Mbegu Mwongozo 2074
Foundations & Mechanisms: The Behavioural and Contractarian Spine
Framework Page 2 — the mechanism layer: why cooperation is rational, how institutions make it real, and the forces that drive ascent or collapse · Working draft v0.1 · June 2026
Companion to Page 1 (Theoretical Context & the African Institutional Inventory) · Conceptual research instrument — not a predictive econometric system
On the status of this layer. This is a model, not a claim to completeness or truth. What follows is the set of mechanisms the framework judges most load-bearing for development trajectories in the region under study, each named, sourced, and defended — together with the failure mode that mirrors it. It is not the complete set of all forces that act on a society, and it does not pretend to be. The architecture is deliberately built so that further mechanisms can be added in the same form (a cooperative force, its negative mirror, the model variable it governs) without disturbing what is already here. Where a mechanism is well-evidenced experimentally or historically, that is noted; where it rests on theoretical argument, that too is marked. A model earns trust by showing its joins, not by hiding them.
1. The Spine: Why Cooperation Is Rational, Made Real, and Humanly Possible
Before the individual mechanisms, one argument runs through all of them and holds them together. It is a single chain with three links, drawn from three different traditions that turn out to be stating one underlying logic. The framework presents it as a spine, not a roof: the empirical mechanisms in Sections 2–7 stand on their own evidence, and a reader who does not grant the contractarian premises can still enter the model from the experimental and historical side. But for the reader who wants to know why these mechanisms and not others, the spine supplies the derivation.
North
Institutions emerge because they lower the cost of cooperative exchange. Constrained behaviour, once enforced, pays — so the institutions that enforce it are selected and persist.
←
Gauthier
Why a rational, self-interested agent submits to the constraint: constrained maximisation outperforms straight maximisation once dispositions are legible. Morality as a rational constraint derived from non-moral premises.
←
Sen
The constraint and the institution are inert unless the agent has the capability to participate — to read dispositions, enter agreements, command the resources that make cooperation worth entering.
1.1 Gauthier: the rational derivation of constraint
David Gauthier’s Morals by Agreement (1986) sets out to show that cooperation is rational on the cold premises of utility maximisation, with no moral assumptions imported. Three of its components are load-bearing for this framework, and each does distinct work.
Constrained maximisation. Gauthier’s agent is not the straight maximiser whose dominant strategy in a one-shot dilemma is to defect. She is a constrained maximiser: she adopts the disposition to cooperate on a joint strategy with others who are similarly disposed, and to withhold cooperation from straight maximisers so as not to be exploited. Crucially, this disposition is chosen by maximising reasoning — over a population in which dispositions can be read, the constrained maximiser does better than the straight maximiser. Gauthier’s own compliance condition is that a constrained maximiser follows a joint strategy when it is beneficial (better for her than non-cooperation), fair (about what minimax relative concession would yield), and she expects it actually to pay. This is the philosophical foundation of the framework’s reputation system: recorded history, transaction reputation, and credible refusal exist precisely to let agents identify who is a constrained maximiser and who is not. Enforcement infrastructure, in Gauthier’s terms, is the legibility technology that makes constrained maximisation rational — which is why Page 1’s engine dimension and this page’s behavioural spine are the same argument from two directions.
Minimax relative concession (MRC). Gauthier’s contribution to bargaining theory, and his answer to how the cooperative surplus is divided. Each party begins by claiming the most it could get; the rational, stable division is the one that minimises the largest relative concession any party must make from that claim, measured against the maximum it could have gained. Gauthier explicitly rejects the Nash bargaining solution (which maximises the product of gains) in favour of MRC. THEORETICAL For this framework, MRC is the fairness condition that makes cooperation stick: an arrangement that forces one party into a disproportionate concession is unstable, because that party defects back to the non-cooperative baseline. The dairy mechanism, graceful failure, and the rule that reward must reach the one who earned it are all minimax-relative-concession constraints in operational form — they fail exactly when someone’s concession grows too large relative to their contribution.
