The Supreme Court upheld Obamacare’s insurance subsidies on June 25, 2015, a huge victory for President Obama’s signature policy accomplishment.

King v. Burwell was arguably the Affordable Care Act’s greatest existential threat since the Supreme Court upheld the individual mandate in 2012. The lawsuit, had it succeeded, would have ripped the subsidies out of 34 of the law’s state insurance exchanges — effectively destroying much of Obamacare in those states.

Without subsidies, private insurance would have become unaffordable for many of the 6.4 million Americans currently using federal subsidies to help pay for their coverage. Healthy people would likely have dropped their coverage, and only the people who were very sick — and therefore very expensive to insure — would keep their plans.

This would set up the classic insurance “death spiral.” By putting coverage out of financial reach for so many people, it would have undermined the entire purpose of the Affordable Care Act.

If the Supreme Court ruled against Obamacare, most observers say there would have been no quick fix. The White House wouldn’t have the legal authority to dole out the insurance subsidies. Congress would need to pass new legislation to allow the financial help to start flowing again, but it’s unlikely the president and Republicans would settle on an Obamacare plan they both liked.

States could have decided to build their own marketplaces, but doing so for the next open enrollment period (which begins in November) would be a logistical challenge. And, politically, many Republican governors — particularly those who oppose Medicaid expansion — would be unlikely to help implement a major Obamacare program.

With no clear solution in sight, a ruling against Obamacare in King v. Burwell would have meant a ruling where millions of Americans lost health insurance coverage.

An estimated 6.4 million Americans currently receive insurance subsidies through Healthcare.gov — and these are the people who had the most at risk in the King decision.

The states with the most to lose in the King v. Burwell decision were those with the most Obamacare enrollees. And these states tend to have two things in common: They have a population that is bigger and lower-income than the average state.

Florida, Texas, and North Carolina are only three of the 36 states using the federal marketplace. But taken together, their residents account for more than one-third of Americans getting tax credits through Healthcare.gov. About 2.6 million of the 6.4 million people who get financial help through the marketplace live in one of those three states. And this means a decision against Obamacare would have been acutely felt in specific pockets of the United States, not evenly across the country.

There are currently 36 states that use Healthcare.gov, and there, all residents’ Obamacare subsidies were at risk. But in a ruling against the health-care law, it’s possible that some states might have built their own marketplaces — allowing financial help to keep flowing.

One good proxy for understanding a state government’s openness to expanding an Obamacare program is whether the state has opted into the health law’s Medicaid expansion. This map might be the most plausible post-King landscape.

It shows the states that have been open to participating in Obamacare by building an exchange or expanding Medicaid (or having a plan to expand Medicaid) — and the states that have outright refused.

There are about 5 million Healthcare.gov enrollees who live in states that have refused to participate in the law’s Medicaid expansion or build their own exchange. The vast majority of them get financial help from the government.

If the Supreme Court had ruled against Obamacare, these people would still have paid taxes that help finance the insurance subsidies that people in other states continue to receive. And their tax dollars are already going to pay for Medicaid expansions in other states, even though their own government had decided against participating in that program.

“I noticed something peculiar about the tax credits. There will be no tax credits for tax payers who live in non-capitulating states,” Christina told the audience there.

Following that, other legal scholars began to examine this as a potential issue. Among them was Jonathan Adler, a law professor at Case Western Reserve University. He featured the quirk in a lecture he gave at the University of Kansas in March 2011. In August of that year, Adler mentioned the issue to Michael Cannon, director of health policy at the Cato Institute, a libertarian think tank.

Shortly after that, the IRS issued regulations clarifying that subsidies will be available on all exchanges, whether established by the state or by the federal government. These regulations were the specific target of legal challenges; in the eyes of the challengers, they contradicted the plain text of the Affordable Care Act.

Cannon reached out to reporters hoping to generate some attention around the issue, though no suits had yet been filed. David Hogberg at Investor’s Business Daily was the first journalist to pick up the story.

Cannon and Adler co-authored a November 2011 editorial in the Wall Street Journal, characterizing the problem as a “glitch” in the way the law was written.

The two started writing an extensive paper on the problem, posting a draft online (at SSRN) in July 2012. A final version of the paper was published in HealthMatrix in March 2013. In the course of their research, Cannon and Adler’s position on the issue evolved: rather than seeing it as a drafting error, they said they believed there was strong evidence that this was what Congress intended all along.

The Affordable Care Act provided for the creation of different types of insurance exchanges. Fourteen states and DC established “state-based” exchanges, which give them more flexibility and authority in controlling their Obamacare markets.

In the event that a state chose not to establish its own exchange, the Affordable Care Act dictates that the federal government would step in and create a “federally facilitated” exchange. There’s also a middle ground for partnership exchanges, where states and feds share authority. Healthcare.gov is the face of both federally facilitated and partnership exchanges.

