Credit scoring in post brick-and-mortar banking Bookmark has been added
Credit scoring in post brick-and-mortar banking
AI and new data sources can enable banks to meet changing customer expectations.
The nature of financial services is changing: customer expectations, technology and regulation are driving traditional banks to rethink their overall lending strategy and business model. A core capability in the lending cycle is the ability to measure credit risk to match risk and return on loans. Deloitte envisions that cognitive automation and unlocking untapped (unstructured) data sources can enable a shift towards instant credit scoring.
Go directly to
- Customer expectations
- Enablers to meet expectations
- Credit scoring to the next level
- What if ...?
- Ready to innovate the lending cycle?
Customer expectations: faster, digital and personalised offerings
The way customers interact with banks has changed substantially in the last couple of years with financial business being executed more and more online. Customers demand faster processes and more personalised and digital services which require banks to provide indicative loan terms (e.g. maximum loan size and interest rate) instantly online. Innovation is essential in order to ensure that banks can meet these changing customer demands.
Enablers: Cognitive automation and unlocking untapped (unstructured) data sources
Cognitive automation and unlocking new data sources can enable banks to meet these changing customer expectations. Traditional credit risk models, however, mostly concern client and product characteristics that are static and quickly outdated such as income statements, loan to value ratios for residential mortgages or financial statement information for corporates. This data is often outdated as it is collected during the loan application or only updated in yearly client reviews. Nowadays a lot more data is available and collected and advanced analytics such as machine learning are increasingly adopted within banks. Additional data types providing opportunities for banks are for example online data such as news feeds, and payment data. New sources of possibly unstructured data give banks the opportunity to adopt analytical methods such as cognitive automation and analyse more dimensions such as payment patterns, allowing for a more comprehensive client view.
"Untapping new sources of data provides an opportunity to attain a more comprehensive client view and spot changes earlier".
The expected Payment Services Directive 2 (PSD2) fastens the process of unlocking the currently untapped payment data. The aim of PDS2 is to ensure an integrated, more efficient and secure payment market within the EU. Additionally it enables innovation by allowing new entrants in the payment market, stimulating competition. More innovative third parties such as FinTech start-ups, Telcos and e-commerce platforms will be able to benefit greatly from PSD2 as it will allow them to gain information of clients’ bank data and provide a unified view of the customer’s financial activities. Therefore, it is important for traditional banks to use PSD2 to their benefit as well and reposition themselves in the lending industry. The data obtained from PSD2 will allow banks to gain a total payment profile of its clients which can be used to estimate the creditworthiness of a client.
Another mostly untapped soure of data is online news, such as news articles, blogs and social media. Cognitive automation is maturing to distill relevant and risk based indicators from theses unstructured sources. Key technologies for automated source discovery and natural language processing have evolved to viable products. Deloitte research has shown that these technologies can increase the discriminatory power of corporate credit risk models up to 10 percent.
Credit scoring to the next level
Credit scoring is one of the first areas where banks started using analytics in their core processes. Since the 1960’s banks have started to quantify credit risk of new and existing client bases. First generation scorecard models have since evolved into instruments for not only client facing activities. Examples include loan application, credit capacity checks, pricing and loan renewals, but also portfolio risk management such as loan loss provisioning, capital requirements and capital planning. We are now at the advent of a next level in credit risk analytics.
Deloitte sees two major opportunities in the lending cycle to address changing customer expectations: instant credit scoring in loan applications and continuous credit risk monitoring. The figure above provides an overview of these two types of innovative credit scoring solutions.
… Credit scoring could be done instantaneously
Instant credit scoring refers to the use of advanced data sources and analytics in the loan application process. The loan application process can be automated via analysis of the customer’s payment data together with other sources of data (e.g. online/social media, smartphone and utility bills). This data can be used on top of the traditional risk drivers used by banks nowadays. The otherwise time consuming loan application process can be done nearly instantaneous. Banks save costs as less manual tasks are involved in the loan application process. Only the non-standard applications might require the intervention of a bank employee. Additionally, this potentially increases customer satisfaction as loan applications can be done much faster.
… Credit risk monitoring could be done continuously
Continuous credit risk monitoring refers to benefits from using online data to improve credit monitoring for existing loans. Whereas conventional credit ratings are based on (historical) financial ratios and are therefore reflecting performance of a company’s past, Deloitte envisions potential in combining traditional credit risk drivers with additional data sources to improve credit risk monitoring. In such a way credit monitoring can be done in a continuous way. Deloitte has a unique an application, called Eagle Eye (see text box) which leverages on this increasing data availability using more sophisticated techniques. Using online data such as news feeds from online articles in credit monitoring has proven to significantly improve the credit rating. Additionally, analysts are able to identify clients having financial troubles much earlier and on a continuous basis. This provides banks with real-time, risk-based insights on their client portfolio throughout the year instead of a periodic manual review. Risk can therefore be integrated in your strategy and power your performance.
Ready to innovate the lending cycle?
Banks face exciting challenges in meeting customer’s demands. The use of unstructured data and cognitive automation can enable banks to improve and innovate the lending cycle, thereby reducing costs and increasing customer satisfaction.
The question whether banks are able to use client specific payment data or news is also dependent on regulation on the topic of privacy. This regulation can differ significantly between countries and has large implications for what banks can and cannot do. On the other hand, these same regulations also impede the possibilities for FinTechs to disrupt the lending market. A different challenge for banks is to accommodate their IT structure to the needs of these innovations. As banks still have very fractured IT architecture based on legacy systems which cannot be disposed of easily, it might be hard to unlock the full potential of the data they have in house.
For more information on Financial Risk Management? Please contact Koen Dessens or Roald Waaijer via their contact details below.