Overcoming the Liability of Smallness: Export Associativity and Hypothesis Testing in the Peruvian Alpaca Sector

Abstract Global trade dynamics create scale, financing, information, and technology challenges for Micro and Small Enterprises (MSEs), making inter-firm cooperation a potentially important mechanism for export participation. This study examines the structural and relational determinants of export-oriented business associativity among specialized alpaca-fiber garment MSEs in Arequipa, Peru. A quantitative, non-experimental, cross-sectional design with an explanatory-evaluative focus was applied to a complete census of 15 export-oriented MSEs participating in PROMPERÚ regional export articulation programs. Data were collected with a structured questionnaire containing 16 multi-item operational variables measured on a five-point Likert scale. Analytical procedures were conducted in IBM SPSS Statistics v.27 and combined heuristic Principal Component Factor Analysis (PCFA), Kendall's τ_b and τ_c non-parametric rank correlations, k-means cluster analysis, and exploratory discriminant analysis. PCFA indicated factorability of the census data (KMO = 0.714; Bartlett's test χ² = 142.36, p < 0.001) and extracted three components explaining 72.2% of cumulative variance: Trust and Mutual Cooperation, Export Capacity and Vision, and Governance and Opportunistic Behavior. The hypothesis tests showed statistically significant positive associations between mutual trust and joint export performance (τ_b = 0.642, p = 0.008) and between institutional leadership and joint export performance (τ_b = 0.589, p = 0.012). Individual achievement orientation was not significantly associated with lower opportunistic behavior (τ_c = 0.215, p = 0.183), while operational interdependency was not significantly associated with associative commitment (τ_b = 0.148, p = 0.363). Cluster analysis identified High Associative Maturity (40%) and Reactive/Individualistic MSEs (60%). The results support an interpretation of associativity as a governance and social-capital problem as much as a resource-pooling problem. Given the census design and small population, the statistical findings are interpreted as evidence within the target stratum rather than as population-wide causal estimates.