BPP Political Science

From edegan.com
Jump to navigation Jump to search

This page is devoted to Rui de Figueiredo's section of the political science class. For Ernesto Dal Bo's section, see his class slides on the course page.

Lookup Table

The following lookup table references concepts in papers in de Figueiredo's section.

                                A&D  AS&F  B(VR)  B(2P1A)  B&D  B&F  dF  F  F&R  G&K  T
Political Science Attributes:                      
Recognition rule                                                 Y    Y  M      
Amendment rule (Open or Closed)                                  Y                Y  
Specialization                                                                    Y   Y
Vote Buying                                  Y     Y              
Activist                              M                     Y            
Violation                             Y                     M            
Agenda power                                 M     M             Y          
Action/Policy Space:                      
Uni-dimensional action                       Y     Y        Y    Y       Y        Y   Y
Multi-dimensional action              Y        
Game Type:                      
Strategic Form                                     Y        Y            Y      
Extensive Form                        Y      Y                   Y    Y       Y   Y   Y
Continuous time                  Y                    
Repeated game                    Y                               Y    Y  M            M
Game Theory Concepts:                     
War of attrition                 Y                                       Y      
Bargaining                                                       Y    Y  Y      
Indivisible                      M                                    Y  Y    M       M
Types                                 Y                     Y         Y    
Cooperation                                                 M         Y        
Reputation                            M                               Y            
Insulation                                                            Y               M
Action                                M                               Y  Y    Y   Y   Y
Outcome                          Y    M           Y         M         Y  Y    Y       Y
Information/Cost assumptions:                      
Private information              Y    Y                     Y            Y    Y  
Common knowledge                             Y    Y              Y    Y               Y
Costly information                                                            M   Y  
Unequal costs                    Y    M                                  M    Y    
Private costs                    Y    M                                  Y    Y    

Note that Y indicates Yes, and M indicates Maybe (or somewhat).

Paper Summaries

Alesina and Drazen (1991)


  • war of attrition
  • continuous time
  • rational delay
  • concession function


  • exogenous shock
  • private knowledge about costs/gains during instability (drawn from a common distribution)
  • disproportionate burden after stablization (exogenous and not subject to bargaining)

Story: Constant output through out (GDP), but ratio (Debt/GDP) changes. Constant Ratio -> Shock -> Rising (or falling) Ratio -> Increasing pain -> Differential willingness to bear pain and that concession result in disproportionate burden, leads to different concession points Private knowledge about willingness to bear pain => Don't know other parties willingness. Willingess 'revealed' by time. Disproportionality of burden is important - equal burden results in instant concession. Extremely unequal burden increases time to concession.

Austen-Smith and Feddersen (2008)


  • Whistleblowing
  • Violations
  • Fix/Don't Fix
  • Penalties
  • Extensions:
    • Seperating equilibrium
    • Sorting


  • Violations observed by employees only (and probabilistically by nature)
  • Ordering of probabilities of observation (whistleblow,private,silent): [math]q_w = 1 \gt q_p v \gt q_{\phi} v \gt 0\,[/math]
  • The probability of a violation becoming known is increasing in the size of the violation

Story: Violation occurs -> Employees can stay silent, report internally or whistleblow -> Managers (if advised of violation) can fix or not -> Society observes with some probability (dependent on the action of the employee) -> Firm and Society have payoffs -> Employees and managers care about both firm and society to differing degrees depending on their type. The managers are able to give penalties to incentivize certain actions by the employees, and do so according to their type (and the employees actions).

Baron (2001)

Vote Recruitment


  • Ideal Points + Compensation in utility (i.e. bribery, corruption, side payments)
  • Status Quo
  • Agenda Setting
  • Extreme/Moderate
  • Vote Buying


  • Common knowledge of ideal points (of Legislators)
  • Single dimension of contention (with single peaked prefs)
  • Different preferences between Interest and Legislators

Story: Interest sets agenda (chosen policy against status quo) or it is exogenous -> Legislators have additively seperable utility over preferences (with intensity) and money -> Interests buy pivotal legislators making them indifferent to chosen policy with money (or don't bother if they are not extreme) ->

Two Interests with an Executive (2 principals, 1 agent)


  • 2 Principals, 1 Agent
  • Costly transfers
  • Rent Extraction
  • Menu Offers
  • Prisoner's Dilemma
  • Extension:
  • Capacity constraint


  • Differing preferences (between P's, and wrt A).
  • Additively seperable prefs
  • Linear contribution schedules
  • Local Truth Telling (used as a solution concept)
  • Feasible transfers

Story: Principals have opposing interests and money -> Agent has intermediate interests and wants money -> Principals must use agent to enact policy -> Principals make transfers to agent to get as close to their ideal point as their money cost will allow -> Opposing principals means that actual policy is a compromise and close to Agent's anyway -> Principals are stuck in a prisoner's dilemma.

Solution concept: Use local truth telling and linear contribution schedules. Solve by maximizing joint surplus to determine equilibrium outcome, and contrast with maximizing joint surplus without one principal to determine that principals needed transfer.

Baron and Diermeier (2007)


  • Activist
  • Recalcitrant
  • Campaign
  • Reward/Harm
  • Demanded Change
  • Concession
  • Target Selection
  • Self-regulation
  • Reputation
  • Contesting a campaign
  • Extensions: Committment not to act opportunistically


  • Exogenous probability of soft/hard target
  • Additively seperable harm and rewards
  • Concave firm profit function with negative slope, and strictly increasing activist's utility function

Story: Activist targets firm and makes demand, promising reward and harm -> Firm is either strategic or recalcitrant -> If expected gain is positive activist conducts campaign -> firm gets harmed or rewarded according to actions. With commitment not to act -> Firms may take preempive action and self-regulate to avoid harm. With reputations -> soft (responsive firms) may play hard to deter activists, but then activists are more aggressive to hard firms. The firm may also contest the campaign.

