What about those subsidies again ... ? Photographer: Andrew Harrer/Bloomberg
What about those subsidies again ... ? Photographer: Andrew Harrer/Bloomberg

Last week I wrote about the Halbig case, in which a federal appeals court ruled that the Affordable Care Act only allowed subsidies for the purchase of insurance to be given out on state exchanges. Because most states have opted out of building exchanges and others are considering switching to the federal exchange, this would obviously be a big deal for the law’s implementation.

Since the decision was handed down last Tuesday, liberal and conservative commentators have joined the battle over whether this reading of the law was obvious or outrageously absurd. A lot of liberal journalists who had covered the law said that if Congress had ever intended any such thing, they would have heard about it and reported on it, because that would have been big news.

It became harder to insist that this was a clear misreading when extensive audio and video surfaced of Jonathan Gruber, one of the architects of the law, telling people in 2012 that tax credits would only be available on state exchanges. This week, another group found an NPR interview with Jonathan Cohn, the New Republic’s star health-care reporter, who also seems to have thought that states could opt out of the exchanges but wouldn’t, because it would cost their citizens a lot of money:

Basically, where a state could opt out of the exchanges, I find it hard to believe a state would actually do that. You know, it's -- if you think about the history of these sorts of things, Medicaid was set up and is, remains, an optional program for states. States can opt out of Medicaid if they want to.

They don't because there's a lot of federal money they are entitled to if they participate in Medicaid. In addition to that, it helps them cover poor people in their state, and if they don't get that money from Medicaid, they're going to be totally responsible for this on their own.

I can't possibly imagine a state opting out of an insurance exchange, given it's a good deal for the state. And I know a lot of states are nervous about what's going to happen with this, but at the end of the day, I just don't see it happening.

To his credit, Cohn -- who is a top-notch reporter and a careful, thoughtful analyst, for all that he inexplicably disagrees with me about the wisdom of the health-care law -- wrote this up himself.

I wish I could tell you definitively what I was thinking at the time but, more than four years later, I truly don't know.

My best guess? I certainly knew the House and Senate bills contemplated different exchange models -- and that liberals preferred the House version, mainly because they were skeptical that state officials could or would adequately police insurers. But I hadn't investigated further, as I said on the air, most likely because I was busy writing on other topics getting more attention. Once the issue came up in the House-Senate negotiations, I probably started asking more questions, became satisfied that the exchange proposals were actually similar, and moved on to other things. Note that the "Fresh Air" interview took place just three days before the e-mail exchange I wrote about earlier this week -- a communication with Debbie Curtis, a key House aide who was part of those negotiations, who told me that the differences between the House and Senate versions of the exchanges were mostly a "red herring."

Again, I can't be sure about all of this and readers will need to make up their own minds. I'd be lying if I said that seeing and hearing my own words didn't give me pause. Still, I didn't write the law. I wrote about the law. The real-time information I had in early 2010, while negotiations were taking place behind closed doors, was necessarily incomplete and tentative -- a point I tried to get across even then.

There’s been a fair amount of schadenfreude over this, and fair enough; over the five years that I’ve been covering health-care reform, both sides have enjoyed crowing over the other -- and eaten some crow, too. But that’s not what this post is about. This post is about memory, good faith, and why it’s harder than you think to know what happened.

When I wrote about Gruber last week, I took pains to stress that I thought he was telling the truth, as he saw it -- that he genuinely did not recall having believed that subsidies would only be available on state exchanges. A number of my readers accused me of being hopelessly naive, at best, and at worst of bending over backward to cover up for one of my liberal buddies in academia.

The latter is ridiculous: I opposed Obamacare four years ago, still think it was a bad idea, and have never met Jonathan Gruber or, to my recollection, even spoken to him. It is true that I bend over backward to give people the benefit of the doubt, which not only steers me clear of libel law, but also makes for a much more productive debate. “You are a lying liar!” is not a proposition that you can reasonably expect your opponent to debate. And frankly, it’s often just a way to avoid dealing with arguments you don’t want to engage in: Witness the number of people who accused conservatives of a bad-faith misreading of the law here.

