How to make bureaucracies better

The word  bureaucracy  is often spoken with a tone of distaste, as if it means “official rules that just make life difficult for no reason.” It conjures images of reams of paperwork: a tangle of red tape that stands as an obstacle to progress. But the definition of  bureaucracy  is, simply, a body of nonelected government officials that undertakes administration — in other words, the people who actually get stuff done.

United States President Donald Trump contends that the US federal bureaucracy has strayed too far from its goals, and he has signed executive orders citing an aim of making the government “lean, responsive, and accountable.”

Knowable Magazine spoke with political scientist Katherine Bersch of Davidson College in North Carolina about the evolution of bureaucracies, why we need them, how they can become corrupt, and the most responsible ways to untangle red tape.

Bersch, who coauthored an article about bureaucracy with Stanford University political scientist Francis Fukuyama in the 2023 Annual Review of Political Science, is also a research affiliate with the Governance Project at Stanford’s Center on Democracy, Development and the Rule of Law. In addition, she is a cofounder of the Global Survey of Public Servants and is writing a book with the working title Who Governs? Presidents versus Environmental Agencies in the US and Brazil. This discussion has been edited for length and clarity.

What is the history of bureaucracies? Where did the concept originate, and how has it evolved?

It’s always been true that ideas and policies don’t execute themselves. You need a bureaucracy, a group of servants, to do that for you. That is true for nation states, and also for companies and universities.

The major leap in the evolution of bureaucracy, to a place that is not corrupt, happens when the people appointed to it aren’t friends or family or political cronies who can pay for their position, but people who are selected for their expertise, meritocratically. If you go back more than a thousand years, China is one of the first places where we see the development of a really important bureaucracy with entrance exams. It was very serious if you were caught cheating — the punishment could be death. We know from ample literature that corruption is lower where you have meritocratically recruited civil servants.

This has been a tricky transition for states the world over. It’s hard to achieve, and once it’s lost it’s hard to get back. Some have made this transition, and some haven’t. In the US, the 1883 Pendleton Act mandated a system of merit-based hiring and promotion. Up to that point, there was a lot of corruption. I study Latin America, and Brazil is a great example of where people are meritocratically appointed. But it also has a very thick layer of political appointees. In Africa, a strong, meritocratically appointed bureaucracy is very uncommon.

In the US, who makes up the bureaucracy?

There are roughly 2.2 million federal civil servants. It’s worth noting that the size of the federal government workforce has remained largely unchanged over 50 years, even while the population has grown by more than 60 percent. The number of civil servants, I would argue, is quite low. This is maybe controversial to say, but true. There are a lot of contractors on top of that — I used to be one — who are often paid a lot more.

Who these people are varies a lot from department to department, agency to agency. It’s very diverse, from Environmental Protection Agency (EPA) staff to forest rangers or postal workers. It’s often thought that most are in DC, but they are dispersed throughout the country.

People tend to call bureaucrats pencil pushers and claim it is just a machine for generating red tape. Is this perception fair?

I think that perception exists in the US, and perhaps in certain other countries, but not in all countries. When I ask students in the US, which I do, how many are thinking of entering the federal workforce or taking a civil service exam, I get very few hands going up. In Brazil, I get a lot. It’s a very different perception.

There can be a lot of red tape. Take procurement: how the government buys stuff, from tanks to computers. The number of public rules in the US is just astounding. They have to tick so many boxes. Many of these rules are in themselves laudable, setting guidelines for good practices, but in the aggregate it can be problematic.

But many of the rules that govern civil servants come from Congress; they didn’t make the rules themselves.

Many Americans are currently complaining that civil servants have unfair tenure, that they can’t be fired.

Bureaucrats can be fired, for cause. It’s hard but possible. But the idea that the president can’t just come in and fire you is the bedrock of a stable and impartial administration. There’s a tricky tension there. To what extent, and how, are public servants motivated if their jobs are so secure?

In the best agencies you have an esprit de corps, where their reputation among their peers really motivates them. Another place where this happens is universities; among the faculty, there are tenured positions. But you will find very hardworking people, even though we are very difficult to fire.

One interesting variable is how much autonomy bureaucrats have: In other words, whether they can make their own decisions or are under firm direction. How do Americans feel about this?

Americans are on the whole very suspicious of a centralized state. That has shaped our discomfort in giving authority to bureaucrats; because we are concerned about rogue or renegade bureaucrats, there are a lot of constraints.