These different types of exchanges were set up by different parts of the health law. But the part of the Affordable Care Act that authorizes the subsidies specifies that those subsidies are available to people who enrolled “through an Exchange established by the State under 1311” — the section that sets up state-based exchanges.

“Congress could not have chosen clearer language to express its intent to limit subsidies to state Exchanges, and no one has been able to explain why it would have used this language absent such intent,” the King challengers argued in their Supreme Court brief.

Why would Congress do this? In addition to making insurance more affordable for millions of Americans, the plaintiffs argued, Congress wanted them to act as a carrot, encouraging states to set up their own exchanges. If a state set up its own exchange, its citizens would receive subsidies. If the state didn’t, its citizens wouldn’t.

“Using subsides as an incentive for states to create Exchanges,” the challengers’ brief argued, “is perfectly sensible and consistent with Congress’s purposes.”

The government’s defense hinged on two key pieces of precedent:

Both of these legal doctrines, the White House argued, set the bar low. The Obama administration didn’t need to show that their preferred interpretation of the disputed sentence was the most natural rendering of the English-language phrase in question. They merely needed to show that it was one possible reasonable interpretation of the sentence in light of the overall purpose of the law. So the White House argued that the overall purpose of the law is, pretty clearly, to give lots of people subsidized health insurance and that the IRS interpretation is a reasonable effort to fulfill that purpose.

From the government’s viewpoint, the plaintiffs were positive there was a massive — and extremely strange — conspiracy in which a Democratic Congress and a Democratic administration designed a scheme whose purpose was to punish recalcitrant states. Then, once the bill was signed into law, its architects and supporters all changed their minds so quietly and so decisively that when the IRS promulgated an interpretation that was contrary to their initial purpose the only person who raised any objections was a longtime anti-Obamacare fanatic working at a libertarian think tank.

The government’s more straightforward interpretation of events was that things proceeded exactly as they seem to have proceeded — one legislative provision was drafted unclearly, the IRS used its discretion to bring practice in line with the overall intention of the bill, and conservative litigators objected because they simply disagreed with the idea of using taxes, spending, and regulation to provide millions of people with subsidized health insurance.

Those who reported on and participated in the Obamacare drafting process have universally said they meant for all states to have insurance subsidies.

“It was always intended that the federal fallback exchange would do everything that the statute told the states to do, which includes delivering the subsidies,” Chris Condeluci, who worked as tax and benefits counsel for the Senate Finance Committee Republicans during the Affordable Care Act debate, told Vox.

Former Sen. Olympia Snow, a moderate Republican who helped write the Finance Committee’s initial version of the bill, recently told the New York Times, “I don’t ever recall any distinction between federal and state exchanges in terms of the availability of subsidies.”

“Not once in the 16 months I reported on the formal congressional debate did any of the law’s architects suggest they were thinking along these lines,” wrote Jonathan Cohn, a senior editor at the New Republic who closely tracked the health reform debate for over a year before the Affordable Care Act was passed. “It wouldn’t make sense in the context of the law, which depends upon those subsidies to accomplish its primary goal. It’s why assessments of Obamacare’s impact, including those from the Congressional Budget Office, assumed that residents of all states would have access to the tax credits.”

Vox’s Sarah Kliff agrees:

After covering the debate over health reform since it began in 2009, I feel completely comfortable saying Congress meant for residents of all 50 states to have access to financial help. It was never a question, during the five years I’ve spent writing about Obamacare, whether this would be case ...When I worked at Politico in 2010 and 2011, my entire beat consisted of covering state Obamacare implementation. Lots of states began debating whether or not to set up health insurance exchanges; the issue of whether they might lose their subsidies never came up.

The challengers countered these testimonials — and the assertions of members of Congress, John McDonough, Peter Orszag, CBO analysts, and others who were intimately involved in drafting the law — with comments from MIT economist Jonathan Gruber, the chief economist behind the Affordable Care Act. In two videos unearthed in the week following the Halbig and King circuit court rulings, Gruber suggested that subsidies would only be available on state-based marketplaces.

Gruber responded to the videos in an interview with Cohn; he said that he misspoke and may have been considering the possibility that federal exchanges wouldn’t be scaled up in time, which would leave states that didn’t establish their own exchanges without any exchange at all.

Compare all of this to the Medicaid expansion: As the Affordable Care Act was originally written, states needed to expand the program or forfeit all federal Medicaid funding, new and old. This incensed conservative states. It incensed them so much that they brought a lawsuit against the federal government, claiming that expansion requirements were coercive — a lawsuit they won.

None of the states made noises suggesting that they were being similarly pressured to scale up state exchanges. “There was not a breath during the legislative debate suggesting that Congress meant to deprive citizens in states with federally facilitated exchanges of tax credits,” said Nicholas Bagley, a law professor at the University of Michigan.

