Picture the scene: a hopeful homebuyer sits in the almost deserted lobby of a high street bank, waiting for the appointment she booked with the mortgage consultant a week ago – a week ago! It annoys her that she has had to come to a branch she has not visited for years, all because she could not work out how to apply for a home loan on the bank’s website.
Meanwhile, a time zone away, another hopeful homebuyer looks excitedly at his smartphone. After reading about a new mortgage lender on social media, he had downloaded its app and applied online. The process had been slick and the response time short. With a pre-approved loan in his pocket, he could focus on finding a new home.
New Market Entrants
The threat posed by new market entrants, unconstrained by the legacy IT systems of big banks, is as obvious as their appeal to a generation of digital natives. For all the advances in big data, machine learning and computational simulation in the decade since the global financial crisis, incumbent banks, still preoccupied with the twin imperatives of ever-tougher regulatory compliance and boosting shareholder returns, are playing catch-up in their adoption of new technologies.
Yet banks work with the same raw material as new entrants: data, masses of it. The trouble is, banks rarely use it effectively. Too much is still recorded manually, when Artificial Intelligence (AI) could perform the task. Banks talk up the benefits of using AI, but their corporate culture and organisational structure hamper their ability to adopt latest technologies.
Simulation for mortgage analytics solution
Tech has been a catalyst for change for most banks, but just how entrenched is that transformation and are lenders any better at converting data into business advantage? If we look back a decade at Northern Rock, do you think that we’ve actually learned the lessons from that in terms of our analytics and what we’re doing in financial institutions?
The trouble is, mortgage lenders persist in relying on historical macro-economic assumptions in their models so they risk repeating the errors of a decade ago when banks – and their regulators – failed to recognize the warning signs from a far richer source: low-level micro-economic data. The feedback loops may not have been obvious because they were either too complicated to model or the computing power and capacity did not exist.
That may have been the case, but even after sharp declines in property values in the crisis, mortgage lenders persistently fail to track individual agent (in this case, borrower) behaviors resulting in increase mortgage risk. They also fail to model the effects of fear and the risk of contagion. Yet the crisis showed how a loss of confidence in one segment of the market spread rapidly through an unexpectedly interconnected global financial system.
Risk Management 3.0
Agent-based modeling captures these behaviors and the way they drive – or are driven by – crisis scenarios. This would not be feasible with data based primarily on historical assumptions. Simulation models can only be realistic and pinpoint risk hotspots if the dynamics driving them are based on a broad range of real inputs.
These models help lenders to mine myriad inputs within a secure simulation environment to assess the probability of a particular scenario occurring and, if appropriate, prepare a contingency plan in response.
If you really want to consider the likely outcomes of stress tests or visualize the consequences of a distressed Californian housing market or say, Brexit on the other side of the world, you can use these simulations as easily as if you were conducting a conventional internet search, then AI can point to the next best move.
There most likely won’t be a plug-and-play solution to your technology problems. To be useful, AI first requires human intelligence. That, in turn, requires considerable thought within the bank to ensure that the right questions are asked – and the most appropriate data interrogated for mitigating mortgage risk.
There is a need to hone your technology to meet another risk – the risk of losing out to those new market entrants whose apps look set to steal a march on the old-school banks.
Even as incumbent banks check for threats within their loan books, they cannot ignore the existential threat from more agile new entrants. How ironic it would be if, so soon after a crisis, lenders had not dealt with an equally obvious threat: their failure to use new technology in a joined-up way across their business.
Watch this video to see a simulation for mortgage analytics
Editor's Choice
