Democrats need a theory of power, but democracies need a theory of accountability

No, the Trump administration is not a model for how to get things done

We are at a crucial, once-in-a-century moment for American governance. Those on the right are, whether they admit it or not, aligning themselves with an increasingly authoritarian model of governance under Trump. Even before Trump’s second term began, many progressive thinkers were forming their own criticisms of public institutions as too tied down by procedural constraints to actually realize their goals.

I don’t know which vision will win, or what the future of American governance will look like, but the safest bet is that it will look significantly different from the past.

The combination of Trumpist authoritarianism and the many flavors of abundance sit uncomfortably together. A reasonable claim is that in the spaces where progressives still hold institutional power, they need to use it better, as well as plan for how to use federal power when they regain it.

An unreasonable claim is that Trump is teaching Democrats how to govern.

An example of this claim came from a recent essay by Marc Dunkelman in the New York Times, What the Left Could Learn From Trump’s Brutal Efficiency. Dunkelman has worked in Democratic administrations and wrote the widely praised Why Nothing Works: Who Killed Progress and How to Bring it Back.

The essay therefore does not come from a Trump defender, but reveals what I think is a temptation, even among his opponents, to treat Trumpism as a model for the future. This mistakes action for achievement, dismantling for building, and corruption for progress. To be fair to Dunkelman, I am going to quote him directly and extensively. I do so not to dunk on Dunkeman, but because I think it is an argument that deserves a response.

Lesson 1: Breaking the law and building things are different skills

Dunkleman starts:

Those of us opposed to MAGA complain plenty about what the Trump administration has done or is doing to the government — drawing on vast administrative resources for deportations, taking stakes in private companies, destabilizing the Federal Reserve. But we’ve missed a core element of President Trump’s appeal. In many cases, he’s effectively freeing the stuck wheels of bureaucracy.

Liberals may not like what he’s doing, but this is what government looks like when it functions — even if it is for brutal, misguided efficiency.

Trying to fire a Federal Reserve member on trumped-up charges for mortgage fraud that Trump himself has engaged in order to reduce central bank independence? Is this what government looks like when it functions?

Demanding public stakes in private companies as part of a broader pattern of coercion of corporations that also includes donations, public kow-towing, uncritical press coverage, or funding Trump crypto schemes in return for favorable treatment? Under what definition of governance could this be considered functional?

What about immigration? Trump has selectively turned Democratic cities into police states, largely arresting immigrants who are not the violent criminals he claims, with the occasional citizen thrown in. All to implement a policy that is broadly unpopular. featuring public officials who routinely lie to judges and the public, and redirecting law enforcement resources away from actual criminal activity to the point that a former ICE chief of staff said “It’s a good time to be an American-born criminal.”

Trump’s immigration enforcement may be brutal, but it is anything but efficient. It is massively costly not just in terms of money, but public faith in the government.

And these are Dunkelman’s strongest examples, the top of the batting order!

Dunkelman goes on to present what I take to be his central thesis: “The problem stems from today’s progressives lacking a clear theory of how public authority is supposed to work.”

Progressives want to achieve big goals, but also tie down the means to get there with excessive procedure. This is also the claim of Ezra Klein and Derek Thompson in Abundance, and my colleague Nick Bagley. Its a claim I largely accept.

But Dunkelman turns it into something closer to psychoanalysis of boomer progressivism: he says that Democrats lost their nerve when they saw the excesses of men like Robert McNamara, Richard Daley or Robert Moses: “Rather than sanction the Establishment to do more, progressives would work to check institutions prone to abusing their authority.”

This framing is simplistic. The Vietnam War was a disaster, but it is hard to argue that the Department of Defense, the only federal agency never to have passed a clean financial audit, subsequently became mired down in procedure because of it. Daley and Moses are not even federal figures, and the limitations placed on the federal government do not fall on their doorstep.

