Capstone: the AfCFTA as a strategic system

From solving games to using them

Across this course you have built a toolkit: dominance and the Prisoner's Dilemma, Nash equilibrium and coordination, mixed strategies, backward induction and commitment, repeated games and the folk theorem, bargaining, collective action, signalling and screening, mechanism design, and the Shapley value. This capstone asks a different question. Not “how do you solve this game?” but “how do you use game theory to think about a policy you have never seen before?” You will learn a four-question method, watch it dissect the African Continental Free Trade Area (AfCFTA) — the largest cooperation problem the continent has ever attempted — and leave with a playbook you can point at any strategic problem that reaches your desk.

The analyst's four questions

1. Who are the players, and what can each of them do? Fix the decision-makers — governments, firms, regulators, voters — and write down each one's strategy set, the concrete actions available to it.
2. What are the payoffs? Turn each combination of strategies into a number for every player, remembering that a government's payoff blends economic welfare with political-economy weights on particular sectors and voters.
3. What is the equilibrium? Solve the game: strike out dominated strategies, find the Nash equilibria, apply backward induction when moves are sequential, and look for a mixed-strategy equilibrium when no pure one exists.
4. What would change it? Here is the analyst's real craft — identify the lever (a commitment device, repeated play, better information, a redesigned rule or mechanism, a side-payment, or a change in who moves first) that relocates the equilibrium to a better outcome.

The first three questions are description; the fourth is design, and most of the value an analyst adds lives there. So keep a catalogue of levers in mind. If the problem is a Prisoner's Dilemma, ask whether repetition and monitoring can summon the folk theorem. If it is a coordination failure, supply a focal point. If it is a broken promise, look for a commitment device. If it is private information, build a screen. If the pie is positive-sum but unevenly split, engineer a side-payment. The rest of this lesson shows all six levers at work inside a single treaty.

The case: the AfCFTA

The AfCFTA was signed in Kigali in 2018, entered into force in 2019 once twenty-two states had ratified, and saw its first trades on the first of January 2021. Almost every African Union member has signed, and a secretariat sits in Accra. On paper it creates a single market of more than a billion people. In game-theoretic terms it is a machine for solving several hard problems at once — a collective-action problem in tariffs, a coordination problem in standards, a free-riding problem in implementation, and an enforcement problem in compliance. We take them in turn, and at each step name the tool that does the work.

Tariff phase-downs: an n-country Prisoner's Dilemma

Start with tariffs. The players are the member governments; each chooses, sector by sector, whether to liberalise (cut its tariffs toward zero) or protect. Liberalisation by everyone creates the prize — a large integrated market with scale economies, lower consumer prices and cross-border value chains. But each government privately values protection: tariff revenue it can spend, and shelter for politically important producers. So the temptation is to keep your own barriers while enjoying everyone else's openness.

                      Rest of Africa
                  Liberalise      Protect
  You Liberalise      3 , 3         0 , 4
      Protect         4 , 0         1 , 1

  (Row payoff first.) Protecting strictly dominates
  liberalising for you: 4 > 3 and 1 > 0. Yet the
  equilibrium (Protect, Protect) = (1,1) is worse for
  everyone than (Liberalise, Liberalise) = (3,3).

Protecting strictly dominates liberalising for you — it pays more whether the others open or not — so the only Nash equilibrium is universal protection, at (1, 1). Yet everyone prefers universal liberalisation at (3, 3). That gap between the equilibrium and the efficient outcome is the whole drama of trade policy. With two players we call it a Prisoner's Dilemma; with fifty-four it becomes a continental collective-action problem, and the single market is the public good it fails to provide.

Rules of origin: coordination and the focal standard

Suppose the members do agree to open. A second problem appears at once: what counts as “made in Africa”? Without a shared rule, a non-member could ship goods into a low-tariff state and wave them across the continent duty-free, and the whole scheme unravels. Rules of origin fix this by setting a threshold — say a minimum of forty per cent regional value content, or a required change of tariff heading. The structure here differs from the tariff game: several self-consistent standards exist, and any common one beats having none, so this is a coordination game with multiple Nash equilibria. The catch is distributional — a light-processing exporter prefers a lenient rule while an integrated manufacturer prefers a strict one, so members agree that they must coordinate but disagree on where, a battle of the sexes. The analyst's job is not to compute the perfect rule but to supply a focal point everyone can settle on.

Ratification versus implementation: cheap talk and free-riding

Signing a treaty is easy; living by it is not. Ratification is close to cheap talk — a low-cost declaration that signals resolve only weakly. Implementation — publishing tariff schedules, retraining customs officers, dismantling non-tariff barriers — is genuinely costly, and its benefits spill over to everyone regardless of who pays. That is a textbook free-rider problem: each state would prefer the others do the expensive work. The way to tell talk from commitment is to watch for costly signals — a notified, legally binding tariff schedule and an operating one-stop border post reveal a government's type far more credibly than a signing-ceremony photograph.

