Defending democracy is easier when you listen to voters

The worst outcomes are less likely if Trump is less popular. So Democrats need to win the battle of public opinion.

This is a long post, so I’m going to put a subscribe button right up here at the top. But today’s newsletter is free, considering the importance of the debate.

---

Last week, Vice President JD Vance lamented that the “entire American media and left wing industrial complex” had decided that “the most important issue” was the deportation of Kilmar Abrego Garcia, a Salvadoran immigrant who entered the United States illegally but — in what the executive branch described as an “administrative error” — was deported to a Salvadoran prison despite holding a legal status called “withholding of removal” that was supposed to preclude exactly that outcome.

This came after the Supreme Court had partially upheld a lower court order stating that the government should “facilitate” Abrego Garcia’s return to the US. Legal scholars differ on the extent to which Trump’s actions so far constitute defying the Supreme Court, because the ruling was ambiguous (drawing a distinction between “facilitate” and “effectuate”) in ways that some coverage of this case is obscuring; I’ll address that later.

What’s clearer is two things: first, neither the administration nor President Nayib Bukele of El Salvador, who met with Trump at the White House last week, has shown any interest in returning Abrego Garcia to the United States. Although the White House claims its hands are tied, El Salvador has approximately one-thoudsandth¹ of the US’s GDP and less than one-hundredth of the United States’ military might. You don’t have to be versed in The Art of the Deal to understand that Trump holds the cards in this relationship and could facilitate whatever he wanted to.

And second, this is a political fight that the White House thinks it can win, or at least battle to a draw. Vance, despite his protests, was happy about being tossed into this particular briar patch. In the four days following his “left wing industrial complex” tweet, he tweeted or replied to tweets about Abrego Garcia’s case another 9 times. The White House would much rather this story dominate the political conversation than the economy.

Public opinion is the best mechanism to fight Trump’s excesses

Let’s back up a second, because other than what I’ve written in those first four paragraphs, many other facts about the case are murky. We’ll return to those details, but first, I want to establish the stakes.

Republicans, as I’m sure you know, essentially have unified control of the three branches of the federal government. I say “essentially” only because the Supreme Court is by no means entirely on board with Trump, instead belonging to a more old-school, establishment-friendy faction of the GOP. But Trump is president, and the Congress is Republican and so far has been highly deferential to Trump despite his relatively narrow majorities. And even if the Supreme Court turns against Trump, we’ll be in real trouble if he gets in the habit of defying court orders.

By “we”, I mean the country. I can voice strong opinions in this newsletter, but I don’t consider myself a partisan. You’ll find plenty of criticism of Democrats in Silver Bulletin when it’s warranted; indeed, this post contains some of it.

When Trump’s second term began, I issued a series of 113 probabilistic predictions for how it would go, partly as a way to calibrate my expectations. Although it will be four years before we can score those predictions officially, and while I think I’ve gotten some things right — for instance, that there would be a political backlash against Trump and that he’d implement severe tariffs on Canada and Mexico — more things than not are running in the wrong direction. The predictions column didn’t take a Pollyannish view, exactly. Nevertheless, Trump has both been a chaos agent and extremely aggressive in using and abusing the powers of the executive branch. To put this a bit clinically, on a scale from -100 to +100, where -100 is Trump’s term going much better than I expected and +100 much worse, I’d say we’re probably at a +75 or a +80 so far.

So what can Democrats — or non-partisans who don’t like Trump — do about it? Well, they can win elections. Democrats holding even one branch of Congress after the midterms would constrain Trump to some extent. And then there’s another presidential election in 2028 after what’s sure to be an incredibly contentious nomination fight. I don’t think we’re yet at the point where we have to worry about elections being rigged or canceled — that would be more like a +95 on the scale. (Actually, let me revise that: I do think you should be worried, I just don’t think it’s the default outcome.)

