Trump has a real shot of winning

Yet it's important to point out that even if the polls are an accurate representation of the current state of the race, presidential races can shift a lot during the final six months of the campaign.

Biden may be favored, but this race is far from over.

Take a look at every presidential election involving an incumbent since 1940. It's 13 races in total and gives us a good baseline from which to work.

There's been a 11-point difference between an average of May polls and the result in the average election. That would be more than enough to change the course of the 2020 race, if the shift occurred in Trump's direction.

Now some of these races (most notably 1964) had polling leaders with large margins that have never occurred in any presidential election and were bound to shrink during the course of the campaign. If we look at the only races where the polling leader had a 25-point advantage or less, the average difference between the polls at this point and the result has still been a fairly high 8 points.

If the 2020 race moved 8 points in Trump's direction, he'd win.

You could even concentrate on just the most recent incumbent elections of 2004 and 2012. Like 2020, opinions of the incumbents (George W. Bush in 2004 and Barack Obama in 2012) were hardened early. The final national results in those races differed from the May polls by 3 and 4 points.

If the 2020 race moved 3 to 4 points in Trump's direction, he might not win the popular vote, but he would have a pretty good shot of winning the electoral college.

Interestingly, there have been three presidents who have lost since 1940 (Gerald Ford in 1976, Jimmy Carter in 1980 and George H.W. Bush in 1992). Two of them (Carter and Bush) were actually ahead in the polls at this point. Carter was up by 6 points and lost by 10 points nationally. Bush was up by 8 points over Bill Clinton and 6 points over Ross Perot, and Clinton would defeat Bush by 6 points and Perot by 24 points. Meanwhile, Harry Truman and George W. Bush were trailing by small margins in the May 1948 and 2004 polls respectively, and both would go on to win by small margins.

Of course, just because something is possible doesn't mean it's likely.

Even if we saw a swing in the 2020 race, there's no guarantee it would go in Trump's favor. In the 13 races since 1940, seven times the incumbent did better than he was polling now. The other six times, the challenger did better. Just based on that data alone, it's really no better than a 50/50 proposition that Trump will do better in the results than he is polling currently.

Further, polls tend to shift in the incumbent's direction when they're doing worse than their approval ratings would suggest. Richard Nixon in 1972 and Ronald Reagan in 1984 ended up outperforming their May polling and came in closer to where their net approval (approval - disapproval) ratings suggested they'd be. Carter in 1980 and Bush in 1992 trended toward their negative net approval ratings.

Trump's net approval rating is in the mid to high negative single digits, which is right in line with his polling against Biden.

The bottom line is six months is a lifetime in politics, but you'd rather be in Biden's position than Trump's.