But in this case, I am not stretching to my utmost to give Messrs. Gruber and Cohn the benefit of the doubt. I said that I thought Gruber was being sincere because, well, I genuinely think he’s being sincere. I believe this even more strongly of Cohn, who has always been fair and honorable in his analysis and our few personal interactions.

At least some of the liberal analysts who were intimately familiar with the details of the law did indeed believe that premium subsidies would only be available on state exchanges. But I disagree that Cohn and Gruber were being dishonest when they said otherwise. I think they just forgot.

“How could they forget!” you ask. To see how, let me take you on a tour of another, fairly recent controversy: the accusations that Mitt Romney had bullied a young classmate in his high school, setting upon him with a group of friends and cutting off his longish blond hair. That student turned out to be gay, though it’s not clear whether Romney would have known that at the time.

When this news broke, Romney said that he didn’t remember the incident, but he made a general apology for any “hijinks” he pulled in school. The Internet immediately divided, almost perfectly, into two camps: liberals who insisted that of course it must have happened and he couldn’t possibly have forgotten, and conservatives who questioned whether it had happened at all and, if it did, whether it had happened in the way described, because most of the people making the accusation were Democrats. There was a strong implication that they had been lying for partisan purposes.

My opinion on this was my opinion on the Swift Boat veterans who attacked John Kerry, and people claiming to remember details of George Bush’s Air National Guard service in the early 1970s, and anyone else who comes forward with a story 40 years later about someone who was not famous at the time: These sorts of stories are simply too unreliable to be treated as fact -- especially if the results of those memories are politically important.

We tend to think of memory as a video camera, capturing events and faithfully replaying them for us. All the research on this shows that it’s nothing of the sort. Memory is a narrative you tell yourself. Every time you recall a memory, you strengthen it -- but you can also change it. It is frighteningly easy to create false memories.

Have you ever been searching for something important you lost, and suddenly thought you remembered where you put it? You can remember yourself doing it. Then you go look, and it isn’t there.

This effect can get even more powerful when there’s another person there, telling you what to remember. Here’s one of the world’s leading experts on false memories, showing how it happened to her:

Not even Loftus was immune to suggestion. In 1988, after 13 years of testifying about memory's fallibility, she was told by her uncle that she was the one who had found her dead mother in the swimming pool. The sights and sounds of that awful morning came back to her -- the corpse face down, the nightgown, the screaming, the stretcher, the police cars. But within three days, her uncle recanted the story, and other relatives confirmed that her aunt, not Loftus, had found the body. The memories of the memory expert were false.

The incident strengthened Loftus' conviction that such recollections shouldn't be trusted in court. The more cases she saw, the more passionate she became about her work. She saw herself as Oskar Schindler, rescuing as many innocent souls as she could. "The beauty I find in helping the falsely accused is something I like about myself," she wrote in an essay years later. "It's the deeper part of who I am."

Anyone who questioned how Romney could fail to remember having bullied a crying kid should take note: This woman misremembered her mother’s death, one of the worst, and most important, things that could ever happen to anyone.

You should also think hard on this when you assess the stories the other people were telling: A few days after the Washington Post broke the Romney story, it came out that they’d been talking to each other before they talked to the Post. At this point, it seems to me that the story was simply hopelessly contaminated. The Post story made much of the fact that they all agreed about the memory. Yet all that tells you is that they’d agreed upon a version that was internally consistent. Note, however, that I’m not suggesting that they’re lying. I’m suggesting that once they’d spent a lot of time talking about an incident that one of them recalled from 40 years ago, the line between what they actually remembered and what they’d heard other people say was too hopelessly fuzzy to treat it as fact.

And this is true of anyone who is giving you a detailed account of something that happened a while ago. People don’t remember things that happened a while ago; they remember the stories that they have told themselves about it. I vividly remember sticking a key into an electric socket when I was (I am told) about 18 months old. Do I remember it? Or do I remember being told about it? It feels like a real memory, but all the research indicates that that tells you precisely nothing.

