Blog Post

Contagion boosts the case for creditR

The world’s economy has taken a major hit. With millions more businesses and individuals needing to access finance to mitigate the recessionary effects of Covid-19, lenders need to adapt to the new world by adopting new model methodologies that better estimate credit risk.

Whilst the public health crisis created by Covid-19 appears to be easing in the UK, the worry is that the economic crisis is only just starting.

Figures from HMRC highlight that the number of people on company payrolls dropped by 612,000 in May compared with March, and it’s anticipated that unemployment could more than double from the current rate of ~4% - with some economists saying it could hit 10% by the end of 2020.

Businesses are borrowing more to stay afloat amidst the pandemic. As at 31st May 2020, financial institutions in the United Kingdom are owed more than £198 billion by small and medium enterprises.

The sad reality is that some smaller businesses are destined to fail without access to fair and affordable credit. And as highlighted in the hypothetical below - as larger businesses fall into insolvency, many smaller businesses, whose sales are concentrated within that large business, will become insolvent as well through no fault of their own.


The unprecedented economic circumstances have significantly altered the lending landscape as it relates to the estimation of probability of default and it is becoming increasingly impossible for lenders to continue ‘business as usual’ while still relying on the legacy technology and traditional credit decision making.

The need for creditR

This brings us to the concept of creditR, a measure of the increased probability of default – particularly in either a specialist or geographically-isolated business – as a consequence of contagion. creditR is a proprietary product that ezbob are developing that will enhance credit risk estimation, drawing on both supervised and unsupervised machine learning techniques. Machine learning is already disrupting finance, boosting functions such as fraud detection, customer segmentation and client retention. creditR will allow for better identification of potential issues in portfolios as business returns to an approximation of normal.

Regulators remain cautious about transitioning into machine learning techniques, though their appetite will be increased by the need to find new solutions to new problems. During a transformation phase, machine learning algorithms developed as challenger models will likely proceed along with the traditional methods. I believe that machine learning algorithms can produce more robust results than traditional approaches whilst also creating a much more transparent process.

In terms of the above business insolvency example, there are multiple contributing factors for the value associated with creditR, which will likely include:

  • Concentration of sales to the larger business.
  • Possible asset recovery from the larger business (albeit likely insignificant).
  • Existing capital within the smaller business (a measure of how long the business can survive without cash).
  • Sectoral specialisation
  • Geographic isolation of both businesses


The ezbob solution already aggregates information from over 40 different service providers and presents better data that leads to better risk management. This includes the ability to provide loans based on non-traditional online data, such as Amazon, e-bay or PayPal activity. We see creditR as the next step in the post-pandemic world, where there will be a need to assess business’ evolving financial situations more accurately.

Get in touch with the ezbob team today to find out more about how we can help you to transform your lending operations and meet the increasing needs of your customers.