Sunday, May 21, 2023
How the pandemic dealt another blow to productivity
Posted by David Smith at 09:00 AM
Category: David Smith's other articles

My regular column is available to subscribers on This is an excerpt. Not to be reproduced without permission.

Much as I hate being the bearer of bad news on a spring morning, a few days ago we had some grim figures for productivity, the ultimate driver of living standards. Productivity, measured by gross domestic product per hour worked, dropped by 1.4 per cent in the first three months of the year and was down by 0.6 per cent on a year earlier, its biggest annual fall in “normal” times for 10 years.

The other main productivity measure, output per worker, dropped by 0.4 per cent in the first quarter, and was down by 0.9 per cent on a year earlier. Both measures of productivity show no growth compared with the fourth quarter of 2019, before the pandemic struck. A very weak trend for productivity growth before the pandemic – evident since the 2008-9 financial crisis – appears to have turned into a non-existent one.

This is interesting, because it provides a first test of the claims and counter claims that were buzzing around on productivity during the pandemic. Did cramming 10 years of technological change into a very short period, providing a big increase in, for example, working from home, boost productivity? We may now be closer to knowing the answer.

Before I come on to that, let me add that if I chose, I could be the bearer of even worse news. The productivity calculations are based on employment and hours worked figures derived from the Office for National Statistics’ Labour Force Survey (LFS), which show no change in the number of people in work compared with the end of 2019.

Some economists are, however, concerned about the accuracy of the LFS figures. Another measure of employment produced by the ONS, workforce in employment, shows a stronger picture for employment, up 2 per cent since before the pandemic, and thus an even weaker picture for output per worker. PAYE data from HMRC also suggest more buoyant employment than the LFS figures. More people in work but with no increase in GDP equals lower productivity.

The debate over the statistics cannot disguise the fact that over the past three years a bleak UK productivity performance has turned even bleaker. This is very early evidence, but what can we conclude about the impact of the pandemic? A couple of things, I think.

Working from home (WFH), which has now evolved into hybrid working for many people – part at home, part at the place of work – has been a bone of contention between employers and employees, and that continues. Many employers believe that, as well as losing the benefits of in-person collaboration between workers, a key part of the creative process, unsupervised workers are less productive, so WFH reduces productivity. This is why there is what looks like a concerted drive at present to get people back into the office.

Surveys of employees, in contrast, tend to find that, as well as preferring a combination of face-to-face and remote working, most think that their productivity has gone up because a result of WFH. There are fewer distractions, and less time is wasted.

These two views are hard to reconcile, though you would have to say that the evidence so far, while preliminary, tends to favour the view of the bosses rather than the workers. If there was a productivity gain from WFH you would expect to see some early signs of it now.

One possible explanation comes from research on time savings due to WFH by Professor Nick Bloom and others, published by America’s National Bureau of Economic Research. When you cut out commuting, people save time, estimated at an average of two hours a week in 2021 and 2022 (averaged across the whole working population).

Some 40 per cent of those time savings are devoted to work, the rest of leisure and caring responsibilities. But even that 40 per cent means that when working at home, people are putting in more hours to accomplish the same tasks, thus reducing productivity. There have also been plenty of stories of Zoom fatigue and people working well beyond normal hours, and being expected to do so, when at home.

The other negative productivity effect from the pandemic could have come from changes in the workforce. The loss of hundreds of thousands of older and more experienced workers, some to early retirement, most to long-term sickness, may also have had an impact, which is hard to measure but appears to be real.

All this comes as we await what some see as the arrival of the productivity cavalry in the form of artificial intelligence (AI). Commentary on AI tends to fluctuate between “it will take all our jobs”, or it will result in machines that are dangerously powerful in comparison with flawed humans, to the potential economic benefits.
Goldman Sachs recently predicted that generative AI could boost global GDP by 7 per cent and productivity by 1.5 per cent over a 10-year period. Generative AI describes developments such as the now ubiquitous ChatGPT, which can create new content.

A cumulative boost to productivity of 1.5 per cent over 10 years is not that much. When law firms use ChatGPT to draft client letters, investment banks for basic company research, or accountants for run of the mill audit work, capturing the productivity gains is quite hard. Only if the machines replaced the people carrying out these tasks on a large scale, or if AI freed them to do more productive work, would you expect to see a productivity effect, and neither yet seems to be happening on any scale.

There probably will be a productivity boost from AI, however A recent report from Washington’s Brookings Institution, ‘Machines of mind: The case from an AI-powered productivity boom’, suggested that the effects could be very significant, though it noted that some of the impact could be in what it described as “silent productivity”. So, for example, if AI improves the quality of reports produced by professional services firms, or even articles like mine, so far merely human, it would be hard for that to be picked up in the productivity statistics.

But, and this cannot be emphasised enough, this cannot be a case of good things coming to those who wait. Waiting for AI to lift the UK’s dismal productivity performance can never be a strategy. We need more business investment, instead of which the pandemic compounded the weakness since the EU referendum, better skills, more innovation and better infrastructure. We also need to make sure that we do not fall behind while other countries are reaping the benefit of AI and other technologies.