The Insurrection Act is a nuclear bomb hidden in the United States Code. Enacted in the early years of this country’s existence, it was subsequently modified several times, as Congress greatly expanded the President’s powers under the law during the bloody tumult of the Civil War and Reconstruction. It has largely been ignored ever since. But today, the law has garnered renewed attention and raised concerns in many quarters for a reason that should echo ominously this election year: It gives the President virtually limitless discretion to use the military as a domestic police force.

Allowing the President unbounded authority to deploy troops domestically goes against core constitutional principles. It also invites abuse. In a host of scenarios—from political protests to the crisis at the Southern border—the Insurrection Act, in the wrong hands, could be used in ways that are more likely to cause emergencies than to resolve them. In such cases, the courts would likely deem themselves powerless to intervene, and Congress might be unable to muster the supermajority necessary to restrain the President.

This state of affairs is principally a congressional failure, and Congress must act to fix it. Congress should reform the law to narrow and clarify when and how it may be used and to implement meaningful safeguards against abuse. The United States has been extraordinarily lucky that the Insurrection Act has not been misused more frequently. There is no reason to believe this good fortune will continue forever.

The Framers Feared an Army Turned Inward

Mistrust of military interference in domestic affairs has deep roots in our nation’s history. The founding generation was steeped in a tradition of legal and political thought that emphasized civilian control of the military, rejected military enforcement of civilian laws outside of an emergency, and believed that the legislature should control when, where, and how the military could be deployed at home. The American Revolution was motivated in part by the British government’s betrayal of these fundamental principles, which were centuries old in English law. The Declaration of Independence expressed this, accusing King George III of:

[Keeping] among us, in times of peace, Standing Armies without the Consent of our legislatures…affect[ing] to render the Military independent of and superior to the civil power…. Quartering large bodies of armed troops among us [and] protecting them, by a mock Trial, from punishment for any Murders which they should commit on the Inhabitants of these States.

The founders saw the unrestrained use of the British military for law enforcement in the colonies as a violation of their rights, and tragedies like the Boston Massacre, where a group of British soldiers stationed in Boston to enforce unpopular parliamentary legislation opened fire on an angry crowd of civilians, had illustrated for them why those rights were so important.

The drafters of the Constitution hoped to ensure that these abuses would not be repeated under the new system of government they were designing. The delegates to the Philadelphia Convention of 1787 were, as a group, deeply suspicious of military power—so much so that they vigorously debated whether to even allow for the creation of a national standing army. They feared that such a force could easily be turned inward just as the British army had been, becoming an instrument of tyranny that could destroy the young republic. At the same time, the disastrous Articles of Confederation had shown them the opposite danger of a weak central government that could not respond quickly or effectively to emergencies, including insurrections like Shays’ Rebellion.

The Framers thus struck a balance. To guard against a rogue President, the Constitution gives most of the powers related to regulating the military and its activities to Congress. It allows the military to be used in domestic emergencies to “execute the Laws of the Union, suppress Insurrections and repel Invasions,” but it places limits on those operations, most notably by guaranteeing individual liberties that cannot be abridged by any exercise of government power. Perhaps most crucially, rather than directly authorizing the President to deploy the military domestically in a crisis, Article I, Section 8 of the Constitution entrusts Congress with the discretion to grant the President that power through legislation. Since 1792, Congress has done so through an amalgamation of laws, which I will describe in detail below, that we refer to today as the Insurrection Act.

In 1878, when the Insurrection Act was nearly a century old, Congress buttressed these foundational constitutional principles with the Posse Comitatus Act. This law bars the federal military from enforcing civilian laws except when doing so has been expressly authorized by Congress. Today, it is the most important statutory restriction on the domestic activities of the U.S. military, but not an especially strong one—not least because the Insurrection Act was preserved as a broad exception to the rule.

The Insurrection Act exists in fundamental tension with the Posse Comitatus Act and the Constitution. Our constitutional system was designed around the assumption that the President would possess only limited, carefully regulated authority to deploy the military domestically in response to sudden emergencies. Yet under current law, the Insurrection Act gives the President essentially limitless power to use the military at home against Americans, and it includes no safeguards that would allow Congress or the courts to intervene in case of abuse.

A Law Designed for a Nineteenth-Century Country at War

How did we get here? The predecessor of today’s Insurrection Act, the Calling Forth Act of 1792, reflected the founding generation’s intense suspicion of domestic military operations. Under this law’s provisions, if the President wished to deploy the militia to enforce the law (and it had to be the militia, because the use of federal troops would not be authorized until 1807), he first needed to obtain certification from a federal judge that doing so was necessary under the circumstances. Even after obtaining certification, the President was restricted to using only the affected state’s militia unless Congress was out of session; in that case, he could call on other states’ militia forces, but only until Congress was back in session and able to respond to the crisis. The entire law was also set to expire after three years, ensuring that Congress would be forced to reevaluate it.

President George Washington scrupulously complied with these procedures when he invoked the Calling Forth Act to suppress the Whiskey Rebellion in 1794. The next year, Congress replaced the expiring law with a permanent Militia Act. This new statute was broadly similar to the 1792 version but lacked many of its procedural safeguards, including the judicial certification requirement. Rightly or wrongly, Washington’s exemplary conduct during the Whiskey Rebellion had persuaded Congress that it need not be so fearful of executive overreach. The 1795 enactment forms the core of what is today the Insurrection Act. It was amended several times over the next 80 years until, by the 1870s, all the remaining guardrails from the original 1792 law—including the time limits on deployment—had been stripped out.

During this period, the Supreme Court also helped to expand the President’s powers under the Insurrection Act. In the 1827 case Martin v. Mott, the Court was asked to consider whether President James Madison had lawfully invoked the 1795 version of the law during the War of 1812. The Court determined that it could not question Madison’s decision—and neither could anyone else. In the Court’s view, the President’s decision as to whether a given situation justified invoking the Insurrection Act and deploying troops was “conclusive upon all other persons.” In effect, under Martin, an insurrection is whatever the President says it is.

To be sure, Martin is a nearly 200-year-old case, and much has changed in the high court’s jurisprudence. Yet the holding in Martin has never been overturned or even subjected to serious question. Perhaps more importantly, thanks to subsequent amendments, the current Insurrection Act grants even broader authority to the President than the version that was at issue in 1827. There is therefore no reason to assume that the modern Supreme Court would rule differently in a similar case, especially in light of the general tendency of U.S. courts to show deference to the executive in cases related to national security.

In the 45 years after Martin, Congress passed two more amendments that expanded the President’s powers under the Insurrection Act: the first, in 1861, to allow President Abraham Lincoln to effectively prosecute the Civil War, and the second, in 1871, to empower President Ulysses S. Grant to suppress the terrorist insurgency waged by the Ku Klux Klan in the former Confederacy. Congress made one last minor change in 1874, but has not substantively amended the law in the 150 years since. (In 2006, in the wake of Hurricane Katrina, Congress enacted a rushed overhaul of the act, but the governors of all 50 states and Puerto Rico joined together to oppose the changes, which were swiftly repealed.)

Today, the Insurrection Act has three operative provisions. The first, Section 251 of Title 10 of the United States Code, is the oldest, simplest, and least vulnerable to abuse. It allows the President, at the request of a state, to deploy troops to suppress an insurrection in that state. Upon invoking Section 251, the President may “call into federal service,” or federalize, “the militia” of one or more states to subdue the rebellion. In the modern era, this allows the President to federalize the National Guard, placing Guard troops—who ordinarily take their orders from their state governor—under the President’s command and control. In addition to, or instead of, deploying federalized Guard personnel, the President may also deploy members of the active-duty armed forces.

Misuse of this authority is certainly possible—for example, through collusion between a President and state governor with shared interests. But requiring two separate parties to agree that military intervention is necessary significantly reduces the likelihood of abuse compared to the law’s other provisions.

The Insurrection Act’s second provision, Section 252, is significantly broader in the authority it conveys. This is perhaps unsurprising, as it was last updated just after the outbreak of the Civil War to facilitate President Lincoln’s efforts to suppress the rebellion. Section 252 does not require a request from the affected state, or even the state’s consent. Instead, it provides that “[w]henever the President considers that unlawful obstructions, combinations, or assemblages, or rebellion against the authority of the United States, make it impracticable to enforce the laws of the United States in any State by the ordinary course of judicial proceedings,” the President may federalize any state’s militia and deploy them, and/or the federal armed forces, to suppress the rebellion or enforce the law.

