The data hierarchy

Weekly Updates

Economic data quality has been declining for years, and the pandemic accelerated the problem. The UK has torn up its labor data because no one answers its surveys. Other countries face similar challenges.
There is a hierarchy of trust in economic data. The best data comes with legal obligation. We have confidence in bank lending figures, because the banks’ regulator (normally) demands the data. Data derived from tax information, in countries with low evasion rates, is similarly reliable.
Big data can be reliable. Satellite images counting container ships outside a port are very accurate. There is a difference between data and interpretation. Counting sandwich sales is only accurate information about sandwich sales—it may not measure how many people are returning to offices.
Survey-based data make up most economic data releases, but are least reliable. Some surveys are narrowly accurate—consumer price inflation reliably records a selection of prices, but that selection of prices probably does not reflect an individual’s inflation experience. Other surveys depend on people responding, answering the actual question asked, and answering honestly. All three rarely happen together. Surveys clearly detailing the questions asked and response rates generated allow economists to make informed judgements about accuracy. Just giving a survey response without those details is not helpful.