JPMorgan: US Q3 GDP Revised Up to 2.5%, September Rate Hike Depends on Inflation

Overview

This issue of 'US Weekly Outlook' focuses on the interplay between economic resilience and the inflation outlook in the United States. Based on strong business survey data, JPMorgan has raised its Q3 GDP growth forecast to 2.5%, while simultaneously noting that the July employment report sent mixed signals of low unemployment alongside weak new job creation. The report argues that whether the Federal Reserve raises rates in September will heavily depend on upcoming inflation data. It also delves into the statistical implications of the manufacturing ISM index rebound, the long-term suppression of labor income shares by AI, and the structural reasons why the gap between core CPI and PCE inflation is unlikely to fully revert to historical averages.

Core views

The divergence between economic growth and the labor market is the current core contradiction. The research report raises the forecast for actual Q3 GDP growth from 1.75% to 2.5%, primarily based on the upward revision of the July Services PMI final reading to 54.6, an increase of nearly 4 points in the ISM Services Business Activity Index, and expectations for manufacturing inventory restocking. However, labor market signals are mixed: although the unemployment rate dropped to 4.1% (the lowest since early last year), non-farm payrolls decreased by 23,000, the three-month average was only 20,000, and the labor force participation rate declined again. The report suggests this weakness may be partly due to seasonal summer disturbances and lagging layoff effects after the exhaustion of federal pandemic relief funds in the education sector. Nevertheless, slowing wage growth (only 2.3% annualized over three months) indicates that the labor market is not overheated, reducing the urgency for the Fed to raise rates due to tight labor conditions. Regarding the inflation path and Fed decision-making, the report expects the core CPI to rise 0.22% month-over-month in the upcoming release, a level insufficient to prompt the Fed to raise rates at its September meeting unless subsequent readings consistently approach 0.3%. A key observation point is whether core goods prices will rebound after two consecutive months of declines, which is crucial for validating the 'inflation stickiness' hypothesis. Additionally, the report notes that the negative gap between core CPI and PCE (-0.7 percentage points) is an anomaly, expected to narrow in the future but not fully return to historical norms (0.45 percentage points). Due to differences in housing inflation weights and the relative strength of PCE super-core inflation, core PCE may exceed core CPI by 0.1 percentage points next year, meaning that even if CPI falls, the Fed's focus on the PCE metric may remain under pressure. Regarding the recent sharp rebound in the manufacturing ISM index (rising from 47.9 to 55.6), regression analysis in the report indicates this is more of a 'catch-up' from previous excessive weakness rather than a sudden change in fundamentals. Historical data shows that since the 'Great Moderation' period starting in 1985, the ISM Manufacturing Index's predictive power for GDP has significantly weakened, with its prediction errors remaining persistently negative between 2022 and 2025. In contrast, ISM Services and Markit Manufacturing PMIs have stronger explanatory power for current GDP. According to model calculations, the current ISM Manufacturing reading implies a Q3 GDP growth rate of approximately 2.3%, which, while improved from previous periods, does not support the explosive growth signal suggested by the index's surface appearance. On structural issues, the report warns that AI may further suppress labor income shares. The labor share in the non-farm sector reached a historical low in Q1 2026 (53.7%), and this is not merely a statistical artifact but reflects a systemic improvement in capital returns. Most macroeconomic models predict that as AI replaces human labor and companies use AI for differentiated pricing, the labor share will continue to decline, potentially falling below 50% by the end of this decade. From a macroeconomic equilibrium perspective, if national income continues to concentrate among capital owners, public or external sectors may need to expand borrowing further to offset demand shortfalls caused by 'excess savings among the rich' to maintain positive real interest rate equilibria.

