In thinking about what’s ahead and what might have been, I offered a counterfactual take at New York:

It is perfectly natural for Democrats to think of the midterm “wave” election they are hoping for in November as representing a righteous repudiation of the particular characteristics of the 45th president — his constant lies and provocations, his manifest lack of respect for the rule of law, his perpetual racial and ethnic appeals, his chaotic approach to decision-making, and his oppressively omnipresent personality. But in fact, the kind of midterm backlash to the White House that appears to be shaping up is entirely normal, even if Trump is not.

To recapitulate: the party that controls the White House almost always loses House seats in midterm elections. There have been just three exceptions in the last century (in 1934, 1998, and 2002), all of them occurring when the president had unusually high job approval ratings and when something else unusual was going on (the Great Depression, for which Republicans were persistently blamed, in 1934; the impending impeachment of Bill Clinton in 1998; and a security-conscious reaction to 9/11 in 2002). When the president’s party is in theory “over-exposed” in the House by recent success (e.g., the big Republican wins of 2010 and 2014), the odds of losses are even higher. And while there is a less systematic carryover to Senate elections because of the high variability of the particular “class” of 33 or 34 senators up in any one midterm, an anti-White House midterm “wave” affects those outcomes, too — particularly with partisan voter polarization steadily increasing.

So had Marco Rubio or Ted Cruz or John Kasich won in 2016, we’d almost certainly be looking at a very good Democratic midterm performance this year, albeit probably one with lower levels of fear and anger among the Democratic voters flocking to the polls.

And the flip side of that reality is plain as well: Had Hillary Clinton eked out an Electoral College victory in 2016, we would almost certainly be looking at a very good Republican midterm in 2018.

The size of that hypothetical GOP advantage when it comes to the House or down-ballot races is debatable, and largely dependent on things we can never know. Presumably the economy would be doing relatively well, as it was in much of Barack Obama’s second term. But in the partisan climate that existed before the 2016 election (Lock her up! Lock her up!), and that would have been intensified by divided control of the federal government and the certainty of perpetual investigations and accusations aimed at a chief executive the GOP loves to hate — the odds of a President Hillary Clinton being able to accomplish a whole lot, or enjoying robust job approval ratings, would have been vanishingly low. And that would have very likely translated into at least modest, and perhaps larger than modest, House losses, and consolidation of the already large GOP advantages in governorships and state legislatures (a big deal as the next decennial round of redistricting grows nigh).

The Senate landscape for 2018 in a new Clinton administration would have been potentially catastrophic for Democrats. As FiveThirtyEight’s Nathaniel Rakich observes, this is the most vulnerable Democrats have been in a Senate cycle since at least the 1980s:

“2018 could be not just bad, but a veritable armaggeddon for Senate Democrats. They should count their lucky stars that their worst-case map looks like it’s going to coincide with their best-case turnout environment.”

Even so, matching their best post-1990 Senate overperformance (in 2012) would leave Democrats with a net loss of two seats this autumn. And if the wind was blowing the other way, as it undoubtedly would have been had Clinton won in 2016, Democrats could have lost virtually all their red-state senators (heavily concentrated in the class up for reelection this year), leaving the party at a disadvantage in the upper chamber for a long time.

“One bad election cycle for this class could virtually eliminate red-state Democrats from the Senate in one fell swoop. That would give Republicans something like a 20-seat majority in the upper chamber — probably too wide for Democrats to overcome in any single future election cycle. Our current sense that the Senate could switch hands in any given election year would be no more, potentially emboldening the Republican majority to pass more conservative policies. In order to regain control, Democrats would not only need to rebuild their standing in red states from the ground up but also sustain that success over multiple election cycles.”

Add in some marginal considerations like the fact that with a Clinton presidency Jeff Sessions would not have resigned a Senate seat Democrats subsequently captured in a strange special election, and the “Clinton won” scenario could have endangered even routine Senate confirmations for her and possibly for future Democratic presidents. And a bad 2018 midterm for Democrats could have made confirmation of a Supreme Court appointment — the big, permanent prize associated with control of the White House — extremely difficult and perhaps even impossible.

