Tag Archives: 2020 Presidential Election

Four more years? (2 months to go edition!)

Well here we are, with 60 days to go! That frightening image above, by the way, is the “Red Mirage”. We’ll get back to that later. But first, as a reminder, at three months to go, we concluded that things had tightened a little since four months out, but that Biden remained in a very solid position. How do things look now that we’re two months out?

The last thirty days have seen a lot of news:

  • Biden’s VP pick of Kamala Harris
  • Both party’s conventions
  • Trump taking action on a series of executive orders meant to address the economic fallout from the pandemic
  • New national unrest, including murder of protesters, over the police shooting of Jacob Blake in Kenosha, Wisconsin
  • Trump responding by a full-court press to run a “law and order” campaign

In response to all this activity, national polling over the past month has seen…on net, just about nothing. Biden is up 0.2% over a month ago, and Trump is up 0.4%, leading to a net of Biden +7.2% now vs. Biden +7.4% a month ago:

Yes, but the polls in 2016 were wr- Stop! Stop? Stop. It is, however, useful to compare Biden’s year-to-date position in the RCP average to Clinton’s 2016 average for the same period:

One will note that Biden has been above 50% several times. Clinton only ever was for a single day. Biden has almost never been below 48%. Clinton was rarely above 48%. Biden has never had a lead of less than 4%. Clinton was less than 2% multiple times, and Trump was even ahead of her twice. On this same day in 2016, 11% of voters were undecided between Trump and Clinton. Only 8% of voters are between Biden and Trump. Everything about the comparative perspective says that Biden is and consistently has been in a much stronger position than Clinton.

It is worthwhile to note that RCP’s polling average is “naive” in the sense that it just averages all major polls within a certain time period (the last eight days, at the moment). 538’s polling average is a bit more sophisticated as it weights pollsters according to their historical track record of accuracy, and makes adjustments for pollsters that show consistent partisan lean in one direction or the other. Using this methodology, 538 currently has Biden at +7.3% vs. +8.2% a month ago:

In 2016, Clinton won the popular vote by +2.1%. Taking the smaller of the two averages above, Biden at +7.2%, and adding that difference of 5.1% to Clinton’s 2016 state by state margins results in the following electoral map:

That gives us a ballpark of how things might be different this time, but of course state by state dynamics are not necessarily going to be the same as last time. If you look at RCP’s averages of the likely swing states, Biden currently leads in all of them:

This has narrowed within the last 30 days, which is certainly worth keeping an eye on:

However, it should be noted again that Biden maintains a fairly strong and consistent lead in these states. Using its more robust methodology, 538 is currently showing:

  • Wisconsin Biden +7.1%
  • North Carolina Biden +1.8%
  • Florida Biden +4.1%
  • Pennsylvania Biden +3.4%
  • Michigan Biden +6.4%
  • Arizona Biden +4.6%

If we took only those states where Biden has more than a 5% lead according to the 538 averages, we would have this map:

Under this map, if Biden wins any one of Florida, North Carolina, or Pennsylvania, even by a few hundred votes, he’s President. Trump would have to win all three to prevail. If Biden just won Arizona, we’d be at 269-269 and the election would got to the House. May Al’lah forefend against such an outcome.

For their part, RCP’s current “no toss-ups” map shows the following:

There are some other electoral data points worth considering as well. RCP’s “no toss-ups” Senate map currently has Democrats picking up five Senate seats (and losing Alabama) for a net gain of +4, which would flip the Senate:

And the Generic House Preference polling has Democrats up by 7.3%, not far off of the 8.7% that catapulted them to a net gain of 41 seats in the 2018 midterms.

Finally, there are Presidential approval numbers to consider. Trump is currently underwater by nearly 9%:

While that’s actually an improvement from nearly -14% a month ago, on the face of it you might still think this isn’t a good place for an incumbent to be two months out from an election. And you would be right! The only two first term Presidents of the last 50+ years who have been in a similar position were Carter and Bush I, who were both headed for defeat:

What this all boils down to is that national polls, state-by-state electoral college totals, the outlook for House and Senate races, and Presidential approval ratings are all telling us the same thing: Biden remains in a very strong position with two months to go.

Now, a word about our opening graphic. This is what’s being called the “Red Mirage“. Indications are that a large percentage of Biden-leaning voters are planning on voting absentee. A similarly large percentage of Trump-leaning voters are mail voting-averse, and are planning on voting in person. Meanwhile, the individual states have widely varying rules about when absentee ballots must be received by, and the deadlines for starting and finishing counting them.

Because of all this, Election Night returns may show a big lead for Trump, and Biden’s true numbers won’t be apparent until several days, maybe even more than a week, later. There is nothing improper or weird about this, and it doesn’t affect the final outcome at all. But the two dangers it presents are:

  • If you don’t know it’s happening, Election Night could be pretty scary!
  • It presents Trump with a week or more to agitate his followers about the “illegitimacy” of the outcome, something that could be enormously disruptive for the country.

