Democrats don't need more "infrastructure"

They need more freedom and independent-mindedness, too many institutions can hurt

Democrats, from moderates to progressives, have been reconsidering a lot of their electoral assumptions since Kamala Harris’s defeat, and the general tenor of conversations that I’ve been privy to has been pretty good.

Not everyone’s ideas are amazing or perfectly aligned with my sense of the path forward. And in particular, there’s a natural tendency to focus too much on superficial and comfortable changes while avoiding pain points and conflicts. But relative to the discussions of 2017-2022, there is a refreshing lack of wishful thinking about the power of voter registration or magical mobilization tactics. Folks who disagree with each other about a lot of things broadly agree that Democrats need to say and do some things that will make some people who voted for Donald Trump — in many cases two or three times — want to vote for Democrats instead.

The one exception to this, I have to say, is that there is a lot of enthusiasm for building new “infrastructure.”

And I get it. Many of us are incredibly alarmed by what Trump and Musk are doing, and people are eager to spend money fighting back. There are also a lot of people out there hustling to raise money for their own projects. Some of those projects are good, some are bad, and some are mid. But whatever your project, if you’re looking for funding, it’s almost always in your interest to argue that more infrastructure is needed. So I kind of feel like it falls to me to be the skunk at the party who calls this into question. Which is not to say that nothing should be funded! There are good ideas out there, and there are things that the world could use.

But there is also an enormous amount of cruft in progressive politics. And the whole space would really benefit from a zero-based budgeting analysis, because a lot of the existing infrastructure is directly counterproductive.

The keys to electoral politics — finding out what your voters think, and adopting some heterodox views that are popular with the electorate — are just not that complicated. Candidates don’t certainly don’t necessarily need a lot of infrastructure to execute on them. But what anyone working to elect Democrats does need is the freedom to do it without taking massive amounts of friendly fire.

Infrastructure for what?

A much harder question than “What will help candidates win in red-leaning districts?” is “Why has it come to be difficult to execute on this basic playbook?”

And the answer in many cases is not the absence of infrastructure but its excess.

Sometimes this is blatant. Indivisible had incredibly honorable roots in the earliest days of Trump’s first term. There was a tremendous grassroots outpouring of concern. People wanted to know what they could do, and the original Indivisible Guide was a great set of practical suggestions for people who wanted to up their level of engagement. But as Indivisible grew as an institution, it just sort of converged with the highly ideological DC-based NGO borg (see Theda Skocpol and Carolina Tervo), detached from the grassroots anti-Trumpism that inspired record Women’s March attendance and pragmatic 2018 candidate recruitment.

As Liam Kerr points out, this eventually landed with Indivisible explicitly organizing not against election denialists and people trying to cut Medicaid, but against the existence of the moderate Dem caucuses that serve as the bulwarks against MAGA trifectas forever.

This is not to say that Indivisible is a uniquely pernicious actor in the progressive infrastructure. In a lot of ways, they are unusually smart operators. That just means that, in this case, they were unusually explicit about the fact that tent-shrinking and orthodoxy-enforcement are key organizational goals. The Revolving Door Project, similarly, has as its core mission trying to get people kicked out of the Democratic Party coalition, starting initially with executive branch appointees but expanding to encompass outside commentators. The big enemies of their “Hackwatch” project aren’t Fox News personalities or right-wing influencers, but Jason Furman and Catherine Rampell.

Factional infighting is a valid form of activity. If your life is dedicated to helping left-wing Democrats beat moderate Democrats, then we just disagree. But I think we can still respect each other and surely find plenty of causes where we’re actually aligned.

But from a funder and party-building perspective, it’s important to understand that this is the function of a lot of progressive infrastructure. It’s often framed less explicitly than the two examples above. But most progressive advocacy organizations — whether on climate, immigration, or criminal justice — dedicate the bulk of their advocacy efforts to pressuring Democrats to adhere to ideological orthodoxy. They do, at the end of the day, encourage people to vote for Democrats over Republicans. But they don’t do anything to persuade voters to switch sides, and they don’t do anything to build bipartisan legislative coalitions. They just discourage Democrats from doing the kinds of things that could persuade voters and craft bipartisan bills.

