Who sits in the tail

The idea can be put as statistics. A society with fewer rules lets more people try more things, and each attempt can go well or badly. More attempts with uncertain results should widen the spread of outcomes: more people doing something remarkable and more people coming to grief, whatever happens to the average. Freedom, on this view, is a variance machine, and an argument about it that looks at one tail is looking at half a distribution.

American states differ enough to test the shape. Jesse Harrington and Michele Gelfand scored all fifty on "tightness", the strength of punishment and the latitude allowed, and compared the scores with how the states fare.¹ Looser states had more utility patents per person and happier residents; they also had more drug and alcohol use, more homelessness and more social instability.¹ That is both tails at once, as the idea predicts.

The tight states, though, had a bottom of their own. Tightness went with higher incarceration rates (a correlation of 0.62), more charges of employment discrimination per person (0.61) and lower happiness.¹ The authors conclude that "both tightness and looseness have relative costs and benefits", and they say plainly that their data cannot show cause.¹

That changes the picture. Constraint does not remove the lower tail; it changes who is in it. State averages cannot say who anyone is, but they suggest that in a loose state more of the people at the bottom are people who fell, the addicted and the homeless, and in a tight one more of them are people on whom the rules fall hardest, the imprisoned and those who face discrimination. Outcomes may well spread wider where rules are looser, but no set of rules decides whether there is a bottom. It decides who is there.

Some freedoms lift the bottom outright. In 1969 California let one spouse end a marriage without the other's consent, and most states followed in the next decade; Betsey Stevenson and Justin Wolfers used the staggered timing as a natural experiment.² Female suicide, which they used as a measure of "the extreme lower tail of the distribution", fell by around 8% on average and by about 20% in the long run in states that adopted unilateral divorce, with no effect on men, and domestic violence fell sharply as well.²

A freedom to leave works on the bottom specifically: it gives the person in the worst position a way out. A freedom to try adds chances at every level, good and bad. Freedom is at least two dials, and they turn the distribution in different ways.

An argument about freedom is an argument about the shape of a distribution and about who sits where in it. A loose society should expect much of its best and its worst news to have the same cause. A tight one should look at who is at the bottom of its own distribution and ask whether its rules put them there.