The broad, vague, and archaic language of this section is characteristic of the Insurrection Act. Nothing in the law explains (1) what means of civilian law enforcement are included in, or excluded from, “the ordinary course of judicial proceedings,” (2) what constitutes an “obstruction,” “combination,” or “assemblage,” (3) what factors would render one of these occurrences “unlawful,” or (4) what level of interference or disruption would rise to the level of making it “impracticable” to enforce the laws. Against the backdrop of the Supreme Court’s decision in Martin, the implication is that all of these determinations are left solely to the President’s discretion.

The third section of the Insurrection Act, Section 253, is even more sweeping. It was added in 1871, at a time when the Ku Klux Klan was terrorizing and murdering freedmen across the former Confederacy. Like Section 252, this provision allows the President to deploy troops unilaterally, whether the affected state agrees or not. A separate, temporary provision of the same legislation also authorized the President to suspend habeas corpus (i.e., the right of prisoners to seek judicial review of their detention), underlining just how dire the situation in the South had become. The law served its purpose: President Grant used his new authority to crush the Klan within a year. Yet Section 253 has stayed on the books, virtually unchanged, long after the circumstances that justified its creation have faded into history.

Section 253 has never been substantively amended; however, during the 1956 recodification of Title 10, Congress split the provision into two parts. The first allows the President to deploy the military into a state to suppress “any insurrection, domestic violence, unlawful combination, or conspiracy” that “so hinders the execution of the laws” that some portion of the state’s inhabitants are deprived of a constitutional right, provided state authorities are unable or unwilling to protect that right. Presidents Dwight D. Eisenhower and John F. Kennedy famously relied on this authority to deploy troops to desegregate schools in the South after the Supreme Court’s decision in Brown v. Board of Education.

The second part of Section 253 authorizes the President to deploy troops to suppress “any insurrection, domestic violence, unlawful combination, or conspiracy” in a state that “opposes or obstructs the execution of the laws of the United States or impedes the course of justice under those laws.” Once again, the text does not explain what is meant by the terms “unlawful combination” or “conspiracy,” or what would constitute “oppos[ing] the execution of the laws” or “imped[ing] the course of justice under those laws.” If, however, the term “conspiracy” is given its modern legal definition, and any attempt to prevent the law from being enforced would qualify as “oppos[ing] the execution of the laws,” then this provision would, in theory, allow the President to deploy the armed forces against any two individuals plotting to commit a federal crime.

That is such bafflingly broad authority that one might argue the second part of Section 253 cannot possibly mean what it says. Yet there is nothing in the text to refute this interpretation. Moreover, even making the tenuous assumption that a court would feel empowered to review an invocation of the Insurrection Act, judges are, for good reason, extremely reluctant to disregard the plain meaning of an act of Congress. An overbroad law might be found unconstitutional, but the mere fact that a law is exceptionally—perhaps catastrophically—unwise does not empower judges to ignore or invalidate it.

Like Sections 251 and 252, Section 253 allows the President to deploy federalized state militia (i.e., National Guard) troops and/or members of the active-duty armed forces. But Section 253 also authorizes the President to use “any other means” to “take such measures as he considers necessary” to enforce the law. There are essentially no limits on what the President can do under this provision. To be sure, anything the President does must comply with the Constitution and federal law, but the dangers of handing the President a blank check like this are obvious. If nothing else, it could provide cover for measures that Congress almost certainly did not intend to authorize, such as martial law (under which the military does not merely assist civilian authorities, but entirely displaces them) or the suspension of habeas corpus.

In addition, the authority to call up “the militia” provided by all three operative sections of the Insurrection Act may be far broader than it at first seems. The National Guard is the modern form of the state militias, and as a practical matter, it has taken their place. However, federal law defines the “militia” to include more than just the National Guard, describing it as “all able-bodied males” between the ages of 17 and 45 “who are, or who have made a declaration of intention to become, citizens of the United States” as well as “female citizens of the United States who are members of the National Guard.” Within this group, which includes tens of millions of Americans, the National Guard constitutes the “organized militia,” while all other individuals who fit the above definition are part of the “unorganized militia.”

