Discussion of Blume, Cogley, Easley, Sargent, and Tsyrennikov “Welfare, Paternalism and Market Incompleteness” Jonathan A. Parker, July 2013.

Discussion of Blume, Cogley, Easley,

Sargent, and Tsyrennikov

“Welfare, Paternalism and

Market Incompleteness”

Jonathan A. Parker, July 2013

Key ingredients• Exchange economy with finite-state Markovian

uncertainty: P0• Endowments: ei Bounded aggregate endowment• Identical, time-separable choice functions u

satisfying Inada conditions• Dogmatic heterogeneous beliefs: Pi• Different market structures: – complete markets – bonds only– complete markets with borrowing constraint– complete markets with transaction tax

Welfare

• Individual objective

• Individual welfare

• Social welfare

Results

ResultsNice examples: different belief heterogeneity and true probabilities and different market structures.1.Variance and trends in consumption 2.Drift in consumption3.Initial wealth -- Pessimism and patience4.Complete markets leads to immiseration, so other markets structures can do better5.Sometimes all agents prefer one market structure and welfare criterion prefers another

ResultsTheorems:1.Welfare maximized by no betting on sunspots2.Autarky may dominate complete markets3.With symmetric possible disagreement and patient enough agents, autarky is preferred to complete markets.

1. What about alternative welfare measures?

1. Bayesian all the way2. Brunnermeier, Simsek, Xiong (see also Davila

(Harvard))

My reading is that the qualitative theoretical points would be similar. But not clear for:

3. Counting anticipatory utility in welfare or• Agents that learn from mistakes in long run

2. Can we use this framework in practice?

Two answers.1.Yes, as is.2.Only by relying on data on beliefs and wellbeing to provide a social scientific basis for normative economics (a critique of all welfare economics also)

A quick refresher on theoretical content of revealed preference theory (RPT)

•“Theory” is really philosophy – Actual theory (testable content) only for hypothetical

choices– In practice: functional form assumptions

•Static welfare– Assume choice functions equal welfare functions

•Intertemporal welfare – Caplin-Leahy: no RPT way to measure a correct discount

rate for intertemporal welfare– Typically all Pareto weight on time-0 choices

So we have no scientific basis for intertemporal social welfare functions (yet)

Welfare depends on preferences vs. beliefs

– Assume choice functions do not maximize welfare• So what do we assume?• E.g. care about the future today?

– Beliefs vs. preferences• Identical immiseration possible from preferences or from beliefs

• how do we interpret observed immiseration?• does the source even matter?

Solution:- Survey measures of beliefs (RP impossible)- Survey measures of happiness- => Social scientific foundation for welfare

3. Caveats to the paper’s lessons

A. Paper: “Belief diversity is a fact of life”– Can’t disagree– Dogmatic: important for long run – no feedback– Elimination of beliefs is elimination of a person

• But maybe some tiny bit of learning

B. Preferences: Kogan, Ross, Wang, Westerfield– In frictionless, complete-market exchange economies,

it is false in general that traders with inferior forecasts do not survive

– E.g. patient optimists or (priced) risk-loving optimists

D. Idiosyncratic risk taking evolutionarily optimal for population• Robson (see also Brennan and Lo)

E. Is an equilibrium with immiseration necessarily worse for even this welfare criterion?

F. Is the paper about financial regulation or welfare state?– human capital inalienable