Every headline number says the same thing. America has a worse inflation problem than almost anywhere else in the developed world. Goldman Sachs just published research arguing that conclusion may be backward once you actually open up what is inside the data.

The bank’s economists did not dispute the top-line numbers. They dug underneath them, and what they found flips the usual story about which economies are actually struggling to get prices under control.

Goldman’s surprising read on the global inflation gap

Goldman economist Megan Peters laid out the puzzle. Core inflation sits close to target in most countries but remains notably higher in the United States, especially in PCE terms, which on the surface makes America look like a global outlier still fighting a meaningful inflation problem, according to Investing.com.

To test that assumption, Goldman broke core inflation down into three separate pieces: core goods, non-shelter services, and shelter, rather than treating it as one uniform number.

More Economy:

The results were stark. Inflation has already returned to its long-run trend in both emerging markets and non-U.S. developed economies, while the United States remains elevated, running roughly 3 percentage points above its pre-pandemic trend in PCE terms, according to Investing.com.

That gap matters because it reframes the entire debate. If most of the rest of the world has already normalized while the U.S. sits well above trend, the usual assumption that America’s inflation fight is simply further behind everyone else’s starts to look too simple. Goldman’s analysis suggests the composition of inflation matters as much as the headline gap.

Why US goods prices are the real outlier

Peters traced the entire U.S. goods overshoot to two temporary, country-specific factors rather than anything structural. Tariffs alone are adding an estimated 2.4 percentage points to year-over-year core PCE goods inflation, an effect she expects to fade substantially by the second half of next year.

Goldman’s earlier research estimated that pass-through of tariff costs to consumers would eventually reach about 70%, rising from an initial rate of roughly 55% as companies gradually adjust their pricing, Investing.com reported.

The second factor is an AI-related measurement distortion tied to how memory chips and software bundling appear in official price data. AI-driven memory price increases are adding roughly 1 percentage point to core goods inflation through the software and accessories category.

Peters expects that distortion to ease in 2027 as memory price pressures moderate and the Bureau of Economic Analysis adjusts its weighting methodology, Investing.com reported.

The scale of that AI effect is unusually concentrated in the United States. Software and accessories carry a much larger weight in the U.S. PCE basket than in other developed economies.

Goldman’s earlier research estimated that AI-related factors were already lifting overall U.S. core PCE by more than 20 basis points annually and could reach 50 basis points by year-end 2026, far exceeding comparable effects in other developed nations.

Peters traced the entire U.S. goods overshoot to two temporary, country-specific factors rather than anything structural.

Bloomberg / Getty Images

Services and shelter tell a different story

On services, Goldman’s findings cut against the popular narrative just as sharply. Non-shelter services inflation looked elevated across most economies in raw terms. After adjusting for how countries differently measure medical and financial services, core services inflation in the U.S. is actually lower than in other large developed markets.

Labor costs reinforce that conclusion. Unit labor cost growth looks more contained in the United States than abroad, which Peters said points to more risk of sustained inflation pressure outside the U.S. rather than within it.

That domestic labor picture is part of why Goldman has recently told investors to weigh inflation data over jobs data when trying to predict the Fed’s next move.

Shelter inflation adds one more layer of nuance. Rent inflation has fully normalized in both the U.S. and emerging markets, but remains elevated in other developed economies, particularly those that have pulled back less aggressively on immigration, according to Investing.com.

What this means going forward

Peters’ bottom line was direct. “Our results suggest that the U.S. has less of an inflation problem than other countries, despite what current top-line measures suggest,” she wrote.

Goldman has already pushed back its rate cut timeline, now expecting the next reductions in December 2026 and March 2027 as core PCE hovers near 3% through the rest of this year. That stance assumes much of today’s overshoot is temporary rather than structural, TheStreet reported.

It followed months of upward revisions, including one instance where Goldman raised its core PCE forecast after finding a disconnect between how CPI and PCE weigh consumer electronics and used-car prices.

If Peters is right that tariffs fade by the second half of next year and the AI-driven price distortions prove temporary rather than permanent, the more encouraging story buried inside this week’s report is that America’s inflation problem may look considerably smaller a year from now than the current headline numbers suggest.

Related: JPMorgan CEO doubles down on his inflation and economy verdict