Topic: Constitution, the Law, and the Courts

A Victory for Consumer Protections and Health Insurance Freedom

Last year, the Departments of Treasury, Labor, and Health and Human Services worked within federal law to expand consumer protections and restore Americans’ freedom to choose the health insurance that meets their needs. On Friday, a federal court rebuffed an effort to block those protections and force Americans into ObamaCare. First, a little background.

In 1996, Congress exempted “short-term limited duration insurance” from federal health insurance regulations. Congress never defined what “short-term” or “limited duration” meant. So in 1997, the Clinton administration gave meaning to those terms by decreeing that health insurance plans qualify for the exemption so long as they have a contract term, and a total duration, that last less than 12 months. The Bush administration finalized this definition in 2004. 

When Congress enacted the ironically named Affordable Care Act (ACA) in 2010, it tied that law’s copious regulation of the individual health insurance market to the same definition of health insurance Congress created in 1996. That means – you guessed it – “short-term, limited duration insurance” is also exempt from the ACA’s costly regulations.

When the ACA began to make the cost of health insurance soar in 2014, the Obama admininstration noticed that consumers were taking refuge in the short-term market, where premiums could be 50-70 percent lower than ACA premiums. So in 2016, the Obama administration arbitrarily shortened the maximum allowable contract term of such plans to 3 months.

Justice John Paul Stevens, R.I.P.

Justice Stevens, who died yesterday at age 99, was the longest-lived justice in American history and the third-longest serving. A proud son of the Midwest, he lived an amazing American life that included witnessing Babe Ruth’s “called shot” and a valorous WWII stint in the Navy. I never had the chance to meet him but, as the personal accounts and eulogies now attest, he was a consummate gentleman and all class, a bow-tied throwback to an era to which we should all attitudinally aspire.  

Stevens, nominated by President Ford in 1975 as the first new justice after Roe v. Wade, had a confirmation process that lasted all of 19 days and concluded with a 98-0 vote. (A different world, indeed.) From 1994 until his retirement in 2010, he was the senior associate justice, meaning the one who assigned opinions whenever the chief justice was on the other side. A moderate conservative when he was on the Seventh Circuit—perhaps the last of the Rockefeller Republicans—Stevens gradually moved left and ended up as the co-leader, with Justice Ruth Bader Ginsburg, of the Supreme Court’s liberal bloc.  

In his nearly 35 years on the Court, Justice Stevens left a lasting legacy, with majority opinions in Chevron v. Natural Resources Defense Council (granting judicial deference to administrative agencies), Apprendi v. New Jersey (making sentencing guidelines non-binding), Hamdan v. Rumsfeld (striking down military commissions), Kelo v. City of New London (allowing the taking of private property to give to another private owner), and Massachusetts v. EPA (allowing states to sue the EPA over greenhouse gases), as well as famous dissents in Texas v. Johnson (would’ve allowed laws against flag burning), Bush v. Gore (would’ve allowed vote-counting to continue in the 2000 presidential election), D.C. v. Heller (would’ve allowed a complete ban on personal firearms), and Citizens United v. FEC (would’ve allowed certain campaign finance restrictions). 

Many legal analyses of Stevens’s work in the last 24 hours have centered on the above cases, but I want to mention his disappointing view of expansive federal power. Especially when interpreting the scope of congressional authority to regulate interstate commerce, Stevens consistently sided with the government.  

He dissented in United States v. Lopez (Gun-Free School Zones Act) and United States v. Morrison (Violence Against Women Act), cases from 1995 and 2000, respectively, that for the first time in decades found that Congress had exceeded its constitutional power under the commerce clause. Five years later, he authored Gonzales v. Raich, which allows the government to enforce the federal criminal ban on marijuana even against patients who grow and consume the plant for their own personal use consistent with state law.  

These sorts of rulings, when combined with his opinions in Kelo, Johnson, Heller, and Citizens United, show that he really didn’t believe in either structural or rights-based protections for individual freedom, at least with the exception of presidential power and criminal procedure. His jurisprudence was difficult to characterize as a matter of conventional judicial methods and modes, but the results were what we’d now doubtlessly call progressive. In other words, he was a lawyer’s lawyer and a man’s man, a war hero and patriot, but no great friend of liberty.

