The Recession That Hasn’t Arrived

America’s economy is still growing. Increasingly, that tells us less about how Americans are actually doing.

Something strange is happening in the American economy.

Job creation has slowed sharply. Labour-force participation is falling. Inflation remains above target. New delinquencies on credit cards and auto loans remain elevated. For anyone trying to find a better job, buy a home or live principally from wages, much of the economy feels considerably weaker than the headline numbers suggest.

And yet America is not in recession.

Real GDP grew at an annualised rate of 1.5 per cent in the second quarter of 2026, following 2.1 per cent in the first. Consumer spending and investment both contributed to that growth. More strikingly, real final sales to private domestic purchasers — essentially consumer spending plus private fixed investment — grew at an annualised rate of 3.9 per cent.

That is the paradox.

The mistake is to assume the two observations contradict one another.

They don’t.

GDP measures the value of economic activity occurring inside a country. It does not measure how evenly the benefits of that activity are distributed, whether housing remains attainable, how secure employment feels, or whether the median household believes it is getting ahead.

For long periods, those things tend to move together. When an economy grows, employment expands, wages rise, businesses invest and households become wealthier.

But they do not have to.

And increasingly, in America, they aren’t.

The first reason is that the American economy is operating with an extraordinary amount of fiscal support.

Washington is projected to run a deficit of $1.9 trillion this financial year, equivalent to 5.8 per cent of GDP. Over the past 50 years, the average deficit has been 3.8 per cent.

A deficit approaching six per cent of GDP is unusual for an economy that is not officially in recession.

This does not mean government spending is currently driving GDP growth. In the second quarter, government spending actually fell and detracted from growth.

The point is structural.

The federal government is borrowing on a scale that provides substantial continuing fiscal support to aggregate demand.

Whatever one thinks about the merits of individual programs, the scale matters.

America is running a deficit far above its historical average while the economy is still expanding.

The second force is more concentrated.

America is in the middle of an extraordinary capital-investment boom.

AI requires chips. Chips require servers. Servers require data centres. Data centres require land, cooling systems, transmission infrastructure and enormous quantities of electricity.

The scale is now difficult to overstate.

Alphabet now expects capital expenditure of between $195 billion and $205 billion this year. In the second quarter alone, it spent $44.9 billion, with the vast majority going into technical infrastructure supporting AI. Around 60 per cent of that infrastructure spending went into servers and 40 per cent into data centres and networking equipment.

That is one company.

Across the technology sector, investment on this scale is creating a genuine industrial buildout reaching far beyond Silicon Valley.

And all of it counts as economic activity.

But a billion-dollar data centre is not a billion dollars distributed evenly through the American economy.

It is concentrated capital expenditure.

It can produce enormous amounts of investment without producing anything resembling a nationwide employment boom.

That distinction becomes important when we look at the labour market.

Total nonfarm employment fell by 23,000 in July. The Bureau of Labor Statistics described both payroll employment and the unemployment rate as having changed little, which is statistically fair. But the trend beneath the monthly number is difficult to dismiss.

The economy has added an average of only 34,000 jobs per month over the past year. Labour-force participation stands at 61.4 per cent and has fallen 0.7 percentage point since January. The employment-to-population ratio has fallen by half a percentage point over the same period.

Financial activities employment is down 121,000 from its May 2025 peak. Healthcare, meanwhile, continues to add jobs.

This is not a labour market in collapse.

But neither is it one behaving like an economy generating broad-based prosperity.

And that is precisely what makes the current moment unusual.

Private demand remains resilient even as employment creation weakens.

The third reason is found in the structure of American household wealth.

The United States has accumulated enormous household wealth, but it is distributed highly unevenly. Federal Reserve distributional accounts show large concentrations of net worth among households near the top of the wealth distribution. Overall household and nonprofit net worth remained about $183 trillion in the first quarter of 2026, and the ratio of wealth to disposable personal income remained well above its historical average.

That matters because wealthy households possess something wage-dependent households do not: large stocks of assets from which consumption can be financed.

They own equities.

They own property.

Many locked in mortgage debt when rates were considerably lower.

A household whose investment portfolio has appreciated substantially operates under very different constraints from one whose economic position depends almost entirely on its next pay cheque.

This allows aggregate consumption to remain resilient even when economic pressure is distributed unevenly.

Real consumer spending rose another 0.4 per cent in June.

That spending is real.

So is the pressure elsewhere.

American household debt remained near a record $18.8 trillion in the second quarter. Overall debt fell slightly during the quarter, but credit-card balances rose by $21 billion to $1.26 trillion and auto-loan balances rose by $28 billion to $1.71 trillion.

These figures do not describe a consumer economy that has stopped functioning.

They describe something more complicated.

One household watches its investment portfolio rise and books another holiday.

Another spends an increasing share of its income servicing housing, food, insurance and debt.

GDP records both transactions.

It does not adjudicate between them.

This is why the American economy can simultaneously produce resilient consumption numbers and widespread economic unease without either side necessarily misunderstanding what is happening.

They are measuring different things.

The weakness is real.

So is the growth.

The mistake is assuming that growth means the weakness cannot be widespread.

And this is where the usual recession debate begins to become less useful.

None of this means America is secretly in recession.

It isn’t.

The economy is growing. Private domestic demand, on the latest numbers, is growing quite strongly.

That distinction matters.

But it raises a more interesting question than whether the National Bureau of Economic Research will eventually put a date on the beginning of the next downturn.

It asks what economic growth now tells us about the condition of the people living inside the economy.

The modern American economy is increasingly capable of producing enormous amounts of economic activity through institutions and concentrations of capital that touch ordinary economic life only indirectly.

The federal government can borrow trillions.

A handful of technology companies can invest hundreds of billions.

Households sitting on enormous asset wealth can continue consuming.

GDP rises.

Meanwhile, hiring can stall, labour-force participation can fall, housing can remain difficult to attain and millions of households can experience declining economic security.

Nothing about this makes the GDP figures false.

The more interesting possibility is that the relationship between national output and ordinary prosperity is becoming weaker.

Our conventional idea of recession emerged from an economy in which national output, employment and household prosperity tended to move together more closely.

The 21st-century economy is making that relationship less reliable.

Deficit-financed spending can sustain demand without creating broad private-sector prosperity.

Asset wealth can sustain consumption even when wage-dependent households are under pressure.

Hyperscale investment can generate enormous economic output without requiring enormous workforces.

None of that activity is fictitious.

But neither is the deterioration occurring outside it.

America has not entered a recession.

The more uncomfortable possibility is that a modern economy can sustain headline growth long after a substantial part of its population has entered something functionally indistinguishable from one.

The recession hasn’t arrived.

Increasingly, the question is whether we would recognise it when it does.