By Prashant Panjwani , Manager, Business Consulting, Publicis.Sapient
The post-trade processing of over-the-counter (OTC) derivatives and market infrastructure landscape have changed considerably since the first executed interest rate swap in 1981. However, regardless of some advances, the swap trade confirmation process still requires significant human intervention, highlighting the need for automation—and more specifically, robotic process automation (RPA).
Industry advances have enhanced the timeliness of confirmation processing and have mitigated operational risks. First, short-form confirmations have mostly supplanted long-form paper confirmation, lessening the number of field firms needed to exchange and match. Second, the development of industry solutions that leverage industry-standard International Swaps and Derivatives Association (ISDA) documentation presently encourage the electronic exchange and matching of trade details.
Yet in spite of the above industry developments, the swap trade confirmation process relies on time-consuming human intervention mainly due to the following reasons:
* Customized confirmation templates that differ among counterparties
* Lifecycle event management that leads to additional complexities such as cash settlements and partial unwinds
* Operational errors in trade capture especially for non-electronic and non-standardized trades
* Managing nuances of individual asset classes
* Increased regulatory reporting requirements such as MIFID II
Human intervention not only brings the operational efficiency down but also increases the costs proposition. It requires significant manpower to manage the trade confirmations on daily basis. Further, scaling trade volumes provide additional challenges in terms of controlling expense and assets.
Bring on the robots
RPA provides the opportunity to implement intelligent automation solutions while minimizing the time and cost associated with swap trade confirmations.
In simple terms, RPA is a software that helps automate any repetitive, rules-based and high-volume process that either requires human effort or is susceptible to human error and miscalculation. A software “robot” is utilized to capture the steps required to complete a process that may involve interacting with an application. The software works with the application in the same way a human does.
To fully appreciate the range of challenges related to swap trade confirmations, we should start by looking at the generic swap trade confirmation process flow, as depicted in Figure 1.
At a glance, the above process flow might look very simple. However in reality, it is extremely intricate. In the operational review, buy-side firms spend considerable time and manpower to ensure all key trade terms are accurately captured. The operational review is further muddled when broker-dealers have different confirmation templates. Subsequently, the operational users invest more energy and time managing varied formats.
Another challenge lies in the nuances of different asset classes. For example, terms that need to be matched for an interest swap contract differ from an equity swap contract. The operational knowledge of such products is vital to efficiently manage the trade confirmation process.
Finally, timelines are extended with trade disputes. Generally, in the case of a trade dispute, firms ask broker-dealers to send revised confirmations after correcting the trade terms. In this scenario, firms have to engage in a second operational review of revised paper confirmations.
The RPA solution
To extend RPA to the swap trade confirmation process, firms should create a rules-based framework that permits users (using RPA tools) to configure business rules to extract and link counterparty trade data with the firm trade data. Automated bots are deployed that mimic human actions to look-up and reconcile the key trade attributes to arrive at a match.
The logical architecture usually entails three key components:
Centralized configuration of rules—Enables a “create once, used by many” paradigm and also eases the operational maintenance of rules on an ongoing basis
Bots—Mimics the user workspace so they can run the process just like humans by using configured rules. The number of Bots can vary depending upon the operational need, just like adding more people to a process if required
Monitoring—Usually involves specific alerts to avoid process interruptions due to infrastructure and environment issues
Figure 2: The swap trade confirmation process with an RPA implementation
However, extending RPA to the swap trade confirmation process should not be an all or none proposition. The process itself is complex and has many exception paths, including extracting trade data from different counterparty templates, navigating third-party applications and handling mismatches and disputes. The approach should be gradational in nature, automating the use cases that are less complex but high in value first, and then gradually moving up the value chain.
Below is a six-step approach for such an implementation:
#1 Process Analysis—The first step in the RPA implementation entails a detailed process analysis. This is where the process in its current form should be documented and outlined in detail. The different interfaces and touchpoints should be determined and each distinct process path should be listed clearly.
#2 Process Prioritization—This step involves prioritizing each listed use case. At a minimum, every use case should be labelled by its level of complexity and value. Process prioritization is important as it clearly lays out what should be the starting point for automation.
#3 Process Re-engineering—This should be thought in terms of how a current process needs to be tweaked for efficiencies in order for automated bots to perform the same process with minimum interruptions and exceptions.
#4 Process Implementation— In this step, specific business rules are configured within RPA tools to be utilized by the bots to run the process in an automated way. It’s similar to outlining process instructions for bots.
#5 Process Validation—This is the last step from an implementation standpoint. It entails testing the different touchpoints and interfaces, including validation of the configured rules and running the automated bots to perform a process in a simulated environment.
#6 Process Monitoring—It is an ongoing step of monitoring processes and bots for interruptions, failures and overall performance. It’s a feedback loop of the previous five steps that provides real-time information about any deficiencies, paving the way for an incremental approach to implementation.
What does RPA really hold for Swap trade confirmations?
RPA for swap trade confirmation provides a perfect opportunity to automate the operational review of trades. Bots can use a rules-based framework to automatically link trades, match key trade terms and approve the trade for a next level review. The automation of these operational steps could result in significant efficiency gains and improve scalability while minimizing costs and manual effort. And by switching to a rules-based automated operational review, companies can better manage operational risk.
