The next time you ask for a home loan, a robot rather than a bank manager may make the decision.
“Robo-lending” or “algorithmic lending” as it is known in financial circles, is becoming increasingly popular, with mortgage companies claiming it reduces the time it takes to approve a home loan from days to minutes.
Traditionally, only an experienced credit manager could fulfil this role. Such a person would be a good judge of human character, and able to distinguish between responsible borrowers and those unlikely to meet their repayment schedules. They would make their decision following careful consideration of the existing relationship between lender and customer, and the size of the requested loan.
Conversely, robo-lending enables computer programs to rate the creditworthiness of prospective borrowers and whether they are capable of consistent loan repayments. Being data driven, it does away with pleasantries and heuristic judgements, instead rapidly analysing data to assess the individual’s credit rating and delivering a clear response within minutes.
Robo-lending capabilities use data taken from previous loan outcomes together with fiscal parameters to make their credit assessments, matching the applicant’s information to a historical record with incredible accuracy.
“The robo-lending programs utilise a technique in computer science known as ‘machine learning’, where the computer program has vast amounts of data on historical lending practices – each application, decision and whether or not the loan defaulted,” explains Head of Accounting at Curtin University Professor Saurav Dutta.
“Through pattern recognition techniques based on the data, it develops a profile of reliable borrowers and borrowers likely to default. Each new case is compared to the profile generated by the program based on historical data. When the new case is similar to the profile of a reliable borrower, credit is granted.”
Specific data detailed in each applicant’s profile is also identified. For example, if the application is filled in by hand and scanned into the computer, the algorithm may consider whether the application was written in block capitals or in cursive handwriting.
“The algorithm may have detected a pattern that applicants who write in all-caps without punctuation are usually less educated with a lower earning potential, and thereby inherently more risky,” says Dutta. “Who knew that how you write your name and address could result in denial of a credit application?”
Critics contend that quirks such as this leave the effectiveness of robo-lending open for debate as inherent biases may be incorporated in loan parameters set by lending institutions. But Dutta disagrees, believing machines can actually judge human behaviour better than humans themselves.
“A large part of human decision-making is based on the first few seconds and how much they like the applicant. A well-dressed, well-groomed young individual has more chance than an unshaven, dishevelled bloke of obtaining a loan from a human credit checker. But an algorithm is unlikely to make the same kind of judgement.
“Computers make lending decisions based on objective data and avoid the biases exhibited by people. When you look at the research, it doesn’t seem that humans are that great at judging financial risk. Algorithms may be flawed but on balance, computers come out ahead,” he says.
While humans set initial benchmarks for the algorithms, robo-lending machines are able to learn from new loan applications and adjust their assessments to fit different circumstances.
“Machine learning tools analyse data deeply and in detail. They also are capable of learning over time, for example, by making changes to variables as patterns emerge, or incorporating macroeconomic changes into their assessments.”
For banks and credit institutions, the switch to robo-lending makes perfect sense. Lending is risky and minimising those risks is key to maintaining a healthy business.
“Banks have a delicate balancing act – they always want more borrowers to increase their income, but they need to screen out those who aren’t creditworthy. The obvious risk is payment default. The bank lends the money but never collects the interest or the principal. Less obvious is the risk of a missed opportunity – the lender denies a loan to a creditworthy borrower. It is about managing the risk in a systematic way,” says Dutta.
A customer can also benefit from robo-lending. With the Australian Banking Royal Commission uncovering instances of financial malpractice, customers may be more likely to trust a robot that, in theory, will not misinform them about their ability to repay the loans on offer. They may even be provided with a range of realistic credit opportunities they may not have otherwise considered.
Such is the predicted demand, large agencies like the US-based Ford Credit are introducing machine learning loan approvals that operate across the entire credit spectrum.
Other credit agencies are introducing ‘robo-loan-officers’ that take the form of a virtual assistant who is available every day of the week to help a customer identify financial opportunities, and may even initiate conversation with them.
In 2017, online banking provider UBank launched RoboChat, Australia’s first artificially intelligent virtual assistant for online home loan enquiries. NAB followed suit in 2018 with its virtual assistant for superannuation, accessible on a range of Google Home devices.
This trend towards loan automation is likely to continue with financial institutions fine-tuning this technology for rapid implementation and end-to-end loan processing by robot expected to be commonplace within three to five years.
So the next time you want an unbiased assessment of your financial situation or want to buy that block of land, you probably won’t need to dress up for an appointment with your bank manager. Just switch on your smart device and ask for Ms Robot.
Related article:
