The American labour market is frozen. That is a kind of fragility.

AMERICA'S employment data has a disconcerting habit of looking better than it feels. The June jobs report, released on July 2nd, is a case in point. Nonfarm payrolls rose by 57,000, well below the 115,000 economists had expected and less than half the downwardly revised 129,000 added in May. Yet the unemployment rate fell to 4.2%. The headline suggests something has gone right. The mechanism behind it is murkier.

The lower unemployment rate did not come from more hiring. It came from fewer people looking. The labour-force participation rate fell by 0.3 percentage points to 61.5%, the lowest since March 2021. The household survey showed 507,000 fewer people at work, according to the Bureau of Labour Statistics. To be sure, the household survey is noisier than the establishment data and prone to large revisions. But the direction of travel has been consistent for months, and the establishment revisions point the same way: April and May were trimmed by 74,000 jobs in total.

The real question beneath the monthly headline is not whether the unemployment rate is 4.2% or 4.3%. It is what holds the labour market together. The answer, increasingly, is inertia. Hires remain depressed. Layoffs remain depressed. Workers who have jobs are not being recruited aggressively by competitors; workers who do not have jobs are not finding new roles quickly; and neither side is moving fast enough to set off a sharp deterioration or a meaningful recovery. Hiring Lab, a labour research firm, put it plainly: the market is frozen between low hiring and low firing, and the calm on the surface reflects stillness underneath rather than genuine momentum.

Three facts make the stillness more disquieting than the headline implies.

First, the long-term unemployed are accumulating. The share of unemployed workers who have been out of work for 27 weeks or more rose to 27.5% in May, up from 20.4% a year earlier and well above pre-pandemic norms. A month's unemployment is a setback; six months is a structural wound. The low-hire environment means that when someone does lose their job, the pool of openings waiting for them is narrow.

Second, job openings have collapsed from their post-pandemic peak without a corresponding surge in matches. The JOLTS data for June, released on August 4th, showed 7.4 million openings, essentially unchanged from the prior month. Hires were flat at 5.3 million; separations at 5.4 million. Quits - workers voluntarily leaving their jobs - held at 3.2 million. The labour market is neither heating up nor freezing over. It is idling.

Third, sectoral divergence tells a story about where demand actually is. Professional and business services, social assistance and health care together added 83,000 jobs, but losses in other sectors trimmed the net gain to 57,000. Leisure and hospitality lost 61,000 jobs, reflecting what the BLS called weaker-than-usual seasonal hiring. Manufacturing, construction, wholesale trade and retail were effectively flat. The economy is growing through government-adjacent services and demographic headwinds, not through enterprise or consumer-facing industries.

It is tempting to read the soft data as a sign that the Federal Reserve has the inflation problem under control. And on that narrow point, the evidence helps the central bank. Average hourly earnings rose 0.3% in June and 3.5% over the year, in line with forecasts. There is no wage-push inflation in these numbers. The Fed, under Chairman Kevin Warsh, can claim that rate hikes are unnecessary.

But the deeper problem is not inflation. It is the fragility of a labour market held together by low churn. A low-hire, low-fire equilibrium is stable only as long as nothing pushes on it. Should demand soften further - from tariffs, from geopolitical shocks, or from a delayed reaction to tighter financial conditions - the lack of hiring leaves no cushion to reabsorb workers who lose their jobs. The market could tip into rising unemployment without the kind of dramatic headline that usually signals trouble, because the trouble would already be embedded in the participation rate and the long-term unemployed.

The politics of the labour data complicate the policy picture. A falling participation rate can read as discouragement, which is bad news, or as people leaving the workforce by choice - retiring early, caring for family, opting out of a difficult job search - which is at least ambiguous. But the arithmetic is unforgiving: fewer workers mean less productive capacity, which means less growth even if unemployment looks manageable. An economy that achieves a steady unemployment rate by shrinking its workforce has solved the headline but not the underlying constraint.

What should policymakers do? The first task is not to panic about the latest monthly number. The second is to recognise that the labour market's current stability is of a different character than it was in 2023 or early 2024, when high churn and high hiring created a genuinely dynamic economy. The tools that would restore dynamism - lower barriers to business formation, faster permitting, a tax system that does not penalise investment, and an immigration policy that does not throttle labour supply - are all politically difficult. That is precisely why they ought to be pursued. Tariffs and industrial policy may look like answers, but they create constituencies for permanent inefficiency rather than restoring the incentives that generate jobs in the first place.

But the structural picture does not depend on a single month. The American labour market is not breaking. It is stagnating. Those are very different things, and they call for very different remedies.

Better to diagnose stillness as a problem before it becomes a crisis.