Without a doubt, Donald Trump was the political story of 2015, showing more staying power than just about every analyst predicted. Back in mid-September (when I began pondering probabilities), I thought Trump had a 20% chance on the nomination and now give him a 30.8% (4/13) shot (chart below) with a growing risk I am too low. A Trump nomination scares the establishment in part because they believe he cannot beat Hillary Clinton, and most public polling shows him faring worse than either Marco Rubio or Ted Cruz. This begs the question: should he win the nomination, can Trump actually win the general election? While he would start out as a clear underdog, I do see a viable path for Trump to get the needed 270 electoral votes and would peg the odds in the 10-20% range.
Now, two major caveats: first, any chance for Trump to get to 270 assumes that a well-funded moderate republican does not run as a third party candidate. Such a candidate, perhaps aiming to help republicans hold the House and Senate by keeping Clinton’s share of the vote in the low 50%’s, would stop Trump from having a realistic shot, a republican Ralph Nader if you will. This analysis assumes a 2-candidate race. Second, events can overwhelm political candidates; for instance, the failure of Lehman Brothers ended any chance John McCain had of being President in 2008. If on Election Day Barack Obama sports a 30% approval rating, democrats will not win a third term, almost regardless of who Republicans nominate. Conversely in Obama sits at 65%, Clinton will almost surely be President. Presumably, concerns about a Trump candidacy reside not in these “tail scenarios” but assume the dynamics stay somewhat similar, ie Obama approval in the 43-53% range. As such, that is the focus of this analysis. Arguing Trump could win if the Obama Administration totally implodes doesn’t really address the question after all.
Recognizing his flaws (and strengths) as a candidate and the fact he would be an underdog, Trump would need to pursue an entirely different electoral strategy than the “generic” Republican, which would manifest in two ways: the states he focuses on and his VP. Conventional wisdom is that Republicans should not have another all-white male ticket, needing to perform better with women or Hispanics. Along those lines, the path to 270 consists of carrying the Mitt Romney states, adding Florida, Ohio, Virginia, and then one of: New Hampshire, Colorado, Iowa, Nevada, or Pennsylvania. That makes sense and is the most plausible strategy for most Republicans, who carry those states by turning out a few more Republican voters, slightly improving on Romney’s share of the white vote by lessening the gender gap, and doing upwards of 10% better among Hispanics. Thread the needle appropriately, and the party can squeak past 270.
However given his incendiary rhetoric, Trump is unlikely to do better than Romney among Hispanics, and he may also struggle to close the gender gap. In fact, his rhetoric and call for mass-deportations could make it easier for Clinton to turn out even more Hispanic votes than Obama was able to. At the same time (despite being perhaps 40x wealthier), Trump seems to connect with middle and working class white voters far better than Romney, given his focus on returning manufacturing jobs and blaming China, Mexico, wealthy donors, and stupid trade deals for our economic ills. More than anything, Trump is running as a populist. Remember, the country has shed about 5 million manufacturing jobs in the past 20 years, so this message can resonate in particular areas strongly. Doubling down on this rhetoric could help Trump increase white turnout and improve upon Romney’s share of the white vote in some states, meaning he could have a better chance in Pennsylvania than Florida.
Now, that sentence should make clear how poorly positioned Trump would be. Despite getting crushed nationally, Romney nearly held on to Florida (losing by less than 1%) while Pennsylvania is a perpetual “fool’s gold” for republicans. Polls often show a somewhat tight race, but its electorate is just not very elastic (i.e. there are few voters who switch from party to party), meaning republicans almost inevitably come up short. Unlike say New Hampshire where a large swath of voters are independent-thinking, making it possible for a GOP win, there are so few winnable true independent PA voters that is a monumental undertaking for the GOP to get to 50.01%. If you are betting on PA over FL, you are starting at a major disadvantage as a Republican candidate. Below I have placed the two maps that are the most probable Trump paths to victory (note Romney Plus includes all Romney states while Upper Midwest includes all Romney states less North Carolina) as well as each state “in play” by the share of their 2012 electorate that was white and the share of the white vote Romney and Obama received.
Upper Midwest Path (275 electoral votes):
Romney Plus Path (274 electoral votes):
New Hampshire (52-46.4% Obama): 93% white, (51-47% Obama)
Iowa (52-46.2% Obama): 93% white, (51-47% Obama)
Minnesota (52.7-45% Obama): 87% white, (49-48% Romney)
Wisconsin (52.8-45.9% Obama): 86% white, (51-48% Romney)
Ohio (50.7-47.7% Obama): 79% white, (57-41% Romney)
Pennsylvania (52-46.6% Obama: 78% white, (57-42% Romney)
Colorado (51.5-46.1% Obama): 78% white, (54-44% Romney)
Michigan (54.2-44.7% Obama): 77% white, (55-44% Romney)
Virginia (51.2-47.3% Obama): 70% white, (61-37% Romney)
North Carolina (50.4-48.4% Romney): 70% white, (68-31% Romney)
Florida (50-49.1% Obama): 67% white, (61-37% Romney)
Nevada (52.4-45.7% Obama): 64% white, (56-43% Romney)
Now, there are some interesting takeaways here. The states the generic Republican would focus on have on average a lower share of the white vote (NC, FL, VA for instance), which follows from their desire to increase Hispanic support. Longer-term, this is a strategic necessity for the party as the white share of the vote declines every day. However for Trump, the focus lies on the Upper Midwest and Rust Belt where weakness among Hispanics is a near irrelevancy.
