Investor Presentaiton

Credit Scoring, Acceptance levels and Collection of Loans Primarily digital identification via mobile and online bank with no physical meetings required1 Scoring and credit policies are centrally steered by the risk team An application scorecard is used to assess new customers and a behaviour scorecard is used to assess repeated customers ■ Based on credit score, customers are grouped into risk classes that ultimately affect the credit decision ■ In its credit scoring the risk team assess internal big data technology (see next slide), public databases, national credit losses registers, statistics databases, and public tax databases if available ■ The scoring model is based on FICO analytics and further developed by the risk team Monitoring systems are in place to accommodate the early identification and management of deterioration in loan quality (daily, weekly, and monthly checks) Source: Company Filings ■ The Group's stringent credit scoring and identification system resulted in an average approval rate of consumer loan applications of 13% during 2018 Approved loans are paid out via bank account money transfer within minutes from application ■ The credit scoring and loan acceptance process is highly effective ■From Q1-2015 to Q4-2018, the Group has increased its customer base by an average of 70 thousand customers per quarter Primarily internal collection employing a series of text messages, letters, and phone calls to encourage customer payment ■ Collection processes are initiated in- house immediately when a payment is overdue and most often outsourced to a third party collection company when the payment becomes more than 30 days delayed Impaired loans may also be sold to third parties Note: 1) In some instances, face-to-face identification is required due to lack of technology in some markets 1 Identification and Credit Scoring Digital identification based on hard facts as name and date of birth. Handled via: ■ Mobile technology Online banking, ■ Face-to-face ID verification 2 Effective check of the customer by means of: ■ External data bases with information on credit-worthiness ■ Tested and dedicated internal scoring ■ FICO-tools ■ Self-learning software: Increasing approvals while decreasing charge-offs 3 Credit decision within seconds: ■ Less than 0.1% fraud cases Professional frauds even 0.0% Consumer loan approval rate 2016 14% 2017 15% 2018 13% ferratum 15

View entire presentation