Heading into the midterm elections, it’s pretty clear that a blue wave is coming. But is the shellacking the GOP is likely to take in November a sign that Donald Trump, and his brand of ethno-nationalist populism, will soon be on the way out? While many pundits (at least, those who aren’t making appearances on Fox News) are leaning that way, political scientists tell us not to count on it.

First of all, the party represented by first-term presidents almost always takes a beating in the midterms. And those beatings have not served as reliable indicators for what is to come. Republicans were creamed in 1982, but Ronald Reagan won an historic landslide in 1984. Democrats were slaughtered in 1994, yet Bill Clinton went on to win decisively in 1996. Democrats were decimated in 2010, but Barack Obama broke little sweat in winning again in 2012. Republicans won’t be wrong to dismiss a blue wave in November as saying little about Donald Trump’s prospects in 2020.

And that’s particularly the case if we look at the underlying ideological and partisan sentiments of Republican and Democratic voters. Prof. Larry Bartels of Vanderbilt University did just that in a revealing working paper titled “Partisanship in the Trump Era,” released last month. In it, Bartels closely examined a November 2017 YouGov survey of 2,000 people, all of whom were originally interviewed in 2015 and 2016 as part of YouGov’s 2016 Cooperative Campaign Analysis Project. After putting all three of those survey findings through rigorous regression analyses, a great deal of conventional wisdom about what’s going on in the Republican Party was blown to bits:

1) Donald Trump has not remade the Republican Party in his own image. That’s both good and bad news for Never-Trumpers. There is no evidence whatsoever that party demographics or political views have changed significantly since he came on the political scene in 2015. To whatever extent the party has changed (e.g., the migration of working-class whites into the GOP), that change occurred before Donald Trump. Accordingly, if you wish to cast blame on what the Republican Party has become of late, blame probably lies with the strident Republican backlash to Barack Obama and the right-wing media apparatus, not with Donald Trump.

2) Contrary to what we’ve seen in some other recently published surveys, Bartels finds only a tiny change in the number of voters who identify with the Republican Party as a consequence of Donald Trump. The oft-searched-for migration of voters out of the GOP is extremely difficult to spot.

3) Those very few Republicans who have left the party since Trump came along are no better educated than those who have remained, they aren’t especially averse to racially charged cultural appeals, not particularly engaged in politics, and they weren’t significantly more negative towards Donald Trump relative to the those who have stayed. Democrats and Never-Trumpers who are waiting for socially tolerant, well-educated Republicans to bolt from the GOP in horror are still waiting.

4) Despite the incredible nature of the Trump’s behavior and conduct since he’s moved into the White House, opinions about him have hardly changed a bit. What little change we’ve seen appears to reflect confirmatory partisan bias. When asked whether Donald Trump is intelligent, a strong leader, knowledgeable, inspiring, or moral, Trump’s average rating across all five questions declined by less than one point among Independents over the course of time, but declined by 4.4 points among Democrats and increased by two points among Republicans. In fact, the number of all voters who viewed Trump as “moral” increased by 2 points from mid-2016 to late-2017. Chew on that for a while.

5) Only about 19 percent of Republicans are closer to the Democratic Party on cultural issues than they are to the GOP. And Republicans care a lot more about cultural issues than they care about limited government. There’s a reason that the rank-and-file of the various libertarian(ish) Tea Party organizations defected en masse to Trump’s campaign during the 2016 primary season. While most of Trump’s Republican rivals were talking like Charles Koch, Donald Trump was talking like George Wallace. And the latter proved far more compelling than the former.

6) Republicans are nowhere near as hostile to government as you may think. A majority of Republicans support government action to regulate pollution, provide a decent standard of living for people unable to work, and ensure access to quality health care. And a substantial minority of the party favors reducing income differences and helping families pay for child care and college. All told, 24 percent of Republican voters are closer to the Democratic Party on economic issues than they are to their own party. The agenda forwarded by Republican mega-donors simply doesn’t resonate that well with rank-and-file voters.

