The Makkah Pact: Can a Treaty Deter War Without Starting One?

What the Pact Actually Says

On 7 August 2026, Saudi Arabia, Pakistan, and Türkiye signed the Makkah Joint Defence Agreement. The ¹ Saudi Press Agency’s joint statement establishes a simple collective-defence principle: an armed attack on one shall be regarded as an attack on all.

But Turkish Foreign Minister Hakan Fidan immediately complicated that headline. In his account, assistance is request-based — the attacked state must ask for help and specify what it needs. The spectrum runs from intelligence and logistics through ammunition to the deployment of military units. A political-military committee and a permanent secretariat in Saudi Arabia are planned. Interoperability, joint training, and common standards remain works in progress. ²

This means the pact is not a mechanical trigger. It is a contingent chain:

The Contingent Alliance Chain Attack → Invocation → Consultation → Choice of Support Level

That intermediate decision node matters. It prevents entrapment — allies can say no — but it also creates the central problem of this analysis: when the moment arrives, will they say yes?

Key distinction: This is ³ not a released treaty text. It is a joint statement plus subsequent commentary. The full legal and operational details remain classified, which means every parameter in the models below is an analytical variable, not an observed treaty provision.

The Core Idea: Probabilistic Deterrence

Here is the insight that drives everything: deterrence does not require an adversary to believe all three states will automatically go to war. It only requires the adversary to assign a sufficiently high probability to some costly response.

If an attacker believes there is probability q that the pact produces a robust military response, the agreement deters whenever the attacker’s expected gains fall below the expected costs created by q.

How can the three states make intervention credible enough to deter attack… without making it so automatic that they invite entrapment?

The answer is graduated commitment. Make the cheap stuff automatic. Leave the expensive stuff discretionary.

Interactive — Iran Deterrence Threshold

Adjust Iran’s perceived benefit, the cost of a strong response, and the cost of a weak response. Watch how the deterrence threshold q* changes.

B_I (benefit of attack) 4.0

C_H (cost if strong response) 7.0

C_L (cost if weak response) 1.0

E_I (escalation risk) 0.0

Note

The striking implication: The alliance does not have to make intervention certain. If credible preparations raise Iran’s perceived probability of a strong response above the threshold, the attack becomes irrational. A 50% chance of a robust response can be enough.

The Strategic Network

The Makkah pact does not exist in a vacuum. It overlays existing institutions — the GCC, NATO, US-GCC security, and bilateral relationships — creating overlapping, sometimes substitutive, sometimes complementary security networks. The critical feature: almost every player has cross-cutting interests.

Fig. 01 — Strategic network around the Makkah Pact. Copper dashed = pact members. Threat vectors, alliance infrastructure, and diplomatic ties are distinguished by line weight and tone.

Saudi Arabia wants deterrence against Iran but not permanent war. Pakistan values Saudi support but cannot welcome an anti-Iran war on its western frontier. Türkkiye wants autonomy but also benefits from NATO. These cross-cutting incentives make a rigid two-bloc outcome less likely than a multilayered bargaining equilibrium.

Crisis Game 1: Iran Strikes Saudi Arabia

This is the pact’s strongest strategic case. Saudi Arabia reported being targeted by 1,069 drones, 110 ballistic missiles, and 20 cruise missiles between 28 February and 2 April 2026 — ⁴ figures submitted to the UN Security Council. The threat is concrete, not hypothetical.

The Game

Players: Iran, Saudi Arabia, Pakistan, Türkkiye. The US and GCC enter as exogenous support variables.

Game structure (click to expand)

Step 1: Nature draws the alliance’s type — high resolve (H) or low resolve (L). Iran sees signals but doesn’t observe the true type directly.

Step 2: Iran chooses: no attack, limited attack, or major attack.

Step 3: Saudi Arabia invokes the pact. Pakistan and Türkkiye each choose a support level: abstain, limited defensive aid, or robust military aid.

Step 4: Iran observes the response and chooses: de-escalate, continue, or escalate.

