Revised EBA guidelines: How banks and financial institutions can leverage data in loan origination and monitoring processes
Best Practices & Trends
Published by Camille Charlier - June 3rd, 2022
TheEBA’s May 2020 revised guidelines on loan origination and monitoring give banks and financial institutions a real opportunity to start relying more and more on AVM (automated valuation models) based solutions for real estate valuations.
In May 2020, the European Banking Authority (EBA) published new guidelines on loan origination and monitoring with the aim of harmonising standards in these areas across the EU. For the first time, the revised EBA guidelines also included requirements for data and automated valuation models.
Many professionals in the real estate, banking and finance sectors agree on it: this update holds enormous potential for the industry, and is a step towards a more digitalised future for the banking sector. Most importantly, having these topics mentioned paves the way for an increased use of AI-powered solutions for real estate valuation and creditworthiness assessment.
As an example, Ron Hess, CMO & CSO of on-geo GmbH, considers that “the European Banking Authority has taken a clear position: it encourages banks to look at statistical models because they can make their life easier. And that is absolutely the right way to go.” Ron Hess was the guest speaker of our webinar “the future of AVM in banking” during our German-speaking online conference Insights by PriceHubble in April 2022.
“The European Banking Authority has taken a clear position: it encourages banks to look at statistical models because they can make their life easier. And that is absolutely the right way to go.”
- Ron Hess, CMO & CSO, on-geo GmbH
“Automation” and “process optimisation” aren’t foreign concepts to the banking and financial services industry. However, they have not yet reached all areas. As an example, in Germany, the credit granting process and monitoring of loans by banks and credit institutions is still carried out with the help of conventional valuation methodologies by an appraiser or expert. This is necessary when dealing with complex properties, or properties not used for residential purposes, but not always when dealing with residential properties whose value can be standardised. In fact, even before the EBA issued its guidelines, it was possible (in Germany and under certain conditions) to make a decision from a desk in the case of loans that fall below the loan limit of EUR 400.000.
Nowadays, the technological possibilities available to us today - e.g. a steadily growing amount of data, powerful valuation and forecasting algorithms, and unlimited computing power and storage capacity thanks to cloud computing - enable real estate and banking professionals to generate sharp and in-depth valuations within minutes (and thus, make the right decisions faster) and provide a base for appraisers to fine tune.
Nevertheless, it is important to point out that new technologies do not aim to replace the expertise and knowledge of appraisers and real estate professionals. Instead, they aim to complement their skills and make their credit decision-making processes easier. Raja Beurer, Director of Banking, Finance and Insurance of PriceHubble Germany, explains: “New technologies such as AVMs let finance and real estate professionals get the full picture of the potential of each property within minutes. This saves them a significant amount of time.”
“New technologies such as AVMs let finance and real estate professionals get the full picture of the potential of each property within minutes. This saves them a significant amount of time.”
- Raja Beurer, Director of Banking, Finance and Insurance, PriceHubble Germany
Let’s take a closer look at the ways in which an increased use of AVM based solutions can be beneficial for banks and financial institutions:
Assessment and valuation of guarantees and collateral through AVMs
By opening up the process to AI based models, the EBA aims to operationalise collateral valuation in the credit process. This not only creates greater transparency, but grants banks access to up-to-date information on the properties they are financing at all times. For that purpose, the performance of statistical models has taken an enormous leap forward in recent years.
AVMs are now able to determine the statistically most probable market and rental price for any residential property at any location on the basis of available data, taking into account not only traditional characteristics, but also less common criteria and metrics which have a proven impact on the price, such as noise pollution, socio-demographics, accessibility, or even construction projects in the surrounding area. AVMs also capture complex relationships that cannot be accounted for in traditional valuation, and make it possible to determine market value with a valuation accuracy of 5 to 15 percent deviation.
In other countries, such as Switzerland, semi-automated processes have already become more common through the use of hedonic models. It is possible, with the help of digital authentication, to carry out the extension of a mortgage completely digitally without an advisor.
Risk management: an opportunity to get closer to the market
In the risk management department of many European banks, such AVMs are still completely underrepresented today, despite the fact that they offer many advantages.
As an example, in Germany, a conservative price determination such as the asset value method according to the Real Estate Value Determination Ordinance (“ImmoWertV”) takes into account the standard land value and the production costs of the property and are, if necessary, corrected by applying a market adjustment factor.
However, this procedure doesn’t accurately take into account, for example, economic changes at the location resulting from migration, or the increase in construction costs. As a result, the production costs considered in the valuation at the time of the loan no longer correspond to the actual production costs and thus, to the market price.
Since an AVM is able to reflect such changes in real time, it is then even much more conservative in its valuation. If such changes in the market do not flow into the valuation at the time of lending and also not into the ongoing credit monitoring of existing loans, the bank takes easily avoidable risks. On the contrary, if the valuation is too conservative compared to the actual market price, banks may miss out on business by failing to recognise the value of the property and rejecting a worthwhile loan transaction.
The use of AVMs can thus support credit institutions in the form of a pre-due diligence already in the first step of the consumer credit life cycle.
Make the right decisions faster with the help of automated valuation models
The use of AVMs can support banks and financial institutions in many ways when it comes to credit risk assessment and loan origination procedures, and the EBA’s revised guidelines taking into account these new technologies demonstrate how relevant these solutions are in 2022 and beyond.
PriceHubble’s cross-country approach and unique data and explainable AI-driven real estate valuations let users in the banking and financial industry generate extremely accurate valuations in a few clicks only. Today, PriceHubble represents one of the largest data science and engineering teams dedicated to the real estate industry. We strive to provide our clients with seamlessly integrable API solutions and intuitive web applications. We work hand in hand with our clients and industry professionals to provide them with solutions that are ever evolving and fit their needs in the best way possible.
Would you like to know more about our solutions? Don't hesitate to request a demo with one of our experts and learn how you can leverage the power of data to make smarter and faster decisions.
The purpose of an AVM is, given a set of property characteristics, to return the most accurate price estimate for this property. In this article, we explain how Regression Splines can be useful to effectively build such systems.