Theory of Decision Making under Uncertainty

Theory of Decision Making under
Uncertainty
Based on papers by Itzhak Gilboa, Massimo Marinacci, Andy
Postlewaite, and David Schmeidler
IDC Herzliya
Dec 29, 2013
Risk and Uncertainty
I
Dual use of probability: empirical frequencies in games of
chance and a subjective tool to quantify beliefs
Risk and Uncertainty
I
Dual use of probability: empirical frequencies in games of
chance and a subjective tool to quantify beliefs
I
Dates back to Pascal and Leibniz (cf. Pascal’s Wager)
Risk and Uncertainty
I
Dual use of probability: empirical frequencies in games of
chance and a subjective tool to quantify beliefs
I
Dates back to Pascal and Leibniz (cf. Pascal’s Wager)
I
1921 Knight , (Keynes) – risk, uncertainty
Risk and Uncertainty
I
Dual use of probability: empirical frequencies in games of
chance and a subjective tool to quantify beliefs
I
Dates back to Pascal and Leibniz (cf. Pascal’s Wager)
I
1921 Knight , (Keynes) – risk, uncertainty
I
1931 Ramsey, de Finetti – subjective probability
Risk and Uncertainty
I
Dual use of probability: empirical frequencies in games of
chance and a subjective tool to quantify beliefs
I
Dates back to Pascal and Leibniz (cf. Pascal’s Wager)
I
1921 Knight , (Keynes) – risk, uncertainty
I
1931 Ramsey, de Finetti – subjective probability
I
1954 Savage "The crowning glory"
Risk and Uncertainty
I
Dual use of probability: empirical frequencies in games of
chance and a subjective tool to quantify beliefs
I
Dates back to Pascal and Leibniz (cf. Pascal’s Wager)
I
1921 Knight , (Keynes) – risk, uncertainty
I
1931 Ramsey, de Finetti – subjective probability
I
1954 Savage "The crowning glory"
I
Uncertainty = or Knightian Uncertainty
Risk and Uncertainty
I
Dual use of probability: empirical frequencies in games of
chance and a subjective tool to quantify beliefs
I
Dates back to Pascal and Leibniz (cf. Pascal’s Wager)
I
1921 Knight , (Keynes) – risk, uncertainty
I
1931 Ramsey, de Finetti – subjective probability
I
1954 Savage "The crowning glory"
I
Uncertainty = or Knightian Uncertainty
I
Objectivity – interpersonal concept, convincing others
Digression
I
For which audience(s) the economic theorists writes?
Digression
I
For which audience(s) the economic theorists writes?
I
Economics is an empirical science
Digression
I
For which audience(s) the economic theorists writes?
I
Economics is an empirical science
I
Convince, persuade, rhetoric
Digression
I
For which audience(s) the economic theorists writes?
I
Economics is an empirical science
I
Convince, persuade, rhetoric
I
Axioms
theory
from Classical Greece to contemporary decision
Digression
I
For which audience(s) the economic theorists writes?
I
Economics is an empirical science
I
Convince, persuade, rhetoric
I
Axioms
theory
I
von Neumann and Morgenstein, Savage
from Classical Greece to contemporary decision
Rationality
I
Economic decision is rational if it optimizes the agent’s
preferences,
Rationality
I
Economic decision is rational if it optimizes the agent’s
preferences,
I
As long as the preferences are consistent
Rationality
I
Economic decision is rational if it optimizes the agent’s
preferences,
I
As long as the preferences are consistent
I
De gustibus non est disputandum
Rationality
I
Economic decision is rational if it optimizes the agent’s
preferences,
I
As long as the preferences are consistent
I
De gustibus non est disputandum
I
In case of risk or uncertainty the agent should maximize
expected utility with respect to the known or subjective
probability
Rationality
I
Economic decision is rational if it optimizes the agent’s
preferences,
I
As long as the preferences are consistent
I
De gustibus non est disputandum
I
In case of risk or uncertainty the agent should maximize
expected utility with respect to the known or subjective
probability
I
This is the accepted view of economic theory
Rationality
I
Economic decision is rational if it optimizes the agent’s
preferences,
I
As long as the preferences are consistent
I
De gustibus non est disputandum
I
In case of risk or uncertainty the agent should maximize
expected utility with respect to the known or subjective
probability
I
This is the accepted view of economic theory
I
or majority of economic theorists and game theorists.
Rationality and Objectivity
I
The de…nition we use:
Rationality and Objectivity
I
The de…nition we use:
I
A mode of behavior is irrational for a given decision maker, if,
when the decision maker behaves in this mode, and is then
exposed to the analysis of her behavior, she regrets it (feels
embarrassed).
Rationality and Objectivity
I
The de…nition we use:
I
A mode of behavior is irrational for a given decision maker, if,
when the decision maker behaves in this mode, and is then
exposed to the analysis of her behavior, she regrets it (feels
embarrassed).
I
In other words, an act is rational (or objectively rational) if
the decision maker can convince others that she optimized her
goals.
Rationality and Objectivity
I
The de…nition we use:
I
A mode of behavior is irrational for a given decision maker, if,
when the decision maker behaves in this mode, and is then
exposed to the analysis of her behavior, she regrets it (feels
embarrassed).
I
In other words, an act is rational (or objectively rational) if
the decision maker can convince others that she optimized her
goals.
I
Like Objectivity this is an interpersonal concept – convincing
others
Rationality and Objectivity
I
The de…nition we use:
I
A mode of behavior is irrational for a given decision maker, if,
when the decision maker behaves in this mode, and is then
exposed to the analysis of her behavior, she regrets it (feels
embarrassed).
I
In other words, an act is rational (or objectively rational) if
the decision maker can convince others that she optimized her
goals.