The Math

Iran’s expected utility from attacking:

Iran's Expected Utility U_I(A) = B_I − [q · C_H + (1−q) · C_L] − E_I

Deterrence holds when U_I(A) ≤ 0, which requires:

The Deterrence Threshold q ≥ q* = (B_I − E_I − C_L) / (C_H − C_L)

The Continuation Game

After an initial strike, the alliance and Iran face a coordination problem:

Alliance Robust

(3, −2) ★

−4, −5

Alliance Weak

1, 0

(−3, 3) ★

There are two Nash equilibria. But when the alliance moves first and Iran observes, the ⁵ subgame-perfect equilibrium is:

Subgame-Perfect Nash Equilibrium Robust Response → Iranian De-escalation

Note

Why interoperability (λ) is everything: Pre-positioned defences, compatible data links, and established deployment procedures reduce the cost of a robust response and can literally change which equilibrium the game produces. The pact’s first institutional priority should be air and missile defence of Saudi critical infrastructure.

Crisis Game 2: The Houthi Attribution Problem

The Houthi scenario has a different structure. The identity of the attacker is usually clear — it’s the Houthis. What’s unclear is how much operational control Iran exercised over that particular attack. This is a ⁶ Bayesian game of incomplete information.

Fig. 02 — Bayesian attribution game. Response intensity depends on the posterior belief about sponsor control; the response boxes darken as escalation rises.

The coalition uses Bayes’ rule to update its belief about Iranian control after each noisy intelligence signal:

Bayesian Update μ(s) = P(P|s) = [π · f(s|P)] / [π · f(s|P) + (1−π) · f(s|A)]

Direct retaliation against Iran becomes rational only when:

Attribution Threshold for Escalation μ > μ* = (M + E − C_D − L_D) / (B + M)

Here’s the critical comparative static: the cheaper and more effective defensive maritime options become, the higher the attribution threshold for attacking Iran directly. The ⁷ EU’s EUNAVFOR ASPIDES mission (extended through February 2027) provides exactly this kind of alternative — it decreases the marginal value of escalation while increasing defensive capacity.

Note

Repetition matters. Each Houthi attack provides a new observation. If signals consistently indicate sponsor direction, the posterior P(P|s₁,...,s_n) can eventually cross μ*, making a stronger response rational. The equilibrium is one of calibrated defence — not automatic war, but stepwise escalation conditional on accumulated evidence.

Crisis Game 3: India–Pakistan Escalation

This is the most difficult contingency because it presents the strongest abandonment-versus-entrapment dilemma in ⁸ Glenn Snyder’s classic alliance terminology. Saudi Arabia and Türkkiye want the pact to deter an attack on Pakistan — but neither wants to encourage Pakistani risk-taking on the assumption that partners will rescue it.

The Responsibility Game

Nature draws who initiated the crisis: India (I) or Pakistan (P). Saudi Arabia and Türkkiye receive an imperfect signal and form a posterior belief p that India was the aggressor.

Their expected utility of supporting Pakistan:

Expected Utility of Support EU(S) = p(V + M) − M − C

And the threshold for support becomes:

The Support Threshold p > τ = (M + C) / (V + M + R)

This is a powerful result. Saudi/Turkish support should not be predicted simply from whether “Pakistan is a treaty ally.” It depends on whether their posterior belief that Pakistan is the victim exceeds the threshold τ.

Interactive — Support Threshold Calculator

Adjust the value of alliance (V), entrapment cost (M), aid cost (C), and reputational loss (R). See what posterior belief p is needed for Saudi/Turkish support.

V (alliance value) 8.0

M (entrapment cost) 5.0

C (aid cost) 3.0

R (reputational loss if abandon) 6.0

India’s Decision

India factors in the probability of Makkah-alliance intervention q_A and the probability of nuclear escalation q_N:

India's Deterrence Condition q_AK_A + q_NK_N ≥ G − C₀

Because the cost of nuclear escalation K_N is astronomically large, even a tiny probability of nuclear use can dominate India’s entire calculation. This is why the existing India-Pakistan nuclear uncertainty matters far more than speculative claims about a Pakistani nuclear umbrella for Saudi Arabia. ⁹ SIPRI estimates ~170 Pakistani and ~190 Indian warheads as of January 2026. No trilateral nuclear doctrine exists in public.

