HM Revenue and Customs has been urged to make a series of changes to how it produces official statistics after a double-counting error added almost £7bn to its corporation tax receipt statistics.
The probe by the Office for Statistics Regulation was launched after a mistake was uncovered in how tax credits that either reduced a company’s liability to corporation tax or made a payment to the firm, were accounted for, with some being added to the payments received. As a result, corporation rax receipts were £650m lower in 2011-12, £2.3bn lower in 2017-18, and £4bn lower in 2018-19 than previously thought.
This led HMRC to ask OSR director general for regulation Ed Humpherson to review the department’s quality assurance regime for official statistics, and Humpherson published his findings on 30 April.
The report found that “HMRC produces lots of official statistics and most of the time it does this well”, with its knowledge, analysis and intelligence (KAI) directorate is responsible for around 65 official statistics, as well as a broad range of other analytical outputs used internally by HMRC.
However, Humpherson said that quality assurance and quality management could be strengthened, with “several common themes and areas for improvement emerg[ing] from our review”.
The review showed quality assurance was not always adequate across teams in the directorate, while initiatives that would support improvements in quality management were not always prioritised.
KAI needs to create and maintain a culture that invests more in quality, which will require commitment and action from producers at all levels, Humpherson said, including senior managers, and will require a review of the resources available to the group.
The review made nine recommendations “to strengthen the quality culture in HMRC”, which included improving quality assurance across KAI and creating a more effective relationships between KAI and other parts of HMRC.
Further recommendations included adding stronger leadership and oversight of quality management in KAI, and improving the information KAI publishes about the quality of its official statistics.
The review warned that “If KAI does not create this culture, the production of statistics could resemble a series of mechanical processes that do not always identify significant quality issues”.
It added: “KAI does not operate in isolation from other parts of HMRC and long-term improvements to quality management will rely on understanding, values and responsibility being shared across many areas of HMRC, especially those that own or process administrative data.”
In his letter to HMRC chief executive Jim Harra setting out the findings, Humpherson said most of HMRC’s IT systems had been designed for operational delivery, rather than for statistical analysis, as was the case in many other large government departments.
This means “it can be difficult for analysts in KAI to ensure that they have access to the right data, in the correct format, for their analytical work”, he said.
“Long-term improvements to quality management will rely in part on IT system and data owners in HMRC understanding and prioritising the department’s statistical requirements alongside operational considerations,” Humpherson said. “Investment in underlying IT systems will also be important to enable HMRC to continue to confidently exploit the value of the data it holds.”
Responding to the report, a HMRC spokesperson said it welcomed the publication of the report.
“The review has found there are no systemic errors in how HMRC produces its official statistics and manages the quality of its data and statistics." the spokesperson said.
“There are nine recommendations in the review, which HMRC accepts.
“KAI will lead a programme of work to implement the recommendations, working closely with data providers from across the department.”
In his written response, Harra asked Humpherson to meet with HMRC's executive committee to discuss the review, while Ruth Stanier, HMRC's director general, customer strategy and tax design, said HMRC accepted the recommendations in full and had developed a substantive work programme to address the recommendations. The work strands will cover: working with data providers to increase understanding of input data, including mapping data pipelines; improving quality assurance, reviewing HMRC's official statistics’ output to ensure resources are used most effectively, and governance and communications on the response. The department will publish an update on this work next year, Stanier said, "and I look forward to further engagement with the OSR on HMRC’s progress".
The full report can be read here, and Harra and Stanier's responses can be viewed here.