Paternal policies fight cognitive bias, slash information costs, and privelege responsible subselves

All of us are of diminished capacity when cognitive biases degrade our thinking, and one way to fight cognitive biases is to expose our thinking to the scrutiny of others, as was done in the recent debate on Paternalism. In all of our thinking, whether we are thinking of taking cocaine or Dr. Quacko’s Snake Oil Miracle Tonic, or gambling with futures options or roulette wheels, we are subject to cognitive biases. For example, for gambling we humans are subject to the following biases: gamblers fallacy, clustering illusion, availability heuristic, attentional bias, illusory correlation, ludic fallacy, optimism bias, overconfidence effect, positive outcome bias, rosy retrospection, and the Texas sharpshooter bias.

To help us avoid these biases we have hired representatives, who create agencies (like the FDA and the SEC) with committees and subcommittees that debate the issue: they put on seminars and conferences and write working papers and white papers and cost/benefit analyses, and invite comments, etc, in short, consider the matter in depth, and then decide to ban certain drugs or activities, and not others. We voters ratify this every 2 years, except when we change our mind, as with alcohol, tobacco, or thalidomide.

Besides saving our lives, minds and fortunes, a side benefit is savings for us citizens in information costs, because we citizens don’t have to read all the papers, and a good thing too, because we have to use our time to make profits, the froth from which is used to pay our representatives. (We also save on decision costs.)

All this we do because we know we have many selves within ourselves, including a short term self who wants to snort cocaine, and guzzle Dr. Quacko’s Tonic because ‘everyone’s doing it’ (the bandwagon bias), and try our luck with porkbellies, who has a bad case of the Bias Blind Spot (the meta-bias which makes us think we don’t have to compensate for our cognitive biases) VS. a long-term self, who knows it has cognitive biases and knows it can’t overcome them, not alone, it needs help.

GD Star Rating
Tagged as:
Trackback URL:
  • Hal Finney

    Your history seems to involve a bit of spin. “To help us avoid these biases we have hired representatives…”. That’s not really what happened, historically. The situation is far more complex than that. I certainly never set out to “hire” any representatives in order to avoid biases, so I don’t know who this “we” is. And I suspect much support for paternalistic regulation comes from people who think of it applying to others than themselves.

    I am confused about your examples when you write: “We voters ratify this every 2 years, except when we change our mind, as with alcohol, tobacco, or thalidomide.” I gather that changing our minds in the case of alcohol refers to Prohibition and its repeal? What are you referring to with tobacco and thalidomide?

  • Bruce Britton

    Thanks Hal,

    I meant that the voters hire representatives, as I did by having a vote on the hiring committee for my Member of the House of Representatives, Senator, and President, and they in turn hire the members of the FTC, FDA, SEC, etc, who in turn issue regulations and propose legislation, which is then considered by the House, Senate, President, etc. Some of which regulations and legislation are intended to help me not do stupid things because I have biases or lack relevant information.

    Tobacco was first not regulated, then later it was regulated federally, and then the states won cases to regulate it further. Thalidomide was first allowed, then forbidden, then allowed again.

  • Robin Hanson

    Yes, those who want to overcome biases might better do so by seeking the advice of specialists who focus on looking for biases and offering arguably debiased-advice. Those who want to overcome other peoples biases may offer unsolicited arguably debiased-advice. And those who want to interfere with others for other reasons may pretend to only be trying to help them overcome their biases.

    The question is whether each person should be able to choose who to trust for debiased advice, or whether a democratic majority should be able to force others to follow their advice.

  • Douglas Knight

    All this we do because we know we have many selves within ourselves

    The title of this post is one thing, but this causal claim seems absurd to me.

  • Eric Crampton

    Ummm…doesn’t the incentive system in the FDA automatically induce bias? I mean, you don’t get fired for causing statistical deaths by being a bit slow, but you do get fired for approving one you shouldn’t have, right? Doesn’t that automatically make the FDA really really risk averse about one type of risk? Mightn’t we expect that kind of bias to be worse than we might expect from individual consumers? Didn’t Gieringer estimate that while the FDA might be saving something like 5-10K deaths per decade by keeping bad products out, it’s also causing somewhere between 21-120K deaths per decade because of their delays in releasing drugs? If the FDA is killing an order of magnitude more people through its delays, it seems a bit obscene to laud them for protecting us.

  • Telnar

    I believe that the information costs involved in selecting representatives based on their positions on the FDA are at least as high as those which would be involved in researching medicines directly. Here are a few quick arguments:

    — Candidates for political office generally don’t emphasize their views on safety regulation, so effort will be required to determine what they are.

    — Regulation questions are generally of low importance to someone who can not predict that he will be personally subject to the regulation. That makes it costly to vote based on regulation issues, since the available politicians are likely to also differ in their views on other issues more important to the voter.

