Our client, an international bank, sought to reduce the default rate for small loans in emerging markets. Modeling risk in this environment presented a significant challenge for the bank, as most customers had a limited digital footprint. As a result, less reliable demographic information was used to assess creditworthiness.
We developed a risk prediction engine that more appropriately determined credit risk by basing it on customer behavior. Our models used a variety of data sources to better identify and understand the key behavioral indicators that drive loan repayment.
Integrating our model into the underwriting process cut the bank’s default rate in half, and is projected to create savings of tens of millions of dollars annually. Our team was able to show this company the ways in which behavioral data works as a better indicator of financial reliability than traditional risk measurements.