Reining In Government by Dear Colleague Letter: An Update

For many decades, critics have noted that agencies were using Dear Colleague and guidance letters, memos and so forth — also known variously as subregulatory guidance, stealth regulation and regulatory dark matter — to grab new powers and ban new things in the guise of interpreting existing law, all while bypassing notice-and-comment and other constraints on actual rulemaking.

That’s a problem we at Cato have been concerned about for at least twenty years — the quote itself is from my 2017 post in this space. In financial regulation, for example, as Charles Calomiris argued in a Cato working paper around the same time, agencies’ overreliance on guidance is a systematic failing that “avoids transparency, accountability and predictability.” We’ve published much more on the topic, in both a lighter and a more serious vein.

Since my update post last year, there have been a number of new developments. Soon after then-Attorney General Jeff Sessions’s announcement of the new policy, followed by the revocation of dozens of existing guidance documents, then-Associate Attorney General Rachel Brand issued a January 2018 directive telling Department of Justice attorneys not to rely on allegations of noncompliance with agency guidance, in and of themselves, as reason to initiate civil enforcement actions. And this past winter, DOJ updated its Justice Manual to limit the use of guidance as a basis for direct liability in both civil and criminal enforcement. “Guidance is not law. It’s not binding. And it shouldn’t be given the force or effect of law,” said Deputy Assistant Attorney General Charles Cox in a January speech.

It’s important not to overstate what is changing here: the revised Justice Manual is careful to specify that even if they do not create liability directly, guidance documents can still prove relevant to enforcement in many other ways. They can be used to prove states of mind; as evidence of professional or industry standards, practices, and customs, or that scientific or technical assertions are generally accepted in a particular field; in cases where a regulated party falsely represented itself as having followed guidance; and in various other circumstances.

On April 11, the White House Office of Management and Budget tightened another means of control when it issued guidance (irony alert) about the duty of agencies, including traditionally independent agencies, to notify OMB’s Office of Information and Regulatory Affairs well in advance of publishing new rules to ensure compliance with the Congressional Review Act. The guidance makes clear that “rule” can sometimes include “guidance documents, general statements of policy, and interpretive rules” that do not go through a conventional rulemaking process.

Meanwhile, on a separate track, the unglamorous but often influential federal agency known as the Administrative Conference of the United States commissioned Yale law professor Nicholas Parrillo to write a lengthy study on the guidance issue. (Spinoffs: summary by Parrillo and Lee Liberman Otis, Yale Journal on Regulation and related symposiumFederalist Society panel.) Parrillo agrees that the problems with guidance are real, but argues that most of them arise from structural difficulties within administrative and legal bureaucracy, rather than from a purposeful intent to circumvent process protections. He also finds that they come up unevenly: regulated parties are most likely to feel that they have no real choice but to obey guidance 1) when they need to obtain preapproval before doing business, 2) when repeat interactions with regulators are inevitable and full compliance all the time is unlikely no matter how hard they try; 3) when the consequences of agency enforcement, or even the opening of an enforcement action, are severe; and 4) when the regulated party employs a large dedicated compliance staff.

These might serve as interesting guideposts in looking for ways to revamp regulatory schemes in such a way that agencies’ whims will no longer be received as law.

Whichever Way You Slice It, Courts Have Made a Mess in Applying the ADA to Websites

An old judicial divide over the meaning of “place of public accommodation” in Title III of the Americans with Disabilities Act—which deals with access to private businesses—has in recent years produced inconsistent rulings regarding access to virtual platforms such as websites and smartphone applications. Some federal courts read the text to apply narrowly to physical places like doctor’s offices, while others read it broadly to include non-physical “places” like insurance policies. Before the internet, it wasn’t hard to see which side had the better textual argument. But in an age of omnipresent e-commerce, what was once as simple as a pepperoni pie has since become a fully loaded Chicago deep-dish.

While no court has ruled that virtual platforms are entirely beyond Title III’s ambit, some have limited it to the websites of brick-and-mortar establishments like restaurants and department stores, provided those websites share a commercial “nexus” to discrete physical locations. Others would extend it to website-only businesses (think Netflix). In short, there is no central Title III definition to guide courts on how to fit websites and apps into an analytical framework devised in the analog age. And the Department of Justice (DOJ) has only muddied the waters, offering prevaricating, non-binding guidance, in lieu of long-promised rules and regulations, the latest proposals for which were quietly withdrawn in 2017.

