Interim review on Office of National Statistics says it has to obtain more public sector data and improve its technology for better economic forecasting
All public sector data should be made available to the Office of National Statistics (ONS), but the organisation has to greatly improve its technology and data infrastructure to support its economic forecasting, according to a recently published report.
The interim report on the Independent review of UK economic statistics, led by Professor Sir Charles Bean (pictured), says new data is not shared as widely as it should be and this is preventing the ONS from fulfilling its potential as economic forecaster.
Detailed microdata from other public sector organisations could help it to provide more insightful, timely and accurate economic statistics, such as early estimates of GDP, improved regional statistics and more insight into financial, trade and labour market flows.
Among the report's recommendations are that the ONS make the most of data sources and the technologies for dealing with them, and become better at understanding and interrogating data.
Bean, professor of economics at the London School of Economics and formerly deputy governor at the Bank of England, said: “It’s nonsensical that different bits of the government don’t speak to each other, so that businesses and households have to provide the same information twice. Unlocking the data hoard already held by the public sector will not only save businesses money but also produce more timely and accurate statistics.
“A culture shift at the ONS is also key to producing economic statistics for a modern economy. It needs to become an organisation which is more intellectually curious, open and self-critical, as well as better at engaging with its customers.”
The report says the ONS suffers from significant problems in obtaining and managing data. Some of these are cultural, with government departments often resisting the call to make their data freely available, even to the ONS, and a legislative framework that requires the organisation to specify how the data would be used.
The ONS is also criticised for being slow to grasp the opportunities in new sources and continuing to rely heavily on survey data.
On the technology front, its IT systems are poorly interconnected, with hundreds of applications running on 25 different platforms, many of which are outdated or bespoke and costly to maintain. This has contributed to errors, such as in the International Passenger Survey of 2014 when researchers could not directly interrogate the data.
The report acknowledges that the ONS is in the early stages of transforming its technology in line with standards from the Government Digital Service, with plans to build tools on common platforms. But it warns against repeating past mistakes such as those that hit its Statistical Modernisation Programme in the early 2000s.
It also highlights the need for a flexible data infrastructure that allows staff to remove duplicates and correct errors, and compare data sources to exploit their analytical value. Currently, statisticians have to use complex matching processes to link data sources – a process that could be simplified by the consistent use of unique identifiers.
The Department of Business, Innovation and Skills is currently leading work on reconciling different sources of information across government. The report says that measuring public sector services is highly important as they account for about a fifth of GDP, but that it is difficult to measure their value as most are provided free or with a token charge.
In addition, it emphasises the need for stronger links with data scientists, and says the ONS Big Data team has begun to develop collaborations with groups of academics.
The ONS has responded with a statement that it welcomes the interim recommendations, looks forward to the final report and “will consider its conclusions carefully” in preparing plans for the future.
Bean is due to provide evidence to the House of Commons Treasury Select Committee this week and the final report is scheduled to be published in time for next year's Budget announcements.
Image from the London School of Economics