The framing is also unnecessary: the intellectual case about the need to reverse proceduralism has carried the day already, even before Trump returned to office. Pushing the argument in terms of Trump’s actions conflates dangerous lawlessness with useful administrative decluttering.

Lesson 2: Don’t confuse authoritarian actions with competence

Dunkelman gets to his solutions:

The only path forward is to lean back into centralized public authority. And that’s why the second Trump administration has the potential to slingshot progressivism into a new, brighter future.

Mr. Trump’s actions on deportation, tariffs and the use of military force in the Caribbean are almost certainly abuses of power. But they’re also a break from the public-sector incompetence that has helped erode faith in government to 22 percent, from almost 80 percent in the 1960s.

My argument isn’t that Democrats should ignore the law. Nor am I suggesting that we go back to the Moses era.

This is an extraordinary passage. Within the space of a few sentences it establishes that Trump is violating the law, while also presenting him as a model for progressives to follow, but concludes that Democrats should not break the law, or return to the past. He caveats the central claim of his argument, seeking to duck the obvious rebuttal, without a clear explanation of how this is practical.

And again, what Trump is doing is not “a break from public-sector incompetence.” It is a model of incompetence! Everyday brings a new story laying bare the lack of professionalism in immigration enforcement. Murdering alleged drug smugglers lawlessly and then trying to shift the blame on the military is not competence, but part of a broader pattern of undermining military professionalism.

The writer Zach Beauchamp coined the term “haphazardism” to describe the system of government emerging from Trump’s personalist style:

Haphazardism is authoritarianism without vision, a governing style defined by a series of individual attacks on democracy without any kind of overarching logic, strategic structure, or clear end state in mind. These attacks can do (and indeed have done) real damage to the American political system, but they are often poorly executed and even self-undermining - preventing Trump from ruling in the truly unconstrained manner he seems to desire.

I don’t know if “haphazardism” will catch on but it clearly rebuts the idea that Trump is ushering in some golden age of government competence. Loren DeJonge Schulman, who worked on government performance and national security for multiple Democratic administrations, posted a scathing response to Dunkleman on LinkedIn:

These actions--particularly the murder of civilians in the Caribbean--are the pinnacle of public sector INcompetence, and precisely the sorts of things that erode faith and trust in government. I sometimes see this sort of bizarre admiration of quick brutality that totally misreads both the outputs (which erode institutions where you very much need legitimacy!) and outcomes (which Trump is arguably not achieving, at least not yet).

The tariffs that Dunkelman points to are a tax hike on Americans that also increase inflation, implemented in the most haphazard manner possible, contributing to long-term allies viewing America as unreliable. The administration has reversed itself so many times that most people have no real sense of what Trump’s policies are. It seems likely the tariffs will be overturned in court, while the administration has no idea how to undo them. Incompetence!

Dunkelman argues that Trump contrasts favorably with Biden:

President Joe Biden wanted to save the climate and build green infrastructure. For all his good intentions and legislative savvy, his efforts were bogged down in the execution….Mr. Trump, by comparison, may be doing bad stuff, but he’s doing it fast. Sure, he’s often being dragged into court battles, and perhaps his detractors will win some of those cases. But the new administration still appears much better poised to leave an imprint than the previous administration.

Biden passed an impressive array of legislation with a narrow majority. Much of that legislation was undone by Trump, and the results were not in place by the time of that reversal. Certainly he would have benefited from moving faster. But Trump and Biden’s goals are different. Biden was trying to fix real problems such as climate change. Trump is largely reversing investments in state capacity to solve those problems. Sure, Trump will make a bigger imprint, in the same way that the guy who hits you over the head with a hammer will make a bigger imprint on your skull.

Again, these are the examples that Dunkelman has chosen to demonstrate the competence of the Trump administration, its ability to get things done. They are also domains (immigration, trade, national security) where Presidents enjoy substantially more discretion than in other areas, such as the climate and infrastructure policies that Dunkelman criticizes Biden for failing to deliver on.

