Anyone who knows me well knows that I’m a pretty big data geek. It comes from my years as a competitive debater and from an almost obsessive desire to be prepared for any possible right-wing argument that comes my way (which, given my job as an anti-racism educator/speaker/author) is pretty often. If anything, I over-rely on data, especially since I know the research from social psychology (still more data), which suggests that facts aren’t actually what persuade people politically. Still, having command of data, even if not that helpful for “converting the heathen” so to speak, still serves a function: it gives one the courage of one’s own convictions, which makes one a more effective educator, organizer, etc. Also, it helps you to chill a bit. When you know you’re right about something, you tend to relax, which is better for your own mental, emotional and even physical health.
I can still remember when I forced myself to read Dinesh D’Souza’s horrific book The End of Racism, because he and I were about to debate at Furman University, back in 1996. Before reading it, I was stressed, because I couldn’t help but wonder, “What if there’s something in there that I don’t know how to rebut?” The stress was debilitating. Then I read it, and, well, let’s just say there wasn’t anything to worry about. Once I knew that, I could exhale and felt instantly better. It didn’t mean I was going to be able to change Dinesh’s mind — if anything it has become even more ideologically deranged in the years since — and didn’t even necessarily mean I would change the minds of a single conservative in the Furman audience (though I did win the debate by show of hands afterward). But at least I felt better, and there’s something to be said for that.
Which then brings me to the second thing my friends know about me: namely, that even when I am confident about the facts surrounding certain issues, I am almost never confident about elections. In 2008 and 2012 I was absolutely convinced that Obama would lose, no matter what the polls said; no matter what Nate Silver said. I’ve simply been white and known white people for too long, I kept thinking to myself. Now, of course, in one way I was right: President Obama did lose white people, and not just white men, but also white women (53-46) and white younger voters (52-48). And not just in the South (though the numbers there certainly skewed the national averages), but even in many states that were overall blue. That said, folks of color turned out in large numbers and that, combined with the changing demographic makeup of the electorate, made the difference.
This time around, I was convinced last August that Trump was a threat, and many if not most of my friends on the left said I was being silly. Once again, I reminded them about the white people thing, and especially that for whites in the “anxious middle” and upper-working class, Trump would connect far better than Mitt Romney. They kept laughing, I kept mumbling something about white people (I do this a lot), and well, here we are. Needless to say, I’ve been in a pretty uptight way, worried about the election.
Until now.
See, for months I’d been looking at the Real Clear Politics interactive graphic, which allows you to play with the numbers a bit, adjusting the white, black, Latino/a and Asian vote, relative to their 2012 breakdowns. And even though most every reasonable manipulation produced the same result — there was just no way Trump could win — something about it was unsatisfying: namely, it really was only able to project popular vote. Yes, the map in the graphic would change state-by-state with certain manipulations, as if to mimic an electoral college outcome, but I knew it wasn’t precise. After all, projecting Clinton to win 80 percent of the Latino vote nationally, doesn’t mean she would get that same share in every state. Indeed, in 2012, in some important states, Romney did far better with Latinos than he did overall, and those small differences could be critical this time around. Running up big numbers in California or New York won’t change a thing electorally for Clinton.
But what finally made the difference and has me breathing quite a bit easier is a set of numbers provided by the Cook Political Report, which provides some detailed information on a state-by-state basis in terms of projected turnout by demographic sub-group, combined with information on how those sub-groups voted in 2012.
Let’s start with Florida.
In this table, the first column shows the 2012 percentage of the vote for Obama and Romney, by demographic group: College-educated whites (CEW), Non-college whites (XCW), Blacks (BLK), Latino/as (LAT) and Asian Pacific Islanders/Other (APA/O). The second column shows the projected 2016 share of the electorate in Florida for each demo group, according to Cook. The third column represents my own, worst-case scenario projections of how Clinton would do with each demo group. The final column is the percentage of the overall vote received by Clinton from each demo, if these projections are accurate
Cook Report PROJECTIONS + MY ESTIMATES — FLORIDA
| DEMO |
OBAMA-ROMNEY |
2016 PROJECTED ELECTORATE % |
2016 ESTIMATED VOTE % HRC |
TOTAL FOR HRC |
| CEW |
40-60 |
34.0 |
40.0 |
13.6 |
| XCW |
37-63 |
30.0 |
34.0 |
10.2 |
| BLK |
96-4 |
13.4 |
92.0 |
12.3 |
| LAT |
61-39 |
18.3 |
65.0 |
11.9 |
| APA/O |
61-39 |
4.4 |
65.0 |
2.9 |
|
|
|
|
50.9 |
As you can see, I am projecting that Clinton can hold serve, so to speak, when it comes to college educated whites, and repeat Obama’s numbers from 2012. This seems eminently reasonable to me, given the likely boost HRC will get from white women (who went for Romney last time but will likely split or even go slightly for her this time), and especially college-educated white women. My goal here is to see how she can do, even if she does not build at all on Obama’s share among college-educated whites, which seems to me a worst-case scenario.
Among non-college educated whites, I think it is reasonable to expect Trump to do better with this group than Romney did. In fact, I am projecting a full three point drop here for Clinton relative to Obama, giving Trump 66% of that group, compared to 63% for Romney. That may seem like too small a shift for some of you, given Trump’s polling numbers with such whites nationally. But actually, it is a pretty generous shift that I’m projecting for him — a gain of three points and 5%, roughly, in absolute terms.
Then, among black voters I am being extremely generous to Trump, projecting him to double Romney’s numbers from 4 to 8 percent of the total black vote; and among Latinos and Asians — despite Trump’s horrific anti-immigrant stances and comments — I am being quite favorable to Trump, projecting that he will only lose 4 points relative to how Romney did with these groups. I am giving Trump 35 percent of both Latino/as and Asian Americans here, which strikes me as very generous indeed.
And the fact is, even with these worst-case scenario projections, as you can see in the bottom right hand corner, Clinton wins Florida with about 51 percent of the vote. Not a lot of wiggle room, but likely enough to allow for her to win even if these numbers are off by a bit, and not pessimistic enough.
Although I doubt my vote projections could be much worse for Clinton than these, where things might get a bit sticky is with the projections by Cook about the expected share of the electorate that will be represented by each demo. As they explain on the page linked above, they are making these projections based on two things: 1) 2012 vote shares for each group, and 2) Census changes since 2012, in terms of persons in each group eligible to vote. In other words, they are assuming that turnout among each group will mirror 2012 and that each group’s newly eligible voters will vote at the same rates as those in 2012 did. Neither of these are ironclad guarantees. Black turnout was historically high in 2008 and 2012, at least in part because of Barack Obama; likewise, white turnout was actually lower than normal, and indeed below black turnout in 2012, percentage wise. Neither of those conditions are likely to obtain again. Although black turnout will not drop to pre-Obama levels — after all, there have been significant gains in registration and when people are registered they tend to vote in similar percentages, across racial lines — it will likely fall a bit. Similarly, white excitement over Trump, especially among the non-college educated, might push their share of the overall vote up a bit, relative to what Cook is predicting.
So let’s play with the numbers a bit, making slight tweaks to the projected vote shares of each demo to reflect this possibility. Let’s redo the column for voter percentages by demo, and then look at the total gained by Clinton for each new projection (basically, re-doing just columns 3-5)
Adjusted Florida ProjectionS
| DEMO |
‘16 elecTORATE % |
‘16 ESTIMATED VOTE % HRC |
TOTAL FOR HRC
|
| CEW |
34.0 |
40 |
13.6 |
| XCW |
31.5 |
34 |
10.7 |
| BLK |
12.5 |
92 |
11.5 |
| LAT |
18.0 |
65 |
11.7 |
| APA/O |
4.0 |
65 |
2.6 |
| TOTAL: |
|
|
50.1 |
Even in this scenario, in which I assume black turnout will drop enough to suppress the share of the vote that is black by nearly a full percentage point from the Cook projections, and that non-college white turnout will spike enough to drive that group’s share up 1.5 points above the Cook projections, Clinton still wins Florida in a squeaker. And again, that assumes Trump doubles his vote share among blacks, only drops a few points among Latinos and Asians and does 3 points better than Romney with non-college whites. In short, there is very little way to see Trump winning Florida, under any reasonable assumptions and no matter how hard the folks in The Villages might wish otherwise.
Now, let’s look at Ohio, combining the projected electorate shares per demo provided by Cook with my own worst-case scenario estimates in terms of Clinton’s vote percentages from each demo.
