3. Theories of Risky Choice Flashcards
When does reference dependence arise?
When preferences are defined on a reference point as well as on consequences experienced
How does status quo bias arise?
From loss averse preferences, if reference point is current situation
Three key things for reference dependence
- Consequences are evaluated as deviations from reference point
- Negative deviations distinguished from positive deviations
- For choice between lotteries faced immediately, assume reference point is given by situation when choice is faced
Assumptions for reference dependence
-zero deviations have zero significance
-losses yield negative utility increasing in absolute size of loss
-gains yield positive utility, increasing in the size of the gain
-diminishing marginal sensitivity to both losses and gains
-loss aversion
What is the reflection effect?
Where gambles are opposite in sign with the same magnitude yet people’s answers differ depending on whether it is a loss or gain
Why does loss aversion suggest that the double mirror image property doesn’t hold?
It imposes an additional restriction beyond reference dependence and diminishing sensitivity to capture particular aversion to losses
Give the definition of a preference relation displaying absolute loss aversion
If for all probabilities 0<P<1/2 and all outcomes x>0
(0,1) } (x, p ; 0, 1-2p ; -x, p)
If the certainty is preferred then loss in the gamble must weigh more heavily than the gain in the gamble does
What is the matching task variant? T&K 1992
Experimenter fixes c (>0) and then subject reveals x such that
(x, 0.5; -c, 0.5) ~ (0,1)
Loss aversion x/c>1
What is a common value for the matching task variant?
2 often taken as typical value but even here ratio declines with c
What is Rabin’s calibration theorem?
EUT is fundamentally miscalibrated as a model of small risk aversion. Reference dependent models can avoid the problem
What is the example Rabin gives with “Johnny”
If Johnny always turns down a 50:50 gamble of losing £10 and winning £11 then he will turn down any bet with a 50% chance of losing £100 no matter how big the upside is
How can reference dependence provide a solution for Rabins micallibration critique
Redefine utility to be a function of gains and losses relative to reference point. Suppose that whenever an agent faces a choice, relevant reference point is wealth at moment of choice. This means the utility relevant to the decision is defined on the lottery prizes not the final total wealth
Give two examples of reference dependence
-where preferences are defined on deviations from a reference point
-loss aversion is a special case in which negative deviations have particular significance
What is double mirror image?
Where loss and gains utility functions are identical- ie there is no loss aversion and lambda is 1 where lambda is the ratio of slopes as x goes to 0
Give behavioural definition of loss aversion
Someone who turns down a 50:50 equal gain: equal loss gamble