Note

The ¹⁰ moral-hazard paradox: Better intelligence accuracy θ creates more polarized behaviour. When evidence of Indian initiation is strong, support becomes much more likely. When evidence indicates Pakistani initiation, support collapses. Paradoxically, improving attribution simultaneously makes the pact more credible as a defensive alliance and less exploitable by Pakistan.

What Changes the Equilibrium

The models produce clear predictions about how changes in the pact or its environment shift behaviour.

Variable Change	Effect on Deterrence	Effect on Intervention	Main Danger
Greater treaty clarity	Usually raises q	Raises response probability	Increases entrapment/moral hazard
Greater interoperability λ	Strongly increases credible deterrence	Lowers material cost of action	May appear threatening offensively
Greater nuclear ambiguity n	Raises tail risk in South Asia	Limited Gulf applicability	Miscalculation and uncontrolled escalation
Stronger US reassurance u	Can strengthen deterrence immediately	Lowers operational response costs	Free-riding / slower autonomous integration
US withdrawal	Raises demand for pact	Initially makes action more expensive	Capability gap during transition
Higher domestic cost d_i	Lowers adversary belief in response	Reduces high-cost assistance	Alliance abandonment
Better attribution θ	Improves conditional deterrence	More targeted response	Adversary incentives to obscure responsibility
Strong third-party defence z	Usually raises defence capacity	Can substitute for pact combat	Alliance free-riding

The Hump-Shaped Ambiguity Curve

Treaty ambiguity has a non-linear effect on deterrence. Some ambiguity creates deterrent uncertainty — an attacker can’t be sure a limited operation will stay below the threshold. But too much ambiguity lets adversaries infer that allies will always find an excuse not to respond.

The relationship between ambiguity and deterrence is ¹¹ hump-shaped: a moderate uncertainty generates caution; excessive uncertainty destroys commitment.

Interactive — The Ambiguity-Deterrence Curve

Drag the peak position and width to see how the hump-shaped relationship works.

Peak ambiguity 50%

Curve width 25

The Coalition-Size Paradox

Foreign Minister Fidan has mentioned potential expansion — Egypt was named specifically. But adding members creates a paradox:

The Expansion Paradox N ↑ ⇒ Aggregate Capability ↑ but Preference Heterogeneity ↑ and Decision Speed ↓

A three-member compact with highly credible minimum commitments can be strategically stronger than a ten-member organization whose response requires unanimous bargaining during every crisis.

The Optimal Design: Tiered Commitment

The models converge on a clear prescription. The alliance should make cheap, mutually beneficial actions nearly automatic, while preserving consultation for actions carrying extreme escalation risk.

Tier 1

Automatic Low-Cost Defence Intelligence sharing, early warning data, cyber defence, air/missile-defence information, logistical access, resupply, crisis consultations. Low C_j, high credibility. These actions satisfy the intervention inequality almost always.

Tier 2

Discretionary Reinforcement Pre-designated reinforcement units, ammunition pre-positioning, maritime escort, limited combat deployments. Moderate C_j. Credible when institutionalization λ is high enough to lower costs. The equilibrium is conditional on crisis type and attribution.

Tier 3

Deliberately Uncertain Major interstate war, offensive campaigns, nuclear-related contingencies. Very high C_j, E_j, d_j. The pact deliberately leaves this ambiguous — enough uncertainty to deter, enough discretion to avoid entrapment.

The strategically superior formula is:

The Optimal Commitment Formula Certain Limited Response + Credible Possibility of Larger Response

…rather than an incredible promise of automatic maximal response.

The Seven Institutional Mechanisms

1. Pre-authorized minimum response package

A published or privately pre-authorized package after a verified attack: intelligence, early warning, cyber defence, air-defence data, logistical access, resupply, crisis consultations. Low cost, high credibility. Once an adversary believes something consequential happens automatically, deterrence rises.

2. Pre-positioned forces

An air-defence battery already deployed in Saudi Arabia creates a different game from a promise to send one later. Pre-positioning reduces mobilization time, creates sunk costs, and generates a ¹² tripwire: attacking Saudi infrastructure may endanger Pakistani or Türkish personnel.

