The Small Merchant Fraud Gap: Why US Businesses Under $10M in Revenue Absorb Disproportionate Fraud Losses.
DOI:
https://doi.org/10.70917/ijcisim-2026-6101Keywords:
payment fraud, small and medium-sized enterprises, chargebacks, card-not-present fraud, retrospective audit, synthetic data, Shapley decompositionAbstract
In this last decade, industry surveys, central-bank studies and the information-systems literature converged on a working explanation for why US merchants with less than $10 million in annual revenue appeared to lose a larger share of their sales to payment fraud than larger firms. That explanation was never tested as a whole. This study was designed and written in 2025 as a retrospective; it makes no claim to have been produced in the period it studies. We assemble an era-bounded corpus of 54 sources published between 2017 and 2024, extract eight claims the period treated as settled, and test each against a seeded synthetic merchant panel whose data-generating process instantiates the era's own documented assumptions about channel mix, liability allocation, detection tooling, dispute recovery, false declines and large-event losses. A Shapley decomposition with common random numbers attributes a pooled 19.9 basis-point gap in retained loss ratio between the two size bands to five mechanisms. Channel exposure to card-not-present commerce accounts for 36% of the gap, under-adoption of detection tooling for 21%, weaker dispute recovery for 18%, deliberate targeting for 6%, and EMV-related liability lag for 2%; the remaining 17% is carried by rare large-value events. Six of the eight era claims are held, two are qualified, and none is overturned, but the audit reverses the era's emphasis: the liability rule that the period blamed does not differ by firm size, and the targeting narrative that dominated vendor reports explains almost nothing. Evidence published in 2025 and 2026 is consistent with these verdicts. All data are synthetic, and all code is released.