Robert Mueller and the FBI can’t save us.

The special counsel’s investigation into Russian interference with the 2016 election, possible Trump campaign collusion with it, and possible obstruction of justice related to those are vitally important. I am, of course, in complete agreement with the public letter to the Congressional Republican leadership organized by the Niskanen Center calling for them to protect the integrity of the investigation. In light of the widespread nondisclosure of and outright lies about contacts, meetings, and financial and political connections between Russian interests and members of the Trump campaign and administration, it’s very hard to know what we don’t yet know. There might well turn out to be substantial criminal activity behind all this obfuscation, and certainly members of Trump’s circle shouldn’t be above the law. The investigation moreover seems likely to help ongoing counterintelligence investigations and planning for future elections. The Putin regime is not going to stop disrupting and interfering in elections in western democracies, and understanding how domestic political actors were manipulated, corrupted, or persuaded into acting out Russia’s script can help the US and other western democracies mount defenses in the future.

And, frankly, what’s already on the public record—not only involving Paul Manafort, Rick Gates, and George Papadopoulous, but also Carter Page, Michael Flynn, and members of the Trump family, from the meeting last summer in Trump Tower to the president’s firing of James Comey—already amounts to an unprecedented scandal. Yes, members of the Trump campaign were in private communication with Russian interests as well as being entangled in unreported financial relationships with them. Yes, members of the Trump campaign knew that Russia had access to hacked information from the Democratic Party. Yes, members of the Trump campaign knew that Russia was interested in putting a thumb on the scale and helping them; they were happy to hear it and didn’t report it. Yes, members of the Trump campaign and future administration sought out back-channel communications with the Putin government during the transition. And, yes, when the FBI investigation of Michael Flynn pushed too hard, Trump fired Comey, concocting a false justification for it before acknowledging, then bragging, that he did so because of Russia.

But the Mueller investigation is, and will continue to be, vulnerable to political interference. Most obviously, Trump could still fire Jeff Sessions and then appoint a successor who hadn’t recused himself from the Russia investigation and who would fire Mueller or order the investigation to conclude. Or Sessions might resign, though that seems less likely; he keeps signalling that he’s not interested in a voluntary change of job (most recently, that he is not interested in returning to his old Alabama Senate seat by write-in), and he seems to have weathered Trump’s season of trying to humiliate him into quitting. Moreover, a series of presidential pardons could wipe out Mueller’s investigative leverage, which relies on indictments and threats of indictments against lower-level or more marginal figures to get them to reveal information on more important figures.

The problem of the connections between Russian interference on one hand and the Trump campaign, administration, and family on the other is a genuinely political problem. Suppose that the Mueller investigation turns up evidence of widespread lying to Congress and the FBI, widespread nondisclosure of Russian contacts, financial crimes by Manafort in 2016, and financial crimes related to Russian money by Trump and his family before 2016, but nothing more than the winks and nods already on the public record between the Trump family and Russian interference in 2016. There’s nothing in the criminal law that can dictate what happens next. If he issues a few (or a few dozen) pardons at that stage to tie up loose ends, while some cases might still be viable in state courts, there’s really nothing left in the criminal law to prevent Trump from tweeting his “NO COLLUSION!” triumphalism and trying to barrel on through his term.

The courts can’t save us.

Consider the ongoing legal fights about Trump’s various executive orders on immigration. In response to the highly unusual circumstances of those orders, the federal judiciary has shown itself unusually willing to intervene and limit executive discretion on immigration questions. The headline-grabbing aspects of the first two versions of bans on migration from several mostly-Muslim countries spent longer being blocked by injunctions than they did actually in effect; and the version of the ban that was eventually allowed to go into effect by the Supreme Court was much whittled down from the original attempt. Courts have stretched traditional boundaries in taking into account Trump’s statements (on the campaign trail and on Twitter since his inauguration) about the order’s purpose and intent. It is still possible that the Supreme Court will rule on the merits of the challenges to those first two executive orders; it has not yet decided whether those cases are moot in light of the second order’s expiration and its replacement by a third.

But I would bet against it. The courts have good reason for reticence. They are institutionally reluctant to pick fights they can’t win with either Congress or the presidency. Only deciding live cases and controversies is a fundamental norm of the American judiciary. And the executive branch has constant opportunities to play shell games with its policies in response to judicial challenges. The addition of North Korea (from which the United States gets essentially no immigration) and Venezuela (in an asymmetric way that makes its inclusion misleading) to the Muslim-majority countries on the original list is a good example. If the courts look likely to restrict executive discretion to engage in religious discrimination in immigration, the executive can lightly disguise it.  In the time it takes slow, deliberate courts to reach a final decision about that policy, the policy can change again. The executive’s built-in speed advantage over the judiciary, and its freedom to opportunistically alter particulars while the judiciary struggles to find general principles, make it extremely difficult if not impossible for the courts to keep up.

And even courts stretching their traditional constraints and boundaries can’t substitute for decency in policymaking. No court has interfered with the Trump administration’s slashing of the number of refugees admitted, and the Supreme Court allowed the temporary complete ban on refugee admissions to go into effect. This is almost certainly right. There unavoidably is a substantial domain for discretionary policymaking in the executive, and the judiciary lacks the expertise, institutional capacity, or legitimacy to substitute its own policy judgments. That it’s a moral horror to confine people to war zones or tyrannies, and appalling shirking to shift ever more of the obligation to protect refugees onto the overtaxed countries that immediately adjoin conflict zones, doesn’t mean that it’s illegal as a matter of US domestic law.

Even the Constitution can’t save us.