3. Integrated early warning and C2

Shared sensor data is an obvious public good. The GCC has called for strengthened ballistic-missile early-warning. Integrating Pakistani and Türkish systems with a Saudi/GCC framework lowers defensive costs while giving the pact an operational function that doesn’t require agreeing on who the enemy is.

4. Real deployment exercises

Measure T_decision, T_deployment, T_integration — time from attack to authorization, from authorization to arrival, from arrival to combat effectiveness. Repeated exercises reduce all three and reveal bottlenecks before adversaries exploit them.

5. Shared attribution protocol

Predetermined evidentiary standards: θ ≥ θ_defensive for automatic defensive assistance, θ ≥ θ_offensive for retaliation against an alleged sponsor. This separates rapid defence from potentially catastrophic escalation.

6. Anti-entrapment understanding

For India-Pakistan crises: publicly, aggression against Pakistan carries consequences. Privately, the strongest guarantees apply only when Pakistan has not initiated escalation. Public deterrence plus private restraint — more stable than either an unconditional guarantee or a completely vague pact.

7. Nuclear clarification

Two coherent options: (a) state explicitly that the pact is conventional and does not alter Pakistani nuclear command; or (b) develop an explicit nuclear consultation mechanism. What is dangerous is relying on speculation about a “nuclear umbrella” without established doctrine or red lines.

Seven Testable Predictions

#	Prediction	Observable Evidence
01	Defensive cooperation institutionalizes before offensive war planning	Earliest implementation: intelligence, command-post exercises, missile/drone defence, logistics, communications
02	Saudi contingencies produce stronger alliance behaviour than India-Pakistan contingencies	Overlapping interests clearest on protecting Saudi infrastructure from missiles and drones
03	Repeated attacks produce stepwise, not continuous, escalation	Response stays defensive until posterior belief in sponsor responsibility crosses μ*, then escalates abruptly
04	Physical deployments deter more than communiqués	Stationed units impose sunk costs and create potential casualties — costly signals
05	Saudi/Türkkiye will distinguish between defending Pakistan and supporting Pakistani escalation	Diplomatic messaging during the first serious India-Pakistan crisis after the treaty
06	Strong US reassurance may delay autonomous integration	If US-GCC cooperation remains intensive and Makkah secretariat develops slowly, substitution dominates
07	Alliance expansion will expose a capability-vs-consensus trade-off	Additional members like Egypt raise reach but create decision-making delays

The Bottom Line

The Makkah agreement has already changed the game. Iran, India, Israel, the Houthis, and other actors must now attach some positive probability to Pakistani and Türkish involvement in contingencies that previously would have been treated mainly as bilateral or US-Gulf problems. That alone can modify bargaining behaviour.

But deterrence depends on the size of that probability.

The pact’s most credible equilibrium is likely to be: highly credible for consultation, intelligence, logistics, and defensive military support; moderately credible for reinforcement and limited combat; and deliberately uncertain for major interstate war or nuclear escalation.

That may not sound as dramatic as “Islamic NATO,” but from a ¹³ game-theoretic perspective it may actually be the more rational and sustainable design.

The decisive variable is not what the pact is called. It is whether its institutions transform the political cost of abandonment and the military cost of intervention enough that helping an attacked partner becomes rational after the attack, not merely attractive when leaders sign the treaty.

On present evidence, the pact has the political architecture necessary to attempt that transformation. It does not yet have the publicly demonstrated operational machinery to prove it has succeeded.

q → 1 for defensive support | 0 < q < 1 for offensive combat

The pact succeeds when adversaries face predictable costs for attack, yet partners retain enough discretion to avoid automatic entrapment.

Sources

Saudi Press Agency (SPA) — Joint statement, 7 August 2026.
Turkish Ministry of National Defence — Fidan remarks, August 2026.
James Morrow — “Alliances: Why Write Them Down?”, Annual Review of Political Science.
UN Security Council Document S/2026/283.
Council of the European Union — ASPIDES mission extension, 2026.
Glenn Snyder — “The Security Dilemma in Alliance Politics”, World Politics.
SIPRI — World Nuclear Forces, January 2026.
James Fearon — “Rationalist Explanations for War”, International Organization.