    — Regulation questions are generally of low importance to politicians also, so there is significant risk that even after doing the research necessary to find a politician who agrees with him, a voter will have wasted that effort because his representative will trade a vote on regulation for something more important to the politician.

    — Voters have the time and motivation to do extensive research when they personally face a major decision and the need is likely to be rare (even acknowledging that bias and laziness can prevail here). Politicians are likely to be far less motivated to make a through effort at researching the trade offs involved in various kinds of rules for regulators.

    — The regulators themselves will have time to do careful work, but may have been given inappropriate incentives — for example, incentives to set a uniform rule barring something for ease of enforcement. That incentive can exist even when making it more difficult to get would have been clearly better overall (e.g. if those who can benefit from a substance which might be prohibited would be happy to pay enough extra for it to pay for adequate controls to prevent others from harming themselves with it out of ignorance).

  • michael vassar

    You know, it seems to me that literal paternalism would solve the “multiple internal agents” problems Bruce discusses far better than legislative regulation. In a variety of domains where individuals are suspected of being biased by short term “system one” errors etc, we could, and do until a certain age, give parents a legal right to restrict their behavior. If we believe that this right is necessary, why not simply extend the age of parental authority, or more optimally, designate a discrete set of “parental” authority types that individuals have periodic (annual) opportunities to review and to transfer to some other parental figure. Some sort of licensing might or might not be required for “in-loco parentis” candidates (and for biological parents, or perhaps not). The right to designate parental authority, as well as various reductions in such authority, might be gradually phased in with age, or as the result of various tests or maturity rituals, or possibly a combination of both. The current system, that of, say, designating the power to ban a particular drug for a particular patient to legislative authorities with limited medical expertise and no patient-specific knowledge seems dubious.

    I honestly wonder if the system outlined above would work, and if so, how well. It seems far too alien to modern norms to be usefully advocated, and probably severely subject to memetic infection (but less so than our actual system). I think that there may actually be very large business opportunities for the production of voluntary less comprehensive systems along these lines. The corporation itself and law may be considered related institutional systems in the same class.

  • JMG3Y

    So, if you are a mid-level employee in a regulatory agency with professional scientific or legal expertise in the relevant regulatory area:

    – Your regulatory mandate was written by lawyers for lawyers decades ago in a completely different technical environment under much different economic conditions and that no one dares to attempt to update it for fear of the outcome
    – You have a broad regulatory mandate that is not, never has been and never will be fully funded (enabling vs. authorizing legislation) so you have to establish regulatory priorities with differing impacts on the different stakeholders (FDA’s CPG’s)
    – Your agency boss is a political appointee of the current administration, who will change with next presidential election and who is advancing the current political agenda that is in direct opposition to the agenda of the previous administration
    – You are in the midst of continual tug of wars between powerful stakeholders with opposing agendas that ply the calculus of campaign contributions and votes that are doing battle to exhaustion in the the legislatures, the press and the courts, heading to the other arenas when they loose in one that do not trust each other or you.
    – If you screw up sufficiently, you or your immediate boss will be called before the most sanctimonious bunch of two-faced individuals imaginable attempting to generate evening soundbites who will posture as though they had nothing to do with the situation, were completely unaware of it (despite numerous contacts by constituents but finally being moved to action by a big press expose) but are exceptionally indignant and scornful of your actions.

    Careful work? Like a scared rabbit feeling the predator’s breath.

  • Stuart Armstrong

    All this we do because we know we have many selves within ourselves

    The title of this post is one thing, but this causal claim seems absurd to me.

    Yes, it’s an important part of your argument against libertarianism, Bruce, but I feel that you haven’t defined it sufficiently. Can you explain exactly what you mean? And preciesly how it undermines libertarianism?

  • Matthew

    Yes, it’s an important part of your argument against libertarianism, Bruce, but I feel that you haven’t defined it sufficiently. Can you explain exactly what you mean? And preciesly how it undermines libertarianism?

    I also echo Stuart’s request for more explanation of your claim here and how it invalidates libertarianism and makes paternalistic government a superior choice.

  • Richard Hollerith

    (My previous comment deleted by Robin at my request.)

    Additions to Telnar’s list:

    — A voter cannot hold an elected official to a campaign promise other than via the weak sanction of not voting for them two, four, or six years later.

    — A voter’s vote is diluted by the votes of millions (most of whom get their information about candidates from television news, ads and highly-rehearsed televised debates). In contrast, a decision made by a individual under the nonpaternalistic arrangement does not suffer from a similar dilution.

  • TGGP

    I don’t vote. Neither do most Americans. One vote doesn’t have any marginal impact (you’re more likely to win the lottery than change an election).

    If I wanted to hire someone to protect me from myself it would be Nick Cotrell, not the government.

