On September 3, the U.S. Bureau of Labor Statistics released revised figures for productivity and costs for the second quarter: labor productivity in the non-agricultural business sector increased by 1.4% on a quarterly annualized basis, driven by a 1.7% increase in actual output and a 0.3% increase in working hours. Compared to the same period last year, productivity grew by 2.2%. Meanwhile, unit labor costs rose by 1.2% on a quarterly annualized basis, as hourly wages increased by 2.6%, which was higher than the rate of productivity growth; unit labor costs have accumulated an increase of 1.4% over the past four quarters. These data do not tell a simple story of "increased corporate efficiency necessarily leading to decreased costs," but rather show that output, working hours, and wages are changing at different rates.

Productivity refers to the change in actual output per hour, while the cost of labor per unit roughly reflects the labor remuneration required to produce one unit of output. As long as the growth rate of hourly remuneration exceeds productivity, the cost of labor per unit will continue to increase. The 1.4% increase in productivity in the second quarter mitigated some of the wage costs, but it did not completely offset the 2.6% increase in hourly remuneration. Therefore, the labor cost pressure on businesses continues to rise, although the rate of increase is not equivalent to consumer prices, and it does not necessarily imply that inflation will accelerate in any given month.

Efficiency recovery occurs simultaneously with a decline in workers' purchasing power.

The most striking figure in the report is that the actual hourly wage has decreased by 3.3% on a quarterly to annual basis, and by 0.1% in the past four quarters as well. The fact that actual wages take into account consumer price factors indicates that the increase in nominal wages has not fully translated into an improvement in purchasing power. The share of labor income dropped to 52.8% in the second quarter, the lowest level since the series began in 1947. Increased productivity is generally considered the foundation for the joint growth of wages and profits, but the distribution of these gains does not occur automatically; in the short term, prices, profits, industry structure, and employment composition can all lead to more of the output benefits remaining with non-labor income.

Starting from the fourth quarter of 2019, the current business cycle has seen an average annual growth of 2.1% in non-agricultural business productivity, with output growing at an average annual rate of 2.5% and hours worked increasing by 0.4%. This rate is higher than the 1.5% seen in the previous cycle from the fourth quarter of 2007 to the fourth quarter of 2019 and is comparable to the long-term average of 2.1% since 1947. In other words, recent efficiency performance is indeed better than in the previous cycle, but it is not yet possible to assert that the economy has entered a permanently higher trajectory of productivity based on the data from just the second quarter. Quarterly data can be revised, and changes in capital investment, industry shifts, and cyclical adjustments in working hours can also affect short-term readings.

In the second quarter, manufacturing productivity increased by 2.4%, output grew by 5.4%, and working hours increased by 2.9%. The increase in manufacturing output was the largest since the second quarter of 2021. It is also important to distinguish between "rapid output growth" and "achieving more output with less labor": since working hours also increased significantly, the final increase in output per hour was less than the total output increase. To determine whether a company's expansion is healthy, it is necessary to consider not only factory production but also whether the additional working hours, compensation, and unit costs are sustainable.

Revised values have improved measurements, but they have not eliminated uncertainty.

This release represents the revised figures for the second quarter, not the initial values. BLS has recalculated based on the updated output and labor hour data, and reminds that there is a significant range of possible revisions to the quarterly productivity estimates. Official historical experience shows that there is about an 80% probability that the preliminary estimate of non-farm business productivity will fall within the range of minus 1.1 to plus 1.4 percentage points when revised in subsequent years. Therefore, 1.4% should be considered the current best estimate, rather than a figure that is precise to one decimal place and will not change further.

The annualized rate is also often misinterpreted. A quarterly growth of 1.4% in the second quarter means that if this growth rate is maintained for four consecutive quarters, the annual growth would roughly be at this level; it does not mean that the second quarter itself was 1.4% higher than the first quarter. Similarly, a 3.3% decrease in actual hourly wages, when expressed as an annualized quarterly rate, should be considered in conjunction with year-over-year changes, employment levels, and total income when assessing household purchasing power. Mixing different measures together can easily exaggerate the significance of a single quarter's data.

It should also be noted that the non-agricultural business sector does not cover the entire economy. Government, non-profit organizations, and some activities that are difficult to measure are not included in this core measurement. However, the hours worked by self-employed individuals and unpaid family laborers are taken into account. The output indicators are also derived from estimates of the national accounts, rather than being calculated on a piece-rate basis for each employee. This method is suitable for observing changes in macroeconomic efficiency, but it cannot directly answer whether employees of a particular company or position are “more efficient.” Changes in industry structure, such as an increase in the proportion of high-productivity industries, can also raise the overall figures.

The cost of labor per unit is also not the total cost for a company. Energy, raw materials, interest, rent, and depreciation can all change in different directions. Companies may absorb labor costs by reducing profit margins, or they may adjust prices, reduce hiring, or increase investment in automation. To determine its impact on inflation, it is necessary to consider producer prices, corporate profits, and the strength of demand, rather than mechanically adding a 1.2% annualized growth rate to consumer price forecasts.

In terms of macroeconomic judgment, this revised report supports two seemingly contradictory phenomena: while the hourly output of enterprises continues to increase, the cost per unit of labor is still rising; nominal wages for workers have increased, but after deducting inflation, their real wages have actually decreased. The key question going forward is not only whether productivity can be maintained, but also whether the benefits of growth are passed on to households through actual wages. The preliminary figures for the third quarter are scheduled to be released on November 5th. Until then, the data for the second quarter are more appropriately seen as a cross-section of efficiency, costs, and distribution relationships, rather than providing a single conclusion regarding AI, wages, or inflation.