On the other hand, we don’t like red tape, we like efficiency. So, then you want to give an agency a reasonable amount of discretion, so they can cut through red tape with common sense.

How has the autonomy of US bureaucrats changed over time?

One major change happened in the wake of the Great Depression, when the federal government took a stronger lead. An alphabet soup of new agencies and regulatory bodies emerged, issuing rules that carried the force of law. This sudden expansion prompted public concern: Wasn’t it Congress that was supposed to be making the laws? Members of Congress grew increasingly uneasy. In response, they enacted the Administrative Procedure Act (APA) in 1946 — a piece of legislation that may sound dry but is something all Americans should know more about.

The APA is often described as quasi-constitutional. It established procedures by which administrative agencies may create rules with the force of law. For example, agencies must provide public notice about proposed rules and give people time to comment, and then they must be explicit about how they incorporate those comments, which injects a certain amount of accountability.

This is one of the greatest constraints on agency rulemaking, and one of the greatest constraints on deregulation. It takes a long time to make a rule or take one away. Agencies must provide a justification that can withstand judicial review and isn’t deemed “arbitrary and capricious.” That’s hard.

President Trump is now arguing that agencies should repeal a host of federal regulations, and that they have a “good cause” to do so without providing notice for comment. Plenty of scholars have said this would gut the APA. Are there other moves afoot to change bureaucratic autonomy in the US?

From 1984 up until the summer of 2024, there was something known as Chevron deference. The idea is this: When a statute is passed by Congress, sometimes it can be vague. Like it might tell the EPA: “Go create clean air.” The EPA then has to figure out how to do that. So, Chevron deference means that the courts should defer to a reasonable judgment by the agency as the experts.

But then last summer, we had Loper Bright Enterprises v. Raimondo. Fishing companies had been required by the National Marine Fisheries Service to hire someone to monitor what they were bringing in, and it cost a lot. The Loper Bright fishing company was fighting that rule. The US Supreme Court decided that, actually, the courts have the authority to interpret the statute themselves and they don’t have to blindly defer to agencies.

That struck down Chevron deference. Some conservatives thought that this was good: Their argument was that Chevron gave leftist liberals in the bureaucracy too much free rein. But many liberals who were against the strike-down are now in favor of it, because it might allow courts to constrain President Trump’s changes to federal rules. We’re still getting a sense of the effects of this change.

Is it a good thing or a bad thing if the leader of a nation intervenes in bureaucratic autonomy?

A lot turns on who you believe is acting in the public interest. If you think the president is acting in the public interest, it’s good if they can intervene. If you think the civil service is capable, and acting in the public interest, it’s good if it has autonomy.

Our polarized world in the US right now looks at this through very different lenses. Some people view civil servants as smart people who were willing to take a lower salary because they’re very invested in the state of the country; others see them as lazy and self-interested.

In terms of the legal argument for a leader intervening with bureaucratic autonomy in the US, you’ll see a term thrown around: unitary executive theory. This is what Trump is using as the basis of his power to cut through constraints. It’s the idea that sole authority to the executive branch is given to the president as an elected official.

But is that the only source of legitimacy we have: the electoral legitimacy of the president within the executive branch? A lot of us would argue no. Congress also has electoral legitimacy, and they set up the agencies. There is procedural legitimacy through things like the APA, passed by Congress, which injects democratic accountability into the agencies.

How does the US bureaucracy compare with those in other countries?

For regulation and implementation, national bureaucracies have all sorts of constraints, both before (ex ante) and after (ex post) rules are made. The way that various countries set this up is very different. Compared to the US, for example, France has very weak ex ante but very strong ex post constraints. France, notably, does not have an equivalent to the Administrative Procedure Act. Instead, administrative policies may be subject to legal challenge only after they have been implemented.

Both Trump and Brazil’s former president, Jair Bolsonaro, tried to transform their environmental agencies. What held Trump back in his first term was the notice and comment process required before taking away regulations. Whereas in Brazil, Bolsonaro simply used his pen and signed away — during a moment of distraction, during Covid — a lot of policies that were intended to protect the Amazon. There are very different dynamics around the world about who has power within the state.

There are indicators for democracy that quantify and illustrate how democratic countries are. Are there similar indicators for bureaucracies and how “good” they are?

