Topic: Constitution, the Law, and the Courts

Bump Stock Ban Bumps Up Against the Constitution

After the tragic mass shooting in Las Vegas, the phrase “bump stock” entered the public lexicon. What was, and always has been, a gun-range novelty was suddenly the subject of national discussion. In the months following the tragedy, Congress considered and ultimately rejected a law banning these devices. Eager to seize political capital, however, the Trump administration sought to ban them anyway.

The administration faced one problem, though: the Constitution. As anyone who’s seen School House Rock can tell you, only Congress has the ability to write new laws. So the administration attempted to give itself such a power by “reinterpreting” an existing law: the National Firearms Act of 1934 (NFA), which heavily regulates “machineguns.”

For decades, Congress, the executive branch, and the people shared a common understanding: the definition of “machinegun” in the NFA was clear, applying only to weapons that fired continuously from a single function. Bump stocks, which require substantial user input to fire, had never been considered “machineguns,” with precedent spanning multiple administrations. President Trump announced that his administration was changing course. The president expressly declined to go through Congress, instead directing officials to redefine bump-stock devices as “machineguns.” In turn, the Bureau of Alcohol, Tobacco, and Firearms (ATF) broke from decades of precedent and granted itself a new power to ban a widely owned firearm accessory.

This expansion of regulatory authority, motivated by political expediency, cannot stand. Whether one agrees that bump stocks should be regulated or not, this change is not limited to a ban on bump stocks. ATF has asserted the complete authority to ban any new class of weapons that federal law did not address. This approach impermissibly expands the executive branch’s power to rewrite criminal laws and threatens to stifle new developments in firearm technology.

The new rule, making felons of an unknowable number of Americans, was set to take effect March 21. To prevent this, a group of Second Amendment organizations filed a lawsuit in federal court. They sought a preliminary injunction to stop the government from enforcing the new rule, but were denied. Because the effective date is so close, the appeal to the U.S. Court of Appeals for the D.C. Circuit was expedited. Cato filed a brief addressing issues that no other amicus discusses: that the executive branch cannot use the administrative process to accomplish legislative goals that Congress declined to enact.

The implications of this case extend far beyond bump stocks. Regardless of what public opinion is at this moment, the law means what it says. The executive branch has the power to interpret existing law, not write new ones. The administration argues, essentially, that because the statute did not provide a separate definition of the terms that make up the definition of “machinegun,” that it gets to insert their own meaning. That simply isn’t the case. Administrative interpretations are supposed to do just that—interpret existing law—not give new meaning to an old one.

If the government really wants to regulate bump stocks, it needs to do so by passing a new law, not by assigning new meaning to an old one. The Founders weren’t short-sighted; there’s a reason laws that affect the entire nation have to come through Congress, not through bureaucratic reimagination.

The D.C. Circuit hears argument in Guedes v. BATFE on March 22.

Edgy Rap Lyrics Shouldn’t Land a Man in Jail

Jamal Knox was arrested on drug-related charges and released pending further court proceedings. During the interim, he created a rap video in which he made disparaging remarks about the police and named two of the officers who arrested him. The Pennsylvania Supreme Court held that the video demonstrated subjective intent to harm under the “true threats” doctrine and imposed a criminal penalty on Mr. Knox.

The First Amendment protects the right of every person to speak freely without fear of punishment; it specifically protects the ability to criticize government officials. Indeed, the Constitution’s protection of speech is at its highest when government attempts to prosecute someone for his spoken words. One narrow exception is the “true threats” doctrine, which allows a government to punish a person for what he says if his words amount to an actual threat of harm, such as a bomb threat. But just how narrow this exception is remains a mystery.

Courts are divided over whether to consider only the speaker’s subjective intent, or the speaker’s intent and the objective nature of the threat—whether a reasonable person would view the speech as a threat. The second standard offers more protection to the speaker, because the government must prove show both subjective and objective elements of a “threat.” Cato, joined by the Rutherford Institute, has filed an amicus brief asking the U.S. Supreme Court to review this case and ultimately choose the higher standard of protection.

The ambiguity over whether and how the government may criminally prosecute someone for the content of speech is a serious threat to liberty.  The situation is more alarming given that the United States is undergoing a communications revolution, driven by unprecedented new forms of online expression—and unprecedented new attempts by government to monitor and restrict such expression. This case presents an opportunity for the Court to set clear boundaries for the government’s authority to limit online expression through criminal prosecution. 

