Should the wise pay the price of Paternalism?

Business people in bar.

Luc Bovens – London School of Economics and Political Science


Sarah Conly argues that paternalistic interventions that restrict options or that make options less accessible may be perfectly admissible on grounds that they increase the value of choice. Examples are banning trans fats and enforcing seat belts.

Conly writes: “Having more options need not be better than having fewer options. Indeed, it is hard to see that there has been a substantive loss, if the more valuable option is now more likely to be chosen. (…)  If we consider an option to do a valuable thing that has a 1% likelihood of being exercised and an option to do a valuable thing that has a 99% of being exercised, the second option is more valuable, …”

Let a choice be a set of options. Let a choice situation be the context in which the choice is exercised. Each option has a value as conceived of by the chooser. The chance that a particular option is chosen is contingent on the choice situation. The value of a choice is a function of its expected value, that is the sum of the chances that options are chosen times the values of the options, with the sum ranging over all the options in the choice set.

For most consumers, eating trans fats has low value and refraining from eating trans fats has high value. Without a ban, they are likely to eat trans fats and unlikely to refrain. With a ban, they are likely to refrain and unlikely to eat trans fats. Hence a ban increases the value of choice for the consumers, and this is what justifies a ban on trans fats. A similar argument holds for enforcing seat belts.

I can see two objections.

First, there are certain choice situations in which we do cherish choice for reasons other than its expected value. Suppose that the divorce rate for marriages between particular social groups is unusually high. Exercising one’s choice to intermarry in this manner is risky. It’s a different kind of risk than eating trans fats or not wearing a seat belt, but it’s risky nonetheless, considering one’s future well-being in matters of romance. Imposing a ban or, less radically, discouraging such intermarriages through a social media campaign, would induce people to make less risky matches. If the government were to intervene in this manner, there may be some local heartaches here and there, but the expected value of romantic choice would increase. The value of romantic choice with a ban or a social media campaign is greater than the value of romantic choice without such paternalistic interventions. And yet, few would approve of such interventions.

Conly anticipates this objection. The expected value of the choice is not the only determinant of the value of a choice. The very fact that certain options are accessible may have independent value. A choice may express “our rationality, in being the product of deliberated choice, …,” Conly writes. It is the fact that we get to choose, the fact that we have the liberty of choosing, that increases the value of the choice. Choosing to eat trans fats is not exactly such a cherished liberty, while choosing a spouse is. This is why banning trans fats is permissible while interfering in the nation’s love life is not.

Conly states that contemporary paternalistic theory “endorses the values of the individual who is subject to the regulation.” Granted, it may be hard to see how the liberty to consume trans fats matters could be anyone’s cherished liberty. But we should not be too quick to say that particular liberties could not possibly matter to people, if we wish to respect what they actually care about. West of the Hudson, there are many voices saying that what a man does in his car is his own business—seat belts and all—and the government should have no say in it. And as to food safety, some people did genuinely despair that Obama and the FDA were coming for their raw milk.

Second, there are certain choice situations in which expectations radically differ between people. Think of the opiate epidemic. The unrestrained prescription of opiates has led to massive addiction and early deaths in the US. A restrictive regime of opiate prescription or even a ban on opiates would increase aggregate expectations. But it would not increase the expectations of a person who is able to use opiates wisely. It would not increase the expectations of the person who is in desperate need of opiates. And even without a ban but with a more restrictive regime there will be people in desperate need who fall through the cracks. The value of drug choice with a ban or a restrictive policy may be greater than without for the population at large, but it is not greater for each and every person within the population. What is objectionable about paternalistic interventions is that it fails to respect the expectations of some people. And this is the more worrisome if it restricts the choices of the wiser and more self-restrained among us. Why should they have to pay the price?

A similar concern enters in with “sin taxes.” When England was debating to introduce minimal unit pricing on alcohol in 2014, people fought back, saying “We do not want to become Sweden,” where alcohol is priced high and pubs are scarce compared to in England. The government ultimately backed down. Aggregate expected value may increase with more restrictive alcohol policies, considering the costs of overconsumption, but responsible drinkers pay a price if you take away the comradery over an after-work pint due to pub closures caused by such policies.


Read the full article by Sarah Conly “Paternalism, coercion, and the unimportance of (some) liberties” in the second issue of Behavioural Public Policy for free here


Leave a Reply

Fill in your details below or click an icon to log in: Logo

You are commenting using your account. Log Out / Change )

Twitter picture

You are commenting using your Twitter account. Log Out / Change )

Facebook photo

You are commenting using your Facebook account. Log Out / Change )

Google+ photo

You are commenting using your Google+ account. Log Out / Change )

Connecting to %s