Conversely, I am told that when I was a teenager, a horse reared up and pawed around my ears, miraculously not kicking me in the head. Multiple people agreed that this had happened, but I have no memory of it. Either I’ve forgotten something I certainly ought to remember -- or they are misremembering something that happened to someone else. No way to tell, because we don’t have contemporaneous documentation.

This is particularly true when there is an answer people very much want to get to. It’s called hindsight bias: Once we know the “correct” answer, we tend to believe that we would have figured that out even without being told. In fact, if we’re asked to predict the answer in advance, we will often edit our memories of what we did believe, to show that “we knew it all along.” This is not a conscious attempt to deceive someone else; it is part of the mind’s unconscious mechanisms for deceiving itself. Elizabeth Loftus theorizes that this has self-protective functions, helping to tamp down distressing memories and boosting our self-esteem. But of course this mechanism probably gets a powerful boost when you add in the sort of motivated cognition that we see around political topics.

Misremembering is not lying. Are you more likely to misremember things in politically salient ways? Of course, because when your memory suggests the “wrong” thing, you’re quite likely to go hunting for evidence that you’re wrong … while a memory that bolsters your preferred narrative is not one you’re going to ferociously interrogate. This is just human nature. It doesn’t mean that Cohn and Gruber are being dishonest; it just means they’re human. And it applies just as strongly to conservatives who claim to have some strong memory of what went down in 2010.

Of course, that leaves us with a bit of a problem. We can’t rely on memory to tell us what the intentions were, because as this episode has demonstrated, memories are extraordinarily unreliable. And the written record is extremely thin, because -- as best I can tell -- it simply never seriously occurred to any of us that states would fail to set up exchanges.

That does not mean, as some of the law’s supporters have suggested, that we must therefore defer to the Internal Revenue Service's interpretation allowing subsidies to be offered on federal exchanges. The logic of this seems to be that Congress meant to pass a law that worked; this will make the law not work; therefore, this cannot have been congressional intent. As a friend points out, by this logic, Congress could have just written “ALL WORK AND NO PLAY MAKES JACK A DULL BOY” for 2,000 pages, and whatever the regulatory agencies did would be fine, because hey, Congress wanted the law to work.

But you cannot run a country that way. Nor do political parties get to take a mulligan and effectively rewrite the law because they screwed up the first time. There are very good reasons that we demand that agencies hew to the law that was passed, not the law as it should have been passed or the law as a hazily shining dream in the hearts and minds of the congressmen who voted for it.

So we have to fall back on our written memory: the legislative history and the text of the law. The arguments that this result was “absurd” and that “no one could have thought that” were always bad arguments, strategically as well as intellectually. The case needed to be won in the text.

While my understanding is that the Cohn clip may ultimately end up in the briefs before the U.S. Supreme Court, the Gruber gotcha clips were never going to make a difference in the court case, because for precisely the reasons I have outlined above, courts do not normally consider post-facto statements about the law. (After all, Gruber could have been misremembering in 2012!) They are not legally important, but they matter nonetheless, because they bring the public debate back to where the legal debate always seems to have been: the text of the statute and the bills that preceded it.

Will the plaintiffs' reading of the statute prevail? On that, I have no opinion. I started out pretty skeptical of this case, but after reading the plaintiffs' briefs -- and, funnily enough, after listening to the impassioned arguments of liberal commentators -- I find it very plausible. However, I opposed Obamacare, and I have long said that it was an overly complex monstrosity with too many internal contradictions to survive, so it’s not exactly a big stretch for me to agree with the Halbig plaintiffs. And, more to the point, I am not a lawyer, so I’m not really qualified to judge their arguments about the finer points of administrative law and the two-part Chevron test.

What I think I can say, after all this, is that their view of the law is not some absurd ideological construction. If it is wrong, it is wrong in a way that anyone -- even very smart, very liberal analysts who have spent a whole lot of time thinking about the law -- could be wrong. I think I can also say that we’d all be in much better shape if we were a lot less confident about our own memories -- and a lot more charitable in engaging people across the aisle.

To contact the author of this article: Megan McArdle at mmcardle3@bloomberg.net.

To contact the editor responsible for this article: James Gibney at jgibney5@bloomberg.net.