Civil-service Reform and Republican Government
The federal government employs about 2.4 million workers, the vast majority of whom work in executive-branch agencies. About 4,000 of those employees are political appointees who serve at the president's will. The rest constitute what is generally known as the "civil service." Members of the civil service are protected from their jobs being terminated at will, by their office heads or by the president, through a variety of statutory protections. In what follows, I'll refer to those statutory protections as "tenure," and proposals to limit or abolish that tenure as "civil-service reform."
Civil-service reform is now a serious possibility, sparked by conflicts that have arisen during President Donald Trump's second term. Many civil servants do not like President Trump or the policies he has been pursuing. Early in his second term, Trump appointed as director of the National Institutes of Health (NIH) Jay Bhattacharya; a physician, former Stanford professor, and outspoken critic of the federal government's response to the Covid-19 pandemic. NIH staff walked out of an agency townhall when Bhattacharya didn't rule out the possibilities that Covid-19 was a man-made virus and that the NIH may have funded its development. At the Civil Rights Division (CRD) of the U.S. Department of Justice (DOJ), Trump appointed as assistant attorney general Harmeet Dhillon, an experienced civil-rights litigator and outspoken advocate for race-blind policies. Since Dhillon was confirmed, at least 60% of CRD's career attorneys have resigned.
Trump and his allies have responded to such resistance by attempting to limit tenure in the civil service. In drafting the "One Big Beautiful Bill Act" signed into law in July 2025, Congress considered encouraging newly hired federal civil servants to choose at-will employment over tenure. House Republicans included provisions to that effect in draft versions of the bill, but the Senate stripped those out before it was passed and signed.
But the case for civil-service reform emerged long before Trump's presidency. The federal government greatly expanded the number of agencies and their administrative complexity during the New Deal, and civil servants began securing tenure protections in the post-war era. Since their jobs became more secure, federal workers have been more willing to defy presidents and their political appointees; this has become particularly pronounced in the last 20 years. Also in the last two decades, some federal agencies have disregarded the rights of the American citizens they regulate. Civil-service tenure is far from the only factor contributing to these problems, but it is long past time to consider limits on tenure for newly hired civil servants.
But limiting tenure, or making other permanent and far-reaching reforms to employment in the civil service, will not be easy. Reform opponents have had a lot of success by personalizing debates about civil-service changes; they point to specific civil servants who are competent, and ask why anyone would want those workers fired and thrown out on the street. President Trump himself also undermines the case for reform. An independent, tenure-protected civil service appears attractive when elected and appointed executive-branch officials test the limits of the law.
Civil-service reform is also difficult to pass because it goes against the policy goals that academics and other policy professionals value. In law, public administration, and political science, most scholars agree: To say whether any change to federal government-employment laws is good, the reform must be tested against criteria such as efficiency and rational administration. Those perspectives make tenure seem the presumptive arrangement, and at-will employment appear aberrational.
Current law and conventional wisdom, however, have things backward. No theory of government can say specifically which civil-service positions should be protected by tenure, and which should be at-will positions. But such a theory can evaluate the case for tenure generally. It can also produce a checklist, a series of considerations that can be applied to specific positions in the civil service. According to the theory I'll introduce here, at-will employment is presumptively preferable, and tenure should be allowed only when a particular executive-branch job clearly requires security or continuity.
This normative theory is based in the character of the American regime as an experiment in republican self-government. In a republic, the government is supposed to respect the rights of its citizens, and federal policies are supposed to change as voters cast their ballots for lawmakers, a president, and new initiatives.
As I will discuss below, current tenure protections threaten both of these tenets of republican government. Tenure in the civil service ultimately facilitates bureaucratic and administrative governance. At-will employment facilitates republican government and our Madisonian political order. It is time for Congress to act and ensure that government-employment arrangements reinforce, rather than undermine, republican principles.
TENURE UNDER FEDERAL LAW
Tenure can be guaranteed to a federal employee through one of three legal sources: statutes, executive directives, or contracts (mainly, collective-bargaining agreements). To make a case for reform, my primary focus will be statutory protections. Tenure is provided to most federal civil servants through Title 5 of the U.S. Code. Title 5 exempts probationary employees, workers with less than a year's experience in the federal government. It also exempts employees in some lines of work — most notably, in law enforcement, intelligence gathering, and national security. But federal employees who aren't exempted are entitled to tenure.
