Those (like me) who are addicted to polls have choices on where to go if you want to panic, relax, or start partying like it’s November 9th. We should never underestimate the ability of half of the country to do something blisteringly stupid. After all, in 2000 after 8 years of peace and prosperity, about half of Americans thought “Let’s try something different!” The polls show support for Hillary Clinton and Donald Trump combined to be farther below 100% than is typical this late in the race, so there is some uncertainty as to what will happen on election day. Donald Trump has been careful not to say anything more horrifying than usual this week, and that might be long enough for a few waffling people to think dropping a bomb on American democracy might be a good idea. The reason we should be cautiously confident now is not so much because of the polls or whose statistical model we believe, but because of hard data from early voting, particularly in key states.
Let’s start with the famed “blue firewall”, states that aren’t really being contested and everyone agrees will vote for Clinton. Here they are in approximate decreasing order of blueness: DC, MD, VT, HI, MA, CA, NY, RI, WA, IL, DE, NJ, CT, OR, ME (3 EV), NM, MN, WI, VA, MI (239 EV)
Next is the light blue firewall: states that are being contested but with all polls showing Clinton in the lead: CO, PA (cumulative 268 EV)
At this point, most people add New Hampshire to get to 272 EV, but several recent polls (though not necessarily good ones) have the race tied or Trump ahead. New Hampshire is not very predictable, and remember they were among those who decided to throw peace and prosperity over the rail in 2000 and try Bush instead. I’d say Clinton is still favored, but not by much. The next brick in the blue wall is actually…
Nevada (cumulative 274 EV). It’s not the polls we should be watching here so much as people going to the polls. The ultimate source for data and analysis is Jon Ralston, who is a god in his realm of Nevada elections. Yesterday was a spectacular finish to early voting, pushing Clinton past Obama's early voting margin. It is inspiring to see 1,000 people lined up to vote at Cardena's Mexican grocery, and they are not waiting hours in line to vote for Donald Trump. Democratic early vote turnout is now slightly better than in 2012, when Obama won the state by 7 points. It gets better in the polling data of people who voted early. PPP found that among those who have already voted, Clinton is ahead by 11 points, while CNN says Clinton is up by an unspecified narrow margin among those who have voted. Ralston has a strong record of predictions (including Harry Reid’s win when all the polls said he would lose), and here is what he thinks of Republican prospects after last night:
Donald Trump will be in Reno on Saturday, but the Republicans almost certainly lost Nevada on Friday.
Trump's path was nearly impossible, as I have been telling you, before what happened in Clark County on Friday. But now he needs a Miracle in Vegas on Election Day -- and a Buffalo Bills Super Bowl championship is more likely -- to turn this around. The ripple effect down the ticket probably will cost the Republicans Harry Reid's Senate seat, two GOP House seats and control of the Legislature.
The next likely Democratic state is… Florida (cumulative 303 EV). Steve Schale is the early voting data guru in Florida. The story there isn’t just that early voting is going well for Democrats, but Hispanic voters are turning out in huge numbers.
And it isn’t just that Hispanics are voting, it is the types of Hispanics who are voting. Here is one way to look at it: Right now, statewide, 16% of early voters are either first time Florida voters, or haven’t voted in any of the last three elections. Across party lines, 24% of all the Hispanic votes today come from these first-time voters. Among Hispanic Republicans, it is 14%, among Democrats, it goes up to 26%, and among Hispanic NPAs, a whopping 32% have no previous or recent voting history.
When you expand it out to voters who voted in one of the last three, which is what I define as “low propensity,” it goes up to 53% of Hispanic Democrats and 60% of Hispanic NPAs. That, my friends, is the definition of a surge.
So apparently Hispanic voters are highly motivated this year — gee I wonder why? Those low-propensity voters aren’t going to make it past alot of likely voter screens unless they are polled after having voted. CNN says that Clinton holds a wide lead and Quinnipiac has Clinton 6 points ahead among those who have already voted in Florida, even though until today Republicans had a small edge in early voter turnout. Florida can give us heartburn, but this is a very good sign.
North Carolina (cumulative 318 EV) was all about African-American turnout in 2008 and 2012. Early voting data shows African-American turnout behind where it was in 2012 (when Obama narrowly lost the state) but gradually catching up. Clinton will probably also get a little less than the near unanimous support among African-Americans that Obama got. Most of the polls have Clinton leading, however. PPP claims Clinton leads 59-40 among those who have already voted. It’s pretty hard to believe the margin would be that big, but it’s a good sign at least. This state is a test of whether Clinton gets alot of votes from college-educated white women. Increased Hispanic turnout will also help.
The next most likely state in Clinton’s column is New Hampshire (cumulative 322 EV) It could go either way but the polling average here is still leaning toward Clinton, and that’s about all we have to go on without campaign insider data.
The last state I give Clinton greater than even odds of winning is Arizona (cumulative 333 EV). Yes, really. We are seeing a surge of Hispanic voting across the country, expected to be 25% higher than in 2012. Data from Arizona is showing that although white voters are still overrepresented in the early vote, the percentage of Hispanics among early voters has almost doubled. CNN reports that Arizona early voters narrowly favor Hillary Clinton, even though the early voting population is whiter than the electorate was in 2012. Almost all the polls show Trump winning by a few points, but I think the wave of Hispanic voters are disproportionately “unlikely” (and difficult to poll to begin with) so we could be in for a surprise here.
I think Clinton will most likely fall just short in Maine’s 2nd CD, Ohio, Iowa, and Nebraska’s 2nd CD although her superior ground game could still carry the day. Georgia could theoretically go blue thanks to college educated whites and a surge of Hispanics, and Texas could be shockingly close for the same reasons, but none of these states will be relevant in the end because they won’t be the ones to bring Clinton past 270. Huge accolades are due to those working on the Clinton ground game, but particularly in Nevada and Florida, the latest bricks in the blue firewall which will carry the day.
So what’s the headline November 9th? Well, probably “Clinton Wins”. The subhead will be: “Don't Mess with Mexico!” After the 2012 election, the Republican “autopsy” report said that the Party had to improve its standing with Hispanic voters. Instead, Donald Trump kicked a sleeping giant by attacking Mexico and Mexicans and we will see the result in surprising Democratic performance from Florida to Nevada as Hispanics of all stripes turn out in huge numbers to deny him the presidency.