Because the Insurrection Act refers simply to “the militia,” and not specifically to the National Guard or the organized militia, a President could, in theory, use it to call private individuals into federal service—including members of the Proud Boys, Oath Keepers, and other private militias. Indeed, the Oath Keepers’ Stewart Rhodes defended his actions on January 6, 2021 by claiming that he expected former President Donald Trump to invoke the Insurrection Act and call the Oath Keepers into service to stop Congress from certifying the results of the 2020 presidential election.

The principles of a society founded on respect for individual liberty and the rule of law are incompatible with the amount of power that the Insurrection Act grants to the President. Across the law’s provisions, there are no meaningful criteria for determining whether a given situation warrants using it. Instead, the language of the statute and the Supreme Court’s ruling in Martin both indicate that the President has sole discretion to make that decision. There are also no mechanisms that would allow Congress or the courts to check an abuse of the law.

These flaws are due in large part to Congress’s failure to update the Insurrection Act over the past 150 years. The law was designed for a nineteenth-century country wracked by more than ten years of bloody civil war and insurgency, not for the radically different needs of the twenty-first century United States. It was also written at a time when police as they exist today were a new and developing concept. In the 1870s, even minor unrest could overwhelm the local sheriff or U.S. Marshals and necessitate assistance from the militia. By contrast, America’s modern law enforcement agencies are, for better or worse, massive, heavily armed, and well able to handle most civil disturbances by themselves. It may not be a coincidence that the Insurrection Act has not been used since 1992—the longest the United States has ever gone without an invocation of the law since its first incarnation was enacted in 1792. The country has changed, but the Insurrection Act has not.

The Potential for Abuse

Because the President has sole discretion to decide whether a particular situation justifies invoking the Insurrection Act, the scope for abuse is essentially infinite. In practice, the only real constraints on a President’s use of the law are political. Most presidents would not want to anger the electorate by deploying American soldiers to police American citizens, especially if it meant being the first President to do so in 30 years. Similarly, most presidents would likely balk at issuing orders that might cause senior Department of Defense and military leaders to resign in protest. But these factors will restrain a President only insofar as he or she cares about such political costs. A second-term President not facing the prospect of a reelection campaign, a President who is uninterested in the preservation of norms, or a President who sees chaos and controversy as politically advantageous will feel less bound by these concerns.

It is also true that the President’s constitutional duty to take care that the laws be faithfully executed carries with it an implied obligation to act in good faith when exercising the authorities granted by the Constitution and Congress. In theory, courts should be able to enforce this principle to stop presidents from invoking the Insurrection Act on the basis of an obviously fabricated pretext—such as characterizing a small number of peaceful protestors near the White House as a rebellion against the authority of the United States. But generally speaking, courts are intensely reluctant to inquire into whether an executive action was taken in good faith. This is all the more likely when it comes to the Insurrection Act because of courts’ general deference to the executive in cases that touch on national security and the degree to which the Supreme Court foreclosed any kind of judicial review of the law in Martin.

There are thus a daunting number of ways in which the Insurrection Act could be misused, with few meaningful avenues for recourse. Perhaps most obviously, the law could be used to quell protests against policies or actions that the President either implemented or supported. Even if the protests were overwhelmingly peaceful, the President could cite a few alleged acts of vandalism at the margins—or even a punch thrown by a protester—as a justification for deployment. The consequences of this would ripple far beyond the suppression of a single protest or series of protests. If the President can deploy the military to suppress virtually any demonstration, then many people will be unwilling to face that risk, leading to fewer and smaller protests. As can be seen quite clearly in authoritarian countries around the world, that ultimately reduces the public’s ability to bring pressure against their own government.

