How Finance Uses Natural Language Processing

Crunching text is the new financial frontier

Natural Language Processing (NLP) is a collective name for a set of techniques for machines to uncover the structure within text data. 

Modern banks and investment managers have built their business on crunching numbers. But, with access to information no longer the competitive edge it once was, pockets of value have become much scarcer. Large volumes of text have become the new frontier for hidden market signals.

Moreover, growing volumes of text information is overwhelming employees.  it’s also becoming harder to keep handling text data with the same processes.

Research Outline

The research aimed to educate financial industry insiders on the world of possibilities that NLP now offers. FinText decided to tell this story through powerful case studies showing the different ways NLP was being put to use in large financial companies – and generating tangible rewards.

The report also explains key NLP and machine learning concepts (Topic Modelling, Named-Entity Recognition, Feature Selection etc.), assuming no prior knowledge.

Finally, the research introduced some of FinText’s use of NLP, applying text analytics to improve the processes of creating effective marketing for financial products.

Findings

FinText’s writers researched and wrote a series of case studies. Initially, these were published as gated content, but we’ve since made the information publicly accessible.

Applications in banking and asset management:

• Deutsche Bank uses NLP to weed out greenwashing
• Barings Asset Management uses NLP to track buzzing companies
• JP Morgan Chase used NLP to slash legal costs
• State Street uses NLP to boost its investment research arm
• American Century scans earnings calls for fibs
• N26 created a chatbot on the cheap
• ING Merchant Bank uses NLP to connect corporate entities

FinText Improves Investment Marketing with Text Analytics

“All asset management content looks the same!” is a common concern among industry marketers. But our data shows that different problems can plague companies’ marketing material.

For this case study, FinText analysed 255 articles published by seven investment managers during the first quarter of 2020. The articles analysed were targeting both intermediaries and retail investors. We also benchmarked against 664 financial news stories published during the same period on City A.M., a widely circulated UK financial daily paper.

We benchmarked this content against two measures. The first was complexity (average percentage of words in a text that had three syllables and four syllables or more). The other considered how frequently investment managers used the first person (in average percent of sentences)

Both tests reveal interesting findings. Crucially – and unlike vanity metrics – these are forward-looking metrics, offering actionable steps to increase content appeal.

Revisiting the charts side by side reveals company-specific weaknesses relative to competitors.

For example, investment manager B writes in very clear language, but talks about itself in a way that stands out as self-centred, when compared to its competitors.

By contrast, investment manager G doesn’t refer to itself that much, but uses very complicated language. Faced with other options, readers are likely to prefer the insights of a more accessible investment manager.

Other metrics –  including on quantities published and topics covered, add further detail – and point marketers towards specific actions to improve content success.

Outcomes

1.Media Coverage

The topic’s appeal won the report meaningful media coverage in leading industry publications. Most notably, in Institutional Investor, which wrote:

Asset managers seeking an edge in the uncertainty of the pandemic might do well to turn to natural language processing the way firms including American Century Investments do — and avoid the task of digesting massive volumes of research, according to text analytics company FinText.

In a recent paper looking at the ways finance firms uses the machine learning application, FinText said American Century tries to detect deception in management language during companies’ quarterly-earnings calls. Its sentiment model checks for omission of important disclosures, spin, obfuscation, and blame.

Firms such as Barings Asset Management, State Street Corp., and Deutsche Bank are also using natural language processing, according to the paper. The technology removes “text-related grunt work, allowing employees to focus on higher-value tasks,” FinText said in the paper.

2. Establishing the ‘NLP in Financial Services’ content niche

Prior to this report, AI or machine learning in financial services were already hot topics, but NLP in financial services had yet to emerge as a theme. But interest has been consistently growing.

A companion article to this research was published in established machine-learning journal Towards Data Science. For over two years, the article continues to attracts views daily, mostly through Google search.