So I would recommend fortifying yourself against the first possibility, and all of us being prepared to take appropriate action against the second. Meanwhile, we’ll reconvene in a month to see how things are looking with a month to go!

Four more years? (3 months to go edition!)

countdown

Is it just me, or was that a fast month? When last we gathered with four months to go, Biden’s position was looking very solid, and even better than it had at five months or six months. How are things different at this point?

To start with, the polling numbers have tightened up a little, but not a lot:

BvT

It hasn’t so much been Biden falling out of his usual range as Trump having bounced back a little from a dismal bottom beneath his usual range. Overall, Biden’s position remains very strong. One thing to keep in mind about the RCP average, is that it’s just that- an average. But not all polls are created equal.

Over at FiveThirtyEight, they address this by weighting pollsters according to their history of accuracy (more accurate ones being given more weighting than less accurate ones) and historical partisan skew (adjustments being made to pollsters that are regularly more +D or +R). This more nuanced approach has Biden up by a little more:

538poll

Another useful way to think about the RCP polls is to compare Biden’s YTD numbers with Clinton’s for the same time period in 2016:

This side by side comparison makes immediately apparent that Biden’s 2020 position is much stronger than Clinton’s 2016 position:

  • Biden has never been less than 4 points ahead, compared to Clinton, who was frequently within 2 points, and twice was briefly behind Trump.
  • Biden has been above 48% 2/3 of the time or more, Clinton was above 48% only 1/3 of the time or less.
  • Biden has had several runs above 50%, whereas Clinton was only above 50% for a single day.

Of course, we in America are blessed with that cute little Constitutional anachronism known as the Electoral College, so national polls can only tell one so much. We do have a pretty good idea from 2016 and trends since then what some of the key states will be, and Biden currently leads in all of them. Not only does he lead, he’s well outside the margin of error in every state here except Arizona:

battlegrounds

If we look at RCP’s current “No Toss-Ups” map, Biden is in very strong shape:

NoToss

Even if one just looks at the states where Biden leads by 5 points or more, the map looks like this:

5plus

“But the polls were wrong in 2016!” First of all, no. Polls can be wrong, but they are rarely 5% wrong. Also, in some of the 2016 states, a major problem was that, because they weren’t widely believed to be competitive, the were under-polled. We may be looking at the wrong states now, but for the ones that were surprises last time, they are very well polled this go-around.

Take Michigan as an example. The current RCP average there is based on four major polls that all have data from the last two weeks:

MI

A final useful check-in is on Trump’s net approval numbers in comparative perspective:

NetPop

One can clearly see that, except for a few days right at the beginning of his term, Trump has never had positive net approval numbers, and has consistently lagged every President of the last 50+ years. The only two his current numbers really compare to are Carter and Bush I, first term Presidents who were going to electoral defeat.

So, while the polling numbers have tightened up a  little, by every concrete measure we have, Biden remains in a very strong position. Let’s tune in again in a month (and after the pseudo-conventions and Biden’s VP pick to boot!) to see what things look like two months out!

 

Four more years? (4 months to go edition!)

US-POLITICS-TRUMP-INAUGURATION-SWEARING IN

Look, I get it! Of course I do. To be of a Leftward persuasion, and to have closely followed Presidential politics for years is to know heartbreak. The Supreme Court halting the counting in 2000 and selecting Bush as President. Seeing John Kerry haplessly roll over and play dead, losing a winnable election in 2004. “Losing” 2016 while finishing 3 million votes ahead. Michael Dukakis being up 17 points in late July 1988, and going on to lose by 8 points!

So I get why instinctive panic, and a mistrust of any numbers, creeps in. Instinctive panic, however, is both irrational and tactically unhelpful, and data can be clarifying. When we took our “five months to go” look in June, Biden actually seemed to be in pretty good shape. Friday last week was four months to go. Does that still seem to be the case?

Let’s start by looking at the latest polling averages:

JoeB

Biden has a more than 8-point lead. But that’s not all. As I’ve mentioned before, I find it instructive to look at Hillary Clinton’s averages for the same YTD time period:

HC

A few things immediately stand out from this comparison:

  • Biden has been above 50% for about half the time YTD. Clinton was only above 50% for a single day.
  • Clinton was rarely above 48%. Biden has rarely not been above 48%.
  • Biden has never had less that a 4 point lead over Trump. Clinton was down to less than 2% several times, and Trump even briefly had a lead over her.

To me, the conclusion that the Biden campaign is in much better shape than the Clinton campaign was at the same point is unmistakable.

