The pandemic has demonstrated the value of rapidly-available government statistics, Sir David Norgrove, chair of the UK Statistics Authority, has said.
The Office for National Statistics, which Norgrove oversees, used quickly-available data on how people were responding to Covid-19 including from traffic cameras, mobile phone locations and new surveys and questions to track social distancing. “These data were needed fast, in days not weeks,” he told an online event organised by University College London.
The government’s economic response benefitted from HM Revenue and Customs’ access to real-time data and swiftly-published private sector information such as on credit and debit card activity and job adverts. “We would really have struggled without those,” said Norgrove. “We needed not just to compile historical data but rather present a picture of what was happening right now, day by day.”
Faster statistics would have been useful in previous crises, he said. If the Bank of England had known about the 2008 economic downturn a month or two earlier, it could have taken action saving billions of pounds in lost output. But he added that such data needs to be available openly, with “too many cases” of ministers using unpublished numbers in public followed by too long a gap before publication that reveals sources and caveats: “The public needs to see what the government is seeing.”
Norgrove said that he could only think of a single case of deliberate data distortion during the pandemic. In April 2020, then-health secretary Matt Hancock claimed that a target of carrying out 100,000 Covid-19 tests a day had been met, but only by counting tests that had been sent out as well as those completed. Despite Norgrove complaining, the double-counting continued until August: “This was poor practice, to say the least.”
The government also did a poor initial job of presenting data, with Norgrove describing one chart on the national outbreak from the prime minister’s 23 March 2020 "stay at home" address as being “as clear as mud”. This year, ONS staff have been seconded to help produce clearer, better-labelled graphs for official presentations.
Replacing the census with ongoing data
Norgrove said that he took pride in how the ONS ran the census in England and Wales in March, but added: “I hope it will be the last one.” Sir Ian Diamond, the national statistician, has said that census-level estimates should be published every month rather than every decade and the ONS will present plans for the future of population statistics in the next couple of years.
Such an ongoing model is used for the weekly Covid-19 infection survey run by the ONS based on a random sample of people using oral swabs and blood tests. Norgrove said this kind of research could be used for continual tracking of population health, although the current survey costs a quarter of a billion pounds annually.
He added that combining the results of that survey with data on infections, from the census and from GP health records has had strong results: “Without that work we wouldn’t have been able to confirm what we all suspected, the vulnerability caused by deprivation and related to ethnicity,” he said, also the basis for research on long Covid. “That’s why I get animated about how this kind of analysis is jeopardised by the campaign for opt-out from sharing of health records.”
The use of linked data by ethically-approved researchers in a secure environment has great promise in tackling public health issues, he added.