Exploring the Underlying Logic of the Significant Revision of Non-Farm Data

Core Viewpoints

The downward revision of non-farm data for May and June may accelerate the pace of interest rate cuts by the Federal Reserve. Based on current data, the probability of a "compensatory" rate cut in September is limited, with the core focus being on rate cuts rather than recession. On a deeper level, the non-farm data may impact the data-driven investment research analysis system, further highlighting the value of gold.

1. Non-farm employment data disappoints, accelerating rate cut pace: In July, the U.S. added 73,000 non-farm jobs, falling short of market expectations, with an unemployment rate of 4.2%. The non-farm data for May and June saw significant downward revisions. Following the release of the non-farm data, market expectations for a 25 basis point rate cut by the Federal Reserve in September surged from 37.66% to 80.31%, while fully pricing in at least one 25 basis point rate cut before October. The pace of rate cuts by the Federal Reserve may significantly accelerate in the next phase.

2. Conditions for "compensatory" rate cuts have not yet been met, and rate cut trading may still be mainstream: In September 2024, the Federal Reserve initiated this round of rate cut cycle with an unexpected 50 basis point cut. Based on current data, we believe it may not be sufficient to drive the Federal Reserve to implement another "compensatory" rate cut. Regarding the U.S. economy, we believe the most appropriate definition at present is that the momentum of growth is marginally weakening rather than heading into recession.

3. The data-driven investment research analysis system may continue to be impacted, highlighting the value of gold: In the short term, ADP employment data may serve as a "substitute" for non-farm data, and market attention to ADP employment data may further increase. In the long term, the impact of data credibility issues may gradually ferment, adding more disorder and noise to the financial market, further highlighting the value of gold.**

Main Text

1 Analyzing the Underlying Logic of the Significant Downward Revision of Non-farm Data

On August 1, the U.S. Department of Labor released the latest non-farm employment data for July, but what is more noteworthy is the significant downward revision of non-farm data for May and June. We previously suggested that the main logic of macro trading in the third quarter may gradually transition from TACO trading to rate cut trading, and this non-farm data is expected to further accelerate the pace of rate cut trading. Additionally, concerns about data credibility triggered by non-farm data may gradually ferment, combined with rising rate cut expectations and potential inflation risks, further highlighting the cost-effectiveness of gold.

The newly added non-farm data slightly missed expectations. The U.S. added 73,000 non-farm jobs in July, falling short of market expectations, with the private sector adding 83,000 jobs as the main support, while government jobs decreased by 10,000. The unemployment rate rose by 0.1 percentage points to 4.2%, generally in line with market expectations, primarily due to the resonance of a declining labor force and a marginal increase in the number of unemployed. Looking solely at the July data, it indicates a marginal cooling of the U.S. labor market while still maintaining a certain level of resilience, which may be the mild weakening state preferred by the market.

The newly added non-farm data for May and June has been significantly revised downwards, with both the structure and total amount deteriorating. Along with the release of the July non-farm data, the U.S. Department of Labor simultaneously revised the non-farm data for May and June, showing that the newly added non-farm employment data for June was revised down from 147,000 to 14,000. The initial value for May's newly added non-farm employment was 139,000, which was slightly revised up to 144,000 when the June data was released, but this time it was significantly revised down to 19,000, resulting in a total downward revision of 258,000 for the newly added non-farm employment in May and June. Generally speaking, due to certain delays in survey collection, non-farm data is usually adjusted in the following two months after its initial release. Regarding the reasons for the revisions in May and June, the U.S. Department of Labor stated that one reason was the recovery of additional survey results after the initial data release, and the other was the impact of recalculating seasonal adjustment factors.

The reason this revision has attracted widespread attention from the market is primarily due to the exaggerated total and structural changes.

On the total level, the newly added non-farm employment data for May and June was revised down by 125,000 and 133,000, respectively. If we measure the adjustment magnitude by the absolute value of the revised number/initial published value, the adjustment magnitudes for May and June reached 86.33% and 90.48%, respectively, marking the largest correction since 2021.

