Economics
Economics Benchmarks 2023 report – tentative steps on AI
The benchmarks show some small steps to increase AI use, and shed light on broader reforms to modelling, governance and more
Central banks continuously update their modelling frameworks, but the past year saw a surge in research interest in one particular area: the use of artificial intelligence. Overall, AI adoption in central bank economics departments is limited, but progress looks set to be much faster in the years ahead. The benchmarks also include data on staffing and salaries, economists’ qualifications, research time and output, economists’ use of data, and more.
Charts
Research focus
After a year of inflation surprises, central banks are very focused on research into inflation dynamics – the topic received a score of 4.8 out of a possible 5 from respondents. The monetary transmission mechanism is also a major focus.
Other areas are generally less pressing. The impact of technology and artificial intelligence received a score of just 2.6 across the sample. However, lower-middle income central banks see the topic as much more important.
For the full breakdown, use the benchmarking service’s interactive charts to explore the data.
Model banks analysis – Economics 2023
Data reveals differences in research, forecasting and data use across central banks
Economics data
View the full breakdown of responses to the Economics Benchmarks.
Previous years
Economics 2023
Economics 2022
Economics 2021
Economics 2020
Economics Benchmarks 2023 – model banks analysis
Data reveals differences in research, forecasting and data use across central banks
Economics Benchmarks 2023 report – tentative steps on AI
The benchmarks show some small steps to increase AI use, and shed light on broader reforms to modelling, governance and more
Central banks adapt forecasts with financial sector and other changes
Host of changes to modelling incorporate financial channels, nuanced expectations and more
Inflation forecast errors moderate after surge in 2022
But growth forecast errors are higher in 2023, data shows
High income central banks allocate most time to research
Research economists are more numerous and are allotted more research time in high income jurisdictions
Administrative data tops list as alternative data source
Research and forecasting remain key areas of application for central bank economists
Economists’ rate of access to centralised data services still below par
Central banks in high-income jurisdictions still have more access
Most heads of research are responsible for sign-off on research
Peer-reviewed papers remain most-used metric for measuring research success
Central banks average 11 working papers over the past year
Just over a quarter of central banks run research blogs
Central banks prioritise inflation dynamics research
Middle income institutions tend to have much widest research interests
Time series model is top tool for forecasting and research
Use of machine learning models for research purposes surges year on year
Economic growth and inflation are most-forecast variables
Private debt, commodity prices and market interest rates rank as laggards