A Quantitative Bargaining Theory of War Key concepts in bargaining - - PowerPoint PPT Presentation
A Quantitative Bargaining Theory of War Key concepts in bargaining - - PowerPoint PPT Presentation
Kristopher W. Ramsay Brenton Kenkel September 3, 2016 A Quantitative Bargaining Theory of War Key concepts in bargaining model of war: Military strength Resolve/cost of fighting Prior beliefs/uncertainty How can we measure these
SLIDE 1
SLIDE 2
Introduction
Key concepts in bargaining model of war:
- Military strength
- Resolve/cost of fighting
- Prior beliefs/uncertainty
How can we measure these theoretical quantities in terms of
- bservable characteristics of states?
SLIDE 3
Approach
- 1. Write down bargaining model of war
- 2. Model exogenous parameters as functions of data
- 3. Assume data generated by equilibrium behavior
- 4. Structurally estimate
SLIDE 4
Bargaining model
Sides A and B, each with ≥ 1 constituent states
- 1. Side A offers x ∈ R
- 2. Side B accepts or rejects
- Accept → A gets x, B gets 1 − x
- Reject → each pays θk, war occurs
War costs θA, θB i.i.d. Exponential(λ)
SLIDE 5
War-fighting model
Each state expends effort ei ≥ 0 Probability Side A wins: pA = ∑
j∈A mjej
∑
j∈A mjej + ∑ j∈B mjej
War payoffs: uA = pA − θA − ∑
j∈A
cjej uB = 1 − pA − θB − ∑
j∈B
cjej
SLIDE 6
Empirical parameterization
Crisis level
- Shape of prior beliefs: λ
exp W
- Contiguity
- Preference Similarity
- Rivalry
- Major Power Involvement
- Peace Years
State level
- Military effectiveness: mi
exp Xi
- GDP
- Population
- Military Quality
- Marginal cost of effort: ci
exp Zi
- Imports/GDP
- Democracy
SLIDE 7
Empirical parameterization
Crisis level
- Shape of prior beliefs: λ = exp(Wα)
- Contiguity
- Preference Similarity
- Rivalry
- Major Power Involvement
- Peace Years
State level
- Military effectiveness: mi
exp Xi
- GDP
- Population
- Military Quality
- Marginal cost of effort: ci
exp Zi
- Imports/GDP
- Democracy
SLIDE 8
Empirical parameterization
Crisis level
- Shape of prior beliefs: λ = exp(Wα)
- Contiguity
- Preference Similarity
- Rivalry
- Major Power Involvement
- Peace Years
State level
- Military effectiveness: mi
exp Xi
- GDP
- Population
- Military Quality
- Marginal cost of effort: ci
exp Zi
- Imports/GDP
- Democracy
SLIDE 9
Empirical parameterization
Crisis level
- Shape of prior beliefs: λ = exp(Wα)
- Contiguity
- Preference Similarity
- Rivalry
- Major Power Involvement
- Peace Years
State level
- Military effectiveness: mi = exp(Xiβ)
- GDP
- Population
- Military Quality
- Marginal cost of effort: ci
exp Zi
- Imports/GDP
- Democracy
SLIDE 10
Empirical parameterization
Crisis level
- Shape of prior beliefs: λ = exp(Wα)
- Contiguity
- Preference Similarity
- Rivalry
- Major Power Involvement
- Peace Years
State level
- Military effectiveness: mi = exp(Xiβ)
- GDP
- Population
- Military Quality
- Marginal cost of effort: ci
exp Zi
- Imports/GDP
- Democracy
SLIDE 11
Empirical parameterization
Crisis level
- Shape of prior beliefs: λ = exp(Wα)
- Contiguity
- Preference Similarity
- Rivalry
- Major Power Involvement
- Peace Years
State level
- Military effectiveness: mi = exp(Xiβ)
- GDP
- Population
- Military Quality
- Marginal cost of effort: ci = exp(Ziγ)
- Imports/GDP
- Democracy
SLIDE 12
Empirical parameterization
Crisis level
- Shape of prior beliefs: λ = exp(Wα)
- Contiguity
- Preference Similarity
- Rivalry
- Major Power Involvement
- Peace Years
State level
- Military effectiveness: mi = exp(Xiβ)
- GDP
- Population
- Military Quality
- Marginal cost of effort: ci = exp(Ziγ)
- Imports/GDP
- Democracy
SLIDE 13
Data
Militarized Interstate Disputes, 1816–2001
- N = 2,295 disputes, with 5,451 total participants
- War: 0 or 1
- Winner: A, B, or censored
SLIDE 14
Data structure
Crisis level
Dispute War Winner Contiguity Rivalry ...
- 1
. 2 1 A 1 3 . 1 ...
State level
Dispute Side GDP Population ...
- 1
A 0.4 6.4 1 B 7.8 3.1 2 A 0.8 5.6 2 A 4.2 6.4 2 B 6.2 8.6 3 A 1.3 2.0 3 B 7.9 8.4 ...
SLIDE 15
Parameter estimates
α: ln(1 + Peace Years) α: Major Power–Both α: Major Power–Either α: Rivalry α: Preference Similarity α: Contiguity β: ln(1 + Military Quality) β: ln(Population) β: ln(GDP) γ: Democracy γ: ln(1 + Import Percentage) Prior Beliefs
- Mil. Effectiveness
Effort Cost
- 1.0
- 0.5
0.0 0.5 1.0
Estimate ± 1.96se
SLIDE 16
Equilibrium quantities: USA vs. Russia
0.4 0.5 0.6 0.7 1850 1900 1950 2000
Probability USA wins war
SLIDE 17
Equilibrium quantities: USA vs. Russia
1.0 1.5 1850 1900 1950 2000
Optimal offer by USA
SLIDE 18
Equilibrium quantities: USA vs. Russia
0.00 0.05 0.10 0.15 1850 1900 1950 2000
Probability of war
SLIDE 19
Conclusions and next steps
Conclusions
- Bargaining model has empirical content
- Major powers, similar preferences → more uncertainty
- Rivals, long time at peace → less uncertainty
- No discernible effects of economic/political characteristics on
states’ ability and willingness to wage war Next steps
- Different variables in the effectiveness/cost equations?
- Benchmark models for predictive comparison?
- Other substantive applications of estimator?