Perfect Deterrence: Implications and Conclusions

1
Perfect Deterrence Implications and Conclusions

• Frank C. Zagare
• Department of Political Science
• University at Buffalo, SUNY
• Buffalo, NY 14260

2
Perfect Deterrence Theory

• Aim is to specify necessary and sufficient
  conditions
• Perfect Deterrence Theory grows out of joint work
  first with Kugler and later with Kilgour
• It focuses on parity relationships which Organski
  and Kugler show is necessary for major power
  conflicts

3
Critical Questions

• When does deterrence work?
• When do limited conflicts occur?
• When is a conflict spiral likely?
• What are the best crisis management strategies?
• What defense postures are efficacious deterrents?

4
Deterrence Theory An Alternative
Conceptualization

• Deterrence works best when the underlying
  structure of the game resembles Prisoners
  Dilemma
• In PD, players have mutually credible threats
• In a sequentially played game with PD-like
  preferences, the status quo is the unique subgame
  perfect, Pareto-efficient, Nash equilibrium

5
Perfect Deterrence Theory (1)

• Explores the strategic implications of credible
  threats given capable threats (i.e., parity
• Unlike Classical Deterrence Theory, Perfect
  Deterrence Theory allows the credibility of end
  game threats to vary
• Subsumes conditions presumed by Powell, Gauthier,
  Brams, etc.

6
Perfect Deterrence Theory (2)

• Applies as well to nuclear and non-nuclear
  situations
• More useful for examining extended deterrence
  situations than standard models
• Logically consistent and empirically plausible

7
Capability A Necessary Condition

• Capability is the ability to hurt.
• In Perfect Deterrence Theory, capability emerges
  as a necessary condition for deterrence success
• The lack of a capable threat is sufficient for
  deterrence failure
• Contrary to some theorists, a capable threat is
  not sufficient for success

8
Deterrence Success and Failure Sufficient
Conditions

• Deterrence can break down in a number of ways,
  even when a Defenders threat is capable
• Multiple equilibria almost always exist, making
  sufficient conditions for deterrence success
  almost impossible to specify
• Deterrence is, therefore, tenuous almost
  anything can happen under any conditions

9
What Makes Deterrence Success More Likely (1)

• A highly valued status quo is a major determinant
  of deterrence success
• Neglected in all studies of deterrence breakdown
  (e.g., Lebow and Stein) and underplayed in most
  empirical studies (e.g., Huth and Russett)
• Neglect introduces a case selection bias

10
What Makes Deterrence Success More Likely (2)

• Credibility is the critical variable. It is,
  however, neither necessary nor sufficient for
  deterrence to succeed. Deterrence can succeed
  when a defenders threat lacks credibility and
  fail when it does not.

11
What Makes Deterrence Success More Likely (3)

• In Perfect Deterrence Theory credibility is
  defined in terms of a players belief that its
  opponent prefers to carry out a threat
• Believability Credibility Rationality
• Rationality subgame perfection

12
Credibility (1)

• Defenders credibility is an important
  determinant of deterrence success
• Extended deterrence is most likely when a
  defenders sub-strategic and strategic level
  threats are highly credible
• But a challengers credibility may actually be
  the key

13
Credibility (2)

• When challengers threat is not very credible,
  defender will be undeterred from resisting,
  regardless of its credibility
• In consequence, deterrence succeeds because
  challengers threat lacks credibility, not
  because of the characteristics of defenders
  threat
• Empirical studies that focus on the
  characteristics of a defenders threat miss this
  important dimension of extended deterrence
  relationships

14
Major Finding of Incomplete Information Models

• The conditions of war and peace generally exist
  simultaneously
• Parity is dangerous so dont proliferate
• Overkill capability has limits
• Commitment tactics should be avoided

15
Major Finding of Incomplete Information Models
(continued)

• Conditionally cooperative stratagems (tit-for-tat
  policies) appear to be the most efficacious
  mechanisms for avoiding conflict and war
• Defense postures like Massive Retaliation that
  depend entirely on escalatory threats are
  ineffective deterrents. And when Challenger is
  likely hard, they are useless.

16
Major Findings (continued)

• Limited conflicts and even escalation spirals are
  low probability events. They are outcomes that
  are not anticipated by the players rather they
  are the consequence of gambles taken by
  players who have incomplete information about
  their opponents preferences (intentions).

17
Perfect Deterrence Theory Empirical Expectations
(1)

• Initiators of interstate conflict should
  generally be stronger than their opponents (Bueno
  de Mesquita)
• Satisfaction with the status quo should be
  associated with deterrence success (Reed,
  Rousseau et al.)

18
Perfect Deterrence Theory Empirical Expectations
(2)

• Tit-for-tat strategies should be associated with
  increased probability of deterrence success
  (Huth)
• Firm-but-flexible negotiating styles should be
  associated with increased probability of
  deterrence success (Huth)

19
Perfect Deterrence Theory Empirical Expectations
(3)

• Reciprocity should dominate interstate
  interactions, as states seek to establish
  credible threats (Stimulus-Response Model)
• One-sided victories should be the most frequently
  observed outcome once deterrence breaks down
  (MID-data)

20
Perfect Deterrence Theory General Comments(1)

• Converges with best empirical studies
• Speaks to efficacy of various deployment
  strategies (i.e., all-or-nothing and limited war
  stances)
• Specifies conditions for limited conflicts and
  escalation spirals
• Explains why proliferation is unsafe

21
Perfect Deterrence Theory General Comments(2)

• Supports downsizing of forces (i.e., minimum
  deterrence)
• Consistent with no-first-use deployments
• Consistent with arms reduction initiatives