The Office for National Statistics (ONS) has been underestimating the country’s gender pay gap for the past twenty years, according to new findings released in the British Journal of Industrial Relations.
The report suggests that, since 2004, the UK’s statistics agency has failed to properly account for the fact it received more data from larger employers. The gender pay gap is a core measure of economic equality. If policymakers and businesses have been working with an inaccurate picture of earnings in Britain, official pay recommendations and anti-poverty measures, such as the national minimum wages, may be wrong. Employees and advocacy groups who rely on official statistics to push for fair pay may also feel misled.
This is not the first time the number-crunching body has come under scrutiny over errors, and there are growing concerns that some of the statistics it produces are no longer fully reliable.
Cracks beginning to show
In 2023, Britain discovered its economy was bigger than initially thought. Revised figures from the ONS for 2020 and 2021 showed the economy was larger than earlier assumptions by around 2% – equating to £50bn. In the two years that followed, the ONS has battled to maintain reliable economic data as problems related to labour and growth figures also emerged.
In a Mansion House speech last year, the Bank of England’s governor, Andrew Bailey, said that unreliable labour market statistics have become a “substantial problem” for the UK central bank for policymaking. The bank’s chief economist, Huw Pill, argues that ONS figures are most likely understating employment growth and overstating the high rates of inactivity and joblessness.
It’s to the ONS’s credit that it is open about its weaknesses
This doubt over job statistics has arrived at a crucial moment for the UK labour market. Unemployment is rising while business lobby groups state Rachel Reeves’s tax policies have led job losses.
Problems have also arisen for other statistics produced by the ONS, with new questions about survey data used to formulate GDP readings, including the producer price index (PPI) and services producer price indices (SPPI).
“Mistakes happen in all organisations, but we rarely hear about them. It’s to the ONS’s credit that it is open and honest about its weaknesses,” says Keith Church, head of economic modelling at 4most, a data consultancy. That said, some of the latest issues are “hardly reassuring,” he adds.
Part of the ONS’s core responsibility is to adapt quickly to new data challenges and that’s where it has, at times, moved slowly. Church blames a lack of change management and insufficient oversight. “The issues can be fixed, but the skills needed to address them are in high demand across the private sector,” he says. “Some of the more material problems like falling survey response rates might have been anticipated and tackled sooner.”
Antonio Fatás is a professor of economics at INSEAD and research fellow at the Centre for Economic and Policy Research in London. He agrees that there is an opportunity for the ONS to move much faster. But the ONS has suffered from a lack of funding. Complicating matters further are ongoing disagreements among economists around methodology – how to best measure data and what defines economic indicators such as inflation – a complex and constantly evolving topic.
In his view, the agency must take a much stronger stance in promoting participation in data collection efforts of national significance. The ONS was forced to scrap the publication of the labour force survey last October after dwindling response rates left them so low that they were no longer reliable. Although surveys, tax records and online job adverts give the Bank of England (BoE) alternative indicators of trends in hiring, the survey is the only source for estimates of unemployment and, crucially, of economic inactivity.
While this is not uniquely a UK problem – data revisions are common in many countries – the problems appear to be “particularly prevalent” at the ONS, Fatás says.
In response, the government has launched a review into the agency’s failings, particularly around the frequency and scale of data revisions. But the ONS has admitted these problems that could take until 2027 to rectify. So where does that leave policymakers and business leaders?
How does unreliable macro data impact businesses?
Inaccurate macro data can affect every part of the economy. Insurers look at GDP trends when pricing coverage or modelling long-term risk. Banks use it to shape lending policies and companies rely on it to set expectations for growth. Meanwhile, advocacy groups use pay data to set anti-poverty and gender equality recommendations. If the base data is wrong, the decisions made off the back of it are going to miss the mark.
“Macroeconomic data and forecasts are valuable tools for planning and forecasting,” stresses Church. But he asks: “How can we set spending plans when we aren’t even sure on something as fundamental as population size?”
However, the limitations of ONS data may not be as damaging as they first appear. Fatás, who also works as a consultant for organisations such as the International Monetary Fund, the OECD and the World Bank, says macroeconomic data may set the backdrop for the environment businesses operate in but, in practice, firms should generally detect shifts in conditions long before official data becomes available.
Business leaders should be looking at a wide range of statistics when formulating strategies and forecasts, he continues, including their own data, employment figures and consumption statistics (the use of goods and services).
Furthermore, there is a case that the ONS output is better tuned in to business needs than ever before. Church notes that HMRC’s estimates of payrolled employment are published every month and these real-time indicators provide a good early view of consumer spending.
Data is not perfect: take it with a grain of salt
Some of the criticism levelled at the ONS is not wholly justifiable. Economic data relies on surveys and so will always carry a degree of uncertainty. Economies are complicated and data-driven economic models are guides to, rather than substitutes for, reasoned thinking.
“There is probably an unrealistic view about the extent to which it is possible to fine-tune monetary policy even with better data,” says Church. Problems with the labour force survey have spurred policymakers to broaden the range of indicators they look at, which is “a step in the right direction”.
Measuring an entire economy with millions of people and hundreds of thousands of companies is incredibly hard, says Javier Díaz-Giménez, a professor in the economics department at IESE Business School. Inaccuracies in statistical agencies are a global issue, he adds, stemming from inherent challenges in measuring complex economies, such as incomplete information and the difficulty of accounting for the underground economy – economic activity that is unrecorded and untaxed by governments.
“Business leaders, policymakers and central banks should always take these numbers with a grain of salt,” Díaz-Giménez says. Macroeconomic figures, like GDP, should be viewed as “numerical metaphors” rather than precise measurements, subject to revisions and influenced by events like changing tariffs, he adds. “They provide a broad indication rather than absolute certainty and trying to read too much into small fluctuations (like 0.1% growth) is not advisable.”
An alternative to GDP
There is a growing consensus among economists that GDP alone is not a sufficient measure of economic progress and societal wellbeing. “The concept of GDP is quite old – nearing its 100th birthday – having been invented in the 1930s,” Díaz-Giménez says. “It was originally designed for an industrial analogue economy, which is significantly different from the current digital service economy.”
To measure the digital economy, alternative metrics like GDP-B have emerged as a potential addition to traditional GDP. This accounts for free digital goods and digitisation indices, which assess technological adoption and usage. Analysing internet connectivity and domain names, for example, can provide important insights into the digital landscape.
Boards need more macroeconomic expertise
Given the complexities involved in understanding and interpreting the macro environment, Díaz-Giménez believes businesses would benefit from having a macroeconomist on their board of directors. “Each country’s data landscape is so unique,” he says. “Having specific expertise directly involved in strategic discussions at the board level can provide a significant advantage in navigating the complexities of the global economy and understanding the broader context in which their business operates.”
It would help them to anticipate economic shocks or policy changes and understand how global and national economic trends impact their business on a much deeper level, he continues. “This perspective is crucial for strategic decisions related to market expansion, international trade and overall growth in a globalised world.”
While the chief financial officer is largely responsible for providing some of that macro outlook, Díaz-Giménez believes they too would benefit from having a dedicated professional to help digest the macro information, interpret it correctly and most importantly, tailor that intelligence to the specific sector or industry of the company.
Britain’s economic picture remains murky. Confidence among businesses and consumers is falling sharply as the Trump administration’s trade war combines with rising costs to threaten a greater slowdown. With the reliability of the data used to assess these trends under increasingly under scrutiny, business leaders and policymakers are operating in the dark more often than they should be. And that’s perhaps the most concerning development of all.
The Office for National Statistics (ONS) has been underestimating the country’s gender pay gap for the past twenty years, according to new findings released in the British Journal of Industrial Relations.
The report suggests that, since 2004, the UK's statistics agency has failed to properly account for the fact it received more data from larger employers. The gender pay gap is a core measure of economic equality. If policymakers and businesses have been working with an inaccurate picture of earnings in Britain, official pay recommendations and anti-poverty measures, such as the national minimum wages, may be wrong. Employees and advocacy groups who rely on official statistics to push for fair pay may also feel misled.
This is not the first time the number-crunching body has come under scrutiny over errors, and there are growing concerns that some of the statistics it produces are no longer fully reliable.