In the interim, Trump will be more constrained in various ways if public opinion turns against him:

• How responsive Trump himself is to public opinion is unclear, but the answer is not “zero”. Trump backed down — for now — from the most severe tariffs after a harsh reaction from the markets and in public opinion polls. The White House withdrew its nomination of Elise Stefanik for UN Ambassador out of fears it could lose a Trump +21 district. It might be about to fire Pete Hegseth. And Trump has mostly stood down on abortion, one of Republicans’ worst issues in polling.

• Congress is also responsive to public opinion. Because of political polarization, voter self-sorting, and gerrymandering, it’s less responsive to it than the founders probably intended. (Or, members are more concerned about how the median Republican voter feels than the median American, because they face a greater threat of losing a primary challenge than a general election.) But on tariffs, where polls run strongly against Trump even among his base, Congress might eventually be compelled to action. Trump’s approval ratings bottomed out in the high 30s in his first term, and although his downside is protected by extreme partisanship, I don’t take for granted that’s his floor. If his numbers were to fall much below that, you might get a downward spiral like George W. Bush experienced in his second term.

• Courts, and particularly the Roberts court, are also attentive to public opinion.

• So are elites: not just “fourth estate” actors like the media, but also institutions such as the business community.

• Then there are local officials, such as governors and mayors, who could become increasingly emboldened to challenge Trump as he becomes more unpopular.

• Finally, to the extent that there are disagreements within the executive branch, individuals selected to nonpartisan positions or who are holdovers from previous administrations can also take a stand. I hope it doesn’t come to this. However, if we get into some of the darker timelines, resistance could run both ways: that is, other parts of the system defying the executive, as well as Trump ignoring legal or legislative intent. In Abrego Garcia’s case, there’s an element of “finders keepers”: since he’s currently in El Salvador, it takes active effort to get him back. But that won’t always be the case.

I’d argue, therefore, that the most important indicator to track, at least until a year or so from now when we can start to get a better read on the midterms, is simply Trump’s popularity. And those numbers tell an interesting story:

Trump isn’t popular. His approval ratings are now firmly in negative territory; the only president to be less popular at this stage is Trump 1.0. But the decline in his numbers — initially, most Americans gave him the benefit of the doubt — has proceeded in a few distinct stages:

• There was a sharp decline in the first month or so, which is typical of a new president’s “honeymoon period” wearing off.

• This was followed by a steadier and more linear decline.

• Then there was an abrupt downshift early this month, which was almost certainly the result — there’s enough polling to provide for this sort of precision — of the “Liberation Day” tariffs and the concomitant downswing in the markets.

• Since then, however, his numbers have steadied.

Because I’m a betting man, I’ll tell you that if I had to put my money down, I’d expect a further decline rather than a rebound, especially if and when the tariffs begin to influence the price of consumer goods. And Monday was another terrible day for the markets, with Trump attacking the chairman of the Federal Reserve.

Still, we’ve now seen two salient data points here. First, that Vance — who unlike Trump, isn’t a lame duck — welcomes the fight on Abrego Garcia. Believe me when I say I wouldn’t award Vance an A+ for his political instincts. But it’s something. And second, since the “main thing” has shifted from the tariffs to Abrego Garcia’s case, the decline in Trump’s approval ratings has halted.

There’s no playbook for this

Abrego Garcia’s case has opened up a debate among Democrats about whether and how forcefully to respond. Maryland Sen. Chris Van Hollen traveled to El Salvador to meet with Abrego Garcia last week. Predictably enough, with the Salvadoran government having every incentive to make Van Hollen look bad, the optics of this were not wonderful; instead, concoctions that appeared to be margaritas were mysteriously placed on the table partway through Van Hollen’s meeting with Abrego Garcia. Meanwhile, other Democrats have expressed concern that the case is playing on Trump’s home turf: his approval ratings are much stronger on immigration than on the economy.

I feel conflicted about this. Honestly, I was thinking about sitting this story out. But part of what you’re counting on me for — and in some cases, even paying me for — is weighing in on high-stakes political controversies even when the response is sure to piss off half the Silver Bulletin reader base.

And as I’ve thought more about this story, my opinion has become firmer. Look, nobody knows anything; we’re all just guessing. The Abrego Garcia story is highly newsworthy, and people should protest against whatever motivates them. The case is within the margin of error, so to speak. But as a matter of pure political strategy, I mostly think the critics are correct and that Democrats are picking the wrong fight here.

Let me start by stating what I consider to be an unpleasant fact. Trump instigated an insurrection against the orderly transfer of power on January 6, 2021, and then pressured officials to overturn the results of the 2020 election. Nevertheless, he was reelected, even winning a plurality of the popular vote, four years later.

Joe Biden, and then his replacement Kamala Harris, heavily emphasized “democracy” in their campaigns. I have considerable sympathy for Harris — far less for Biden — and the reasons for her loss were overdetermined. Nevertheless, even if Abrego Garcia’s case is really about democracy — due process, and the separation of powers rather than about immigration per se — I’m not sure that’s a winning issue for the anti-Trump coalition. What sort of voter is going to turn against Trump because of Abrego Garcia when they didn’t because of January 6?

I’ve also been frustrated by what I see as Democratic efforts to put some spin on the ball. Look, it’s politics, so spinning issues in your direction is fair game. However, this can be an indication of a politically tenuous position. I respect Substack GOAT Heather Cox Richardson, but when she writes that “the U.S. Supreme Court unanimously upheld [District Judge Paula Xinis’s] order,” that’s at least a little bit misleading. Let me go ahead and quote at length from Justice Sotomayor’s opinion just because I’m nervous about leaving something out if I summarize it:

The application is granted in part and denied in part, subject to the direction of this order. Due to the administrative stay issued by THE CHIEF JUSTICE, the deadline imposed by the District Court has now passed. To that extent, the Government’s emergency application is effectively granted in part and the deadline in the challenged order is no longer effective. The rest of the District Court’s order remains in effect but requires clarification on remand. The order properly requires the Government to “facilitate” Abrego Garcia’s release from custody in El Salvador and to ensure that his case is handled as it would have been had he not been improperly sent to El Salvador. The intended scope of the term “effectuate” in the District Court’s order is, however, unclear, and may exceed the District Court’s authority. The District Court should clarify its directive, with due regard for the deference owed to the Executive Branch in the conduct of foreign affairs. For its part, the Government should be prepared to share what it can concerning the steps it has taken and the prospect of further steps. The order heretofore entered by THE CHIEF JUSTICE is vacated.

It wasn’t a clean rebuke of Trump. Rather, Xinis’s ruling was “granted in part and denied in part”. I agree with Jack Goldsmith’s interpretation that the court is trying to find a middle ground here and perhaps fire a warning shot to prevent Trump from doing this sort of thing again. Because how a court would go about compelling a president to “effectuate” Abrego Garcia’s return, as the district court ordered, isn’t so clear. In practice, Trump has considerable leverage over Bukele. But if El Salvador was determined not to give him back, is the commander-in-chief then required to conduct a military operation to effectuate his return? SCOTUS probably doesn’t want to go there.

You can see this pattern in online and IRL conversations, too. I haven’t been super active on Twitter lately, and I don’t want to get too deeply into conversations that I’ve had with friends about the case. But I’ve noticed a familiar pattern: a default toward partisan-flavored “stylized facts” about the case that are typical when public opinion lines up about 50/50 on an issue. This is in contrast to the tariffs, where public sentiment runs incredibly and uniquely strongly against Trump.

Let me present you with one further exhibit. This was the New York Post’s homepage when I took a screenshot of it late Friday afternoon. I consider the Post to be an interesting indicator because it’s mostly pro-Trump — but the Post is published in a blue part of the country, and takes itself seriously enough as a journalistic institution that it can occasionally buck him. On the Abrego Garcia story, it was toeing the GOP party line:

Conversely, here was the Post’s homepage the day after the “Liberation Day” tariffs were announced. This is extremely unfriendly coverage for Trump: the headline was literally “LOSING BATTLE”:

So, if you’re a Democrat, which of these fights would you rather pick?

27 minutes per day

Of course, I’ve seen Democrats deny that there’s any sort of choice to be made. They claim that there’s no trade-off in “focus”: the party can walk (pick a fight over tariffs) and chew gum (Abrego Garcia, too) at the same time.

I’ve been in the news business a long time, and I think this is a misdiagnosis of how it works. True, professionals whose job it is to cover politics for a living can walk, chew gum, and tap dance at the same time. That the New York Times is covering the death of Pope Francis doesn’t mean it’s ignoring either the economy or Abrego Garcia, for instance.

But most Americans aren’t like this. Perhaps the biggest bias among people who work in news or politics is the failure to recognize the limitations of normie attention spans. Most people, even those who vote, rely on vibes, snap judgment, partisan cues, and stylized facts. If you’re working, raising a family and so forth, it’s hard to stay on top of one news story at a time, let alone two or three.

How much time does the average voter spend following the news? Well, let’s try to estimate that. The American Time Use Survey, which I’ve cited before in the newsletter, simply asks a random sample of Americans aged 15 and older to track how they budget their time. There’s no designated category for reading or watching news, but there are a few places where it might slot in:

• “Watching TV”: Americans spent 2.67 hours per day watching television in 2023, according to the ATUS;

• “Reading for personal interest”; this is just 0.26 hours (about 16 minutes) per day;

• And “computer use for leisure, excluding games”: this is only 0.20 hours per day, or 12 minutes.

But of course, people watch and read all sorts of crap, celebrity news, sports coverage and so on. I asked ChatGPT² to estimate how much time from each of these buckets are spent “consuming news or current affairs programming”. It ballparked this at 12-15 percent of total TV viewing time, 10-15 percent of non-gaming computer use, and 25-30 percent of personal reading time. Using the midpoint of those estimates, that works out to 0.45 hours per American per day on news coverage, or about 27 minutes. This number hasn’t changed very much over time, even as media consumption patterns have shifted:

Probably about half that is watching Fox News, CNN or MSNBC: on average, about 1 percent of Americans are tuned into one of these networks at any given time, which works out to around 14 minutes per day, per voter. And the median news consumption is undoubtedly lower than the average, since many people, particularly retirees, leave the cable news channels on all day. According to the ATUS, in fact, only about one-sixth of Americans read anything for personal interest on any given day. So you have a large share of Americans who don’t read the news at all, then some who are extremely heavy consumers. In between, the median is 10 to 15 minutes of news consumption per day, ChatGPT estimates.

So if you’re going to try to persuade people, you don’t have that much time to do it. You want the stylized facts and snap impressions to be firmly on your side. The important exception is if the news directly affects people’s lives. That’s another reason why tariffs are such a strong issue for Democrats: people will notice their impact on the prices of consumer goods, and also on their 401Ks, etc.

Abrego Garcia is not an ideal test case

Another claim I strongly disagree with is the idea that, although Trump may enjoy relatively strong approval ratings on immigration overall, people start to side with the Democrats when you ask them about particulars, like about due process for people in deportation hearings.

I have several objections to this. First is what I articulated above: people don’t have time to sort through the particulars, so instead they default to their priors, e.g., that Republicans aren’t trustworthy on health care and Democrats aren’t trustworthy on immigration. Democrats are sometimes accused of being too “poll-driven.” But I think the obsession with “messaging” or the micromanagement of policy positions is the worst form of this — thinking that you can weave straw into political gold if you frame an issue in precisely the right way rather than looking at public opinion in broader strokes.

Trump won an election on immigration less than six months ago: it was probably his single best issue after inflation. It’s going to take a lot to overcome that presumption with a largely inattentive public. Vance’s argument is basically that Trump was elected with a mandate to fix a broken immigration system. The subtext is that if a few mistakes are made along the way — well, he’s not exactly concerned about that.