If your interest in politics is motivated by something like, “Donald Trump seems like a really bad, corrupt, authoritarian person” or “The government should give a damn about poor people” or “Women should have reproductive rights,” then supporting these kind of organization is worse than doing nothing. Note that tent narrowing is generally bad for your cause even if the narrowing is specifically on your issue. If, in 2016, there had been pro-life Democratic senators representing southern states, where (unlike outside the south) abortion rights are less popular than the Democratic Party, they would almost certainly have voted to confirm Merrick Garland to the Supreme Court, and women today would be better off. Trying hard to win elections is really important, and funding groups whose raison d’être is making that harder are deeply harmful.

The scourge of misinformation

A couple of weeks ago, a very smart, very pragmatic moderate Democrat told me that 90 percent of the public supports universal background checks and only the fear of primary challengers can possibly explain why Republicans vote against it.

For this guy’s personal politics, it’s fine to believe this. Gun control is a good issue in his district, and he’s heterodox on issues where it makes sense for his voters. And in terms of his narrative about himself, it’s part of a good moderate-sounding discourse about how both sides are hostage to extremists, whereas he’s smart and sensible.

But there’s something screwy about this vision of background checks being universally popular. You’ve probably noticed that background checks is never wielded as a decisive wedge issue in a campaign against a frontline Republican incumbent. It doesn’t test well in ad effectiveness experiments. When Maine, a state that Hillary won, had a background checks ballot initiative in 2016, it failed by a few points. That same year, a similar initiative passed in Nevada, but again ran a few points behind Clinton.

I’m not going to tell you that universal background checks are unpopular. It seems like they run a bit behind the Democratic Party in rural areas and a bit ahead of it in suburban ones. But it’s also not particularly hard to understand why Republicans are comfortable opposing this idea — it’s low-salience, their base doesn’t like it, and it’s not overwhelmingly popular outside of that.

So why did my guy think it’s a 90-10 issue?

Well, there are a million ways to game an issue poll. And one thing advocacy organizations have learned to do is to invest heavily in polling that leverages acquiescence bias¹ and careful question wording to exaggerate the support for their cause. The people who do this aren’t necessarily saboteurs, tent-narrowers, or bullies. These are often cheerful, well-meaning issue advocates who genuinely are pushing popular causes. But they’re often taking a cause that, in a well-designed survey is a 55-45 issue, and trying to tell you it’s an 80-20 issue.

Because almost everyone working on almost every cause is doing this, there’s an incredible amount of pressure on people who know better to also do it. If you have a genuinely 60-40 issue but everyone else is lying and pretending their 55-45 issue is 80-20, then you have to lie, too, or you’ll be left in the dust. The result is an informational tragedy of the commons, where many actors in the political system are misinformed about the popularity of their own positions. Most seats aren’t competitive and American elections are very polarized, so in practice, it doesn’t really matter if the average member is publicly taking several toxically unpopular positions for no particular reason. But by the same token, precisely because most seats aren’t competitive, the average member has no incentive to fight through this sea of public opinion misinformation that he and his staff are swimming in.

So lots of members who, all else being equal, would prefer to say things that are helpful to their frontline colleagues so they can win a majority are accidentally saying things that undermine the party brand. Not because of a lack of infrastructure but because there is too much bad infrastructure — too many advocacy organizations that are polluting the information environment because the whole incentive structure of the network of progressive infrastructure is disastrous.

Good things are good

I’m not trying to be too negative or contrarian about the value of political infrastructure.

I’ve been shouting into the void for years that rigorous, accurate policy analysis is an underrated investment opportunity. I think you can find non-partisan homes for this kind of analysis and also party-aligned homes for it, and both are valuable and useful in real ways.