Analysis framework

The report employs a multi-dimensional data cross-validation method to assess the economic situation. When judging GDP trends, it does not rely solely on a single indicator but comprehensively compares the historical fit of ISM, PMI, and hard data. After finding that the predictive efficacy of manufacturing ISM has declined, it shifts to assigning higher weight to services surveys. In analyzing the inflation gap, it uses an Error-Correction Framework to test the long-term cointegration relationship between CPI and PCE, finding that convergence speed at the growth rate level is faster than at the price level, leading to the conclusion that the gap will narrow but is difficult to return to normal. For labor share trends, it combines BEA National Income and Product Accounts data with Census Bureau Quarterly Financial Report (QFR) profit margin data for cross-validation, ruling out statistical noise caused by changes in small business tax classifications and confirming the widespread nature of rising capital shares.

Methodology notes

• Macroeconomic framework

  Structural Breakpoint Test of Soft Data Predictive Efficacy

  The report found through structural breakpoint tests that since the 'Great Moderation' began in 1985, the coefficient and goodness-of-fit of the ISM Manufacturing Index predicting GDP have significantly declined. This提示s investors to be wary of signal attenuation when using traditional business survey indicators to predict modern economic growth, avoiding simple application of historical experience.

• Macroeconomic framework

  Error Correction and Cointegration Analysis of Inflation Indicators

  Rather than directly comparing the price levels of CPI and PCE, the report models the long-term cointegration relationship between the 'changes in their growth rates'. This method more accurately captures the convergence speed of the two indicators during dynamic adjustment, helping to determine whether the current difference in inflation readings is a temporary deviation or structural divergence.

• Corporate Fundamentals and Financial Framework

  Cross-Validation of Labor Share Trends Using Multi-Source Data

  To exclude statistical interference from owner income distribution in national accounts, the report introduces Census Bureau QFR manufacturing profit margins as an independent proxy variable and distinguishes between large and small enterprise samples. This cross-validation method effectively eliminates noise such as tax planning, confirming the authenticity of changes in macroeconomic distribution patterns.

Key data

• US Q3 Real GDP Forecast (SAAR)2.5%Upward revised from previous forecast of 1.75%, reflecting strong business surveys and inventory restocking expectations
• July Unemployment Rate4.1%Down 10 basis points from the previous month, the lowest since early last year
• July Non-Farm Payroll Change-23,000Three-month moving average only 20,000; private sector job growth is sluggish
• July Core CPI Forecast (MoM)0.22%Expected to be insufficient to trigger a September rate hike; watch for rebound in core goods
• Non-Farm Sector Labor Income Share (Q1 26)53.7%Historical low, down approx. 3 percentage points from pre-pandemic levels
• Core CPI-PCE Gap Forecast (Next Year)-0.1%PCE expected to be 0.1 percentage points higher than CPI; previously forecast to be 0.2 percentage points lower

Impact & implications

The report believes the US economy remains on a 'soft landing' track, but the mismatch between growth momentum and inflation pressures increases policy uncertainty. The upward revision in GDP means near-term recession risks are alleviated, but implicit weakness in the labor market and continued decline in labor shares suggest that the mid-to-long-term demand foundation may be eroded. For the Fed, the structural differentiation in inflation indicators (PCE relatively stronger than CPI) implies that even if headline CPI cools, the threshold for monetary policy easing may be higher than market expectations. Furthermore, if AI-driven changes in distribution patterns persist, fiscal policy or external imbalances may be needed to absorb excess savings, posing profound impacts on long-term interest rates and asset pricing.

Risks

• If core goods prices continue to fall, it may falsify the inflation stickiness hypothesis and alter the Fed's path
• If labor force participation continues to decline sharply, the healthy signal from unemployment data will be distorted
• Rapid decline in labor share driven by AI may lead to insufficient aggregate demand, forcing the public sector to expand deficits
• If the manufacturing ISM index rebound is merely statistical noise, actual industrial output may fall short of expectations

What to watch

• Whether core goods prices in next week's core CPI data end their downward trend
• July retail sales data (needs to be combined with June data to remove Prime Day shift effects)
• Legal progress of the attempt to remove Fed Governor Cook and its impact on voting at the September meeting
• Whether the strength of PCE super-core inflation relative to CPI persists
• Whether manufacturing inventory data re-syncs with survey signals