First, a dose of reality.
It doesn’t matter that nobody has ever called you or that the polls were off in this or that particular election or that you can cherry pick and find a poll where Biden is up — the truth of the matter is that, taken on average (especially among high quality polls) Biden is clearly down in any reasonable head to head average against Trump. And worse, since Biden needs to win by about 3 points or so in order to win the election, he’s functionally down by about 5.
But the good news is that the polls right now almost certainly reflect a protest or messaging, rather than actual voter intention. Sure, there are people pissed off about the war in Gaza (as though Biden controls Netanyahu), but it is hard to imagine that a year from now, this will be most people’s highest priority. There are those concerned about his age and would like someone else. But this isn’t new information, and especially compared to Trump’s complete incoherence and incompetence, Biden is in great shape — and it doesn’t explain why people have gotten so vehemently opposed to him.
To give you a sense of how completely ahistoric the current polling is with actual electoral outcomes (especially given full employment, exploding GDP, falling gas prices and inflation, rising stock market, and on and on), I’ve taken a look at the validated exit polling (a summary spreadsheet is here) using Roper (pre-2016) and Pew analysis (2016-2020) since 1980. I’ve smoothed the data between age ranges a little, but other than that, basically taken the results at face value. For the most part, they are consistent with the actual electoral outcomes to within a couple of points.
I then asked how any given cohort (the group born, in say, 1970) voted in each subsequent election, and what I found was a little surprising. Except for some weird hiccups (like in 1992 and 1996 when third parties got > 10% of the vote), each given cohort was noisy, but flat. The old adage about getting more conservative as you get older seems to not hold up in national politics. But what’s more, for the cohorts born after around 1970, there is a definite — and persistent — shift leftward. This is, of course, great news. It’s not just that young people are liberal now. If history is any indication, they’ll stay liberal for their whole lives. This is why Republicans are so scared. As a practical matter, they can’t win a national vote ever again.
When you look at how consistent these cohorts are, it’s even more surprising when you compare it to any of the recent polls. For instance, looking at the Fox Poll (chosen because it’s the most recent high quality poll that published its crosstabs, but also has pretty typical results), the idea that voters born in 1980 or later will suddenly fall off a cliff seems inconsistent with how history has actually played out.
The trend is even clearer when we eliminate the Bill Clinton elections from the plots. After all, at least so far, Manchin are Kennedy are polling well as protest votes — as opposed to Ross Perot who (weird as it is to think about now) had a small cult-like following.
The ridiculousness of taking the polls literally is even clearer when we look at polling of black and hispanic voters over this period.
Since 1980, Dems have absolutely dominated with black voters, with their absolute worst performance, in 1980, winning by “only” 69 points. Even in the massacre of 1984, Mondale carried black voters by 82 points. And in the last 5 elections (including, I remind you, one which includes the actual matchup that we’re discussing here), the Dem won by an 80+ point margin in 4 of them (and in the worst, in 2004, by 77). Fox (and the other polls) would have us believe that margin is 40 for next year. This is a whole different universe. Again, given the incredible wide range of candidates and electoral outcomes over the last few years, the idea that Biden, in particular, has collapsed by 40 points overnight beggars belief. The same is true, though admittedly a smaller effect, for the Hispanic vote.
Incidentally, the black vote alone is sufficient to explain his deficit. An additional 40 points on 11 or 12 percent of the electorate improves his margin by 4.4-4.8 points, enough to put him in the lead.
To anticipate cynicism, this isn’t a matter of “unskewing.” I firmly believe that the pollsters did their jobs in the sense that they got a reasonably random sample of respondents and they answered honestly. But I also just as firmly believe that there is no way that they actually vote this way in a year’s time. We just won a major election (and overperformed in special election after special election), and if you haven’t noticed, every election is a referendum on the incumbent. And in places like the PA supreme court (where abortion wasn’t directly on the ballot), the Dem won by nearly 6 points!
My point is that there is enough that is unprecedented about this election (and indeed, about this moment) that to simply say that Biden is in worse shape than Obama in 2011 is to ignore that we have a very different political landscape now than then. But apparently one thing we can learn is that voters have a very consistent voting pattern, and so long as we don’t freak the hell out (or let doomsayers like James Carville freak everyone out), we’ll be fine.