Why was US GDP revised up while consumer spending was revised down?

GDP is calculated as the sum of consumption, investment, government spending, and net exports (exports minus imports). When imports are revised downward, the net exports component improves, which mathematically increases GDP — even if nothing else in the economy changed. In the Q1 2026 third estimate, the downward revision to imports (led by consumer goods and capital goods) added enough to GDP to more than offset the 0.9-percentage-point downward revision to consumer spending. The result is a GDP number that looks stronger but reflects weaker domestic demand. This is sometimes called the "import illusion" — GDP rises not because the economy produced more but because it imported less.

What is US consumer spending growth in 2026?

Real consumer spending grew at just 0.5% annualised in Q1 2026, according to the BEA's final estimate. This is the weakest quarterly consumer spending growth since Q4 2022 and one of the lowest readings outside of a recession. Services spending, which had been the economy's growth engine, was revised down from 1.8% to 0.5%. Meanwhile, the personal savings rate fell to 2.6% — its lowest since June 2022 — and credit card debt reached a record $1.33 trillion. Consumer sentiment, as measured by the University of Michigan index, sits at 53.3, which is in territory historically associated with recessions.

Is the US economy in a recession in 2026?

The United States is not in a technical recession — GDP growth has been positive for several consecutive quarters. However, the composition of growth raises concerns. The Q1 2026 GDP figure of 2.1% was driven largely by business investment (particularly AI-related capital expenditure by a small number of technology companies) and by a statistical improvement from lower imports. Consumer spending growth of 0.5% is barely above zero in real terms, and the Conference Board's Leading Economic Index has shown negative 6-month and 12-month rates for an extended period. The economy is growing, but the growth is narrow — concentrated in AI-related investment while the consumer sector weakens.

The Import Illusion: US GDP Revised Up, Consumer Spending Down

Q: What was US GDP growth in Q1 2026?

Real GDP grew at a 2.1% annualised rate in Q1 2026 (January through March), according to the Bureau of Economic Analysis's third and final estimate released on June 25, 2026. This was revised up from 1.6% in the second estimate and from 1.4% in the advance estimate. Haver Analytics described it as an "unusually large upward revision." However, the revision was driven primarily by a downward adjustment to imports — which are subtracted in the GDP formula — rather than by stronger domestic economic activity. Consumer spending, the largest component of GDP, was actually revised downward from 1.4% to 0.5%.

On June 25, the Bureau of Economic Analysis released its third and final estimate of first-quarter 2026 GDP growth: 2.1% annualised. The headline looked reassuring. It was half a percentage point higher than the 1.6% second estimate published in May, and a full 0.7 points above the initially reported 1.4%. Haver Analytics described it as an “unusually large upward revision.” Markets rallied modestly. The narrative settled into a comfortable groove: the US economy is resilient.

But the composition of the revision tells a sharply different story. Consumer spending — which accounts for roughly 70% of US GDP — was revised downwardby 0.9 percentage points, from 1.4% to just 0.5%. Services spending, the growth engine that had sustained the economy through tariff shocks and inflation, fell from 1.8% to 0.5%. The entire upward revision to GDP came from a single line item: imports were revised lower. In GDP accounting, imports are subtracted. When imports fall, GDP rises — even if not a single additional dollar of domestic economic activity occurred.

This is what might be called the import illusion. The economy did not produce more. America bought less from abroad. And the GDP formula rewarded the contraction as if it were expansion.

The Anatomy of a Misleading Revision

The BEA’s own summary was precise: “Real GDP was revised up 0.5 percentage point from the second estimate, primarily reflecting a downward revision to imports, which are a subtraction in the calculation of GDP, that was partly offset by a downward revision to consumer spending.” Within imports, the revision reflected downward adjustments to both goods (led by consumer goods and capital goods, excluding automotive) and services (led by transport services).

A useful way to understand this is through a simplified accounting framework. GDP equals consumption plus investment plus government spending plus exports minus imports. When the BEA discovers that imports were lower than initially estimated, the subtraction becomes smaller, and the total grows — mechanically, automatically, without any change in what American workers produced or American consumers purchased. In this case, the downward import revision added roughly 1.4 percentage points to headline GDP growth. Consumer spending subtracted 0.9 points. The net effect was the 0.5-point headline improvement.