I
Like Objectivity this is an interpersonal concept – convincing
others
I
An act is subjectively rational if the decision maker can not be
convinced by others that she failed to optimize her goals.
The Bayesian approach
I
Four tenets of Bayesianism in economic theory
The Bayesian approach
I
Four tenets of Bayesianism in economic theory
I
Formulation of a state space, where each state “resolves all
uncertainty”
The Bayesian approach
I
Four tenets of Bayesianism in economic theory
I
Formulation of a state space, where each state “resolves all
uncertainty”
I
Prior Probability: (i) Whenever a fact is not known, one
should have probabilistic beliefs about its possible values.
The Bayesian approach
I
Four tenets of Bayesianism in economic theory
I
Formulation of a state space, where each state “resolves all
uncertainty”
I
Prior Probability: (i) Whenever a fact is not known, one
should have probabilistic beliefs about its possible values.
I
(ii) These beliefs should be given by a single probability
measure de…ned over the state space
The Bayesian approach
I
Four tenets of Bayesianism in economic theory
I
Formulation of a state space, where each state “resolves all
uncertainty”
I
Prior Probability: (i) Whenever a fact is not known, one
should have probabilistic beliefs about its possible values.
I
(ii) These beliefs should be given by a single probability
measure de…ned over the state space
I
Updating of the prior according to Bayes rule
The Bayesian approach
I
Four tenets of Bayesianism in economic theory
I
Formulation of a state space, where each state “resolves all
uncertainty”
I
Prior Probability: (i) Whenever a fact is not known, one
should have probabilistic beliefs about its possible values.
I
(ii) These beliefs should be given by a single probability
measure de…ned over the state space
I
Updating of the prior according to Bayes rule
I
When facing a decision problem, one should maximize
expected utility
The Bayesian approach
I
Four tenets of Bayesianism in economic theory
I
Formulation of a state space, where each state “resolves all
uncertainty”
I
Prior Probability: (i) Whenever a fact is not known, one
should have probabilistic beliefs about its possible values.
I
(ii) These beliefs should be given by a single probability
measure de…ned over the state space
I
Updating of the prior according to Bayes rule
I
When facing a decision problem, one should maximize
expected utility
I
(ii)* Sometimes the prior is posited on the consequences.
Background
I
Undoubtedly, the Bayesian approach is immensely powerful
and successful
Background
I
Undoubtedly, the Bayesian approach is immensely powerful
and successful
I
It is very good at representing knowledge, belief, and intuition
Indeed, it is a …rst rate tool to reason about uncertainty
(cf. “paradoxes”)
Background
I
Undoubtedly, the Bayesian approach is immensely powerful
and successful
I
It is very good at representing knowledge, belief, and intuition
Indeed, it is a …rst rate tool to reason about uncertainty
(cf. “paradoxes”)
I
Used in statistics, machine learning and computer science,
philosophy (mostly of science), and econometrics...
Background
I
Undoubtedly, the Bayesian approach is immensely powerful
and successful
I
It is very good at representing knowledge, belief, and intuition
Indeed, it is a …rst rate tool to reason about uncertainty
(cf. “paradoxes”)
I
Used in statistics, machine learning and computer science,
philosophy (mostly of science), and econometrics...
I
However, in most of these, only when the prior is known.
Background
I
Undoubtedly, the Bayesian approach is immensely powerful
and successful
I
It is very good at representing knowledge, belief, and intuition
Indeed, it is a …rst rate tool to reason about uncertainty
(cf. “paradoxes”)
I
Used in statistics, machine learning and computer science,
philosophy (mostly of science), and econometrics...
I
However, in most of these, only when the prior is known.
I
Typically, for a restricted state space where the set of
parameters does not grow with the database
Background
I
Undoubtedly, the Bayesian approach is immensely powerful
and successful
I
It is very good at representing knowledge, belief, and intuition
Indeed, it is a …rst rate tool to reason about uncertainty
(cf. “paradoxes”)
I
Used in statistics, machine learning and computer science,
philosophy (mostly of science), and econometrics...
I
However, in most of these, only when the prior is known.
I
Typically, for a restricted state space where the set of
parameters does not grow with the database
I
By contrast, in economics, it has been applied to very large
spaces
Non-Bayesian decisions
A
I
a
b
c
7
0
3
B = AC
0
7
3
Ellsberg’s Paradox
I
One urn contains 50 black and 50 red balls
I
Another contains 100 balls, each black or red
I
Do you prefer a bet on the known or the unknown urn?
I
Many prefer the known probabilities. People often prefer
known to unknown probabilities
I
This is inconsistent with the Bayesian approach
Ellsberg’s Paradox
I
One urn contains 50 black and 50 red balls
I
Another contains 100 balls, each black or red
I
Do you prefer a bet on the known or the unknown urn?
I
Many prefer the known probabilities. People often prefer
known to unknown probabilities
I
This is inconsistent with the Bayesian approach
I
Still, many insist on this choice even when the inconsistency
and Savage’s axioms are explained to them
Symmetry and Reality
I
Ellsberg’s paradox may be misleading
If one wishes to be Bayesian, it is easy to adopt a prior in
this example (due to symmetry)
Symmetry and Reality
I
Ellsberg’s paradox may be misleading
If one wishes to be Bayesian, it is easy to adopt a prior in
this example (due to symmetry)
I
But this is not the case in real life examples of wars, stock
market crashes, etc.
Symmetry and Reality
I
Ellsberg’s paradox may be misleading
If one wishes to be Bayesian, it is easy to adopt a prior in
this example (due to symmetry)
I
But this is not the case in real life examples of wars, stock
market crashes, etc.
I
Indeed, my critique was based on the cognitive implausibility
of the Bayesian approach, and not on the results of an
experiment