E-Day Minus 16

Fivethirtyeight.com is the best in the business of aggregating and analyzing election polls. So, what does the site tell us at this point in the campaign?

A candidate needs 270 Electoral College votes to win. With just over two weeks to go to Election Day, 538.com shows Biden ahead substantially, i.e. by 6.8% points or more in states with 274 Electoral College votes, including the closest of the “battleground” or tipping point states, Minnesota (+9.1%), Michigan (+7.9), Wisconsin (+7.8%) and Pennsylvania (+6.8%). Fivethirtyeight.com gives Biden an 88% chance or better of winning in each of these states.

With leads in New Hampshire of 11.6%), Florida (+4.0%), Arizona (+3.9%) and North Carolina (+3.2%) Biden is rated respectively at an 86%, 72%, 67% and 67% chance of winning those states. They provide backup if, for some reason, one of the battleground states falters (or is “stolen”). Georgia and Texas, both rated as toss-ups or leaning slightly for Trump, provide further backup.

So here’s your guide for watching things unfold in these last weeks.

Keep an eye on the national polls. Today Biden has a 10.6% lead nationally. Don’t panic until Biden’s lead falls to below +3%, which it shows no sign of doing at this point. To be safe you might want to add a percentage point or two as insurance against skullduggery. Last time, Hillary Clinton won the national vote by +2.1% and lost by less than 1% in Wisconsin, Michigan and Pennsylvania, giving Trump the more than 270 ECV he needed. Round up Clinton’s 2.1% national lead to 3% add one for good luck,  and call it the 4% rule: as long as Biden is more than 4% ahead in the 538.com aggregate of national polling, he is virtually certain to win the states necessary for a Electoral College majority of 270.

So subtract Trump’s built-in Electoral College lead of 3% plus 1% for good measure to cover a small leakage for Republican voter suppression, and you have Biden’s likely lead in the key swing or tipping point states, Michigan, Wisconsin and Pennsylvania, that in 2016 were virtually even. If Biden does better than % more than Trump nationally he should be fine in the Electoral College.

Right now a multiplicity of polls on aggregate show Biden ahead by 10.6%. By subtracting 3% from 10% one drags Biden’s vote into the 7% range, precisely his lead in the states he needs to get 270 ECV. It’s a useful rule of thumb for reassurance as the last days of the campaign unfold.

If you’re interested in the fine detail, keep an eye on Pennsylvania, Wisconsin, Minnesota and Michigan. If Biden’s lead in those states start falling below 2 or 3% resume worrying because now we are within the standard polling margin of error. Check Florida and North Caroline. If Biden continues to lead in those states he’ll be fine, even if things are tight in the other four battleground states. The rest, like Georgia and Texas, will be icing on the cake.

Huge numbers are already voting. Voting intentions have been extraordinarily stable, though it must be said that the events have twice moved noticeably the polls to new plateaus, both times in Biden’s favour: once in June/July after the murder of George Floyd up from a 4% +/-Biden lead in the previous three months to 7-8%; recently following the debate and Trump getting infected with the corona virus the national polls  went from a +7-8% lead for Biden to +10% and more. Trump’s underlying poll numbers are terrible and give him little chance at this late date to change the environment, if he even knew how.

People ask me about the chances of Trump stealing the election. He’s announced he will try, though his track record for carrying through on his threats is mediocre. The Republican Party however has demonstrated some success in voter suppression. Neither the GOP nor Trump himself have shown any inhibitions in trying to use their power legally or otherwise to interfere in democratic elections. It is a tradition that goes back to the first elections in the U. S., and was used extensively to keep freed Black citizens from voting. In various forms, it continues to this day. Violence is possible and another tradition in certain parts of the country.

Be assured, however, that the Democrats will be preparing to offset any steps the Republicans and Trump may try. They will have played out any number of scenarios and planned offsetting moves. They have recruited an army of lawyers. In many states and jurisdictions the election process is overseen by Democrats or honest Republicans. Law enforcement will in most cases will play by the book and be planning to deal with incidents that arise. In the end the best defence to these attempts will be a massive turnout of Democratic voters – the early signs are that this is happening. If the results are close then all bets are off.

If Trump does try to stay in office after losing the election, I think U.S. institutions are strong enough to deal with him. The threats he makes are just noise. Moving to another country isn’t. It’s what all authoritarian leaders do. Maybe he has already booked his flight.

Elections, like all other human activity, are fundamentally unpredictable. This one won’t be over until it’s over. At this stage, it's like landing a 747. A victory for the Democrats needs to be brought home carefully.

Postscript

After I published this post, a reader asked me to comment on the 2016 polls in relation to the current situation. This was my answer.

"Last time, Hillary was at 4% two weeks out. There was a much higher percentage of undecided voters. Her lead then was ‘soft’ or ‘lukewarm’ — people were not committed to support her but were merely leaning toward her. Biden is much more popular today than Hillary was then. The undecideds and third party voters were around 15% in 2016 whereas they are at 5% today. Last time, they broke 3:1 for Trump and this time Biden is viewed more positively over Trump by a significant margin. Worst case scenario is a 50-50 split. 

In two crucial states (Wisconsin and Michigan) the state polls failed to note the significant difference between white, non-college and college-educated men. The former voted substantially for Trump and were over represented in these polls. This time, pollsters have corrected for this. Biden is polling better with this group because he’s a man — they simply did not want to vote for a woman last time. 

In sum, Biden’s lead is significantly bigger, there is a much smaller pool of undecideds who prefer Biden over Trump, and state polling errors have been fixed."