Late Friday evening, Judge Alex Kozinski of the U.S. Court of Appeals for the Ninth Circuit filed an opinion dissenting from the denial of en banc review in Washington v. Trump, joined by four of his colleagues. Judge Kozinski’s opinion (or as he would say, his dissental) is noteworthy because it marks the first time that an appellate judge (and a stellar one at that) has signaled disagreement with the emerging judicial consensus that President Trump’s revised travel ban violates the First Amendment’s Establishment Clause on the ground that it was intended to disfavor Muslims.
Importantly, Judge Kozinski does not actually say that he thinks the ban was motivated by something other than religious animus. Nor does he go so far as to say that he would uphold the ban (though he certainly leaves that impression). Instead, he makes two points regarding the ban’s constitutionality under the Establishment Clause, which together provide a roadmap for how a court might uphold the ban in the future.
The first point, found in a footnote, is his belief that the typical test for determining whether a law violates the Establishment Clause—a test that “requires a showing of secular purpose”—has no application to “laws affecting immigration.” But he does not cite anything to support this assertion, nor offer up what he thinks the test should be (or even whether, in his view, the Establishment Clause has any application in the immigration context). It is one thing to defer to the president’s judgment on immigration and national-security matters; it is another thing entirely to permit religious discrimination in the name of national security, simply because the policy affects immigration. Imagine a law that said, on its face, that only Christians may be admitted to the United States. Would this be constitutional in Judge Kozinski’s view? If so, why? If not, why not? Or suppose that Trump had followed through on his campaign promise of a “total and complete shutdown of Muslims entering the United States.” Constitutional?
Although he doesn’t say, my guess is that Judge Kozinski would find that these laws violate the Establishment Clause because they discriminate on their face and cannot withstand scrutiny. But if you concede that the Establishment Clause has some application in the immigration sphere, then why should it forbid only express discrimination? After all, the Supreme Court has repeatedly held that the First Amendment’s religious clauses “extend beyond facial discrimination.” As the Court explained in a 1993 opinion by Justice Anthony Kennedy: “Official action that targets religious conduct for distinctive treatment cannot be shielded by mere compliance with the requirement of facial neutrality.” The logic of that principle holds true beyond the purely domestic context. If the Establishment Clause applies to immigration policy, then it should forbid more than just facial discrimination, or else it would be too easily circumvented.
Which leads to Judge Kozinski’s second point: his view that, in determining whether an executive policy was motivated by discriminatory intent, the president’s campaign statements should not be considered. This is Judge Kozinski’s real objection to the reasoning of the courts that have blocked the ban’s enforcement, all of which considered Trump’s statements as a presidential candidate to be relevant to the inquiry. That includes the Ninth Circuit panel in Washington v. Trump, which halted enforcement of Trump’s first ban on due-process grounds. Although the court did not rest its opinion on the Establishment Clause claim, it indicated that courts are permitted to consider all of Trump’s “numerous statements” about Muslims—including those made before he took office—in conducting the analysis. And just this past week, two courts did exactly that, holding that the ban was likely motivated by religious animus, based in part on Trump’s statements as a candidate.
Before addressing Judge Kozinksi’s specific objection, it’s worth stressing the stakes. If Trump’s campaign statements are fair game, then the case for the ban’s unconstitutionality is both strong and straightforward: In December 2015, then-candidate Trump called for a “total and complete shutdown of Muslims” entering the country (a proposal that Mike Pence said at the time was “offensive and unconstitutional”). Seven months later, when asked whether his policy had changed, Trump said: “No, I—call it whatever you want. We’ll call it territories, OK?” Six months after that, barely a week into his term as president, Trump signed an executive order covering seven countries—all of which have overwhelming Muslim majorities. Shortly thereafter, former New York City mayor Rudolph Giuliani publicly explained that Trump wanted to do as he had said—impose a “Muslim ban”—but he “wanted to do it legally,” by focusing on “the areas of the world that create danger for us.” Then the revised ban, issued six weeks later, sought to achieve (in the words of senior Trump adviser Stephen Miller) “the same basic policy outcome,” while narrowing its scope. And when the courts blocked enforcement of that effort as well, President Trump drew a direct link between the two versions, telling a crowd: “This is a watered down version of the first one,” because “the lawyers all said, oh, let’s tailor it.” He declared that “we ought to go back to the first one and go all the way.”
