Pedro Rojas holds a sign directing people to an insurance company where they can sign up for the Affordable Care Act, also known as Obamacare, on Feb. 5 in Miami. (Joe Raedle/Getty Images)

The Affordable Care Act hangs in the balance in the Supreme Court for the second time in three years, but the public has rendered a judgment ahead of the court's ruling. By a margin of 55 percent to 38 percent, more people say the court should not take action to block federal subsidies in states that didn't set up their own exchanges, according to a new Washington Post-ABC News poll.

At issue in King v. Burwell is whether low- and moderate-income Americans living in certain states should be allowed to receive tax subsidies to buy health insurance. The challengers say a straightforward reading of the law means the credits are available only for those who buy insurance in marketplaces, or exchanges, that are "established by the state" -- rather than on the federal marketplace. If the court agrees, millions of people in 34 states that didn't set up their own exchanges could lose federal subsidies, which could undermine the entire law.

Public opinion on providing subsidies splits in predictably partisan ways -- but not overwhelmingly so. Nearly two-thirds of Democrats (65 percent) say the court should not take action to block health insurance subsidies. Fifty-five percent of Republicans say the court should rule against the subsidies. Independents side with keeping subsidies, 57 percent to 36 percent.


Support for keeping the subsidies comes despite the law polling as poorly as ever. The survey finds opinion on the health-care law among the worst in Post-ABC polling; 54 percent oppose, up six percentage points from a year ago. Support ties the record low of 39 percent, which was last hit in April 2012.  These results, though, contrast with other recent polls finding softening opposition and support above record low levels.

Although the law has generally been more unpopular than popular, the public has not indicated a willingness to unravel all of the legislation. Polling from the Kaiser Family Foundation has shown that about three in 10 would prefer repealing the entire law, rather than just scaling back, keeping it as is or expanding it. (When polls provide only two options -- repeal the law or keep it -- support for repeal increases.)

The Affordable Care Act is facing another challenge at the Supreme Court in King v. Burwell, which deals with subsidies for health insurance. The case could cut out a major provision of Obamacare, causing the law to unravel. Here's what you need to know about the case. (Julie Percha/The Washington Post)

The disconnect in the new poll -- that a majority opposes the law, while a nearly equal majority does not want the Supreme Court to rule against it -- is driven by political independents as well as Republicans. Independents oppose the law 56 percent to 35 percent. But they also want a ruling in favor of subsidies by almost exactly the same margin. Likewise, although only 19 percent of Republicans support the law, 34 percent say the court should not block subsidies to low- and moderate-income people.

Among independents, the seemingly incongruent stances are explained by opponents of the law who nonetheless favor subsidies. Among this group, 49 percent want the court to block subsidies, while 45 percent say it should not.

The poll also indicates that a Supreme Court ruling to block federal subsidies could prove unpopular in the states in which it matters most -- those without a state exchange. Despite these states being disproportionately under Republican control, people in them say 53 percent to 40 percent that the court should not block subsidies for low- and middle-income people.

The margin is wider -- 58 percent to 33 percent -- in states that have their own exchanges and are unaffected by the law. Again, these are mostly blue states.

Scott Clement contributed to this report. 

The Post-ABC poll was conducted May 28-31 among a random national sample of 1,001 adults, including landline and cellphone-only respondents. Overall results have a margin of sampling error of 3.5 percentage points.