Relative benefit. The measure throughout is not absolute payoff but gain relative to the non-cooperative baseline — how much better off cooperation makes each party than the disagreement point. This reframes the model’s central question. Development is not “reach a wealthy endpoint”; it is “does each participant receive a fair share of the surplus that cooperation generates, relative to their fallback?” A modest economy that shares a growing surplus fairly is stable; a richer one whose surplus is captured is not. This is the framework’s trajectory-robustness measure stated in Gauthier’s own currency, and the reason the composite index must never be calibrated as distance from a high-income reference society.
1.2 North: the empirical emergence
Douglass North supplies what Gauthier’s derivation points toward but does not itself document: the historical record of institutions actually emerging to make cooperative exchange possible, lowering transaction costs and enabling impersonal trade. North tells us that institutions emerge and what they do; Gauthier tells us why a self-interested agent builds and obeys them. The two are complementary halves of one account — and Page 1’s inventory of indigenous African institutions is North’s emergence observed directly in the substrate the model builds on.
1.3 Sen: the capability precondition
Amartya Sen supplies what both North and Gauthier presuppose but neither provides: the agent must be capable of participating. Constrained maximisation assumes an agent who can read dispositions, enter and honour agreements, and command resources enough to make cooperation worth entering. Sen’s capability approach makes this explicit and gives it two roles in the framework, both load-bearing.
First, capability is the model’s true output — development as the expansion of what people are actually able to do and be, not aggregate income. This guards against the failure mode in which a society posts healthy headline numbers while the substrate and human freedoms erode (Page 1’s surface-versus-substrate decoupling, now stated in Sen’s terms). Second, and more sharply, capability is a gating input: an intervention engages only among agents capable of using it. A programmable transaction instrument deployed to people without literacy, numeracy, or agency delivers nothing; a tiered-certification ladder is unclimbable if the skills rungs do not exist. Capability deprivation is therefore a precondition failure — the “develop and deploy” constraint that ties the human-capital and education layers into the engine rather than leaving them as soft additions. Sen’s entitlement theory of famine completes the picture: famine enters the model as a collapse of entitlements (the exchange relations through which people command food), not as a production shortfall — a specification that fits the framework’s starvation-to-prosperity purpose and is grounded in Sen’s most consequential empirical work.
The plain-language hinge. The spine’s middle term — Gauthier’s claim that constraint is rational once dispositions are legible — has a one-sentence statement that the framework adopts as its through-line: reasonableness is a rational strategy. Cooperation, fair dealing, and restraint are not moral demands imposed against self-interest; under the right institutions they are self-interest, correctly calculated over a long enough horizon. Everything the model builds — reputation, enforcement, fair reward — exists to make that sentence true in places where it currently is not.
2. The Mechanisms: Cooperative Forces and Their Negative Mirrors
Each mechanism below is presented as a pair: the cooperative force that drives ascent, and the failure mode that mirrors it and drives collapse. This pairing is not decorative. The model must show both directions — ascent and reversal — and which column a given scenario falls into is precisely what the simulation makes visible. The same structural conditions that switch a society from the left column to the right are the levers the framework studies. For each pair, the source tradition is named, a plain-language term is given (the academic label is retained in the written theory), and the model variable it governs is identified. Operational thresholds and coefficients are specified on Page 3; this page establishes why each force is real.
2.1 Conditional cooperation & credible refusalOstrom 2000; folk theorem; Axelrod 1984
Cooperative forceOstrom’s experimental work establishes that self-interest is not inevitable: real populations contain conditional cooperators (who cooperate if others do) and willing punishers (who discipline defectors even at personal cost). Cooperation holds when monitoring lets conditional cooperators see that others are complying, so they need not fear being exploited. The folk theorem gives the formal condition — cooperation is a rational equilibrium when the shadow of the future is long enough — and Axelrod’s tournaments identify the winning disposition: nice, retaliatory, forgiving, and clear. Credible refusal (declining an exploitative deal at one’s own short-term cost) is the enforcement act that makes the norm hold.
Negative mirrorWhen the shadow of the future shortens, detection is absent, or defection is cheap, the folk theorem runs in reverse: defection becomes the rational equilibrium and the cooperative order unravels into a defection trap — the predatory regime in which it is individually rational to cheat because everyone else does and no one is watching. Willing punishers stop punishing once the cost exceeds the now-vanishing benefit of a future that will not arrive. This is the structural form of the “predation pays” society.
