Can you explain your recent research about the value of privacy in online bank loaning?
My recent research paper, “The Value of Privacy: Evidence from Online Borrowers”, studies the monetary value consumers attach to the privacy of their personal data. Whether and how much consumers value data privacy are not obvious questions. For example, previous studies based on lab experiments and survey data often find small to zero value for privacy. This is commonly referred to as the “privacy paradox”, i.e., the claimed intentions to protect privacy divert from how people actually behave online.
I investigate this question in a real life context, where online borrowers have to relinquish personal data in exchange for consumer loans.
I investigate this question in a real life context, where online borrowers have to relinquish personal data in exchange for consumer loans. Because there is no official credit score in China, online lending platforms commonly require applicants to disclose extensive amount of information, in order to assess their creditworthiness. This implies that individuals borrowing from such platforms face a clear tradeoff between data privacy and credit access. In this context, their preferences for privacy can be truthfully revealed by their actual choices.
How do you quantify the monetary value of privacy?
To quantify the monetary value of privacy, I exploit large-scale field experiments, which involve around 320,000 participants, on a major Chinese peer-to-peer lending platform. These experiments randomly vary the amount of information required from applicants and the cost of borrowing. Applicants’ decision to disclose or not then truthfully reveals their preferences for privacy. In particular, the experiments allow me to measure applicants’ sensitivities to disclosure requirements and to cost of borrowing. By contrasting these two sensitivities, I back out the monetary value of privacy. The $33 value in the headline result can be interpreted as the fee reduction required to exactly offset the drop in loan demand cased by additional data questions in the application procedure.
What have you found?
The headline result is that online borrowers in China are willing to share their social network ID and employer contact in exchange for a $33 reduction in loan origination fees. This is a non-trial amount considering that it represents 70% of the average daily salary in China.
In addition, the platform makes 10% less in expected revenue by requiring applicants to disclose these additional data. This is because such data policy discourages a significant fraction of applicants from completing the applications.
These results suggest that maybe there is no “privacy paradox” after all. When we measure the value of privacy in a real-life context, the number can be far above zero. This paper also contributes to the ongoing debates about whether the recent privacy protection regulations, such as the General Data Protection Regulation (GDPR) and the California Consumer Privacy Act (CCPA), benefit consumers or not.
What are the implications for FinTech companies and for customer’s privacy?
A key innovative aspect of the FinTech business model is to use customer data to generate revenue, either by using the data to improve the risk model, by profiling customers and sending them targeted advertising, or by selling the data to third parties. The overall impact of data collection on firm profit depends on how much consumers value data privacy and how useful the data is for generating revenue.
The overall impact of data collection on firm profit depends on how much consumers value data privacy and how useful the data is for generating revenue.
My paper shows that there are cases where collecting more data is socially inefficient. This is when the data items are valued to a large extent by consumers but have little economic benefit for the platforms. Therefore, it is a win-win situation if FinTech firms can identify such data items. The framework provided in this paper can be easily generalized and exploited by firms to conduct this type of analysis.
My research paper shows that there are cases where collecting more data is socially inefficient.
A word by Professor Johan Hombert, Huan Tang’s thesis supervisor
Huan’s research on privacy is making an impact because it provides a methodology to put a number on how much individuals value privacy. For example, Huan has been able to quantify that borrowers on a Chinese peer-to-peer lending platform needs to be given more advantageous loan terms worth $33 in exchange for revealing their ID on Tencent QQ, the Chinese equivalent of WhatsApp.
Researchers previously calculated such numbers using data from surveys and lab experiments. However, such data suffer from well-known methodological problems such that results derived from these data can only be taken with a grain of salt. Instead, Huan’s methodology rely on “field data”, that is, data from actual borrowers taking actual loans, which are considerably more reliable, especially for quantitative analysis.
Huan’s methodology can be used by FinTech platforms and more generally by B2C internet platforms to assess how much business they could gain by optimizing the personal information they require from customers.
The HEC Paris Finance Department is proud to see the impact Huan’s research is making. Huan is now joining the London School of Economics as an assistant professor of finance, thereby continuing the exceptional track record of placement of HEC finance Ph.D. graduates (in recent years, HEC finance Ph.D. graduates were hired by universities such as MIT, Harvard Business School, Princeton and Wharton).