Fraud costs in Asia Pacific has grown 10%-16% for businesses over 2019 pre-pandemic levels, said LexisNexis Risk Solutions recently.
For every US $1 lost to fraud costs an organisation an average of $3.99, compared to $3.50 in 2019, according to the company’s study that surveyed 387 risk and fraud executives in Malaysia, Philippines, Singapore, and Thailand in the February and March this year.
- The cost of fraud per transaction was higher than average, costing digital banks and alternative lenders $6.33.
- Other new payment channels, such as buy now pay later (BNPL) and digital wallets, cost businesses $4.75 for every dollar lost to fraud.
- BNPL providers recorded a 65% jump in new account creations, which aligns with the strong growth in remote online and mobile transactions.
- However, BNPL providers also account for more than one-tenth of payment losses, which is disproportionately higher than the average volume of transactions through other payment channels.
- The top contributor to the rising fraud costs in Asia Pacific remains the inability to identify synthetic identities and verifying and authenticating identities using attributes such as phone numbers, email addresses, behavioural analysis and devices.
- Ecommerce merchants in particular find identity verification challenging since it requires finding a balance between providing a seamless customer experience and implementing step-up authentication and security measures.
Management framework to lower fraud costs in Asia Pacific
According to the company, almost all digital banks and alternative finance providers including BNPL and digital wallets have not yet fully integrated cybersecurity and operations into fraud prevention processes.
In addition, survey findings show that organisations are not widely using AI and machine learning (ML) models for fraud detection, weakening mitigation efforts, LexisNexis pointed out.
The firm said that the percentage of organisations using various capabilities to fight fraud are as follows:
- Rules-based approaches – 52%
- Crowdsourcing – 36%
- Social media intelligence – 33%
- Cybersecurity alerts – 25%
- AI/ML models – 21%