U.S. Markets close in 4 hrs 55 mins

Brazil's Bradesco widens data trawl in bid to boost lending

By Carolina Mandl

SAO PAULO, Dec 20 (Reuters) - Brazil's Banco Bradesco SA expects to boost many clients' credit limits in 2019 thanks to new tools that will better measure outside income from sources like freelance jobs and rental properties, an executive said on Thursday.

The new measures will boost the assessed value of account holders' total pay by an average of 41 percent, Bradesco Executive Vice President Eurico Fabri said in an interview.

Banks and credit reporting firms increasingly are looking beyond conventional tools to understand clients' default risk.

Brazilian banks, which are struggling to increase their loan books amid a gradual recovery in Latin America's largest economy, traditionally use measures such as paychecks, personal investments and historical loan repayments.

Bradesco, the country's second largest private lender, is also factoring in new data it collects from clients' checking and savings accounts as well as their cable TV and electricity bills.

Every month, Bradesco processes 369 million checking account transactions and 296 million bill payments.

Still, Fabri said it is difficult to determine how much the higher credit limits derived from the new income measurements would impact the bank's loan outlays in 2019.

"Loan book expansion depends not only on supply, but also on demand," he said. "So far it is unclear how much demand will increase."

In a test run within a controlled group in the third quarter, Bradesco said the number of clients eligible for personal loans increased 36 percent. On average, that meant loan outlays nearly doubled compared with a similar group of clients screened using the conventional method.


Data science is particularly important for Brazilian banks as the country does not have an effective credit score system.

Although banks have been discussing with regulators the creation of credit scores similar to the ones used in the United States, so far each bank has had to rely on its own system for determining creditworthiness.

"The absence of this kind of credit score hinders the bank's ability to determine credit limits for individuals," said Robert Duque-Ribeiro, managing director and lead of Accenture Analytics for Brazil and Latin America.

Larger rival Itaú Unibanco Holding SA said it has boosted available loans by up to 25 percent over the past two years thanks to big data and machine learning. (Editing by Christian Plumb and Paul Simao)