Governs: willing-punisher threshold (enforcement holds only if enough participants will refuse bad deals); couples to EIS (monitoring/legibility) and the time-horizon logic. Axelrod’s four properties are the design template for the framework’s graceful-failure and asymmetric-liability rules.
2.2 Bridging-trust conversion & in-group boundednessPutnam; Tajfel & Turner; parochial-altruism literature
Cooperative forceTrust can widen from bonding (within the kin or ethnic network) to bridging (across networks, among strangers) — the conversion the whole model is about. Encouragingly, the minimal-group findings show that group boundaries form on trivial cues and can therefore be re-formed: experimental recategorisation around a shared identity raises cross-group cooperation. Trade complementarity across groups has, historically, normalised cross-group trust over long periods. Bridging conversion is constrained maximisation extending its membership set outward.
Negative mirrorThe default is in-group favouritism (social identity theory; parochial altruism) — “in-group preference and out-group suspicion” as a co-evolved disposition. Its mechanism is precise and corrosive: the ultimate attribution error, by which the same act is judged leniently when an in-group member does it and harshly when an out-group member does. This walls the radius of trust, blocks bridging conversion, and underlies what is loosely called tribal or ethnic favouritism — better named, analytically, bounded reciprocity. Cross-national evidence shows it is measurable and varies with the rule of law, so it is an endogenous variable the model can move, not a fixed cultural constant.
Governs: trust radius / bridging-conversion rate; couples to KT (the claim network is the bonded set) and to the conflict gate. Constrained maximisation running inside a wall: the policy task is extending disposition-legibility across the boundary, not removing the in-group bond.
2.3 The tunnel effect & its reversalHirschman & Rothschild 1973
Cooperative forceHirschman’s tunnel effect: people tolerate others’ advancement — even rising inequality — so long as they expect their own turn to come, just as a driver in a stalled tunnel lane takes the other lane’s movement as a sign the jam is breaking. Tolerance buys the time during which uneven early development can proceed without losing legitimacy. It is a patience reservoir, and it is real: societies do extend it during credible take-offs.
Negative mirrorThe reservoir drains. In Hirschman’s words, gratification turns to indignation when one’s own lane stays jammed while others move — tolerance reverses into destabilising resentment, and “the greater the tolerance, the greater the scope for reversal” if inequality is not corrected in time. Critically, social fragmentation blocks the tunnel effect entirely: where a society is segmented by ethnicity, language, or religion, another group’s advance never reads as a signal of one’s own coming turn — it reads immediately as exploitation. This is the bridge between this mechanism and 2.2: in a fragmented society the patience reservoir never fills.
Governs: legitimacy / patience reservoir — whether inequality is motivating or destabilising; couples to fragmentation (2.2) and to the conflict gate’s endogenous risk. A minimax-relative-concession violation made dynamic: tolerance holds while expected relative benefit stays positive.
2.4 Cooperative surplus & distributive captureGauthier 1986; Olson 1993
Cooperative forceThe core state transition the model studies: moving from zero-sum distributive conflict over a fixed pie to cooperative generation of surplus, divided by minimax relative concession from a non-cooperative baseline. This is Gauthier’s morally-free zone of mutual advantage, and it scales: trade at the market level, settlement at the political level, financing structure at the programme level. When parties move from fighting over shares to growing the pie and sharing the growth fairly, cooperation becomes self-sustaining.
Negative mirrorDistributive capture / rent-seeking: the surplus is fought over rather than grown, and division violates relative benefit — a party (the bandit, the rent-holder) takes a share wildly out of proportion to contribution. Olson’s logic explains when this is stable: the secure-but-extractive ruler with a short or contested horizon, who suppresses prosperity he cannot control because independent wealth funds rivals. Capture holds only as long as exit and voice can be suppressed.
Governs: the zero-sum → surplus transition, the model’s framing concept; couples to RI (rents enable capture by feeding elites off-baseline) and to FDP (broad productive taxation ties elites to surplus growth). Gauthier’s proviso — do not better yourself by worsening others — is the boundary capture crosses.
2.5 Capability expansion & capability deprivationSen 1999; entitlement theory
Cooperative forceDevelopment as the expansion of capabilities — what people are substantively able to do and be. Capability is both the model’s honest output measure and the precondition that lets every other mechanism run: an agent must be capable to read dispositions, transact, and climb the tiers. Building capability (literacy, numeracy, agency, skills, the institutions that produce certifiers and adjudicators) is what makes interventions deployable.
Negative mirrorCapability deprivation: where people lack the substantive freedoms to participate, interventions do not engage regardless of design quality — the programmable instrument reaches no one, the certification ladder strands its lowest rung. Entitlement collapse is the acute form: famine as the breakdown of exchange relations through which people command food, not as a failure of production. Deprivation is a precondition failure that turns otherwise-sound interventions inert or negative.
Governs: capability as gating precondition on every intervention, and as the composite’s ultimate output; couples to the human-capital/skills layer and the university-as-enforcement-producer thread. Develop-and-deploy: capability is the substrate that determines whether instruments engage at all.
2.6 Long shadow of the future & the short horizonfolk theorem; Axelrod 1984; Olson 1993
Cooperative forceThe formal engine under everything above. The folk theorem: in repeated interaction, cooperation is sustainable as a rational equilibrium when players weight the future heavily enough — the “shadow of the future” is long. Axelrod’s winning strategy operationalises it: nice (never defect first), retaliatory (defection is met, so it does not pay), forgiving (return to cooperation after a single defection, avoiding the death-spiral), clear (legible enough that others can recognise and reciprocate). These four properties are, independently arrived at, the framework’s graceful-failure-plus-asymmetric-liability design.
Negative mirrorWhen the horizon is short — political instability, expropriation risk, no expectation of repeated dealing — defection dominates and cooperation cannot get started. Olson’s short-horizon ruler is this at the level of the state. An unforgiving variant (pure grim retaliation) is also a trap: a single error triggers permanent breakdown, which is exactly why graceful failure (forgiveness) is a required design property, not a softness.
Governs: the time-horizon parameter at agent and elite levels; couples to EH (elite horizon) and to every reputation mechanism. THEORETICAL Note: the literature qualifies that tit-for-tat’s success is not universal across all populations (Binmore); the framework takes the four properties as design guidance, not tit-for-tat as a universal law.
3. How the Pairs Connect
The mechanisms are not independent. Read together through the spine, they form one system, and the connections are where the model’s explanatory power lies.
The defection trap (2.1) is the general failure, and the others are its specific routes. In-group boundedness (2.2) is constrained maximisation that works, but only inside a wall — cooperation is real yet cannot scale, so the society sits in a bounded-trust equilibrium that looks cooperative locally and predatory across groups. Tunnel reversal (2.3) is what tips a tolerant society toward the trap when relative benefit turns negative, and fragmentation (2.2) is what prevents the patience reservoir from filling in the first place — so 2.2 and 2.3 are coupled: ethnic segmentation both walls trust and blocks the tolerance that might buy time for conversion. Distributive capture (2.4) is the trap at elite level, stabilised by rents (which sever the elite from the cooperative surplus) and broken by fiscal dependence on that surplus. Capability deprivation (2.5) is the precondition failure that keeps agents out of the cooperative game entirely. And the short horizon (2.6) is the formal condition under which all of these collapse at once.
The unifying claim, in the spine’s terms: a society escapes the defection trap when enforcement infrastructure makes dispositions legible (so constrained maximisers can find each other — Gauthier), when that infrastructure is institutionally durable (North), and when enough agents have the capability to participate (Sen) — all under fair division (minimax relative concession) that keeps the patience reservoir full and denies capture its hold. Each mechanism in Section 2 is one face of that single transition, and each negative mirror is one way it fails. The simulation layer (Page 3) turns these into thresholds and lets a scenario be run to see which way it tips.
What this layer deliberately omits. Honesty about scope: the framework does not model individual psychology below the level of these aggregate dispositions, does not claim the six pairs exhaust the forces acting on a society, and does not treat its mechanisms as culturally universal in magnitude — the strength of in-group boundedness, the depth of the patience reservoir, and the length of typical horizons are calibrated per scenario, not assumed. Further mechanisms (for example, gendered intra-household bargaining, already used in the instrument layer; or religious and normative-ethical commitment, present in the Nordic re-reading on Page 1) slot into the same pair-plus-variable form. The model is an argument about which forces are most load-bearing, offered for testing — not a census of all forces that exist.
