Since the 1990s, consumer financial lending has been based on automated decision processes. But for businesses, obtaining credit remains largely a manual process with significant human interaction, which brings with it the incumbent human biases. In 2019 that will change, with data democratising businesses’ access to capital. As a result, we will see increased growth among small businesses and, subsequently, the global economies they significantly impact.
Civil rights legislation has ended the explicit exclusion of protected groups from holding certain professions, obtaining credit, owning businesses or buying land. But when it comes to business lending, informal networks of reputation – better known as selection bias – still often dictate whether or not someone is given a loan. Someone who goes to a bank manager’s church, for example, or coaches their children’s football team, can seem like a sound investment, meaning that less formal, and less diverse, social networks function as referral and recommendation services. When human beings are making decisions about who they should lend to, familiarity and relatability between lender and borrower often play an outsized role in risk assessment.
Major changes in demography are accelerating the need for diversification and modernisation of financial services across all users. Millennials, for example, are now much more transient in the jobs market, preferring flexible careers that are often supplemented by gig economy incomes. This makes it difficult for them to develop a reputation with a local community lender, especially when they are trying to set up a business.
But geographic movement is not the barrier to trust it once was. Wherever small businesses are operating, they are generating creditworthiness via millions of data points on online services such as Xero, which automates accounting, and digital-payment company Zelle and on sites such as Alibaba, not to mention on social media.
This data will allow us to remove human assumptions from otherwise objective lending criteria, which in turn will increase access to financial products, including working capital, that businesses need. However, we still need to be cautious. Not all data is objective.
Volatile cash flow, for example, may be a predictive factor for the success of a business, but if the difficulties a company faces are the product of macro- and micro-economic factors, which can be overly influenced by state and local authorities, lending decisions based on those inputs will perpetuate injustices baked into the data. Algorithms are designed, at least initially, by programmers teaching a computer what to value; traditional lending models can perpetuate and reinforce ethnic, racial and other discriminatory biases.
Once we have replaced human biases with algorithmic decisions, automating risk assessment will result in an expansion of access to capital by extending credit to deserving small businesses, who will be judged on their objective business performance data rather than on who they are, what products they create or their customer profile.
Algorithms are able to detect fraud, verify a business’s identity and assess the character of its owner through social data – all in milliseconds. My company, Kabbage, for example, does not meet its borrowers in person nor do we require collateral, but despite the size of our customer population (currently about 150,000 businesses across the US) we have rates as low as 12 per cent, similar to conventional loans. That means that we have taken on sound investment risks, with reliable repayment and returns, which were previously overlooked by traditional banks and lending policies.
In 2019, we will see this technology-led democratisation of access to capital for small businesses give new opportunities to every community where borrowers live and work.
Kathryn Petralia is president and cofounder of Kabbage
– Meet the companies fixing depression by stimulating neurons
– An e-bike revolution is about to upend urban transport
– How companies will use AI tackle workplace harassment
– The blockchain needs protecting from quantum hackers
Get the best of WIRED in your inbox every Saturday with the WIRED Weekender newsletter
This article was originally published by WIRED UK