The paradox is that lower imports can reflect weakness, not strength. When consumers cut back — buying fewer imported electronics, clothing, and household goods — imports fall. This makes GDP look better at the precise moment when the demand signal is deteriorating. In Q1 2026, that is exactly what appears to have happened: consumer spending on goods grew at just 0.5%, services at 0.5%, and the import decline was concentrated in consumer goods. The trade arithmetic flatters the headline while the demand data flashes yellow.

Consumer Spending at 0.5%: The Weakest Since the Pandemic

Real consumer spending growth of 0.5% annualised is not just weak; it is historically unusual outside of recessions. For context, consumer spending averaged 2.5–3.0% annualised growth through 2023–2024 as the post-pandemic recovery boosted spending on services, travel, and dining. The 0.5% print in Q1 2026 represents a deceleration of roughly 80% from that pace.

The downward revision was broad-based. Durable goods spending held at 0.5%, roughly unchanged from the second estimate. Nondurable goods were similarly flat. But services spending — the category that includes healthcare, housing, financial services, food services, and recreation — was slashed from 1.8% to 0.5%. This is significant because services account for roughly two-thirds of consumer spending and had been the most reliable growth contributor since 2022. A services slowdown of this magnitude, if sustained, would signal a fundamental shift in the consumer economy.

The adjacent data points reinforce the picture of consumer strain. Credit card debt reached a record $1.33 trillion in February 2026, up from $1.25 trillion the previous quarter. Late-payment delinquency climbed to 4.8% of all household debt. Student loan delinquency hit 9.6% at 90+ days past due. The personal savings rate, which had been falling steadily, dropped to 2.6% in May 2026 — its lowest since June 2022 and well below the pre-pandemic norm of 6–8%.

The University of Michigan consumer sentiment index stood at 53.3 in March 2026, deep in territory historically associated with recessions, despite unemployment holding at a relatively modest 4.3%. More than half of consumers — 53% — now carry credit card balances to cover essential expenses, not discretionary purchases. Retail sales set a nominal record of $752.1 billion in March, but nominal figures disguise the inflation adjustment: in real terms, spending is barely growing.

What Is Actually Driving GDP Growth?

If not the consumer, what is keeping the US economy above water? The answer, as documented in the AI capital expenditure analysis and the semiconductor supercycle report, is business investment — specifically, the AI-driven capital expenditure of a small number of technology companies. Equipment investment grew 15.8% annualised in Q1 2026. Intellectual property products — a category that includes software and R&D — grew 13.8%. These are strong numbers, and they are almost entirely attributable to the five hyperscalers — Amazon, Google, Microsoft, Meta, and Oracle — collectively spending approximately $725 billion on AI data centres, GPU clusters, and supporting infrastructure.

Pantheon Macroeconomics has calculated that without AI-related capital expenditure, US corporate equipment investment would be negative. TECHi, which tracks technology investment contributions to GDP, estimates that AI capex drove 75% of Q1 GDP growth. The Conference Board’s Leading Economic Index, a composite of 10 forward-looking indicators, stood at 99.3 in May with both its 6-month and 12-month rates still negative — not a profile consistent with an economy growing at 2.1%.

Government spending contributed positively, driven by defence outlays and state-level expenditure. Exports added modestly, partly reflecting the front-loading of shipments ahead of tariff deadlines. But the composition narrative is clear: strip out AI investment and government spending, and the private-sector domestic economy is barely expanding.

The Tariff Dimension: Did Trade Policy Distort the Data?

One factor that complicates the Q1 data is the tariff regime. The Supreme Court struck down IEEPA-based tariffs on February 20, 2026. The administration immediately imposed Section 122 tariffs (10% global), which were in turn struck down by the Court of International Trade on May 7 but stayed by the Federal Circuit pending appeal. Throughout Q1, businesses were facing an extraordinary level of trade policy uncertainty — tariff rates changed multiple times, supply chains were being restructured in real time, and front-loading of imports (buying ahead of expected tariff increases) may have shifted trade flows between quarters.

This matters for the import revision. If businesses front-loaded imports in Q4 2025 to beat anticipated tariff increases, Q1 imports would have been naturally lower — not because demand weakened, but because inventory was already built. The BEA does not adjust for this directly. The result would be a GDP number inflated by tariff-related inventory timing rather than genuine economic activity.

