P2P대출 차입자의 특성과 의사결정 관계 연구 : 언어표현 정보를 중심으로

This study investigates the influences of diverse sources of borrower information on the decision-making of both borrowers and lenders and actual loan repayment performance, especially focusing on language use factors, a sort of important atypical information in peer-to-peer (P2P) lending market. To this end, I collect data for 33,650 individual loan requests from Moneyauction, the largest and oldest P2P lending platform in South Korea, and analyze the determinants of loan cancellation and funding success adopting binary logistic regression models, and try to use the opportunism theory, trust building theory, and the cognitive load theory for interpretation. In addition, this study suggests prediction models assessing whether potential borrowers will withdraw their loan requests on the way of bidding process or not and whether loan requests will be fully funded or not. The research findings show that the borrower’s demographic traits (e.g., gender, age, and marital status) and linguistic style (e.g., word count, number of money words, and causal words) have as much of an influential effect on participants’ decision making (i.e., loan request cancellation and funding success) as do financial and loan request profiles (e.g., credit grade, loan amount, interest rate, and duration), which are generally considered in credit assessments done by conventional financial institutions. This study also suggests that participants’ decision making is significantly influenced by language factors including description length and the frequency of words expressing financial situation and causal relation.