This post was co-authored with Eli Savit, an attorney and adjunct professor at the University of Michigan Law School.

In a New York Times op-ed this week, we argued that a Michigan bill that would impose work requirements on Medicaid beneficiaries violates Title VI of the Civil Rights Act of 1964, which prohibits any program receiving federal funds from discriminating on the basis of race. As currently written, the bill exempts people who live in counties with high unemployment from the work requirement, but doesn’t extend the same accommodation to cities with high unemployment. Because Michigan’s black communities are concentrated in cities, the facially neutral exemption will disproportionately benefit white, rural Michiganders over their black, urban counterparts.

But so what? In its 2001 decision in Alexander v. Sandoval, the Supreme Court held that Title VI doesn’t supply a private right of action to vindicate a disparate impact claim. So if Michigan passes this bill, and the Trump administration declines to enforce Title VI, is there any legal remedy? Suing the Trump administration for declining to take disciplinary action against Michigan won’t work. Under Heckler v. Chaney, HHS’s decision to decline to enforce Title VI is unreviewable.

We nonetheless think the Michigan bill is vulnerable to legal challenge. The key, in our view, is that Michigan can only impose work requirements if it gets a waiver from HHS under section 1115 of the Social Security Act. And the Second, Third, and Ninth circuits have all held, consistent with the presumption favoring judicial review of agency action, that the decision to grant an 1115 waiver is reviewable under the Administrative Procedure Act. The APA thus supplies the cause of action that Title VI doesn’t.

It would be arbitrary within the meaning of the APA for HHS to grant a waiver that licenses Michigan to violate HHS’s own rules governing Title VI. At a minimum, HHS would have to offer a reasonable explanation for why, in the face of the exemption’s disparate impact, there’s a “substantial legitimate justification” for what Michigan wants to do. As we argued yesterday, it’s doubtful that HHS could meet that burden, even under a deferential standard of review. The disparate impact is blatant and a substantial justification for it is nonexistent.

Trying to use the APA as a vehicle to vindicate Title VI isn’t novel. Prior to Sandoval, however, the approach was a non–starter. The APA makes reviewable “final agency action for which there is no other adequate remedy in a court.” Before Sandoval, courts regularly entertained private disparate-impact claims under Title VI, so there was an “adequate remedy” that precluded APA review. As the D.C. Circuit reasoned in an opinion that ended the long-running educational dispute of Adams v. Richardson, “Congress considered private suits to end discrimination not merely adequate but in fact the proper means for individuals to enforce Title VI and its sister antidiscrimination statutes.”

Sandoval changed all that by eliminating a private right of action for disparate impact claims. Absent that “special, alternative remedy,” an APA claim to enforce an agency’s compliance with its Title VI regulations should now be viable.

The post-Sandoval case law on the question is thin, but what little there is reinforces that conclusion. As the Sixth Circuit has observed (albeit outside the Title VI context):

There is a major difference between a plaintiff attempting to obtain a remedy against a person subject to federal regulations, and a plaintiff attempting to hold an agency accountable for alleged violations of its own rules. Sandoval spoke to the former situation—alleged misconduct by a regulated person. But this case involves the latter situation—review of agency action.

In line with that reasoning, a California district court ruled last month in an environmental case that “neither a Title VI nor an equal protection claim constitutes an adequate remedy to an APA claim” when it comes to disparate impact. Indeed, the U.S. Justice Department has itself raised that possibility of APA suits to vindicate Title VI, and Justice Breyer made a related point in a concurrence in a case about the enforcement of the Medicaid statute.

In our view, the APA theory is sound. We’re less sure that the Supreme Court still believes, as it once did, that Title VI authorizes agencies to issue rules that prohibit policies that have a disparate racial impact. If the Court were to rethink that position, an APA challenge to the waiver couldn’t succeed. But who knows? Lower courts can’t indulge in armchair predictions about what the Supreme Court might do. They’re bound unless and until the Supreme Court reverses itself. And, under existing case law, the Michigan bill looks really dubious.

One final note. Because of persistent income disparities across race, minority groups are overrepresented in the Medicaid population. As a result, any waiver that allows states to constrain Medicaid eligibility will have a disparate racial impact—and, as such, is vulnerable to the argument that HHS granted the waiver in contravention of its own Title VI regulations.

The argument hasn’t been pressed in the high-profile lawsuit challenging Kentucky’s work requirements, but maybe it should be. Sure, it’s possible that HHS can offer a “substantial legitimate justification” for work requirements—at least one that can survive APA arbitrariness review—notwithstanding their racial disparate impact. But given the fragility of the justifications for work requirements, an APA-cum-Title VI claim has a decent shot at succeeding—at least as good a shot, we think, as the arguments currently being raised.

@nicholas_bagley & @EliNSavit