Opinion on the draft EBA guidelines on loan origination and monitoring: Please accept technology as part of banking!

Leitlinien zur Kreditvergabe | guidelines on loan origination | PayTechLaw

The European Banking Authority (EBA) has carried out a consultation on draft guidelines on loan origination and monitoring. These guidelines reduce the role of technology to something that needs to be monitored and present it as inferior to traditional processes. Furthermore, the guidelines on loan origination appear to emphasise traditional methods of data collection and data points and completely ignore the technological progress that is already a reality.

In June 2019, the EBA issued draft guidelines on loan origination and monitoring that set out a standard on how CRR institutes and other incumbents issuing loans are to manage governance, loan origination procedures, pricing, valuation of property and monitoring with regard to loans. The new guidelines on loan origination are also set to replace the EBA guidelines on credit assessment which set out guidelines with regard to consumer creditworthiness checks.

Technology as a risk

What I would like to focus on is the short section on the use of technology that is contained in the draft guidelines on loan origination. In a nutshell, the EBA’s guidance on the use of technology for credit granting purposes states that:

  • the risk associated with technology-enabled innovation must be managed;
  • the potential for bias in the credit decision-making process must be managed and the integrity of systems must be ensured;
  • the institute must be able to understand the underlying model and explain the outcome;
  • the output must be verified and monitored and the performance must be compared to the output of traditional methods/tools;
  • processes and models must be properly documented and reviewed ;
  • the management body must be able to understand how the technology is used and what its impact is on credit granting procedures;
  • the credit risk management function must be able to understand and explain the behaviour of the technology-enabled innovation.

To summarise: Your credit department and top  management should understand what the company is doing, manage risk, not take biased decisions and also be able to explain this  to others. – Really?? Should this not also apply to what the guidelines on loan origination call “traditional tools and methods”? The underlying assumptions are that technology is a risk, nobody really understands what they are doing and that the traditional methods are better.

Technology should be the standard

In my opinion, the EBA guidelines on loan origination do not adequately reflect the role of technology in today’s world anymore. Technology is already an essential part of many credit business models and is often able to provide quicker and more accurate credit assessments than the traditional methods. Rather than focusing on the risks and thereby implicitly voicing a preference for traditional methods, the EBA should foster the use of technology to take credit granting to the next level. Why not make it mandatory to use technology where appropriate instead of relying on man-made decisions that we all know are limited by the amount of data considered and analysis carried out to what is humanly possible? This would help guide the banking industry into the next century instead of dragging  it back to the days of pen and paper.

Back to pen and paper

The fact that the EBA is implicitly basing its guidelines on loan origination on antiquated procedures becomes even more apparent when looking at Annex 2 of the guidelines on loan origination. This Annex sets out the information that the lending institute should collect, and requires, for instance, financial statements and accompanying notes of the respective individual entity as well as at a consolidated level (balance sheet, profit and loss account, cash flow), a business plan for the borrower as well as in relation to the purpose of the loan, age debtor statements at borrower level and evidence of tax status and tax liabilities. All of this evidence is usually paper-based and is often only available for larger companies and not for small or medium-sized businesses or start-ups. While it needs to be noted that the guidelines on loan origination emphasise that the principle of proportionality should be applied to assess adequate credit-granting processes and thus not all information may be needed for all types of credit, there is still a bias for paper-based evidence.

In reality, more and more loans are granted based on the ability of the lender to make use of the access to the borrower’s accounts now available due to PSD2. Technology enables the lender to access the payment accounts of the borrower, to analyse incoming and outgoing payments and to base its credit-decision on actual income and spend. This type of information collection is not even mentioned in the guidelines on loan origination but rather the call for paper-based evidence perpetuates the “traditional methods” because it appears to set them as the standard.

The double standard of the guidelines on loan origination

It is true that recent research has shown that also artificial intelligence may have a bias due to the selection of data that its decision is based on. The dream of a bias-free decision-making process has yet to be realised. However, to assume that this is only a risk for technology is something else. Rather, we all know that traditional methods that involve human decision-making produce even more biased decisions. Yet the guidelines on loan origination do not require bias to be managed  for traditional methods of  granting credit.. So I guess if a bank does not give out loans to female founders because its all-male credit-decision committee does not believe in the ability of women to run a company, this is not an issue. But if an algorithm prefers tech-companies (more likely to be run by men) to start-ups in the healthcare or education sector (where a higher number of female founders typically exist) this is an issue. Don’t get me wrong. Both are not ok and it is important to look into how the choice of data sources can produce biased-results. But  it is also hypocritical  to call out this issue only with regard to technology.

The same applies to information that needs to be considered in the credit-granting procedure as listed in Annex 2 of the guidelines on loan origination. It makes sense to consider various data points when making a credit decision but to limit the list to more or less static types of information due to a preference for paper-based evidence which is only available with respect to certain customers is in itself biased and hinders innovation.

The guidelines for loan origination require that technology is compared to traditional methods to analyse if technology delivers results that are as good. But shouldn’t it also require that traditional methods are compared to technology to see if they still stand the test of time compared to technology??

There is a better way

The consultation period for the guidelines on loan origination expired on 30 September 2019. The final guidelines are due to be issued shortly and will apply from 30 June 2020. Since the final guidelines on loan origination have not yet been issued there is still hope. I would ask the EBA to reconsider their take on technology and amend the guidelines on loan origination to make technology part of credit granting. Here’s my suggestion on how to amend the guidelines on loan origination:

  • Technology-enabled innovation is a strong driver for quicker and robust credit-granting processes as it is able to consider many data points for a holistic view of the borrower. The institute should thus consider making use of technology where appropriate to be able to process large amounts of data quickly.
  • Information used for credit-granting procedures may come from different sources, including internet-based sources and access to accounts and is not limited to paper evidence or corporate documents. How much information is collected is subject to the principle of proportionality.
  • Borrowers should regularly review their credit decision procedures for any bias in the decision-making process.

Will this happen? Hope dies last.


Cover picture: Copyright © fotolia /Chan2545

Leave a Reply

Your email address will not be published. Required fields are marked *

You May Also Like