Endorsement, Coercion, and the Nature of Legal Tests

by Michael C. Dorf

Constitutional law is replete with tests that are not literally found in the text of the constitutional provisions the tests implement. For example, time, place, and manner restrictions on speech on government property that is a public forum are permissible so long as they are content-neutral and leave open adequate alternative opportunities for speech. Race-based classifications by government are impermissible unless they are the least restrictive means of advancing a compelling governmental interest. Police investigative activity constitutes a search within the meaning of the Fourth Amendment if it violates a reasonable expectation of privacy.

None of the italicized phrases in the foregoing paragraph appears in the relevant constitutional text. And while conservative scholars and jurists sometimes complain (selectively) about courts reading into the Constitution concepts that its text does not contain, they routinely apply existing tests and often formulate new ones. Tests that go beyond the literal text of the Constitution are essential to, in the words of Professor Richard Fallon, implement the Constitution.

Why are there so many constitutional tests that go far beyond the literal language of the text? Partly because the U.S. Constitution is short, but also because life is complicated. Given enough time and cases, even a constitution that partook of the prolixity of a legal code would generate numerous tests.

Because tests play such an important role in U.S. constitutional law, the most hotly contested issues often play out as fights over tests. Should race-based affirmative action programs have to meet strict scrutiny or merely intermediate scrutiny? Should infringements on firearms possession be measured by strict scrutiny, intermediate scrutiny, or a direct examination of text and history? Etc.

And yet, tests may be less important than meets the eye, for at least three reasons.

First, as Holmes famously stated in his Lochner dissent, "general propositions do not decide concrete cases." Even though constitutional tests are typically more specialized and thus less general than the constitutional text they implement, they are invariably somewhat general. That, after all, is what it means for a test to be a test--whether formulated as a rule or a standard. Even if the application of some test is straightforward in the context in which it is initially formulated, new cases will arise as to which the test will be indeterminate; or the test will be determinate but it will determine a plainly wrong (because unanticipated) answer, calling for the test's modification. Tests are necessary but insufficient to implement the Constitution.

Second, losers in legal debates typically do not give up. They regroup and fight on another front. Thus, prior to the Supreme Court's 1989 decision in Richmond v. Croson, justices who thought race-based affirmative action generally permissible argued for intermediate scrutiny. After they lost that battle, they argued--successfully--that the kind of strict scrutiny applicable to such affirmative action programs was different from conventional strict scrutiny in that it is not "fatal in fact." Opponents of affirmative action have been pushing back since then and will likely triumph next term in the UNC and Harvard cases. One sees a similar dynamic in the Court's free exercise jurisprudence. Since Employment Division v. Smith, the Court has officially rejected claims for religious exceptions from neutral laws, but increasingly it finds laws that seem religion-neutral are discriminatory.  In these two examples we see that once the Court settles on a test, the debate over what test to apply morphs into a debate over what the applicable test means.

Third, it is sometimes assumed that the choice between two (or more) tests poses a question of whether to adopt a liberal or conservative approach. We talk about strict, narrow, and forgiving tests, for example. That's sometimes true but not always. This week's oral argument in Kennedy v. Bremerton School District was a nice illustration.

Paul Clement, arguing on behalf of the praying football coach, told the Justices that the school district had no legitimate cause to worry that the district would be seen as endorsing religion when Coach Kennedy kneeled and prayed at the fifty-yard line following games, because he was doing so after the games were over and that, given the district's very public distancing, the reasonable observer would conclude that the coach and any players or others who joined him in prayer were acting in their personal capacity, not as official representatives of the school.

Put aside the question whether that's right. Here I want to focus on the pushback that Clement got from various Justices, including Justice Kavanaugh. Even if the coach's prayer didn't amount to official endorsement of religion, they said, there was at least a risk of coercion: student-athletes wanting to start or get more playing time would feel subtly or not-so-subtly pressured to impress the coach by joining him in prayer. Because both the Free Exercise and Establishment Clauses of the First Amendment bar coercing prayer--and because in the school context the court is especially sensitive to psychological coercion--it is possible to conclude that there's a constitutional violation here.

Thus, in this case, the school district will have an easier time establishing that it was justified in terminating the coach to avoid an Establishment Clause violation by arguing that it wanted to avoid coercion than by arguing that it wanted to avoid endorsement. That's interesting because the coercion test first emerged as a proposed substitute for endorsement, with conservative Justices favoring coercion.

And indeed, in some cases the coercion test is more accommodating of religion than is the endorsement test. That's especially true in the context in which the coercion test and the endorsement test initially did battle--challenges to religious displays on public property. The more separationist view there is endorsement: if a reasonable observer infers official endorsement of Christianity from the placement of a creche in front of city hall in mid-December, that's an Establishment Clause violation. By contrast, merely observing a religious display on public property doesn't coerce anybody to participate in or say anything contrary to their conscience.

But the fact that the coercion test is more accommodationist in the religious display context does not mean it's more accommodationist in every context. As the oral argument in Kennedy illustrates, endorsement and coercion are different tests, but whom that difference favors will vary depending on the particulars. 

So, will the school district win? Will conservative Justices who are skeptical of the endorsement test and prefer the coercion test follow their stated principles where they lead? Or will it turn out that the coercion test was only ever a convenient means for rejecting Establishment Clause challenges, so that once it is invoked to strengthen such a challenge, the conservatives will cast the coercion test aside or find some way to say it doesn't mean what it previously was thought to mean? That question will test the Court.