Why fintech is may not be ready to disrupt lending
Automated lending decisions might one day replace old-fashioned credit evaluations, but not yet
Article Published in Financial News Jan 2018
A lot of claims are being made by a variety of fintech banking industry disrupters. Conferences echo to the sounds of visionaries expounding on great plans for creaming off profitable areas of banking activity by employing more powerful and efficient technology.
But of the many things that banks do, lending to small and medium sized companies (‘SME’ lending) is one of the more complex areas to emulate. It is also politically charged. As such, there is some evidence that the disruption attempted in the area of lending is itself being disrupted by the practicalities of a complex area of finance.
The relatively small amount of sector lending derived from startups faces a variety of practical hurdles. The smallest companies and businesses often have restricted access to finance. This is the area probably ripest for new lending sources. However, some of the current fintech hype glosses over the inherent riskiness of SME lending and high failure rate of a lot of borrowers in the sector. This riskiness is reflected in the high capital charges bourn by banks against their lending, although politicians keep pushing to reduce these, seeing SME lending as the motors for employment growth.
Understandably, where bankers feel reluctant to lend on account of the capital cost, risk or even lack of capacity, the fintech sector jumps in, arguing that lending risk just needs a good set of algorithms to fix the issue. Evidence so far, however, suggests that algorithms have not yet safely cracked the mysteries of lending to a diverse, and often speculative sector of the economy where ideas, dreams, hype and numbers often end up in a shapeless pile of financial wreckage. Moreover, the cost of lending can end up being high for the borrowers.
While SME lending can certainly be profitable, losses in economic downturns can make this an unpredictable business. Add poor lending decisions, inadequate surveillance or weak advice to struggling borrowers and substantial losses can pile up. Even when employing old-fashioned loan officers and credit systems, mistakes are easily made. HBOS failed as a bank, in part, on account of its poor lending practices, encouraged by a technocratic management of non-bankers and very weak risk controls.
Credit evaluation is not just a question of analysing numbers, but evaluating business plans, competitive situations, quality of cashflows and assets, capital structure, financial controls, management calibre and a host of other parameters. Some Fintech lenders now have reintroduced experienced risk officers to supplement their algorithms. Others find that they need to constantly re-write the algorithms to correct for deficiencies. This iterative process comes at a cost and involves risk.
The tech firms are themselves SMEs with funding models that have not yet demonstrated an ability to scale banking funding and importantly, leverage. This limits their reach. Indeed, should they really try to do what banks do, it is only a matter of time before they end up looking like and being regulated as banks, sending their core costs rocketing. Being light on regulatory oversight, long on imagination, low on overhead and fleet of foot is part of being an innovative fintech company. Yet investors might wonder if this is compatible with much of financial services. There are also questions over the cost of the funds the new lenders are able to source to on-lend. Banks may have expensive business models, but they can source liquidity at reasonable rates.
While automated lending decisions might one day replace old-fashioned credit evaluations, that moment is not quite here. Rising defaults and losses are a hallmark of unsound lending practices. Moreover, reliance on algorithms remains controversial in view of the experience of economic projections, investment decisions and market reactions during the financial crisis. More work is needed to get the right business model for this line before disruption becomes meaningful.
Expect investor faith in 2018 to continue to show some support for the technologist/ coder trying to challenge the old lending order. New lenders may offer fresh insights to stale banking incumbents, which will provide fresh clues to how disruptive fintech lending and crowd funding might potentially become. But there is more to this numbers game than coding to lend credibility to the sector and investors may struggle to spot the successful business models at this stage.
Tim Skeet is a career banker in the City