7) Republicans have significantly cooler feelings towards their party than Democrats have towards their party. Republicans even have somewhat cooler feelings towards Donald Trump than Democrats had for their two past presidents when they were in office (Bill Clinton and Barack Obama). And only 44 percent of Republicans “strongly approve” Trump’s presidency. Even so, Republican disapproval of Donald Trump is relatively nonexistent. There is no sign of an emerging split between pro-Trump and anti-Trump sentiment in the GOP.

8) However, rank-and-file Republicans overwhelmingly side with Donald Trump when he gets into fights with Republican congressional leaders. That has less to do with their warm feelings for Donald Trump (which, as noted, really aren’t all that warm), and more to do with their remarkably low opinion of the Republican congressional leadership. And in these fights, it’s the conservatives—not the moderates—who are most likely to support Trump. If it’s Trump’s party, Bartels notes, it is largely by default.

Bartels also shatters some of the conventional wisdom about what’s going on in the Democratic Party:

1) There were no important ideological differences between Bernie Sanders voters and Hillary Clinton voters. Nor is there any evidence that a surging progressive tide of Sanders-style democratic socialism is rolling through the party. Looking forward to 2020, those who have warm feelings toward Bernie Sanders were actually slightly more conservative than those who had warm feelings for Joe Biden!

2) The fault lines in the Democratic Party pertain primarily to social identities, interest group affiliations, and internecine antipathies, not ideology. Struggles are about which elements of the party coalition deserve more time and energy than others. Bernie Sanders’ appeal to young people was more about him being a maverick outsider (which always appealing to the young) than about any particular element of his campaign platform.

3) Unlike Republicans, Democrats care a lot more about government action to address economic problems than they care about cultural issues. And only 11 percent are closer to the Republican Party on economic issues than they are to their own party. That said, 26 percent of Democrats are closer to the Republican Party on cultural issues.

And there are even a few important things we learn about Independents:

1) Once we remove from the sample those Independents who usually vote for one party or the other (that is, those who are functionally no different from Republicans or Democrats but who, for whatever reason, eschew the party tag), Independents are all over the lot when it comes to questions surrounding economic and cultural issues. Their opinions are not clustered in the center of American politics.

2) True Independents (that is, voters who are not really Republicans or Democrats) are not drifting towards the Democrats because of Donald Trump.

3) Independents are unlikely to change the partisan balance of power. 31 percent of them are closer to Republicans on both economic and cultural issues, while 28 percent were closer to Democrats on both of those dimensions. If Independents made peace with their true partisan identities, it would prove a wash. The remainder are such a heterogeneous lot that neither party would have an advantage in competing for their votes.

How do we square Bartels’ findings with the big blue wave that appears to be coming? Well, intensity counts, especially in low turnout midterm elections. Amongst those animated enough to go to the polls in November, hatred of Donald Trump burns with the white-hot intensity of a thousand suns. Trump’s hard-core supporters are still there, of course, but their intensity burns less hot (for now).

Moreover, Donald Trump is not on the ballot. His supporters, as noted, are markedly less enthusiastic about the Republican Congress than they are about him. So Republican support for Donald Trump does not fully translate to support for Republican office-holders, especially if they’re anything less than uber-Trumps.

The big take-away is that, despite the political and cultural earthquakes triggered by the rise of Donald Trump, the political terrain as it existed in 2016 has remained remarkably stable. The political sentiments that allowed Trump to win the White House are still there and strong as ever.

Nor do those sentiments and party attachments show signs of changing any time soon. Sure, a sizeable number of the Republicans who are closer to the Democrats on economic and regulatory issues could bolt. But so could a sizeable number of Democrats who are closer to the Republicans on cultural issues. Even if a resorting of party allegiances were to occur, it would likely prove a wash.

If the saliency of economic or cultural issues were to change, however, things could get interesting. If economic issues were to rise in importance, that would greatly strengthen the Democrats. If cultural issues were to rise in importance, then a resorting of partisan allegiances would greatly strengthen the Republicans.

