All shocks are different: insights from sentiment and topic analysis using LLMs

Iulia Bucur and Ed Hill

Modern language models – think OpenAI’s GPTs, Google’s Gemini or DeepSeek – are powerful tools: but how can we use them in economic policymaking? Economic analysis often relies on decompositions to understand macroeconomic data and inform counterfactuals. But these decompositions are typically obtained from numerical data or macroeconomic models and so may overlook nuanced insights embedded in unstructured text. We propose decomposing the metrics which Large Language Models (LLMs) can derive from text data to offer insights from large collections of documents in a highly interpretable format. This approach aims to bridge the gap between natural language processing (NLP) techniques and economic decision-making, offering a richer, more context-aware understanding of complex economic phenomena.

Continue reading →

Balancing complexity and performance in forecasting models: insights from CHAPS volume predictions

Tom Davies

CHAPS is a critical element of the UK’s payments landscape, handling 92% of UK payment values despite comprising 0.5% of volumes. CHAPS is used for high-value and time-critical payments, including money market and foreign exchange transactions, supplier payments, and house purchases. We forecast CHAPS volumes to help CHAPS participants in making staffing decisions and support our long-term planning including system capacity and tariff setting. While advanced forecasting methods can capture subtle, non-linear patterns, a tension arises: should we use complex models for the most accurate prediction, or use simpler, transparent approaches that stakeholders can quickly grasp? In practice, forecasting isn’t as straightforward as picking whichever model maximises performance; it is the combination of computation and domain expertise that shapes success.

Continue reading →

Using causal inference for explainability enhancement in the financial sector

Rhea Mirchandani and Steve Blaxland

Supervisors are responsible for ensuring the safety and soundness of firms and avoiding their disorderly failure which has systemic consequences, while managing increasingly voluminous data submitted by them. To achieve this, they analyse metrics including capital, liquidity, and other risk exposures for these organisations. Sudden peaks or troughs in these metrics may indicate underlying issues or reflect erroneous reporting. Supervisors investigate these anomalies to ascertain their root causes and determine an appropriate course of action. The advent of artificial intelligence techniques, including causal inference, could serve as an evolved approach to enhancing explainability and conducting root cause analyses. In this article, we explore a graphical approach to causal inference for enhancing the explainability of key measures in the financial sector.

Continue reading →

Payments without borders: using ISO 20022 to identify cross-border payments in CHAPS

James Duffy and James Sanders

Understanding a payment’s journey around the globe can be difficult. As the operator of the UK’s high-value payment system (CHAPS), the Bank is all too familiar with this challenge. By leveraging the benefits of the newly introduced ISO 20022 standard for messaging, we have devised a new methodology to identify and classify cross-border CHAPS payments more effectively. This method reveals that international transactions form over half of CHAPS activity, and offers new insights into the global payment corridors for CHAPS payments. Gaining a deeper understanding of payment flows could assist policymakers in prioritising their efforts to reduce global barriers as they implement the G20 roadmap for enhancing cross-border payments.

Continue reading →

Beyond the average: patterns in UK price data at the micro level

Lennart Brandt, Natalie Burr and Krisztian Gado

The Bank of England has a 2% annual inflation rate target in the ONS’ consumer prices index. But looking at its 700 item categories, we find that very few prices ever change by 2%. In fact, on a month-on-month basis, only about one fifth of prices change at all. Instead, we observe what economists call ‘sticky prices’: the price of an item will remain fixed for an extended amount of time and then adjust in one large step. We document the time-varying nature of stickiness by looking at the share of price changes and their distribution in the UK microdata. We find a visible discontinuity in price-setting in the first quarter of 2022, which has only partially unwound.

Continue reading →

Can data science capture key insights in news articles?

Itua Etiobhio, Riyad Khan and Steve Blaxland

The volume of information available to supervisors from public sources has grown enormously over the past few years, including unstructured text data from traditional news outlets, news aggregators, and social media. This presents an opportunity to leverage the power of data science techniques to gain valuable insights. By utilising sophisticated analytical tools, can supervisors identify hidden patterns, detect emerging events and gauge public sentiment to better understand risks to the safety and soundness of banks and insurance firms? This article explores how data science could support central bank supervisors to discover significant events, capture public trends and ultimately enable more effective supervision.

Continue reading →