Week 9-10 - Uncertainty and Information Flashcards
what are two ways for individuals to deal with external risk?
- diversification
- risk pooling
what is diversification?
the spreading of risk; opting for a large number of small risks
what is risk pooling?
spreading risk across multiple players to reduce your share of a large risk
how can a subsidization agreement work?
1) the outcomes for the parties are publicly observable
2) there is a way to enforce the contract
how to create asymmetric information in a game?
manipulate the other player’s information about your abilities and payoffs to affect the game outcome
good info in a strategic interaction?
if others knew this info, they would alter their actions in a way that would inc your payoff
bad info in a strategic interaction?
info that, when revealed, makes other players act in a way that will lower your payoffs
what is signal jamming?
confusing your opponent by convincing them that your info is good not bad
what is an incentive scheme?
a strategy that attempts to influence an unobservable of another player
what is cheap talk?
communicating information that has no associated costs
what is the cheap talk equilibrium?
the equilibrium from rolling back after using cheap talk to align your interests w that of your opponent
signaling in a constant / zero sum game?
players should never give away private info because their preferences are in direct opposition
babbling equilibria?
all communication in zero/constant sum games are disregarded bc players will not think they are credible
how a player should determine whether to believe info given or not?
the player must decide the other’s utility function and see how it interacts with their own
signals if players’ utility functions are perfectly positively correlated?
the player should act as if the signal received is true
signals if players’ utility functions are perfectly negatively correlated?
the player should act as if the signal received is false
if the cost of a signal is directly related to the private info conveyed by the signal?
the signal is evidence that supports how true it is
participation constraints?
n must not be so high that students just decide to choose their alternative
incentive compatibility constraints?
n (number of courses) must be:
a) high enough to disincentivise C-types from trying to pass as A-types
b) low enough not to disincentivise A-types from taking high courses
for participation constraints to work?
players must want to signal and we want them to earn more from signaling than from their alternative option
when does a separating equilibrium exist?
when both participation and incentive-compatibility constraints exist such that players can signal their type only if their info is true
what does the separating equilibrium do?
causes separation of types / self-selection
what do we call it when someone’s payoff is decreased by the cost of signalling?
negative externality
what is the pooling equilibrium?
when the only equilibrium available is one where no signaling is possible and all types are treated alike
in a market with a separating equilibrium, those who do not signal are assumed to have ___ info?
bad information
what are moral hazard problems?
situations in which there is an efficiency loss due to people’s incentives not to consider the utility of other players
what is a game of perfect information?
if only one player moves at a time and if each players knows every action of the other players who moved before her before at every point
when is there a refinement of NE that removes non-credible threats?
when the SPNE matches one of the NEs we found from the game table
what does it mean when the SPNE matches one of the NEs we found from the game table?
there is a refinement of NE that removes non-credible threats
when is something a non-credible threat?
when something is a mutual best response (NE from table) but is not the SPNE
what happens when a player uses a non-credible threat?
they say they are going to play something, but do not (to trick the other player). the strategy that includes the action that is a non-credible threat cannot be trusted.
what does the SPNE do to non-credible threats?
it removes them
what is the equilibrium selection problem?
there are multiple NE and the NE concept is silent on which will be played