The Court really does need to step in to avoid chilling protected expression. The subjective-intent-only test fails to protect defendants who are prosecuted for their speech, insulates “true threats” convictions from appellate review, and leaves controversial speakers unprotected even with respect to political or artistic expression.

Senator Hawley’s Apostasy and the Substantive Due Process Problem

A week or so ago, the nomination of George Mason Law’s Neomi Rao to fill Justice Brett Kavanaugh’s seat on the D.C. Circuit ran into some unexpected headwinds when Missouri’s freshman Republican senator, 39-year-old Josh Hawley, raised several concerns about her views, all centered around his opposition to abortion. Fearing that the nomination might fail in committee, the Wall Street Journal’s editorial board took the extraordinary step last week of running not one but two house editorials questioning Senator Hawley’s “judgment.” In the end, the senator came around. On Thursday, Prof. Rao, since July 2017 the head of the Office of Information and Regulatory Affairs and a highly credentialed critic of the administrative state, was voted out of committee on a straight party-line vote of 12-10. 

But questions linger about the motivation and thinking behind Sen. Hawley’s opposition, not least because he himself is highly credentialed (Stanford, Yale Law, clerk for Chief Justice Roberts), and he came of age when the issues he raised were being hotly debated on the Right. He was quoted initially, for example, as saying that “I am only going to support nominees who have a strong record on life”—the “litmus test” approach to nominations more often associated with the Left. But he was also cited as concerned, more broadly, that Rao “might be comfortable with substantive due process,” the doctrine the Supreme Court employed in 1973 when it found a right to abortion. It seems, however, that he may have finally reconsidered that larger concern, for the Journal’s second editorial tells us that “Mr. Hawley now claims he doesn’t object to using this method to incorporate the Bill of Rights to states, only to progressive abuses of substantive due process.” 

Few constitutional doctrines have more vexed conservatives than substantive due process, so a word is in order on the subject, especially given that it’s likely to reemerge with future nominations. Justices Scalia and Thomas have called the doctrine an “oxymoron,” yet that hasn’t stopped the Court’s conservatives from employing it variously, as in finding that the Second Amendment binds the states pursuant to the Fourteenth Amendment’s instruction that no state shall “deprive any person of life, liberty, or property, without due process of law.” And a fortnight ago, joined this time by the Court’s liberals, they again invoked the Fourteenth Amendment’s Due Process Clause to apply the Eighth Amendment’s Excessive Fines Clause against the states, although here, as elsewhere, Justices Thomas and Gorsuch expressed their concerns that the Fourteenth Amendment’s Privileges or Immunities Clause might be the better way to go: “No State shall make or enforce any law which shall abridge the privileges or immunities of citizens of the United States.” 

Thomas and Gorsuch are right, but they’re up against stare decisis. In one of its worst decisions ever, especially considering its context and implications, the Reconstruction Court in the infamous Slaughterhouse Cases of 1873 rendered the Privileges or Immunities Clause “a vain and idle enactment,” as Justice Stephen Field put it in dissent. Meant to protect “the natural and inalienable rights which belong to all citizens,” the Court majority read the clause as protecting only a few federal rights, like “the right to use the navigable waters of the United States.” There the clause has sat, all but “idle” ever since. One might think that result an embarrassment for conservative originalists, constitutional textualists as they generally are.

What then explains their refusal to override stare decisis, a principle at its weakest in constitutional adjudication. Hawley put his finger on it: progressive abuses of substantive due process. After the Slaughterhouse debacle, Fourteenth Amendment cases continued to come before the Court, of course. With the Privileges or Immunities Clause having been reduced to a nullity, however, the Court turned to the Due Process Clause, finally incorporating its protection for property rights in 1897, for contractual freedom in 1905, free speech in 1925, and so on, one right at a time. But after the New Deal constitutional revolution, the Court was largely deferential to the political branches and the states, until the mid-1950s when it got its second wind, and not a moment too soon in the case of civil rights, criminal procedure, and certain other rights. But at the same time, the Warren and Burger Courts were finding other “rights” that were nowhere to be found, even among the unenumerated rights the Ninth Amendment was written to protect. And that led to a conservative backlash and a call for “judicial restraint.” 