The Insurrection Act would also allow a reelected Donald Trump to follow through on most of his campaign’s promises to use the military to carry out an immigration crackdown. Under President Joe Biden, federal troops are already patrolling the Southern border, just as they have been under every President since Ronald Reagan. But by using unlawful migration at the U.S.-Mexico border as an excuse to invoke the Insurrection Act, Trump could direct the military not just to provide logistical and surveillance support to the Department of Homeland Security, as it is doing now, but also to participate directly in immigration enforcement. Soldiers could arrest migrants at the border, man checkpoints to search cars along highways in border states, and even take part in enforcement raids deep within the interior of the country.

Cities and states that refused to cooperate with federal immigration enforcement policies would do so at their peril. The President could frame such resistance as obstruction of the execution of federal law and invoke the Insurrection Act, sending soldiers into those places whether residents want them there or not. In short, while the Trump camp’s plans to deport millions of people might be logistically implausible, the legal authority for using the military as part of this effort is readily available thanks to the Insurrection Act.

A President could also use the Insurrection Act to try to interfere in an election or disrupt the peaceful transfer of power. We have already experienced one near miss. The January 6 insurrection at the Capitol, which aimed to prevent Congress from certifying Biden’s victory in the 2020 presidential election, was in many ways a paradigmatic example of why the Insurrection Act exists. On that day, civilian authorities were manifestly overwhelmed by a violent mob that posed a clear threat to the physical safety of one of the federal government’s three branches and obstructed its ability to fulfill its legal duties. There is no question that it would have been appropriate for Trump to invoke the Insurrection Act and deploy troops to protect Congress. But if he had instead ordered those troops to shut down the building for days or longer—in the name of keeping the peace, but in truth to prevent or delay certification of his opponent’s victory—it would have been difficult for courts to argue that he had exceeded his (nearly limitless) discretion.

This sort of gambit would not have reversed the results of the 2020 election, not least because Trump’s term as President would have ended automatically at noon on January 20 by operation of the Twentieth Amendment to the Constitution. However, it could have plunged the country into a constitutional crisis, thrown further doubt on the integrity of the election, and ultimately sparked even more violence than that which took place on January 6. And this is just one of any number of ways in which an Insurrection Act invocation could throw a wrench into the workings of democracy. For instance, in a contentious election, a President could cite combative rhetoric on social media as a reason to deploy troops near strategically chosen polling sites on Election Day—a move more likely to intimidate voters than to protect them.

To be sure, the Insurrection Act does not allow the President to direct the military to violate the Constitution or other federal law. Soldiers cannot be punished for refusing to obey an unlawful order, and they have a duty not to follow an order when it is patently obvious that doing so would be a crime. But suggestions that the military would therefore refuse to carry out many or all of the above directives miss the fundamental problem at the heart of the Insurrection Act: Many of these orders would likely be legal. Only a handful would be clearly unlawful, and even fewer would be obviously criminal. In the face of any doubt, there is a strong presumption in military law that any order is legal and must be obeyed. This reflects the necessities of military service: Good order and discipline require a culture of obedience.

In practice, only senior military leaders would plausibly have sufficient knowledge of the circumstances on the ground, the considerations behind the President’s order, and the relevant law to reasonably argue that they know for certain that an order from the President is unlawful. Even then, those who choose to disobey that order would do so at the risk of not just their careers but their personal liberty. That is as it should be. Under any circumstances, expecting or hoping for members of the military to disobey orders from their elected, civilian commander in chief sets a grave precedent that risks undermining the same fundamental principles that the Insurrection Act itself threatens.

How to Fix the Insurrection Act

The Insurrection Act is a problem that only Congress can fix. To do so, lawmakers must enact legislation that either completely overhauls the law or replaces it with something entirely new. In either case, this legislation will need to address three main problems with the current law: its vagueness and overbreadth, its failure to specify what actions are and are not permitted once the law is invoked, and the absence of any procedural safeguards against abuse.

First, reforms to the Insurrection Act must eliminate the vague, archaic, and overbroad language used throughout the law. Outdated terms like “unlawful obstructions, combinations, or assemblages” are difficult for twenty-first-century readers to interpret and offer little guidance as to the kind of scenarios the law is meant to cover. Reforms must also remove phrases like “[the President] shall take such measures as he considers necessary,” [emphasis added] which suggest to both courts and executive branch officials that the President has unlimited discretion to decide when, where, and how to use the powers granted by the law. In place of these terms, the new Insurrection Act should lay out clear, narrow, and judicially interpretable criteria for whether a given situation justifies invoking the law and deploying troops.

