The brinkmanship trap

The Chicken game places two players on a collision course, each hoping the other will swerve. In the trade dispute between Northland and Eastmark, both governments have announced escalating tariff packages and face a binary choice: follow through with the next round of escalation, or back down and accept a worse bilateral deal. Neither side has a dominant strategy. Each prefers to escalate if the other backs down, and to back down if the other escalates. This structure, anti-coordination, is what defines the Chicken game and separates it sharply from the prisoner's dilemma.

Northland and Eastmark are not symmetric. Northland runs an export surplus and its government faces domestic political pressure to appear resolute. Its electorate reads any concession as weakness, and its ruling coalition depends on nationalist support. Eastmark is import-dependent, with a manufacturing sector that relies on Northland's components. A prolonged trade war would hurt Eastmark's output more quickly than Northland's, which is precisely why Northland chose to escalate first.

The payoff structure captures this logic in four outcomes. If both escalate, the result is a full trade war: both economies contract, both governments absorb political damage, and the payoff is (-3,-3). If Northland escalates and Eastmark backs down, Northland secures its trade objectives and Eastmark retreats with face damage: (5,1). The mirror holds for the reverse. If both back down, there is a mutual de-escalation in which both lose some face but avoid the economic crash: (3,3).

The name Chicken comes from the original formulation of two drivers heading toward each other on a narrow road. The rational move, given the other driver swerves, is to hold the wheel straight. But if both hold, the crash destroys both. The game teaches that rationality is not enough. What matters is whether your commitment to holding straight is believable. In a trade dispute, the equivalent of holding the wheel straight is announcing irreversible implementation steps: binding parliamentary votes, pre-signed contracts with alternative suppliers, public statements by senior ministers that back down is politically impossible.

Neither player has a dominant strategy, and that is the central fact of the game. Map the incentives carefully. If Eastmark escalates, Northland's best response is to back down: a payoff of 1 beats the crash at -3. If Eastmark backs down, Northland's best response is to escalate: a payoff of 5 beats 3. Northland's optimal move tracks Eastmark's move exactly in reverse. The same symmetry applies from Eastmark's side. This is the defining feature of an anti-coordination game: each player's best response is the opposite of the other's choice.

The BATNA for each side matters. Northland's outside option if negotiations collapse entirely is a regional trade agreement with third parties that partially substitutes for the Eastmark market. Costly, but survivable over a 12-month horizon. Eastmark's outside option is worse: alternative supply chains would take two to three years to develop, and its export-dependent manufacturers cannot wait that long. This asymmetry in BATNAs means Eastmark has stronger incentives to back down from a prolonged conflict. Northland's strategists know this, which is why the initiating party chose to escalate.

The incomplete information dimension compounds everything. Each side holds private information about its own resolve. Northland does not know whether Eastmark's counter-tariff threat is genuine or theatrical. Eastmark does not know whether Northland's commitment is locked in by parliamentary procedure or merely announced. This information gap is where brinkmanship operates: the rational move is to make your resolve as legible and irreversible as possible, so the other side faces the calculation clearly.

Crash outcome (to avoid) Asymmetric Nash equilibria Retreat outcome

Figure 1. Commitment space for the brinkmanship game. The danger zone (shaded) is where both players have committed so completely that backing down is politically or physically impossible. Both pure-strategy Nash equilibria lie on the axes, not at the crash.

Four outcomes, two players, two strategies each. The full strategy space is a 2x2 matrix with Escalate and Back Down as the pure strategies for both Northland and Eastmark. The outcomes are not symmetric in value or in stability.

Notice that Back Down is not a dominated strategy. If Eastmark believes Northland will escalate, Eastmark's best response is to Back Down. Back Down is the best reply to Escalate, which means it survives iterated elimination. This is the key difference from the prisoner's dilemma, where both players have a dominant strategy of defection. In the Chicken game, no dominant strategy exists for either player.

The mixed-strategy Nash equilibrium fills the gap left by the two asymmetric pure equilibria. Let p be the probability that Northland escalates. Eastmark is indifferent between its two strategies when the expected payoff from escalating equals the expected payoff from backing down. Setting those equal: p(-3) + (1-p)(5) = p(1) + (1-p)(3). Expanding: 5 - 8p = 3 - 2p, which gives 2 = 6p, so p = 1/3. By the symmetry of the payoffs, Eastmark also escalates with probability 1/3 in the mixed-strategy equilibrium. At this equilibrium, each country escalates one time in three. The probability that both escalate simultaneously, producing the crash, is (1/3)(1/3) = 1/9. This is not a small number in a dispute between nuclear-armed or economically entangled states. A one-in-nine chance of a severe bilateral contraction is the price of playing the mixed equilibrium.

