Gedragseconomie
INTRODUCTION
What is behavioural economics?
It is about understanding economic behavior
o Why do we make the choices we do?
It is about testing the standard economic model
o What does it do well and what not so well?
It is about applying insights from lab experiments, psychology and other social sciences
o BE is not new, going back to Adam Smith: “The theory of moral sentiment 1759”
Some economic principals were based on the psychological theory at the time
At the end of the 20th century Vilfredo Pareto argued that economics should break from
psychology and the focus should be on choice rather than desire “if people are rational then
they will reveal their desires through their choices, and so we need to focus only on choice.”
Economics was then dominated by models of rational choice
In the second half of the 21st century behavioral economics made a comeback:
o Herbert Simon questioned the purpose of rational models and proposed the concept
of bounded rationality: Rationality is bounded because there are limits to our
thinking capacity, available information, and time.
Kahneman and Tversky provided evidence that the assumptions of the standard economic
model are limited (cognitive bias, framing effects, reference point)
STANDARD ECONOMIC MODEL
The neo-classical economic model is the way most economists think about consumer
welfare and consumer choice.
Neoclassical economics provides a unified vision of the economy based on some common
rationality assumptions.
This sets economics apart from other social sciences such as psychology, where ‘theories’
describe empirical regularities
Agents are generally assumed to maximize their utility (which may, but must not
necessarily, coincide with their profit)
It is generally assumed to have access to complete information, and to be able to
process such information
Agents are assumed to be fully rational and be driven purely by their self-interest.
✓ People act with full information full external knowledge
✓ People have known preferences full internal knowledge
✓ People choose the best option available rational choices
Theories are usually normative and descriptive at the same time; this may lead to tensions if
they fail descriptively
, Definitions
o Normative theories: tell us how we should behave to obtain a certain goal (usually
utility maximization)
o Descriptive theories: how people do really behave, and may or may not be the same
as the normative theory.
We will talk a lot about risk, i.e. problems in which probabilities are objectively known
(vb. roulette)
A probability is a number between 0 and 1 that indicates a likelihood that a particular
outcome will occur (0 means impossible, 1 means it is certain).
BEHAVIOURAL ECONOMICS
People often tend to satisfice rather than to maximize (Herbert Simon, 1959; this also seems
to be healthier)
Information is not generally available; information about the existence of information may
also not be available
Where information is available, people may not obtain it; they may also not be able to
process it if they obtain it.
Systematic deviations from the self-interested rational agent model exist not only for
individuals, but also for firms
EXPECTED UTILITY
Example:
P (Heads) = 0.50
P (Rain tomorrow) = 0.30
The probability of all possible events must sum to 1:
P(Heads) + P(Tails) = 0.50 + 0.50 = 1
Pr (Rain) + Pr (Cloudy) + Pr (Sunny) = 0.30 + 0.10 + 0.60 = 1
Let’s focus mostly on binary prospects (i.e. lotteries), with two
outcomes x > y and probability p; we can write this as (x,p;y)
Alternatively, choices will be represented using decision trees
We are then interested in preference relations between such
prospects; ∼ indicates indifference and ≻ strict preference (do not confuse this with
inequalities >)
The first theory used to model decision making under risk was expected value theory: EVT
Under EVT the value of a prospect is simply taken to be its mathematical expectation:
𝐸𝑉 (𝑥, 𝑝; 𝑦) = 𝑝𝑥 + (1 – 𝑝) 𝑦
The expected value of a gamble is the value of each possible outcome times the probability of that
outcome.
,EXPECTED VALUE
Example:
You have the option to participate in a game where:
• With a 50% chance (probability p=0.5) you win €100.
• With a 30% chance (probability p=0.3), you win €50.
• With a 20% chance (probability p=0.2), you win nothing (€0).
You want to calculate the expected value of your winnings:
EV(game) = (0.5×100) + (0.3×50) + (0.2×0) = 50 + 15 + 0 = € 65
ST PETERSBURG PARADOX
Consider the following prospect:
A fair coin is tossed. If it comes up heads, you
are paid €2. Then the coin is tossed again. If it
comes up heads again, you are paid €4 = 2^2;
and so on.
When the coin comes up tails, you are paid the
accumulated outcome up to that point (e.g. €2 +
€4 = €6 if tails comes up on the third toss) and
the game ends.
LIMITS OF EVT: ST PETERSBURG PARADOX
✓ If somebody offers you this prospect, how much would you be willing to pay for the possibility to
play it?