In short, Dunkelman is picking examples where Trump has relatively few procedural constraints, and contrasting them with Democratic failures in domains where there are a lot (too many!) such constraints. As Trump has destroyed American soft power, and seems determined to use military force to pursue regime change in Venezuela, we might wish he was more constrained.

Dunkelman cannot bring himself to present DOGE, the President’s signature government reform, presumably because of its obvious failures to achieve its own goals, and the net negative effect it will have on government’s ability to solve broader public problems. Dunkelman does not mention Trump’s headline piece of legislation, HR1, which is designed to impose new administrative burdens on state governments and the public. These examples are not anomalous; they fit quite well with Trump’s broader record of weakening public services.

Lets take just one example of the conflict between the ethos of “move fast and break things” and the need for accountability, which is USAID. Trump’s Chief of Staff, Susie Wiles, revealed how haphazard the process was, and how disengaged Trump was from one of the most consequential actions of his administration:

I was initially aghast. Because I think anybody that pays attention to government and has ever paid attention to USAID believed, as I did, that they do very good work…When Elon said, ‘We’re doing this,’ he was already into it. And that’s probably because he knew it would be horrifying to others. But he decided that it was a better approach to shut it down, fire everybody, shut them out, and then go rebuild. Not the way I would do it…The president doesn’t know and never will. He doesn’t know the details of these smallish agencies…Elon’s attitude is you have to get it done fast. If you’re an incrementalist, you just won’t get your rocket to the moon. And so with that attitude, you’re going to break some china. But no rational person could think the USAID process was a good one. Nobody.

Brutal? Yes. Efficiency? No. The absence of rational people to check Musk will cost millions of lives.

Lesson 3: Trust in government is not a great barometer of effective government

Dunkelman says that trust in government has declined partly because of public sector incompetence. And hey, trust in government has actually increased since Trump returned to office, so the public is recognizing Trump’s competence, right?

Wrong. The increase in trust in government is driven by partisans. Basically, Republicans increase in trust when Trump retook office more than offsets Democrats loss in trust. It’s not really about performance or competence, but about identifying with the person in the White House.

This is one reason why broad measures of trust in government are not very useful as a measure of how well a government is doing. Such measures are more driven by partisan identity rather than agency effectiveness.

Trust has also been declining in other public institutions not responsible for delivering public services, such as Congress and the Courts, and for private institutions, such as the media. There are different explanations as to why, such as polarization, increased transparency revealing institutional flaws that were previously missed, or the rise of post materialist values. But the key takeaway is that it is very difficult to make a serious claim that declines in trust are driven by changes in government performance.

This doesn’t mean we shouldn’t take the relationship between government performance and trust seriously, but that taking it seriously means understanding it in a way that is different from the type of claim Dunkelman makes. Look at trust in agencies, or in specific services if you want to gain insights into how people’s experiences with public services shapes their beliefs about those services.

For example, Veteran’s Affairs has made trust in services their north star for performance for more than a decade. They also measure every aspect of their services — which is easier when they are an actual service delivery organization, unlike most of the federal government — and have the largest customer experience team in the federal government. That is what it means to take trust in government seriously.

Lesson 4: The Supreme Court is a political body that doesn’t treat Democratic and Republican Presidents the same way

Dunkelman believes that progressives can take advantage of the Supreme Court becoming newly invested in an empowered executive:

At the same time, new conservative jurisprudence appears poised to enhance the power of the executive branch. A series of cases at the court suggest that the court’s conservative majority is prepared to give the executive branch more deference. Progressives have railed against this jurisprudence, but if the long-term effect of these rulings is to give the administrative state more discretion to act with greater alacrity, then progressives, once elected, should be able to use it to much the same effect.

This assumes the court is relatively even handed and apolitical in how it treats presidential power. It is very hard to square this with the reality of how the court has treated Biden and Trump when they pushed the boundaries of executive power. Biden was greeted with checks, justified by major questions doctrine, while Trump has been mostly shown a green light for more radical actions, often with barely written opinions provided via the emergency docket that upends lower court decisions, and decades of laws and norms.

Maybe SCOTUS will find it hard to reclaim the executive power it is now handing to Trump. Even so, the SCOTUS approach might empower the President, but it weakens the institutions that work for him, which bodes poorly for implementation.

The repeal of Chevron, combined with the erasure of public employee protections from politicization will make it harder for the administrative state to function. Which is the goal of a conservative court, and suits the purposes of a leader interested in capturing power rather than making the executive branch more effective. Allowing Trump to dismantle agencies offers no insights about how to build them. Allowing him to fire competent people is going to make it harder to hire them in the future. Strongman, weak state.

This may reflect an asymmetry in how Republicans and Democrats think about power that needs to be recognized. Democrats largely want a government that delivers (even if they sometimes hamper it); Republicans largely want a constrained government. A unitary executive President with a weak administrative state is better suited to the latter task than the former.

Lesson 5: Democrats might need a theory of power, but democracies need a theory of accountability

It can be both true that proceduralism is a real problem, and that abandoning the law and constitution is not a solution.

Government is an enterprise of the delegation, directing and constraining of power for some collective good. I‘m sympathetic to the idea that progressives need to rethink procedural barriers to get things done. But we are witnessing a shift to authoritarianism right now.

“Unitary executive theory - but for good” is not a theory consistent with a democracy. It is certainly a theory of power: it was created specifically to allow Republican Presidents to unshackle themselves from accountability, and Trump is exploiting every opportunity to do so. His dismantling of the civil service system follows a pattern of other leaders engaged in democratic backsliding.

Authoritarianism is not just bad for democracy. It is also bad at delivering what the public wants. Mussolini never actually made the trains run on time. The type of centralized authority Trump is pursuing eventually fails for the banal reasons why personalist authoritarian regimes always fail: they do not tolerate dissent, they become bubbles insulated from reality, they rely on loyalists rather than competent and honest officials, they abhor real accountability, and they feed on corruption.

Once you take away the false claim that authoritarians offer “brutal efficiency” you are just left with the brutal part.

Again, from Loren DeJonge Schulman:

“In many cases, he’s effectively freeing the stuck wheels of bureaucracy.” In the space of national security and tariffs, he’s using emergency powers and engaging in extrajudicial killing in spaces where he can direct, not enabling bureaucracy to act toward the ends of good government. That’s not stuck wheels of bureaucracy - that’s autocracy.

One online wag posted the following visual metaphor for how Trump is “freeing the stuck wheels of bureaucracy.”

Lesson 6: People value accountability, including procedural checks

Again, I’m broadly sympathetic to Dunkelman’s broader critique of proceduralism, just not his solutions, which feel vague. To be fair, such solutions will depend on the specifics of a situation, which Dunkelman does not really have the room to get into in an op-ed. But using Trump as a starting point leads only to a dead-end.

Even when Dunkelman articulates a solution — which seems to be that people get their say, but a leader decides — he uses an unfortunate example:

If the F.B.I.’s top counterterrorism official and the special agent in charge of a regional office differ on how that office’s agents should be deployed, the bureau’s director, having heard both sides, will ultimately decide.

The problem with this example is that the actual FBI director is engaged in an unprecedented deprofessionalization of the agency, and only appears to be listening to the White House and online commentators, rather than the experts who work for him. He lied to Congress when he said he would not politicize the agency. Dunkelman’s hypothetical scenarios simply does not account for the reality of what is happening right now, right in front of us.

You could make the same point with one Trump agency after another. For example, the Department of Health and Human Services has a decisive leader; the problem is he is ignoring the scientists who work for him, or pushing them out the door. Cabinet officials are systematically refusing to appear publicly before Congress to account for extraordinary uses of power, like the use of immigration forces, changes to vaccines, or the use of military forces to destroy vessels off the coast of Venezuela.

