Why Might Productivity Be Surging?
Cost Pressures Make Firms Do More With Less
There are some tentative signs of a rapid increase in productivity in both the US and the UK. This is easy to see in the US figures. Even amid all the tariff turbulence, output per hour has climbed higher.
The productivity growth acceleration in the UK is less obvious, and its confirmation would be a lot bigger news given the endemic low British productivity in the post-GFC period. Official statistics suggest that GDP per head has largely gone sideways over recent years; however, a new report from the Resolution Foundation claims that errors in official figures cloud a major uptick.
The official productivity figures are premised on the ONS Labour Force Survey, which has had some major issues with reliability in the post-pandemic period. LFS data have not recorded the contractions in employment in the last 18 months as appear in the ONS’s two other main employment data sources (WFJ and HMRC PAYE RTI data), which many find suspicious. If we use alternative data to calculate productivity, i.e. if we divide GDP by the lower number of employees as implied in the HMRC PAYE payrolls data (assuming stable average hours per worker), we get a larger number for output per hour, representing a very sharp rise in productivity recently. This adjustment is represented in the red line below.
At this point, we can conclude that AI is driving a productivity boom and get on with our day. Right? Sorry, wrong.
It might be easy to conclude that AI is causing a nascent surge in productivity growth, and (as we have highlighted before) general purpose technologies such as AI are likely the main source of productivity growth. However, the invention of AI in itself does not instantly make workers more productive.
