US monetary policy, neutral rate, and macroeconomic outlook: UBS questions whether a single neutral rate can guide policy in an AI-driven, highly uneven US economy.

Overview

This weekly US macro report examines whether a meaningful neutral policy rate exists amid AI-led investment, weak rate-sensitive sectors, elevated inflation and repeated supply shocks. UBS concludes that policy restraint may already be sufficient over time, but the FOMC’s mechanism, desired speed of disinflation and reaction function remain unclear.

Core views

UBS argues that the usual concept of r-star—a single real policy rate that balances output and inflation—is poorly suited to the current US economy. A single overnight rate cannot simultaneously offset oil-price inflation, restrain AI capital expenditure and avoid raising interest income for asset-rich households. The report notes that the Laubach-Williams 2026Q2 estimate of the real policy rate was 1.65%, but argues that sectoral imbalances, continuously changing shocks and differing rate sensitivities leave no single economy-wide equilibrium rate. The September FOMC projections add to the ambiguity: the projected policy-rate midpoint at the end of the forecast horizon was 1.6 percentage points above inflation and 0.4 points above the longer-run median; longer-run dots ranged around a 3.0% mode, 3.2% median and 3.4% average. UBS challenges the view that stronger potential growth has raised r-star. Population growth has slowed, labor-force participation has fallen, and productivity rose only 1.27% annualized over the past three quarters, below the roughly 1.6% pre-Covid low-productivity trend and well below the more than 2.5%-3% pace associated with high-productivity regimes. Real GDP grew 2.1% over the four quarters through Q2 despite the AI impulse, partly because much AI-related equipment is imported. UBS therefore sees little evidence that potential growth warrants a higher neutral rate. The report’s central evidence is an extreme split between AI-related activity and the rest of the economy. Residential investment contracted in four of the last five quarters and was expected to contract again in 2026Q3; real house prices were falling, nationwide nominal house prices were only 1.5% higher year-on-year through June, and nonresidential structures investment had fallen in nine of the prior ten quarters. Equipment investment outside information processing and data-center inputs was down 1.5% year-on-year. In contrast, high-tech manufacturing rose 12.5% over the 12 months through August while the remaining 97% of manufacturing rose only 0.3%. UBS estimates that each 1 percentage point contribution from AI-related business fixed investment to real GDP growth generates a further 0.5 percentage point contribution from the remaining components of business fixed investment through spillovers. AI-related investment and equity wealth are, in UBS’s view, less clearly rate-sensitive than housing and conventional investment. The top 20% of households hold liquid assets roughly one standard deviation above their pre-Covid norm, and the top 10% own most equity wealth; higher rates therefore increase interest income for a group supporting consumption. The bottom 80% have historically low savings, while the bottom 90% primarily hold housing wealth whose real value is falling. UBS estimates that the AI impulse accounts for more than all real GDP growth over the past four quarters, with the rest of the economy contracting 0.7%. This leaves policymakers facing a difficult trade-off: if AI investment slows, little else is supporting growth; if rates must reduce inflation through traditional labor-market slack, the report estimates a Phillips-curve coefficient of 0.3 would imply unemployment near 8% to lower inflation from above 3% to 2%. UBS expects the September 30 annual National Economic Accounts update to revise data back to Q1 2021 and potentially alter the recent inflation and activity narrative. New price methodologies for portfolio management, software and legal services are estimated to lower August core PCE inflation by 0.18 percentage point versus the old methodology, to 3.18% year-on-year, from 3.34% in the latest release. Portfolio-management and software changes are expected to reduce 12-month core PCE inflation by 0.08 and 0.15 percentage point respectively, partly offset by a 0.05-point increase from legal services. UBS cautions that details remain undisclosed, the effects will not be a uniform downward shift, and lower current inflation could be offset by upward revisions to 2022-24 inflation. Revisions may also reshape the composition of growth and income: real GDP and real GDI were both up 2.1% year-on-year in Q2 2026, while revisions to income, inventories and equipment remain key uncertainties. For the near-term policy outlook, UBS lifted its September headline CPI forecast to 0.54% month-on-month and 3.61% year-on-year because of energy prices, while retaining a 0.18% core CPI forecast and 2.40% year-on-year core inflation forecast. It removed a projected June 2027 rate cut after the September FOMC meeting, aligning with the dot plot, and now sees two further hikes this year followed by a prolonged hold, with risks tilted upward. However, UBS believes the current funds rate is sufficient to return inflation to target over time and doubts four more hikes are required over the next few years. It identifies the Chair’s intended transmission mechanism—whether policy is meant to restrain AI capex, equity valuations, inflation expectations or broader demand—as the key unresolved issue. UBS expects a reasonably firm September employment report: nonfarm payrolls up 95K, including 90K private and 5K government jobs; unemployment down 6 basis points to 4.08%; average hourly earnings up 0.21% month-on-month and 3.1% year-on-year; and average weekly hours down to 34.3. Low claims support the forecast, but underlying hiring measures are subdued. UBS estimates published job gains have been overstated by 10K-20K per month, implying a more fundamental pace near 55K, and flags unusually large seasonal-adjustment uncertainty from the late Labor Day calendar configuration. It expects the September residual net birth-death adjustment near -170K on a non-seasonally adjusted basis and warns preliminary adjustments may later revise lower. Other data point to resilience but unevenness. UBS raised Q3 equipment-investment tracking to 20.1% and real GDP tracking to 3.1% after August core capital-goods orders rose 1.6%. It expects August nominal spending to rise 1.0%, real spending 0.6%, and nominal income 0.4%, adding pressure to the saving rate. It forecasts August headline and core PCE increases of 0.32% and 0.27%, respectively. Housing remains soft despite an August rise in new-home sales: months’ supply was still elevated at 8.5 and the median new-home price was down 5.8% year-on-year. Rising mortgage rates—7.12% for the MBA 30-year fixed rate and 7.03% for the Freddie Mac measure—could further tighten affordability. The report also highlights energy and policy risks. Brent was $106 per barrel on September 25, versus a recent $79 trough, while regular gasoline was $4.49 per gallon and diesel $6.50. UBS notes that a diesel-export ban could lower US diesel prices in the short run but could also reduce refinery operating rates, gasoline and jet-fuel output, crude demand and producer revenues, while pressuring renewable-diesel margins and soybean prices. The US-China trade truce was extended to January 10, 2027, with a possible broader economic package still under negotiation.