To give a sense of how much Congress has been sidelined, the Senate Foreign Relations Committee had two hearings with administration witnesses in Trump’s first year back in office. In Biden’s first year, the same committee had 24 hearings with administration witnesses. The House Foreign Affairs Committee had 12 such hearings in in 2025, compared to 47 in Biden’s first year. While there was a special DOGE committee created by the House, no DOGE official testified about an extraordinary destruction of government that occurred with almost no oversight.

Maybe Dunkelman has is an implicit theory of accountably, which is accountability for results. This idea is appealing, but it has a couple of problems. One is that governments have already constructed outcome-based models of government. I spent much of my early career studying performance management processes, such as the Government Performance and Results Act. My conclusion is that performance data is mostly not used by the public or policymakers, and at best it is useful as an internal managerial tool. It is a less potent model of accountability than reformers hope.

Another issue is that people genuinely value procedural constraints. They object to corruption. They cannot personally watch over the shoulders of government officials, but want to know that someone is doing this job. These objections are not merely the outdated holdovers of the spoils system. They also reflect entirely reasonable concerns today about how wealth and political power is concentrated in the hands of the few.

For example, the richest man in the world was handed over the reins of government because he gave vast sums of money to the President and asked for the job. He made a lot of truly awful decisions, and then returned to the private sector to demand a trillion dollar pay package. In such a system, its not irrational for people to want more procedural constraints.

A new paper by Dan Walters and Brian Feinstein uses survey experiments to present evidence on how people tradeoffs between participation and speediness of decisions, and shows that people place a high premium on participation opportunities, even when they result in considerable delays.

For an abundance-minded reformer’s perspective, this might seem to be exactly the problem: the public can be easily persuaded to buy into attractive notions of participation that slow outcomes. Their cheap talk is they want accountability, but their revealed preference is that they don’t actually engage in the type of participation opportunities they say they value, which are dominated by interest groups, leading to agency capture.

Lesson 7: You can borrow tools from authoritarians, but not their model of governing

One of the core debates in the field of public administration theory is about Woodrow Wilson’s claim in his 1887 essay The Study of Administration.

If I see a murderous fellow sharpening a knife cleverly, I can borrow his way of sharpening the knife without borrowing his probable intention to commit murder with it.

Wilson’s claim was about lesson-learning: that you can borrow the practices of non-democratic states (or the private sector) to improve democratic governments.

I think it broadly holds for narrow managerial tools, or specific tasks. For example, Russia and Ukraine are learning warfare tactics from each other as a dictatorship tries to overturn a democracy. But a core reason why Ukrainians are fighting is that they do not want to borrow the experience of living under a dictatorship. They value their system of government enough to die defending it.

One categorical mistake that Dunkelman makes is to conflate taking on proceduralism with building a personalist authoritarian government. The Trump administration has certainly been more effective in its second term. It has planned better, and is using the levers of power to get what it wants. But mostly what it wants, what it has been sharpening the knife for, is to establish a personalist system and gut the administrative state.

It is not merely that it has different policy goals from progressives. It wants a different model of government. All of its understandings of governmental power start with a transactional and loyalist perspective. This is nothing to emulate.

When Russ Vought says “we’re going at it with everything we’ve got and trying to bulldoze the bureaucracy where it exists” he does not mean he is freeing the government to be more effective; he wants it to make it disappear to the greatest degree possible, while ensuring that what remains is a docile tool for Trump’s democratic backsliding. That is the lesson Trump is offering, one that only that an aspiring strongman should want to learn.

---

My hope is that the conversations among those opposed to Trump continue to focus on how to make government work better, but that they also explicitly engage in discussions of accountability as something other than a constraint to be overcome. Those discussions should not assume previous systems of accountability have to be maintained. Lets look at the evidence of their effectiveness, and lets look at alternatives. But lets also recognize that people value accountability as central to democratic legitimacy.