Workout Recovery of Platform-Originated MSME Loans in Bangladesh: A Loss-Given-Default Analysis

Abstract

Bangladeshi business-to-business (B2B) marketplaces disburse instant working-capital loans to micro, small and medium-sized enterprise (MSME) merchants at the point of purchase, yet the amount of cash that is actually collected after default has never been disclosed. This study was designed to close that information gap and to equip platform managers with a practical, same-day collection rule. I extracted administrative closed-settlement files from a leading nationwide marketplace covering every merchant account that was charged off during 2023 and that later reached a negotiated pay-off. After institutional ethics clearance and full anonymization, 107 files were analyzed with simple regression techniques. The data show that merchants who had provided a post-dated cheque at origination return markedly more cash than those without such an instrument, while larger exposures also yield modestly higher recoveries. Importantly, settlements that were concluded within three months of default, even at discounts of 30-40%, generated a higher present-value cash inflow than files that were pushed for full recovery but remained stuck in follow-up for a year or more. The findings hold across alternative estimation methods and remain unchanged when collection costs are added. Operationally, the results support a zero-cost policy: discounts should be capped at 20% when a post-dated cheque is held and may be raised to 40% only for unsecured files. Platforms can adopt the rule immediately, expect faster cash-in. The study supplies the first loss-given-default benchmark for platform-based MSME finance in South Asia and offers a replicable template that other marketplaces can apply to their own recovery data without additional software or regulatory approval. Artificial intelligence was used for statistical modelling, language polishing, and APA formatting; analytical decisions remain the authors’ sole responsibility