UK regulator raises concerns about labour data quality

LONDON (Reuters) – Britain’s statistics regulator raised concerns on Tuesday about the Office for National Statistics’ (ONS) approach to calculating unemployment and other key labour market data, following a big drop in response rates for its main survey.

The ONS switched at short notice to a new, experimental methodology for data published from October onwards, after response rates for its Labour Force Survey fell below 15%, down from 48% a decade earlier.

Although the new unemployment estimates are similar to previous ones, they come at a sensitive time as the Bank of England seeks to assess how fast the economy is slowing, and they no longer include reasons why people are not working.

The Office for Statistics Regulation said the ONS should have done a better job explaining the changes, as it was hard to judge how reliable the new numbers are.

“ONS should publish suitable information on the methodology underpinning the new experimental method, its approach to quality assurance and whether international standards are being met,” it said in a report.

The new unemployment and employment numbers are currently being calculating by using separate tax and unemployment benefit claims data to extrapolate from previous LFS numbers.

The ONS has been planning a major revamp of its Labour Force Survey for early 2024, but it introduced interim measures due to more urgent concerns about the data.

The regulator said the ONS should have acted sooner.

“ONS could have done more to prevent these circumstances from arising,” it said. “ONS should identify what lessons can be learnt to more effectively and transparently manage and pre-empt quality issues in the future.”

The ONS said falling survey response rates were a problem for statistics agencies in other countries too, and that it had already responded to some of the regulator’s concerns.

“We will now take some time to consider the report’s findings before responding in full,” an ONS spokesperson said.

Source link