Substack

Wednesday, January 9, 2019

"Big data" ain't "strong data"

This is very significant. Evidence from Alibaba's experimentation with Sesame credit scoring system raises doubts about the value of big-data based credit-scoring systems. Since Ant Financial launched it in January 2015, it still has not made lending decisions using Sesame credit scores. 
Sesame, which is an opt-in feature of the Alipay mobile payments app, draws upon the biggest pool of non-traditional ratings data in the world. It synthesises details from hundreds of sources — ranging from purchases on Alibaba’s Taobao marketplace to subway fares — into a single trustworthiness number for each user, called a “Sesame score”. But one Ant Financial employee conceded there was a difference between “big data” and “strong data”, with big data not always providing the most relevant information for predicting behaviour, and analysts say the best predictor of whether someone will default on a loan in future is often their previous loan repayment history, rather than their likelihood of returning a rental car. Banking and transaction data remain fundamental to predictive credit scoring... Martin Chorzempa, a fellow at the Peterson Institute of Economics think-tank, agreed, saying trustworthiness is “very context specific”. “Someone evading taxes might always pay back loans, someone who breaks traffic rules might not break other rules,” he said. “So I don’t think there is a general concept of trustworthiness that is robust.”... Critics have also questioned how Sesame crunches thousands of data points into a single score, saying there needs to be a strong correlation between hundreds of different behaviours — from trashing a hotel room to stealing a mobile charger — in order for the metric to be meaningful.
And the Chinese authorities have even barred private credit rating agencies, 
earlier this year, China’s central bank stopped allowing independent companies to provide credit ratings, and required all credit ratings to be given by a new public body called Baihang — effectively ending Sesame’s credit ratings business altogether.
This has forced Sesame to pivot to other use cases,
Sesame’s business now relies on tie-ups with hundreds of other companies. When a company joins the Sesame platform, it gives Sesame some of its user data, which helps inform Sesame scores. In return, Sesame helps the companies decide how likely different users are to violate their contracts. The companies often give perks to users with a high Sesame score, such as deposit-free rentals for umbrellas, bicycles or even apartments. Ant Financial will regularly promote the discounts on the Alipay mobile payments app... Analysts have compared the platform to a glorified loyalty card, rewarding users for buying Alibaba’s and its partners’ products, but not predicting behaviour.
This is a cautionary tale for the hundreds of start-ups in developing countries, supported by venture capital, seeking to use various types of digital activity trails to develop credit worthiness measures for poor people and for small entrepreneurs. If Alibaba with its vast treasure of data of all kinds is not able to establish credible enough correlations to make lending decisions, then it does not bode well. 

No comments: