Let’s Get Ready to Rumble!!

Here’s my opening bit for my debate with Robin tonight.  Needless to say since we’re both all about Overcoming Bias, there will be no cheap point-scoring or sophistry and much gentlemanly conceding of valid points.  And maybe a little hitting each other with metal folding chairs.  At least in my day college kids liked that sort of thing.

Aside from basic functions like enforcing contracts and maintaining order, there are three justifications for governmental intervention in people’s lives:

1. Correct market failures.
2. Redistribution (make Peter do or not do something, or take from Peter, for the benefit of Paul).
3. Paternalism (make Peter do or not do something for Peter’s own benefit).

This discussion is about #3.  The traditional arguments against paternalism are:

A. You can’t help someone by restricting options; doing so will either have no effect or it will hurt.
The idea here is that people do the best they can, by their own lights, subject to the constraints that they face.  You can help them by relaxing a constraint (say by giving them some money), but you can’t help them by taking away a choice; if they weren’t going to choose the choice you took away, then there is no effect, but if they were, then you have made them worse off by forcing them into a less preferred choice. 

B. People have a fundamental right to make their own mistakes.
It may be that some people would be better off if someone took part of their freedom away from them, but freedom is an irreducible right, so it’s not OK even if it would help.

C. You can’t trust the government.
Even if you think that it is possible to make people better off by taking away part of their freedom, and even if you don’t have a fundamental philosophical problem with the idea, you shouldn’t give the government the power to do so because it is the nature of government to either screw it up or to use it as the thin end of the wedge to take more and more freedom away from people or both.  Even if it seems to work in the short run, it will fail in the long run as the policies get more and more entrenched and hard to get rid of and as citizens get less and less used to taking care of themselves.

Despite its usefulness for some kinds of economic theorizing, I don’t think many people really take Argument A seriously.  The idea that it’s impossible to make a decision that is in some meaningful sense bad for you just flies in the face of common sense.  Argument B has some appeal to it, and would probably have a lot of appeal beyond a certain level of paternalistic intrusiveness, but I think most people who invoke it really mean some version of Argument C; it’s not that they have an irreducible philosophical opposition to paternalism, it’s just they they think that beneficial paternalism is pretty much never on offer, at least from the government.

So Argument C is where the action is.  In order to believe that the government should paternalistically use its coercive powers on Peter, you have to believe both that the government knows better than Peter what’s good for him and that it will use this power in his interest, and do so without causing too much dimunition in his ability to look after himself, or to raise kids who are capable of looking after themselves.  By no means are these conditions guaranteed to hold, and that’s why paternalism should be limited in scope and in intrusiveness, more or less in the spirit of "libertarian paternalism."  But it is not hard to imagine that they could hold.  The government can sometimes know better; it wouldn’t have to be staffed with impossible geniuses in order to know better than, say, the least intelligent 20% of the population (probably much more) the consequences of investing in a pyramid scheme, or of having no savings for retirement, or of ingesting some dangerous snake oil miracle cure.  As Eliezer Yudlowsky has pointed out, without paternalism, these people will just suffer, and sometimes suffer a lot.  And even smart people get benefit out paternalistic regulation as it allows them not to have to become knowledgable about each and every issue for themselves (firm reputation effects and private monitoring agencies will only get you so far). 

Nor is it impossible, in a well-functioning society, for the political stars to be lined up so that governments and bureaucracies have an incentive to engage in paternalism and to do at least a decent job of it.  It is well to remember not to fall into Tyler Cowen’s "libertarian vice," which is to believe that the quality of government is fixed (and very low) rather than what it is, which is highly variable and dependent on how well the society has its act together.

Robin has suggested that instead of using coercive power to ban bad things, the government could instead operate "would have banned" stores that offer the government’s advice (and require that you pay it some heed), but don’t actually compel you to do or not do anything.  Of course, there are a lot of cases where this won’t work (say Ponzi schemes or Social Security), and there are a lot of practical issues (would the store be located in prime real-estate or in a slum?, would it be allowed to advertise?, how strict would be the rules about who could shop there?), but let’s leave all that aside and examine the principle.  In my view, if such a program actually served to warn off a lot of people who would have swallowed poison while giving a lot of smart people access to, say, drugs that have been approved and used successfully in Europe, then great.  But in my view that’s not an alternative to libertarian paternalism, that’s a kind of libertarian paternalism.  The question of whether it works remains, and I’m skeptical that it would, but I don’t see any major question of principle.

So far I’ve just laid out the standard case for an (I would say) mild pro-paternalism position.  And that pretty much is my position. But there is something else, something that is absent from most debates on this subject, which somewhat increases the scope of the paternalism program that I favor, and which very much increases my zeal for the central bits of it.  And that is the role of capitalism.  Markets are attractive in that they are an efficient way to allocate resources to satisfy human wants, if you take those wants as given.  But in modern capitalist societies, firms don’t just provide you with the means to satisfy a set of pre-existing wants, what they do is use advertising and other tactics to tell you what to want, and they do so incessantly and remorselessly and from extremely early childhood. There is nothing in economics that says that the effect of this has to be benign, and in my view it is catastrophically bad in a lot of ways, and this is one of them; they very often tell you to want things that are bad for you.  When the government abandons the field, they don’t just leave behind a bunch of fallible humans who might stumble into making an unfortunate mistake from time to time.  They leave behind a bunch of overmatched, outgunned marks who will be manipulated and cajoled and badgered into making extremely costly mistakes, again and again and again.  It’s not a fair fight, and only the government can even hope to even up the sides.

GD Star Rating
Tagged as:
Trackback URL:
  • Another problem: other-regarding actions. It’s quite possible to make a bad decision whose consequences affect persons other than yourself.

  • conchis

    you kind of edge towards this at the end, but it seems that one often overlooked aspect of the debate about paternalism concerns the distribution of its effects. too often, individuals are treated as suffering from identical failings. in contrast, an (incredibly simplified) argument could be run as follows:

    (1) the segment of the population most likely to benefit from paternalism is that segment most likely to make bad decisions;
    (2) the segment of the population most likely to be harmed by paternalism is that segment most likely to make good decisions;
    (3) the segment most likely to make bad decisions is also likely to be worse off in in general than the segment most likely to make good decisions;

    if we care more about the worst off, then this strengthens the case that the tradeoffs involved in paternalism are one’s we should, in some cases, simply be willing to wear.

  • David, you say “paternalism should be limited in scope and in intrusiveness.” I thought we would be debating typical actual government policies which are justified via paternalism. If you want to limit the scope, could you give us some idea which paternalistic policies you do support?

    You focus on how government might know more and be more intelligent, and how firms use a lot of ads to push their views. I’ve been trying to argue this is not enough to justify paternalism; for rational agents warnings and further ads should be sufficient to deal with such issues.

  • David J. Balan

    Alex, That would go under redistribution in my typology.

    Conchis, There is little doubt that the people who are most likely to benefit from paternalism are likely to be the least well off in other ways. And that is an argument in its favor: sometimes the well-off just have to take one for the team. But at least some paternalistic regulation benefits almost everybody.

    Robin, I basically support the things that are currently done in the U.S. for paternalistic purposes: social security, pure food and drug regulations, regulations of some kinds of potentially predatory financial practices, and so on. My position is not that everything is perfect as it is; I could easily be talked into believing that one or the other of them could be made more effective and/or less obtrusive, or even that some of them should be scrapped. But for the purposes of our debate, I think I can be thought of as a reasonably conventional supporter of paternalism as it’s currently practiced in the U.S. As for the latter point, I looked at your paper, and while I didn’t invest enough time to fully understand it, I think I got the flavor of it. I haven’t fully absorbed the implications, but I’m pretty sure that ultimately my defense of paternalism rests mostly on agents being irrational and therefore systematically wrong and/or “just having no idea” (whatever the faults of the latter concept).

  • Matthew

    Robin, I basically support the things that are currently done in the U.S. for paternalistic purposes: social security


    Social security in practice has worked out as a pretty classic ponzi scheme.

    The earliest retirees got wonderful returns, while my cohort will get negative real returns from this “retirement program”.

    Now I am not the most savvy investor in the world, but I suspect my 401K invested mostly in index funds will net me 200-300%+ total returns over my working lifetime. Versus net negative real returns from the 12.5% extracted from most people’s incomes for the mandatory “retirement program” of social security. Of course, if I happen to be a lower-income African American male, Social Security is an even worse deal, since my life expectancy is much lower. Then again, if I am a professor making $250K a year, social security isn’t such a bad deal, since it maxes out at 90K or so.

    Explain to me now why the current Social Security system is a good program. . .

  • You have to believe both that the government knows better … and that it will use this power in his interest. … The government … wouldn’t have to be staffed with impossible geniuses in order to know better than, say, the least intelligent 20% of the population.

    Ultimately my defense of paternalism rests mostly on agents being irrational and therefore systematically wrong and/or “just having no idea.”

    When two agents are rational, have the same interests, and do not communicate, then the one who knows more and is more intelligent will make the better decision. But if the agents are not rational, can communicate, and have somewhat differing interests, then it is not at all enough that one of them knows more and is more intelligent. You seem to be trying to have it both ways – assuming rationality in order to prefer the person who knows more, and assuming irrationality to ignore the possibility of communication.

  • Stuart Armstrong

    Explain to me now why the current Social Security system is a good program. . .

    I’d assume that David would be in favour of the “Security” aspect of it…

  • Tom Crispin

    It seems to me that much of this discussion is distinguishing advertising (Ride our Bus!) from paternalism (Use Public Transportation!) based on the motives of the actors rather than on the content of the message. And that judging a person’s behavior based on the quality or purity of his motive is a bias we should avoid. Road/Intentions/Hell anyone?

  • Paul Gowder

    I think there’s another angle in all of this: bounded rationality and organization. Consumers are individual boundedly rational agents, regulators aren’t individual people, they’re institutions. Suppose it’s the case that regulatory agencies have checks to minimize the risks of the sorts of errors that individual consumers, or individual regulators, would make? (Those checks might include, e.g., notice and comment, legislative oversight, committee work leading to applications of the condorcet jury theorem, lots of extra data, division of labor in accord with skill, multiple layers of approval required for decisions, etc. etc.) Then overall, there might be reason to believe that regulatory *agency* decisions will be closer to the truth than consumer decisions.

    I think Robin misses this by modeling regulators as single individuals.

  • I hope this debate will lead to a Disagreement Case Study or two! I’d be interested to hear about the meta issues which you use to justify continued disagreement. Do you think you would eventually come to agreement if you spent enough time together?

  • Stuart Armstrong

    It seems to me that much of this discussion is distinguishing advertising (Ride our Bus!) from paternalism (Use Public Transportation!) based on the motives of the actors rather than on the content of the message.

    I think it’s more a question of distinguishing “eat our big mac” from “ensure a healthy diet”.

  • 1. Information cost is one reason why people value paternalism. It is far too time consuming for a person to get the relevant information about every possible drug, etc, and usually the consumer doesn’t have the domain knowledge to understand the information for a particular drug even if they could get it. So we delegate that task to the FDA, and similarly for other bans.
    2. Decision cost is another reason why people value paternalism. If we have ‘would-have-banned’ stores, people would have to decide whether or not to buy the product, or decide not to decide. But we already have too many decisions to make, so we delegate the decisions.
    3. The libertarian idea is based on liberty, but that concept only makes sense for a unitary self. Only a little introspection is needed to see that we have multiple selves, not only at different points in time, but at same time. Decisions that are preferred by one self are not preferred by others. Literature (e.g., Proust), social psychology, and behavioral economics converge on this.

  • Bruce – (1) Don’t warnings fix this information cost as well as bans? (2) Even under bans, people still have to decide to obey the law. Consider how much effort people spend to decide when and whether to speed on roadways, even investing in laser and radar detectors. I once ordered a pharmaceutical from an overseas pharmacy rather than go to the doctor for a prescription. Bans and regulation don’t necessarily take away the need for a decision. To the extent that you can decide, once and for all, to obey the law, you can decide not to violate government recommendations. (3) I think you are rebutting some argument which I don’t see spelled out in enough detail to comment on. I’m not sure anyone here has made an argument that relies on a unitary self.