Title 5 tenure gives civil servants two basic protections. First, civil servants may not be removed or disciplined without considerable process. Civil servants facing removal or discipline are entitled to written notice that they are facing adverse actions, representation by lawyers, and written justifications for such actions against them. When government departments start disciplinary or removal proceedings, civil servants are also entitled to contest those moves — orally, in writing, and in appeals before the Merit Systems Protection Board (MSPB).
Second, civil servants are protected by substantive guarantees, which limit the reasons their supervisors may invoke to discipline or remove them. Most civil servants are protected from removal except for cause. For-cause guarantees shield employees from being fired unless they engage in "unacceptable performance," or if they give their supervisors "such cause" for discipline or termination "as will promote the efficiency of the [civil] service."
These two protections apply to most civil servants, but not to the highest-ranking employees in the civil service. Those are the 7,000 members of the "Senior Executive Service" (SES), who work between the president's political appointees and the rest of the civil service. SES members in some cases receive shorter notice of impending discipline or removal than ordinary civil servants. SES members may not challenge adverse actions before the MSPB. And the substantive standards for removing or disciplining SES members are easier to satisfy than the guarantees for ordinary civil servants. But even though the SES's protections are not as strong as those enjoyed by most civil servants, their safeguards resemble the rest of the civil service in broad outline. And of course, those protections are far stronger than whatever guarantees at-will employees have.
Many observers assume that civil-service tenure originated more than a century ago, when post-Civil War reformers abolished the spoils system. (Robert Beschel, Jr., suggested as much in these pages.) But this is a misconception. In 1883, Congress started dismantling the spoils system in the Pendleton Act. The act changed the spoils system on the front end by instituting competitive hiring and promotion-review policies. It applied to only about 10% of employees in the federal executive when first passed; Congress gradually widened the coverage of its competitive-hiring model in later laws. But the Pendleton Act did not alter federal manager-employee relations on the back end (save for banning the requirement of political service or donations). In other words, under the act, when federal employees were selected competitively, they could still be fired at will.
Tenure did not become a regular feature in the federal civil service until more than half a century after the Pendleton Act. Congress first instituted a version of tenure in 1944 for military veterans. The Veterans' Preference Act gave veterans priority in obtaining jobs in the civil service, and it also entitled them to tenure once they had such jobs. Once military veterans had tenure, non-veteran civil servants wanted it, too. Post-World War II presidents (especially John Kennedy) gave many other civil servants tenure in collective-bargaining agreements with public-sector unions. Congress then created a general statutory right to tenure — establishing the basic protections summarized above — when it passed the Civil Service Reform Act of 1978, nearly a century after the Pendleton Act.
QUESTIONING TENURE
Title 5 tenure protections are certainly valuable job benefits. But should they be standard features of employment in the federal civil service?
Most private-sector jobs do not come with tenure, though some do (e.g., many university professors and some workers on union contracts). Some employees enjoy job security through guaranteed long-term contracts, such as elite athletes and corporate officers. Generally, though, when an employment contract does not specify how or when it ends, it can be terminated at the will of the employee or the employer. In virtually all employment, whether private or public, anti-discrimination statutes make it illegal for employers to fire or discipline employees on the bases of race, gender, disability, or other similar classifications. But anti-discrimination laws are not as comprehensive as tenure protections. Employers only need to show that they are firing or disciplining for genuine job-related reasons and not discriminating on the basis of a certain characteristic.
Tenure isn't an automatic entitlement of government service, either. Many state governments do not offer tenure as a standard feature of employment. In about 20 states, at-will employment is the accepted model for state-government work. A few states (most notably, Texas) never made tenure a meaningful feature of government employment. Although many other states instituted civil-service tenure in the mid-20th century, some have since eliminated it. Georgia, in particular, removed state civil-service tenure in 1996, and since then more than a dozen states have expanded at-will employment.