Congress intended the ADA to reshape the cultural and architectural landscape of American society to make it more welcoming to disabled people by compelling businesses to construct ramps, include braille signage, and provide countless other aids to ensure that the disabled have equal access to goods and services. But the ADA doesn’t define “access” with precision. It certainly doesn’t advise businesses on what they must do to avoid Title III liability. It does, however, list the types of “places of public accommodation” to which it applies, places that are alike only in their physicality (e.g., a concert hall and a barber shop).

That brings us to Guillermo Robles, a blind man who sued Domino’s Pizza, alleging that the company’s website doesn’t allow him to order pizza online. The federal district court ruled that, while Title III did apply to websites and apps, the absence of a formal rule meant businesses didn’t have sufficient notice of how to comply. The U.S. Court of Appeals for the Ninth Circuit reversed, holding that existing DOJ guidance is sufficient. Domino’s has asked the Supreme Court to review that ruling and Cato has filed an amicus brief supporting that request.

Courts that define website-only businesses as “places” of public accommodation in themselves have gone a bridge too far. So too have those courts that define the “nexus” between a website and a physical location so broadly as to require every page of a commercial website—even those pages that don’t involve access to a physical location—to have a Title III-compliant interface. And since DOJ remains largely in an advisory role, courts have invented their own compliance criteria, or adopted as binding certain guidelines offered by international standard-setters—even though many businesses have warned of the significant expense these approaches entail.

The compliance costs of this regulatory morass are too great for the Supreme Court to ignore. The confusion emanating from a rudderless bench and a reticent DOJ has opened the floodgates of litigation, driven largely by a plaintiffs’ bar more interested in attorneys’ fees than improving their clients’ lives. By one count, the number of Title III lawsuits rose from 7,663 in 2017 to 10,163 in 2018—a more than 30% increase. Without the Court’s intervention, the costs of this “regulation by litigation” will continue to rise, like so much dough in a wood-fired pizza oven. Although the facts of each case vary, the recent onslaught of claims target businesses both big and small. Today, its Domino’s Pizza or Netflix. Tomorrow, it’s a local contractor with far-more limited resources. The ADA was never meant to be a business-killer, so courts shouldn’t make it into one unless Congress so specifies in no uncertain terms.

Eleventh Circuit Grants Immunity to Officer Who Shot Child Lying on the Ground

The Eleventh Circuit’s decision in Corbitt v. Vickers, handed down last week, constitutes one of the most grotesque and indefensible applications of the qualified immunity I’ve ever seen. The case involves a claim of excessive force against Michael Vickers, a deputy sheriff in Coffee County, Georgia, who shot a ten-year-old child lying on the ground, while repeatedly attempting to shoot a pet dog that wasn’t posing any threat. Without even deciding the constitutional question, a majority of the Eleventh Circuit panel granted qualified immunity to Vickers, simply because there was no case on point with this particular set of facts.

The key facts as alleged in the complaint are as follows: Vickers and other officers were pursuing a criminal suspect, Christopher Barnett, when Barnett wandered into the backyard of Amy Corbitt (who had no relation to Barnett). At the time, one adult and six minor children were in the yard, and the officers demanded they all get on the ground. Everyone immediately complied, and the police took Barnett into custody.

But then, the family’s pet dog Bruce walked into the scene. Without provocation or any immediate threat, Vickers fired his weapon at Bruce. His first shot missed, and Bruce retreated under the home. About ten seconds later, Bruce reappeared and approached his owners, and Vickers fired again – missing once more, but this time striking Corbitt’s ten-year-old child, who was at the time still lying on the ground only 18 inches away. The bullet tore through the back of the child’s knee, causing serious injuries. The child suffered severe pain and mental trauma and has to receive ongoing care from an orthopedic surgeon.