Second, there’s inevitably some degree of cherry-picking when people cite these polls, which sometimes contain leading questions or are published by interest groups. If you poll a bunch of specifics about immigration and people side with Democrats, but then turn around and tell you they nevertheless trust Republicans more on the issue overall, perhaps the survey isn’t asking the right questions, or there’s something else it’s not capturing.

And third, in Abrego Garcia’s case, I’m not sure that voters actually will become more sympathetic to Democrats if they spend more time studying it.

At a minimum, voters are left to evaluate some strongly competing claims. Vance described Abrego Garcia (not even mentioning him by name) as an “MS-13 gang member (and illegal alien)”. The Atlantic, conversely, described him as a “Maryland father with protected legal status” who “came to the United States at age 16 in 2011 after fleeing gang threats in his native El Salvador.”

I’m just going to be honest here: when I first started reading about this story, I assumed that Vance and conservative news sources were, at best, grossly exaggerating. But having invested the time to learn more about the underlying facts, I’m not sure where the truth lies beyond that his deportation was both mistaken and illegal.

Liberal news outlets often take claims made by Abrego Garcia’s attorneys at face value and downplay some unfavorable aspects of the case. The BBC has one of the more neutral summaries about the facts, and some of them aren’t so flattering. Take this, for instance. Two judges concluded that there was sufficient evidence to conclude Abrego Garcia was a gang member:

According to his lawyers, Mr Abrego Garcia has never been convicted of any criminal offence, including gang membership, in the US or in El Salvador. He lived in the US for 14 years, had three children and worked in construction, according to court records.

But the judge who presided over his 2019 case said that based on the confidential information, there was sufficient evidence to support Mr Abrego Garcia's gang membership. That finding was later upheld by another judge.

Or this: Abrego Garcia’s wife filed a protective order against him, alleging domestic violence, which she’s since declined to pursue:

In 2021, his wife, Jennifer Vasquez Sura, filed a protective order petition against him, alleging that he had physically attacked her on multiple occasions, according to documents shared by the US Department of Homeland Security.

Ms Vasquez Sura said in a statement on 16 April that she had decided not to follow through with the court process at the time and that she and her husband "were able to work through this situation privately as a family, including by going to counseling".

Or that the reason Abrego Garcia wasn’t deported is that he feared persecution from a rival gang to MS-13 if he was sent back to El Salvador, a claim the government found credible enough to give him “withholding of removal” status, a considerable step short of asylum:³

In October 2019 he was granted a "withholding of removal" order, court documents show - a status different from asylum, but one which prevented the US government from sending him back to El Salvador because he could face harm.

Mr Abrego Garcia's lawyers say that he was granted the status based on his "well-founded" fear of persecution by Barrio-18, the main rival gang of MS-13.

He said that prior to him entering the US, his family and their business had been threatened and extorted by Barrio-18.

When political groups seek to influence judicial precedent, they look for ideal “test cases” where the facts are incredibly sympathetic to their plaintiffs. For instance, in the Obergefell decision that resulted in the legalization of same-sex marriage, Jim Obergefell’s partner was terminally ill and suffered from ALS. Abrego Garcia is far from the ideal test case, conversely. That might or might not matter to the courts, but it helps to explain why Vance and Trump are willing to battle in the court of public opinion.

Trump will give his opponents plenty of better opportunities

What if Trump deports an American citizen to a Salvadoran prison, as he’s threatened to do? Or what if he more clearly violates a Supreme Court order, with no plausible deniability about the daylight between “facilitate” and “effectuate”?

Well, it’s pretty simple: that’s when you pull out all the stops, call your member of Congress until her phone is ringing off the hook, and take to the streets in protest if it comes to that.

Because when even Trump’s supporters call him a “bull loose in a china shop”, when he’s 78 years old and doesn’t face re-election, and when he’s surrounded by people like Hegseth rather than more competent advisors, he’s going to do any number of things that are both really bad for democracy and really unpopular, where public opinion runs 60/40 or even 80/20 against him.

