Transcript Lecture II

Lecture II
Comparing the hypotheses of
discovered preferences and
preferences construction
Wrap up of the previous lecture
• Three main reasons for the assumptions of exogenous
preferences: 1) division of labor between economists
and sociologists; 2) axiomatic RCT; 3) after a close
analysis (meta)preferences are stable.
• Raising of the interdisciplinary exchange between
economics and cognitive sciences: are preferences
endogenous?
• Need of a behavioral foundation of economic theory:
Discovered preferences hypothesis (DPH) and
preferences construction hypothesis (PCH).
RCT axiomatic core (preferences)
• Asymmetry axiom: > is an ordering relation: (x > y)  ¬ (y > x)
• Axiom of transitivity: (x > y & y > z) (x > z)
• These two axioms are sufficient for a simple preference ordering on
the chosen acts.
• Continuity axiom: (x > y > z)  [px + (1 – p)z > y > qx + (1 – q)z ]
for any p and q strictly in between 0 and 1
• Independence axiom: for any p such that 0 < p ≤ 1, (x > y)  [px +
(1 – p)z] > [py + (1 – p)z]
RCT axiomatic core (beliefs)
• Given a simple lottery Lk = (pk1,…, pkn), K = 1,…, K and a
probability αk ≥ 0 Ʃk αk = 1
• Compound lottery: (L1,…, Lk; α1,…, αk)
• Risky alternative giving Lk with probability αk for K = 1,…,K
• Reduced lottery (multiplying the probability of each lottery αk – by the probability pnk of the outcome n in lottery Lk and
adding over K)
• pn = α1 pn1 +…, αkpnk
Illustration
Representation theorem
• If an ordering relation (>) satisfies all the
axioms, then it exists a real valued utility
function U(.)
• For all X and Y, X > Y  EU(X) > EU(Y)
• The expected utility EU is given by the sum of
the utilities multiplied by the probabilities of
the outcomes of a lottery: EU = Ʃpi U(xi)
Basic claim of EUT
• An individual whose preferences satisfy all the
axioms of rationality behaves as if (s)he were
maximizing his or her own expected utility
function.
• This assertion
assumptions.
is
devoid
of
any
causal
• It is a systematic description of behavior and not a
causal explanation of behavior.
Empirical content of the theory
• Basic assumption: the theory describes the
behavior of individuals who know which action
best satisfy their preferences.
• The theory abstract away from learning process.
• RCT has empirical content by virtue of
operational criteria for identifying circumstances
in which the learning process can be expected to
be almost at an end (Bruni & Sugden 2007).
Problem of anomalous preferences
• Use of empirical methods coming from
psychology to test the predictive power of the
theory (Kahneman and Tversky 1979).
• Evidence
of
non
(endowment
effect,
reversal).
convergent
behavior
framing,
preferences
• Is the theory to be revised or can anomalous
behavior be accounted for within the same
theoretical framework?
Behavioral foundation
• Causal assumptions on agents’ behavior.
• Substantive
structure.
assumptions
on
preferences
• Substantive assumptions on learning process.
Discovered preferences hypothesis
• Individuals are endowed with a well-defined
structure of preferences hardwired by natural
selection (Plott 1996).
• Individuals discover their “true” preferences
through interactive learning (Binmore 1999).
• Discovered preferences are rational, that is,
axioms of rationality describe an existing
preferences structure.
Learning process
• Learning causally affects temporary preferences. Thus
in the short run preferences are endogenous to choices:
anomalous preferences are a systematic short run
phenomenon.
• There is no causal link between learning and hardwired
preferences. Thus in the long run preferences are
exogenous to choices.
• Economic theory performs well only in repeated
interactions that are simple enough and incentive
compatible (Binmore 1999).
DPH as a conservative behavioral
foundation of RCT
• DPH provides a behavioral foundation which is
coherent with standard comparative statics:
individuals preferences are held constant such that
they result to be consistent across equilibria
(Bruni & Sugden 2007).
• It avoids any reformulation of the theoretical
axiomatic core.
• It not required to open the black box of
preferences formation.
Evidence supporting DPH
•
Consider the WTA/WTP gap in evaluating
an
unpleasant
substance
in
an
experimental setting (Coursey et al. 1987).
• Participants
evaluate
the
consequences of their choices
after a series of iterative trials.
•
Value elicitation through a second price
auction. To elicit WTA the subject with the
lowest bid wins the auction and gets paid
with the second lowest bid. To elicit WTP,
the subject with the highest bid is selected
but he pays the second highest bid.
• Participants learn that truthful
preferences revelation is a weakly
dominant strategy.
•
•
Each subject submits a form expressing
his WTP and WTA.
The mean values of WTP and WTA
converge across auctions after repetitions.
• The erosion of WTP/WTA gap is
interpreted as a support for DPH
(Braga and Starmer 2005).