On the structural level, this revision reflects a simultaneous weakening of employment in both the private and government sectors in the U.S. Before the revision, the newly added non-farm employment in May was primarily driven by the private sector, with government employment relatively weak. In June, private employment fell while government employment rose, creating a certain hedge. After the revision, we see that the significant downward revision of non-farm data is not due to anomalies in individual sectors but shows a widespread decline. For example, in June, the newly added employment in the private sector was revised down from 74,000 to 3,000, with significant declines in cyclical industries such as retail and leisure and hospitality. Government employment was also revised down from 73,000 to 11,000, with both state and local government employment weakening.

We believe that the impact of this non-farm data may be quite profound and could significantly disrupt the macro trading logic for the third quarter.

First and foremost, has the Federal Reserve acted too late? The timing of the release of this non-farm data is quite special, as on July 31, the Federal Reserve held a meeting to announce that it would maintain the federal funds target rate unchanged, which aligned with market expectations. However, Powell's subsequent statement was relatively hawkish, not providing clear guidance on a rate cut in September, pointing out that commodity inflation is rising, and a relatively balanced labor market may constitute the Federal Reserve's strongest justification for holding steady However, looking at the revised labor data, the non-farm payrolls in the U.S. for May to July were 19,000, 14,000, and 73,000 respectively, with an average of only about 35,000 over the three months. Excluding the special circumstances of 2020, the last time similar employment data appeared was during the financial crisis from 2008 to 2010. From an absolute scale perspective, the employment data for May to July can hardly be considered balanced; it can even be deemed quite poor. If the Federal Reserve were to see the revised employment data before the interest rate meeting, we believe that the number of Fed officials supporting a rate cut in July may not be limited to just Waller and Bowman. If we take the latest non-farm employment data as the objective reality, we might conclude that the Fed's actions have come too late.

The pace of Fed rate cuts is expected to accelerate significantly. Following the release of the non-farm data, market expectations for Fed rate cuts have undergone a substantial adjustment, with the probability of a 25 basis point cut in September jumping from 37.66% to 80.31%. A rate cut by the Fed in September has likely become a high-probability event. Meanwhile, according to cumulative probability calculations, the market has fully priced in at least one 25 basis point rate cut by the Fed before October, and there are signs of an upward revision in the forecast for the total number of rate cuts for the year. The pace of Fed rate cuts is likely to accelerate significantly.

In this context, we believe the following two questions may be more worthy of attention:

First, will the Fed's "compensatory" rate cut reoccur in September 2024? In September 2024, the Fed initiated this round of rate cut cycle with an unexpected 50 basis point cut, which the market generally believes included some compensation for the rate cut that should have occurred in July 2024 but did not. The current situation has both similarities and differences compared to the third quarter of 2024.

The similarity lies in the fact that the marginal weakening of the labor market is the main reason for the rising rate cut expectations. On August 3, 2024, the U.S. Department of Labor unexpectedly reported a disappointing non-farm payroll figure for July, with the unemployment rate rising by 0.2 percentage points to 4.3%, pushing the Sam Rule indicator to 0.53%, exceeding the 0.5% warning line, indicating that the U.S. economy may be heading into recession. As a result, the macro trading theme in the third quarter of 2024 revolved around the interplay of rate cut expectations and recession expectations. The Fed also primarily considered the marginal weakening of the labor market and unusually chose to initiate this round of rate cut cycle with a single 50 basis point cut. On August 1, 2025, U.S. non-farm data again disappointed; although the unemployment rate did not spike significantly, the substantial downward revision of non-farm payroll figures may also indicate a certain degree of deterioration in the U.S. labor market, becoming a major factor in the rising rate cut expectations The difference lies in the fact that the current inflation environment imposes certain constraints on the Federal Reserve's interest rate cuts. In the third quarter of 2024, U.S. inflation is generally in a downward trend, with the year-on-year growth rate of the U.S. CPI continuing to slow from March to September 2024, providing a relatively favorable macro environment for the Federal Reserve to cut interest rates. The current U.S. inflation environment is relatively unclear, as Trump's tariff policy will officially take effect on August 7, with most countries facing reciprocal tariffs ranging from 10% to 20% on U.S. export products. The inflation structure data from June also shows that the impact of the tariff policy on core goods prices such as furniture and clothing is gradually becoming apparent. The Federal Reserve has continuously paused interest rate cuts since January 2025, with the risk of secondary inflation being a core consideration.