It is that fear that lingers today among a fair number of conservatives, although in the last few decades, the debate has shifted on the Right, with ever more conservatives, prodded often by classical liberals and libertarians, recognizing the political infirmities of judicial restraint and, more important still, its constitutional inconsistency. For if textualism is originalism’s bedrock methodological principle, then judges cannot ignore the plain text of the Ninth and Tenth Amendments—they must be read together—or the Privileges or Immunities Clause of the Fourteenth Amendment. Rather than restrained, judges must be engaged. And that means, when necessary, informing the text with the rich natural rights and common law theory that stands behind it.

So Senator Hawley was right, eventually: The problem is with progressive abuses of substantive due process—with court’s finding rights nowhere to be found (and powers nowhere granted). Thus, the Griswold Court was right. The Connecticut statute that criminalized the sale and use of contraceptives, like the Texas statute years later in Lawrence that criminalized private same-sex sodomy, was enacted under the state’s police power, the power mainly to secure our rights. But in those cases, there were no rights to be secured—no plaintiffs who might have brought a civil suit or criminal complaint against the defendants—yet there were rights the statutes did infringe, our basic natural right to liberty, a right “retained by the people.” 

Roe v. Wade, however, is a very different case, for the question there was whether the Texas statute at issue was indeed protecting rights, the rights of the unborn child. Here too the scope of the police power was at issue. Just after Samuel Alito’s 1985 Justice Department memo on abortion came to light—his 2007 Supreme Court confirmation hearings about to begin—I argued in the Wall Street Journal that because abortion at bottom is a criminal law matter concerning where to draw a line about which reasonable people can reasonably disagree, this police power question belongs with the states, which in 1973 were already drawing that line, in different ways, as they should have been left to do. That is a point that no less than then-Judge Ruth Bader Ginsburg made in her NYU Madison Lecture in 1993, two months before she was nominated for the High Court—and it almost nixed the nomination! 

Thus, contrary to what too many conservatives have too long believed, it is hardly difficult for judges to read the text of the Constitution—the whole text—in light of the document’s structure and background assumptions and theory. Griswold, Lawrence, and similar cases, especially many involving economic liberty, are straightforward rights cases, easily decided by courts. Roe was more difficult, but in the end, because the inherent line-drawing belongs properly to the people and their state legislatures, absent egregious examples like those that would sanction infanticide, it was a federalism case. 

Unfortunately, with the Slaughterhouse Cases, the Fourteenth Amendment got off to a bad start, and it’s never fully recovered. And the debate over substantive due process, a concept with roots in Magna Carta, has often only clouded matters. The Due Process of Law Clause will do the job, but it takes more work than many judges are prepared to do. As I wrote above, Justices Thomas and Gorsuch are right. The Privileges or Immunities of citizens of the United States Clause is the better route, because it takes us more directly to the Constitution’s text and, especially in light of the ample debates that produced it, to the natural rights theory that lends substance and legitimacy to our political and legal order. 


Unconscious People Can’t Consent to Police Searches

A reasonable expectation of privacy is one of the most fundamental rights people hold in a free society. Accordingly, the Fourth Amendment prohibits warrantless searches, with few exceptions. Police officers in Wisconsin violated that right when they drew Gerald Mitchell’s blood while he was unconscious—to test his blood alcohol content after a drunk-driving arrest. The state has attempted to excuse the officers by citing an implied-consent statute, which provides that simply driving on state roads constitutes consent to such searches.

The right to privacy is not absolute; police are allowed to search for evidence of a crime. But in doing so, they must follow procedures that comport with the Constitution. Before police conduct a search, Johnson v. United States (1948) indicates that the evidence should be judged by “a neutral and detached magistrate instead of being judged by the officer engaged in the often competitive enterprise of ferreting out crime.” The Fourth Amendment contains a simple requirement for law enforcement that is an effective bulwark against unreasonable searches: get a warrant first.

Unfortunately for Mitchell, the Wisconsin Supreme Court upheld this unconstitutional search under the “pervasively regulated business” exception, which allows for warrantless administrative inspections of certain highly regulated businesses. But this exception is quite narrow and designed to ensure regulatory compliance, not to facilitate evidence-gathering in cases of suspected of criminal activity.

