The powerhouse behind the strong September non-farm payrolls: record seasonal adjustments

The U.S. September non-farm payroll growth data released this Friday initially appears very strong. In September, the addition of non-farm employment was 254,000, the largest monthly increase in half a year, exceeding economists' consensus expectations by more than 150,000, an increase of 69.3%. However, a detailed analysis of the data reveals that such a sharp increase is largely due to data adjustments by U.S. government agencies.

Why the adjustment? In theory, when compiling employment reports, the Bureau of Labor Statistics (BLS) of the U.S. Department of Labor adjusts the collected raw data to better account for seasonal factors, such as summer fluctuations in teacher vacations, booming construction industry jobs, and increased hiring in restaurants, making month-to-month comparisons easier.

However, the seasonal adjustment in September this year set a new record. Bloomberg reported that if the seasonally adjusted employment is divided by the unadjusted employment, the result is 1.0. This is the largest seasonal factor calculation value since September 2002.

Derek Holt, an economist at Scotiabank, stated in a report that in recent years, September has seen historically significant seasonal adjustments, indicating that we are in a different era where BLS's seasonal adjustment of employment data has changed. This means that we should pay attention to the employment growth in the report and smooth the data over time.

Stephen Stanley, Chief U.S. Economist at Santander Capital Markets, a subsidiary of Santander Bank in Spain, stated that the growth in non-farm employment in September was mainly driven by the food service, construction, and retail industries. The strong growth in these industries "suggests that some combination of favorable weather and/or seasonal factors that brought tailwinds is unlikely to persist."

Chris Low, Chief Economist at FHN Financial, pointed out in the report that September is a difficult month for BLS because there are significant changes in the composition of the workforce at the end of summer and the beginning of autumn. "Any seasonal distortions in this report represent an overcompensation for the opposite seasonal distortions in the previous months, so it can be concluded that either this report is strong, or the previous reports were not as weak as previously thought. In any case, stronger than expected is the correct description."

Financial blog Zerohedge pointed out an eerie point. Considering the return of teachers to school, the number of unadjusted government workers in September usually increases significantly, and BLS typically smooths this growth with seasonal adjustments. However, this was not the case in September this year.

As shown in the chart below, historically, regardless of whether the unadjusted number is high or low, the adjusted number in September has always increased relatively moderately. However, this September, the number of unadjusted government workers surged by 1.322 million, setting a new high for government workers after adjustment in September After seasonal adjustments, the government workforce surged by 785,000 in September, the largest monthly increase in government employees on record, excluding the anomaly in June 2020 caused by the COVID-19 outbreak.

While there was a significant increase in government workers, private sector employment only grew by 133,000 in September, a figure that may be more reliable. However, it indicates that the recent labor market slump is still ongoing.

Not only household surveys but also institutional survey data showed a sharp increase in government employment: the unadjusted number of government workers increased by 918,000, while the unadjusted number of private sector workers decreased by 458,000.

Why the significant increase in government employment? Zerohedge believes it is to bring the unemployment rate down to 4.1% instead of rising to 4.5%, as triggering the Sam Rule predicting an economic recession, and to prevent a market collapse. After all, this is crucial data released a month before the U.S. election. BLS's motivation is clear, aiming to increase "government positions" as much as possible, not only to maintain continuous growth in the labor force but also to further reduce the unemployment rate.

In addition to seasonal adjustments, some analysts believe that the unusually strong non-farm employment report is also influenced by the response rate of a crucial component in the report - the Establishment Survey. The Establishment Survey in the report mainly focuses on the survey of employment data in enterprises and government departments, while the Household Survey collects employment information by surveying individual household members.

In the September report, when the initial estimate was made, only 62.2% of businesses submitted responses on time, the lowest percentage since September 2002. Looking at two subsequent surveys over a period of time, the response rate is usually close to around 95%, so the revised data may fluctuate significantly.

Samuel Tombs, Chief U.S. Economist at Pantheon Macroeconomics, believes that the alarmingly low response rate in the employment survey is a cause for concern. The response rate from small businesses is too low, and the extent of layoffs in small businesses exceeds that of large enterprises. Therefore, it is believed that the September employment data to be released in the coming months will see a significant downward revision