With the Section 122 tariffs set to expire on July 24 and no Congressional action to extend them, the trade policy landscape remains unsettled. If the 10% global tariff lapses, the effective tariff rate — currently estimated at 7–11.8% by Penn Wharton, the highest since the 1940s — would fall to approximately 8.2%. This uncertainty alone is enough to distort business purchasing decisions, making quarter-to-quarter GDP figures less reliable as a gauge of underlying momentum.

Three Numbers That Tell the Real Story

For anyone trying to understand the health of the US economy in mid-2026, three numbers matter more than the 2.1% GDP headline.

0.5%: Real consumer spending growth, annualised. This is the figure that captures what 330 million Americans are actually doing with their money. At 0.5%, the consumer economy is barely growing in real terms. Given that consumption is 70% of GDP, this level of spending growth, if sustained, implies trend GDP growth of well under 1% before investment and government spending are added. The K-shaped divide — the top 10% of households spending freely on asset wealth while the bottom half depletes savings and runs up credit card debt — explains how retail sales can set nominal records while real spending barely moves.

$1.33 trillion:Total US credit card debt. This is not a flow variable; it is a stock that accumulates. Households are not just spending less; they are financing their reduced spending with debt at average interest rates above 22%. The Federal Reserve’s consumer credit data shows revolving credit growing at 6–8% annualised — roughly three times the rate of real income growth. This is the classic late-cycle pattern: consumers maintain spending levels by drawing down savings and borrowing, until a shock — a job loss, an interest rate increase, a medical bill — triggers a retrenchment.

53.3:The University of Michigan Consumer Sentiment Index. This reading is lower than it was during the 2022 inflation peak, lower than any point during the COVID recovery, and comparable to levels last seen during the 2008–2009 recession. Consumer sentiment has historically been a lagging indicator — people report feeling bad about the economy after things have already deteriorated — but a reading this low, sustained for this long, has never occurred alongside GDP growth above 2%.

What Comes Next

The Q1 GDP data is now final and will not be revised again until the annual benchmark revision in 2027. Q2 data — covering April through June — will be released in late July. The Atlanta Fed’s GDPNow tracker currently estimates Q2 growth at approximately 2.5%, but this model has been consistently overstating growth in 2026 relative to final estimates, partly because it struggles with the AI capex concentration that distorts the investment data.

The Conference Board projects full-year 2026 GDP growth at 1.8%, down from 2.1% in 2025. Guggenheim Investments forecasts growth “around 2 percent,” noting that the economy is “underpinned by robust artificial intelligence capital expenditures and near-term fiscal tailwinds” but facing headwinds from “elevated gas prices and cooling incomes.” The Federal Reserve continues to hold the federal funds rate at 3.50–3.75%, with markets pricing no cuts before mid-2027.

The structural question is whether AI capital expenditure can continue to carry an economy whose consumers are weakening. At $725 billion in planned hyperscaler capex, the scale is unprecedented — Apollo estimates it at roughly 2% of GDP — but it is also concentrated in five companies. If any of the five major spenders pulls back, or if AI revenue growth disappoints the models justifying the investment, the US economy would face a Q1 composition in reverse: strong consumer spending nowhere to be found, and the investment engine slowing simultaneously.

The 2.1% headline is real, in the narrow sense that the arithmetic is correct. But GDP is a formula, and formulas can be technically right while telling a misleading story. In Q1 2026, the formula says the world’s largest economy grew at a healthy pace. The composition says something different: consumers are weakening, investment is concentrated in a single sector, and the headline was rescued by an import decline that may itself reflect falling demand. For the global GDP rankings, the US position is secure. For the question that matters more — whether American households are better off — the revision tells a less comfortable truth.

Data sources: Bureau of Economic Analysis (BEA) GDP third estimate, Q1 2026 (June 25, 2026); Haver Analytics; Federal Reserve Consumer Credit (G.19); Conference Board Leading Economic Index (May 2026); University of Michigan Consumer Sentiment Index (March 2026); Guggenheim Investments June 2026 Economic Outlook; Penn Wharton Budget Model (June 16, 2026); Pantheon Macroeconomics; TECHi; Apollo Global Management; Morgan Stanley; EY US GDP analysis. Country economic data from Statistics of the World, sourced from IMF WEO April 2026 and World Bank.

The Import Illusion: US GDP Was Revised Up to 2.1% But Consumer Spending Collapsed to 0.5%