America needs more "bureaucrats," not fewer

Capable, competent government matters

Vivek Ramaswamy used to talk about his desire to lay off 75 percent of the federal workforce, which attracted earnest fact-checks questioning the ability of any President of the United States to accomplish this. In response, Ramaswamy moderated his position. He is now calling for only half of “federal bureaucrats” to be fired, something he believes can be achieved legally by randomizing the mass layoff.

I don’t think this won over any voters who weren’t so enthusiastic about his call to fire 75 percent of the federal workforce, but “fire everyone whose last or second to last SSN digit is odd” doesn’t make for a particularly snappy tweet. And wildly altering policy positions based on what makes for a snappy tweet helps explain why he was so intensely and briefly appealing to a bored media industry that needed a story to cover this past summer.

But the notion that we should have willy-nilly reductions in the number of federal employees is an important current in Republicanomics.

The overarching pseudo-idea on the right is that you can address policy problems purely through vibes. Rather than try to assess fiscal tradeoffs at a time of high interest rates, Republicans assert that budgetary issues can be solved by targeting things they don’t like. They avoid specifics at all costs because specifics call to mind tradeoffs, and even something relatively minor like “lay off half the park rangers at the national parks” sounds pretty bad. And while a lot of people rarely if ever visit national parks, a lot of people do visit them, and many of those people are Republicans.

Laying off half the federal workforce is, of course, a much more radical proposal than laying off half the park rangers. But retreating into abstraction gives it a veneer of plausibility.

The truth is that personnel is a very low share of federal spending. And while there are problems with the way the federal bureaucracy works, solving those problems is going to require not a weaker civil service, but a stronger one — one that has more (or at a minimum, better paid) employees so that the government is less reliant on contractors and litigation to get things done.

Where the workforce is

Compensation to federal employees cost $291 billion in 2019, which is a lot of money in absolute terms, but adds up to just 6.6 percent of that year’s total federal spending.

Some businesses, like daycare centers and fast food restaurants, are labor-intensive, while for others, wages and salaries are a relatively small share of the cost. A public school district is generally a highly labor-intensive public sector entity. There’s more to a school than the teachers and staff, but paying them takes up the bulk of the budget. The federal government is way on the other end of the spectrum: The government buys things like aircraft carriers, and it also cuts a lot of checks to individuals, most notably Social Security, the largest non-military program in the federal government.

One very expensive thing that the federal government does is pay for senior citizens’ healthcare. That could be achieved by the government employing a bunch of doctors and nurses and running hospitals, but we don’t do it that way. Instead, the government reimburses private providers. In 2019, Medicare spent about $800 billion, which is more than double the total cost of federal wages and salaries. It’s true that some fraction of that spending was on wages and salaries for the people who run Medicare. But it’s a tiny fraction — the bulk of the money goes to reimbursing people who are not federal employees.

By contrast, the VA does directly employ medical professionals and is by some metrics the largest federal agency. Alternatively, you could aggregate all the military departments into a single “civilian DOD” metric, in which case it would the biggest at 36 percent of the federal workforce and the VA would come in at number two with roughly half that share (note again that this is purely the civilian workforce at the Department of Defense).

So with the mass layoffs plan, we are really talking about two possibilities.

One is that the country would just have to accept a huge diminution of our government’s ability to execute military missions and take care of veterans. I’m sure Republicans would have no problem deliberately kneecapping the ability of regulatory agencies to enforce rules or the IRS to collect taxes. But I think Republicans probably do want the military and the VA to keep functioning. So in practice, the government would have to contract more stuff out.

The contracting model is bad

The problem is that contracting-out core government functions is really bad.

To be clear, it’s not universally bad. There are many situations in which the government really should just to go buy something. The government uses a lot of computers, for example, but it doesn’t make computers. Because generally speaking, the best time to contract something out is when a robust private market exists for the thing you need — the government buys lots of computers, but so do companies and private individuals all across America. In these cases, it’s easy to verify whether the price is fair, and buying things that private buyers also buy to some extent guarantees the quality of goods.

But over the past generation, there’s been a strong trend toward contracting-out basic government functions.

At worst you end up with, as Alon Levy observes in mass transit construction, “consultants supervising consultants” because even the ability to manage contractors has been lost. The more common problem, as John Gravois wrote way back in his 2011 Washington Monthly article “More Bureaucrats Please” is that the contracting paradigm leads to ripoffs. Indeed, government contractors arguably have an obligation to their shareholders to rip the public off.

Because civil service rules can be annoying, contracting offers a tempting short-term patch. If there’s something that you want done and you don’t have a civil servant on hand who is able to do it, it is a lot easier to hire a private company to hire the people to do it for you than it is to create a civil service role and make the appropriate hire in-house. It’s easier for a contractor to offer the high salary it might take to get the right person and it’s easier for a contractor to fire someone who needs firing. Flexibility is valuable, and contracting-out can secure it.

