Harry Enten's new column focuses on a very interesting pattern: the amazing consistency across time of Biden's lead over Trump. Enten conducts some historical analysis to show how unusual this is and that this consistency enhances the likelihood the lead will hold through the election. This would obviously be to Biden's advantage.
"Biden's lead is about as steady as it can possibly be. Not only is he up 6 points over the last month or so, but the average of polls since the beginning of the year has him ahead by 6 points. Moreover, all the polls taken since the beginning of 2019 have him up 6 points.
The steadiness in the polls is record breaking. Biden's advantage is the steadiest in a race with an incumbent running since at least 1944. That could mean it'll be harder to change the trajectory of the race going forward, though this remains more than close enough that either candidate could easily win.
To know this, I went back and looked at how all the national polls deviated from each other during January to early May of the election year.
This year, 95% of all the individual polls so far have shown a result within 6 points of the average. That's basically what you'd expect if you took a lot of polls and the race wasn't moving (i.e. the only shifts are statistical noise from sampling). It's a ridiculously small range historically speaking....
This year, we've seen pretty much none of [the large shifts seen in previous cycles] in the national polls, even as we've had a once-in-a-lifetime health pandemic. That might make you doubt that the polls will move a lot going forward in 2020."
Note that Enten uses a different basket of polls for his estimates than RCP's throw-everything-in-a-blender approach. I guess I'd be happier if his methodology included a way to extract the signal from the low-quality polls he excludes. That's really what 538 does, but of course their moving average is not available.
Judging from 2016, 538 will probably starting putting up a horse race average this yeat around the end of the first week of June. In that cycle, Clinton started out about 4 points ahead in the 538 average, a lead that disappeared by the end of July and then re-opened after that before narrowing again in early to mid September.
It'll be interesting to see where the 538 average starts out this year and how steady it is after that.