What about deterrence or the notion of drawing a line? Yes, that’s important if you’re the courts or Congress — which mostly means just the courts, since Congress has abdicated its responsibilities.

However, the idea that we must take action now, or else things will get worse, is a notoriously hard sell to the public, whether it comes in the context of climate change or the national debt or the erosion of the rule of law. Usually, something has to break before a mass bipartisan consensus forms to fix it. The difference with Trump is that, for better or worse, we may not have to wait very much longer: many Americans think he’s already broken the economy.

But I don’t think you can circumvent this problem with the right speeches or photo-ops. Instead, Democrats suffer from a fallacy that I’ve called The Big Cope, which is their belief that if only the public knew all the facts, it would become incredibly hostile toward Trump. There are two big problems with this. First, however informed or misinformed the public is, that isn’t going to change much until and unless they believe their material interests are threatened, as they are by tariffs, but, with some exceptions⁴, not by Abrego Garcia’s case.

Some righteous indignation about 2024

Also, Democrats’ views can be shaded by their partisan blinders. Most pointedly, extensive reporting on Joe Biden’s condition has revealed that it was actually worse than even critics like me assumed, and was progressively deteriorating; Harris should probably have transitioned into the presidency at some point in 2022 or 2023. There’s been a trope lately that what I sometimes think of as the “resistance libs” were right about everything: that America was approaching a moment of crisis, that Trump is a hard-core authoritarian, and so on. But while they’re looking smarter about this diagnosis, I’m not sure they’ve identified the right cure or have canny instincts for political strategy.⁵ I’m sure there are exceptions, but almost to a person, this group was extremely supportive of Biden’s bid to run for re-election, for instance.

It would be far too self-serving, however, to suggest that the “sensible center” has been right about everything, either. This group’s tendency toward moderation and incrementalism, even if it’s right more often than not, can make them frogs in boiling water. And waiting for the perfect moment can be the enemy of good opportunities to be more politically aggressive. I surprised some readers recently by disagreeing with the center on the issue of shutting down the government — I thought Democrats should have done it.

However, the reason I thought that is because I gamed out the scenario and believed that Trump and Republicans would eventually bear the lion’s share of the blame, and that this would also increase the salience of other unpopular actions they were taking.

Precisely because the stakes are so high in terms of how Trump could undermine the Constitution and the separation of powers, I think his political opponents ought to be fairly consequentialist in their thinking. Initially re-nominating Biden — and to a lesser extent, replacing him with Harris rather than a figure who had more distance from an unpopular president — was a mistake of the highest order if you think four more years of Trump is as much of a threat to the Republic as the resistance libs do.

Yes, these are confusing times for which there’s no good playbook. And when the political consequences are ambiguous, I say go ahead and default to principle. Maybe Abrego Garcia’s case falls into that bucket. I wouldn’t say I think it’s a winning issue for Republicans so much as that tariffs are a massive loser — and that I think there are real trade-offs in the public’s attention span and some risk of boy-who-cried-wolf syndrome.

But compare and contrast those New York Post headlines once again, or Vance’s eagerness to discuss Abrego Garcia’s case against his relative silence on tariffs. Or how Trump’s approval rating suddenly stopped declining when the subject changed from the economy to immigration and democracy. The whole point of a democracy is rule by the public, not by experts or oligarchs. So when public opinion provides clues like these, even in an uncertain environment, Trump’s opponents ignore it at their peril.

1. CORRECTION: This originally read as “one-hundredth”.

2. Yes, I think ChatGPT is good at this sort of task where there’s no hard data to work with so instead it has to synthesize an estimate from various sources.

3. For instance, with “withholding of removal” status as opposed to asylum, Abrego Garcia could be deported to a third country.

4. E.g., students on foreign visas.

5. I almost wonder if there’s some intrinsic trade-off between accuracy and precision here. Resistance libs might accuse moderate incrementalists for missing the forest for the trees, while the incrementalists might turn around and accuse the resistance libs of getting a lot of things wrong at the margin.