The recommendations differ by player, because their positions in the game are not symmetric. Northland is the initiating party with the stronger BATNA and the greater domestic pressure to hold firm. Eastmark is the responding party with more economic exposure and a longer recovery path from a crash. The moves available to each reflect these differences.

For Northland: The primary objective is to make the commitment to escalation as costly to reverse as possible, without actually crossing into the crash. Concrete moves: (a) push the tariff package through a parliamentary vote before the next bilateral meeting, so that backing down requires another parliamentary reversal rather than a unilateral executive decision. This raises the political cost of retreat to a level that Eastmark can observe. (b) Simultaneously construct a visible face-saving off-ramp for Eastmark. The off-ramp should be framed as a reciprocal step, not a concession: if Eastmark agrees to a specific technical measure on dumping standards, Northland will delay implementation of the next tariff tranche by 90 days. This reframes Eastmark's retreat as a negotiation outcome rather than a defeat. (c) Set a public deadline after which the off-ramp closes, so Eastmark's decision window is clear and Northland's commitment becomes time-bound. A floating deadline is not credible; a specific date with visible implementation triggers is.

For Eastmark: The first analytical task is to assess whether Northland's commitment is genuine or theatrical. Signs that it is genuine: parliamentary votes already taken, pre-signed contracts with alternative suppliers announced, senior ministers who have explicitly staked personal credibility on the outcome. Signs that it is theatrical: announcements without implementation steps, statements conditional on Eastmark's response, evidence that Northland's coalition is not unified on escalation. If the commitment appears genuine, back down early enough to negotiate the off-ramp before the window closes. If it appears theatrical, calling the bluff is rational. But this depends on one further condition: Eastmark must be able to tolerate the crash payoff of -3 for at least a quarter if the bluff turns out not to be a bluff. If the crash would cause structural damage to Eastmark's manufacturing sector that persists for years, the asymmetric BATNA already answers the question. Back down, negotiate the best available terms, and protect the off-ramp for future rounds.

The universal recommendation for both parties is to keep at least one private channel open at all times. Public commitments are necessary for credibility, but they eliminate flexibility. A private channel between senior envoys allows both sides to signal real versus theatrical positions without forcing premature public retreats. The public game is for commitment; the private channel is for coordination. Thomas Schelling's original insight applies directly: the communication that matters most is the communication that can be denied.

The game changes as the commitment structure changes. Four shifts are worth modelling, because each alters the payoff matrix or the information structure in ways that can move the equilibrium.

Public irreversibility. Once Northland's parliament ratifies the tariff package by a recorded vote, the commitment moves from announced to enacted. The political cost of reversal rises sharply: the government would need to face a second vote, acknowledge that the first was a mistake, and absorb the domestic political cost. From Eastmark's perspective, this is a qualitative shift. Before ratification, the threat is cancellable at low cost. After ratification, cancellation requires a second legislative act and is therefore far less likely. Eastmark's rational response to ratification is to accelerate the decision: take the off-ramp before the next escalation tranche is also ratified, or commit to counter-escalation if the off-ramp is not acceptable. The payoff matrix has not changed, but the effective probability that Northland executes its strategy has risen dramatically.

Third-party arbitration. An external mediator, whether the WTO dispute panel, a regional trade body, or a third country with relationships on both sides, can provide a symmetric off-ramp that neither side can provide unilaterally. The mediator's value is precisely that it resolves the equilibrium selection problem: both sides can retreat to a mediator's ruling without either side being seen to have backed down. This transforms the game from a bilateral Chicken into a tripartite coordination problem with a focal point. Historically, successful trade dispute resolutions have almost always required this kind of external reference. The 1986 US-Japan semiconductor agreement and the 2019 US-China Phase One deal both required elaborate face-saving architecture that neither side could have constructed without a third-party framework.

Domestic political change in Eastmark. An election result, a coalition shift, or a change in the ruling party's economic policy priorities can alter the effective payoffs for Eastmark. If a new government places higher weight on the crash payoff, perhaps because its electoral base is not concentrated in the import-dependent manufacturing sector, the previously losing Nash equilibrium of (Back Down, Escalate) becomes less acceptable to Eastmark and the (Escalate, Escalate) crash becomes more tolerable. Northland's strategists must update their model of Eastmark's resolve when political conditions change, because the game they are playing changes with them.

Exogenous shock. A global recession, a shared external threat, a financial crisis that hits both economies simultaneously, or a security event that requires bilateral cooperation all alter the crash payoff from -3 to something far more severe. When the crash becomes catastrophically costly for both sides, the mixed-strategy equilibrium shifts: the probability that each player escalates falls, because the expected cost of the mixed equilibrium rises. At some threshold, both sides find the mutual retreat outcome at (3,3) preferable to any risk of the crash. This is the dynamic that historically resolves the most dangerous brinkmanship episodes: not a change in resolve, but a change in what mutual destruction actually costs.