基于信号理论与社会学习理论的P2P网络借贷模式研究

The past decade has witnessed a growing number of business models that facilitate economic exchanges between individuals with limited institutional mediation. One of the important innovative business models in digital finance is online P2P lending, which have received widely attention from government, industry, investors, and researchers. According to prior literature, one of the most critical problems for P2P lending is information asymmetry which leads to many trust issues. Our research attempts to shed some lights on trust mechanisms in P2P lending. In particular, our research focuses on three specific issues: general trust building towards P2P lending platform, lender’s trusting beliefs and behavior towards borrowing requests, and trust-assuring mechanism provided by the P2P lending sites. First, a customer’s trust in an online P2P lending site is considered to be an important factor for P2P lending business success. However, the question of where customers collect information about an P2P lending site to base their trust on has not been addressed. Hence, we investigate various processes through which evidences are collected from different sources and trust is built. A framework of four learning processes is proposed based on social learning theory and signaling theory. Second, a large majority of prior research on P2P lending are based on data from western countries. Despite many critical differences between China and other markets, studies addressing the P2P lending in China are limited. Drawing upon the dual processing theory and regulatory focus theory, our study investigates the trust-building mechanisms based on which investors trust specific funding requests. Transaction data from PPDAI, an online P2P lending site in China will be employed for data analysis. Finally, by integrating Toulmin’s (1958) model of argumentation and regulatory focus theory, our research investigates the effects from various trust-assuring mechanisms on investor trust, and proposes regulatory focus and personal innovativeness as key moderating variables for different trust-assuring mechanisms. A series of survey and experiments are carefully designed to test our research model. We believe that the rise of online P2P lending will bring new competition to the industry, propel the financial sector in China into a new era by integrating finance with the internet, and promote reform and advancement in business philosophies and models. Our research findings would provide some theoretical and practical implications for the sustainable development of online P2P lending in China.

P2P网络借贷模式是信息技术在金融领域的重要创新。当前P2P 网络借贷存在的核心问题为信息不对称以及由此引发的信任问题。因此，本课题的核心研究目标为：基于信号理论和社会学习理论，科学地分析P2P 网络借贷模式中平台与借款请求的信任机制，为互联网金融实践提供更具解释力的新理论。本课题的研究包含三个模块：1）借助信号理论与社会学习理论框架，采用问卷调研方法，探索P2P网络借贷平台信任的关键影响因素（平台信任）；2）基于双重处理理论与调节定向理论，构建P2P网络借贷平台借款请求模型（借款信任），基于“拍拍贷”平台的交易数据对模型进行分析与验证；3）聚焦于P2P网络借贷平台信用认证的研究（信用认证机制），基于Toulmin的说服理论，探索信用认证机制的不同元素组合对于投资人决策的影响。本课题的研究成果将从理论与实践层面上解释P2P网络借贷模式中信任的建立机制，为该创新模式的发展提供建议与指导。

P2P网络借贷模式是信息技术在金融领域的重要创新。P2P 网络借贷存在的核心问题为信息不对称以及由此引发的信任问题。因此，本课题的核心研究目标为：基于信号理论和社会学习理论，科学地分析P2P网络借贷模式中平台中借款请求的信任机制，在此研究基础上进一步探究P2P网络借贷中的提前还款与异常借款，从而为互联网金融实践提供更具解释力的新理论。具体地，本课题的最终研究内容可概括为四大模块：①以信号理论与社会学习理论为总体理论框架，通过问卷调查法，探索P2P网络借贷平台初始信任建立机制；②基于拍拍贷平台的实际交易数据，探究P2P网络借贷平台中，影响不同类型借款请求达成（信任建立）的关键因素。包含借款人首次借款，无历史投资（出借）经验的借款人重复借款，以及有历史投资（出借）经验的借款人重复借款三类借款类型；③引入内部收益率作为衡量投资者实际收益的指标，基于lending club实际交易数据，探究影响 P2P借贷内部收益率水平的因素，并对内部收益率进行预测，并总结出不同内部收益率水平的借款具有哪些特征，预估实际能够带给投资者的回报率，从而为构造有效组合投资优化提供进一步研究的可能；④基于以上研究结果，构建特征工程，提出“多层次分类”的异常借款识别算法，以lending club发布的交易数据为研究对象，分层次进行数据分析，在保证分类器整体性能的前提下，更加有效地识别出P2P网络借贷中异常还款的借款标的。整体而言，本课题的研究成果将从理论与实践层面上解释P2P网络借贷模式中信任的建立机制（包含初始信任与借款达成信任），并尝试将研究结果应用在P2P网络借贷的实践中，基于P2P网络借贷大数据，进一步为识别提前还款的借款标的和异常还款的借款标的提供了较为有效的理论支撑。最后，基于国内P2P实践的调查，本研究识别了P2P模式的风险，为互联网金融创新模式的发展提供建议与指导。