✓ How much is the expected value of this bet? Do your preferences conform to expected value
theory?
EXPECTED UTILITY
Expected utility theory was first proposed as a solution to the St. Petersburg paradox by
Daniel Bernoulli in 1738
The paradox is the discrepancy between what people seem willing to pay (WTP) to enter the
game and the infinite expected value.
The simple idea is that one’s WTP for that bet does not need to be equal to infinity if one
subjectively transforms outcomes.
The determination of the value of an item must not be based on the price, but rather on the
utility it yields.
Utility:
Utility refers to the satisfaction or pleasure a person derives from consuming a good, service,
level of wealth.
We can use it to order a person’s preferences over a set of options numerically by assigning
larger numerical values to more preferred options.
If you prefer eating apples (A) to eating chocolate (C), we can assign U(A)=2 and U(C)=1
Choice:
, Under the standard economic model choice is assumed to be revealed preference
Given the alternatives, X and Y, if you choose X then this “reveals” that you prefer X to Y
( X ≻ Y)
We use utility to explain how individuals make choices to maximize their overall happiness or
satisfaction based on their preferences.
Decreasing marginal utility:
Utility increases as consumption increases but at a diminishing rate.
Definition: marginal utility quantifies the added satisfaction that an agent gets from
consuming additional units of goods or services.
In the context of decision-making under uncertainty, utility functions can reflect different
attitudes toward risk: risk aversion, risk-seeking, and risk neutrality.
In these cases, utility function can take different forms.
Formally we can represent the utility as:
where the assumption is that U′(x) > 0 and U ′′(x) < 0, i.e. utility increases in outcomes
(money), but at a decreasing rate.
Bernoulli proposed that u(x) = ln(x) would solve the St. Petersburg paradox
The expected utility of prospect xi is given by:
Note the difference with expected value:
CERTAINTY EQUIVALENT
In order to measure the risk preferences of people; we generally need to elicit some kind of
preference relation
One convenient way is to find the sure amount of money that
makes a decision maker indifferent between playing the prospect
and obtaining that amount:
We call this amount certainty equivalent (CE). We can formally represent the above relation
as follows: 𝑈 (𝐶𝐸) = 𝑝𝑈(𝑥) + (1 – 𝑝) 𝑈(𝑦)
NATURE OF THE STANDARD MODEL
INTRODUCTION
What is behavioural economics?
It is about understanding economic behavior
o Why do we make the choices we do?
It is about testing the standard economic model
o What does it do well and what not so well?
It is about applying insights from lab experiments, psychology and other social sciences
o BE is not new, going back to Adam Smith: “The theory of moral sentiment 1759”
Some economic principals were based on the psychological theory at the time
At the end of the 20th century Vilfredo Pareto argued that economics should break from
psychology and the focus should be on choice rather than desire “if people are rational then
they will reveal their desires through their choices, and so we need to focus only on choice.”
Economics was then dominated by models of rational choice
In the second half of the 21st century behavioral economics made a comeback:
o Herbert Simon questioned the purpose of rational models and proposed the concept
of bounded rationality: Rationality is bounded because there are limits to our
thinking capacity, available information, and time.
Kahneman and Tversky provided evidence that the assumptions of the standard economic
model are limited (cognitive bias, framing effects, reference point)
STANDARD ECONOMIC MODEL
The neo-classical economic model is the way most economists think about consumer
welfare and consumer choice.
Neoclassical economics provides a unified vision of the economy based on some common
rationality assumptions.
This sets economics apart from other social sciences such as psychology, where ‘theories’
describe empirical regularities
Agents are generally assumed to maximize their utility (which may, but must not
necessarily, coincide with their profit)
It is generally assumed to have access to complete information, and to be able to
process such information
Agents are assumed to be fully rational and be driven purely by their self-interest.
✓ People act with full information full external knowledge
✓ People have known preferences full internal knowledge
✓ People choose the best option available rational choices
Theories are usually normative and descriptive at the same time; this may lead to tensions if
they fail descriptively
, Definitions
o Normative theories: tell us how we should behave to obtain a certain goal (usually
utility maximization)
o Descriptive theories: how people do really behave, and may or may not be the same
as the normative theory.
We will talk a lot about risk, i.e. problems in which probabilities are objectively known
(vb. roulette)
A probability is a number between 0 and 1 that indicates a likelihood that a particular
outcome will occur (0 means impossible, 1 means it is certain).
BEHAVIOURAL ECONOMICS
People often tend to satisfice rather than to maximize (Herbert Simon, 1959; this also seems
to be healthier)
Information is not generally available; information about the existence of information may
also not be available
Where information is available, people may not obtain it; they may also not be able to
process it if they obtain it.