Above all, civil-service tenure seems incongruous when compared to the job security available to other actors in the federal government. Civil servants are mid-level employees in the federal executive branch. Tenure gives them more job security than most of the highest-level officials under the U.S. Constitution — though of course not as much security as the life tenure federal judges are granted under Article III of the Constitution. Civil servants work in the executive branch, and to oversee executive agencies and departments, presidents appoint "principal officers" under Article II of the Constitution. Those officers rely on presidentially appointed staff designated political appointees under chapter 35 of Title 5. All political appointees are removable at will under chapter 35.
As for the principal officers, U.S. Supreme Court precedents require that the president be able to remove most of them at will, except for the commissioners on multi-member commissions at so-called "independent" agencies. The general rule comes from the Court's 1926 decision in Myers v. United States and Roberts Court cases such as Seila Law LLC v. CFPB in 2020; the exception comes from the 1935 ruling in Humphrey's Executor v. United States. But Humphrey's Executor could be overruled soon, in Trump v. Slaughter or one of several other cases the Trump administration has brought to remove agency heads statutorily guaranteed their positions for fixed terms and with for-cause removal protections.
TENURE AND THE CONSTITUTION
When considering all of these employment arrangements in the federal executive branch, an obvious management problem emerges. Presidents have four years to carry out their jobs before they could be replaced in the next election. They rely on about 4,000 principal officers and political appointees to carry out what they promised voters, and they can expect obedience from those employees because most of them work at will. But presidents and their key aides must also strive to implement their goals by working through a civil service of more than 2 million. And most of those civil servants have job security closer to that of Article III judges than to any of their supervisors in Article II. One might ask whether this a good way to run an executive branch.
Opponents of civil-service reform often say that such proposals insult and demean the many civil servants who have tenure. That argument equates tenure with an entitlement such as Medicare or Social Security. But tenure shouldn't be an entitlement. In any line of work, tenure should be offered with a job only when it serves the interests of the employer. And when the employer is a government, the decision to offer tenure is an intensely political question. No matter how competent and decent most government employees are, they work for the people. They should not be eligible for tenure unless it serves the people's interests.
With this understanding in mind, we should ask: In what circumstances is it desirable for a government to make tenure a standard feature of employment in its executive branch? This question deserves a Madisonian answer in accord with the republican government established by the framers.
The Constitution, it should be noted, doesn't say anything explicit about civil servants. And because civil servants are mere employees, not "officers" in Article II's sense, neither the Constitution nor cases such as Myers or Seila Law stop Congress from legislating on the terms or structures of civil-service jobs.
Still, the Constitution does give lawmakers guidance about civil-service employment. It authorizes Congress to legislate on the civil service through the Necessary and Proper Clause at the end of Article I, Section 8. Under that clause, legislators may create jobs in the federal executive whenever doing so seems necessary and proper for carrying into execution the Constitution's enumerated powers. It follows that Congress may also designate those jobs as at-will or tenured positions, depending on which employment structure seems necessary and proper in relation to the enumerated powers the employees will exercise.
Assume for the moment that tenure helps civil servants perform their job functions more efficiently, and that it also helps them administer federal statutory policies more rationally. Even if tenure has those good consequences, they only go to show that tenure is constitutionally "necessary" — not that it is "proper." A "proper" job condition means the position contributes to the Constitution's overall structure, so that the employee helps the constitutional separation of powers work well. A proper employment arrangement also promotes the purposes of government set out in the Constitution's Preamble — to "establish Justice" in the United States, and to "secure the Blessings of Liberty" to all Americans now and in the future.
Going further, one can't understand what "Justice" and the "Blessings of Liberty" mean in the Preamble without consulting the Declaration of Independence. According to the Declaration, government has two main goals. One is protecting its citizens' "unalienable Rights" — "Life, Liberty, and the pursuit of Happiness." The other is promoting republicanism, which takes its bearings from "the consent of the governed." Whenever any particular arrangement of the federal government "becomes destructive of these ends," the people have the right to reorganize federal powers "in such form, as to them shall seem most likely to effect their Safety and Happiness."
