By Dr. Roboadvice
Banks have been growing their cost/income ratios with serious cost reductions since 2008. But from now on, the real breakthroughs will primarily have to come from income growth. Online retailers like Amazon are great at growing their income, but
this is not what banks usually excel in! New target groups, new product market combinations, ecosystems that drive longer customer lifetimes, personalization of the (price) offering are not exactly core banking key success factors.
How can a simple personality test at the core be a solution to that and activate the growth engines of banks and drive income breakthroughs? In this current era of big data and artificial intelligence new data will be at the forefront of serious
income growth. But how?
The thin file issue of poor credit information
Worldwide, about 37% of adults are listed in a private credit registry with information on repayment history, unpaid debts or outstanding credit. They are “framed” as prime customer segments. Companies like Experian and Equifax collect data on payment behaviour
of these consumers to build their credit scores. The differences per country are huge, e.g. in Spain only 18% are registered at a private credit bureau and in the Netherlands 76% according to The
World Bank data (2016).
Nowadays, the United Kingdom counts about 13 million sub-prime consumers. And although near-prime and sub-prime lending was practically banned after 2007’s credit crunch, these near-prime customer segments are increasingly being seen as an interesting target
group. The recent interest rate rise in the UK makes near-prime lending an actual topic in bank’s boards as this group will most likely grow fast, but are also a higher risk as the interest rates and inflation is expected to rise in the upcoming years. The
impact of the overall poor available credit information on people’s financial life is enormous. If granted a loan, extremely high APR’s of up to 39% are granted to these thin file borrowers.
Credit Unions like MyCommunityBank therefore target these customer groups, but are moving heaven and earth to do it at much fairer rates to not bring these people into problems if interest rates would increase further. The lack of data (or maybe the willingness
to think out-of-the-box in screening loan applicants with other views) excludes large groups from building and living their life...
The size of the market opportunity is much bigger. Around two billion people globally don’t use formal financial services and more than
50% of adults in the poorest households are unbanked. These people do not have access to appropriate and affordable financial services. If these people could be sustainably provided access to
financial services, global GDP will grow dramatically and poverty will decrease significantly. Not only households, but also businesses, especially small enterprises (SMEs), still struggle to obtain financing that meets their needs, as a recent
OECD report shows. The opportunity to actually help people and to support sustainable GDP growth even further increases. Therefore, the negative connotation on near-prime and sub-prime will probably slowly disappear as big data driven tools are perfectly
capable of bringing best of breeds together: doing fair creditworthiness assessments that help people build their lives and businesses,
and un-tap a very interesting growth opportunity for banks.
Psychometric Predictive Underwriting
Since I have been doing a lot of research on robo advice over the past six years, I publish academic papers on the targeting, business and operational models of robo advice solutions. Driven by the international research and the insights developed along
the way, I also have been growing tech companies that apply artificial intelligence to robo advice, financial services and algorithmic risk & marketing software. In doing so, I’ve become passionate on the power of psychometric predictive underwriting. Psychometric
predictive underwriting is a fundamental new way of underwriting that applies new data around the data group coming from psychometrics. Rather than looking at someone’s past payment behaviour, it assesses the personalities, traits and values of people and
translates them into risk profiles. These can then be used for predictive underwriting for sub-prime, near-prime but also prime customer segments. Psychometric predictive underwriting is still in its early days, but this new machine learning-driven solution
already enables lenders to grow their customer bases, with 18% and at 20% lower delinquency rates. 18% in the UK alone, would lead already to financially include 2.3 million people. But more benefits can be seen. For example, it enables financial institutions
to specifically target their robo advice solutions to the personalities of their target customers; it drives ecosystem development for longer customer lifetimes, higher customer lifetimes values; and, when automated, it helps to reduce billions of overhead
costs resulting from risk assessments on difficult customer segments currently being done by hand.
From application scoring to continuous behavioural monitoring
As an example, based on motivational insights, one of my companies,
AdviceRobo, developed a psychometric virtual interview to measure creditworthiness of thin-file customer segments like self-employed individuals, start-ups and millennials. Applicants will answer a 5-minute virtual interview and instantly the lender will
receive the applicant’s psychographic credit score and profile. The combination of the psychometric credit score and profile delivered through digitized algorithms enable lenders to make a decision and to predict borrowers’ behaviour on default risks. That’s
what we call psychographic predictive underwriting. Understanding customers’ psychometric profile delivers both a far more intimate understanding of the customers and the possibility to better influence them to avoid behaviour with high risk of default consequences.
This psychometric predictive onboarding solution currently leads to on average 18% higher acquisition at up to 30% lower default rate. And this type of solution becomes even better with the self-learning character of the machine learning software. The thing
becomes smarter every single time it’s being used.
The continuous monitoring of the consumer also provides insights that enable lenders’ marketeers to activate financial healthy behaviour in their customer bases. ING research recently showed that 48% of Europeans have never had any financial education. According
to this ING research, 44% of Germans calls himself financially illiterate. That’s 35 million people in Germany only! With psychometric predictive underwriting lenders will get insight in people’s financial knowledge and financial motivations. It will help
them to apply personalization strategies in digital and robo advice environments that drive customer happiness and life time value.
Costs, costs and costs
So, there are many income growth opportunities. But also on the cost side this kind of predictive robots will seriously drive further cost reductions. The applications of robot process automation from companies like
Blue Prism so far focus primarily on the robotization of back office processes. Huge performance improvements at much lower costs are the result.
The next level of the development in the upcoming years will be the application of cognitively intelligent systems. Software robots like psychographic predictive underwriting as discussed in this blog that digitize more complex consumer computer interaction
are a good example. Forecasts from AdviceRobo and KPMG predict that with psychometric context, behavioural driven software agents will on average reduce IT risk legacy with 50% over the upcoming 5 years. And those cost reductions, on top of the income breakthrough
applications from the cognitive intelligent psychographic predictive underwriting agent will really drive cost/ income breakthroughs!
Smart, algorithm driven technology, called cognitive intelligence, can and will make the difference during the upcoming years to achieve cost/ income breakthroughs. But not at the cost of the customer. No, these applications will, on top of intense cost
reductions, seriously drive financial inclusion and financial health.
The structural trends that are driving many of these substantial shifts stem from multiple sources. Regulation will continue to broaden and deepen as public sentiment becomes less and less tolerant of any seemingly preventable errors and inappropriate business
practices. Simultaneously, customers’ expectations of banking services will rise and change as technology and new business models emerge and evolve. Risk functions will also have to cope with the evolution of newer types of risk (e.g., model, contagion, and
cyber)—all of which require new skills and tools. Fortunately, evolving technology and advanced analytics are enabling new products, services, and risk-management techniques, while debiasing approaches that improve decision making will help risk managers make
better choices about risks. However, the marketing and risk function of the future will probably be expected to deliver against all these requirements and deal with these trends at a lower cost, because banks in all likelihood will have to reduce their operating
costs substantially. And even more important, they will seek for cognitive intelligent applications like psychometric predictive underwriting that cause income breakthroughs and make their customer financially healthier and happier.