According to data released by the U.S. Department of Labor on August 7th, the U.S. economy added -23,000 nonfarm payroll jobs in July, marking the first negative reading in months and significantly missing the consensus estimate of around +80,000. Furthermore, the prior two months' data were substantially revised downward, with a net downward revision of 103,000 jobs for May and June combined. In a contrasting signal, the unemployment rate fell from 4.2% to 4.1%, defying expectations for a rise to 4.3%.

The three-month average payroll gain of just 20,000 is not only far below the monthly average seen in 2025 but is also approaching the lower bound of the breakeven level needed to keep the unemployment rate stable. Slowing immigration and an aging population are rapidly lowering this breakeven point. The Federal Reserve Bank of Dallas estimates that net illegal immigration turned negative in early 2025 and is now flowing out at a rate of about 89,000 per month by July 2026. The Federal Reserve Bank of St. Louis estimates the 2026 breakeven range to be between 15,000 and 87,000 per month, with the lower bound corresponding to a scenario of net immigration outflow of 77,000 and the upper bound corresponding to a CBO estimate of net immigration inflow of 48,000. The wide range reflects significant disagreement among institutions regarding the pace of immigration outflow. Using this as a benchmark, the three-month average of +20,000 still barely falls near the lower bound of the range—implying that even with weaker job creation, the unemployment rate may not necessarily rise.

From a sectoral perspective, the primary drags on employment in July were in local government education (-50,000), leisure and hospitality (-40,000), retail trade (-19,000), and financial activities (-14,000). Positive contributions came from healthcare (+22,000) and construction (+22,000). The data may include distortions from the survey period, such as the decline in local government education jobs being linked to the end of the school year, a trend typically reversed in the fall. The decrease in leisure and hospitality employment may include a reversal of the "World Cup effect." However, the weakness in sectors like finance and retail could be related to factors such as AI displacement and a structural cooling of consumer spending. The supportive effect of the healthcare industry on employment data has also weakened compared to trend values. Notably, the diffusion index of employment breadth fell to 51.8% from 53.2%.

The weakness in interest rate-sensitive and consumer-sensitive sectors was more pronounced in this data. The internal structure of the -19,000 decline in retail trade generally reflects a cooling in consumption: warehouse clubs and supercenters lost 21,000 jobs, and gas stations lost 5,000, partially offset by a gain of 10,000 in sporting goods and miscellaneous retail stores. Financial activities lost 14,000 jobs, with credit intermediation losing 9,000 and insurance losing 7,000, potentially linked to AI replacing back-office processes. Professional and business services, a proxy for AI capital expenditure, added 18,000 jobs in July, while manufacturing saw little change, adding 5,000 jobs. Overall, there are few industries within the private sector with genuinely endogenous expansionary power. Industry breadth and the aggregate total are converging simultaneously. The narrowing of industry breadth means that employment expansion is no longer broad-based, a structural pattern that historically has often preceded a trend of weakening in the aggregate total.

The unemployment rate data performed better than the nonfarm payrolls figure. Dissecting the data, the number of unemployed persons fell by 178,000 to 6.916 million, but the number of employed persons also decreased by 87,000 to 162.18 million. Consequently, the total labor force shrank by approximately 260,000, meaning the decline in the unemployment rate resulted from the numerator (unemployed population) contracting faster than the denominator (total labor force). Correspondingly, the labor force participation rate fell to 61.4%, the lowest in 50 years excluding the pandemic period, confirming a net contraction in labor supply. Breaking down the reasons for unemployment, the decrease in the number of unemployed people came almost entirely from labor force exits, not from improved employment conditions. The number of reentrants to the labor force who were unemployed fell by 114,000 to 2.123 million, the largest source of decline, often implying that people who had previously re-entered the job search have given up again and exited the statistics. Permanent job losers decreased by 54,000 to 1.715 million, and new entrants decreased by 11,000 to 761,000. Conversely, temporary layoffs increased by 153,000 to 921,000, and job leavers increased by 17,000 to 793,000. Put simply, the decline in reentrants was the main reason for the drop in the unemployment rate this month, while the jump in temporary layoffs is a negative change. If these temporary layoffs later convert to permanent job losses, it could lead the unemployment rate to rise again. A few positive structural signals came from the decline in permanent job losses and long-term unemployment. Permanent job losses fell from 1.769 million in June to 1.715 million, continuing their decline from the February high, suggesting that businesses have not yet engaged in recessionary-scale permanent layoffs. However, the figure of 1.715 million is still significantly higher than the pre-pandemic (2019) norm of about 1.1 to 1.2 million, indicating that the number of permanently unemployed remains elevated. Long-term unemployment (27 weeks or more) fell from 1.937 million to 1.771 million, and its share of total unemployment decreased from 27.3% to 25.6%, providing a marginal easing of structural damage. Overall, the unemployment side has not shown a recessionary deterioration, but the improvement is primarily due to labor force exits rather than demand expansion.