Democrats’ fundamental problem in the Senate, of course, is the institution’s small-state bias; given the current demographic foundations of both parties, there is a built-in GOP advantage. Rakich estimates that all other things being equal, Republicans should control 62 Senate seats. It is rarely discussed, but the same factors give Republicans a natural majority in state governments, which in turn have a disproportionate control over U.S. House and state legislative redistricting. All the interminable and redundant talk about Democrats “collapsing” outside Washington to a considerable extent reflects an illusion caused by the disproportionate power of conservative states. At the moment, a fifty-fifty national electorate logically means about a sixty-forty GOP margin in the Senate and in state governments controlled.

In that context, Democrats really need a great 2018 midterm, both to avoid an almost irreversible disadvantage in the Senate, and to give it a fighting chance to do well during the momentous redistricting cycle just on the horizon. For reasons that have almost nothing to do with her leadership abilities and other sterling qualities, a President Clinton would have made a great 2018 midterm for her party extremely unlikely, and a debilitating defeat quite possible. (The same would have been true, to be clear, with a President Sanders or a President Biden). A President Trump could make an inevitably bad midterm for his party really bad.

After yet another big overperformance by a Democratic candidate in another special election, this time in Arizona, I looked at a question that a lot of observers are asking, and wrote it up at New York:

There have been two story lines this year that offer contrasting impressions about what’s likely to happen in the November battle for control of the U.S. House: One is a seemingly unending string of big-time performances by Democrats in special elections (including the one earlier this week in Arizona). The other is a Democratic advantage in the generic congressional ballot (a polling question about party preferences in House elections), which is a shadow of what it was at certain points last year.

Nate Silver posed the problem directly after the Arizona results:

“The bigger question is what to make of the disparity between the overwhelming swing toward Democrats so far in special election results — which would imply a Democratic wave on par with the historic Republican years of 1994 and 2010 — and the considerably more modest one suggested by the generic congressional ballot, which shows Democrats ahead by only 7 points and implies that the battle for House control is roughly a toss-up.”

After musing that maybe the generic ballot was a lagging indicator that might change as November approached, Silver essentially concluded that both special elections and generic ballots were data points that should both be considered, instead of choosing one exclusively.

Cook Political Report’s Amy Walters is looking at the same question:

“If a so-called “blue wave” is about to hit in 2018, why isn’t the generic ballot showing a bigger margin for Democrats? The latest Real Clear Politics average shows Democrats with a 6.5 percent lead. The FiveThirtyEight.com average has Democrats with a 6.9 percent lead. If Democrats are cruising to victory in the fall, why does the generic not look more like it did over the summer when it showed Democrats with a double-digit lead?”

Like Silver, Walters figures the numbers may turn bluer later this year. But she has a specific theory for why that could happen: the bulk of the “undecided” generic vote is among self-identified independents, and “here’s what we know about them: they don’t like Trump.”

“In the latest Marist/NPR/PBS poll (April 10-13), for example, Trump’s job approval rating among independents is 38 percent. On the generic ballot question in that same poll, the congressional Republican gets 32 percent of the independent vote. A late April Quinnipiac poll showed Trump with a 33 percent job approval among independents, and 36 percent of independents say they will vote for a Republican in the fall.”

As more independents begin to make up their minds about their midterm choices, their anti-Trump leanings will probably push up Democratic margins — an anti–White House dynamic that tends to happen during midterms anyway.

Both Walters and Silver also think the superior Democratic enthusiasm so evident in special elections hasn’t fully manifested itself in generic polls at this point. As Nate puts it:

“One plausible answer is that the generic ballot will shift further toward Democrats once voters become more engaged with the campaign in their respective districts and pollsters switch over to likely voter models.”

Now that’s certainly a switcheroo from the 2010 and 2014 midterms, when the shift to likely voter screens usually boosted GOP margins. But that does raise one point of caution about assuming Democrats will benefit from the same dynamics this November: the kinds of people who currently tend to vote Republican — older and whiter voters — have eternally been more likely to show up for non-presidential elections than the kinds of people who currently tend to vote Democratic — younger and minority voters. It’s so familiar a phenomenon that there’s a name for it: the Democratic “midterm fall-off” problem, which has been exacerbated by this decade’s exceptional polarization of the electorate by race and age.

It’s pretty clear by now that Democrats have found at least a temporary solution to the “midterm falloff” problem, and his name is Donald J. Trump. But there remain two questions: will the enthusiasm among Democrats he has created totally erase the traditional disparities in non-presidential turnout? And does the evidence that it has in so many special elections since Trump became president mean we can assume it will carry over to a regular midterm election, when Republican turnout will likely return to its “normal” levels?

Keep these questions in mind as November approaches.

I’ve been immersed in the politics of abortion policy for so long that I felt the need to take a step back and analyze how the demographics of the RTL rank-and-file have changed over time. I wrote it all up at New York:

Looking at an article in the Washington Post about the frenetic activity in many states since 2010 aimed at enacting abortion restrictions, some in order to set up a legal challenge to Roe v. Wade, the American Prospect’s Harold Meyerson noticed a pattern, which he discussed in a subscription email to readers that I happen to receive.

“Thirty-three states have enacted abortion restrictions since [2010], while just 17, plus the District of Columbia, have not.

“What interested me about those two lists was the degree to which they didn’t align with the share of Roman Catholics in the states. The eight most heavily Catholic states—in order, Rhode Island (42 percent Catholic), Massachusetts (34 percent), New Jersey (34 percent), New Mexico (34 percent), Connecticut (33 percent), New York (31 percent), California (28 percent) and Illinois (28 percent)—were among the 17 that had not passed legislation curtailing abortion rights. Conversely, the 13 states with the lowest percentage of Catholics—in order, Mississippi (4 percent), Utah (5 percent), West Virginia (6 percent), Tennessee (6 percent), Alabama (7 percent), North Carolina (9 percent), Georgia (9 percent), South Carolina (10 percent), Kentucky (10 percent), Idaho (10 percent) and Virginia (12 percent)—were among the 33 states that have curtailed access to abortions since 2010.

“In sum, the relationship between the number of Catholics in a state and the intensity of the state’s anti-abortion policies is completely inverse.”

This fact might come as a surprise to people who still think of Catholics as the bedrock core of the right-to-life movement, as they undoubtedly were in the days immediately following Roe.

In fact, Catholic public opinion on abortion policy (as on most political topics) is pretty close to that of the country as a whole, which means marginally pro-choice. Here’s how the Public Religion Research Institute put it in a 2015 survey:

“On the issue of abortion, Catholic attitudes generally mirror Americans overall. A majority (53%) of Americans say abortion should be legal in all or most cases, while 43% say it should be illegal. Among Catholics, a slim majority (51%) says abortion should be legal in all or most cases, compared to 45% who say it should be illegal.”

A more recent survey from Pew showed Catholics favoring the “legal in all or most cases” position by a slightly slimmer 48/47 margin. Both surveys showed that white Catholics — i.e., those significantly more likely to identify with the anti-abortion Republican Party — were more likely to be pro-choice than overwhelmingly Democratic Latino Catholics.

This is not — repeat, not — to say that there aren’t a lot of passionately active RTL adherents in the U.S. Catholic ranks, who can rely on the consistent support of the hierarchy and the Vatican (and yes, despite some RTL angst about his recent statement that defending the poor was as important as defending the “unborn,” Pope Francis hasn’t given much aid and comfort to pro-choice Catholics).

But there’s no question the religious community that is far more solidly in the anti-abortion camp is white Evangelical Protestants. In a 2017 survey that broke out this particular segment of the population, Pew found that 70 percent of white Evangelicals thought that all or most abortions should be illegal. Less than half of Catholics (44 percent), black Protestants (41 percent), white mainline Protestants (30 percent), and the unaffiliated (17 percent) agreed with this position.

This is remarkable in no small part because unlike Catholics, white Evangelicals have little traditional investment in the anti-abortion cause. They have no formal hierarchy, no teaching tradition, no papal encyclicals, and no “natural law” philosophy leading them in the direction of regarding abortion as grievously sinful. They purport to follow only the Bible, which never mentions abortion and only obliquely refers to fetal life. Evangelicals, moreover, were not as a group actively engaged in state efforts to keep abortion illegal prior to Roe; many (particularly among Southern Baptists, the largest white Evangelical denomination) favored “liberalized” abortion laws back then.