Baron and Ferejohn (1989)


  • Bargaining
  • Split the pie/dollar
  • n members
  • Recognition rule
  • Amendment rule: Open or Closed rule
  • Simple majority
  • Unanimity
  • yes/no voting
  • 2 period
  • Infinite periods
  • Agenda power
  • First mover advantage
  • Lowest continuation value partners selected
  • Randomly choosen partner


  • Risk neutral
  • Exogneous recognition probabilities
  • No allocation as status quo
  • Common discount factor
  • Stationary Equilibrium

Solution concept:

  • Mixed Strategies
  • Subgame perfect
  • Continuation value

Story: Play split the pie with n (odd) members -> Strategies and outcomes depend on voting rule (e.g. majority), amendment rule (open or closed) and number of periods (and players) -> under the closed rule, 2 periods, majority rule -> agenda power gives first mover advantage -> recruit cheapest votes in terms of recognition probabilities (lowest) -> in terms of ex-ante expected utility those with higher recognition probabilities will have the lowest ex-ante value for the game -> with equal probability can randomly choose partners.

de Figueiredo and Edwards (2007)

Empirical paper - test of Baron (2001) - Two Interests with an Executive (2 principals, 1 agent).

de Figueiredo (2002)


  • Reciprocity
  • Insulation
  • Multi-dimensional policy
  • Cooperation
  • Infinitely repeated game
  • Political uncertainty
  • Reelection
  • Exogenous probability of moving each turn
  • Value of policy continuity
  • Risk aversion
  • Sustaining cooperation
  • Inefficient action

Story: In Reciprocity Game: Can implement own program -> If other players program in place, can leave or overturn -> payoffs to leaving are less than to own program alone (so conflict) -> Each turn a player is choosen to move with the same probability -> Cooperation can be sustained if payoffs are high enough.

In Insulation Game: Same as above but can insulate and recieve a reduced payoff for all time -> Can trade benefits in power for benefits out of power -> more uncertainty over who will move next leads to more insulation if costs of insulating are not too high. Insulating is inefficient.

Fearon (1994)


  • War of attrition
  • Rationally inefficient
  • Split the pie
  • Private costs
  • Preemptive action
  • Commitment problem


  • Private knowledge of costs/benefits for at least one party
  • Random outcome in 'war' state is common knowledge (or is private knowledge, but then costs can be common knowledge)
  • Risk neutral (works with risk aversion)
  • Continuous range of settlements
  • Private incentives to misrepresent information! (some costly signals can solve this)

Story(for the private info version): Status quo to start -> A makes unilateral choice of outcome -> B can acquiese or go to war -> if acquiese end, if war, outcome decided by common knowledge random outcome p and parties pay costs -> A wants to push B back to reservation level (p+c) but doesn't know c (or p) -> faces trade off of more territory against greater risk and makes it -> positive risk of war.

Fernandez and Rodrik (1991)


  • Inefficient outcomes
  • Majority rule
  • Ex ante uncertainty over whether a winner or loser
  • Reforms rejected under uncertainty (but with risk neutrality), even though accepted under certainty
  • Reforms overturned
  • Status quo bias


  • Risk neutral (though risk aversion makes it stronger)
  • Majority rule
  • Uncertainty over who will benefit
  • Ex ante workers are identical and atomistic
  • Two sectors, with labour and goods
  • Cost of relocating made up of general cost and sector specific cost

Story: Two sectors of the economy -> Reform will make one sector's workers better off and the other worse off, but overall a majority better off -> individuals don't know their cost of switching draws and so don't know if they can switch -> worker in the sector that is made worse off may vote against the reform even though they could pay the switching cost and be made better off by the reform.

Gilligan and Krehbiel (1987)


  • Specialization
  • Commitee/Floor
  • Efficient
  • Costly information about the state of the world
  • Open rule/Closed rule
  • Beliefs (perfect Bayesian)
  • Inefficient overspecialization
  • Endogenous specialization
  • Endogenous choice of proceedure


  • Outcome is linear in policy and state of the world
  • Utilities are negative quadratic about ideal points
  • The informed party (the committee) has an ideal point greater than the decision making party (the floor)

Story: One party can specialize at a cost and learn the true state of the world -> Another party (with different preferences) makes the decisions and set the proceedural rules (open vs. closed) -> Supposing open and no specialization, then no information is learnt and an expectation is used as a posterior; Supposing open and specialization then partitions can be used to communicate some information as in a cheap talk model; Supposing closed and no specialization, the committee can force the agenda relative to the status quo; and for closed with specialization there is perfect communications for extreme values and noisy signalling for non-extreme values. Over specialization is possible.

Ting (2009)


  • Specialization
  • Fungible investments - Generalized capacity
  • Specialized capacity
  • Inefficient overinvestment


  • Complete and perfect information
  • Two stages, solved by SPNE
  • Particular assumptions on the utility functions cross partials - see the paper

Story: There are two versions of the game, the generalized capacity version and the specialized capacity version. The difference is solely in whether the capacity can be appropriated by the decision maker for any policy, or is specifically bound the policy announced by the agent. In the GC game: The agent chooses a capacity -> the principal chooses a period 1 policy, and then a period 2 policy and capacity that can't exceed the period 1 capacity -> Agents incur a cost for specializing and principals and agents both get utility from policies close to their ideal points -> In the GC game the principal appropriates the capacity for his own purposes, in the SC game the agent strategically invests in certain policies to get the principal to use them (and benefit from the capacity) rather than throw them out and use his own capacity-less policy.