Suspicious Discontinuities

Or it’s self reported and on average people fibbed up their height a smidge

The perverse incentives around these discontinuities is one of the worst misfeatures of programs intended to help people.

Here is a real example. My niece wound up with 3 kids, divorced, with inconsistent child support payments. So she went on food stamps. She was looking for a better job. She found one, interviewed, and they wanted to hire her. But the salary that they offered was $0.50/hour too much for her to stay on food stamps, and was less than her current job. Thanks to union regulations, the person who wanted to hire her couldn't pay her less than the standard salary. The result is that she did not take the job.

A common result when Congress takes note of perverse incentives like this is that they introduce a more complex program with specific terms that address specific problems that emerged. The problem is that by creating more brackets and more complex rules you usually create MORE cliffs with perverse incentives (though the perverse incentive at each cliff tends to be less).

Another solution is to introduce rules that attempt to identify people who are "abusing the system" and punish them. So, for example, you can only be on welfare for a limited time because people got outraged that single mothers wound up on welfare for a long time. However those single mothers were acting that was in part because they were better off on welfare looking after their own kids than they were in a low end job and paying for daycare.

Stereotypically the first type of solution is characteristic of Democrats and the second of Republicans. But the truth is that both parties do a lot of both solutions.

The Polish test distribution is an interesting counterexample: it demonstrates that there is _not_ a discontinuity. Instead of failing immediately at a score of 29, the graph suggests that you start failing probabilistically, with greater probability as your score drops. The bump at 30 is merely the mechanism by which this happens. Thus, this is in fact an example of a continuous threshold of the type Dan is recommending.

Government paternalism is terrifying. People are going to screw up, and make bad choices, but having choices is better than being treated like a child your entire life.

...and if you do qualify for a subsidy, you aren't out of the woods yet as far as year to year surprises go. There was a nasty one in my county for people getting subsidies.

There were only two insurers available through the ACA exchange in my county in 2019. They were Kaiser and Premera. The cheapest plans from Kaiser were a pair of bronze level plans (one with an HSA, one without) for about $770/month. The cheapest Premera were bronze plans at something like $1000/month.

An individual in their late 50s just under the subsidy cutoff would get a subsidy of around $660/month, so if they went with a Kaiser bronze plan would be paying around $110/month. Not too bad at all.

For 2020, the same Kaiser and Premera plans are available at about the same cost (actually, they may even have come down a few percent). But there are also two new providers. One of those has a couple plans slightly cheaper than Kaiser, but comparing side by side the Kaiser bronze plans are better in most ways, and Kaiser also has more locations.

That other company's plans, though, are silver plans, not bronze.

The subsidies are keyed to the cost of the second cheapest available silver plan and so these "no better than bronze but for some reason silver" plans greatly lowered the subsidies. What would have been a $660/month subsidy in 2019 is now only about $400/month in 2020, meaning you'd pay this year around $370/month for the cheapest plans in my county.

Going from $110/month to $370/month in a year when the cost of plans went down slightly and the number of insurers on the exchange doubled is pretty damned annoying.