US - ICE melts GDP growth

The decrease in labour supply growth dampens the impact of the decrease in labour demand growth. Annual growth in non-farm payrolls has steadily cooled from an average rate of over 4% in 2022 to slightly above 1.1% as of June 2025, and a further cooling of the US economy is likely to continue to provide a further drag on job growth. We already saw a decrease in immigration, in conjunction with a cooling labour market last year, but the decrease in net immigration is accelerated by the Trump administration. From a strict objective full employment view, the timing is the most benign it could be. A few years ago, this immigration crackdown would have made the historically tight labour market even worse, likely leading to higher wages, elongating the inflation surge. Now it comes at a time of weakening labour demand, which, without the supply might have pushed the unemployment rate up more, while now it keeps a lid on it. The June labour market report is a perfect illustration; despite at best moderate job growth, the unemployment rate dropped against all expectations, mostly driven by a labour force that counts half a million fewer workers compared to the start of the year. Miran wasn’t wrong; native-born workers increased by 300k, but foreign workers dropped by 800k.

This lack of immigration hurts GDP growth. As the left-hand chart below shows, there’s a clear positive relationship between economic growth and labour force growth. As the trend line shows, even without labour force growth, you still get GDP growth through increases in productivity, i.e. output per worker. The chart on the right shows that the absolute and relative contributions of these components have changed over time. Over the four decades starting in 1960, half of GDP growth came from labour force growth, and half from productivity gains. As labour force growth started to weaken, partly through birth rates, the period from 2000-2014 saw lower growth, despite productivity gains still averaging 2% per year on the back of the internet revolution. Since then, we’ve seen productivity growth slowing. Comparing the pre- and post-pandemic periods, growth increasingly came from labour force growth, and predominantly immigration. Looking at 2025 year-to-date, the native-born labour force continues to expand, but overall, labour force growth has stalled (see also the labour supply section here). Native-born labour force growth has averaged about 0.3% per year since the 2007 when the decomposition became available. In order to maintain the Trump administration’s 3% real GDP growth target, we’d need consistent productivity gains of 2.7% per year. While no weight should be attached to the year-to-date productivity decline (it’s based on a single quarter of data and is generally noisy), not even the advent of AI is likely to push productivity to those levels.