• Does confirmation of theory
predictions (WTP = WTA) imply
a support for DPH?
Methodological issues on DPH
• Unreasonable restriction of economic domain to
repeated interactions. i.e. Consider the economic
relevance of one shot choices about education or career.
• The hypothesis does not provide a theoretical
explanation of how discovered preferences satisfy
consistency axioms.
• The hypothesis does not provide an objective measure
of rationality: do we maximize a hedonic quantity of
utility?
Epistemological issues on DPH
• Economics does not need any intellectual
integration with cognitive sciences. Anomalous
preferences are eroded by market forces and
economics does not need psychological
explanations for long run phenomena.
• Cognitive sciences can provide some explanation
of anomalous dynamics only in the short-run.
• However anomalous behaviors fall out of the
economic domain.
Preferences construction hypothesis
• Agents are not endowed with a well defined
preferences structure.
• Agents stabilize their behavior by a process of
context-dependent learning.
• Preferences might not be consistent across
equilibria: there are no “anomaly-free” markets.
Learning
• Preferences are learnt and endogenous to choices both in the
short and long run.
• Short run: framing effects, anchoring, endowment effect.
• Long run: shaping effect that consists of an affiliation of
agents’ valuation with market prices even when such
affiliation is excluded by experimental design.
• Shaping effect is relevant as it postulates the existence of
market specific patterns of behavior emerging in logically
equivalent interactions: preferences might not be consistent
across equilibria.
PCH is a non-conservative behavioral
foundation of economic theory
• Standard comparative statics is a necessary but
not sufficient condition to infer individuals’ true
preferences.
• The theoretical prediction might be (in part)
confirmed but this does not necessarily imply a
support for DPH.
• We do need to open the black box of preferences
formation.
Evidence supporting PCH
•
Consider three median price auctions
where people express their WTA money to
drink an unpleasant liquid (Tufano 2009).
•
•
•
•
Participants affiliate their valuations to the
market price emerged in the previous
round of the auction.
The median price is selected as the market
price and the individuals with bids lower
than the market price drink the liquid and
get paid with the market price.
•
Participants learn to correct their choice as
they come to be aware that truthful
bidding is a weakly dominant strategy.
Standard theory prediction: reduction of
the within group error variance and
convergence of the mean bids variance
between the three auctions.
•
Results: reduction of the mean bids
variance within a single auction and
increase of the mean bids variance
between auctions: emergence of marketspecific patterns of behavior.
We observe a process of error reduction
within a single market that does not imply
a convergence of the bids of different
markets towards the same value.
•
The DPH-based prediction of equality of
mean bids between markets is not
confirmed.
Methodological issues on PCH
• It does not impose a restriction of the economic
domain to the range of repeated interactions.
• PCH endorses an instrumental concept of
rationality such as to question if and how real
preferences satisfies the axioms of rationality.
• It aims at an objective measure of rationality as
maximization of a hedonic quantity.
Epistemological issues on PCH
• Economics might develop through an
intellectual integration with cognitive sciences.
• Anomalous preferences
explanatory relevant.
turn
out
to
be
• The hypothesis implies that there are not
“anomaly free” markets.
Conclusions
• The two hypotheses pose a decision problem:
either extending standard RCT through DPH or
re-elaborating standard RCT through PCH.
• The solution of the problem is empirical in
character.
• The
empirical
solution
has
crucial
epistemological implications as PCH and DPH
concerns the status of economic theory: a separate
or dependent science?
References
Binmore, K. (1999). Why experiment in economics?, Economic Journal, vol. 109, pp. F16–24.
Braga J and Starmer C. (2005) Preference Anomalies, Preference Elicitation and the Discovered
Preference Hypothesis, Environmental and Resource Economics Vol. 32, N. 1, 55-89, DOI:
10.1007/s10640-005-6028-0.
Coursey, D.L., Hovis, J.J., Schulze, W.D., 1987. The disparity between willingness to accept and
willingness to pay measures of value. The Quarterly Journal of Economics 102 (3), 679–690.
Kahneman D. & Tversky E., Prospect theory: an analysis of choices under risk, Econometrica,
Vol. 47, No. 2. (Mar., 1979), pp. 263-292.
Plott, C. (1996). Rational individual behaviour in markets and social choice processes: the
discovered preference hypothesis, in (K. J. Arrow et al, eds.), The Rational Foundations of
Economic Behaviour, pp. 225–50, Basingstoke: International Economic Association and
Macmillan.
Sugden R. & Bruni L. (2007) The Road not Taken: How Psychology was Removed from
Economics and How it might be Brought Back. Economic Journal 117: 146-173.
Tufano F. (2009). “Are ‘true’ preferences revealed in repeated markets? An experimental
demonstration of context-dependent valuations.” Experimental Economics, 13, 1–13.