COOK REPORT PROJECTIONS + MY OWN ESTIMATES - OHIO
| DEMO |
OBAMA-ROMNEY |
2016 PROJECTED ELECTORATE % |
2016 ESTIMATED VOTE % HRC |
TOTAL FOR HRC |
| CEW |
42-58 |
36.0 |
42 |
15.12 |
| XCW |
44-56 |
42.0 |
40 |
16.80 |
| BLK |
97-3 |
15.0 |
94 |
14.10 |
| LAT |
58-42 |
3.5 |
63 |
2.21 |
| APA/O |
69-31 |
3.5 |
71 |
2.49 |
|
|
|
|
50.72 |
Here, even if Clinton loses 4 points off Obama’s totals for non-college educated whites (a 10 percent drop in real terms), and even if Trump doubles Romney’s black percentage (from 3 to 6 percent of the total black vote), and even if Trump only does 5 points worse with Latinos than Romney, thereby managing to still capture 37 percent of the Latino/a vote, and only 2 points worse with Asians, she will still win so long as she maintains Obama’s share of college-educated whites. If she were to nudge up slightly with college-educated whites, or simply hold Obama’s numbers with African Americans she would win even more convincingly.
Now, sure enough, if the Cook projections for each demographic (in terms of what share of the state’s electorate they will likely be) are off, things could get dicey here. For instance, if black turnout falls to any appreciable degree from 2012, or if non-college educated white turnout surges, Trump could win the state; but even then, only if every single one of the above worst-case scenario projections play out. It is very hard for me to imagine that happening. It seems reasonable to assume that HRC will win at least 65 percent of the Latino/a vote, and that even if Trump does better with black voters than Romney a doubling isn’t likely. A 50 percent jump (from 3 to 4.5 percent of the total? Sure, but not a doubling). So long as everything doesn’t go to shit in Ohio, it appears very hard for Trump to carry the state.
Now let’s look at Pennsylvania.*
COOK REPORT PROJECTIONS + MY OWN ESTIMATES - PA
| DEMO |
2012 OBAMA-ROMNEY |
2016 PROJECTED ELECTORATE % |
2016 ESTIMATED VOTE % — HRC |
TOTAL HRC |
| CEW |
43.5-56.5 |
44.2 |
43.5 |
19.23 |
| XCW |
45.5-55.5 |
33.1 |
40.0 |
13.24 |
| BLK |
94.5-5.5 |
13.2 |
89.0 |
11.75 |
| LAT |
82.5-17.5 |
7.2 |
80.0 |
5.76 |
| APA/O |
69.5-30.5 |
2.3 |
70.0 |
1.61 |
|
|
|
|
51.59 |
* NOTE: I have made a slight adjustment here in a way favorable to Trump. Cook’s projected demo breakdowns came to slightly more than 100 percent (over by one-tenth of a point). So as to round down to a flat 100 percent I revised the estimated black percentage of the vote from 13.4 to 13.2 and revised non-college whites upward from 33.0 to 33.1 just to solidify the likely “worst-case” scenario nature of this table.
As you can see, even if Clinton collapses among non-college whites by a full 5.5 points from 45.5 down to only 40 percent of this group; and even if Trump doubles the share of black voters voting GOP from 5.5 to 11 percent; and even if Hillary does 2.5 points worse with Latinos than Obama (meaning that Trump actually gains 2.5 points among them, relative to Romney); and even if Clinton basically only holds the Asian vote where it was four years ago, so long as she holds steady with college educated whites she wins PA. Honestly, I cannot fathom Trump doing better than Romney with Latinos there, nor holding steady with Asians, nor doubling the share of blacks. So the above scenario seems absolute worst-case to me. Seriously, if things actually are worse than this for Clinton, then Trump is likely riding some wave to victory nationwide, and we’d best start building our underground bunkers.
You can tease out these scenarios for the other battleground states as well if you’d like, by using the data at the Cook site. I’ve done it, and although I won’t bore you with the details here, suffice it to say that from what I can tell, the worst, worst case scenario is that:
HRC wins Nevada, Colorado, Florida, Michigan, Pennsylvania, Virginia and Wisconsin;
Trump wins Arizona, Georgia, Iowa, Missouri, Ohio, New Hampshire, North Carolina.
In which case, the electoral college vote is 304 to 234.
Even if we give Florida to Trump, which given the above data looks like an unlikely outcome, it would still leave HRC winning 275-263.
Obviously, in the end it’s all about turnout and mobilization, and anything can happen. But given the edge Clinton has in terms of mobilization, GOTV etc., and given Trump’s reliance on his personality and name recognition to somehow drive people to the polls for him even without that kind of machine in place, it seems unlikely that she would falter on that front.
And finally, let me be clear, I do not mean to suggest by this diary that I am in the mood to rejoice. After all, as a leftist, far more committed to defeating Trump(ism) than electing Clinton, I am not breaking out champagne bottles over the prospects of her presidency. But as someone who also understands the principle of harm reduction, I take my victories where I can find them. I will consider this to be one...and then it’s back to work, building the progressive-left-radical movement for change, without illusion or pretense that real and transformative change is possible under the current system.