Analysis framework

UBS first tests the neutral-rate concept against sectoral transmission and potential-growth evidence, then decomposes activity into AI-related and non-AI components. It combines macro releases, surveys, household balance-sheet data, policy communications and forecast tracking, while explicitly adjusting interpretations for methodological revisions, seasonal effects and data-vintage uncertainty.

Methodology notes

• MacroeconomicsIS-LM model

  Neutral-rate discussion using the equilibrium of investment-savings and money markets.

  UBS describes r-star as the rate consistent with equilibrium output and monetary conditions, then argues that this single-rate framework is inadequate when sectors respond differently to policy.

• MacroeconomicsPhillips curve

  A Phillips Curve coefficient of 0.3 is used to illustrate the employment cost of faster disinflation.

  UBS uses the relationship between labor-market slack and inflation to show why forcing inflation from above 3% to 2% through conventional demand restraint could be painful.

• Industry AnalysisUpstream-Midstream-Downstream Transmission

  AI-related investment spillovers into non-tech business investment.

  UBS estimates that AI capex supports demand for other investment categories, separating the direct AI impulse from broader domestic spillover effects.

Key data

• Laubach-Williams real policy-rate estimate1.65%2026Q2 estimate cited in UBS’s discussion of r-star.
• Non-AI economy contribution-0.7%UBS estimate for the rest of the economy over the last four quarters, while the AI impulse accounted for more than all GDP growth.
• August core PCE inflation forecast after methodology changes3.18%Year-on-year; estimated 0.18 percentage point below the old-method result and below the latest 3.34% reading.
• September nonfarm payroll forecast95K90K private and 5K government; UBS flags unusually wide seasonal-adjustment uncertainty.
• September unemployment-rate forecast4.08%A 6-basis-point decline from 4.14%.
• Q3 real GDP tracking3.1%Raised after the August durable-goods report.
• Brent crude price$106 per barrelAs of Friday, September 25; up from $79 on August 4.

Impact & implications

UBS sees the lack of a credible neutral-rate anchor as making policy expectations less stable. AI-led growth may limit the effectiveness of further tightening while already weak housing and conventional investment bear more of the restraint; forthcoming inflation revisions, employment data and Fed communications are therefore central to judging the policy path.

Risks

• AI-related investment or equity wealth may prove less rate-sensitive than conventional sectors, forcing a more painful policy trade-off if the FOMC seeks faster disinflation.
• Energy-price and Middle East supply shocks could sustain inflation while being difficult for interest-rate policy to offset.
• Annual BEA revisions may materially change measured inflation, income and activity, creating unusual near-term forecast uncertainty.
• A continued rise in Treasury yields could lift mortgage rates further and worsen housing affordability.
• September payroll data face elevated seasonal-adjustment, calendar and later-revision risk.

What to watch

• The September 30 NIPA annual revisions, especially the new PCE price methodologies and their effect on current versus historical inflation.
• August PCE inflation, income and spending data, followed by the next CPI release.
• The September employment report, including payrolls, unemployment, wages, hours, revisions and seasonal-adjustment effects.
• FOMC speeches and indications of whether policy is intended to restrain AI capex, equity valuations or broader demand.
• Oil, gasoline and diesel prices; developments around a possible diesel-export restriction; and the Jones Act waiver.
• Whether AI-related investment spillovers continue to support non-tech capex and broader growth.