In short, whenever members of Congress — the people's representatives — decide whether a particular employment arrangement is "proper" in the executive branch, they should consider two factors more than any other: rights, and republican self-government. Those are "Madisonian" priorities because Madison himself invoked them in Federalist No. 37. Members of the Constitutional Convention walked a fine line when they drew up the new federal government, Madison argued: They needed to "combin[e] the requisite stability and energy in government, with the inviolable attention due to liberty and to the republican form." What Madison said about the Constitution's design generally makes for good advice about civil-service reform specifically. "Stability" and "energy" are proxies for all of the values associated with a well-functioning bureaucracy. But "stability" and "energy" must be reconciled with the two values set out in the Declaration — unalienable rights (for Madison, "liberty"), and respect for consent of the governed (via "the republican form").
Some executive-branch jobs do require considerable stability and energy, enough to warrant tenure in office. But tenure can also undermine those qualities. The most obvious danger is the one Alexander Hamilton hinted at in Federalist Nos. 70 and 76: Tenure might encourage civil servants to undermine the unitary executive.
Overprotective job guarantees can also threaten liberty and the republican form. To respect what Madison called "liberty," the federal workforce must carry out its responsibilities in a fashion that respects the rights of American citizens. For the civil service to accord with what Madison called "the republican form," it must be responsive to presidents and their closest advisors. For the executive branch to play its part in our constitutional order, presidents must be able to translate electoral mandates into real-life policies.
The justifications for giving Article III judges tenure are clear and reasonable. As Hamilton explained in Federalist No. 78, life tenure attracts lawyers with "requisite knowledge" of the law, and it gives judges some protection against presidents and the executive's power over the "sword." But those reasons don't extend to tenure for civil servants. Federal judges apply law while deciding cases, and they shouldn't answer to voters because ordinary citizens do not have legal expertise. Civil servants, by contrast, help the president and executive-branch officers execute the law. Far more discretion goes into the execution of law than in its interpretation during adjudication, and that discretion can be used to threaten American citizens' liberties. Furthermore, voters do not need to be consulted about how executive discretion should be understood and applied. Presidents answer to voters; civil servants do not because they answer to presidents.
TENURE IN AN AT-WILL CIVIL SERVICE
The Madisonian case against overprotection of the federal workforce doesn't necessarily mean that tenure is always undesirable in a civil service. But it should inspire a shift in thinking about whether tenure is the presumptive arrangement.
Specifically, when tenure is granted, it should be justified as an exception — not as the presumptive arrangement it now is under Title 5. Tenure makes it difficult for the federal executive to operate in a "republican form." Not everywhere and not always, but often enough that it should be limited.
If tenure might be appropriate for certain civil servants, the case for it should be made position by position, not for broad classes of employees. For example, Madisonian "stability" justifies making tenure part of the package for jobs that require professional skills; these might be representation of the United States abroad (diplomats) or technical expertise (patent examiners, or accountants in departments that administer federal tax or securities law). "Energy" justifies tenure as part of the package for jobs that cannot be carried out effectively without considerable experience (prosecutors). But those justifications are narrower than what we now have under Title 5 — an entitlement for all civil servants who have worked for the federal government for more than a year, in fields that haven't been exempted.
And even when federal jobs require attributes associated with stability or energy, it still might not be proper to create tenure for them. Tenure is inappropriate for jobs that come with powers to threaten the liberty or privacy of U.S. citizens. If there is a serious risk that tenure might encourage civil servants to harass American citizens while carrying out official job duties, the safer course is to make such a job an at-will position. Think again of the tenure exemptions included in the 1978 Civil Service Reform Act — for civil servants who work in law enforcement, intelligence, and national security. Anyone with the powers to arrest, spy, or kill can threaten core liberties. But so can tax or securities staff; they can use their official powers to violate citizens' privacy and charge them with heavy penalties. The same goes for administrators who process applications for valuable permits, as well as those officials who recommend policies as sweeping as the ones the federal government put in place to encourage Covid-19 lockdowns.
