Thursday, February 20, 2020

Math for Non-Mathematicians

Okay, this blog is not supposed to be just about politics and one of these days--hopefully soon--I am going to post something having nothing to do with politics.

But yeah, that won't be today.

Today the topic is the idea that many people seem to have about Donald Trump being a formidable candidate who will be difficult to defeat in the general election.  For some, this is a reason to support candidates in the Democratic primary whose views do not align with their own, or even diverge quite dramatically, just because they perceive that candidate as more capable of beating Trump.  For others, the expectation of Trump's reelection is a foregone conclusion that has them making plans for everything from coping with prolonged DT-induced depression to remote years or even full-blown emigration.

But how formidable is Trump, really?

One fact that is commonly used to support the belief in Trump's strength as an incumbent POTUS is that the 2016 presidential election featured "record turnout." That is, over 138 million people voted that year, topping the previous record of 132 million voters from the 2008 presidential election.  Of course, only 46.1% of those voters actually voted for Trump--which was not even a plurality of the votes cast.  And really it was closer to 136 million, since an amazing 2.4 million voters went to the polls and voted in other races but left the line for president blank.

Nonetheless, the reasoning behind the argument from record turnout is that if Trump still wins even when the greatest conceivable number of voters show up, then you can't realistically beat Trump just by turning out more of your own voters.  Rather, you'd need to convince some of Trump's supporters to switch sides--which, yeah.

Okay, but here's the thing about the supposed "record turnout" in 2016.  While the total number of voters was a record, it was an average presidential election year by percentage of the electorate.

Image from:
The 60.1% of voters who cast ballots in 2016 was about average compared with the preceding three presidential elections and a full 1.5 percentage points below the 2008 turnout.  Since the end of WWII there have been six elections with higher turnout percentages than 2016--with the highest being 63.8% turnout for the 1960 race between JFK and Richard Nixon.

This means it's fair to assume that the celebrated "record turnout" in 2016 was simply a function of a greater U.S. population and not Donald Trump's historic appeal to voters.  I could probably support this point with a link to inauguration crowd size photos, but maybe I shouldn't.  Yeah, of course I should.

Accordingly, it is totally possible to defeat Trump by turning out more voters.  The difference between turning out 2016-level turnout and 2008-level turnout is now well over 2 million votes.  So if Trump maxed his turnout in 2016, then a Dem who can match Obama's turnout percentage-wise should bury him.

And I mean bury his ass.  We already mentioned how Trump managed to win the Presidency despite getting smoked in the popular vote.  HRC topped the Donald by almost 3 million votes, so now we are talking about almost a 5 million-vote margin in the popular count.   When was the last time that happened?  Probably never, right?

Oh, it happened in 2012 when Obama trounced Romney by 4,982,291 votes.

But yeah, that pesky electoral college.  Bagging another 2 million blue votes doesn't do any good if they are all concentrated in dependable blue states.  And frustratingly enough, most of them probably would be.  The electoral college is an anti-democratic piece of trash that absolutely must be abolished as soon as possible--but until then we are stuck with it, and that gives rise to the second argument in favor of Trump's supposed formidability: the argument from the electoral college.

The reason the electoral college favors Trump of course, is that every state is guaranteed at least one representative and two senators no matter how small the population may be--and sparsely-populated western states tend to vote overwhelmingly red.  So a state like New Jersey, with 8.9 million residents, gets 14 electoral votes--and that comes out to roughly one EV for every 635,000 people.  But a state like Wyoming, with fewer than 578,000 residents in the whole state, still has 3 electoral votes-- approximately one EV per 193,000 residents actually.  One person, one vote?  Not so much.  And those little tiny states add up.

So the electoral college isn't fair.  It isn't just.  It's something we would call "unamerican" if any other country had it.  And it handicaps the race in favor of the Republicans.

Still, the EC factor is far from insurmountable, and tends to be decisive only in close elections.  A Dem who can pull another 2 million voters to the polls should have no trouble overcoming it.

