Rapid growth in peer-to-peer lending has introduced significant default risks that are anchored in platform operations under restricted and varying supervision regimes, as well as in the growing dependence on algorithm-driven lending practices that may increase loan mispricing and underestimation of systemic risk. Hence, to better understand current practices and challenges, the paper presents a systematic literature review on credit risk management in P2P lending. The review is based on 453 articles, 428 of which were analyzed in a bibliometric analysis, and 280 analyzed in a content analysis. The analysis includes thematic mapping and identification of trends and gaps in terms of theoretical, contextual, and methodological aspects of the studies reviewed. The findings suggest that the research is dominated by two streams: one focused on predictive optimization of default risk using machine learning, and the other on identifying predictors of default likelihoods using regression analyses. Both streams often rely on public platform data, focus on consumer lending in few contexts (75% of studies in China and the USA), engage in quantitative analyses (97% of studies), and have weak connection to theory (56% of studies). Opportunities for future research relate to strengthening theoretical anchoring, using alternative data sources and study contexts, and enhancing methodological plurality. Finally, the study presents a new integrative framework summarizing the key factors found to be related to default in consumer and business lending, while aggregating them by their level of analysis. This framework can both inform and shape future research efforts.
Zhu et al. (Sat,) studied this question.