Before the data arrive: a pre-entry transferability framework for alternative credit scoring in emerging economies

Abstract Despite rapid digital adoption, approximately 51% of Indonesian adults remain unbanked, and access to formal credit is further constrained by limited credit bureau coverage—a gap that alternative credit scoring can help close, but only if the behavioral signals developed in more advanced markets can be reliably transferred to the local context. This study addresses the fundamental challenge facing financial institutions that seek to extend credit access in emerging economies: how to evaluate the deployability of alternative credit scoring variables across national markets before any local performance data are collected. We propose the Cross-Market Variable Transferability Score (CMVTS), a composite index that combines Jensen-Shannon divergence over variable distributions, Spearman rank correlation over macro-economic indicators, and cosine similarity over a comparative statistics vector. Their weighted combination yields a single score bounded in [0, 1] that maps directly to a three-tier decision rule—HIGH, MEDIUM, or LOW—prescribing whether direct transfer, partial adjustment, or local redevelopment is warranted. Applied to the Korea–Indonesia market pair using a synthetic credit bureau dataset of 3,129,036 records and 34 macro-economic indicators, the CMVTS yields 0.9225–0.9338 across all 47 tested specifications (0.9338 under the baseline specification; 0.9225 under the refined empirical-prior specification), placing the verdict firmly in the HIGH tier. This verdict rests on a single, structurally comparable market pair; whether it reflects genuine market similarity as opposed to favourable pair selection can only be settled by cross-pair empirical validation, which we identify as the principal direction for future work. Nine transferred variables yield projected portfolio figures of a lift of 262.8 and an approved-population bad rate of 0.024% at the 550-point threshold, illustrating the expected risk differentiation achievable through direct variable transfer prior to local recalibration. To our knowledge, CMVTS is the first framework to operationalize a pre-transfer quantitative transferability assessment for credit scoring variables as a composite of information-theoretic, rank-order, and structural similarity measures, applied prior to any target-market data collection. By enabling financial institutions to screen variable transferability prior to market entry, the CMVTS framework can accelerate the deployment of alternative credit scoring in underserved markets—advancing the goals of financial inclusion aligned with SDG 10 (Reduced Inequalities) and SDG 8 (Decent Work and Economic Growth)—two of the United Nations’ 2030 Sustainable Development Goals targeting equitable economic inclusion.