700,000 Jobs Lost? The Truth Behind May’s Jobs Report

Last Friday’s data showed a 139,000 increase in May nonfarm payrolls, slightly above expectations, with the unemployment rate steady at 4.2%. On the surface, the report seemed quite positive—cheered by equity markets.

But is the truth really as good as the numbers suggest?

The U.S. Bureau of Labor Statistics previously announced that it would revise the April employment report when releasing May data. The reason: a minor error introduced by a redesign in the weighting of the Current Population Survey (CPS) sample. The BLS emphasized that the correction would not affect key indicators like the unemployment rate.

However, the actual impact was far from minor. After applying the new methodology, the total U.S. labor force shrank by 625,000 people in just one month.

This massive reduction has created a structural distortion in the headline figures and misled the market.

For example, such an adjustment can “artificially suppress” the unemployment rate:

Unemployment rate = unemployed population ÷ labor force.

When the labor force shrinks while the number of unemployed remains stable or rises slightly, the rate can mechanically stay flat or even fall—masking real labor deterioration. This time, while the labor force dropped by 625,000, the number of unemployed rose by just 71,000, keeping the unemployment rate at 4.2%.

The real situation may well be that around 700,000 people lost their jobs, most of whom were simply “excluded” from the labor force under the revised definition. This is a classic case of statistical distortion, and perhaps the most misleading part of the May report.

“Economic Barometers” Flash Red

Within the nonfarm report, several structural risk indicators—considered key “economic barometers”—painted a grim picture.

1. Decline in Temporary Help Employment (-20,200):

These roles are typically the most sensitive to future labor demand. A sharp contraction is a leading indicator of recession.

Temp workers are like “canaries in the coal mine”: before cutting full-time staff, businesses usually cut temps first. This continued drop signals a lack of confidence in future demand and corporate preparation for tough times ahead—recognized as one of the most reliable early indicators of economic downturn.

2. Negative Manufacturing Employment (-8,000):

Weakness in manufacturing directly reflects declining business investment and goods consumption—core drivers of the economy. As a cyclical sector, sustained contraction indicates weakening real economic activity.

3. Data Revisions (Downward by 95,000 jobs over prior two months):

This implies that recent economic conditions were worse than previously believed. It also undermines confidence in the current headline data and points to a continued downward trend.

To keep the unemployment rate stable, the U.S. needs to generate 100,000–120,000 jobs monthly. The recent 3-month average gain of 135,000 is barely above that threshold. Any shock or revision could push job creation below the equilibrium point.

4. Extremely Fragile and Concentrated Job Growth:

The entire “growth” in jobs relies on just a few sectors—a deeply unhealthy structure.

Healthcare (+62,000) and Leisure & Hospitality (+48,000) together contributed nearly 80% of net job gains.

Conclusion

If June data are revised downward again, it would mark a rare “three consecutive downgrades”—unlikely to be random error, and more likely a delayed reflection of real labor weakening.

The fact that job growth is concentrated in non-cyclical sectors like healthcare and leisure, while sectors such as manufacturing, government, and temp help are contracting, suggests declining job quality.

In reality, “pre-recession” risk is already surfacing. Continuous monthly revisions of nearly 100,000 jobs, coupled with structural weakness, suggest the labor market is approaching a recession edge. If job creation fails to rebound or the unemployment rate rises in coming months, it would signal a real recession.

The balance of risk in the U.S. economy is tilting toward recession.