4. The Crosswalk: Mechanisms to Structural Variables
This page and Page 1 describe one model from two sides. Page 1’s variable table (§7) is the machine’s state — the quantities that rise and fall; this page’s mechanisms are the forces that move them. The table below makes the bridge explicit: for each structural variable, the behavioural force that drives it and what that force explains about the variable’s otherwise-asserted behaviour. This is the resolving move — it shows that the foundations layer is not a parallel essay but the explanation of why the structural model behaves as it does. Read it as: Page 1 said the variable does X; here is the mechanism that makes X true.
| Structural variable (Page 1, §7) | Driving mechanism (this page) | What the mechanism explains |
| EIS — enforcement infrastructure stock | Legibility under constrained maximisation (§1.1); monitoring under conditional cooperation (§2.1) | Why building EIS raises cooperation rather than merely recording it: EIS is the technology that lets constrained maximisers identify each other and lets willing punishers see whom to discipline. No legibility, no way to tell cooperators from defectors. |
| IX — impersonal exchange share | Bridging conversion (§2.2); folk-theorem cooperation (§2.6) | Why IX is the conversion variable: it is bonded reciprocity extending across the in-group wall into exchange among strangers. The logistic inflection is the moment recorded reputation becomes readable beyond the bonded set. |
| FDP → SB, EH — the committed-thesis arrow | Cooperative surplus vs. capture (§2.4); long horizon (§2.6) | Why the arrow holds: elites fed by the productive surplus acquire Olson’s encompassing interest (horizon lengthens) and a reason to admit producers into the bargain (breadth widens). The reverse — rents — feeds capture instead. |
| RI — rent inflow | Distributive capture (§2.4, negative mirror) | Why rents are corrosive rather than merely neutral income: they sever the coalition from the cooperative surplus, so its fortunes no longer depend on the producing economy — stabilising capture and shortening horizon. |
| FBM — formality-betrayal memory | Defection trap (§2.1, negative); Gauthier’s proviso violated (§1.1) | Why betrayal carries a long tail: a sovereign defecting against the legible signals that dispositions are not safely readable, collapsing the basis on which constrained cooperation rests. Rebuilding legibility is slow. |
| KT — kin-tax intensity | In-group boundedness / bounded reciprocity (§2.2, negative) | Why KT suppresses visible accumulation independent of EIS: the bonded set taxes legible success, so rational agents keep income illegible. This is constrained maximisation running inside the wall — cooperation is real but cannot scale. |
| SB — settlement breadth | Surplus vs. capture (§2.4); tunnel-effect fragmentation (§2.3) | Why breadth widens or narrows: it tracks whether the surplus is shared (admitting groups into the bargain) or captured (closing it). Fragmentation blocks the tolerance that would let breadth widen peacefully. |
| EH — elite horizon | Long shadow of the future (§2.6); surplus vs. capture (§2.4) | Why horizon lengthens or collapses: the encompassing interest engages only when the future is both valued and expected to be captured — long shadow plus a stake in the surplus. Rents and instability shorten it. |
The single sentence the crosswalk encodes. A society escapes the defection trap when enforcement infrastructure makes dispositions legible (so constrained maximisers find each other — Gauthier, driving EIS and IX), when that infrastructure is institutionally durable (North), and when enough agents have the capability to participate (Sen) — all under fair division (minimax relative concession) that keeps the patience reservoir full (sustaining SB and EH) and denies capture its hold (starving RI’s corrosive path). Every structural variable on Page 1 is one quantity in that sentence; every mechanism on this page is one force in it.
