Labor Productivity Is Not Your Performance Review
Labor productivity is the subject of this series, and its first question is easy to state but easy to miss: what does “more productive” mean when the number describes an entire sector rather than one person? The answer supports a practical judgment. Output per hour can tell us something important about how an economy is producing goods and services, but it cannot by itself tell us whether a particular employee worked hard, whether a manager organized a team well, or whether the gains were shared fairly.
That distinction matters because the same word is doing two jobs. In a workplace, a performance review may ask whether someone met goals, solved problems, supported colleagues, or exercised sound judgment. Labor productivity, as measured by the U.S. Bureau of Labor Statistics, describes the relationship between real output and the labor time used to produce it. It is an aggregate accounting measure. Treating it as a personal verdict is not merely unkind; it changes the question being answered.
The same word, two different judgments
Suppose a company produces more units in a quarter. It might have installed better equipment, reduced downtime, redesigned a workflow, received stronger orders, or benefited from a change in the mix of products. Employees may also have learned new skills or worked with greater care. If hours worked rose more slowly than real output, output per hour would increase. The statistic would record the result, but not assign credit among those causes.
The reverse is also possible. Employees may be doing excellent work while measured productivity falls. A factory can keep its staff during a temporary drop in orders. A service organization can spend time training people for future work. A business can encounter an energy disruption, a shortage of materials, or a production bottleneck. The hours remain in the denominator while current output weakens. The measure then falls, even though the immediate problem is not a sudden collapse in individual effort.
BLS makes the scope explicit. For the nonfarm business measure, labor productivity is calculated by dividing an index of real output by an index of hours worked by all workers in the sector. Those hours include employees, proprietors, and unpaid family workers. Productivity growth reflects the joint effects of technology, capital investment, the level of output, the use of capacity, energy and materials, the organization of production, managerial skill, and the characteristics and effort of the workforce.
“Joint effects” is the crucial phrase. The measure is not designed to separate every contribution. It tells us how much real production is associated with an hour of labor in a defined part of the economy. That is useful precisely because it captures the performance of a production system. It is also limited precisely because a system result cannot be cleanly converted into a personal score.
The word “real” matters as well. A business can collect more dollars because prices rose, without producing more goods or services. A productivity measure tries to express output in a way that separates production from price changes through a real-output index. But even a real-output increase is not a complete account of experience. It does not, by itself, tell us whether quality improved, whether the work became safer, whether customers value the result more, or whether the benefits reached the people who produced it. Those are important questions, but they require additional measures and evidence.
The denominator is not a head count. Two sectors could employ the same number of people and record different hours, or keep the same number of people while changing schedules, overtime, part-time work, or training time. Because the measure uses hours worked, not simply the number of jobs, those changes affect output per hour. The inclusion of proprietors and unpaid family workers also means that the statistic is built for a sector's labor input, not for the narrower question of how much a salaried employee accomplished during a shift.
Time and revision add a further boundary. A quarterly percentage change in the BLS release is presented at a compound annual rate, while a year-over-year change compares corresponding quarters. The figures are calculated from indexes and rounded for presentation. More complete source data can later change the estimate. A responsible reader therefore treats a reported rate as a carefully defined observation at a point in the statistical process, not as a timeless fact about the people or organizations inside the sector.
The boundary is not an argument against measuring productivity. It is an argument for measuring the right thing. A restaurant manager might use a team-level output measure to investigate scheduling, equipment, service design, or training. That can be a legitimate operational exercise. But it still requires more evidence than a national productivity rate. The national statistic cannot reveal which employee should be promoted, which team member caused a delay, or whether a workplace's targets are sensible.
Why the distinction changes what the number can mean
The boundary becomes more important when a productivity figure is connected to wages, prices, or employment. These relationships are real, but they are not interchangeable. Higher productivity can help a producer make more output with a given amount of labor time. Unit labor costs, however, compare hourly compensation with labor productivity. If compensation rises, unit labor costs tend to rise; if productivity rises, unit labor costs tend to fall. Real hourly compensation asks what workers can buy after accounting for consumer prices. Labor share asks how much of output accrues to workers as compensation. None of these is a personal performance rating.
