Goldman Sachs Says US Inflation Is Overstated by Tariffs and AI Data Quirk
TREE NEWS reports: Goldman Sachs economists argue that the headline gap between US core inflation and that of other major economies is largely an illusion created by two temporary, US-specific distortions: tariffs and an AI-driven statistical quirk. In a global economic analysis published on September 22, the bank estimates that current tariffs are adding about 2.4 percentage points to year-on-year core PCE goods inflation, an effect expected to fade substantially in the second half of next year. A separate AI-related bias, stemming from how the PCE basket weights software and accessories, is contributing roughly 1 percentage point to core goods inflation and should unwind gradually by 2027.
Why the Data Is Distorted
The report notes that core goods inflation has largely returned to pre-pandemic trends in emerging markets and non-US developed economies, yet remains about 3 percentage points above trend in US PCE terms and about 1 percentage point higher in CPI terms. Goldman attributes the entire excess to the two temporary factors. On tariffs, the drag is mechanical: as higher import costs pass through to consumer prices, they lift goods inflation, but the effect should diminish once the level shift is absorbed. On the AI side, the PCE category for software and accessories carries a larger weight and partly captures business purchasing through retail channels, magnifying this year’s memory price surge. Because the category is not quality-adjusted, it fails to reflect the real performance gains delivered by AI tools. Memory prices contribute less than 0.1 percentage point to core goods inflation in other developed economies, versus about 1 point in the US.
Services and Housing Tell a Different Story
Goldman also challenges the view that US services inflation is uniquely sticky. After harmonizing definitions across countries for categories such as healthcare and financial services, the bank finds US core services inflation is actually lower than in other major developed economies. PCE treatment of financial services prices differs sharply from other nations and is highly sensitive to equity market swings, while medical insurance is also measured differently. More importantly, unit labor cost growth — a key forward-looking driver of non-housing services inflation — is close to its pre-pandemic average in the US but notably higher elsewhere, implying greater persistent inflation pressure outside America. On housing, US and emerging-market rent inflation have normalized, while other developed economies remain well above long-term trends. Economies with larger declines in net immigration, such as the US, Canada and New Zealand, have seen sharper rent disinflation. Market rent data, which is not included in official inflation statistics but offers the timeliest signal, points to a more favorable US outlook.
Market Implications
- Rates and bonds: If US inflation pressure is genuinely weaker than headline data suggest, the case for the Federal Reserve to ease policy or at least avoid further tightening strengthens, supporting Treasuries at the front end of the curve.
- Equities: A softer underlying inflation picture reduces the risk of a policy-driven earnings squeeze, a supportive backdrop for rate-sensitive growth and technology shares.
- US dollar: Markets pricing in a less hawkish Fed could weigh on the dollar, particularly against currencies where underlying inflation is more stubborn.
- Commodities and crypto: Lower real-rate expectations and a weaker dollar have historically been tailwinds for gold and bitcoin, though tariff-driven goods price swings may keep near-term inflation prints noisy.
Key Takeaways for Investors
The central message is that investors should look through the tariff and AI distortions in US inflation data rather than extrapolate them into the policy path. The fading of these effects into 2026 and 2027 suggests the Fed may have more room to maneuver than the headline numbers imply, while other developed economies may face more entrenched inflation problems. That divergence argues for relative-value positioning across rates, currencies and equities rather than a single directional bet on US inflation.
