Welcome back. What is a recession? The popular, yet overly simplistic, shorthand is two consecutive quarters of negative economic growth. The US, however, has a more refined approach, determined by technocrats at the National Bureau of Economic Research.
“A recession involves a significant decline in economic activity that is spread across the economy and lasts more than a few months,” says the private non-profit. Annualised US GDP grew in the second quarter of this year, following a first-quarter contraction, meaning the country avoided a recession by the basic definition. The same appears true according to the NBER’s broader standard.
However it is defined, in this edition, I outline why the recession label isn’t particularly meaningful — especially when applied to the world’s largest economy. The main six indicators used by the NBER to determine the onset of a recession are shown in the chart below.
All six indicators were either in or close to contraction in May on a month-on-month basis. In recent months, all have remained weak, but not perhaps bad enough to clear the NBER’s criteria.
However, Pascal Michaillat, professor of economics at the University of California, Santa Cruz, cites two main shortcomings of relying on the research organisation’s approach. “It places too little emphasis on unemployment and vacancy rates,” he says. “Also, as it waits for data and revisions, recessions are often declared several months after they have begun.” Using a real-time recession detection algorithm trained on a century of broader labour market data, Michaillat — who is also a research associate at the NBER — estimates a 71 per cent probability that the US economy was already in recession in May.
“While the other NBER data might seem useful, the labour market reflects a more fundamental reality. Falling vacancies and rising unemployment are, in my view, a more reliable indicator of widespread economic distress,” he says. Indeed, in July the number of job seekers exceeded the number of openings for the first time since 2021.
Many Americans would agree with Michaillat’s assessment. In an early August poll by the Economist and YouGov, almost half of respondents said the US economy was “getting worse”. Close to one-third thought the US was already in recession; 28 per cent weren’t sure.
Clearly the NBER’s task is difficult and subjective, not to mention fraught with data complications. (I wrote about America’s dodgy jobs numbers for FT Alphaville last year.) Nor is the recession nomenclature that useful, anyway. In an economy of the US’s size — which is greater than the other G7 nations combined, on a purchasing power parity basis — many states, sectors and households can be facing a downturn, while expansions elsewhere raise activity in aggregate.
A more informative question to ask, then, is what is keeping the US afloat despite weakness across several national-level indicators?
In US President Donald Trump’s second-term, economic resilience has come largely from four narrow sources: a handful of states, artificial intelligence, healthcare and the wealthy. Using a similar methodology to the NBER, Mark Zandi, chief economist at Moody’s Analytics, finds that US states making up close to one-third of the country’s GDP are either in or at high risk of recession.
“States experiencing recessions are spread across the country,” he says. This includes the midwest and parts of the rustbelt. “California, Texas and New York, which together account for close to a third of US GDP, are holding their own, and their stability is crucial for the national economy to avoid a downturn.”
The geographical distribution of industries is a key determinant of how US regions are performing.
Zandi’s calculations find that the agriculture, manufacturing and construction sectors are in recession, which is hitting rural and industrial states. These are goods-producing sectors most likely to be affected by Trump’s tariffs and trade partners’ retaliatory measures. A downturn in the federal government sector, linked to the administration’s cuts, has also pushed Washington DC into a slowdown “Healthcare, technology and real estate are expanding, while financial services, retail and hospitality are treading water,” he adds.
Relatedly, the AI boom has provided an uplift to particular states and industries. “AI is primarily a bigger-city office technology. Most of its superstar and star hubs are in California’s coastal cities, Texas’ business cities and the Boston to Washington corridor,” says Mark Muro, senior fellow at the Brookings Institution. “Data centre construction is also booming across the US, but this leads more to a short-term surge in building activity.” Pantheon Macroeconomics estimates that US GDP would have grown at a mere 0.6 per cent annualised rate in the first half were it not for AI-related spending, or half the actual rate. Beyond the surge in AI-related spending, private investment elsewhere in the US economy has started to shrink this year under the weight of higher interest rates and uncertainty. Total private fixed investment rose by about 3 per cent year on year in the second quarter, but it would have fallen by around 1.5 per cent if AI-related components were excluded, says Pantheon in a recent research note.
The annual rate of data centre construction has grown over Trump’s second term — but residential, manufacturing and other commercial building work have declined.
As Zandi’s research suggests, the healthcare industry is also driving the economy. Friday’s non-farm payrolls report, showing a small gain of 22,000 jobs in August, would have been negative were it not for the sector. So far in Trump’s second term, 598,000 jobs have been created. A staggering 515,000 — or 86 per cent of them — are healthcare and social assistance jobs.
Since April, the Bureau of Labor Statistics’ payroll diffusion index for the private sector fell below 50, signifying that more sectors are shedding jobs than gaining them. That is a rarity outside of a recession.
Healthcare jobs are less a reflection of a booming industry and more one of sickness, as I outlined in the May 18 edition of this newsletter. Indeed, household expenditure on healthcare — which accounted for 12 per cent of US GDP last year — prevented the annual growth rate of real consumer spending from going negative in the first quarter.
More broadly, real consumer spending — which comprises around two-thirds of US annual GDP — has slowed into this year.
Other than healthcare expenditure, Morning Consult’s new Consumer Health Index, which combines labour market dynamics with consumer sentiment to provide a gauge on near-term expenditure, suggests that high-income households have driven spending, while demand from lower-earning cohorts has tailed off.
Part of this reflects lower-earning households’ greater exposure to higher interests rates and slower-growing sectors. US stock market gains — linked to the AI boom — have also largely accrued to richer individuals.
“Wealth effects will go from being a drag on consumption in the wake of the ‘liberation day’ tariff announcement, to a boost, as the S&P 500 is near all-time highs,” says Bernard Yaros, lead US economist at Oxford Economics. “High-income, wealthy households will ride the coattails of this positive financial wealth effect.”
So, whether or not the NBER eventually declares a US recession is, in the end, a technicality. For now, a handful of narrow growth engines are propping up an otherwise broadly slowing economy.
This highlights America’s economic diversity and collective resilience, but it also shows that the absence of an official nationwide recession verdict does not spare large parts of the country from recession-like conditions.
In a system as vast, complex and uneven as the US economy, a binary judgment obscures more than it reveals.
infinite growth engine achieved