The ratio of job openings to unemployed persons remains above 1 and is tightening again, but this is a result of supply contraction rather than excess demand. The June JOLTS report showed 7.4 million job openings (down from 7.5 million in May), with an opening rate of 4.4%, a quits rate of 2.0%, and a layoff rate of 1.1%, all at low levels. Based on the 7.4 million June openings and the 6.916 million unemployed in July, the V/U ratio is approximately 1.07 (up from about 1.06 the prior month), above 1 but still below the 2019 level of about 1.2. From a supply-demand gap perspective, labor demand (household employment of 162.177 million plus job openings of 7.4 million, totaling 169.577 million) still exceeds the labor supply (169.094 million) by about 480,000. It is important to note that this current tightness is completely different from the excess demand of 2021-22. It is a supply-side tightening where openings are flat and the labor force is contracting. This suppresses the unemployment rate and maintains some wage resilience but does not correspond to an overheating economy. If the change in the participation rate is excluded, most of the improvement in the unemployment rate disappears. Using the July civilian noninstitutional population of about 275.4 million, if the labor force participation rate had remained at June's 61.5%, the labor force would have been about 169.37 million, which is about 280,000 higher than the actual figure. If these 280,000 people had remained in the labor force and been counted as unemployed, the number of unemployed would have risen from 6.916 million to about 7.19 million, implying an unemployment rate of about 4.25%, which is about 0.15 percentage points higher than the reported 4.1%. The labor force participation rate falling to 61.4%, the lowest in 50 years excluding the pandemic, is the fundamental backdrop for the current decline in the unemployment rate. The participation rate fell by 0.1 percentage point from June's 61.5%, and the employment-to-population ratio also fell from 59.0% to 58.9%, both confirming a net contraction in labor supply. The structural decline in the participation rate is primarily driven by supply-side factors, not by workers voluntarily exiting. The number of discouraged workers in July was 476,000, essentially flat compared to June's 477,000, suggesting that the exits are not due to frustration in finding work. The participation willingness of the prime working-age population (25-54) is also relatively rigid historically. The true pressure comes from two slow-moving variables: first, slowing immigration, even turning to net outflow (the Dallas Fed estimates July 2026 net illegal immigration outflow at about 89,000 per month), directly draining labor supply; second, the aging population driven by the ongoing retirement of the Baby Boomer generation. The labor force participation rates for the 20-24 age group and those aged 55 and over fell from 70.6% and 37.1% to 70.2% and 36.9%, respectively. The combination of these factors has significantly lowered the breakeven point for employment—in an environment of contracting supply, even if nonfarm payrolls turn negative, the unemployment rate could still fall or even rise, which is the fundamental reason why the 4.1% reading needs to be interpreted cautiously.