Corbitt, individually and on behalf of her child, filed a lawsuit against Vickers under Section 1983, the text of which guarantees that any state actor who violates someone’s constitutional rights “shall be liable to the party injured.” Vickers filed a motion to dismiss, but the district court held that he wasn’t entitled to qualified immunity, emphasizing that the facts as alleged in the complaint indicated that there was no conceivable safety threat or any need to discharge his weapon at the family’s dog.

But in a split decision, the Eleventh Circuit panel reversed, holding that Vickers was entitled to qualified immunity as a matter of law. Judge R. Lanier Anderson, for the majority, said that there was no prior case law involving the “unique facts of this case,” in which a child was accidentally shot while the officer was intending to shoot someone (or something) else. Although the majority dutifully recited Supreme Court precedent purporting to say that overcoming qualified immunity does not require that “the very action in question has previously been held unlawful,” it is clear from the rest of the opinion that the majority was, indeed, requiring this level of specificity:

No case capable of clearly establishing the law for this case holds that a temporarily seized person—as was [the child] in this case—suffers a violation of his Fourth Amendment rights when an officer shoots at a dog—or any other object—and accidentally hits the person.

We Need More Speech about the Census

Facebook has pledged to remove misinformation about the Census from its platform. Inevitably this removal involves suppressing speech that would be protected by the First Amendment if uttered in a public forum. After all, there is no misinformation exception to the First Amendment.  However, Facebook may remove the speech because as a private firm, it is not obligated to enforce the First Amendment. Nonetheless, we might ask: why speech about the Census? What’s different about misinformation about that project? I see four possibilities as to why speech about the Census might merit closer attention by content moderators.

Facebook might feel obligated to protect basic political institutions. For example, Facebook tries hard to combat misinformation around elections including the 2018 American midterms. The Census and elections are mentioned in the U.S. Constitution, with authority over the latter largely given to the states. But so are Congress, the presidency, the courts, and the Post Office. I doubt anyone expects Facebook to remove misinformation about all these institutions.

And the slope might be slippery. Facebook faces pressure to adopt an expansive, militant understanding of “democracy” in Europe. It has largely resisted doing so, relying on account authenticity, advertising transparency, and fact-checking to support election integrity. Facebook’s restraint makes sense. An expansive agenda of protecting institutions might lead to suppression of speech that should be seen.

The distinction between facts and opinion offers a second general justification for treating “Census speech” differently. “Misinformation” would contravene clear, not controversial, facts. This justification appeals to many. Consider an election example. Polls in a district are open for a set period that is public and easily confirmed. Spreading the wrong hours for voting is incorrect beyond doubt. In contrast, opinions about who should win and hold office cannot be confirmed by some external authority. Disagreements about that question are why we have elections.

Some questions about the Census are clearly not facts. One might believe that ours is a Constitution of No Authority, that the more efficient exercise of state power is deleterious of freedom, and that, as James C. Scott writes, to be made legible is to be governable. From this perspective, the census is a dangerous tool of control. Whatever you may think of this position, it is coherent, internally consistent and plausible even if ultimately wrong. Articulations of this belief ought not be thrown out with the bathwater of misinformation.

Some propositions look like facts but are really contestable claims (i.e. opinions). Consider the apparently factual claim that “the census will help the government deport illegal immigrants.” The government denies this. Some say that this is misinformation intended to depress census completion rates. Notice this is a “fact” about the future. Perhaps the past might inform our assessment of this “fact.” While census data may not be legally shared with law enforcement, this prohibition was lifted during the Second World War to allow the FBI to more efficiently apprehend Americans of Japanese ancestry. In other words, the past suggests a reasonable person might have doubts about future uses of census data. The government’s future use of census data is more a matter of trust than of hard facts. Facebook cannot use the “clear facts” criterion to side with the government here.  

Political science suggests one important difference between “census speech” and other social media talk. Organized groups and the Census Bureau are concerned about misinformation on Facebook about the Census. Misinformation that matters little to organized groups or the government may be overlooked by social media content moderators. But Facebook should avoid at all costs acting in response to specific demands of organized groups. The more Facebook is seen to cater to specific demands, the more demands will be made of it. And being responsive to interest group campaigns may delegitimate Facebook’s content moderation in general.