  • David, thanks for debating with me today. I will try to summarize your position and you can correct me. You are wary of most of the paternalism in most societies at most times, such as that banned alternative religions, political groups and sexual orientations, and that limited the freedoms of women and ethnic minorities. But, you argue, since our society today is the most “successful” of all, our paternalisms should be presumed to be good on balance, and so the people who run those paternalisms should be presumed better judges than those whose acts are limited.

  • I would agree that on the whole, the paternalism of today is morally superior to that of the past. Of course, I am a creature of the present day. People of the past, if they could see our government paternalism in action, might be horrified to see what they would have judged as immoral actions being encouraged and even mandated. Nevertheless, I judge that humans have made moral progress over the centuries, just as we have made progress in other areas of knowledge.

    It’s likely that in the future, government paternalism will be substantially different from today, and that some of its policies will seem to be very questionable from our present-day moral perspective.

  • TGGP

    So was it rational for the society of the past to support those old forms of paternalism? By the standards of that time it was the most “successful” society in the same sense that ours is now.

  • Roy Haddad

    You made A too strong. A weakened form of it would be another considerable argument against paternalism:

    A) People tend to choose what is best for them, and as they are educated, they are more likely to do so.

    A.1) People learn (i.e. become educated) through their choices.

    A.2) There are markets for education as well as for informed third-party decision making (e.g. doctors, lawyers, contractors), reducing the complexity of many choices.

  • Marcus

    When the government abandons the field, they don’t just leave behind a bunch of fallible humans who might stumble into making an unfortunate mistake from time to time. They leave behind a bunch of overmatched, outgunned marks who will be manipulated and cajoled and badgered into making extremely costly mistakes, again and again and again. It’s not a fair fight, and only the government can even hope to even up the sides.

    Corporations manipulate, cajole and badger people? OK, but what of the government? Such as making children pledge allegiance to its flag or otherwise promoting nationalism. Doesn’t the government have an unfair advantage over its ‘marks’?

  • Bruce Britton

    Hal, thanks for your comments. They really made me think even more deeply about the subject.

    1. It seems to me that if it’s just a warning, then you have to make a decision about whether to follow the warning or not, and you have to invest time and effort in gathering the relevant information, so you have not eliminated the information or decision costs.

    But If it’s a ban, then your work is done, you don’t need to gather any information because you don’t have to decide.

    Of course, if you decide in advance to treat all warnings as bans, then you avoid the costs, but that’s just changing the name of ‘warnings’ to ‘bans.’

    2. I was assuming people make a meta-decision to obey the law. If you decide not to be bound by law, then bans (about anything) are no longer bans, they are just information, and if you have to decide what to do, you have information and decision costs as above. But note, if you are not a law-abiding citizen, then other costs come into play.

    3. I thought I detected a libertarian strain in Robin’s arguements on this and on other issues, but perhaps I am wrong about this. If he is using libertarian arguements, then such arguements rely on a unitary self, as follows:

    Libertarianism is based on a person’s self-ownership, with liberty to do whatever he wants with his self, as long as he allows others the same liberty.

    But with multiple subselves within the same person, when one subself takes an action it will often restrict the liberty of another subself, thus making a nonsense of the libertarian project.

    For example, suppose one subself spends money or time or effort on something: the liberty of all the other subselves is restricted because they can’t spend that money or time, because it is gone.

    It is just like if I spent YOUR money on something; that would restrict your liberty, because you couldn’t spend it on anything, because it is gone.

    Now suppose your self were in here with my self (and I mean that in the most abstract possible sense, Hal), and the Bruce subself spent some money.

    Then the Bruce subself would have restricted the liberty of the Hal subself, because Hal would not be able to spend the money, because it is gone.

  • Two different cognitive biases that color people’s view of issues are an easier identification with individuals than with masses and a stronger sentiment stirred by pictures rather than words. A third bias is a built-in preference for narrative over statistics.

    All three biases are on display here, thus proving that slimy techniques can be deployed for good causes as easily as for evil ones.

    Bias-harnessing = Jedi Mind tricks

  • TGGP

    The me of right now and the me that has put his hand on a hot stove are different, but nevertheless I keep my hand off the stove. Another person might place my hand on the stove and when I tell them it burns they might continue to believe it is for my own good.