If we assume the 2020 Dem nominee can at least hold the Hillary Clinton states, then getting the magic 270 electoral votes needed to reach the White House requires flipping 38 more electoral votes. The most likely places to find those 38 electoral votes are in the so-called battleground states, i.e., those that went for Trump by narrow margins.  Just looking at states Trump carried by less than 1% of the vote gives us:
  • Michigan (16 electoral votes): Trump won by 10,704 votes in 2016 (0.2%)
  • Pennsylvania (20 electoral votes): Trump won by 44,292 votes in 2016 (0.7%)
  • Wisconsin (10 electoral votes) Trump won by 22,748 votes in 2016 (0.8%)
Just those three states account for 46 electoral votes between them--enough to swing the result to the Democrat.  Now, there are several other states that Trump also won by slightly larger margins in 2016 and that realistically will likely be battlegrounds in 2020, such as Florida (Trump won by 1.3% in 2016), Arizona (Trump won by 3.5%), North Carolina (Trump won by 3.8%), and maybe even Ohio (Trump won by 8.6%).  But for our purposes, let's just focus on the top three.

Hillary received 65,853,514 total votes in the 2016 election.  Of those, 2,279,805 were cast in Michigan (3.46%), another 2,912,941 were cast in Pennsylvania (4.42%), and 1,382,210 were cast in Wisconsin (2.1%).  Totaling those up, we see that in 2016, our three key states accounted for approximately 9.98% of the total Democratic presidential ballots in 2016.  Man, that's real close so to simplify the math I am just going to say 10%.

What happens if the Democratic nominee draws 2 million new voters to the polls, and 10% of them are in Michigan, Pennsylvania, and Wisconsin?  That's 200,000 more votes in a set of states the Dems lost in 2016 by a total of less than 80,000 votes.  I think you see where this is headed.

I'll do the math anyway.  Michigan's share of the additional 2 million blue votes would be about 69,200.  Add that many to Hillary's total from 2016, and she beats Trump by almost 60,000 votes in Michigan.  Pennsylvania's and Wisconsin's additional shares would be 88,400 and 42,000 more Dem votes respectively -- or enough to win those states by about the same margins as Trump did.  

I know you are curious about Florida too--so am I, so what the heck.  In 2016, Trump won the state by just under 113,000 votes.  Hillary received 4,504,975 votes there, representing 6.84% of her total.  So Florida's additional share of 2 million more blue votes comes in around 136,800 votes: that's more than 113,000, so enough to take Florida and its 29 fat electoral votes--which we don't even actually need because we already have cars, cheese, and whatever Pennsylvania does.

That is, just by matching Obama's 2008 turnout level (61.6%), a Democratic challenger should be able to flip Michigan, Pennsylvania, Wisconsin, and Florida--collectively representing 75 electoral college votes.  That drops Trump to 229 EVs against the challenger's 302, and puts a Dem in the White House.  And all of this is without persuading a single red-faced Trump voter to confront his or her cognitive dissonance.   

Now, there are plenty of other dynamics in play this time around, as in any election.  But the fact is that most of them do not redound to Trump's advantage.  As a Pew study observed last year:
  • "nonwhites will account for a third of eligible voters – their largest share ever"
  • "one-in-ten eligible voters will be members of Generation Z"
  • "one-in-ten eligible voters in the 2020 election will have been born outside the U.S., the highest share since at least 1970"
  • "hispanics will be the largest racial or ethnic minority group in the electorate, accounting for just over 13% of eligible voters"
Probably the one factor that favors Trump is that voters aged 45 and over still outnumber younger voters (54.1% of the electorate).  But even this figure is slightly smaller than the 56% of voters aged 45 and over from 2016. 

And then there was the massive, 53% voter turnout in the 2018 mid-term elections--the big orange spike at the end of the chart above.  It was the highest participation for a mid-term election in four decades and tipped control of the House of Representatives to the Democrats.  Maybe they didn't show up with torches and pitchforks.  But does anyone think they were happy?

Trump is not a lock.  Far from it.  In fact he damn well should lose, as long as the Democrats can just nominate somebody worth voting for.

No comments:

Post a Comment

Letter to University of Michigan President Santa Ono on Pro-Palestine Protests of March 24, 2024

 Dear President Santa Ono: Although I was generally in agreement with the content of your letter regarding the disruption at the honors conv...