Relax. Biden will win.

     "Democrats have a massive case of PTSD about 2016, so much that they often overlook how this election is entirely unlike the one four years ago."

       Peter Lemieux, Political Scientist with a quantitative orientation.

Getting beat.

Peter Lemieux: 

Don't worry. Biden has got this one.

My blog post yesterday got pushback. 

Yesterday I wrote that there are some frightening parallels between 2016 and today. Biden is doing what Hillary Clinton did: lay low, let Trump have center stage to self-destruct, and flood the market with a hundred proposals for change.

I warned in 2016, and again yesterday, that this gives Trump a better story than it gives the Democrat. Trump has a simple brand; Biden is left being the off-stage guy who isn't Trump. I worried that this wasn't enough.

Peter Lemieux said I was wrong.

Peter Lemieux is a college classmate who went on to get his Ph.D. in political science from MIT. His career is at the intersection of information technology and quantitative analysis, which gives him a different perspective than mine on politics. (I start with partisanship and tribalism, then add narratives and messages. I think people are choosing sides in a drama.) He looks at granular data on voting behavior: what certain demographic groups voted for in the past, then he adds them up, adjusts figures based on new data, and makes predictions. He puts his own analysis up for public review at his website: www.politicsbythenumbers.org.

Guest Post by Peter Lemieux

Peter Lemieux

"I see Biden winning 300+ electoral votes this fall."

"Democrats have a massive case of PTSD about 2016, so much so that they often overlook how this election is entirely unlike the one four years ago.

Most importantly, 2016 was a open-seat election between two highly-disliked candidates.  2020 is a referendum on Donald Trump. Biden is a good candidate in a race like this since he is so obviously not-Trump.  (People who disliked both candidates in 2016 voted for Trump; in 2020 it's the reverse.)

Second, the margins in polling today generally surpass those seen in 2016.  Biden's lead of 10-15 points is about the same as Reagan's lead over Mondale at this time in 1984.  I recall few pundits expecting Mondale to stage a comeback of that scale.

Third, Hillary ran a poor campaign, especially in the upper Midwest. Not visiting Wisconsin once was a major mistake.

Fourth, Trump won by fewer than 80,000 votes across three states.  I see little chance he will win Pennsylvania or Michigan this time around, and while he might still carry Wisconsin, Biden will likely pick up Arizona.  (I expect astronaut Mark Kelly's campaign for Senate to provide a "reverse-coattails" boost for Biden.)  Flipping just PA, MI, and AZ gives Biden an electoral college majority.

Fifth, people worry that massive campaign spending by outside groups could tilt the balance in the Senate races.  I found no effect for spending by such groups in 2016 and 2018 though spending by the campaigns themselves does matter.  I suspect that massive television advertising, which is the primary instrument available to outside groups, has little effect these days.  Remember how unsuccessful both Bloomberg's and Steyer's campaigns were during the primaries.

Sixth, the polls are better this time around.  Even state pollsters are weighting by education which mattered greatly in 2016.  (Remember the national polls were only off by one percent, giving Clinton an average three-point lead in the popular vote when she won by two.) College-educated people are now more likely to respond to pollsters so they need to be "down-weighted" when estimating figures for the entire electorate.

A plausible way to win

Seventh, Trump is behind in key groups like people over 65.  He carried them by seven points in 2016.  In the latest Quinnipiac poll, Biden leads Trump 54-40 among 65+.  Trump's margins among whites overall, evangelicals, and even white, non-college voters have all declined compared to 2016.

Finally, I haven't even mentioned COVID, Black Lives Matter, or the recession, all of which favor Biden.

I see Biden winning 300+ electoral votes this fall.  My model, based on polling through June 21st, predicts an eleven-point Biden win in November, which corresponds to a 390-148 margin in the Electoral College.  Biden's lead has widened since that article was written.

https://www.politicsbythenumbers.org/2020/06/25/some-observations-on-bidens-margin-in-presidential-polling/

I expect Trump to lose badly in November, and the Democrats to end up flipping the Senate by a couple of seats.  My Senate model includes Trump's approval at the state level, and again, he is only trending in the wrong direction.

https://www.politicsbythenumbers.org/2020/03/05/senate-update-march-2020/ "


People ask, what about attempts to steal the election?

These are all common fears on the left, but I don't think any of them will play a big role in 2020.

Let's take voter suppression first.  Yes, there are a few states where governors and state legislatures make it harder for people to vote.  But there are vastly more states where that is not true.  Suppression cannot overcome the kind of margins I expect this fall.

As for electronic voting, many more people (half? more?) will be casting mail-in ballots this fall.  I'm actually more concerned about mail-in ballots being lost by the Postal Service than I am about them being thrown away by local officials.

As for any Trump effort to remain in office after losing, that's what we have the Secret Service for.  I'm pretty confident he would be escorted out of the Oval, by force if necessary, were he to try such a thing.

Finally, please have some faith in the local and state election officials who care a lot about the validity of the elections they oversee.  The program on election administration at the Humphrey School has been extremely active this entire year. https://twitter.com/HHHElections