From the current data alone, we believe it may not be sufficient to drive the Federal Reserve to implement "compensatory" interest rate cuts again, but potential paths may have already emerged. Before the Federal Reserve's September monetary policy meeting, officials will still see an August labor data report and inflation data for July and August. If the labor market further deteriorates in August, and the subsequent inflation rebound is relatively low or turns downward again, the resonance of both may drive the Federal Reserve to make a more accommodative monetary policy decision.

Secondly, should we trade on interest rate cuts or on recession? After the U.S. non-farm payroll data was released on August 1, expectations for a September interest rate cut by the Federal Reserve rose significantly, but the U.S. stock market experienced a substantial adjustment, with the Nasdaq falling by 2.24% in a single day, marking the largest single-day decline since May. The reason for this is that moderately weakening labor data can benefit the U.S. stock market through interest rate cuts, but if labor data weakens significantly, it may raise investor concerns about whether the U.S. economy will fall into recession, thereby driving down risk appetite and adversely affecting risk assets such as U.S. stocks.

The U.S. economy may still be some distance from recession. With the "import rush" effect from the first quarter receding, the U.S. GDP growth rate for the second quarter rebounded from -0.5% to 3.0%, showing a seesaw state of marginal recovery in net exports and weakening investment. Domestic consumption is the core driving force of U.S. economic growth, and the year-on-year growth rate of total domestic purchases in the second quarter fell to 1.9%, showing a certain degree of decline compared to the previous several quarters, which maintained a central level of around 3% Retail and food service sales overall maintained a certain level of resilience, recording a year-on-year growth of 3.9% in June. Considering the downward revision of non-farm data and the marginal rebound in inflation, we believe that the most appropriate definition of the current U.S. economy is that growth momentum is marginally weakening, which is not sufficient to conclude that the U.S. economy is about to enter a recession.

The significant downward revision of non-farm data may subtly impact the data-driven investment research analysis system. On a direct impact level, the substantial downward revision of non-farm data has led to a significant increase in the expectation of a rate cut by the Federal Reserve in September, and the pace of rate cut trading may be anticipated to advance. However, looking deeper, the non-farm data for May and June both experienced substantial adjustments of nearly 90%, which may impact the current financial investment research analysis system based on objective data. If the underlying data is no longer credible, the research results derived from it may also be grossly misleading.

In the short term, ADP employment data may serve as a "substitute" for non-farm data. From the perspective of final results, prior to the release of non-farm data for May and June, ADP employment data provided certain warning signals. The new jobs added by ADP in May and June were 29,000 and -23,000, respectively, which corresponded relatively well with the revised non-farm data of 19,000 and 14,000. At that time, the market placed much greater emphasis on non-farm data than on ADP employment data, so after the relatively good initial non-farm data was released in May and June, the market was more willing to believe that the labor market was still in a relatively balanced state. After this significant adjustment of non-farm data, we believe that the market's attention to ADP employment data may further increase, especially when discrepancies arise between the two sets of data. Whether to trust ADP or non-farm data may become a dilemma for market investors and even Federal Reserve officials.

In the long term, the impact of the issue of data credibility may gradually ferment. After the release of non-farm data, Trump indicated that he had instructed officials to fire the head of the U.S. Bureau of Labor Statistics, McIntosh, which may carry the implication of using him as a "scapegoat" to soothe market sentiment. The formation of trust is gradual, while the breaking of trust is abrupt. For the upcoming non-farm data in August, regardless of whether the data itself is good or bad, it may raise reasonable doubts among investors about its authenticity. This may add more disorder and noise to the financial market, and the pricing of financial assets such as equities and bonds may be affected to varying degrees. As the best tool for hedging against chaos, gold, combined with rising expectations of rate cuts and potential inflation risks, may further highlight its value.

Authors: Qin Han, Cui Zhengyang, Source: Qin Han Research Notes, Original Title: "Exploring the Underlying Logic of the Significant Downward Revision of Non-Farm Data"

Risk Warning and Disclaimer

The market has risks, and investment requires caution. This article does not constitute personal investment advice and does not take into account the specific investment goals, financial conditions, or needs of individual users. Users should consider whether any opinions, views, or conclusions in this article are suitable for their specific circumstances. Investing based on this is at one's own risk