I always wanted to get my hands on indicators of that — of bureaucratic capacity (or capability) and autonomy. You know, sometimes you go to an agency and you think, “Wow, these people know what they’re doing, they’re fast and efficient.” Can we quantify that, and compare across and within states?

When I met Frank Fukuyama I realized we shared this vision to compare nation states and their abilities to get things done. One approach was to survey public servants themselves: How are they appointed to their positions? What motivates them? This project has developed over time. I joined Frank’s Governance Project in 2015. We teamed up with colleagues and put our minds together to come up with a list of core questions for the Global Survey of Public Servants. Now we have millions of responses from civil servants in more than 25 countries.

So far, we have tracked things like job satisfaction or whether people feel merit matters for promotions. In the UK, for example, just 28 percent of respondents agreed that their pay adequately reflects their performance. The next step will be creating indicators of capacity and autonomy. That’s still coming.

Have other states struggled or suffered when trying to fix their bureaucracies?

In Brazil, it was a very difficult moment under Bolsonaro. There were red lists of civil servants who were not in line with Bolsonaro’s agenda, and a lot of trumped-up legal charges. It created a real sense of fear. Although they have very strong tenure protections in their civil service, in some cases they were fired; sometimes they were locked out of the rooms where decisions were being made.

In another example from Brazil, although under a different leader, there was an infrastructure agency that was always considered corrupt, and so they destroyed the agency. But you needed an infrastructure agency. So a few years later, they recreated it again, but it had very weak capacity.

In Argentina, there have been times when politicians cut the public sector to the marrow. They privatized the national rail service, for example. A few years later, it was so broken and so in debt, they ended up pouring money into it. They had no civil servants left who understood what was even going on, and so they ended up trying to rehire many of the same civil servants.

I think people hope that you can break some things and then you fix them to make them better. But some of these systems are very old; you need to understand them before you break them.

Are there examples of bureaucracies being reformed in a better way?

Many successful cases of incremental reform exist. In Brazil, programs like public health and Bolsa Família (a conditional cash-transfer program for poor families) were developed gradually through experimentation and adaptation. Chile and Uruguay built strong civil services through institutionally grounded, step-by-step reforms. South Korea and China — despite their historical authoritarian contexts — have relied on incremental, experimental approaches to governance. These cases may not be headline-grabbing, but they demonstrate that patient, long-term reform can produce capable and trusted public institutions.

But whoever said, “I want to be an incremental reformer”? That’s the political challenge.

Is the US system in need of reform?

I would say yes. Many of us who study the public sector would say there are many common-sense reforms we could do to make things more efficient. Frank leads the Working Group to Protect and Reform the U.S. Civil Service.

One of the areas that I think is ripe for reform is the information technology (IT) system in the government. I remember living in Canada and thinking, “Wow, this is so much faster.” The US has always had more siloed databases because of a fear of government knowing everything about you. In Brazil, there’s one database for all civil servants, for the federal budget. It’s so much easier.

I think we would benefit from the ability to more quickly hire high-level experts in, say, IT. And reducing procedures in certain areas — absolutely. Many civil servants would cheer for not having to fill out so many forms.

Many of us started out hopeful for the DOGE (Department of Government Efficiency) effort — that it would work with Congress to make change. There’s a whole exciting agenda of things we can do to reform the public sector, but that takes thought, time and expertise, not a chainsaw.

Is what’s happening with the US bureaucracy right now unexpected?

I teach a course on corruption. This has been a very interesting year to teach it. Things that we just didn’t expect to see in the US are now occurring rapidly. The idea that the richest man in the world, Elon Musk, would be advising the president on what seems to be the entire makeup of the federal bureaucracy, including agencies that regulate his own business interests, is — I don’t know how to say this — surprising.

Stay in the Know
Sign up for the Knowable Magazine newsletter today

What do you think will happen next in the US?

It surprises me that Congress has been so silent about DOGE. Capacity of government takes a lot of time to develop, especially in the US where civil servants are not paid much. My big concern is this is going to weaken our capacity for the long haul.

What will happen to the use of the APA, to the period of notice and comment? I’m for reducing proceduralism in many areas, but not in policymaking. We want Congress to handle it, or when agencies do it, we want it done carefully. If the Trump administration finds loopholes in the APA, that would be a big issue. Because if you control personnel and procedures, including the ability to make rules with the force of law without procedural constraints, you have no constraints on your presidency. And that’s dangerous territory.