Forecasting
Paper presents forecasting method for bank losses
Bank of Finland paper attempts to forecast both “expected” and “unexpected” banking losses
The changing data landscape: Part 1
Central Banking speaks to Eyal Rozen, Ramūnas Baravykas and Wanpracha Chaovalitwongse about whether there is a need to change underlying infrastructure to bolster data-driven policy-making
ECB outlines changes to forecasting during Covid
High-frequency data, GDP-at-risk, changes to use of PMI and stringency index all useful, authors say
Economics Benchmarks 2020 – presentation
Central Banking’s economics subject matter specialist Daniel Hinge speaks with Christopher Jeffery about how central bank economists fared in a year where the Covid-19 pandemic upended the usual business of forecasting, analysis and research
Administrative data is most popular alternative data source
Central banks make use of a wide range of non-traditional data sources
Central banks use alt data mainly for research
Almost all respondents make use of alternative data in at least one application
Few central banks forecast policy rates
Economics Benchmarks 2020 highlights wide variation in variables forecast by central banks
Central banks rotate economics staff frequently
Economists typically spend only small portion of time on own research; opportunities for secondment abound
Climate change on the agenda for most economics departments
Central banks explore a wide range of topics, with some differences between advanced economies and EMEs
Agent-based models remain rare among economics departments
Economics Benchmarks 2020 highlights the varied applications of different model types at central banks
Forecast errors marginally higher for growth than inflation
Data shows contrast between advanced and emerging economies
Semi-structural models are the forecast weapon of choice
Flexible modelling approach comes out on top; around half of central banks include a financial sector
Currency Benchmarks 2020 report – the data behind the cash cycle
Perspectives on staffing, circulation, forecasting, fraud, substrate choice, outsourcing and climate risk
Bank of Italy paper compares US recession-forecasting methods
Measure of uncertainty out-performs even yield curve in short term, researchers say
IMF warns of permanent economic damage from Covid-19
Pandemic could pitch 90 million people back into extreme poverty, fund says in new economic outlook
Haldane makes case against ‘economic anxiety’
Communication essential to avoid caution “morphing into fear and fatalism”, says BoE chief economist
‘Combination’ forecasts can beat standard methods – RBI paper
Traditional inflation-output relationship seems “broken”, authors say
Deep learning can beat other forecast methods – Bank of Korea research
Deep learning models have smaller margin of error than conventional methods, research finds, but data quality is key
RBA expects long period of low inflation – minutes
Policy set to remain on hold, as board members say term funding is having desired effect
Bundesbank paper uses neural networks for rate forecasting
Artificial neural networks outperform linear and non-linear functions, author says
Covid-19 policy-making and the need for high-speed data
High-frequency data holds the promise of speed and adaptability, but the rush to find alternative economic indicators has the potential to create problems
Book notes: Radical uncertainty, by Mervyn King and John Kay
The one certainty we have faced is that we have to confront uncertainty, which is precisely the point of this wonderful book
ECB paper looks at Bayesian nowcasting
BVAR models have several advantages over Dynamic Factor Models, researchers say
RBNZ boosts asset purchases and prepares new policies
Negative rates and funding for lending programme under “active preparation”