BuildBackBetter: Portfolio Management Review

BuildBackBetter – Portfolio Management Review

Author: Daniel Mclean
Contributors: Stuart Leak and Torgunn Ringsjo

Unprecedented is a term that we have become familiar with since the emergence of Covid-19 and it is an accurate description of the time we are all living through.

The word ‘unprecedented’ implies that there is an absence of a historical reference to inform the experience, and in the case of portfolio managers and credit risk practitioners, this uncertainty creates a conundrum. The conundrum being that the industry has developed its tools largely on the premise that past behaviour is predictive of future performance.

As an industry, with the direction of the UK government and the Bank of England, we have taken drastic steps to support UK consumers and businesses. By the end of May, there were 8.7 million workers on the Coronavirus Job Retention Scheme (CJRS)[1], and 1.86 million mortgages, 0.88 million credit cards and 0.60 million loans on Payment Holidays[2].

One thing we are acutely aware of is the challenge this experience is placing on traditional data sources used in credit models and strategies, as we discussed in our recent article – ‘Scores on Ice’. Income and repayment data are but two of the characteristics where data will be compromised, both in the short term and the foreseeable future. Both the initial phase of Payment Holidays and the furlough scheme (whereby employers were not expected to input any payment) are scheduled to end on 30th June 2020, which, at the time of writing is only a matter of days away and at a time when many lenders will be will be delivering their half year results. What can be done to navigate the uncertainties that will follow and how best can we, as an industry, continue to ensure appropriate decisions for customers?

With all the uncertainty this brings lenders are keen to understand how their customers and portfolios will be impacted, and how best to adjust their portfolio management approach to build back better.

The key questions being posed are:

How to build a model where there is an absence of historical data? Without suitable historical data to draw on, a new way of thinking is required; a new way to understand how this period has, and will, effect new and existing customer decisions. Initially, the focus has been on the expected performance of customers whose Payment Holidays are coming to end, and where traditional metrics may no longer give the same insight as to how a customer is likely to perform. One approach some clients taken is in the use of 4most’s Knowledge Elicitation Process (KEP) tool which leverages proprietary techniques to extract expert judgement from underwriters and collectors to develop statistically robust models and segmentation. It is an approach that requires materially less data, and hence is significantly faster to develop and deploy than traditional scorecards.

How relevant do incumbent models remain? Existing model suites (both proprietary and generic Credit Reference Agency) will include characteristics that incorporate payment history. As these characteristics are compromised, the scorecards will inevitably lose predictive power and accordingly the quality of decisions will be diminished. Model and strategy owners need to understand this, to ensure that their models remain appropriate and effective. 4most are actively assisting clients enhance model monitoring with the view to empower proactive strategy update through a revised focus on characteristic construct, weights of evidence and the new risk exposures that have emerged.

Do policy rules ensure an appropriate decision? Many organisations are engaged in reviewing their policy rules with the view to confirm that the relationship between behavioural drivers and expected performance remains consistent with expectations. These relationships, which determine acceptance, line assignment and pricing outcomes for credit agreements, are likely to be challenged as consumer behaviour adapts to the changing macroeconomic environment. As such, it is important to monitor and identify any change, or to proactively respond, to mitigate any deterioration in predictive power. Within the industry portfolio managers are implementing simplification agendas, reviewing challenger strategies with the view to strip back complexity to ensure clarity and reduce ambiguity in a post Covid-19 world.

Are there lead indicators of financial difficulty? If traditional missed payment / forbearance flags are not working, what other measures can be adopted to understand those customers that may be in, or about to experience financial difficulty. And what do they mean for lenders’ customers and portfolio. New data made available by the Credit Reference Agencies or captured as part of the treatment of customers on Payment Holidays, alongside the repurposing of existing internal data is being considered by lenders in the development of early warning indicators and reporting. Taking all these points into consideration, clients are looking into the refinement of their affordability and creditworthiness policy rule sets, updating logic and enhancing existing monitoring to facilitate proactive portfolio management.

By ensuring early identification and tracking of customers likely to be harder hit over the short/medium term, lenders can ensure they are proactive in their customer management and make prudent and responsibly sound decisions for new business.

4most leverage a broad range of expertise ranging from consultants who have worked within portfolio management teams, from Analysts right through to CRO. We have experience from across a very wide spectrum of financial institutions, prime and sub-prime lending portfolios as well as the credit reference agencies. Drawing on this experience and our unique skillset, we can offer comprehensive support through these challenging times, from recommending best in class processes/strategies to identifying potential pitfalls and offering solutions on how to steer through the economic uncertainties ahead.

[1]https://assets.publishing.service.gov.uk/government/uploads/system/uploads/attachment_data/file/891249/Coronavirus_Job_Retention_Scheme_Statistics_June_2020.pdf

[2] Source:  UK Finance

Do you have any questions? Please contact Stuart on stuart.leak@4-most.co.uk.