Over the longer-run, though, this dynamic leads to selective exit of a lot of the best civil service employees to higher-paid contracting roles, and eventually the loss of the ability to supervise the contracts effectively. And because there’s no private market for what you’re buying, there are no checks on quality or price. I think this all mostly starts from a well-intentioned place of sincere frustration with civil service rules that create real problems. But there is a serious time-consistency issue around state capacity. In the moment, it’s always easier to work around the lack of capacity rather than try to address it. The longer that goes on, though, the more state capacity decays and the greater the reliance on hacks and workarounds, until one day someone tosses out the idea of arbitrarily firing half the civil service with no regard to quality or job role. What we need to do is actually fix the problem.

Leviathan by proxy

John DiIulio, a kind of idiosyncratic conservative, wrote a book in 2014 called “Bring Back the Bureaucrats: Why More Federal Workers Will Lead to Better (and Smaller!) Government.”

He introduced an idea he calls “leviathan by proxy,” wherein the federal government, instead of doing things, tries to tell other people to do things.

A good example comes from the Build Back Better childcare proposal that wasn’t enacted. The basic goal was to get people access to subsidized, high-quality child care. The simplest-to-describe way to do that would be to have the federal government open and operate child care centers that provided a high standard of care and that operated at a loss while charging families tuition on a sliding scale. But while that may be simple to explain, actually doing it would not be simple at all. America is a large, diverse country with lots of local variation in conditions. You’d need to set up some kind of management structure, you’d need to hire lots of people, and you would face constant questions about whether you were providing services equitably across a varied terrain. Because that task is so challenging, the authors of the legislation preferred not to take it on.

Instead, they constructed a mechanism that’s incredibly difficult to summarize (you can read all about it here), but the basic upshot was to provide a large pool of money to states to construct systems that would contract out the provision of subsidized child care.

How were they going to ensure that the quality of care was good?

Well, they weren’t exactly. What they were going to do instead was mandate that states who received the money “create a tiered system to measure the quality of child care providers and help providers to reach higher quality standards,” with the proviso that the highest tier standards should be equivalent to Head Start Program Performance Standards. The plan was full of this kind of thing. Instead of answering difficult implementation questions, it required other people to try to figure them out.

That proposal didn’t pass in the end, thanks to a mix of Joe Manchin’s inflation concerns and Democratic Party leadership’s assessment that climate change is a more important issue. But it very much reflects how modern policymaking is done. The federal government doesn’t hire people to do things, it hires people to supervise grants that have rules that purport to require the people who get the money to do things. It doesn’t work very well as a model for delivering services, but it does make life easier for politicians to do position-taking in favor of various things (universal childcare, free college, whatever) without needing to worry too much about implementation details.

We need actual reform

All of which is to say that conservatives are not wrong to perceive some shortcomings in our current administrative state.

As Jennifer Pahlka wrote this week in The Washington Post, the basic act of hiring people to fill civil service vacancies is far too cumbersome, and due to run-amok proceduralism, “too often acting responsibly in government has come to mean not acting at all.”

Conservatives have their own set of concerns about the civil service that Trump wants to address with his Schedule F plan to make it easier to fire people.

That’s a bad idea on its own terms. But it could make sense as a component of a larger bargain that would make it easier to hire people and easier for civil servants to earn competitive salaries. Public sector employment at all levels tends toward a myopic focus on fiscal cost, leading to a tradeoff in which the pay is low, but it can be easy to get away with being lazy. Then in a pinch, you rely on contractors, and if you have an ambitious goal, you make sure it involves a minimum level of direct government action.

But there are positive counterexamples. About 15 years ago, at the initial behest of Republicans in congress — though with the eventual enthusiastic support of the Obama administration — the United States government invested more money in monitoring Medicare overspending. And according to research Maggie Shi published earlier this year, every dollar Medicare spent on monitoring generated $24-29 in savings with no evidence of deleterious impacts on patient health. The savings overwhelmingly came from deterring unnecessary low-value treatment rather than from actually clawing back claims. Merely the knowledge that someone was watching greatly improved the cost-effectiveness of the program.

Most of what the government does — military, law enforcement and border security, taking care of the sick and elderly — isn’t as controversial as conservative rhetoric would lead you to believe. If conservatives could admit to themselves that they don’t actually have a vision to drastically shrink the scope of federal spending, they might be willing to think harder about finding ideas like that Medicare monitoring initiative — ways to build a stronger, more competent state that’s also less wasteful. Alternatively, they might lay off half the civil service for no reason.