The 2022 midterm elections have already begun, with voters casting early ballots in some states. With Americans so closely divided -- see the tight Senate races and the tied generic congressional ballot -- even the slightest change in voter sentiment can make a big difference.
In other words, it's not only about who wins what blocs of voters (men, women, college educated, non-college educated, etc.), but also about the margins by which they win them.
That's why I was struck by a recent finding about Black voters in an ABC News/Washington Post poll. And that's where we begin our look at the week of politics that was.
Biden and the Democrats have a smaller edge with Black voters
Black voters tend to be a forgotten part of the discussion about general elections. The reason is simple: They have been one of the most Democratic parts of the electorate for a long time. While Joe Biden in 2020 likely did slightly worse than Hillary Clinton among African Americans, he still won them by an overwhelming margin.
That doesn't mean Black voters are unimportant, though. Black voters put Biden over the top in a number of swing states -- e.g., Georgia, Pennsylvania and Wisconsin. All of those states have competitive gubernatorial and/or Senate races this year.
Looking at the national polling, it seems possible that Democrats might not be able to count on nearly as much support from Black voters as they have in previous elections.
An average of the final five live interview polls of the 2020 election showed Biden with an 84% to 9% lead over then-President Donald Trump among Black voters -- a big 75-point advantage. But this year, an average of the last five live interview polls I could find gives Democrats a 74% to 12% advantage among Black voters -- a 62-point edge -- on the generic congressional ballot, which usually asks respondents some form of the following question: "If the elections for Congress were held today, would you vote for the Democratic or Republican party?"
This represents a larger swing toward Republicans by Black voters than the swing we have seen among all voters from the 2020 baseline.
Notably, the final five live interview polls I could get from the 2018 midterms showed Democrats with an 85% to 9% lead on the generic ballot with Black voters. Again, what we're seeing this year is clearly different.
This shift among Black voters isn't something new. Back in April, I pointed out that Democrats had a 73% to 11% advantage among this demographic on the generic ballot. To me, this consistency indicates that it isn't statistical noise we're witnessing. It's something more concrete than that.
As the sample size of Black respondents in any of these polls tends to hover around 100, it's difficult to say why Black voters seem less pro-Democratic than they were in 2020 or 2018.
The simplest explanation is the same one I had back in April: Biden's standing. His average approval rating among Black adults in these polls is 64%. In a compilation of Gallup polls I looked at earlier this year, it was 67%. Biden had an approval rating among Black Americans in the high 80s at the beginning of his term.
Biden's positive standing with Black voters is still significantly higher than his average approval rating with all adults (in the low 40s). But it's also lower than the lowest rating of the last Democratic president (Barack Obama) with this group.
The question going forward is how Democrats' lower standing with Black voters on the generic ballot translates into midterm election results. About 14% of Black voters are, on average, undecided or leaning toward backing a third-party candidate. Many of them could end up either voting Democratic or staying home.
But if Democratic House candidates' current margin among Black voters on the generic ballot -- 62 points -- holds, it would be the smallest margin they've won this group by since 1990.
Public House polls match partisan House polls
All of the polls cited in the above analysis of Black voters come from reputable nonpartisan public pollsters, including ABC News/Washington Post, New York Times/Siena College and Quinnipiac University. There's a simple reason we use this data: We know it hasn't been selectively released or manipulated.
Of course, we have the luxury of having a lot of public national polls to detect trends. We do not have the same luxury when it comes to polling of individual House districts. Most of the polls in the public domain at this level are from campaigns or partisan groups with a vested interest in the outcome.
It turns out, though, that these polls, on average, are showing a similar thing about the House playing field: It's tight, but with a clear shift to the GOP from 2020.
There have been 36 House polls conducted and released since the beginning of August from partisan sources in races including one Democratic and one Republican candidate. I took the result of each of these surveys, took away 4 points from the Democratic or Republican side (because that is the average bias for partisan polls) depending on whether the sponsor was sympathetic to Democratic/liberal or Republican/conservative causes, and then compared the new margin to the 2020 presidential vote in the district.
Republicans are, on average, doing about 3 points better than the 2020 baseline. Given that Biden won the 2020 election by 4.5 points nationally, this would indicate a national environment that is about 1 point in favor of Democrats. For reference, the generic ballot currently has Democrats and Republicans tied.
The fact that partisan House polls (once adjusted) suggest basically what the generic ballot suggests shouldn't be too surprising. House elections have become increasingly nationalized in the last few decades, so district-level polls should broadly resemble national polls.
Still, the similarity points to something rather important. The people inside these House campaigns likely aren't seeing anything related to the overall political environment that we're not seeing in national public polling.
The same was true in the 2020 campaign. Private polling, like public polls, pointed to a sizable Democratic victory, and both overestimated the ultimate Democratic performance.
We'll have to wait for about five weeks to see what ends up happening in November. Maybe Democrats will end up holding the line, or maybe Republicans, like in 2020, will end up doing a lot better than the data at this point suggests.