However you choose to explain the white Evangelical shift toward strongly anti-abortion views — as a moral “awakening” after Roe; a general rejection of liberalism and feminism; a nostalgic embrace of cultural conservatism in all its elements (including patriarchy); or a byproduct of a growing alliance with conservative politics — it’s unmistakable, and it has offset the gradual drift toward pro-choice views among Catholics.

Getting back to Meyerson’s observation, most of the states he notes as having small Catholic populations along with virulently anti-abortion policies also have large white Evangelical populations (there’s also Utah, with an LDS majority that is culturally conservative and also has a strong church hierarchy doctrinally opposed to abortion). And not coincidentally, they all (with the partial exception of Virginia) are currently Republican-run states.

The polarization of the two parties on abortion policy stems from multiple sources, but none is so powerful as the shift in the anti-abortion “base” from a Catholic population that is more or less split down the middle between the two parties (and if anything leans Democratic) to a white evangelical population that has become aligned with Republicans on a broad range of issues from civil rights to taxes to “size of government” to the cultural issues like abortion and LGBTQ rights that we associate with the Christian right.

So the archaic view of abortion as primarily a “Catholic issue” needs updating for those who want to understand why some places are so hospitable to anti-abortion politics.

During a week that featured a heavy dose of Trump scandal coverage, one of my favorite journalists, Ron Brownstein, wondered how it was all going over in different demographic groups. It was an important enough of a question that I wrote about it at New York:

[A] lot of things about the man who became the 45th president that worry upscale Republicans (and their elite #NeverTrump representatives) just don’t matter as much to the white working-class folk who have provided Trump’s sturdiest base of support. Some of it may have to do with news consumption habits: If you watch Fox News rather than read National Review, you got a very different impression of the options available to conservative-leaning voters in 2016.

It’s also entirely possible that white working-class voters are more cynical than their more highly educated counterparts about the moral tone of politicians who are not named Trump; “They’re all crooks” is a pretty common sentiment in those circles. In any event, this is a question that is not important strictly as a matter of retrospectively figuring out how a man of Trump’s character and background managed to get himself elected in the first place; as Ron Brownstein observes, it may well determine the political impact of the continuing Trump scandals we are hearing about nearly every day:

“All three national polls released this week placed Trump’s approval rating among whites without a college degree below his commanding two-thirds in 2016. But he remained positive with those voters overall, and in each survey they preferred Republicans over Democrats for Congress by at least 13 percentage points. That’s despite last week’s nonstop news about Comey’s new book; the continued sparring between Trump and Daniels, the adult film star; and the FBI’s raid on Cohen, the president’s longtime ‘fixer.'”

Trump is taking much more of a beating among college-educated white voters, who are also an important part of his coalition, and that’s not surprising. They are to some extent Comey voters:

“Comey embodies precisely the voters the GOP has been shedding under this president—even despite his unusually personal reasons to recoil from a Trump-led party. The former FBI director, after all, is a white man with a post-graduate education who’s long leaned Republican.”

Brownstein thinks this is a problem for Democrats not just because white working-class voters are relatively indifferent to evidence that Trump is a little bit piggy and a little bit thuggy. The saturation media coverage of the president’s scandals is also interfering with anti-Trump messaging about his broken promises to precisely this element of the electorate. To put it bluntly, if all these voters hear is the familiar tale they’ve heard for years about Trump’s womanizing and shady business practices, they may not hear more compelling information about Trump selling them out to Wall Street and gorging himself and his rich friends on the perks of public office.

A vote is a vote, of course, and losses among college-educated voters may (particularly if supplemented by less dramatic losses among non-college-educated voters) be enough to give Trump a black eye and Democrats control of the U.S. House. But as Brownstein notes, a significant erosion of support among Trump’s white working-class base could represent the difference between a modest and a large Democratic victory: “For a sunny outcome this fall, Democrats probably need more health care and taxes—and less Comey and Stormy.”

As we continue to absorb data on the larger-than-originally-realized size of the white working-class portion of the electorate, this is a dynamic worth watching closely. As much as the chattering classes may marvel at the ever-increasing evidence of the president’s corruption, outrage doesn’t earn the outraged any extra votes.

