multi attribute choice
one must select between 2 or more options that differ in 2 or more attributes
inter-temporal choice
one of the attributes that varies is time
e.g., 10£ today or 25£ in a month
risky choice
one or more of the possible outcomes are probabilistic
sometimes the probabilities are not known precisely, in which case the decision may be referred to as “under certainty” or “under ambiguity”
expected value
the expected value (EV) of an option is the sum of each possible outcome weighted by its probability
EV = P1A1 + P2A2 …
A= value of outcome, P = probability of outcome
n = number of total outcomes
most people are ‘risk averse’
expected utility
- a replacement for expected value
- utility can be thought of as the subjective value of an outcome and is some transformation u(a) of the objective amount
EU = p1u(a1) + p2u(a2) + …. - u is the identity function
- utility function is concave, people have diminishing sensitivity to increasingly large gains
reference dependence
Kahneman and Tversky 1979
- presented 2 decision tasks, one gaining money, one losing
- found people less risk averse when losing money (reversed conditions)
- violates rationality and the expected utility account of decision making
prospect theory - K + T 1979
- posits an S shaped value function which is concave for gains and convex for losses
- defined with respect to reference point
- peole show diminishing sensitivity to progressively large decreases from the reference point, and diminishing sensitivity to progressively larger decreases from the reference point too
the endowment effect
people value an item they already own more that they would be prepared to pay for the same item if they did not own it
demonstrated by Knetsch, 1989,
- used chocolate bars and a coffee mug finding endowment effects for both
economists argue these effects are rational - transactional cost
decision weights
reflects a person’s subjective interpretation of an objective probability
allais’ paradox
certainty effect
peole disproportionately weight outcomes which are guaranteed to occur so that “a reduction of the probability of an outcome by a constant factor has more impact when the outcome was initially certain than when it was merely probable” K+ T
non-linear treatment of probabilities
people seem to overweight extreme probabilites and under-weight moderate to large ones
e.g., Gonzalez and Wu
2 lotteries to win 250, 1 offers a 5% chance to win, the other 30%
A: you can improve the chances of winning first to 10%
B: improve second to 35%
75% of people said A seemed like a more significant change
core components of prospect theory
- an editing stage, where a number of principles are used to simplify the options and ready them for evaluation.
- a reference point is selected which determines whether outcomes are construed as gains or losses, with the value of these outcomes being determined by the S-shaped value function
- The subjective value of outcomes are multiplies with the decision weight transformations of their associated probabilities
valuation vs choice critique - prospect theory
- peoples valuations of 2 gambles can contradict their choices when asked to pick between them
- preferences are constructed by elicitation procedures rather than reflected in people’s responses in those tasks. Imply no stable value function relating objective and subjective value
explanation
- in the task used responses are on the same scale as the rewards/losses offered by the bet, so that aspect of the gamble will dominate
decoy effects critique - prospect theory
asymmetric dominance effect
Ariely, 2009
- describes economist ad
A : 1 year online only 59$
B : 1 year print 125$
C: 1 year both 125$
16% chose A, 84% chose C
when print option removed
68% A, 32% chose c
these context effects are not captured by probability theory
the similarity effect
when the decoy is similar to option A it draws choice share from A and boosts relative preference for B
the compromise effect
when the decoy is more extreme than option A on both dimensions, it boosts choice share for A
empirical issues with prospect theory
- when outcomes have high probabilities people are less risk averse when stakes are low - peanuts effect
- prospect theory cant readily accommodate this, we’d have to assume that the value function is less concave when values are low, but it doesnt really make sense to posit that the subjective value of an amount of money depends on the probability of receiving it
conceptual problems wirh prospect theory - the lack of mechanism
the prospect theory is purely descriptive; it says decisions will be “as if” people combine amounts and probabilities in a particular way, but makes no claims or predictions about the processes
