Everything banks think they know about risk is wrong

Haley Tippmann

The risk models that banks use to help inform their commercial-lending decisions have been dealt a blow by Covid-19. The pandemic has presented a crisis where historical correlations do not hold. In 2021 we will find new ways of assessing risk that use forward-looking as well as backward-looking data. This will make lending smarter and better for everyone.
Traditional risk models are based on historical data, but the dynamics of the Covid-19 crisis mean that extrapolating from the past may now be a less helpful approach. As is the case with trade wars, natural disasters or, indeed, climate change, pandemics are by their very nature situations that are hard to predict or plan for. We can make assumptions based on what we have seen with similar events in the past, but no two are the same, so any view of them needs to be supplemented with forward-looking data, which takes into account future challenges that may arise.


In the context of commercial lending, forward-looking data, such as projections of revenues, provide an additional means of understanding future risks. As these offer a glimpse of a possible outcome under certain assumptions, they can never be as accurate as historical numbers, but they do give banks and borrowers the opportunity to act with foresight. In a fast-changing world, a timely change of course based on imperfect data could be better than 20/20 hindsight when it’s too late to avoid a problem.
Risk models are also less useful when a situation is changing as rapidly as it is now due to the pandemic. Take, as an example, two contrasting restaurants. Under lockdown, a fine-dining restaurant is unlikely to experience any business, whereas a pizza delivery chain may see an increase in trade as people stay at home. When lockdown eases, however, and restaurants are allowed to reopen (but with strict social-distancing and cleaning measures in place), the situation then is quite different. The formerly empty fine-dining restaurant, which had always spaced customers far apart with a lengthier turnaround service, is now experiencing a surge in reservations as many diners look to make their first meal out in months “special”. Meanwhile, the pizza chain may see demand for deliveries shrink slightly as people rush to enjoy the outdoors and take advantage of their freedom. This is also in a context where we are seeing unprecedented government support of the economy.
A situation like this is so dynamic and ever-evolving that banks need to be able to rerun scenarios on loans on a regular basis, using new, real-time data as they receive it.
Lending models will also need to provide an understanding of the portfolio at the granular-loan level, taking into account the individuality of each business and how metrics such as their cashflow may be affected by the crisis. A luxury boutique that specialises in made-to-measure gowns is likely to see its revenues obliterated as few customers would have an occasion to wear one of its creations. E-commerce businesses that specialise in yoga wear, by contrast, may see a substantial increase as customers practise more yoga at home and order more online.
As the consultancy McKinsey has noted in its report on lending after Covid-19, “to evaluate creditworthiness properly… banks must go beyond analyses of sectors or subsectors and assess individual borrowers.”


As a result of this crisis, banks are having to change the way they lend to businesses. They will need to use forward-looking as well as backward-looking data, rerun analysis on an ongoing basis, rather than annually, and take a granular, loan-by-loan approach. The best banks will expand on these practices that have emerged because of Covid-19 in 2021 and beyond.
Rishi Khosla is co-founder and group CEO of OakNorth

Like this article?

Share on Facebook
Share on Twitter
Share on Linkdin
Share on Pinterest

Leave a comment

Why You Need A Website