**3. Data, Model and Variables**

The data we used is from one of the world's pioneer P2P platforms, RenrenDai, which was established in 2010. By October 2016, the total amount of its transactions exceeded 21.2 billion yuan. The platform targets microloans, 71,000 yuan being the average loan amount. The platform consisted of 251,887 listings from 2010 to 2014. Borrowers fill out a loan application online to be published on the website. Peer investors conduct their own credit analyses and choose which loans to invest in. The funding process is completed when the entire loan amount has been filled by investors. Like crowdfunding, a single loan may have multiple investors. Thus, among the total listings, only 65,394 loans were funded. The borrowers can repay the loan in full or in monthly installments until it matures. Among the funded loans, 50,819 loans are still in the repayment process and 14,575 loans have reached maturity. In the finished loans, 13,901 loans completed the repayment process while the other 674 defaulted, representing a relatively modest default rate of about 4.2%. Detailed variable descriptions are presented below.

Since the dependent variable is binary, we use the logit model to test the determinants of loan funding and default in P2P lending. Our models are presented below: *Model I:*

Logit (Funded*i*) = β<sup>0</sup> + β1*Hard In f ormtioni* + β<sup>2</sup> *So f t Inf ormtioni* + ∝ *Control Variablesi* + ε (1)
