Last week, I wrote that “there would be no legal basis for entering a stay” if the Supreme Court inKing v. Burwell invalidates the IRS’s rule governing the distribution of tax credits. I received some thoughtful pushback on that point, most notably from Will Baude and Andy Grewal. In light of their comments, I want to offer a qualification, although my bottom line stands: it’s very unlikely that any stay will issue from the Court.

As both Will and Andy pointed out, the Supreme Court has, on rare occasions, muffled the impact of decisions that might have proven socially disruptive. In Northern Pipeline, for example, the Court held that the Constitution prohibited Congress from assigning to bankruptcy judges the power to resolve certain controversies. Similarly, in Buckley v. Valeo, the Court held that members of the Federal Election Commission had been appointed unconstitutionally.

Yet the Court declined to invalidate any past decisions of the bankruptcy judges and the FEC. Instead, the Court’s judgments would only apply going forward—in the lingo, Northern Pipeline and Buckleywould only have prospective effect. In addition, the Court entered brief stays to enable Congress to fix the constitutional problems.

In one respect, King doesn’t implicate reliance interests as serious as those in Northern Pipeline andBuckley. Even if the government loses, King wouldn’t call into question the validity of tax credit payments made before June 2015, when the Supreme Court is likely to decide King. Under 26 U.S.C. §7805(b)(8), the IRS has the power to “prescribe the extent, if any, to which any ruling (including any judicial decision …) relating to the internal revenue laws shall be applied without retroactive effect.” In plain English, that means the IRS need not claw back tax credits that were paid out before the Court’s decision.

But what happens after June? Unless the Supreme Court stays its decision, the IRS would have to abruptly stop issuing advance tax credits. The effects would be felt almost immediately: 4.5 million people would see their insurance rates surge; many of them would drop coverage; and insurers’ risk pools would skew toward the unhealthy, leading to enormous losses. In a word, there would be chaos on Healthcare.gov.

Could the Supreme Court draw on Northern Pipeline and Buckley to enter a stay? Sure, in principle. But don’t count on it. Decided more than thirty years ago, the cases display more comfort with the Court’s power to tailor its remedies than do cases decided more recently. In the 1993 case of Harper v. Virginia Department of Taxation, for example, the Court held that it would no longer countenance giving only prospective effect to its judgments. Doing so, the Court suggested, was “incompatible with the judicial role.”

Entering a stay in King might similarly smack of legislating from the bench. That’s especially so given that King is quite different from Northern Pipeline and Buckley. There, federal officers whose appointments were technically invalid were temporarily allowed to continue doing their work. Here, in contrast, the Court would have to allow federal officials to withdraw money from the U.S. Treasury, even without statutory permission to do so. As I explained in my last post, enabling such withdrawals raises especially serious constitutional concerns.

All that said, I can see a narrow path to getting a stay. If the Court were to split 5-4 along ideological lines, the four dissenting justices would probably be open to staying the judgment. The dissenters would have to peel off just one conservative vote—presumably Roberts or Kennedy—in order to get a stay. Now, it doesn’t strike me as especially likely that any of the conservative justices would be inclined both to invalidate the IRS rule and to keep it in place for a while. But it’s possible. Given that slim possibility, maybe the government should at least ask for a stay. Temporary relief from the Court’s judgment would give the states a narrow chance to consider afresh whether to establish their own exchanges now that they understood the consequences of failing to do so.

In the meantime, it would be foolish to bank on a stay. The states have seven months to develop contingency plans before the Supreme Court rules in King. If they squander the time, they’ll lose their best chance to avoid massive disruption for millions of people.