If we didn’t properly appreciate the weakness of what Madison described as “parchment barriers” before 2017, the example of the Emoluments Clause under Trump means we surely must now. Trump has defied, ignored, or shredded the whole previous system of norms about avoiding financial conflicts of interest and the use of public office for private enrichment. He has kept his own finances secret, most conspicuously by not releasing his tax returns. He has not put his personal fortune into a blind trust, has not separated his family from his business, and despite an pro forma gesture toward doing so, has not even sharply distinguished between the members of his family working in politics and those running the business. He routinely relocates to properties owned by his companies, especially the Mar-a-Lago resort whose membership fee was doubled (to $200,000) after his election. And he does not release information about the identities of the paying guests at those properties while he relocates to them. He has, in short, drawn a very clear map to foreign interests about how to enrich him and his family and how to gain direct access to him in the process. His administration is mounting an all-fronts resistance to any attempt to apply conflict of interest norms in general or the Emoluments Clause in particular to him. The defiance is impressively forthright, amounting to saying “those rules are for little cases, and cannot be applied to huge cases.” There is no way to separate Trump from his family or his family from their family business; there is no way to separate his family business from foreign money or from personal access to Trump; the rules are impotent in the face of this complete mixture of man, brand, business, and office, and so don’t apply. This is all false, but it has become clear that the ethics rules on the books and the Emoluments Clause in particular are powerless in the face of this level of defiance.

Liberals—and here I do not only mean classical or market liberals, but we are very much included—have a deep-rooted attraction to law and worry about politics. The rule of law is a defining liberal institutional value; procedural regularity, formal equality, and the enforcement of rights all draw the liberal mind to legalism. Indeed, justice and rights are both legalistic concepts in their etymology and history, and they’re central to liberal thinking about governance.

Law aims at certainty, the definitive and correct protection of those who hold rights against those who would violate or undermine them. Politics offers no such certainties. Even at its best it is a domain of contestable judgments that never stop being contested. There is no final settlement; there is always another election. Liberals worry about majoritarianism, and think law can, as politics cannot, protect individuals and minorities from it. We imagine that constitutional settlements can tame politics, confining it within the boundaries of law, ensuring that it complies with justice and respects rights. But they can’t.

The critique of “liberal legalism” has been a major theme in political and constitutional theory in the past two decades. The NYU philosopher Jeremy Waldron has been one of the most prominent voices here. He has argued for many years that we can believe in justice and rights while still also recognizing permanent good-faith disagreement about their precise content. Once we see that clearly, he maintains, we will see the need for those disagreements to be debated politically—in democratically-elected legislatures, for example—rather than short-circuited by recourse to law.

I’ve slowly become persuaded by some of this, and also by some related worries about the political implications of legalism. The American willingness to subordinate everything else in politics to the fight for control of judicial appointments is extraordinarily unhealthy. The most conspicuous examples right now are the ongoing opportunistic rewriting of the procedural rules of the Senate and the idea that Alabamians should elect as Senator a child molester who was twice removed from the bench for disregarding the law, in order to ensure a Republican vote for judicial confirmations. But I’m also tremendously troubled by the classical liberal legalists who seem willing to tolerate any amount of substantive authoritarianism in politics for the sake of friendly judicial appointments.

The current administration shows why the defense of freedom and of the liberal society can’t be an exclusively legal concern. Rules can be manipulated and danced around by the powerful. Legal proceedings are much slower than changes in political circumstances. And executive power is in its nature somewhat lawless. John Locke described executive prerogative as necessary in any system that separated the executive and legislative powers, and defined it as the “power to act according to discretion, for the public good, without the prescription of the law, and sometimes even against it.” (Emphasis added.)

If it’s true that the contemporaneous legislature can’t legally plan for every eventuality, it is of course even more true of a generations-ago constitutional convention. And so the U.S. Constitution does not try to constrain all possible executive misconduct by law, but gives Congress the authority to review executive conduct and to judge officials for “those offences which proceed from the misconduct of public men, or, in other words, from the abuse or violation of some public trust. They are of a nature which may with peculiar propriety be denominated political, as they relate chiefly to injuries done immediately to the society itself,” in the words of Hamilton in The Federalist. The pardon power is arbitrary and lawless; so is the veto. Neither requires justification as a matter of law. For their abuse, and more broadly for the abuse of executive prerogative, the Constitution envisions a political, not a judicial, response: impeachment.

Only politics can save us.

If the independent executive cannot be successfully bound by law, then there is nothing else for it but politics. I’ve argued several times in this space that we need to understand the defense of the liberal society as a political project, one that is dependent on political resources from motivations for popular mobilization to organizational capacity to institutional counterbalances. (See also Michelle Schwarze’s fine essay.) The liberal order of free and open commerce, of religious liberty and freedom of speech and the press, and of rule-of-law constraints on state arbitrariness and violence requires strong political foundations; while law is a crucial part of that order, it can’t pull itself up by its own bootstraps. The liberal society needs an electorate, and elected officials, who are willing and able to stand up for it.

Executive authoritarianism and lawlessness can be hemmed in and checked but not fully constrained by courts, the criminal law, or the written Constitution.

They ultimately have to be confronted by elected officials: co-partisans willing to exercise serious restraint, or if not, an opposition voted into office who will do so instead.

Jacob T. Levy is Tomlinson Professor of Political Theory and Director of the Yan P. Lin Centre for the Study of Freedom and Global Orders in the Ancient and Modern Worlds at McGill University; author of Rationalism, Pluralism, and Freedom and scholarly articles including, most recently,”Contra Politanism”; a blogger at Bleeding Heart Libertarians; and a Niskanen Center Senior Fellow and Advisory Board Member.