Missing EPC data is a growing frustration for mortgage lenders and property managers.
Our own analysis of EPC open data concluded that 49% of UK homes are missing a valid EPC.
Many of these missing EPCs will be for homes that have had the same owner occupier for decades – including many who have already fully repaid their mortgages.
That’s because EPCs are only a legal requirement when selling or letting out a home and last for 10 years once created. So, homeowners who have owned the same home for longer than a decade have little incentive to renew their EPC.
For this reason, a typical mortgage book is likely to have a lower rate of missing EPCs than this incredibly high overall figure of 51%. Yet for some, particularly later life lenders (as discussed later in this article), this can be even lower.
In any scenario, with climate risks and reporting requirements on the rise, the amount of missing EPCs in a typical mortgage book is still a growing concern.
Why is missing EPC data a problem?
Missing EPC data means a big blind spot and obstacle when it comes to the energy efficiency of residential properties in a portfolio.
That’s important because for financial institution that work with housing, an inaccurate analysis of energy efficiency means:
- Inaccurate transition risk quantification and stress testing – needed for internal climate risk assessment and management processes, to fulfil climate-related financial disclosure obligations, and to prepare for minimum energy efficiency standards which are likely to push for a minimum EPC C standard.
- Inaccurate financed emissions calculations and net zero planning – without a full view of carbon emissions from properties it’s impossible to understand how it contributes to financed emissions and, therefore, to effectively set targets and mitigation plans to reduce those financed emissions.
- Untapped potential for green lending – because the key way to reduce the financed emissions from energy inefficient properties is to support customers with retrofitting, which brings with it the opportunity to build new revenue streams from green retrofit improvement loans.
- Untapped potential for investment in green assets – similarly, by supporting customers with retrofit and increasing the number of energy efficient homes in the portfolio, the opportunity to qualify those homes as green assets for investment arises.
💡 Does missing EPC data hit later life lenders the hardest?
In our conversations with lenders, we’ve found that missing EPC data is a particular issue for later life mortgage lenders.
Later life mortgages are typically only available for homeowners aged 55 or over, who are more likely to have owned their home for a long time. EPCs were only introduced in 2007 so many of these homeowners will never have needed an EPC. Further, if they did have an EPC when they bought the home, it’s likely that it’s now invalid due to the 10 year validity.
Because of this, later life mortgage lenders typically have a higher proportion of missing EPC data – Just Group, for instance, were missing EPCs for 63% of their mortgage portfolio before they worked with Kamma to fill the gap.
Later life lenders also have more stake in the quality and valuation of the homes, because the mortgage is repaid by capital from the sale of the house when the customer eventually dies – if energy efficiency is not improved, it’s possible that the value at sale will be lower than expected.
This explains campaigns such as the Just Group and Vibrant collaboration to offer free EPCs to all lifetime mortgage customers – supporting them with seeing the potential for energy efficiency improvements, whilst also ensuring more up-to-date EPC data for their mortgage books.
How could predictive modelling be harnessed to estimate missing EPC data?
Predictive modelling is a great solution to close the gaps in missing EPC data and increase the accuracy of climate analysis for a property portfolio.
This is one of the ways we work with mortgage lenders and property managers at Kamma.
Kamma’s EPC modelling combines geospatial data expertise with machine learning to predict EPC data to a high level of accuracy.
There are three key steps to our EPC modelling approach:
- Geospatial analysis of any shared building data. This identifies any properties that are missing an EPC but share a building with other properties that do have an EPC (usually flats). Shared buildings are a strong predictor for EPCs due to consistent construction methods, materials, fittings, and so on, so this is the most accurate approach for these property types.
- Machine learning model. Our machine learning regression model is trained on all the predictive drivers within a given database for the EPC band of a property, such as property age, number of habitable rooms, heating system, and so on. This model is applied to a mortgage book or property portfolio to predict any missing EPCs.
- Geospatial analysis of the local area. If any EPCs remain that have not been successfully predicted, we analyse the local area to determine the most similar properties in the local areas that do have a valid EPC – because property characteristics are strongly correlated with EPC data.
The result? A much more accurate picture of EPC ratings for a mortgage portfolio.
In fact, we recently worked with Just Group to predict EPC ratings for the 63% of properties in their later life mortgage portfolio which do not have an active EPC, a key piece of information to inform their 2024 climate transition plan.
Kamma’s approach was able to achieve 80% of missing EPCs accurately mapped to the correct EPC band.
A further 19% of properties were off by a single EPC band – still enabling a much more complete view of the energy efficiency and carbon emissions of their mortgage portfolio.
“Working with Kamma is all about confidence. Confidence in the Kamma team, confidence in the data and analysis they provide, and confidence that we’re on the right track with our emissions calculations and plans for reducing those emissions.”
Tom Kenny, Group Property and Credit Risk Director at Just Group
If you’re interested in exploring predictive EPC modelling for your portfolio, we’d love to chat. Get in touch and a member of the team will get back to you ASAP.