Tenure is also inappropriate when it appears likely to interfere with the republican character of the federal executive. Today, that limit matters most in federal agencies that administer open-ended enabling statutes. At those agencies, every presidential election is a referendum on how the relevant enabling statutes should be interpreted and enforced for another four years. Voters expect presidents to change agency policies as they campaigned to do, and presidents are entitled to expect that agency staff will implement the proposals on which they ran. If there is a serious risk that tenure might encourage civil servants at such agencies to disregard directions from presidents and their key aides, the safer course is to make the relevant jobs at-will positions.
It should be noted that this more restrictive approach to tenure has a downside: It is a lot easier for presidents to violate the law when civil servants work at the president's will, rather than being insulated from political direction by tenure. The second Trump administration is raising obvious concerns on this score. It has raised tariffs without clear authority under federal statutes, and it has instructed federal lawyers and immigration officers to skirt injunctions ordered by federal courts.
Those actions are troubling. But it's impossible to address both that danger and the opposite danger — civil servants slow-walking or resisting lawful and reasonable presidential policies. If a legal system cannot address both problems, it is better to avoid the harms of slow-walking and resistance. As unfortunate as presidential lawbreaking is, it can be contained more effectively. Trump's transgressions of federal tariff and immigration laws have both been challenged in federal courts; the Supreme Court will rule on the former in Learning Resources, Inc. v. Trump. Other such instances of presidential lawlessness can be stopped by hard-hitting reporting, congressional investigations, impeachment, and elections.
THE DANGERS OF TENURE IN PRACTICE
The argument above offers a theory for why civil-service tenure lacks a broad justification under our republican form of government, as well as some cases where job protections are appropriate. But it's important to look beyond theory and consider the available evidence. And it turns out that the theory fits the evidence quite well. Interested readers should consult David Bernhardt's 2023 book You Report to Me, which reflects on his service in the Trump administration as deputy secretary and secretary of the Department of the Interior. A few examples (some from Bernhardt, some not), will illustrate the dangers of tenure and the relevance of my republican argument.
To start, sometimes tenure does not concentrate and instead dissipates energy in the federal executive. In the early 2010s, the Department of Veterans Affairs (VA) was the subject of embarrassing news reports. VA hospitals were routinely delaying checkups or treatment for military veterans, and some said patients were dying as a result. An investigation by the acting VA inspector general concluded that at three-fourths of facilities, managers falsified patient wait-time data. The delays led to denied or late care for more than 120,000 veterans, and critics of the VA hospital system allege that hundreds of those veterans died waiting for medical care.
A report co-authored by Rob Nabors, President Barack Obama's deputy chief of staff, blamed VA's problems on "corrosive cultures" and "significant and chronic systemic failures" at department offices, and he further attributed those cultural issues and failures to VA's uppermost leadership. But it would have been reasonable to suspect that the problems went further down VA's organization chart than Nabors concluded. And a closer look would probably have revealed that tenure dragged down the culture throughout VA, making it difficult to remove employees behind the failures.
Conversely, tenure can also embolden civil servants to violate the rights of American citizens. Government service attracts idealists, who are often far too certain that the causes they support are just. When those idealists obtain government jobs, they can violate the legal rights of citizens without even realizing that they might be infringing such rights. Tenure, it must be said, encourages that extremism.
That danger has stoked conflicts between the Department of Justice's Civil Rights Division staff and Harmeet Dhillon, as well as with her Republican predecessors who served as assistant attorneys general for the CRD. In civil-rights policy, conservatives argue that anti-discrimination laws should be applied on a colorblind basis, while progressives contend that they should be enforced mainly to protect minority groups who have historically been the victims of discrimination throughout American history. Career CRD staff trend to the left, and they tend to oppose the litigation priorities of Republican presidents. And when CRD lawyers resist those priorities advocated by GOP presidents, they deny the rights of Americans who haven't been traditionally discriminated against, and who enjoy the same protections under federal civil-rights laws.
This has been a long-running problem in the CRD, especially in its Voting Section. In early 1990s litigation about redistricting in Georgia, attorneys for the American Civil Liberties Union (ACLU) sued the state to make it adopt a racial-gerrymandering plan favoring the election of black lawmakers. A three-judge federal court complained that ACLU attorneys were colluding with CRD lawyers in the Voting Section — like "peers working together, not...advocate[s] submitting proposals to higher authorities."
