Evaluation of Contestable Counterfactual Explanations for Artificial Intelligence Agriculture Credit-Scoring Appeals Under Cooperative Patronage: A Governance Recourse-Access Estimation Strategy

Artificial intelligence credit-scoring models increasingly mediate smallholder access to agricultural finance through cooperative patronage, yet the appeals mechanisms attached to these models remain weakly specified as governance instruments. The central problem is that a counterfactual explanation can be technically faithful and still fail as a governance device if the actions it recommends are unavailable to the appellant within the cooperative's patronage structure. The paper formalises recourse access as the joint probability that a counterfactual recommendation is feasible, affordable and institutionally admissible, and proposes a before-after-control-impact design for estimating whether granting appellants access to contestable counterfactuals changes appeal outcomes. Operating characteristics are characterised under a declared variance structure with an intraclass correlation of 0.64, between-cluster standard deviation of 6.0 and residual standard deviation of 4.5; power was estimated across 2000 design cells spanning cluster counts of 8, 16, 24 and 32 and effects of 3, 5 and 8 percentage points. Fifty-two clusters are required for 80 per cent power at a five percentage point effect. The contribution is a governance-oriented evaluation framework that treats explanation contestability as an institutional property rather than a model property, and that specifies the design conditions under which recourse access can be estimated rather than assumed.