Of course, as we all too sadly know from 2000 and 2016, winning the popular vote doesn’t win you the election, winning the electoral college does. However, being more than 8% up almost certainly wins you the electoral college. If we add the 6.6% higher Biden’s current numbers are compared to Clinton’s final 2016 margin (2.1%) to her state-by-state totals, we get this electoral map:

plus67

2020, of course, isn’t 2016, and dynamics in the individual states may be quit different this time. So what happens if we look at current polling in the potential swing states? As it happens, Biden is currently leading in all of them:

AZFLMINCPAWI

In addition to this, Iowa, Ohio, and even Texas are all close enough to be in play. Polling inherently has a margin of error, even when using polling averages, but just keeping the states where Biden’s lead exceeds 5% results in this map:

moe

One more statistical check-in is helpful here. This is the President*’s net popularity vs. other Presidents of the past 50+ years at this same point in their terms:

net

Lest we forget, Trump is, and has been the whole time, historically unpopular. The only two Presidents his current net (un)popularity looks comparable to are Carter and Bush I, who were both headed to first term defeat.

One last sanity check is to think about these numbers in the larger electoral context. In latest figures, Biden and the DNC are out-raising Trump and the RNC. The Generic Congressional preference numbers are about as positive for the Democrats as they were in 2018, when they gained around 40 seats. Democratic candidates are polling strongly in the swing Senate races. All of this fits together into a consistent picture.

Up above, I asked if Biden’s four-month outlook still looks as strong as it did at five months. The answer, given the above, is a clear yes. If anything, it looks a little better. Take nothing for granted about this- donate, phonebank, textbank, write letters, organize on social media. But be encouraged in doing so, and know that bed-wetting for other than recreational purposes is not called for.

Onward!

 

Four more years? (5 months to go edition!)

mappie

About 6 weeks ago, I did my latest update on the election outlook. At the time, I concluded that Biden seemed to have a pretty strong chance based on prevailing indications, and there were plenty of signs that Trump was in trouble. Today marks five months out. Is there anything to add to our previous examination? I think so!

Let’s start with the basics. Yes, Biden is still the all-but-certain nominee. Not only have all rivals dropped out, all endorsed him including chief rival Bernie Sanders, the prior nominee endorsed him, and the last Democratic President endorsed him, as of last night’s Super Tuesday, he is sitting less than 100 delegates off of making it official:

StX

Basics number two- an average of recent polls gives Biden a healthy lead:

rcp

More than that, year-to-date Biden has always had a healthy lead. It’s instructive in this regard to compare his numbers to Clinton’s 2016 numbers for the same time period. A key thing to note is that she was rarely above 48% (where she ultimately finished) and spent only a single day above 50%:

clinton 16

In contrast, Biden has mostly been above 48%, and has spent significant time over 50%:

bidenover50

Of course, as 2016 painfully demonstrated, the popular vote isn’t what ultimately matters. The election is determined by that cute little antiquated Constitutional mechanism, the Electoral College. To do an initial “face value” analysis of that, Clinton “lost” in 2016 by being at +2% nationally, and Biden’s average is currently +8%. If you add 6% to Clinton’s state-by-state 2016 margins, you get this map:

Clintonplus six

An interesting ballpark, but I’ll grant you, things are not that straightforward. 2016 really came down to a handful of swing states, and the dynamics of those swing states will be different in 2020 than they were in 2016. What do current polling averages of these states currently show?

swing

That’s a Biden lead in 5 out of 6, and Trump’s lead in the 6th is razor-thin. If we just take only the states where Biden’s lead is outside of a 3 -point margin of error, we get the following map:

MOEplus3

If we took the “no toss-ups” approach for all current state polling averages, the map looks like this:

notoss

The other way to look at this is to consider the President’s approval rating according to an average of recent polls. It’s currently more than ten percent net negative, which is where it’s been for much of his Presidency:

approval

More than that, except for literally the first few days of his term, he’s never had a net-positive rating. Comparing his net approval to other recent administrations reveals the stark difference. The only recent Presidents who his net numbers are similar to at the moment are Carter and Bush I, who of course were both first term Presidents who were about to be defeated:

net

My final appeal is to common sense. Consider the following:

  • The Libertarian and Green parties had well-known candidates who had previously run nationally in 2016. This time they have comparatively unknown newcomers.
  • The 2016 nominee had been the target of decades of character attacks. Literal generations of Conservatives hated her. This campaign will get very nasty, but there just isn’t that kind of base of animosity to activate in regards to Biden.
  • Clinton was the subject of active investigations which continued to dog her as late as 10 days before the election. The attempts of the Administration to stir up something similar for Biden have fizzled out, and their most concrete result so far was to get Trump impeached.
  • A national crisis has killed 100,000 of us and counting in just a few months, and the  widespread perception is that the current administration has badly mishandled it.
  • The economy is in recession, and unemployment is higher than it’s been at any time since the Great Depression. That’s never good for an incumbent.
  • The country is, actually, on fire. It’s true that “law and order” issues often rebound to the Republican candidate’s favor, but more than that, they almost always come to the detriment of the party in power.