This case keyed in on what appeared to many outside observers to be poorly crafted legislative language. And there’s a reason for that: The final version of the Affordable Care Act was a messy marriage of two Senate bills that passed in a hasty, unusual process.

Typically, the way a law like this would pass is that the House would pass one version, the Senate another, and then the two bodies would meet in a “conference committee” to clean up the language, iron out their disagreements, and generally edit the bill. Then the House and Senate would both pass the final draft — known as “the conference report” — into law.

And that was the plan for Obamacare, too. But when Senator Ted Kennedy died in 2009 and was replaced by Republican Senator Scott Brown, Democrats no longer had a filibuster-proof majority in the Senate. That meant they couldn’t break the inevitable Republican filibuster of the conference report.

Yet even though this is what many people who followed the legislation think happened, neither side asserted that there was a “mistake” in the way the law was written.

The government was hesitant to argue that this lawsuit was premised on an error in how the law was drafted, because it was possible that the court would hold the government to the text of the law — even if that text was flawed and contrary to Congress’s original intent — so framing it as an “error” could have resulted in losing the suit.

Two legal scholars widely acknowledged to be the architects of the lawsuit initially described this as a “glitch” in the law’s text. They have since revised their opinion, asserting that subsequent research uncovered legislative history that provides evidence that this was what Congress planned to do.

On July 22, 2014, the US Court of Appeals for the DC Circuit and the US Court of Appeals for the Fourth Circuit — both courts one step below the Supreme Court — handed down conflicting rulings. The DC Circuit struck down subsidies in Halbig v. Burwell. The Fourth Circuit upheld those same subsidies in King v. Burwell — and the Supreme Court affirmed their ruling.

“A provision thus may seem odd without being absurd, and in such instances it is up to Congress rather than the courts to fix it, even if it may have been an unintentional drafting gap,” Judge Thomas Griffith wrote in the majority opinion.

The court pointed to two other examples of potentially “absurd” outcomes from the text of the law, including language that would have locked Guam and other US territories into an unsustainable health insurance system.

The major departure between the DC Circuit and the Fourth Circuit is characterized by this line from the majority opinion: “The fact is that the legislative record provides little indication one way or the other of congressional intent, but the statutory text does.”

The DC Circuit agreed on September 4 to hear the case “en banc,” which means the court’s full panel of 13 judges will issue a ruling. Eight of the judges on the court are Democratic appointees. The en banc rehearing could reverse the initial decision in Halbig, which was issued by a three-judge panel, two of whom had been appointed by Republicans.

The Fourth Circuit judges were not wholly persuaded by the argument from either side. Both the plaintiffs and the government advanced reasonable interpretations of the law, but the court said it could not confidently conclude which was the correct interpretation.

Judge Roger Gregory wrote in the majority opinion:

Although the plaintiffs offer no compelling support in the legislative record for their argument, it is at least plausible that Congress would have wanted to ensure state involvement in the creation and operation of the Exchanges. Although Congress included a fallback provision in the event the states failed to act, it is not clear from the legislative record how large a role Congress expected the federal Exchanges to play in administering the Act. We are thus of the opinion that nothing in the legislative history of the Act provides compelling support for either side’s position.

There are three actors who arguably could have restored subsidies in Obamacare: the White House, Congress, and the states. All available evidence suggests none of them would have moved quickly, had the Supreme Court struck down the law.

The Obama administration says that subsidies would have had to stop flowing had the Supreme Court found them to be illegal.

What the Obama administration was essentially saying here is that if they had lost in court, their hands would have been tied. There were two other possible remedies: Congress could have passed new legislation to fix the law, or states could have established their own exchanges. But the executive branch says it could not have continued doling out subsidies barring those scenarios playing out.

Republicans lawmakers also came up with five alternatives plans to keep the dollars flowing. The question is whether they would have done much good. Most of the plans would have extended the availability of subsidies while dismantling other parts of Obamacare. The result would likely have been a world that looked much more like America before Obamacare — where fewer people were enrolled in coverage and were paying higher premiums.

Without a requirement to purchase insurance coverage, health economists roundly expected that young, healthy people would no longer buy coverage. That, then, would lead to a spike in premiums as only the really sick people, who use their coverage a lot, opted to buy insurance plans.

The transitional period Johnson’s bill imagined is one where the individual market is smaller and a more expensive place to shop.

These types of problems turned up again and again in all five Republican plans. When you try to repeal Obamacare and maintain the law’s subsidies, it turns out you end up with some very bizarre policy outcomes that are not good for the individual insurance market.

This is very much a work in progress. It will continue to be updated as events unfold, new research gets published, and fresh questions emerge.

So if you have additional questions or comments or quibbles or complaints, send a note to Sarah Kliff: sarah@vox.com.