There was some buzz this week about the possibility that Trump might have an easier time getting reelected in 2020 than we’ve generally assumed. I looked at some historical precedents and offered a take at New York:

The belief that 2017 to 2021 is the danger zone for really serious Trumpian damage to the republic has lent some additional urgency to the Democratic drive for a big midterm victory that will neuter Trump until such time as he departs — voluntarily or under compulsion — the White House, cursing and boasting at every step.

Reinforcing this one-term assumption is the remarkable number of Republicans who will not commit to supporting Trump’s reelection, despite the GOP’s largely supine surrender to his takeover of their party.

But Kyle Kondik of Larry Sabato’s Crystal Ball has published a warning to the complacent based on predictive models (mostly drawn from international samples) that suggest the power of incumbency is a bigger factor in reelection contests than is normally believed.

“[A]ssuming Trump is on the ballot, and assuming his approval rating stays around the 40% mark, it would probably be wrong to assume he’s an underdog for reelection. That’s not to say he would be a sure winner, but he wouldn’t be a sure loser, either.”

Placing a higher-than-normal value on incumbency has the benefit of helping explain why Trump managed to win in 2016: There was no Democratic incumbent, and moreover, it was (to use the title of political scientist Alan Abramowitz’s model, which predicted a Trump win) “Time for a Change,” since Democrats had held the White House for two terms. Here’s Abramowitz’s general take on how incumbency matters initially and then erodes as a party continues to hold the presidency:

“[C]andidates running for reelection after only one term in the White House enjoy a substantial advantage. In fact, in the past hundred years there has been only one election in which a party lost the White House after only one term — the 1980 election in which Jimmy Carter lost to Ronald Reagan. After two or more terms in the White House, however, it appears that this advantage disappears. Even if the incumbent president is a candidate, there is no incumbency advantage in third or later term elections. As a result, these third or later term elections tend to be highly competitive. And of course 2016 is another third term election.”

And 2020 is another second-term election — the kind an incumbent party rarely loses.

That’s one way to look at the historical record. But there are others.

Jimmy Carter wasn’t just the only second-term incumbent to lose; he was also the only one (since the 1940s, when presidential-approval-rating polls became available) to have Trump-like approval numbers. According to Gallup, here are the preelection job-approval numbers for presidents facing voters after their party had just one term holding the White House: Eisenhower ’56: 68 percent; Johnson ’64: 74 percent; Nixon ’72: 56 percent; Reagan ’84: 58 percent; Bush ’04: 53 percent; Obama ’12: 51 percent. Carter’s final job-approval number before the 1980 election was 37 percent.

Using the same source (Gallup), Trump’s highest approval rating was registered on the week of his inauguration, at 45 percent. Unlike his predecessors, his approval ratings appear to have a very limited range. So it is not at all clear, unless you really value non-U.S. examples, that Trump would be an even bet for reelection if he doesn’t become more popular.

There are, of course, four crucial variables we cannot possibly know this far away from the 2020 election: the possibility of a disabling scandal like the one that swept away Richard Nixon’s presidency less than two years after he carried 49 states; the performance of the economy (crucial to many reelections); the possibility of intra-party opposition (Johnson ’68, Ford ’76, Carter ’80, and Bush ’92 were all incumbent candidacies damaged badly by primary opposition); and the identity of the Democratic nominee.

You’d have to say at this moment that Trump’s reelection prospects are most definitely threatened by ever-emerging scandals, a likely pre-2020 economic downturn, and a Democratic opponent not as thoroughly unpopular as Hillary Clinton was in 2016. A primary opponent is less likely unless Trump otherwise looks like a loser.

On top of everything else, it’s worth remembering that Trump lost the popular vote in 2016 by more than two percentage points, winning basically the same percentage that big-time loser John McCain won in 2008. Yes, he won the electoral college via the political equivalent of an inside straight, but pulling that off a second time is significantly less likely.

Anyone would be foolish to write off Trump’s reelection prospects entirely. But even with the advantages of incumbency, he’s going to have to do a lot better than he has so far to stay in the White House beyond 2021. And the sense that this strange man is never more than an inch from the political precipice is not entirely the product of his critics’ wishful thinking.