Is analytics the missing piece in a complex settlement jigsaw?

By Daniel Carpenter, head of regulation, Meritsoft (a Cognizant company)

29 October 2020

At a time when the costs of doing business are skyrocketing, is it sustainable for banks to persevere with the current approach to trade settlement failure?

While there will always be trades that fail, the tolerance levels of banking boardroom execs must be at breaking point right now if the latest Esma Trends Risks and Vulnerabilities (TVR) report is anything to go by. The study shows a dramatic surge in the level of settlement fails during the second half of March, with fails climbing to around 14 percent for equities and close to six percent for government and corporate bonds.  

Even once the pandemic volatility subsides, and assuming the percentage of fails come down, this does not mean the issue goes away. On the contrary, with every day that goes by banks are increasingly hamstrung by reduced profitability, which means they can ill afford not to take a deeper dive into the root causes of their settlement issues. And while CSDR has been delayed, penalties for fails will be significant when this is introduced.

The problem really starts with trying to understand exactly where and why trades are failing to settle. If, as a bank, you are constantly experiencing no trade confirmations back from a particular counterparty, then there is a clear need to reassess the nature of this relationship.

What happens if, for example, a trade confirmation does not leave a firm by the end of the working day? Or if a trade is completed in the last hour of the trading day but is unable to get out of the building by 10am the next morning? If it is suddenly 11am and the trade still has not left the building, is this a clear warning sign that the trade is at a high risk of failing to settle? Even if the trade is on time, there is always a risk that the trade information does not match up with the information that the counterparty has. Unless the bank understands the depth and breadth of the problem – through the right analytics – it is nigh-on-impossible to address these real challenges.

Certain trades failing to settle will, of course, always be accepted part of the trading and settlement process. But this does not mean that the unnecessary costs can’t be removed. Huge strides are being made to improve communication between counterparties, so they know what to expect, but this is fundamentally not going to resolve the issue by itself. 

This is where AI and analytics could be deployed to look into the likelihood of and sources of the most costly and frequent fails. With a single pane of glass, through which all the necessary transaction data can be viewed, banks can then layer AI and data analytics capabilities over the top to figure out which counterparty trades are most likely to fail and when and what they can do about it to limit that number. With this information, banks can then adjust their counterparty relationships based on more detailed insights, as opposed to simply accepting the fail as the nature of the business. Some banks are already moving in this direction, with one very well-known bank recently announcing the upcoming launch an early warning securities settlement service that utilises machine learning to minimise the total number of failed trades. 

The issue of trade fails is not going away and there is a collective responsibility, as well as a regulatory requirement with CSDR in 2022, for the industry as a whole to address this issue. Analytics has to be front and centre of the settlement conversation and identifying how your counterparties are performing through more in-depth analytics and how you should react to their behaviour through predictive AI should be a focus for all concerned.

Become a bobsguide member to access the following

1. Unrestricted access to bobsguide
2. Send a proposal request
3. Insights delivered daily to your inbox
4. Career development