Wage pressures continue to subside. Average hourly earnings rose by 0.1% month-over-month in July, and the year-over-year increase slowed to 3.2% from the prior month's 3.4%. This is now close to the pre-pandemic norm of 3.0% to 3.2%, sitting at the lower end of the 3.0% to 3.5% range generally considered compatible with the 2% inflation target. The cooling of both wages and employment suggests the risk of a "wage-price spiral" is further diminishing. The average weekly hours worked remained stable at 34.3 hours for the third consecutive month, and manufacturing hours were 40.4 hours, indicating that businesses are responding to slowing demand by freezing hiring rather than cutting hours. The fact that hours have not been cut and permanent layoffs have fallen suggests that the recessionary chain of first cutting hours, then laying off workers, has not yet systematically started. On an aggregate level, with the number of employed persons falling and hours worked flat, the year-over-year growth rate of nominal labor income (hours worked multiplied by hourly wages) has slowed from over 4% earlier. After accounting for inflation of around 3.5%, the year-over-year change in real aggregate labor income is near zero, meaning the real purchasing power of the household sector is treading water.

In summary, the July U.S. employment data was weak, shown by the nonfarm payrolls figure missing expectations and the downward revisions to the prior three months, the decline in the employment diffusion index, and the cooling of wage growth. However, it is important to note that U.S. nonfarm payroll data is historically characterized by a "sawtooth" pattern and high noise, with significant single-month volatility. It is an incremental data series, inherently less smooth than stock data; it is an absolute value, less smooth than diffusion indices or ratio data; and the "birth-death model" it relies on can amplify data fluctuations. Therefore, the typical sampling error for nonfarm payrolls can be around ±100,000, and it relies on repeated revisions to confirm the underlying trend. Given that the July PMI data was relatively high, the assessment is that the labor market situation is neither as optimistic as the unemployment rate suggests nor as pessimistic as the nonfarm payrolls figure indicates. For the Federal Reserve, a prudent approach of waiting and observing is the rational choice. Subsequent inflation data will be another important indicator to watch. Following the FOMC meeting in July, officials reiterated the commitment to "deliver price stability," noting that prices are "still too high," systematically suppressing expectations for rate cuts under the constraint of still-elevated inflation. The July labor report fits this narrative perfectly: the weak payrolls prevents an overheating narrative and removes the empirical basis for a hawkish rate hike, while the headline 4.1% unemployment rate gives the dovish camp no reason to cut rates before inflation is under control. The assessment maintains the view that the policy rate will remain unchanged at 3.5% to 3.75% for the year, with the subsequent focus remaining on the inflation path.

Following the data release, the implied probability of a rate hike as measured by the CME FedWatch Tool declined overall. The probability for a September rate hike fell to 44%; the cumulative probability for an October rate hike fell to 59.2%; and the cumulative probability for a December rate hike fell to 77.1%. In terms of market reaction, major assets reacted relatively mildly to the reduced probability of a rate hike. The 10-year U.S. Treasury yield fell 4 basis points to 4.65%, and the 2-year yield fell 6 basis points to 4.19%. The U.S. Dollar Index fell 0.34% to 99.59. The three major U.S. stock indices rebounded, with the Dow Jones Industrial Average rising 0.28%, the S&P 500 rising 0.62%, and the Nasdaq Composite rising 1.3%. The positive macro backdrop, combined with strong earnings reports from the software sector, led to broad gains for large-cap technology stocks, and the Philadelphia Semiconductor Index rose 2.56%. U.S. stocks ended the week higher, with the software sector's earnings reports standing out and leading to widespread gains in large-cap tech stocks. Other outperforming sectors included internet, semiconductors, industrial metals, engineering & construction, homebuilding, healthcare, freight, auto parts, and cosmetics. Underperforming sectors included insurance, payments, credit cards, airlines, restaurants, and food & grocery retailers. U.S. Treasury bonds generally strengthened, with the 10-year yield down 4 basis points to 4.65% and the 2-year yield down 6 basis points to 4.19%. The U.S. Dollar Index fell 0.34% to 99.59. Gold rose 2.3% and was up over 7% for the week. Silver rose 3.1%.

Risk Warnings: Inflation may not fall as quickly as expected, or fiscal easing could reignite demand overheating, forcing the Federal Reserve to maintain high interest rates for a longer period. Geopolitical uncertainties and potential changes in tariff policies could pose a supply-side shock to supply chain recovery. If macroeconomic data deviates from the baseline soft-landing path, asset prices currently pricing in rate cuts and a soft-landing economic scenario could face significant valuation correction risks.