6. Coordination Games Flashcards
What are coordination games?
Represents situations where groups of independent agents have:
-common interest in coordinating their actions
-more than one way to coordinate that, if attained, would be robust against incentives to deviate
Coordination equilibria
Nash equilibria with the extra property that there is no deviation by one player that any player gains from
What is a pure coordination game?
-A game with multiple pure strategy NE. Each is a coordination equilibrium
-Both players are indifferent about how to coordinate
What is a mixed coordination game?
Where there are coordination equilibria but players have opposed preferences over now to coordinate, but they still have a common interest to coordinate
What is a coordination game with Pareto rank able equilibria?
There a more coordination equilibria and both players rank the equilibria in coordination
Describe the Cooper et Al 1992 experiment
Subjects play repeated 2 player games against different players with no communication. Their payoffs are (L8,L8), (L8,R0), (R0,L8), (L10,R10)
What are the results of the cooper et al 1992 experiment?
By the second phase 97% of play is (L8,L8) giving the lower of the two equilibria. This shows that disequilibrium payoffs matter
Describe the Van Huyck, Battalio & Beil 1990 experiment
Weakest link games
-14-16 players in a group
-players simultaneously and independently pick numbers between 1 and 7 which interpreted as effort
-the payoff depends on the lowest amount of effort provided by anyone in the group
What are the results of Van Huyck, Battalio & Beil 1990?
-game is played 10 times in constant groups
-initially 7 is the most common effort but this quickly changes to 1 since everyone receives a lower payoff
Thomas Schelling hypothesis
People can coordinate with higher frequency than 1/k and they do so by following salience
Difference between the primary salience hypothesis and the more subtle salient effects
Primary salience hypothesis says labels make strategies stand out to individuals which lead them to coordinate even when they don’t intend to. More subtle salience effects says players deliberately and reflectively use labels as tools for coordination
Describe the design features of Mehta, Starmer & Sugden 1994
-20 questions asked to individuals in group p and c
-group p are control group and have no incentive to coordinate
-group c get paid if they coordinate with others in their group
What is the coordination index?
The proportion of all possible pairings in which subjects have the same answer
What were the results from Mehta, Starmer & Sugden 1994?
In group p, coordibation index is greater than 1/k which suggests schelling’s hypothesis is correct. In group c, coordination index is far far higher than group p
What was the premise of the Brandts and Cooper experiment 2006?
Is it possible to disturb a bad equilibrium once established, and move to a better one
Describe the Brandts and Cooper 2006 experiment
Subjects grouped in “firms”. Employees choose effort level and they are paid in a weakest link payoff structure. A bonus is applied and changed across rounds to see its effects on effort.
Findings of Brandts and cooper when increasing the bonus and then withdrawing it?
When bonus is increased, effort increases. This isn’t necessarily proportional to the amount the bonus increases by, if just matters that the bonus is increased. When bonus is withdrawn effort levels are maintained at a higher level
What is the premise of the Brandts and cooper follow up paper in 2007?
To see if the results hold up when limited communication is allowed
What are the results of the Brandts and Cooper 2007 paper?
Two way communication and one way communication from a manager both increase effort across the group
Payoff irrelevant
Not concerning consequences or availability of particular strategy profiles
What is the premise of Isoni et Al 2013?
To study whether payoff irrelevant clues help coordination by making some ways of coordinating “salient” or “focal”in bargaining situations with resources to divide
Describe the Isoni et al 2013 experiment
There is a bargaining table where each player has a base and there are 2+ disks scattered on the table. Each player gets points from gaining disks but only if the other player doesn’t also pick the disk. Players must try to coordinate to collect all the disks but not to pick any of the same disks. Initially no communication
What is the proximity rule?
The idea that each player should pick the disks closest to their base
How applicable is the proximity rule to everyday bargaining?
Not as relevant because we can normally communicate. However bargaining context may provide payoff irrelevant contextual cues to coordination and when this happens they facilitate agreement