Compliance you cannot see: screening and the audit game

Even a committed member can quietly cheat. A state can honour its tariff schedule on paper while smuggling protection back in through non-tariff barriers — arbitrary standards, licensing delays, discriminatory charges. Its trading partners cannot directly observe whether it is truly complying; this is asymmetric information and moral hazard. The AfCFTA's answer is a screen: an online non-tariff-barrier reporting and monitoring mechanism (tradebarriers.africa) that lets exporters flag obstacles and forces states to respond. Monitoring turns unobservable effort into a public signal — which matters enormously, because the enforcement that follows can only be triggered on signals everyone can see, the problem of imperfect public monitoring in a repeated game.

Enforcement is itself a game. Consider a customs authority deciding whether to audit a consignment claiming AfCFTA origin, and a trader deciding whether to cheat on the rules. If audits were free the authority would always inspect and nobody would cheat; but audits are costly, so there is no equilibrium in pure strategies — the authority mixes between auditing and waving through, and the trader mixes between honesty and fraud. Solving for this mixed-strategy Nash equilibrium yields a famous, counter-intuitive result: the trader's equilibrium cheating rate is pinned down by the auditor's costs, while the audit rate is pinned down by the trader's penalty. Raise the fine and you change how often inspectors inspect, not how often traders cheat. To cut cheating you must lower the cost or raise the payoff of monitoring — which is exactly what a shared digital notification system does.

Dispute settlement: turning one shot into many

If the tariff game were played once, defection would win and the market would never form. But trade is not a one-shot game — members interact year after year, across thousands of products. That is what the dispute-settlement mechanism exploits. Its protocol, modelled on the World Trade Organization's, lets an aggrieved state bring a complaint, have a panel adjudicate whether a violation occurred, and — if it did — obtain authorisation to retaliate in proportion. In this course's language, the mechanism institutionalises a trigger strategy: defect, and you lose a measured slice of future cooperation. By the folk theorem, if members value the future enough — here, once the discount factor δ is at least one-third for our payoff numbers — universal liberalisation becomes a subgame-perfect equilibrium that sustains itself. Proportionality is the subtle part: bounded, adjudicated retaliation punishes defection without igniting an all-out trade war, keeping the players on the cooperative path rather than in a spiral of mutual punishment.

Moving the equilibrium: phasing, carve-outs, side-payments

Now watch the treaty edit the payoff matrix on purpose. Liberalisation is phased: members cut roughly ninety per cent of tariff lines over five years — ten for least-developed members — with seven per cent of “sensitive” products on a slower schedule and three per cent excluded outright. Phasing matters twice over. It shrinks the one-period gain from defecting relative to the stream of future benefits, which lowers the discount factor needed to sustain cooperation. And by letting each government shelter its politically hottest sectors, the sensitive and exclusion lists cut the domestic temptation to defect, so that liberalising the other ninety per cent becomes individually rational rather than politically suicidal.

Some members still lose in the short run — states that lean heavily on tariff revenue, or whose infant industries face new competition. Left uncompensated, they have every reason to walk, and their exit shrinks the market for everyone. The AfCFTA's Adjustment Fund, anchored by Afreximbank, is the side-payment that keeps them in: a transfer from the winners of integration to its losers. Read it as a bargaining solution — a split of the cooperation surplus that satisfies every member's participation constraint, so that staying dominates leaving. Coalition theory sharpens the point: compensation tied to each member's marginal contribution to the market, in the spirit of the Shapley value, keeps the grand coalition of all fifty-four in the core, with no subgroup able to do better by forming a bloc of its own.

The analyst's playbook

1. Name the players and their strategies. Be concrete; the wrong player set hides the real conflict.
2. Write the payoffs, political weights included. If you cannot rank the outcomes, you cannot solve the game.
3. Solve for the equilibrium. Use dominance, then Nash, then backward induction for sequential games, and mixed strategies when no pure equilibrium exists.
4. Diagnose the pathology. Is it a Prisoner's Dilemma, a coordination failure, a commitment problem, free-riding, or asymmetric information? The diagnosis dictates the cure.
5. Choose the lever. Add repetition and monitoring; supply a focal point; build a commitment device; design a mechanism that makes truth-telling optimal; arrange side-payments; or change who moves first.
6. Check credibility. Verify the fix is incentive-compatible and self-enforcing — an equilibrium nobody wants to deviate from — not merely a promise on paper.

The method travels. Central-bank credibility is a commitment problem: a government tempted to inflate for a short-lived boost ends up in a high-inflation equilibrium that fools no one, so the cure is to delegate to an independent central bank — the same backward-induction logic a mining investor relies on, and the discipline behind the West African CFA monetary zone. Continental climate policy is a Prisoner's Dilemma in emissions, answered by the same levers: a club with border penalties to change payoffs, repetition to enable trigger strategies, and climate-finance side-payments. And every further step of African integration — the Pan-African Payment and Settlement System, a single air-transport market, regional power pools — is a coordination-plus-collective-action problem met by a focal standard, credible monitoring, and compensation for those who lose. One method, many games.