Utilising Financial Market Data and AI for Statistics Compilation

The Census and Statistics Department (C&SD) is responsible for compiling Hong Kong’s Balance of Payments (BoP) statistics, which include Direct Investment and Portfolio Investment statistics. Over the years, these statistics have primarily been gathered through statistical surveys. As financial markets become more globalised and complex, there is a rising demand for more timely, comprehensive, and high-quality data to support this compilation.

To address these challenges, C&SD has integrated financial market data and artificial intelligence (AI) technologies into its BoP statistics compilation operations since 2023. By utilising financial market data, C&SD gains convenient access to extensive datasets, including corporate financial information and transactions in debt and equity securities. This approach reduces limitations in data collection, improving both the efficiency and accuracy of statistical work.

To analyse the vast volumes of financial market information, C&SD has applied large language models (LLMs) to automate data extraction, processing and analysis. Integrating LLMs into the statistical processes has significantly reduced manual processing time and improved the quality of statistics produced. This multi-source approach also enhances data quality by cross-validating survey data with independent financial sources.

C&SD will continue to integrate new data sources and apply AI technologies to provide high-quality statistical services that meet the needs of society.