These criteria need not be extremely detailed—in fact, they probably should not be. No statute, no matter how well crafted, can anticipate every possible emergency that might arise. Moreover, in today’s tumultuous political environment, it is important that presidents retain the authority to act quickly and decisively in the event of an actual insurrection or widespread, violent civil unrest. The revised criteria should make clear, however, that the Insurrection Act may be used only when (1) there is a rebellion against government authority, massive civil unrest involving violence or the threat of violence, or obstruction of the enforcement of the law under circumstances that threaten constitutional rights or public safety; and (2) the use of civilian law enforcement personnel—whether local, state, or federal—is, or would likely be, insufficient to address the crisis.

Second, the law must clearly specify what actions are and are not allowed once it is invoked. To begin with, the authority under Section 253 to use “any other means” must be removed. This kind of a blank check, even when constrained by the inviolability of constitutional rights, is incompatible with the rule of law. The law should also not permit deputizing private militias to serve as federal forces—a power that is more likely to cause chaos than restore order. It should clarify that invoking the act does not give the President authority to declare martial law or suspend habeas corpus. And it should specify that federal troops remain bound by applicable federal and state laws while deployed under the act.

Third, and perhaps most important, the new Insurrection Act must include mechanisms that allow either Congress or the courts—and preferably both—to act as checks against misuse of the law. To start, the law should include a time limit on deployments that can only be extended by Congress. That time period should be long enough to not cut short the President’s response to a genuine crisis before Congress has a chance to act, yet short enough to function as a meaningful safeguard against abuse. One week would be a reasonable starting point. At the expiration of this period, the President’s authority to use the Insurrection Act would automatically terminate unless Congress had enacted a joint resolution approving the President’s invocation of the law. A resolution would extend the President’s authority for a longer period (e.g., two weeks), and that could then be extended further, as necessary, by subsequent joint resolutions. This system would force Congress to periodically reevaluate the deployment, ensuring that the extreme measure of using the military inside the United States is continued only so long as both elected branches of the federal government agree that it is necessary.

As for the courts, Congress should make clear that it does not intend for judges to apply the rule from Martin to the new Insurrection Act. On the contrary, the new law should expressly provide for judicial review of the President’s decision to invoke it. Importantly, however, the standard of review in these cases should be quite deferential. The goal of judicial review under the Insurrection Act should not be for unelected judges to second-guess the decision-making of the President in a crisis. Rather, it should be to allow the courts to invalidate obviously pretextual and abusive invocations of the law.

Congress Must Act

Up to now, U.S. presidents have been remarkably restrained in their use of the Insurrection Act. On the one hand, presidents in the late nineteenth and early twentieth centuries frequently used the law to suppress labor movements by deploying troops to break strikes on behalf of employers. But the law also has a long and praiseworthy history of being used to protect marginalized communities. President Grant used it to suppress the Ku Klux Klan and, more broadly, to defend the rights of freedmen in the former Confederacy. In the 1880s, President Grover Cleveland used it to protect Chinese immigrants from violent white mobs in Washington Territory. And, of course, Presidents Eisenhower and Kennedy used the law to enforce federal court orders desegregating Southern schools in the 1950s and ’60s.

But the Insurrection Act has not been used for civil rights enforcement in 60 years. Since the late 1960s, it has instead been used exclusively to suppress civil unrest—principally in cities with large African American populations suffering from systemic racism and oppression at the hands of local and state authorities. The last time the law was invoked was in 1992, in response to the violent unrest that gripped Los Angeles after four police officers were acquitted of beating African American motorist Rodney King.

There is no reason to assume that good fortune and presidential self-restraint will continue to prevent abuse of the Insurrection Act. Congress has ignored the law for 150 years, all while the country and its needs have changed dramatically. In the current political climate, where threatening political opponents and questioning the integrity of elections has become normalized, continuing to turn a blind eye ultimately endangers the United States’s future as a republic ruled by law.