Over the past decade, investors have increased their use of natural language processing (NLP) technology to analyze text and other forms of unstructured data to unearth signals to boost investment returns. Today, institutional investors are also looking to extend NLP technology to other core functions like surveillance and compliance and its ability to deliver accurate sentiment analysis.
A quarter of the quantitative investors participating in a recent study from Greenwich Associates believe has important implications in the surveillance of internal and external communications, with one-third indicating that such use would cover voice as well as other communications (e.g., social media).
A new report from Greenwich Associates analyzes how investors in North America, Europe and Asia are using NLP technology and how they prioritize NLP functions and benefits, and provides an examination of how investors are implementing NPL—including a look at the top external vendors in the field and in-house requirements for resources and talent.
NLP Vendors Provide Value…
Achieving the benefits of NLP requires an in-depth understand of data sets, tools and techniques for integration. Of course, not all firms will be able to dedicate or develop the internal resources to build out such an infrastructure.
“Financial firms continue to face challenges in terms of resourcing and budgets due to low interest rates and fee compression, a situation that the COVID-19 crisis can only be expected to exacerbate,” says Shane Swanson Senior Analyst for Greenwich Associates Market Structure and Technology and author of Sentiment, Signals and Surveillance Natural Language Processing for Financial Markets.
As a result, many firms will rely on outside vendors to make it possible. NLP vendors often can bring more resources to bear on these problems than a financial firm in isolation. In particular, breadth and standardization are areas where vendors provide tremendous value. Consultative engagements with vendors have also helped firms generate or enhance unique alpha signals.
…But NLP Will Always Require In-House Resources and Talent
However, even firms that use external vendors should be prepared to invest significant in-house resources in the NLP process. In trading and investing, alpha signals must be monitored on an ongoing basis due to their tendency to lose predictive value as the information becomes available across a broader segment of the market. In compliance and surveillance, firms must be on guard for any faulty results that could put the organization at risk.
“In other words, there is no escaping the need for some level of in-house expertise for the proper implementation of an NLP system,” says Swanson. “This demand has created a highly competitive market for data analytics and technology talent.”