The U.S. Supreme Court has only recognized four types of business to which the exception applies: liquor sales, firearms dealing, running an automobile junkyard, and mining. None of these resemble the simple act of driving a vehicle. The state court thus erroneously conflated the licensing of a driver with a highly regulated business order to justify an otherwise unreasonable search.

Gerald Mitchell is thus asking the U.S. Supreme Court to overturn the Wisconsin Supreme Court and find that this warrantless, non-consensual search violated his Fourth Amendment rights. Cato has joined the Rutherford Institute in filing an amicus brief in support of his petition. We argue for the basic notion that unconscious people can’t consent to anything, especially police searches, and that inspecting a coal mine for safety compliance—a justified exception to warrantless searches—is not the same as searching a driver’s blood in an attempt to convict him of DUI.

John Dingell and the Death of Institutional Pride in Congress

John Dingell, who represented southeastern Michigan for 59 years in the House of Representatives, died on 7th February. While I disagreed with virtually all of Rep. Dingell’s policy preferences, I very much admired the man for the institutional pride he took in Congress. It’s a quality sorely lacking in contemporary lawmakers.

Rep. Dingell was a master parliamentarian. He is famous in Washington D.C. for his steady accumulation of power as chairman of the Energy & Commerce Committee during the 1980s. Due to his efforts, Energy & Commerce presently exercises the broadest jurisdiction of any authorizing committee in Congress. 

From this perch, he regulated the regulators. His use of agency “details”—i.e., commandeering executive branch officials to assist in oversight—was legendary. His investigations were feared. Again, I disagreed with his policy preferences, but, undoubtedly, he was effective in competing with the president (and his colleagues) for control over agency policymaking.

Today, alas, the president calls the shots at regulatory agencies, largely unfettered by Congress. As described by then-Prof. Elena Kagan in a celebrated law review article, we live in a time of “presidential administration.”

Over the last few decades, the president has developed powerful mechanisms, including budget management and White House regulatory review, to superintend the administrative state. Yet, in the face of growing executive power, Congress declined to compete. As I explain further below, party allegiance trumps institutional pride in today’s Congresses. Most congressional Democrats looked the other way, or even cheered, when “their guy” used a “phone and pen” to circumvent Congress and unilaterally impose a deferred action immigration policy. And most congressional Republicans looked the other way, or even cheered, when “their guy” declared a “national emergency” to circumvent Congress and unilaterally build a wall. I’m confident that such obsequiousness would’ve been anathema to a Congress full of John Dingells.

What made Rep. Dingell different? He was a product of his times. Whatever else their flaws, lawmakers of his era abided certain norms that made for a more competent Congress. Under the “apprenticeship” norm, new members of Congress were expected to keep a low profile and instead devote themselves to committee work. Under the “specialization norm,” lawmakers were expected to master one to two narrow issues within their committee’s jurisdiction. To be sure, Dingell’s generation engaged in overbroad delegations of lawmaking functions to agencies, but these lawmakers at least felt a duty to counterbalance such grants by taking the time to understand the inner workings of the agencies they created and funded. The 1946 Legislative Reorganization Act, for example, requires congressional committees to exercise “continuous watchfulness” over regulatory agencies.

Moreover, lawmakers of Dingell’s era genuinely cared about Congress as an institution. To the Congresses that legislated the administrative state into existence, the president was a competitor, and oversight was a primary means for ensuring that laws were faithfully executed. To a large extent, the legislature’s oversight role was codified in various enabling statutes, which often required agencies to consult with appropriate committees before implementing proposed policies. Back then, agencies usually acknowledged and honored objections voiced by members. Congresses of yesteryear also authorized lawmakers to check regulations after they already had been issued. From the New Deal until the early 1980s, Congress legislated 295 veto provisions in 196 statutes, pertaining to every field of governmental concern. In 1983, however, the Supreme Court ruled that this sort of legislative oversight is unconstitutional.

At about the same time that the Supreme Court rejected the legislative veto, Congress began a structural change that further undermined its capacity to compete with the president. Specifically, there was a shift from committee-control to party-control over both chambers of Congress, such that, again, party affiliation currently takes priority over institutional affiliation. In large part, this centralization of power resulted from demographic shifts that homogenized the Democratic and Republican parties, primarily the decline of northeastern Republicans and southern Democrats. Another major factor contributing to this trend is the “fiscalization” of politics in Washington D.C., by which scholars mean that congressional affairs are now dominated by periodic high-stakes negotiations over budget deficits, debt ceilings, and government shutdowns. In these recurrent crises, congressional leadership naturally assumes a greater hand. Congressional leaders accelerated the centralization of power by slashing professional support staff at committees and Article I agencies.