So, has the outlook changed since our last check in? Yes, in that it’s gotten better for Biden and worse for Trump. It could be that the current crisis reverses this trend, but my instinct is that it won’t. More than anything, the thing that most stands out is how durable the dynamics of the race have been year-to-date no matter what external events have been in play.

I’m not saying take it as a sure thing. Not even remotely. I am saying be very encouraged as you go out to do the work that needs to be done over the next five months to make it so.  Tune in some time in early July to see if it’s looking different with four months to go!

 

Four more years? (6.5 months to go edition!)

2020 election

Well, here we are. Over the past few weeks, I had some personal priorities to attend to, as well as, you know, the disruption of a global pandemic. My plan had been to update my election series on the Democratic nomination after Ohio, but due to these two things, and also Ohio and almost every other primary being delayed until June (or later!) I never got there.

And now, whereas we previously knew:

there-can-be-only-one

We also now know who the one is!

biden smiling

I have come neither to praise Biden nor bury him, my goal was always to provide a data-driven outlook on who the likely nominee was. Now there’s no “likely” to it. All the other candidates have dropped out and endorsed him. There could still be some shocking outside event, but, barring that, Joe Biden is the Democratic nominee.

So now it’s time for my analysis to pivot to our next electoral outlook, be it still more than six months out:

ev

Actually, I’ve dabbled a bit in this already, here and here. Both times, I came down on the side of Trump’s re-election being very very far from a  sure thing, so much so that we on the Left should stop bed-wetting ourselves into surrender. That was based on general outlook, things are a little less abstract now that we have actual names on both sides of the ledger. So how do things look currently?

The first stop is to observe that Biden actually has a pretty healthy and persistent lead in year-to-date polling averages:

YTD average

I can hear a chorus of “so did Clinton” coming, but I think that’s actually a pretty useful comparison to make in some ways. What immediately strikes me in comparing Clinton’s numbers for the same period is that she spent a lot of time under 48% (where she actually ended up finishing) and almost no time above 50%. Biden, by contrast, has been above 48% almost the whole time, and not infrequently above 50%. His “floor” is higher.

Clinton

Yes, but Chris, the polls were wrong! Before continuing, I urge you to read this from the lead analyst at Real Clear Politics, and this from the same at FiveThirtyEight. If you’re now convinced that national opinion polls were actually in range, and the real problem was polling in particular states, and pundits generally understanding probability, then we can continue. If you’re not, well, data-driven analysis probably isn’t for you.

Nationally, if Biden did turn out to have around a 5.8 point lead, the quickest and crudest way to look at that would be adding 3.7% to Clinton’s 2016 closest state margins (since she won nationally by 2.1%):

margins

That would get you the following map:

EVapril

Admittedly, that’s a crude way of estimating. It would be better to look at actual surveys of some of these states tagged as possible flips from last time (Arizona, Florida, Michigan, North Carolina, Pennsylvania, and Wisconsin) to see if that outlook matches the national outlook. So let’s do that!

states

At first blush, it certainly looks plausible to think that Biden, currently, is competitive in a lot of this territory.

The other major way to look at this is to examine the incumbent’s popularity. He had a brief “rally around the flag” moment around his handling of the COVID pandemic, but that seems to have almost entirely reverted now. Overall, since inauguration, his popularity has been underwater, and he’s never consistently had a run above 44%, two percent less than what he drew in 2016:

unpop

We’ve kind of gotten used to this in some ways, so it can be helpful to look at it compared to other recent Presidents’ net approval rating at the same point. Trump never had the kind of early term approval bump that the others had. He’s almost never been above the trend line of any of them. The two his ratings at this point look most similar to are Carter and Bush1, both of whom were headed to first term defeat amid economic turmoil.

comppres

To me, all of this says that the data is saying the same thing that makes immediate intuitive sense: Any President hobbled by economic turmoil and botched reaction to a crisis is probably facing a difficult reelection. And Clinton had so many unique headwinds against her- decades of slurs aimed at her character, a last-minute FBI intervention, an at the time underappreciated massive foreign intelligence operation against her, and still only “lost” because of around 77,000 votes in three states. Any halfway decent Democrat was likely to be competitive.

And, indeed, there is every reason to think at this point that Biden has a very solid shot at winning the election. There are miles of twists and turns to come. So many that I’m not even going to bother to do an update until June 3rd, which puts us five months out. But nothing we’re seeing now dissuades from the idea that Trump is in serious trouble, and Biden’s chances are good.