Another example of such resistance was reported in 2013 by the DOJ inspector general. According to that report, career staff at the Voting Section "ostracized" their conservative career colleagues in the mid-2000s, frequently emailing one another with calls for "resistance" and a "guerilla war" against President George W. Bush's political appointees to the CRD. Of course, many career CRD staff have chosen to quit rather than resist President Trump and Dhillon, despite enjoying the cover of tenure. But tenure shielded those CRD staff members far longer than they should have been protected.
As already alluded to, sometimes tenure encourages civil servants to resist the policy priorities of elected presidents. No recent experience illustrates that better than the Covid-19 pandemic. In early 2020, President Trump established a White House Coronavirus Task Force, appointing Deborah Birx (an SES member) as its response coordinator. One of the task force's responsibilities was to advise states, through regular reports, about how to reduce the likelihood of Covid-19 infections. The task force's draft reports were reviewed by White House staff before going out, and presidential aides removed some recommendations proposed by Birx.
Birx, however, thought she balanced public health and competing policies better than Trump's advisors did. So Birx and her staff reinserted recommendations (deep in the new drafts, where Trump's aides wouldn't see them), and she gambled that White House advisors would miss the reinsertions. As Birx recounts in her 2022 book Silent Invasion, even though White House advisors caught that strategy after the reports were published, Birx had no regrets: She "did what [she] needed to do."
Of course, not all civil servants act in such a high-profile and contested manner as those described here. But the federal executive branch is supposed to work for the American president, so the president can serve the people. The burden lies on the supporters of tenure to show that it doesn't encourage civil servants systematically to underperform; or to harass, resist, or box in political appointees. There are too many examples here for tenure supporters to carry that burden.
RESTORING ACCOUNTABILITY
But again, we return to the reality that today's legislators and policy thinkers assume that tenure is necessary or inevitable. That assumption has subtle effects on federal lawmaking, as illustrated by the recent attempt at tenure reform in the One Big Beautiful Bill Act. It is telling that the Senate removed the House's proposals to broaden at-will employment; the Senate often prefers the respectable status quo against populist ideas that come out of the House. But even more revealing is how cautious the House's proposals were. Those reforms would have offered new civil servants a choice: tenure plus a higher required contribution rate to their retirement plans, or at-will employment plus a lower contribution rate. The authors of that language accepted that tenure was the norm; they hoped to persuade incoming civil servants to give up that presumption of job guarantees with better contribution benefits.
The same presumption of tenure weakened the last bill that Congress did pass on civil-service reform a few years ago, the VA Accountability and Whistleblower Protection Act. That act was passed in 2017 as a response to the waitlisting scandal. If Congress had wanted to reform VA root and branch, it could have changed personnel rules so that, going forward, no new VA employees would be granted tenure. But the 2017 law's employment provisions focused only on the uppermost SES members at VA. The act gave those SES officials at VA weaker substantive protections than those enjoyed by non-VA SES members, and it established a speedier process to settle VA personnel matters compared to standard MSPB procedures.
Unfortunately, President Joe Biden's VA Department eventually gave up on enforcing the new rules in the 2017 act. (VA lawyers were confused about whether certain SES employees at the department were covered by the act or more general civil-service protections, and federal courts construed the law's language more narrowly than its supporters had expected.) But the problems in what Congress did pass are not as revealing here as what lawmakers didn't do. Members of Congress instituted modest limits on tenure for VA staff they targeted, and they made the changes apply to the narrowest set of VA employees possible. Again, lawmakers accepted that tenure was the norm by adopting the smallest necessary changes to that norm.
Why do members of Congress accept that tenure is good and here to stay? There are several answers, but the most important one is this: The most respectable academics, who influence the decisions of our elected leaders, think that tenure is good because it encourages government by centralized administration. Among these academics and policy professionals, the values that matter most in civil-service design are efficiency and rational administration.
