Predictive model of traceability adoption in small-scale coffee systems: the case of Nicaragua (2021–2025)

Abstract

Traceability has become a condition for accessing premium international coffee markets, yet few studies provide models capable of anticipating its evolution or guiding public policy. This study develops and validates a predictive model of traceability adoption in the Nicaraguan coffee sector, aiming to identify the factors with the greatest explanatory power and to project their behavior through 2028. A purposive sample of 50 value-chain actors was used, on which a multiple linear regression model with actor-type dummy variables was estimated, verifying the assumptions of normality, homoscedasticity, independence, and multicollinearity, and validating the model through 5-fold cross-validation (mean out-of-fold R²=0.777). The model explained 86.5% of the in-sample variance (R²=0.865; adjusted R²=0.843). Technical training was the strongest and most precisely estimated predictor (β₁=0.3815), followed by the growth of the specialty-coffee niche (β₄=0.2912, estimated with wider uncertainty); belonging to a cooperative or an exporting firm added, respectively, 12.45 and 16.78 percentage points of traceability relative to individual producers, and under moderate continuation of current trends, forecasts place average traceability at 87.3% by 2028. Technical training and strengthened associativity emerge as the levers with the greatest predictive impact, and their implications for public policy are discussed.