When Fintech Lending Meets Privacy Protection : The Role of Privacy-Insensitive Device Attributes in Identifying Loan Default

We study the role of privacy-insensitive data on device attributes in identifying loan default, under the context of ongoing tension between privacy protection and credit efficiency. Using a unique applicant-level data from a leading Fintech credit institution in China and merging two basic device attributes extracted from applicant’s mobile phone with public information of device itself, we find that device attributes can provide incremental information beyond basic demographics and currently available credit scores. The default probabilities are lower among applicants who use mobile phones with higher price, more famous brand and later released year, and who update the operating system timely. Further, we explore the underlying explanations and find that the value-related attributes are a quantitatively important signal of personal economic status, while the update behavior indeed serve as informational cues of personal cognitive ability. In response to the usefulness of device attribute data, not only do sophisticated applicants have incentives to manipulate the data, but also Fintech lenders will correctly use the information through multiple decision adjustments even though they cannot make full use of all the information in the initial decision making. Our findings highlight the growing importance of non-traditional data in the digital age and the responses of affected participants