Limitations of algorithms

When it comes to business lending, lenders rely heavily on sophisticated technology to identify the safe bets and automatically manage risk by rejecting those who don't tick the right boxes.

In the not-too-distant past, financial services firms made decisions based entirely on human judgement, however today the majority use different types of algorithms to predict financial markets, identify prospective employees, assess potential customers and offer loan applicants an immediate yes-no decision. Despite the speed and various advances in financial technology, algorithms do impose limitations and have the potential to decline an opportunity based solely on someone’s digital footprint.

Computers excel at separating black from white. While it makes perfect sense for finance firms to employ algorithms to manage risk and automate management of tasks, what happens when it restricts too many people? What happens if the computer says no to you? Algorithms are trained to make recommendations based on data that’s not always representative of the wider picture. So, in a world where there is a great deal of grey, we need to rethink how reliant we are on technology and the impact this has on businesses looking to borrow. 

When assessing potential borrowers, lenders have historically focused on limited types of data that directly relate directly to the likelihood of repayment, such as debt-to-income and loan-to-value ratios and individuals' payment and credit histories. More recently, the emergence of big data analytics has prompted many lenders to consider non-traditional types of data that are less obviously related to creditworthiness to support decision making. The trouble with all of this is that decisions are made on data from the past rather than the potential in the future. 

The goal of an algorithmic system is to eliminate the subjectivity and cognitive biases inherent in human decision-making. While this makes sound business sense to lenders, it means those who don't score well are left with limited options when they look to borrow. If your client has a strong business case, experience and expertise in a chosen sector and an opportunity to exploit, it can be devastating to be told that an algorithm has said no. 

At Accredo, we have embraced technology and seek to use it to maximise our efficiency and manage risk, however, we haven't forgotten the value our team members can add to ensure we lend to the right people. As such, we don’t rely on automated credit scoring; instead, we consider each application on its individual merit. We listen to your customer’s story and weigh up the opportunity to make a rational decision on lending. As a result, this gives a better, more balanced decision that benefits more businesses and means we may offer business finance solutions when other lenders won’t.

Our specialist team is always on hand to answer your queries regarding secured business loans so please do call us on 01444 255915