The great productivity boom is about to take a break

THE great productivity boom of the 2020s isn’t over, but it’s having an interlude. And it could become a long one if US President Donald Trump remains committed to protectionist policies that make the economy less efficient. Ultimately, artificial intelligence will help, but it may take a while.

After an impressive streak in the past two years, labour productivity – or non-farm employee output per hour – declined at a 0.8 per cent annualised rate in the first quarter, the first such decline since 2022. For all the caveats about Thursday (May 8)’s statistic (the quarter-to-quarter data is notoriously noisy), it’s sure to burst the bubbles of fellow productivity bulls, who hoped it could sustain economic growth even as reduced immigration slows the expansion of the labour force.

Productivity growth is the magic pixie dust behind great economies. It can produce economic booms while also keeping a lid on inflation, lifting the fortunes of workers and business owners alike. For much of the post-financial crisis period, labour productivity grew at a frustratingly lacklustre pace of around 1.5 per cent annually, but something extraordinary happened in recent years. Four of the past five years saw labour productivity growth above that level, including the fantastic 2.8 per cent increase of 2024. And the boost helped the economy to stay afloat even as the Federal Reserve raised policy rates to rein in inflation.

Don’t let guard down

That progress has not been derailed by a single negative quarter, but we shouldn’t let down our guard either. Trump’s trade war risks snarling supply chains and delaying corporate investments in advanced technologies that allow workers to do more with less. We have some evidence of this from the high-tariff period under president William Mckinley, who Trump likes to cite as inspiration for his policies. Research from Alexander Klein and Christopher Meissner suggests that Gilded Age tariffs lowered labour productivity by attracting smaller, less productive firms into various businesses. Tariffs purport to protect domestic firms from outside competition. Even when the strategy works as intended, it reduces the incentive for firms to stay at the vanguard. We see this at work in the auto industry, where a mix of protectionism and an aversion to electric vehicles threatens to consign Detroit’s big carmakers to outright irrelevance.

Theories abound about what drove the productivity boom of the past several years. Economists including Ryan Decker and John Haltiwanger have found that the pandemic produced a surge in new businesses concentrated in higher productivity sectors. Relatedly, some people think that the Covid-19 shock forced firms to do more with less and that they carried those lessons forward. As Goldman Sachs Group Inc has shown, the industries that cut jobs in 2020 were also among the ones that produced the greatest pickup in productivity in subsequent years. A third intriguing theory – supported by the research of Nick Bloom – holds that the switch to work-from-home may have helped productivity growth, though some academics and top executives have argued the opposite. To the extent that any of these theories hold true, they were likely to have been one-time boosts to productivity rather than persistent sources of growth.

The most enticing narrative revolves around artificial intelligence, which economic and financial commentators started breathlessly writing about after ChatGPT was released in November 2022. In a speech in February, Federal Reserve Bank of Chicago President Austan Goolsbee noted that many of the industries contributing the most to productivity growth in recent years were “tech or AI intensive”, and he surmised that the gains from AI might already be in the data (that thesis has its detractors, too).

Whether or not the AI boom is upon us, it’s clear is that the technology seems to have the potential to usher in a period of significant and sustained productivity growth. Great innovations including electricity and computers provided powerful boosts to the economy, even if the impact wasn’t instantaneous. For instance, Hewlett Packard released its desktop computer in 1968 for around US$44,000 in today’s dollars. It took over a decade for the technology to become affordable and widespread enough to shape the macroeconomic data through the 1980s and 1990s. Goldman estimates that the AI productivity effects might start to be felt around 2027, with the strongest impacts in the 2030s.

Stay optimistic

All in all, that leaves plenty of reason to stay optimistic about the longer term, though the next few years could prove more challenging. Whatever caused the post-pandemic productivity boom, it’s likely that the positive effects are now fading. In their place is an unpredictable trade war that stands to put productivity gains in reverse, which will make it feel like the AI revolution can’t come soon enough. BLOOMBERG