WEEK 1
---------------------------------------------------------------------------------------------------------------------------
• We first largely follow Camerer and Loewenstein (2004);
• We then treat criticisms of behavioural economics by Tirole (2002) and by
Pesendorfer (2004) (Note: parts of Pesendorfer’s paper refers to detailed models;
these parts of his paper will be picked up once we have treated these models
further in the course.)
---------------------------------------------------------------------------------------------------------------------------
Behavioural economics purpose (has become mainstream in recent economic literature):
making the psychological underpinnings of economics more realistic:
- To generate theoretical insights
- Making better predictions
- Make better policy suggestions
Neoclassical theory is useful and completely rejected. And behavioural economics has
become a standard in recent economics.
Bounded rationality models: a smaller group of behavioural models which departs more
radically from the rational paradigm -> the level of sophistication to invest in decision making
should itself be the subject of a cost-benefit analysis -> the effort of making a cost-benefit
analysis should be included in the analysis itself
Most behavioural models depart from the standard neoclassical model only by changing one
or two assumptions -> the assumptions changed do not necessarily contradict rationality,
just the textbook simplified version of it
Criteria for evaluating behavioural economics:
1. Realism: should have realistic psychological underpinnings -> you want something
which can be systematically observed -> e.g. loss aversion -> can be observed in
thousands of experiments
2. Generality: ideally, should have the neoclassical model as a standard case. In this
way, the behavioural can be tested against the standard model -> e.g. when there is no
loss aversion, prospect theory in essenced coincides with expected utility theory -> can
compare now well what effect loss aversion has e.g.
3. Tractability: models should remain tractable, and continue to be applicable in a wide
range of settings. (loss aversion creates a discontinuity in the utility function around the
reference point).
When economics arose as a discipline -> psychology was not a discipline -> only some
psychological principles were included into economic analyses -> in the 20th century,
psychology arose as a discipline, yet not very scientific or measurable, so got erased from
the economic discipline -> group on economists began to publish on so-called anomalies,
refuting expected utility theory and exponential discounting -> then psychology developed as
a scientific discipline -> they used economic models as a benchmark to develop
psychologically-based models:
1. take standard model
2. identify anomaly against predictions of standard model, and rule out non-
psychological explanations
, 3. build alternative theory -> third wave behavioural economics: use new theory to
develop economic models, and test these models
Methodology of behavioural economics: has traditionally been based on laboratory
experiments to make sure no other factors influenced decision making, but this shifted to
field experiments since people were not fully able to link it to real world because it was too
abstract.-> in real life examples, there are a lot of explanations for anomalous behaviour ->
with experiments, such can be excluded
more recently: field experiments -> problem with laboratory experiments is that they
are highly abstract and that participants may not link these to any real-world context
Inductive- build theory that fits empirical data
Deductive- impose axioms which rational decision maker should obey (completeness,
transitivity, continuity) and build from there.
What is the standard model we have been talking about? Utility maximization:
t: period
ct: consumption vector in period t
ut: utility function
It: information in initial period t
d: discount factor, representing that you may care less about the future than about the
present (exponential discounting);
E: expectation. Beliefs are attached to every possible state of the world; expected utility.
At: action to be taken.
Pesendorfer criticisms (1)
Behavioural economics often introduces an extra variable into a model (a reference point in
prospect theory which you use to assess outcomes in terms of gains and losses) -> variable
may be unobservable outside of the laboratory -> empirically, the extra variable constitutes a
free parameter, that is calibrated to fit the data. Yet, a problem is that the way in which the
free parameter (e.g., the reference point) is calibrated, may not be consistent across
applications -> each time you can have plausible explanations for a reference point
(reference point based on past or future e.g.) -> yet for different explanations, different
reference points will be picked -> problem
Pesendorfer criticisms (2)
Typical behavioural economics research is based on an anomaly (a violation of ‘ behaviour
in line with standard theory’) -> > the anomaly is often dealt with that decision makers do
some part of their decision process incorrectly
Anomaly: assuming people make mistakes in their decision process. People over the years
learned to do their best in maximization problems but did not always fully succeed,
behavioural economics changes a specific part of the maximization problem.
Criticism: if people systematically make mistakes when solving maximization problems, why
should they maximize in the first place? -> yet, a justification of the standard model is often
,that it is an ‘as if’ model: people are not literally assumed perform such maximization, but
have somehow learned to behave in their own interest -> when someone is very good at
pooling, he does not know all things about gravity etc (maximization), just has had a lot of
practice and became good at it
Pesendorfer criticism (3)
Behavioural economics often becomes economic psychology, describing how individuals
think or feel in a particular instance.
Criticism: not focused on broad applicability, or on economic implications.
Tirole criticisms (1)
Much of behavioural economics consist of adding additional elements, (such as altruism,
reference point) to standard utility function.
Becomes problematic if we add a new element to the utility function each time we cannot
explain some behavior. Frank (2001): we should not try to explain everything that happens
since some things don’t need to be explained.
Tirole criticisms (2)
Problem of contradictory behavioural models. E.g., altruism and spite; toning down one’s
expectations not to feel disappointed vs boosting up one’s expectations to benefit from
anticipatory utility.
Tirole criticisms (3)
People may not have direct preferences over elements that are added to utility function ->
these elements may be only instrumental in achieving some aim -> need to find underlying
motivation.
Short-term publication strategy: look for new biases, and add new elements to the utility
function (“reduced-form model”);
long-term strategy: look for parsimonious model, that takes into account a larger number of
biases, using a few premises (“structural models”).
Tutorial week 1 Robson, A.J. (2002) Evolution and human nature
- preferences = the rankings over the set of gambles that are based on such beliefs over the
outcomes
- beliefs = the probabilities with which people think various outcomes will occur, conditional
on information that might be available
- rationality = the extent to which people succeed in choosing optimal actions in the light of
their beliefs and preferences and the extent to which their beliefs are appropriately modified
by their information
Put beliefs and preferences together to make rational choice
Biological fitness, the ability to survive to reproductive age, find a mate and produce
offspring (quality and quantity), survival of the fittest.
Gene selection can cause altruism since you try to keep ‘your’ genes as good as possible
Individual fitness instead of group selection since individual know more about their
environment and circumstances while for the group nature mostly has the information but it
cannot tell how to maximize utility since people need to adapt to their environment.
, People compare to others in their environment but not to overall wellbeing since you do not
have to do with that on a daily basis.