In this environment, Congress has had insufficient political will to reinstitute a viable mechanism for checking administrative policymaking. In 1996, Congress partially revived the legislative veto, but its scope was limited to resolutions of disapproval rather than resolutions of approval, which meant that a sitting president could use his constitutional veto power to protect regulations promulgated during his term. As a result, Congress gained veto power over only a small subset of some presidents’ regulations—basically, those rules promulgated in the final five months of an outgoing presidency, where the incoming president is from a different party. The crucial point is that the legislative veto is a shell of former self.

Obviously, norms like specialization and apprenticeship have fallen by the wayside—just look at the AOC phenomenon. Less than two months into her first term, she has basked in the political limelight and even introduced a sweeping proposal that would fundamentally overhaul the American economy. “Apprenticeship” plainly does not apply. Nor does “specialization.” In modern Congresses, members tend to be generalists when they enter, and they are content to remain generalists while they stay.

Compared to members from Dingell’s time, lawmakers today know less and have fewer resources. They also work fewer days. Effective oversight has suffered accordingly. Consider the House Natural Resources subcommittee hearing on climate change last Tuesday, which was adjourned by the minority party because too few Democrats bothered to attend. Mind you, it’s a Democrat party plank that climate change is the most pressing problem of our time.  Quality of oversight isn’t any better in a cleanly divided government, where both chambers are held by the party in opposition to the president, as was demonstrated amply by the Republicans during the Obama era.

To be clear, I’m wide-eyed about the realities of power. I understand that Dingell wielded his authority to benefit his constituents—in particular, the Big Three automakers in Detroit. While there are those who genuinely believe that “what is good for GM is good for America,” I’d no more want Rep. John Dingell than anyone else to wield the powers of the modern president. Rather, my preference is for Congress and the president to battle to manage the administrative state—like they used to—and thereby tame the leviathan. That’s how the system is supposed to work. 

What will it take to make Congress great again? We’re going to need a few score more members like John Dingell, in both houses and on both sides of the aisle. R.I.P.

Even Something as Great as School Choice Should Not Be Federalized

Today, Sen. Ted Cruz (R-TX) and Rep. Bradley Byrne (R-AL), in conjunction with U.S. Secretary of Education Betsy DeVos, will unveil a bill to create a $5 billion scholarship tax credit, an unprecedented federal school choice effort. An op-ed all three have in USA Today spells out both the good of federal school choice, and inadvertently, the potential bad which makes it too dangerous to justify.

First, the good. DeVos, Cruz, and Byrne argue, quite rightly, that “education isn’t about school systems. It is about school children.” If you recognize basic reality, you’ll know that all children and families are different—different talents, values, dreams—hence it makes no sense to say all should get uniform education. But opposing school choice is de facto endorsing the idea that education should be largely uniform. One size must fit all.

They also make another crucial point, one that is starting to elicit push-back from public schooling advocates who insist that public schooling and public education are synonymous. DeVos, Cruz, and Byrne write that their proposal is not “an attack on public education.” Of course it isn’t. For one, they say their proposal would allow credit-eligible funds to be used for public school options. More broadly, just as public assistance doesn’t mean every recipient of help must go to a government grocery store, nothing about public education implies government must supply the schools. Indeed, we’ve been moving away from things like government housing projects for decades.

Now the bad. School choice is about individualization and freedom, and almost certainly that is what DeVos, Cruz, and Byrne want. But federal initiatives are a terrible way to deliver that. The reality is that what the feds fund, even indirectly, they inevitably want to control. DeVos, Cruz, and Byrne specifically acknowledge that historical reality in federal education policy. They write, “A series of administrations on both sides of the aisle have tried to fill in the blank with more money and more control, each time expecting a different result.” Note that the primary vehicle for that control, the Elementary and Secondary Education Act, started aimed just at funding low-income districts. It eventually became the uber-controlling No Child Left Behind Act.

