who might hold their noses and vote Trump against Hillary in the privacy of the voting booth would blanch at writing him checks. Trump’s only real fundraising asset would be his reputation for vindictiveness, but against the Clintons, he’d be facing masters of that art, and big-dollar donors are nothing if not able to read the polls.
Trump has managed the feat of overcoming all this in the primary by means of getting vastly more free media coverage than his divided and lesser-known opponents. But that trick can’t be repeated in a general election where there are only two major-party candidates and the other one has been just as famous for almost as long as Trump.
Third, who will organize and get out the vote for Trump? After his GOTV failure in Iowa, he has had more success at getting his voters to show up, but the number of people you have to turn out for a general election is vastly, vastly larger. Trump shuns traditional political tools (like conducting internal polling), and even if he adds those, a candidate needs a lot of help. Where will it come from? Establishment organs like the RNC and NRA will probably help, although it seems unlikely that a Trump nomination would be followed by anything other than a financial crisis in RNC fundraising. Downstream, though, GOTV relies on grassroots activists pounding the pavement as well as outside groups ranging from the Koch brothers’ network to issue groups like pro-lifers, all of whom will be ranging from dispirited to open revolt at Trump. The Koch network is not an asset to write off lightly:
If Trump becomes the nominee and he faces self-declared socialist Bernie Sanders in November, the senior Koch official explains, members of the donor network are likely to hold their noses and back Trump’s candidacy. But there’s another scenario that could prove far more controversial and possibly damaging for the network: a Trump-versus-Clinton matchup. There is absolutely no love between the Clintons and the Kochs, whose company experienced one of the most traumatic periods in its history as it fought off regulators during Bill Clinton’s presidency. But, so strong is the dislike for Trump within the Koch network, that a Clinton-Trump race is a tough call. “I could see the network not participating in the presidential election at all,” says the senior Koch official.
Of course, Trump would probably benefit from increased white working class turnout. But that alone is not enough to overcome the 1-2 punch of the Democrats’ huge advantages with non-white and young voters and the likely deterioration of the rest of the GOP base vote. Indeed, Sean Trende’s original “missing white voters” analysis assumed that adding these voters would not by itself be enough to reverse the outcome in 2012, let alone make good substantial losses among Republicans and independents who voted for Romney.
Rubio would not face a single one of these hurdles: aside from the most hard-shell anti-immigration zealots, he’d have the full and complete backing of basically the entire party apparatus, the large and small donor networks, the activists…with so much at stake in this election, a candidate who is simply “generic Republican” would have vastly more to work with than Trump, and that’s even before you consider Rubio’s strengths as an individual candidate. True, Rubio or another more conventional Republican (even Cruz) would face some losses among Trump’s base, who would resent Trump’s defeat. But while Trump’s problems involve pervasive distrust on the issues combined with major personality and temperament issues personal to Trump, Rubio would have time to woo back some of these voters with general election policy initiatives, plus adding on the current Trump voters who are just attracted to him because he seems like a winner.
Could Trump beat Hillary in November? It’s possible – she could have a stroke, or be indicted, or a terrorist attack could level Los Angeles, or we could face a sudden financial crisis, or some other totally unpredictable event that would hand the election to any Republican. And of course, it is always possible that the world has changed so much that 70 million people will vote for Trump for no particular reason at all – that all the polling will be wrong, that demographics won’t matter, and organization won’t matter, and media coverage won’t matter. You never know!
But even in light of everything that has brought us to this pass, there is no rational reason whatsoever to think Donald Trump will be able to beat Hillary Clinton, and every reason to think nominating him would throw away an election Republicans otherwise would have a very strong chance to win.
Don’t do it. Don’t surrender. A vote for Trump is a vote to lose to Hillary. A vote for Rubio (or for Cruz) is a vote to defeat her. After 8 years of battles to get ourselves to the point where we might actually be able to beat the Democrats and do something with the victory, let’s not just choose to roll over lose on purpose.