Behind the PCE "technical squeeze-out of froth": Market sentiment is warming, and expectations for falling U.S. Treasury yields are rising.

Key points of the August PCE data 1. Core PCE price index (excluding food and energy, the Fed's most closely watched inflation gauge): +3.0% year-over-year (below the market expectation of 3.3%, previous value 3.3%), +0.2% month-over-month (below the market expectation of 0.3%, unchanged from the previous value of 0.2%). 2. Headline PCE price index: +3.4% year-over-year (below the market expectation of 3.7%, previous value 3.7%). Immediate market reaction After the data release, as both core PCE and headline PCE growth came in broadly below prior market expectations, market concerns about inflation stickiness were noticeably alleviated: 1. U.S. equity index futures rose in a straight line: S&P 500 futures briefly climbed 0.37%, Nasdaq 100 futures rose 0.35%, and Dow futures rose 0.31%. 2. Rate expectations repriced again: previously, the market worried that the Fed would continue to aggressively hike rates at its October meeting because of rebounding energy prices and consumer resilience, but the below-expected PCE data sharply reduced the pressure for near-term urgent rate hikes. Our judgment logic We believe that in the midterm election results, the Democratic Party will most likely control the House of Representatives, which will form an effective check on the Trump administration (for example, fiscal discipline, launching wars, etc.), and market uncertainty about the Trump administration will be soothed. Therefore, the 10-year U.S. Treasury yield is expected to move lower, and the market may begin to warm up starting in November. If next month's CPI continues today's PCE slowdown, U.S. Treasury yields may decline earlier. 1. Disinflation regains momentum: compared with the high stickiness of inflation in previous months, August core PCE year-over-year fell to 3.0%, indicating that the short-term shock from tariffs and rising goods prices is beginning to fade, and the growth rate of housing services prices has also slowed further. 2. It gives the Fed breathing room for monetary policy: as Williams said in his latest remarks, there is "no need to act in haste"; one more hike within the year remains the baseline, but October is not a must. This below-expected inflation reading provides strong data support for the Fed to stay on hold and pause rate hikes at its October meeting. 3. The market's focus shifts to consumption and the labor market: although disinflation exceeded expectations, subsequent developments in household personal consumption expenditures and real income still need to be closely watched. If U.S. household consumption remains resilient, the process of inflation returning to the long-term 2% target may still be accompanied by periodic reversals. However, alongside this August PCE data release, the BEA also published its Annual & Comprehensive Revision, mainly involving corrections to statistical methods, rebasing of consumption weights, and recalculation of price deflators for detailed components. Main detailed components of the specific adjustments According to the technical details published by the BEA for this revision, changes in component weights and fine-tuning of calculation methods are mainly concentrated in the following four core areas: 1. Financial and insurance services Adjustment content: the BEA revised the price deflator calculation model for financial services without explicit fees (such as bank custody services and fee-free transactions). Weight/change: this category has a weight of about 1.7% in core PCE. The new model corrected the previous bias that overstated price increases in this area, lowering the annualized increase for this detailed component from 14.3% under the old method to 12.3% under the new method. 2. Healthcare services Adjustment content: based on the latest actual settlement data from the Centers for Medicare & Medicaid Services (CMS), the actual expenditure weights for hospital services and outpatient medical care were slightly adjusted. Weight/change: because the share of medical costs in GDP and the expenditure approach had previously been recalculated, the weight of medical services was slightly raised. However, since actual medical service prices have changed moderately year-over-year in the past two quarters, increasing the weight of this item instead diluted the pull from high-inflation items (such as housing). 3. Housing and owners' equivalent rent Adjustment content: although housing accounts for as much as about 34% of CPI, its weight in PCE has long remained between 15% and 18%. Based on annual population mobility and consumption survey data, the BEA rebalanced the actual consumption share of housing services. Weight/change: the slight downward adjustment to the housing weight further weakened the pull of "sticky housing inflation" on the overall PCE index. 4. Computer software and accessories Due to differences in statistical scope, this category accounts for only about 0.035% of the CPI basket, but its weight in the PCE basket is as high as about 1.2% (more than 30 times higher). Weight/change: the BEA repackaged broad IT services, financial data processing, and cloud hosting and incorporated them into a new deflation algorithm. Although the total weight of this broad component slightly expanded to about 1.7% in the new accounting system, its internal price growth rate was sharply reduced. The BEA recalculated the supply-based bridge table, compressed the actual weight of physical storage hardware to less than 10%, and used the PPI for software publishing and cloud infrastructure (90%+) as the core price deflator. Fed research points out that after this revision, the annualized contribution of this category to core PCE will fall by about 0.16 percentage points (from 0.66 percentage points to 0.50 percentage points). If the "old weights/old method" were used, what would this PCE reading be? The BEA's method update and weight revision this time had a downward effect of about 0.2-0.3% (that is, 20 to 30 basis points) on the year-over-year decline in core PCE: Summary: the real implications for policy judgment If the BEA's "technical revision" is not taken into account, August core PCE year-over-year should actually have been around 3.2%-3.3%. This means: 1. Nominal upside surprise vs. actually moderate cooling: on the surface, inflation fell sharply from 3.3% last month to 3.0% (far exceeding the market expectation of 3.3%), but about half of that was due to the "technical squeezing out of froth" from statistical tools and weights. 2. The Fed's real perspective: when assessing inflation trends, Fed officials usually filter out this kind of one-time base effect caused by changes in statistical methods and focus more on the true inflation momentum after removing technical fluctuations. Therefore, although the 3.0% reading is a big boost to market sentiment on the screen, the Fed will still maintain a certain degree of prudence when setting interest rate policy. This article is reprinted from "Zhang Yidong Strategy World"; GMTEight editor: Zheng Yuyang.