DeVos, Cruz, and Byrne are looking to skirt the control problem, sticking with tax credits instead of vouchers, and letting states opt in. But not only is this unconstitutional—taxes are authorized to execute specific, enumerated powers, not to lightly engineer state policy—it won’t, ultimately, prevent encroaching federal control. If enacted, the credit would spur people to demand their states participate, and as more schools benefited from federally connected scholarships all schools would be financially pressured to use them. But the federal government will have the power to decide which state programs are or are not eligible, and on what grounds. As Corey DeAngelis and others have noted, what happens when, instead of a President Trump, we have a President Sanders or Harris and they don’t like the policies of religious schools, or maybe how economics is taught? Suddenly lots of private schools and other options will be federally pressured to look very similar—shape up or credit eligibility goes away—and true choice will be curtailed.

Even the roll out of the proposal raises the specter of federal control. Though the great benefit of tax credits is they do not use government money, and hence are less prone to regulation than vouchers, DeVos, Cruz, and Byrne write that through their proposal they “are putting forward a historic investment in America’s students.” That sure sounds like the federal government is doing the funding, and what government funds it tends to control. Also, that Secretary DeVos is so prominent in the proposal release at least symbolizes not only federal intervention in education policy, but a strong connection to the executive—the dangerously regulatory—branch of the federal government.

School choice is great, and DeVos, Cruz, and Byrne recognize that. But as with so many policies, we cannot let our hearts overcome either our adherence to the rule of law—the Constitution—or make us underestimate the potentially crushing unintended consequences that the product of our pure motives may have.

A Rose by Another Name Still Merits First Amendment Protection

Barronelle Stutzman owns and operates Arlene’s Flowers, where she designs floral arrangements for a variety of occasions, including weddings. Mrs. Stutzman is also a practicing Christian; she believes that marriage is a spiritual union between a man and a woman and will not create floral arrangements for same-sex ceremonies. For this reason, when long-time clients Robert Ingersoll and Curt Freed asked Mrs. Stutzman to create floral arrangements for their wedding, she respectfully declined and referred them to several nearby florists.

To be clear, Mrs. Stutzman serves everyone. She gladly created Valentine’s Day and anniversary floral arrangements for Messrs. Ingersoll and Freed for nearly a decade before this litigation, all the while knowing they were a same-sex couple. She just has a sincere religious objection to creating her expressive floral works for same-sex weddings.

Nevertheless Ingersoll and Freed sued Stutzman, which suit was later consolidated with another one brought by the state of Washington. The state trial court ruled against Arlene’s Flowers and the state supreme court affirmed, holding that floral design did not constitute First Amendment-protected artistic expression. Stutzman took her case to the U.S. Supreme Court, which held it pending its decision in the factually similar Masterpiece Cakeshop case last year. The Supreme Court then remanded Arlene’s Flowers v. Washington back to the Washington Supreme Court for reconsideration. As it has in previous stages of this litigation, Cato has filed an amicus brief supporting Arlene’s Flowers, urging the Washington Supreme Court to revise its earlier ruling and hold that floral design is constitutionally protected expression.

The plaintiffs here simply fail to recognize the difference between discrimination based on sexual orientation and refusing to create messages that violate one’s conscience, which is an important First Amendment right that the Supreme Court has repeatedly affirmed. Floristry, like painting, dance, or music, is art, and art is speech protected by the First Amendment. The Court declared in Wooley v. Maynard (1977) that the government can’t force people to speak, even when the message is simply a state motto, and the means of speaking is just displaying the motto on one’s license plate. The First Amendment protects “freedom of the individual mind,” which the government violates whenever it tells a person what she must or must not say. Forcing a florist to create a unique piece of art similarly intrudes on that freedom of mind.

Just last year, the U.S. Supreme Court in NIFLA v. Becerra and Janus v. AFSCME reaffirmed its commitment to striking down laws that compel speech. Justice Clarence Thomas’s concurrence in Masterpiece Cakeshop also illustrates the unique danger of forcing someone to speak against his or her conscience.

While Wooley provides important constitutional protection, it also offers an important limiting principle to that protection: Although florists, writers, singers, actors, painters, and others who create speech must have the right to decide which commissions to take and which to reject, this right does not apply to others who do not engage in protected speech. The court can rule in favor of Arlene’s Flowers on First Amendment grounds without blocking the enforcement of antidiscrimi­na­tion law against denials of service by caterers, hotels, limousine drivers, and the like.

In sum, the government should not be allowed to persecute expressive professionals for declining to create the government’s preferred messages.