UKRI Innovation Fellowship: Intelligent Language Processing for Understanding Financial Text

The project will develop and use corpus and computational linguistics methods and tools to study two important corporate financial disclosures: preliminary earnings announcements (PEAs) and annual reports (ARs). The research is co-produced with the UK financial reporting regulator – the Financial Reporting Council (FRC) and KPMG as industrial partner. The following objectives define the project:

Objective 1: Analyse the properties and economic impact of performance commentaries in financial disclosures and evaluate the need or otherwise for regulatory guidance

The project will develop tools to facilitate automated analysis of performance commentaries in UK financial disclosures. The resulting tools will be used to provide the first large sample evidence on the properties and usefulness of financial reporting narratives, and their alignment with other disclosures. Such evidence is long overdue given the importance of ARs and PEAs as corporate reporting channels in the UK. The research will also inform FRC discussions on the need or otherwise for regulatory guidance in the area of PEA reporting.

Objective 2: Provide insights on a set of contemporary policy-relevant themes relating to the content and structure of UK financial disclosures

The project will provide large sample evidence on trends in PEA and AR narratives. Building on prior ground-breaking work by the research team (ES/J012394/1), new methods and tools will be developed for analysing complex UK financial disclosures. This aspect of the project will contribute to research on financial disclosure content and structure, where extant large sample work focuses on disclosures by US companies. The research will also inform policy discussions on annual report content and design.