The latest BLS release illustrates why context matters. In the second quarter of 2026, revised nonfarm business labor productivity increased 1.4 percent from the previous quarter at a seasonally adjusted annual rate. Real output increased 1.7 percent, while hours worked increased 0.3 percent. The output increase was faster than the hours increase, so output per hour rose. From the same quarter a year earlier, productivity increased 2.2 percent. Those are both valid descriptions, but they answer different questions: one describes a recent quarter using an annualized rate, while the other compares two corresponding quarters.
The same release reports that nonfarm business unit labor costs increased 1.2 percent at an annual rate in the quarter, reflecting a 2.6 percent increase in hourly compensation and the 1.4 percent productivity increase. Real hourly compensation decreased 3.3 percent in the quarter. The figures can coexist without contradiction. Producers can face higher compensation costs per unit while workers' purchasing power falls, because nominal compensation, productivity, and consumer prices are different parts of the accounting picture.
The report also puts the quarterly result beside a longer view. From the fourth quarter of 2019 through the second quarter of 2026, nonfarm business productivity grew at a 2.1 percent annualized rate, based on 2.5 percent output growth and 0.4 percent hours growth. That current-business-cycle rate was above the previous cycle's 1.5 percent and equal to the long-term rate since 1947. A single quarter, a year-over-year comparison, and a multi-year trend should not be compressed into one undifferentiated statement that “productivity is up.”
These distinctions affect public arguments. If productivity rises, it does not follow that every worker's pay should rise immediately or equally. Prices, bargaining, labor-market conditions, wage-setting institutions, and the distribution of profits all matter. If productivity falls, it does not follow that workers became less diligent. Output may have weakened, hours may have been retained, or a sector may be operating under a different production technology. The number is a signal about the relationship between output and labor time, not a complete explanation of either.
A wider question than “who worked harder?”
The evidence also warns against treating a national average as a universal experience. In the same BLS release, manufacturing productivity increased 2.4 percent in the second quarter of 2026 as output rose 5.4 percent and hours rose 2.9 percent. But manufacturing productivity grew only 0.5 percent at an annualized rate over the current business cycle, well below its long-term rate of 2.1 percent. BLS notes that the concepts and methods used to construct manufacturing output differ from those used for business and nonfarm-business output, so the measures are not directly comparable without qualification.
The OECD adds another layer. Its 2026 Employment Outlook reports that employment and labor-force participation remained at historic highs while unemployment and labor-market weakness showed signs of rising. It also finds that labor productivity growth across the OECD averaged 1 percent annually from 2019 to 2025, compared with 2 percent from 1995 to 2019. In Europe, labor hoarding—firms retaining staff while output or hours adjust—helped keep employment resilient but mechanically reduced output per hour. Structural shortages could persist at the same time because population ageing, digital change, the energy transition, and poor job quality affect the supply and allocation of labor over a longer horizon.
That is why this series will keep returning to the relationship rather than the label. First comes the construction of the measure: what counts as real output, whose hours are counted, and which time comparison is being used. Then comes the practical test of reading a revised U.S. release. After that, productivity can be placed beside compensation, unit labor costs, labor share, and profits; beside employment, vacancies, and labor hoarding; and beside the uncertain evidence about young entrants and artificial intelligence. Each connection changes the interpretation, but none erases the subject's boundary.
For now, the essential rule is simple. When someone says that labor productivity rose or fell, ask: in which sector, over which period, measured as what kind of output, against whose hours, and with what other conditions changing? Only after those questions are answered should the number enter a conversation about wages, prices, jobs, or workplace decisions. The next step is to examine the ratio itself—what output per hour actually measures—so that the statistic can remain a tool for understanding production rather than a shortcut for judging people.