But (and this is the hard part) the analysis has to actually be good. Not just anyone can do that. You need to be careful and selective about building your team, and you need to be able to stand by them, even if they piss people off. But high-quality analytic work is incredibly useful, both in the form of “here’s an idea we came up with and it’s good” and also in the form of working closely with elected officials to refine their own ideas into a viable form.

For all the pixels spilled on electoral strategy, the positioning part (project moderation, open-mindedness, big-tent vibes, and say popular stuff) is easy compared to governing. Politicians genuinely need help with that.

Closely related things, like magazines or websites where smart people say true things about the world and conferences where people can network and discuss, are also useful. Putting smart political ideas into media outlets with broad reach among people who aren’t political obsessives would be helpful. It’s good for like-minded elected officials in different spheres of life but with aligned values to find places to connect. Providing talented people media training so they can be a little more effective on television, on podcasts, and on vertical video can be useful. There are bits of good infrastructure out there that could use more money and bits of good infrastructure that it’s worth building.

The point I want to make as someone who’s not personally pitching right now is that this is not a situation where everything helps at least a little.

The infrastructure itself can be harmful by formulating policy proposals that are bad, spreading misinformation about public opinion, threatening candidates or elected officials who try to break from the heard with negative attention, and deliberate narrowing of the tent.

Thinking eight years forward

One reason this is on my mind is that while many people are extremely agitated about Trump right now and desperate to do something, the odds are very high that Democrats will take back the House in 2026. There are no guarantees. But history, the map, the current generic congressional ballot polling, and the shifting nature of the parties’ voting bases all point in that direction. Of course, Democrats could still screw it up. Or they could over-perform with smart tactics.

Whatever Democrats spend the next 18 months doing, though, they stand a good chance of looking really smart over the winter of 2026-27. What’s important is that they actually be smart.

Looking back at recent history, I think this is what happened during the “resistance” to Trump’s first term. You had a big, grassroots, anti-Trump movement. Democrats did well in the 2018 midterms by riding normal political backlash, plus sensible recruiting of mostly pragmatic frontline candidates. But a lot of infrastructure investment during this period wasn’t aimed at assembling a big tent anti-Trump coalition. It was aimed at funneling anti-Trump sentiment into a push to replace Clinton/Obama liberalism with a new ideology. This project was only partially successful at capturing the party, but its ideas were influential enough to generate significant backlash.

What I want everyone who wants to beat Trump to think about now isn’t just what kind of infrastructure will look good after the midterms or might be “good enough” in a 2028 world where things have gone badly wrong and there’s anti-Trump backlash.

We should be thinking about what infrastructure sets Democrats up to compete for state office everywhere so we can expand Medicaid? What would make it so that high-income blue states are also high-growth? So that the “party brand” is healthy enough that people aren’t casting about for Osborn-style gimmicks to get to a Senate majority? There are investments that can advance those goals, but I think the beginning of wisdom is to recognize that the average piece of infrastructure is probably counterproductive to them.

It’s also probably worth recalling that Donald Trump in 2016, for better or worse, completely remade the image of the Republican Party just by being good on television and saying stuff. There’s a good amount of MAGA-aligned infrastructure now, but that was all built after Trump refocused the party on a different set of issues and disavowed a lot of Bush-era policy commitments. It’s hard to pull that off, in the sense that the odds of failure are relatively high. But it’s not actually a massive resource-intensive undertaking — it takes people with talent and good ideas and a bit of luck.

1. If you run a poll that asks “Do you like bananas?” and then you run a separate poll that asks “Do you dislike bananas?” and add up the people who answered “yes” in each poll, the total will come to more than 100 percent. People who answer surveys are biased toward saying “yes” unless you specifically structure the question (“Would you say that you like bananas or that you dislike them?”) to avoid the bias.