Many believe, however, that the GOP’s day in the sun has an expiration date that will be come due with the full arrival of Millennial voters. Maybe, maybe not. Bartels finds that, contrary to conventional wisdom, Millennial Democrats and Millennial Republicans aren’t any more liberal than older members of their respective parties when it comes to issues surrounding the role of government. They are much more liberal than older generations, however, when it comes to cultural issues. Millennials, however, care a lot more about economic issues than they do about cultural issues. So how they will break is difficult to say. “There is no evidence here of broad secular change in views about the role of government,” reports Bartels, “or any hint of the existence of a ‘really quite left-wing’ cohort of ‘young progressives’ poised to deliver the Democratic Party ‘to the promised land.’”

Likewise, how long Millennials will hold their beliefs is unclear. Surprisingly enough, the political science literature is rather thin and somewhat contradictory regarding how much “stickiness” there is to partisan and ideological sentiment as voters age over time.

Now it should be noted that other surveys of Millennial attitudes, most notably those conducted by Pew, detect a greater leftward tilt than does Bartels in the course of his examination of the YouGov surveys. Regardless, given that young people always have, and likely always will, turn out at lower rates on election day than older voters, it will take at least a few more election cycles for Millennials to significantly change the political landscape. And when they do, the Republicans could accommodate tomorrow’s voters (if they so wish) by deemphasizing cultural warfare and returning to bread-and-butter issues pertaining to the role of government, a place they’ve certainly been before.

All in all, the picture for the #Resistance doesn’t look that great. But while the landscaping exercise undertaken by Bartels is convincing, there’s more to electoral outcomes than navigating elemental public sentiment.

As political scientist James Stimson has convincingly demonstrated, America has long been, and continues to be, a slightly center-left nation. Voters positively respond to conservative platitudes and issue framing at the 10,000 ft. level, but are operationally liberal when it comes to concrete policy matters. That’s why Republicans campaign and govern in ideological broad strokes, while Democrats campaign and govern as policy wonks. Even so, Democratic administrations generally give voters more government than they want, and Republican administrations give voters less government than they want. Voters tend to respond to this thermostatically. The further away from the center-left the White House is perceived to be, the stronger the political counter-response.

Donald Trump is forwarding the most hard-right policy agenda the nation has ever seen. If past is prologue, and all things remain equal, a day of reckoning will come. The public, however, usually takes a while to notice that the White House is giving them more or less government than they want (which isn’t surprising given that most people aren’t very interested in politics and politicians usually labor to sound as moderate and nonthreatening as possible). Trump’s dominance of the news cycle and determination to stomp moderation into bloody bits (a determination shared by an increasing number of his congressional allies) will likely reduce the signal-to-noise ratio and give the Democratic nominee some extra muscle in 2020.

A second caveat is that economic performance (in presidential elections, anyway) swamps almost everything else. Political landscaping exercises, while interesting and useful, are less important than economic fundamentals. That’s because voters have traditionally chosen presidents based (largely) not on ideological or policy criteria, but on how well the economy is perceived to be doing. The election forecasting models built by political scientists are premised on these considerations, along with other non-ideological matters like how long an incumbent party has been in power (politicians and political parties tend to wear out their welcome), the president’s approval rating (which is usually heavily affected by how the public feels about the economy), etc. Those models, which pay no direct attention to the ideological landscape, have far better track records at forecasting electoral outcomes (even in 2016) than forecasts that put a lot of weight on polling data and/or where the public stands on this issue or that. In fact, when these political science models are re-run for past elections, political scientists find no evidence that presidential nominees have ever paid a price (or received a bonus) for ideological extremism.

Last but not least, there may be tipping points ahead. Donald Trump is taking us into uncharted political territory. Never before has such a divisive, scandal-plagued figure occupied the White House. Nor have we ever witnessed a president openly flirt with authoritarianism. So far, the public has either not noticed or not cared about matters that would have sunk previous presidents. But how long can that last?

President Donald Trump is no ordinary President. He may prove to be such a singular outlier in the annals of American politics that voter responses will likewise break historical patterns. What’s sobering about Bartels’ paper, however, is that we see no signs of that. Yet.