Mortgage-related market data used by investors isn’t always clean, particularly on older loans. A new report produced by Greenwich Associates and sponsored by Refinitiv examines how technology and data analytics is helping to improve risk decision-making on mortgage-backed securities.
- The mortgage market is characterized by volume and complexity, and market players need quality data if they are to make accurate and profitable decisions.
- The breadth and depth of data available in the mortgage market continues to grow, but incoming data may be ‘dirty’, disconnected and non-standardized.
- Technology can harness and make sense of this data, delivering insightful analysis and empowering professionals in the complex world of mortgage-backed securities.
U.S. mortgage debt currently stands in excess of US$12 trillion, and the majority of this debt ($9.4 trillion in Q2) is held by institutional investors via mortgage-backed securities, collateralized mortgage obligations and agency (i.e. FHLMC, FNMA, GNMA) bonds.
These institutional investors and mortgage dealers are tasked with developing accurate risk models to help them make decisions relating to, for example, price, yield and risk.
Robust and meaningful analysis relies on quality data, but given that the majority of mortgages involved in these complex computations were initiated before financial markets embraced digitization, there remains a vast quantity of data that is neither standardized nor usable.
Added to this the often poor due diligence that was carried out when creating many of these mortgage agreements means that much of the data is incomplete or inaccurate, which can result in suboptimal decisions and rising risk.
If we are to avoid any future mortgage crisis, the industry needs to focus on new ways to access better quality data and more insightful data analysis.
Making sense of mortgage data
Several perennial challenges surround the process of ingesting, collating and making sense of data.
Not only has the fixed income market experienced an explosion of available data (which brings its own challenges), but incoming data may be ‘dirty’, non-standardized and presented in different formats, meaning that complex data technology is needed to store and analyze data relating to millions of mortgages and mortgage securities.
Moreover, institutional mortgage investors need to make risk decisions at speed, and relying on overnight batch jobs, although still common in mortgage investing, has become outdated.
The good news is that cloud-based solutions are coming to the rescue, and now mean that the necessary data and analytics can be accessed quickly and efficiently to enable real-time calculations and promote faster, better decision-making.
A complete view of mortgage-backed securities
Over the past decade, much progress has been evident:
- Innovation has meant that individual loans within mortgage-backed securities can now be scrutinized, meaning that investors can form a more complete picture of a potential investment.
- Nearly two-thirds of mortgage-backed securities are now traded electronically, which brings greater transparency and benefits for all stakeholders. Increased data spurs more electronic trading, which generates more data in what is described as a virtuous cycle.
- Industry agreed standards have improved and agency mortgage pools now consist of more specific groupings (‘spec pools’) for greater clarity and more targeted investing.
As technology continues to overhaul traditional solutions to risk management, the industry can look forward to further progress in the years ahead.
For example, machine learning technology that can digest mortgage documents will mean that information is more readily accessible.
In the complex world of mortgages and mortgage-related investments, the role of analytics is set to become ever more crucial as an indispensable tool to help traders form a complete view and make better, faster decisions that will ultimately benefit all market players.
