Thread by @webdevMason

The main reason I don't consider myself much of a libertarian anymore: the most reckless people are generally the least capable of mitigating, covering or compensating for the damage they can cause, and modern communities don't really have a way of expelling people like that

I generally think people ought to be able to do things that carry the potential risk of externalities for others as long as they're able to insure against those risks or cover damages, but nobody really wants to bite the attendant bullet: that wealthier people have more freedoms

We're comfortable with doing this with limited scope, e.g. car insurance. If you can't afford car insurance, sorry, no car for you. But we don't require addiction insurance to drink or use drugs, or concussion insurance to play sports that cause socially harmful brain damage

It's good for society to subsidize *productive* risk-taking, and arguably take the meta-risk of subsidizing some risk-taking that may *turn out* to be productive. But it's hard to make that case for purely recreational risk-taking

https://twitter.com/PurpleWhale12/status/1383849491703365637

And we haven't socialized the risk associated with many productive activities that are also personally beneficial, e.g. driving. We expect companies that run oil tankers to tap their own coffers to prevent and clean up spills, even if many of us benefit from having access to gas

Where our intuitions run the other way — communities should subsidize risk — some unsavory stuff tends to happen under the hood. For example, less wealthy towns buy bus tickets for indigent addicts and the mentally ill homeless and hope they can make them someone else's problem

For better or worse, I think it's basically inevitable that you'll see various forms of "soft banishment" anywhere you have people causing a lot of social problems that communities are expected to subsidize but don't have the power or resources to mitigate/control

Insofar as communities have been able to solve their own social problems cost-effectively, it's been via two mechanisms: (1) very firm norms and enforcement around behavior that's dangerous to the community, (2) technologies that reduce the behavior or mitigate its costs

So, a given community is looking at five real options for dealing with costly social behavior:
1) Enforced personal responsibility for damages, e.g. insurance
2) Cultural norms, status manipulation
3) Legal regulation, imprisonment
4) Technology for mitigating harms
5) Banishment

In dealing with child abandonment, for example, a community might:
1) Force the abandoning parent to pay child support
2) Shame the abandoning parent
3) Make abandonment illegal
4) Encourage high-efficacy birth control/abortion
5) Remove abandoners from the community

Some options are more practical than others. Some abandoning parents may have no ability to pay. Imprisonment and banishment don't solve the problem of the unsupported child. Cultural norms and technology are difficult to control and may cut against each other in complicated ways

What I've noticed is that people usually have pretty strong moralized aversions to some of these strategies (and preferences for others) generalized across circumstances. And frankly, people don't generally seem to care all that much about what will *work*

From my perspective, these are all just social tools, and depending on the specific circumstances some of them will work better than others. When people fixate too much on the strategy they *want* to work, they're prone to make bad assumptions about the nature of the problem

Yes, and people who are attached to particular policies that *have* been tried (with poor results) very commonly move the goalposts away from their initial descriptions of success/failure, or argue that the execution of the policy didn't go far enough

https://twitter.com/DavidDeutschOxf/status/1383876084161122311

In general, people demand less apparent success from policies they like than from policies they dislike, attribute failures to their opponents' efforts, or insist that their failures are evidence that they need *more* power, resources, and time to proceed without scrutiny

Of course, it's true that no policy is executed in a vacuum (it's never isolated from myriad factors that may influence the outcome) and that counterfactuals can't definitively prove anything. It's just not a good argument for rewarding consistent failure and staying a bad course

The problem with allowing uncertainty to excuse poor results is that it ultimately favors policy proponents who confidently make the most attractive promises. For accountability to exist at all, it has to price in uncertainty. Ideas have to face judgment on contact with reality.

Because I picked on libertarianism in the first tweet, I want to mention that I think the libertarian mechanism is often very sound. For one thing, it's hard to arbitrarily punish harmless behavior if the punishment is simply being forced to compensate those you materially harmed

Problems arise when the perpetrator *can't* compensate the victim, or refuses to. But there are always circumstances where the various strategies can't be expected to work very well. Imprisonment is expensive. Cultural norms are hard to manipulate, blunt, and unevenly enforced

And it should go without saying that behavior caused by desperate need or severe mental illness probably won't be very responsive to punishment of any kind. I think people overestimate how much criminality can be solved with social services alone, but it's not *none* of it