It is also unwise to think of “the white vote” as monolithic, just as it is wrong to lump all Hispanics or women together. Trump is unlikely to do demonstrably better among upper class whites or those in the service sector than Romney; instead, his focus would be on the working class and those in manufacturing. His economic rhetoric is directed at these individuals, and much of his support in primary polls comes from these voters. These voters are far more prevalent in PA, MI, OH than in VA, CO, or FL. In other words, there is more room to “grow” the white share in the upper Midwest. In states like FL (61%), VA (61%), NC (68%), there is a case to be made that Romney did nearly as well as possible among white voters while leaving some on the table in MI, WI, IA. For instance, Romney’s strong performance with whites and the risk of higher minority turnout due to Trump mean North Carolina could plausibly fall to the Democrats while stronger turnout and support in the Western Half of the state could put Pennsylvania in the GOP column.
Additionally in the “near GOP” states like Florida, Romney did a better job closing the gender gap, winning 58% of white women whereas he on average he fared worse in the Upper Midwest (losing 38-60% among women in Minnesota for instance). If the gender gap widens nationally, that likely hurts disproportionately in FL vs. MN where there are fewer women left to lose. It also means Trump can’t pick up Virginia, especially with Northern VA increasingly being the place where Republican dreams go to die, but his anti-trade rhetoric could have appeal in Michigan, which has been ground zero in lost manufacturing jobs.
The path to 270 for Trump relies on recognizing he will likely do worse than Romney with Hispanic and women voters, making states like CO, FL, and VA very difficult, but has room to grow among working class whites in PA, MI, MN, and WI. That means focusing on those states and continuing an economic message focused on lost manufacturing jobs and offshoring. Clearly, that message has resonated with GOP-leaning voters in that cohort, irrespective of whether Trump is actually offering solutions that will fix the underlying problems (another issue entirely). If Trump can hold serve on African Americans (ie keep the pathetic ~4% Romney got), lose Hispanics by another 5-10%, but pick up another 5% of the white vote (ie go from 49% to 54% in Minnesota) by focusing on a message that clearly has some resonance with disenchanted members of the middle and working class, he could pull off a Midwest sweep and sneak just past 270. Is it easy? No, relying on PA and MN (30 EVs) to offset FL (29 EVs) is far from ideal, but it isn’t impossible. In fact, one could argue that Cruz doesn’t have a much better chance than Trump in the General because while Cruz should be able to hold Romney’s Hispanics, he doesn’t (at this point at least) seem to have the same appeal to Americans earning under 50k.
Last, Trump needs to pick a VP in-line with this electoral strategy. Given his outsider message, he also probably couldn’t pick an established politician or DC insider. There would be a logic to say, Governor Mike Huckabee, whose evangelical credentials shore up the right flank and who also has a populist message. However, when you are the clear underdog, it calls for going for the Hail Mary. Going for a 60 yard TD is never the percentage play on a typical down but when on the 40-yd line with 3 seconds on the clock, a big pass is the only shot at winning. With this “play to win” mindset, I would argue General David Petraeus is the wisest VP choice for Trump.
Yes, Petraeus comes with substantial risk (it wouldn’t be a Hail Mary otherwise); namely, providing classified information to his girlfriend over email. However, Clinton is uniquely poorly positioned to make this attack given her far worse email indiscretions. Putting Petraeus on the ticket dares Democrats to get into a fight about emails, and it is almost certainly a fight that will not go well for them. With Hillary as the nominee, his biggest liability is greatly diminished.
On the positive side, his military bonafides would go a long way to give Americans confidence Trump won’t be “trigger-happy” in military affairs and will be getting real advice. That could make a Trump Presidency palatable to many more Americans. Last, as CIA Director on 9/11/12, Petraeus knows what happened in Benghazi and what the Administration knew and when it knew it. He also knows how many security warnings were given prior to then that were ignored. In the past, there have been some incendiary claims that Petraeus wanted the talking points to include more references to terrorism but was silenced by the Administration, which preferred the since debunked video explanation. If these claims are true, could you imagine him beginning his VP acceptance speech at the GOP convention in Cleveland by saying, “Hillary Clinton and President Obama lied to you to win an election”? If that happened (no sure thing), the whole campaign would be shaken. Even without any Benghazi re-litigation, Petraeus, by giving the Trump ticket military credibility, would be a wise VP pick.
Is Trump well positioned to beat Hillary Clinton? No, he isn’t, given the easiest states to pick up, like FL, would be uniquely challenging for Trump assuming likely problems with Hispanic voters materialize. However with much of his campaign focused on the working and middle class, who have been left behind under Presidents Obama, Bush, and to a degree even Clinton, Trump could have a path through the Upper Midwest by doubling down on populist economics and attacks on bad trade deals. It would be a different campaign with different battlegrounds with the potential of being eminently entertaining. A Trump nomination doesn’t ensure a GOP washout, it just means pursuing an Upper Midwest strategy, with maybe a 15% likelihood of success.
Ideal? No. Panic-inducing? Not really, especially if he picks someone like Petraeus to run with. Trump has proven the prognosticators and GOP establishment wrong thus far. Just maybe, he can continue to do so.