Systematic deviations from the self-interested rational agent model exist not only for
individuals, but also for firms
EXPECTED UTILITY
Example:
P (Heads) = 0.50
P (Rain tomorrow) = 0.30
The probability of all possible events must sum to 1:
P(Heads) + P(Tails) = 0.50 + 0.50 = 1
Pr (Rain) + Pr (Cloudy) + Pr (Sunny) = 0.30 + 0.10 + 0.60 = 1
Let’s focus mostly on binary prospects (i.e. lotteries), with two
outcomes x > y and probability p; we can write this as (x,p;y)
Alternatively, choices will be represented using decision trees
We are then interested in preference relations between such
prospects; ∼ indicates indifference and ≻ strict preference (do not confuse this with
inequalities >)
The first theory used to model decision making under risk was expected value theory: EVT
Under EVT the value of a prospect is simply taken to be its mathematical expectation:
𝐸𝑉 (𝑥, 𝑝; 𝑦) = 𝑝𝑥 + (1 – 𝑝) 𝑦
The expected value of a gamble is the value of each possible outcome times the probability of that
outcome.
,EXPECTED VALUE
Example:
You have the option to participate in a game where:
• With a 50% chance (probability p=0.5) you win €100.
• With a 30% chance (probability p=0.3), you win €50.
• With a 20% chance (probability p=0.2), you win nothing (€0).
You want to calculate the expected value of your winnings:
EV(game) = (0.5×100) + (0.3×50) + (0.2×0) = 50 + 15 + 0 = € 65
ST PETERSBURG PARADOX
Consider the following prospect:
A fair coin is tossed. If it comes up heads, you
are paid €2. Then the coin is tossed again. If it
comes up heads again, you are paid €4 = 2^2;
and so on.
When the coin comes up tails, you are paid the
accumulated outcome up to that point (e.g. €2 +
€4 = €6 if tails comes up on the third toss) and
the game ends.
LIMITS OF EVT: ST PETERSBURG PARADOX
✓ If somebody offers you this prospect, how much would you be willing to pay for the possibility to
play it?
✓ How much is the expected value of this bet? Do your preferences conform to expected value
theory?
EXPECTED UTILITY
Expected utility theory was first proposed as a solution to the St. Petersburg paradox by
Daniel Bernoulli in 1738
The paradox is the discrepancy between what people seem willing to pay (WTP) to enter the
game and the infinite expected value.
The simple idea is that one’s WTP for that bet does not need to be equal to infinity if one
subjectively transforms outcomes.
The determination of the value of an item must not be based on the price, but rather on the
utility it yields.
Utility:
Utility refers to the satisfaction or pleasure a person derives from consuming a good, service,
level of wealth.
We can use it to order a person’s preferences over a set of options numerically by assigning
larger numerical values to more preferred options.
If you prefer eating apples (A) to eating chocolate (C), we can assign U(A)=2 and U(C)=1
Choice:
, Under the standard economic model choice is assumed to be revealed preference
Given the alternatives, X and Y, if you choose X then this “reveals” that you prefer X to Y
( X ≻ Y)
We use utility to explain how individuals make choices to maximize their overall happiness or
satisfaction based on their preferences.
Decreasing marginal utility:
Utility increases as consumption increases but at a diminishing rate.
Definition: marginal utility quantifies the added satisfaction that an agent gets from
consuming additional units of goods or services.
In the context of decision-making under uncertainty, utility functions can reflect different
attitudes toward risk: risk aversion, risk-seeking, and risk neutrality.
In these cases, utility function can take different forms.
Formally we can represent the utility as:
where the assumption is that U′(x) > 0 and U ′′(x) < 0, i.e. utility increases in outcomes
(money), but at a decreasing rate.
Bernoulli proposed that u(x) = ln(x) would solve the St. Petersburg paradox
The expected utility of prospect xi is given by:
Note the difference with expected value:
CERTAINTY EQUIVALENT
In order to measure the risk preferences of people; we generally need to elicit some kind of
preference relation
One convenient way is to find the sure amount of money that
makes a decision maker indifferent between playing the prospect
and obtaining that amount:
We call this amount certainty equivalent (CE). We can formally represent the above relation
as follows: 𝑈 (𝐶𝐸) = 𝑝𝑈(𝑥) + (1 – 𝑝) 𝑈(𝑦)
NATURE OF THE STANDARD MODEL