As this history shows, Trump first sought to ban all Muslims from entering the U.S., and then worked backward to gerrymander a category that would cover the largest possible number of Muslims (and the smallest number of people of other faiths), while keeping the law religiously neutral on its face. That’s unconstitutional. If Trump’s campaign statements are off limits, however, then the story does not begin until after he is elected, and the case for discrimination is harder (but still provable). In other words: a lot may ride on this.
Why does Judge Kozinski think that Trump’s statements as a candidate can’t be considered? Because they’re unreliable, he says, and it would be unworkable to consider campaign statements as a basis for finding religious animus. But neither is a reason to exclude these statements altogether. As to reliability: Yes, “[c]andidates say many things on the campaign trail,” and yes, “they are often contradictory or inflammatory.” That is especially true for Trump. But that’s a reason to give less weight to campaign statements—not to categorically exclude them from the inquiry.
To see why, take an example David Cole gives in his brilliant piece in the New York Review of Books: “Imagine if the mayor of a town promised repeatedly during a campaign for office that he would keep African Americans out of the town, and then, upon election, adopted a policy barring entry from six cities with populations that were 90 percent African-American. Would anyone doubt that the policy discriminated on the basis of race?” The question answers itself. When candidates for an executive office say that they will do something to target a minority group—particularly something central to their campaign—and then they enact a policy one week into their tenure that carries through on their promise, their campaign statements are relevant to their intent in crafting that policy.
If anything, a recently elected president’s campaign statements are more relevant to his intent than the statements of individual legislators are to a law enacted by a collective legislative body. And yet courts, including the Supreme Court, have not hesitated to consider the subjective motivation of individual lawmakers in determining whether a law intended to disfavor a particular religious minority. What sense does it make for a court to probe the statements of individual lawmakers for evidence of what the legislature as a whole thought, but then shut its eyes to what the president himself said about the very policy in question simply because he said it as a candidate? As the Supreme Court explained in 2005, “the world is not made brand new every morning,” and there is no reason “to ignore perfectly probative evidence” or “to turn a blind eye to the context in which [the] policy arose.” Just so here. Judge Kozinski makes strong arguments for why a court should generally discount a president’s campaign statements. But he offers no compelling reason why a court should ignore them entirely.
Judge Kozinksi also objects to the use of campaign statements as unworkable. He contends that no one can know how much weight to give each particular statement. But courts already have to consider all sorts of informal “circumstantial evidence” to uncover a law’s true purpose, including “the historical background of the decision under challenge, the specific series of events leading to the enactment or official policy in question, as well as the legislative or administrative history, including contemporaneous statements made by members of the decisionmaking body.”
Does this totality-of-the-circumstances test suddenly become unworkable if courts are also permitted to consider highly relevant campaign statements? Judge Kozinski thinks that it does, because that would improperly expand the universe of evidence that courts may review. In the real world, however, campaign statements will rarely make the difference. And when they do, it will be because they reveal a clear desire to discriminate overtly, as in this “highly unique case”—where the President himself made a “Muslim ban” a central plank of his campaign, telegraphed that he would achieve his ban by focusing on “territories,” and then a week into office put a ban in place that did just that (which both Trump and his advisers then repeatedly connected back to his campaign promises). Examining this history—and calling it what it is—does not require any “judicial psychoanalysis” of President Trump’s “heart of hearts,” as the Supreme Court has cautioned against. It requires only that courts look to Trump’s explicit, “openly available” statements that support the “commonsense conclusion” that this ban was religiously motivated.
If ever a candidate’s statements can be used to help establish discriminatory intent, it is here. The Establishment Clause seeks to prevent the government from “mak[ing] a divisive announcement that in itself amounts to taking religious sides,” because this sends the signal to members of the disfavored religion that they’re not full participants in our democracy. President Trump made that announcement as a candidate—loudly and repeatedly—and then he enshrined it into official governmental policy almost immediately upon taking office. Courts can, and should, take this history into account in assessing whether his ban comports with the Constitution’s command of religious neutrality.