The perspective of these academics begins with the notion that a civil service is a bureaucracy. In a well-run government, their argument goes, bureaucracies receive broad directions from elected political leaders. Administrators then implement those directions into real-life policies, and values such as efficiency and rational administration are meant to guide bureaucracies as they convert broad political guidance into concrete actions. To make policies efficient and rational, bureaucratic work needs to be insulated from politics after agency staff receive political input — hence the importance of tenure. It keeps party hacks out of administration, and it protects a state from poorly thought-out proposals by inexperienced political actors.
These arguments have been popular among political scientists and policy scholars for some time, and were recently rehearsed by Joseph Heath in his 2020 book The Machinery of Government. Heath derided "the standard textbook story" that "the people, via their elected representatives, exercise sovereignty and decision-making authority," and he celebrated a model of government in which "the civil service" has broad "autonomy."
In his college textbook The Politics of the Administrative Process, Donald Kettl further elucidated such a civil-service model. According to Kettl, public administration should be guided by three basic policy goals — "politics" (or "the choices among values"), "performance" (also described as "effectiveness" or "efficiency"), and "accountability." In this analysis, the importance of high-performing civil servants can outweigh accountability often enough to justify tenure.
Accountability is a prominent term in scholarship and policy papers on the civil service; civil-service reformers cite the goal nearly as often as reform opponents do. (See, for example, Philip Howard's 2023 book Not Accountable.) But "accountability" is a vague and poor substitute for the corresponding items on Madison's list of good-government values — the effects of government administration on individual liberty and republican self-government.
Accountability leaves unanswered an important question: To whom or what are administrators accountable? Are they accountable to voters, to a state's laws, to the state's legislature or relevant oversight committees, to elected executive officials, or to their professional peers and those peers' standards of conduct? In the law of agency, an "accounting" doesn't get ordered until a court knows who the principals are to whom agents owe an accounting. In studies of administrative governance, however, it isn't at all clear who the relevant principal is who stands to benefit from the accounting. That ambiguity makes it easier to argue that administrative agencies should be independent of political supervision. Accountability could support executive oversight more than any other factor on Kettl's list. As the number of actors or offices to whom bureaucracies are accountable grows, though, agencies become less and less accountable specifically to elected executive politicians. Accountability is repurposed so as not to threaten autonomous administrative governance.
But in the American regime, in which the consent of the governed is one of the highest priorities, "accountability" describes a relation that is more specific and forceful. The Declaration of Independence makes the entire federal government accountable to the American people, so that it respects their rights and encourages republican government. The Constitution makes each of the three departments of government accountable in law and to the American people for specific functions: Congress should legislate, the president should execute federal law, and federal courts should adjudicate disputes about the law. Presidents' responsibilities, in turn, make them accountable to the people — to administer federal laws as written when they are clear, and to interpret and apply them as presidents campaign to do when the laws are not so clear. Thus, every non-officer job Congress creates in the federal executive should be accountable to the president, so the president can be accountable to the American people.
Construed in that spirit, accountability is not the weakest value relevant to civil-service tenure. It is instead the strongest value, and it justifies making at-will employment the presumptive employment arrangement in a civil service.
A MADISONIAN REFORM
When James Q. Wilson received the James Madison Award from the American Political Science Association in 1990, he suggested in his acceptance remarks that Madison himself would never have been chosen to receive the James Madison Award. Whether Wilson's joke was a good one or not, his point has been vindicated by contemporary civil-service scholarship. In today's thinking on administrative governance, the most important policy values are the ones that make tenured civil servants seem necessary or even inevitable, such as efficiency and rational administration. When the relevant goals for the civil service are framed that way, Madisonian values for good government — stability and energy, but also a respect for liberty and the consent of the governed — seem unimportant or downright irrelevant, and Madison clearly isn't deserving of the James Madison Award. Madisonian arguments, however, are not so easily refuted, as the many dangers of our current system of tenure reveal.
Members of Congress and policymakers have conceded too much to what contemporary academics believe about good government — in particular, the notion that the civil service should be largely independent from the political branches and especially from the president. But if we return to Madison and other founders of the American political tradition, we find that the main goals of government are to secure the people's rights and to facilitate republican self-government. At-will government employment better protects liberty and encourages government by popular consent most of the time, while tenure should be reserved for those with special jobs and skills. Congress should act to make this Madisonian reform a permanent feature of the civil service.
