How to Tackle Complex Libor Issues with Machine Learning

The deadline for the end of Libor is fast approaching and many financial institutions are struggling to address the complexities that this creates. Libor is ubiquitous in modern financial markets and this raises the stakes for managing the process of identifying and remediating the situation. While modern technology processes like machine learning can be extremely useful as a tool in this case, there are nuances and pitfalls that can be avoided to insure success. 

The recent Maven Wave white paper, Libor as a Template for Digital Document Transformation in Financial Services, takes a comprehensive look at the Libor challenge, including opportunities, methodologies, and potential gains. In this blog post, we take a look at areas that may be problematic and explore ways in which they can be addressed.

Complex Data Challenges Call for Careful Consideration

Implementing change is always a challenge and doing so in a situation with high stakes is even more difficult. In the case of Libor and the application of machine learning and best data practices in place, it pays to be aware of the complexity of the task at hand and to adopt a holistic approach. By doing so, you will not only avoid costly delays and mistakes, you’ll be able to harness the full potential of new technologies and processes. 

Four broad areas to concentrate on are the mindful application of business processes, a focus on the quantity and quality of data, utilizing best practices for machine learning, and a realistic approach to managing change.

Step one: Aligning business process with objectives

When it comes time to address a large challenge such as the demise of Libor, it’s essential to make sure that business is aligned with technology, and this starts with clarity around business objectives. Too often, sufficient time isn’t devoted to securing buy-in from stakeholders and this leads to both a lack of commitment and energy as well as unclear or incomplete objectives. 

Table stakes = both the right data and good data 

It may seem obvious, but the most common problems with machine learning initiatives stem from inadequate or poorly managed data. First, it’s important to take the time to ensure that the issues to be tackled can be addressed with the data at hand. Next, time should be taken to confirm the quality of the data. In fact, data quality is more important than the quantity of data. In any case, solid outcomes are predicated on devoting time, money and other resources to data at the onset.

Do what’s best: Following machine learning best practices is critical

While the application of machine learning can produce amazing results, it is by no means a one-size-fits-all proposition. The science itself is actually quite complex and varied, with different methodologies being better suited for particular circumstances and desired outcomes. A multi-phase, layered approach is best. It is important to maintain best hygiene when tracking results as well as the utilization of accuracy scores to check on performance metrics at each step along the way.

Don’t boil the ocean: best results come through “TIAS” – try it and see

Unfortunately, new technologies such as machine learning often suffer from magical thinking: sprinkle some ML over a challenge and it will be solved. In fact, nothing could be further from the truth. Machine learning produces the best results when its application is focused and the issues at hand are well-defined and limited in scope. Big issues can indeed be tackled but they should be addressed in concise, logical steps where dialogue and experimentation are consistently applied. 

Possibilities and Pitfalls for Machine Learning in the Libor Transition

The end of the Libor era is both complex and consequential. In this context, it is critical to adopt remediation efforts that have sufficient scope, utilize the best in technology and are consistently afforded the constant and consistent resources and attention that they require. Machine learning is an excellent candidate for inclusion is such an effort but care should be taken to make sure that the project has the greatest opportunity for success. To see how, download the Libor as a Template for Digital Document Transformation in Financial Services white paper now. 

Maven Wave helps drive the future of financial services with innovative business outcomes, fueled by cloud, with risk top of mind. To help organizations maximize economic outcomes and advancements, Maven Wave brings a rich blend of industry-specific technological expertise, agile-integrated design, and best practices for transformation.  Click here to download the whitepaper or contact us to learn more.

About the Author

Andrew Dunmore
Andrew Dunmore
With over 15 years of solution delivery experience, Andrew Dunmore leads the financial services consulting practice at Maven Wave. Mr. Dunmore is one of Maven Wave’s thought leaders in capital markets innovation and delivery excellence. Prior to Maven Wave, Mr. Dunmore served as a Vice President at Bank of America in their global private equity investment office and an Assistant Vice President at LaSalle Bank in their credit derivatives servicing division.
September 22nd, 2020
FINANCIAL SERVICES

Get the latest industry news and insights delivered straight to your inbox.

2020-10-22T13:39:19-05:00