  • Stuart Armstrong

    The question is whether each person should be able to choose who to trust for debiased advice, or whether a democratic majority should be able to force others to follow their advice.

    The question is also which situation offer the most debiased advice – a governmentally imposed system or a free-market one (unless you are morally a libertarian to the extent that that trumps all other considerations, in which case that question is irrelevant). In both cases you are effectively at the mercy of groups of people who have a huge information asymetry with you, and their own goals in mind.

    We’ve been over some of the weaknesses of governmental advice, but what do you need for free-market advice to be reliable? Which advice markets are efficient, and which aren’t? I’m thinking of trying to put together some criteria, but there probably already are some already; anyone know where I can find them?

  • Bruce k Britton

    To TGGP: Thanks for your comment.

    I think the model you are referring to is the Rational Choice Model of Voting Participation, which is that people will vote if

    p times B > C

    where p is the probability that your vote will change the outcome, B is the Benefit you would obtain from your candidate winning, and C is the cost of voting ( inconvenience, shoe wear, etc,)

    Since p is vanishingly small, maybe 10 to the -90th power for a US Presidential election, the model predicts much lower turnout than actually occurs, for which reason it has been suggested that instead of calling it the Rational Choice Model, it should simply be called the False Model of Voter Participation.

    Fortunately for the reputation of Behavioral Economics, there is a related model which predicts much better, which is that people will vote if

    p times B + E >C

    where E is the Expressive Value of voting, to include the satisfaction derived from such things as: compliance with the ethic of voting, affirming one’s allegiance to the political system, affirming a partisan preference, affirming one’s efficacy in the political system, and other satisfactions, such as the pleasures of gathering information to decide which way to vote, social interaction associated with voting (discussion and debate, etc) and similar factors.

    This model is probably to be preferred, since it predicts the results better than the other. I wonder what we should call it.

    (All this comes from Keith Dowding’s 2005 paper in the British Journal of Politics and International Relations, 7(3), 442-459.)

    I saw the Nick Cotrell video, and it’s very funny, thanks. My son liked it too. But I have to say that if I had a choice I would actually pick the government over Nick Cottrell. Actually, I wouldn’t be surprised if you would too, unless you are a serious masochist ( I don’t mean to get personal here). The government has not beaten me to a pulp and defenestrated me, not yet. Nor has it banned smoking, yet. And yet I gave it up, without the help of either Nick or the Government, in a previous century. The Dual-self Model of Impulse Control is probably more relevant here than either Nick or Uncle Sam.

    To the other commentors, I plan to get to the other comments just as soon as I wash the dishes and check the homework.


  • Bruce K Britton

    To Stuart Armstrong, Robin Hanson, and other commentors:

    I think Robin Hanson’s question along with Stuart Armstrong’s proposal goes right to the heart of the matter

    — how can we decide paternalism vs. the libertarian alternative on an empirical basis, instead of bloviating hot air? as I have been doing —

    and let me propose an answer to the question: Where do we find the most unbiased and knowledgable advice market?

    I bet you see where I’m going already… IT’S THE PROFESSORS! That’s why they give us tenure, right? And by way, if you have spent some time in the bowels of government, as I have (as an outsider at DARPA, ONR, AFOSR, US Forest Service, mostly) did it seem to you, as it did to me, like nothing so much as a University Department, inside of which I spent several decades: it’s full of seminars, and papers, and reviewing papers, and discussion groups, and committees and subcommittees (the only thing missing is the teaching, but there are a lot of add-ons, like annual day long refresher courses on defensive driving, and mandatory training on Total Quality Management, etc). I’ve spent only a little time in business environments, but not ones that function as advice markets, so I can’t say anything useful about those.

    I suppose though that Professors shouldn’t be trusted to give advice on things in which we have a possibly biased self interest, like University Governance — but wait! Isn’t that exactly the thing they do ask us to give advice on? Oh Dear.

    As I read this over, it sounds like it might be interpreted as ironic, because there may be an ironic edge to the writing, but be assured I am serious, I do think the Professors are the best advice market available. And I’m not saying that with an ironic edge, though I can see that people who are not Professors might wonder …..

  • Bruce K Britton

    TO Michael Vasser:

    Your proposal is very interesting, and I have some questions.

    What are ‘system one’ errors, I’m not familiar with that term?

    Also, isn’t this what we do with old people, when they begin to have what we see as diminished capacity, we children try to persuade them not to drive, and the state takes away their licenses, etc, though of course there it is the children who have a legal? right to restrict the behavior of their aged parents, as well as minor children being restricted by their more competent parents, as you point out. As you also point out, in loco parentis is used in college, before the children are 21.

