Being able to accurately predict how a property asset will perform over a long period of time sounds like a panacea for any property lender, but sadly there isn’t a silver bullet and different markets have developed a range of different approaches.

On Friday, Knight Frank hosted a seminar for Young Professionals in Real Estate Finance (a sub-committee of CREFC Europe). We heard from a panel of speakers regarding the German system of Pfandbrief valuations, the different approach used by UK lenders and then the potential trends for future-proofing valuations.

What is the Pfandbrief valuation approach?

Harvey Kasin from Deutsche Pfandbriefbank discussed the standardised valuation model of Pfandbrief and how this is a more rigid approach to valuation than the RICS Red Book's flexibility. This approach is highly trusted by investors, consequently allowing a fixed proportion of the debt to be secured by covered bonds issued under the German Pfandbrief Act and Pfandbrief banks therefore to provide very competitive margins.

However, the Pfandbrief approach is designed for the German property market, so it is debateable whether it is correct to apply the same rigid model in other real estate markets on which Pfandbrief banks lend, including the UK. In particular, market valuations in Germany tend to remain relatively stable, in contrast to the cyclical nature of the UK property market.

How does the UK valuation approach compare?

Klaus Betz-Vais of Lloyds discussed how, although UK loans have different sources of funding which tend to be more expensive than Pfandbrief (e.g. deposits, AAA CMBS, unsecured paper and other types of covered bonds), the Pfandbrief funding advantage does not apply universally. In particular, where the loan to value ratio is high and/or the difference between the mortgage lending value and market value is large, the portion of the debt on which covered bonds cannot be issued, which is consequentially subordinated and is therefore more expensive, can outweigh the pricing benefits provided by the portion secured through Pfandbrief.

To assess the risk weighting of real estate finance, UK lenders apply standardised questions, which then determine how much capital is to be held against potential losses on each loan (“slotting”).

Although the regulators seem comfortable with this approach, the panel discussed whether “slotting” was overly cautious and could eventually reduce the debt available in the market (especially for property in the UK regions outside London) due to the higher equity amounts banks are required to hold for loans with lower “slotting” scores.

Where next for valuations?

Neil Crosby from The University of Reading explained that another property crash seems inevitable (previous cycles suggest around 2024), but it is unclear what will be the cause, given that the 1990 crash was (in simplistic terms) triggered by a boom and bust in rents and the 2007 crash, by a boom and bust in cap rates.

After each crash, the real estate sector has adapted in an attempt to stop history repeating itself, and forecasting has been updated to factor in the causes of previous crashes and anticipate similar issues.

However, for long term property valuations to be useful to lenders, the ideal valuation methodology needs to be more than just reactionary in light of prior issues.

The Investment Property Forum (IPF) has commissioned in-depth research into developing new methods of long term property valuations, and their findings are due to be published later this year.

It will be interesting to see how the sector reacts to IPF’s proposals, and how this affects finance documents, such as loan-to-value covenant tests in facility agreements.

Overall, there are clearly flaws in all of the current valuation systems for UK property, but it is debateable what the appetite is to radically change the basis of property valuations in the current market.

Perhaps one of the hardest – as well as the most important – judgments faced by CRE lenders and regulators alike is assessing the level of risk arising from CRE loans.