5. Epistemic Status of This Layer
Following the framework’s convention, the claims on this page are classified by evidential weight. Experimentally grounded: the existence of conditional cooperators and willing punishers and the role of monitoring (Ostrom’s and others’ public-goods experiments); in-group favouritism and the ultimate attribution error (the social-identity and parochial-altruism experimental literatures, including cross-national measurement); the Axelrod tournament results and the four properties of robust strategies. Historically grounded: North’s institutional emergence; Sen’s entitlement analysis of famine; Hirschman’s tunnel effect and its determinants (with the caveat that its empirical magnitude varies and is contested). Theoretical: the folk theorem (a result, not an empirical claim about any society); Gauthier’s constrained maximisation and minimax relative concession (a philosophical derivation, marked THEORETICAL where it appears); the precise mapping of these mechanisms onto the model’s variables, which is the framework’s own construction and is offered for testing.
The honest qualifications matter and are stated rather than buried. Tit-for-tat’s tournament success is not universal across all opponent populations; the framework therefore takes Axelrod’s four properties as design guidance, not tit-for-tat as a law. Social identity theory is strong at explaining intergroup behaviour after the fact and weaker at predicting it in isolation, so the framework pairs it with the material (resource-competition) account rather than relying on the psychological one alone. Gauthier’s minimax relative concession is one bargaining solution among several (it competes with, and Gauthier argues against, the Nash solution); the framework uses it because its relative-concession logic matches the fairness-and-stability behaviour the model needs, not because the bargaining-theory question is settled. None of this weakens the spine; it shows where the spine is argument and where it is evidence.
6. Forward to the Specification
This page has established the mechanisms and why they are real. Page 3 — Specification & Criteria — turns them into the operational model in the sequence the framework follows: start conditions (country and REC profiles), the simulation mechanics (how the variables move and couple), the criteria (fallbacks, tipping points, performance measures), the instructions, and the applied use-case set that anchors the abstractions in documented reality. Each mechanism named here acquires there its threshold logic and its behaviour in the simulator: the willing-punisher threshold, the bridging-conversion rate, the patience reservoir’s drain function, the surplus-versus-capture condition, the capability gate, and the horizon parameter. The named strategies a scenario can deploy — credible refusal, surplus-building, capability-first sequencing — are drawn from this page’s mechanisms and exercised there. The interactive tool (Page 4) is built strictly on Page 3’s published criteria, so that nothing in the simulation is hidden from the specification that justifies it.
7. References
Axelrod, R. (1984). The Evolution of Cooperation. Basic Books.
Binmore, K. (1994). Game Theory and the Social Contract, Vol. 1: Playing Fair. MIT Press. (On the limits of tit-for-tat’s robustness and a critique of Gauthier.)
Bowles, S. & Gintis, H. (2011). A Cooperative Species: Human Reciprocity and Its Evolution. Princeton University Press. (Strong reciprocity; parochial altruism.)
Fehr, E. & Schmidt, K. M. (1999). A Theory of Fairness, Competition, and Cooperation. Quarterly Journal of Economics, 114(3), 817–868.
Gauthier, D. (1986). Morals by Agreement. Oxford University Press. (Constrained maximisation; minimax relative concession; relative benefit; the proviso.)
Hirschman, A. O. & Rothschild, M. (1973). The Changing Tolerance for Income Inequality in the Course of Economic Development. Quarterly Journal of Economics, 87(4), 544–566.
North, D. C. (1990). Institutions, Institutional Change and Economic Performance. Cambridge University Press.
Olson, M. (1993). Dictatorship, Democracy, and Development. American Political Science Review, 87(3), 567–576.
Ostrom, E. (1990). Governing the Commons: The Evolution of Institutions for Collective Action. Cambridge University Press.
Ostrom, E. (2000). Collective Action and the Evolution of Social Norms. Journal of Economic Perspectives, 14(3), 137–158. (Conditional cooperators; willing punishers; design principles.)
Sen, A. (1981). Poverty and Famines: An Essay on Entitlement and Deprivation. Oxford University Press.
Sen, A. (1999). Development as Freedom. Oxford University Press. (The capability approach.)
Tajfel, H. & Turner, J. C. (1979). An Integrative Theory of Intergroup Conflict. In W. G. Austin & S. Worchel (eds.), The Social Psychology of Intergroup Relations. (Social identity theory.)
Pettigrew, T. F. (1979). The Ultimate Attribution Error. Personality and Social Psychology Bulletin, 5(4), 461–476.