BlogJuly 20, 2018
How Natural Language Processing is Changing Financial Risk and Compliance and Why You Should Care
Monitoring New and Changing Regulations
A 2017 survey indicated that compliance officers' foremost concern for the year was the "volume and pace of regulatory change." Traditionally, FIs have been reliant upon both internal staff and external vendors to track regulatory changes and analyze their impact. It is almost impossible to apply a manual solution to this problem at any reasonable cost, and the historical reliance on labor-intensive monitoring and review of multiplying and ever-changing regulations is untenable.
Luckily, document and process reviews ─ a key component of regulatory change management ─ are among the first domains where natural language processing algorithms have been successfully applied. RegTech companies have been quick to embrace this technology, as their solutions are capable of reading and reviewing multiple regulatory sources and automatically identifying and alerting FIs of any recent changes or additions. Depending on the specific software, the solution can also automatically determine the relevant internal stakeholders who need to be alerted of the change, as well as analyze how the change may impact current internal policies and procedures.
For example, efficient regulatory change management software may be able to automatically detect a change in a certain country's financial regulations and alert both the relevant country-level compliance officers, in addition to any relevant global compliance staff. Certain RegTech companies already provide AI-based regulatory change management platforms, with some monitoring regulatory changes in over 180 countries and 60 languages.
Although regulators have not been overly forthcoming with specific approvals or guidance for the use of technological solutions, there is some indication that they are becoming increasingly aware and comfortable with the use of RegTech. In fact, some regulators are even looking at using RegTech themselves. The most prominent is the Securities and Exchanges Commission (SEC). The SEC first began experimenting with RegTech by employing NLP algorithms in the review of tips, complaints, and referrals (TCR) data looking for undiscovered patterns. Following their success with reviewing TCR data, the SEC began utilizing NLP algorithms to identify commonalities in disclosures produced by firms charged with wrongdoing, resulting in the discovery of consistent language patterns which could be applied to future disclosures. The SEC has since spread its use of NLP algorithms to review investment advisor prospectus. '
In a speech on the SEC's use of RegTech, Scott W. Bauguess, former SEC Acting Director and Acting Chief Economist, stated, "The results are impressive. Back-testing analyses show that the algorithms are five times better than random at identifying language in investment adviser regulatory filings that could merit a referral to enforcement." Better than random may seem like a low threshold, but it is also the reality SEC analysts face. Without assistance from the NLP algorithms, they have no way of knowing whether an unreviewed filing contains suspicious language, and therefore must select randomly from their backlogs (excluding reviews prompted by tips or previous investigations). Implementing NLP algorithms, however, allows the SEC to conduct more targeted reviews.
Regardless of the specific RegTech applications financial institutions choose to implement, RegTech offers the most scalable and effective solutions for tackling increasing regulatory change and enforcement. This is not to say the compliance officer will disappear. RegTech still has its shortcomings, and regulators are unlikely to ever trust an entirely automated solution for managing key regulatory risks.
However, as Ed Sibley, director of Credit Institutions Supervision, Central Bank of Ireland said in March 2017, "We need to be alive to the disruptions that are coming, to be flexible and adaptive and recognize that successful implementation of new technologies can drive significant efficiencies and greater robustness." With efficient and responsible implementation of regulatory technology, financial institutions will be able to reduce regulatory burden, better allocate internal compliance resources, and improve overall internal compliance.