借款人風險特性與總體經濟變數對金融科技借貸利率的影響

本研究評估借款人風險特性及總體經濟環境對P2P Lending平台利率的影響，並進一步分析P2P Lending平台和傳統銀行之間是否存在競爭或互補的關係。實證上，採用美國最大P2P借貸平台—Lending Club的借款人資料作為研究對象，樣本期間為2016年第一季至2018年第二季，共計1,014,791筆觀察值。選取負債所得比、工作年資、貸款金額、年收入、信用評等、借款目的、貸款期限及房屋居住狀況等八大項目作為衡量個人風險特性的解釋變數，以及S&P500股價指數報酬率及聯邦基金利率作為總體經濟環境變數的解釋變數。本研究首先進行橫斷面異質性檢定，檢定誤差項是否存在異質變異。確定誤差項存在異質變異後，再採用廣義的最小平方法（generalized least squares, GLS）估計P2P借貸的利率決定模型。實證結果得知，借款人的居住狀況（租屋）、負債所得比、信用評等、借款期間等四項風險特性變數對P2P網路借貸平台訂定之借款利率有正向的影響；借款人的借款目的為償還信用卡債務及債務整合、借款人居住狀況為不動產有抵押，以及年收入等四項個人風險特性變數則對P2P網路借貸平台訂定之借款利率有負向的影響。此外，在總體經濟變上，S&P500股價報酬率及聯邦基金利率皆對P2P網路借貸平台訂定之借款利率有正向的影響。最後，本研究提供相關的建議，供P2P網路借貸平台業者、政府及傳統銀行制定決策及訂價策略的參考。在P2P網路借貸平台方面，可以參考本研究所採用的八項個人風險特性變數，並對照各變數對利率影響的結果以進行利率訂價；在政府方面，在風險能有效控管的情況下，政府應儘量維持利率與股市的穩定性，並監督P2P借貸平台的風險評估情形，以免造成借貸市場風險擴大；在傳統銀行方面，可將P2P網路借貸平台所使用之利率定價模型列入銀行貸款訂價策略之參考，且可評估是否可創造與P2P網路借貸平台進行技術合作或數據共享等新商業模式。

關鍵字

金融科技；借貸利率

並列摘要

This study evaluates the impacts of borrowers’ risk characteristics and macroeconomic environment on the P2P Lending rates and further analyzes whether there is a competitive or complementary relationship between the P2P Lending platform and traditional banks. Empirically, the borrower data of the Lending Club, the largest P2P lending platform in the US, are used. The sample period spans from 2016: Q1 to 2018: Q2, with a total of 1,014,791 observations. Eight explanatory variables are used to measure individual risk characteristics, including debt-to-income ratio, working years, loan amount, annual income, credit rating, loan purpose, loan term, and housing living status, and the return rate of S&P 500 and Federal funds rate are used to measure macroeconomic environments. This study first performed a cross-sectional heterogeneity test to determine whether there was a heteroskedasticity in the error term. To resolve the problem of heteroskedasticity, the generalized least squares (GLS) method is used to estimate the determination model of P2P lending rates. The empirical results show that the four risk characteristics－borrower's housing status, debt-to-liability ratio, credit rating, and loan period－have a positive impact on the P2P lending rates while the remaining four risk characteristics have a negative impact on the P2P lending rates. In addition, in terms of overall economic changes, both the S&P500 return rate and Federal funds rate have a positive impact on the lending rates. The associated policy suggestions for P2P lending platforms, governments, and traditional banks to make decisions on pricing are reported as follows. For the P2P lending platforms, the eight individual risk characteristic variables and two macroeconomic environment variables can be used to determine the rates of the P2P lending. For the governments, in the case of effective risk control, it is necessary to maintain the stability of interest rates and the stock markets and to supervise the risk assessment of the P2P lending platforms so as not to expand the risk of the lending market. For the traditional banks, the pricing model of the P2P lending analyzed in the study can be used as a reference for their loan pricing strategies. Furthermore, the traditional banks also can assess whether new business models can be created through technical cooperation or data sharing with the P2P lending platforms.