Separating Hyperplanes
As I've argued from the beginning, the ACA doesn't actually ban any insurance plans. That isn't even within the scope of federal policy—the federal government leaves it up to the states who have their own regulators to determine which insurance products can and cannot be sold within the state. But the ACA does require everyone to have a certain type of insurance or pay a tax, and many existing insurance policies didn't qualify for this requirement. Most analysts thought this mandate would kill demand in the pre-ACA market of non-qualifying plans, giving rise to the whole "if you like your health insurance you can keep it" controversy. That's not the same as banning those non-qualifying insurance plans, but alas no one would listen to me.

Turns out they should have, because the market for pre-ACA-style non-qualifying insurance plans is booming. I certainly didn't expect that! Apparently more people are willing to sign up for cheap low-coverage plans now that they know that if anything serious happens, they'll be able to sign up for comprehensive coverage through the ACA exchanges.

Generally these plans have small networks, high cost-sharing, and a lifetime cap at around $1 million in health expenses, along with risk-rated premiums and exclusions on people with pre-existing conditions. Prior to the ACA, relying on these plans was extremely risky—not only could you easily breach your lifetime cap if you develop a serious health condition, but you could find your self unable to buy any other insurance after this because you now have a preexisting condition, and even if you could get insurance it will be risk-rated and therefore much more expensive after you develop a known health condition. Under the ACA, however, guaranteed issue means you'll be able to buy insurance with a preexisting condition, while community rating means it won't cost a dime more no matter what conditions you've developed. Hence, the ACA has counter-intuitively increased the appeal of non-conforming health insurance.

The downside of all this is that it means we can expect the ACA risk pool to be sicker and therefore more expensive for everyone who does buy it. There's a heafty tax penalty for not buying an ACA conforming plan, and there are substantial subsidies for some people who do, which together with the fact that non-qualifying plans have very crappy coverage means only the healthiest individuals will chose this route.
4/26/2016 05:38:00 PM
One of the narratives about the New York primaries has been that Hillary Clinton won the democratic primary only because New York primaries are closed, so independent and third-party voters couldn't vote for Bernie Sanders. This feeds into a long-running theme in Sanders's rhetoric that the system is rigged. It also feeds into popular rhetoric by democrats in general against Republican voter suppression tactics such as burdensome voter ID requirements.

I suspect the Sanders camp is wrong on the counterfactual—they would not have won in an open primary. For one thing, Clinton was always up in the pre-election polling by a hefty margin, which isn't what you'd see if there was a mass of voters intending to vote Sanders but then found out they couldn't. However, I find myself agreeing with Matt Breunig:
"The story is supposed to be that Republicans like to suppress the vote and Democrats don’t. I’ve long suspected this story to be false. The real story is that basically everyone likes to suppress the votes of people they don’t like and abhors suppressing the votes of people they do like. This is not to say that everyone would endorse literally any measure to suppress disfavored voters, including explicit laws categorically suppressing entire groups from voting. It’s just to say that most would endorse measures that reduce the voter turnout of their opponents, provided those measures have at least some kind of halfway plausible-sounding justification."

Arrow's Impossibility theorem says that it is impossible to design a "fair" election system, but one principle that we would all agree on, at least in the abstract, is the principle of "one person, one vote." Given that, clearly regulations which aim to prevent people from voting are wrong.

Primaries have always harbored an awkward relationship with this abstract principle. The least suppressive rules might be to allow everyone to vote in every primary, but if we did this, then Hillary Clinton would have won both the Democratic and Republican primaries! Obviously, excluding some voters is the whole intention of having party primaries.

Over the past decade or so, states have dramatically opened up their primaries to help enfranchise more people. Not long ago nearly every state was more restrictive than New York is now. As I recall, when I registered to vote, Ohio primaries were totally closed—you had to declare your party when you registered to vote, and getting that changed was kind of a pain. Incidentally, in the time between when I registered and my first ballot, Ohio changed the rules and made party registration automatic upon voting in a primary, effectively allowing anyone to show up and vote in whichever party they choose. Other states followed suit, some going even further and others stopping somewhat shorter. All of this liberalizing of the primary rules followed from a growing ease among party insiders, who thought that more open primaries would help elect moderates with more cross-party appeal since supposedly hardcore party insiders wouldn't be the only ones at the polls anymore.

2016 has upset the conventional wisdom on open primaries. There's now a strong sense that the openness they've cultivated over the past decade have enabled Donald Trump and Sanders—two radicals who are not only not actually members of their respective parties but who harbor views much more extreme than their parties' platforms.

In response, I expect most states will now move to slowly claw back some of the new open-primary rules.
4/22/2016 05:40:00 PM
Here's how I do taxes. Starting in January, a handful of W2, and 1099 forms are sent to me. I don't check them for errors. I accept the lack of a form being sent as evidence that I have no reportable income from that institution. I fire up the IRS website, open their 1040 form in one browser tab and their tax table in another tab. I copy and paste crap from my W2's and 1099's into the 1040 form, then copy and paste crap from the IRS's tax table into it as well. Before you can submit the form you have to click the "do the math" button which inexplicably takes 5 seconds despite the fact it is just adding about 4 numbers—I once built a javascript DSGE simulator that was faster than this tax form. Ok, now that several javascript epochs have passed, it will now finally let me submit it.

The really absurd part of all this is that the IRS already has copies of all the same W2 and 1099 forms. They already know literally the same amount about my income as I do. Having me send them this info again doesn't increase it's accuracy at all—I have merely copied and pasted. My role in this process adds absolutely zero value to IRS auditors.

Oh, and once I do that, I have to do it again on the Ohio website.
4/14/2016 05:08:00 PM
Slate has an article offering a defense of the expression over its alternative construction "I couldn't care less." It's on the right track but I think both that piece and the one before it are missing some obvious points.

First of all, "less" is a word that requires a comparison of two or more things. If I say "This cup has less juice in it," without context there's absolutely no way to know what I meant. Probably the most common interpretation would be that there exists at least one other related cup containing more juice in them than the one that I singled out—as in, "this cup has less juice in it than those cups." But I could also have meant, for example, "this cup has less juice in it than is required by these specifications." When using comparison words like "less" and "more" we often leave the necessary comparison clause unspoken because it can usually be easily inferred by context.

Which brings us back to the expression "I could care less." What are we comparing to? People who argue that "I couldn't care less" is the only correct formulation make an implicit but unnecessary assumption that the comparison clause is always "than I currently do," as in "I couldn't care less than I currently do." But the comparison could just as easily be "than anyone else." If we take the latter as the comparison clause, it's the "couldn't" that is incorrect: "I couldn't care less than anyone else" specifically rules out the possibility that another person exists who cares less than you.

But the other thing we need to bear in mind is that "could" is a statement about probabilities and not exclusively reserved for counterfactuals. People who argue that "I could care less" is incorrect assume that the expression implies that you do not currently care less when in fact it specifically includes this as a possibility. That is, "I do care less" is a subset contained within the range of possibilities implied by "I could care less." To give a similar example, "I could be faster than you" means that at the present moment, me being faster than you is a real possibility—it would be unreasonable to interpret that statement as saying that I am definitely not currently faster than you.

The fact is that both idioms, "I could care less" and "I couldn't care less" are equally correct. They differ in what they compare your current amount of caring to. If the comparison is to other people, then "could" is correct, whereas if the comparison is to your own current state of caring, then "couldn't" is correct.

But then, you really didn't need that spelled out for you, did you? Because you already knew exactly what is meant by "I could care less."
4/06/2016 07:17:00 AM