An Economic Theory of Paternalism

Individuals sometimes make bad choices.

As a result, there are cases where an outcome would have been improved with a bit of paternalism.

Paternalism means I think I know better than you and I am going to try to make you do things my way. My efforts might manifest through coercion or persuasion. Both are paternalistic.

If paternalism is a good then some quantity of it belongs to the optimal stock of human behavior. This article will argue that paternalism is a contingent social good and therefore an economic good.

There is a minority view in economics which is that individuals truly are rational agents. We can call this Naive Capitalism.

An individual never makes a mistake under this view. If an individual did something apparently moronic, they merely had a preference for it. If two men find themselves in a car wreck it was either that they both demanded a car wreck, or else that one of them demanded it more than the other was willing to pay to avoid it. Or that there were some dubious externalities and Coase was absent.

The mainline view is that individuals can be modeled as rational agents, but that this is a simplifying assumption we know to be false.

We construct a simplified model as a baseline, compare it to reality, and observe patterns of deviation in order to continually improve our models.

Behavioral economics is a key field which identifies many of these deviations, mainly at the individual level.

Fama‘s efficient market hypothesis is a key alternative to behavioral economics.

The efficient market hypothesis deals with the market. Markets are importantly better at information processing compared to an individual.

Tangentially, I do have a solution to the Joint Hypothesis Problem. We should expect that both the model and the market are malformed to some degree. That is a straightforward application of risk minimization. In the absence of any information they should each be expected to share 50% of the burden.

On to my real criticism. Essentially, I disagree that individuals are omniscient.

Naive Capitalism is incompatible with the Human Condition. The human condition may be defined as the imperfect ability to ascertain, retain, process, and apply information.

Naive Capitalism describes a theoretical best-case economy. This is useful for benchmarking but without further aid it is not useful for individual or social prescription.

One consequence of this view, which is notable in the libertarian paradigm, is that no individual can be said to have a malformed preference. So there is never any gains to be had by coercion. We can never force someone to do something against their will and find them better for it.

Coercion is a strict loss for welfare under this view. Even evangelism is somewhere between wasteful and conspicuous.

I have no issue with individuals as rational agents in that humans are teleological and in other ways. I assert that individuals do not have complete knowledge of their own preferences across time and often even at some point in time. Diminishing marginal gains from complexity and additional choices

I think that individual fallibility creates scenarios where paternalism is necessary for efficiency in a straightforward way.

So far:

  1. There is a minority view in economics which is that individuals truly are rational agents.
  2. Such a view is wrong.
  3. Even if that view was not wrong, it would be an invalid application to consequently rule out paternalistic efficiency in all microeconomic cases.
  4. Theoretical and empirical cases can be constructed and observed which demonstrate paternalism as a mechanism for superior outcomes in all the traditional senses
    1. Utility
    2. Quantity produced
    3. Technical efficiency
    4. Growth rates

Some other reasons we shouldn’t expect individuals to obtain optimal choices without paternalism, and/or paternalism is/can be a social good:

  1. TCE allows a macroeconomic situation in which a central authoritarian regime can be efficient through an organizational structure through reduced transaction costs.
    1. At least in theory, central or hierarchical control can reduce transaction costs to such a degree that social surplus can increase compared to the free market arrangement.
    2. This is perhaps more plausible at a micro level.
  2. Behavioral economics in general and Nudge Theory in particular give empirical evidence for some cases where paternalism is efficient.
  3. If information is systematically disbursed in society, instead of randomly disbursed, then it follows simply from individual fallibility and the human condition that outcomes are not ensured to be efficient at the individual level. Reasons to think information is not randomly disbursed:
    1. Education systems are constructed systematically and they distribute information systematically
    2. Those with access to lucrative information are incentivized to prevent the dissemination of that information
    3. Rational ignorance, rational irrationality, and motivated search are all reasons that an individual won’t obtain information in a non-biased way
  4. Unexpected events and technological progress are at odds with an omnipotent rational agent.
  5. The existence of wicked problems is at odds with an omnipotent rational agent.
  6. The Bible advises parents to teach and discipline their children.
    1. Child rearing gurus sometimes advise disciplining children.
    2. Anecdotes about times where I’m retrospectively glad my parents disciplined me, although I would not have preferred it at the time.
    3. The thought experiment where a young child runs into a street and the parent yanks the child away from an oncoming car, even though some call that coercion.
  7. Everyone wants better education, but education presupposes an imperfectly informed student and a better-knowing teacher.
  8. Complexity and efficiency have an endogenous relationship.
    1. More complex information is more costly to obtain
    2. More complex decisions are harder for people to make
  9. An argument from Doritos as given below.
    1. Consider that a person wants to lose weight so they stop eating Doritos and drinking Coke.
    2. This person then begins to eat veggie chips with ranch dip and drinking diet coke.
    3. The person thinks they are now healthier, but in reality they aren’t. They gain weight and other ailments not out of preference, but due primarily to misinformation.
    4. Did this person really have a revealed preference for bad health? No. This was a normal person, not a theoretical construct.
  10. Preference aren’t really stable
    1. Contra Stigler and Becker, De Gustibus Non Est Disputandum.
    2. Pro Caplan.
    3. Consider a person who is emotionally volatile. This person went through a recent break up. This has led to periodic depression and binge drinking.
    4. Or the well documented way taste literally changes with age
      1. Taste buds can be trained. Here is an MD arguing for paternalism on this fact alone.
      2. Senses overall weaken with age, which has important implications for preferences.
    5. Or the way preferences plausibly change with the introduction of systematic information changes
    6. Instead, individual preferences are in a constant state of fluctuation. Sometimes minor, sometimes major, and usually due to as few as 5 factors:
      1. Factors of mental state
        1. Biochemical / Emotional
        2. Information obtained and the structure of beliefs
        3. Heuristics and cognitive biases
      2. Physical state, other than mental state
        1. Disability, illness, hunger, etc
        2. External physical state / the environment.

Two interesting measurements are factor elasticity of supply and demand. If I am a bit more emotional, does this significantly alter my demand for some good or my willingness to work?

Some emotions might boost productivity, others would detract.

To conclude:

  1. Sometimes dad really does know better than you.
  2. Sometimes the police do good things.
  3. Sometimes you should listen to them.
  4. Sometimes society is better off when they punish you for not listening to them.
  5. Sometimes the above 4 do not hold, but you are sometimes not the best judge of that.

Further lines of inquiry:

  1. Given that agent X knows better than Y, is X more or less disposed to paternalism?
    1. Related: Experts tend to overvalue their own expertise
    2. Police have a higher rate of spousal abuse
  2. Given that agent X knows better than Y, what are the factors of Y’s receptiveness?
  3. Given that paternalism is sometimes ideal, is it normally ideal?
    1. Hint: No, and this is the key to being a sophisticated free market proponent that can pass an Ideological Turing Test.
    2. The benefits need to outweigh the costs: Paternalism and coercion are often expensive. Coercive change management strategies don’t always optimize adoption.
    3. Speaking of being a sophisticated free market proponent, consider the terms polycentrist vs anarcho-capitalist
    4. Paternalist bidding wars: Should vigilantism ever be legal? (cough the free market)
    5. If it’s not normally ideal, what discriminators exist?
    6. If it’s not normally ideal, what systems allow for the most frequent coincidence of ideal factors? (cough the free market)

1 thought on “An Economic Theory of Paternalism”

Leave a Comment