    But you seem to be proposing such a scheme for those between 21 and 65, right?

    if I read you right. I see why you say that the law is a system of this kind, but I am puzzled what you mean when you refer to the corporation, what do you have in mind there?

    As far as business opportunities, I guess life coaches, psychotherapists (often with licensing requirements), and similar professions are voluntary systems along the lines you have in mind. I wonder if a system more aligned with modern norms might be developed along the lines of discussion groups, things like men’s groups ( in which I participated for a while some years ago, and got some excellent advice based on other’s life experiences, etc.), and other voluntary social groupings.

    Please let me know the what you think about these questions.

  • Matthew


    I agree that the “self” is a fiction. I would not say that we have multiple “selves” though. What is actually “self-like” about us is the continuous stream of consciousness, but the idea we have of agency is an illusion (in fact the “I” itself is simply a thought which is believed in, which is what Buddhism means by anatta).

    I still don’t see how this actual lack of selfhood means that treating each individual person as an autonomous agent is worse than relying on government agents to make all the decisions they care to make about how everyone should live. That is the part of your thesis I think is not well developed, and the one that historically governments have made a complete hash of. . . E.g. consider cost/benefit studies of the FDA, and how many lives it costs through drug delays vs. lives saved by keeping bad drugs off the market. . .

  • michael vassar

    System 1 vs system 2 is standard heuristics and biases terminology for fast automatic cognitive systems and slow deliberative but potentially debiasable systems respectively.

    I think that corporations serve to dilute aspects of responsibility asymmetrically, often in order to encourage relatively consequentialist behavior. Employees can be obliged to do things that they might otherwise not have the stomach for, some of which need to be done.

    Sadly, no time to go into detail into my institutional ideas.

  • TGGP

    I think the expressive model of voting is true. I think Bryan Caplan’s theory of “rational irrationality” is the best description of how things are that I’ve heard of. It does not make me confident in our democratically elected government.

    It is true that I would not hire Nick. I don’t smoke and I don’t want to be protected from myself. The thing about that scenario though is that people are actually choosing Nick, just as people can choose to limit their freedom through other contracts like marriage or non-disclosure agreements. Stirner is one of my significant influences, but I think this post from Will Wilkinson at Happiness & Public Policy on the “Stirnerite fallacy” (and the Virginia Postrel review of “Paradox of Choice” it links to) do a decent job of explaining how we do choose to limit our choices or commit to things. Many of the paternalistic things the government does we could do as well, like Odysseus having his men tie him up. We don’t because we don’t want to. The government is enacting these policies not on people who have requested them for themselves (even different versions of selves) as in the Nicotrel commercial, but on people who don’t want it. For a government that considers abortion to be a constitutional right, claiming that they are protecting the senior citizen me who will get diabetes but cannot speak for himself due to his current (and possibly permanent) non-existence rings hollow.

    Judging from some of his other posts here, I doubt Robin Hanson would concur with your faith in professors / University departments. They aren’t so much a market for advice/information as prestige/credentials and once in government aren’t a market at all. I suppose I have less confidence in the first X names in the phone-book than William Buckley but they would still have many advantages over the faculty of Harvard for reasons discussed in The Wisdom of Crowds. The nature of the Harvard faculty would of course also change (as would the American Economic Association) if it were made into a governing institution, making the random phone-book page more attractive.

  • ChrisA

    The best defense of parternalism I think is that most western countries have a partly parternalistic system, and it works well enought in that these countries are not, by and large, hellish places. If you propose to change a system that works to get something you believe is even better, you should proceed carefully and with small steps.

  • Nick Tarleton

    ChrisA, this ignores the above-cited hypothesis that paternalism in the FDA kills far more people than it saves.

  • ChrisA

    The argument for “inertia paternalism” does not require we keep the FDA or an FDA like entity or say that all patnernalism is good. The argument is that, on balance, overall, we have a working system (not necessarily perfect) and therefore downside risks should be weighted more strongly than upside ones.

    The inertia paternalist would say that we should take each case on a case by case example and determine whether the existing status is on balance positive or not. If not, then they would agree to replace it, perhaps after trying to modify it. But they would not agree to elimination of all paternalism at once, without the case by case debate. This recognises the complex integrated nature of human society and its systems, it may seem like it is a no brainer to eliminate, say, the FDA, but until we do we can’t say whether it is actually a good idea (because of second and third order effects).

  • Stuart Armstrong

    paternalism in the FDA kills far more people than it saves.

    The question is not how many people die because of the FDA, but how many people would die if the FDA did not exist. Would deaths go up or down? (And, from another angle, would people prefer the system with FDA or without?) I can make convincing-sounding arguments either way on both cases, but do we have any evidence to decide?

  • Nick Tarleton

    Stuart, look at Eric Crampton’s comment above. You’ll have to ask him for a source.

  • Chris Salmon

    What a load.