The Census Bureau seems to be different. Facebook generally enforces local laws criminalizing speech. The law requires responding to the Census which itself, as mentioned, is constitutionally required. But speech misinforming the public about the census is not illegal in the United States. The Bureau is asking Facebook to help facilitate law enforcement regarding the census. And the law requiring answers to the census is, in fact, a constitutional exercise of legislative power under the necessary and proper clause (John Marshall edition). The same can be said of laws criminalizing sodomy prior to Lawrence v. Texas. Would Facebook have facilitated enforcing Texas’s law prior to Lawrence? Something that might seem as simple as helping the Census Bureau turns out to require Facebook to take sides on controversial political questions.

Facebook might do better by distinguishing between misinformation about the Census that can be dealt with by “more speech” and misinformation that cannot. Most errors about the Census may be corrected by “more speech.” After all, unlike elections, the census does not take place on one day at determinate places. The government and its allies can make their case against misinformation, disinformation, and controversial claims. There is time for “more speech” to bring about the best outcome for the society.

Elections may be different. Let’s say it’s widely and wrongly reported that a candidate has dropped out of a race. That’s a hard fact though perhaps temporarily unclear. It matters a lot whether the report circulates two days or two weeks prior to election day. “More speech” can handle the latter; avoiding harm to the candidate might be much harder two days before the voting. Speech inciting violence may be criminalized because in limited situations there is no time for “more speech” to rebut the appeals to force.

But there’s an important difference between the two harms, violence and false beliefs about elections. I cannot avoid being punched in the nose as a result of incitement. I can avoid false beliefs by modest research regarding facts. Here’s a (hardly obscure) place to start. Our freedom of speech does require that citizens take some responsibility for their beliefs and the reasons for them. Facebook should not protect us from our sloth.

Knight Institute v. Trump

The U.S. Court of Appeals for the Second Circuit has affirmed a district court’s ruling that Donald Trump may not block Americans from viewing or responding to his tweets. In an opinion largely mirroring that of the district court, Judge Barrington D. Parker determined that Trump’s use of his account in his capacity as President of the United States rendered it a limited public forum. As such, Trump’s decision to block the accounts of individual critics amounted to government viewpoint discrimination, rather than a private exercise of associational rights.

The case hinged upon whether or not Trump used his Twitter account in an official capacity. Initially, I expressed skepticism, writing;

The question turns upon our understanding of the purpose and character of Trump’s Twitter presence. Is @realDonaldTrump simply the private Twitter account of a man who happens to be president, or does it constitute a designated public forum, as asserted by Jaffer?

When Knight Institute v. Trump was first filed in June of 2017, @realDonaldTrump looked more like a leftover campaigning tool than an instrument of policy promulgation. Trump had inherited @POTUS, the official presidential account, from Barack Obama, and to the extent that Trump used his personal account to conduct official business, his proclamations seemed untethered from the positions officially taken by his administration.

However, over the past two years, as Judge Parker’s opinion notes in granular detail, @realDonaldTrump has been repeatedly employed to announce the policy positions, and Trump’s tweets are popularly understood to be the official positions of his administration. President Trump has used @realDonaldTrump to announce the withdrawal of U.S. troops from Syria, the selection of Dr. Christopher Waller for the Federal Reserve Board, and, just this past weekend, to reveal that the administration aims to re-include a citizenship question in the 2020 census. Parker also cites the White House’s designation of Trump’s tweets as “official statements,” and other government agencies’ affirmation of the official nature of Trump’s tweets, writing; “we note that the National Archives, the agency of government responsible for maintaining the government’s records, has concluded that the President’s tweets are official records.” Given this evidence, it would seem that @realDonaldTrump is far from a private, personal account, engaging the usual First Amendment prohibitions on content-based restrictions in public fora.

While no one in American politics utilizes Twitter quite so adroitly as Donald Trump, this decision will nevertheless bind other politicians and state officials who make use of social media, particularly those who comingle public and private functions within the same accounts. The decision cites Packingham, but only to establish that speech on social media receives no less First Amendment protection than other speech transmitted via other mediums. While the outcome of Knight Institute v. Trump prevents Donald Trump, and other public officials, from making use of Twitter’s block feature, it has no bearing on  ongoing debates about the limits of private moderation or recent attempts to categorize platform access as a “civil right”.