THANK YOU FOR SUBSCRIBING
The pandemic has revealed how capital markets have been vital in helping some firms survive, but it is the ones who invested heavily in tech beforehand which have stood out and remained strong during this time. Capital markets technology help streamline processes, provide greater transparency for investors, and unlocks the potential of AI and machine learning to forecast the performance of investments. In effect, future proofing businesses.
The need for greater transparency
One of the major issues that has been brought to the forefront because of the pandemic is investor demand for transparency. Investors want easy access to information and documentation. They also want that information to be very granular so that they can monitor the risk in their investments. A significant amount of the securitisation market regulation brought in after the global financial crisis was designed to provide for this. This is probably one reason why the market has performed better through this crisis. There are several providers, notably the European Data Warehouse and EuroABS that provide this service to investors for a wide range of securitisation deals from different issuers. These repositories contain extensive deal documentation and so called ‘data tapes’ that provide information on each loan or exposure that make up the collateral of the securitisation.
As well as these repositories, several issuers have tried to answer this concern with the creation of their own investor portals to display information about their own transactions in one centralized hub. These dashboards contain a huge amount of data, but this does not mean the information is always displayed in an easy to understand format.
At Kensington, we have tried to convert this information into a more user-friendly and graphical way, through the creation of a state-of-the-art data dashboard within our investor portal.
Enabled by Microsoft Power BI to give the latest data visualization capability, investors can view specific information about Kensington’s securitisations. This includes the production of stratification tables and visualisation of historical performance trends, for single or combined deals, in order to analyse mortgages that secure the Kensington RMBS programmes. The data dashboard manages over £6billion in mortgage assets.
This type of data enriched platform provides investors with an easy way to take a deep dive insight into the loans behind the investments they are making. If investors wish to see how mortgage payment holiday borrowers are performing in the loan book by a particular region –investors can select by geographic region, see how many loans are there, and even narrow down to a particular loan-to-value (LTV) range.
We allow investors to take a decisive view on the risk and if they are able to see what that risk is, with their own eyes, it allows them to make the best possible decision for their assets. As more securitisation issuers start to embed data visualisation tools like Power BI into their IT architecture they may also choose to roll out this capability.
The three A’s: Automation, analytics, AI
Data is one of the most important assets that currently exists. And businesses rely more and more on data models to help predict where they will be in three, fiveand ten-years’ time.
Stochastic financial models that use historical data to forecast the performance of mortgages have been used in the US mortgage market for many years (for example the models developed by Kay Giesecke from Stanford). Over time the calibration of these models has been moved to machine learning frameworks. The US market has provided loan level performance data since before the financial crisis and so there is plenty of data to help calibrate these models. The most commonly used database is the one provided by Core Logic.
However, access to such data for the UK mortgage market has not been so good. There are a few mortgage companies that have access to large amounts of data and have invested in the data infrastructure to determine that.
Kensington is one of those, and we have developed a scenario model called Vector, and this is used in every aspect of the business. Developed over the last decade, the latest version uses machine learning (AI) to calibrate itself on our proprietary database and helps us predict how borrowers and the market may behave in a certain way in the future.
If house prices go up (as the majority of recent HPIs are showing) or go down (as predicted after the stamp duty break finishes), we want to see how each mortgage portfolio will perform. We then aggregate this data together to see how the company will perform over time. Vector tracks exactly this, historic data, and predicts patterns. As with all thing’s technology, the more data you input, the more you get out.
The algorithms look at performance data and allows us to see what the greatest chances, or outcomes, are of a certain scenario. It can show us the primary factor that makes the difference to how that loan book performs. Whether that’s LTV range, house prices, deposit amount, or other significant drivers of loan performances.
If investors are looking to buy a book of loans, Vector can be used to assess what the book of loans is likely to generate over time and forecasts when repayments will take place and securitisations will be redeemed, or the percentage in arrears. On a macro level we can see how the company and portfolio will perform whether we are in a V, L, or U-shaped economic recovery and how many people may fall into arrears in each.
The goldilocks formula
Ultimately, the mortgage industry needs to fully embrace technology to remain competitive and there is no doubt that technology is one of the biggest opportunities to address industry challenges and drive future growth. Behind the scenes, analytics and technology play a vitalrole in manual underwriting and helping us support borrowers who struggle to get help on the high street, while offering investors greater transparency and predictions in the hunt for yield.