Unlock the Editor’s Digest for free
Roula Khalaf, Editor of the FT, selects her favourite stories in this weekly newsletter.
Anyone who has taken “Econ 101” will know that most economic models assume decision makers have perfect information. In reality, that is rarely the case. Data can be incomplete, one party may have more detail than the other, or everyone might be equally misinformed. Indeed, official national statistics — which are vital for policymakers, businesses and investors — can sometimes miss the mark. Right now, many countries are having a particularly hard time counting the number of people in work.
The Resolution Foundation think-tank recently claimed that Britain’s Office for National Statistics may have lost almost a million workers from its job figures since 2019. It reckons the statistics body could have wildly overestimated the extent of worker inactivity. If true, it would undermine the narrative of “missing workers” that has been the basis of several policy pledges from the UK government as well as interest rate decisions by its central bank.
There is a puzzle in America too. Job creation has flatlined over the past two years, based on its household survey. Non-farm payrolls, compiled from a survey of businesses, however, show relentless jobs growth. The data has been volatile too, and subject to hefty revisions. In August the Department of Labor said that the US economy created over 800,000 fewer NFP jobs than initially reported in the year to March.
One reason for the uncertainty may be that over the past decade, response rates to data questionnaires across the UK, US and EU have been trending downward, exacerbated by the pandemic. Lockdowns have also messed with projections of population, immigration, and business births and deaths, which help statisticians aggregate employment data up from survey samples. This has led to biases, bad estimates and the conflicting stories from different data sources.
Counting workers is also a problem in the developing world, albeit for more systemic reasons. Estimating India’s unemployment has long been a challenge given that a significant portion work in the informal sector. In China, opacity is another limitation.
Bad jobs data leads to bad decisions. Employment numbers underpin tax, spend and welfare choices, and are central to monetary policymakers’ assessments of how hot the economy is. Businesses use it in hiring and salary decisions. Investors lean on it too. America’s NFP numbers drive global financial markets’ pricing of interest rates.
What can be done? Governments should ensure funding keeps pace with the demand for greater, timelier and more precise data. In June, America’s Bureau of Labor Statistics said budget cuts may mean it has to cut the sample size of its household survey. Data wonks are also getting snapped up by higher-paying tech firms. The authorities should do a better job of holding statisticians to account, too. The ONS has been particularly sluggish in acting on falling responses, and moving to online-first surveys.
Even with incentives or better survey design, declining response rates may be hard to reverse. Some studies suggest people are fatigued by too many questionnaires. Either way, capturing labour market data through other sources will be important. Statistical bodies should partner more with the private sector — including jobs boards, such as Indeed and LinkedIn — for real-time statistics to support their estimates. Governments also need to share more timely administrative numbers with data agencies. In some countries, national identity cards have helped agencies get a better grasp of population and workforce data.
National statistics bodies need to do better at counting jobs numbers. More resources, effective oversight and wider access to other data sources will not make the numbers perfect. But at least it will give a greater sense of how imperfect they really are.
