How to Read a Productivity Release
A labor-productivity release is not a verdict about workers; it is a compact report about the relationship between real output and labor time in a defined sector. This chapter uses the revised U.S. release for the second quarter of 2026 to show how to read that relationship as evidence. The goal is not to memorize the figures, but to learn how a headline changes meaning when its period, denominator, sector, revision status, and companion measures are made visible.
A good reading starts with the number that attracts attention and then immediately moves behind it. What happened to output? What happened to hours? Is the rate quarterly and annualized, or a comparison with a year earlier? Does the measure cover nonfarm business, manufacturing, or nonfinancial corporations? Was the estimate preliminary or revised? Each question narrows the possible interpretation. Without them, a precise-looking percentage can invite an imprecise story about individual effort.
Read the headline as a decomposition
The revised BLS release reports that nonfarm-business labor productivity increased 1.4 percent in the second quarter of 2026 from the previous quarter. That sentence contains three boundaries. “Nonfarm business” names the sector. “From the previous quarter” names the comparison. “At a seasonally adjusted annual rate” describes the convention used for the quarterly percentage change. The figure does not mean that output per hour literally rose 1.4 percent during the three months, nor does it describe every business, job, or person in the United States.
The release gives the components that make the result intelligible. Real output increased 1.7 percent from the previous quarter at the annual rate, while hours worked increased 0.3 percent. Because output grew faster than hours, real output per hour increased. The components are not background decoration. They are the numerator and denominator of the reported relationship. If output had grown more slowly than hours, the productivity result would have been lower; if hours had fallen while output held steady, the result could have been higher for a different reason.
This is the first translation a reader should be able to make: “In the defined nonfarm-business sector, the latest seasonally adjusted quarterly annualized rate indicates that real output grew faster than hours worked.” That sentence is longer than the headline, but it is also more honest about what the data show. It describes a measured relationship without claiming to know whether a new technology, better management, stronger demand, or greater employee effort was responsible.
The year-over-year comparison changes the window. From the same quarter a year earlier, nonfarm-business labor productivity increased 2.2 percent, while output increased 2.5 percent and hours increased 0.2 percent. This comparison accumulates the difference across four quarters. It is useful for reducing the weight placed on one quarter, but it is not a substitute for the quarterly annualized rate. A reader who says “productivity rose 2.2 percent in the second quarter” without saying “from a year earlier” has removed the time boundary that makes the claim interpretable.
The longer comparison changes the question again. From the fourth quarter of 2019 through the second quarter of 2026, nonfarm-business productivity grew at an annualized rate of 2.1 percent. BLS attributes that relationship to 2.5 percent annualized output growth and 0.4 percent annualized growth in hours. The current-business-cycle rate is above the 1.5 percent rate of the previous business cycle and equal to the long-term rate since the first quarter of 1947. This tells us how the period fits into a broader pattern. It does not tell us that every quarter was equally strong or that the same cause operated throughout the period.
A release therefore contains several valid stories about time, and they should not be collapsed. The quarterly rate is sensitive to the latest movement. The year-over-year rate describes a four-quarter distance. The business-cycle rate places a period beside an earlier cycle and a long-run benchmark. An index level records where the series stands relative to its base. A careful reader chooses the comparison that matches the question instead of selecting whichever percentage sounds most dramatic.
The annualization convention deserves special attention because it can make a modest quarterly movement look large. If a ratio rose by roughly 0.35 percent in one quarter and that pace continued for four quarters, compounding would produce something close to a 1.4 percent annual rate. The reported rate is a standardized way to express pace, not a claim that the entire annualized amount occurred at once. Seasonal adjustment and annualization help analysts compare movements, but they do not remove the need to say what period was actually observed.
The number also needs a causal boundary. The BLS definition says productivity growth reflects the joint effects of technology, capital investment, output levels, capacity use, energy and materials, production organization, managerial skill, and workforce characteristics and effort. The release does not identify a unique contribution for each influence. Thus, a productivity increase can be evidence that the output–hours relationship improved, but it is not evidence by itself that workers became more diligent or that one management decision caused the change.
The revised nature of the number matters too. The current release uses more recent source data from BLS, the Bureau of Economic Analysis, and the Federal Reserve Board than were available for the preliminary report. That information can change output, hours, compensation, or cost estimates. A headline should therefore be read as the current estimate of a defined relationship, not as a permanent inscription on the period being measured.
BLS describes a regular revision sequence. The first estimate is published within about 40 days after the reference quarter, the second approximately 30 days later, and the third approximately 60 days after the second. Looking at reference quarters from the first quarter of 2001 through the first quarter of 2026, BLS found that the third estimate of quarterly nonfarm-business productivity growth differed from the first by between minus 1.1 and plus 1.4 percentage points about 80 percent of the time. That historical range is not a prediction for the current quarter. It is a reason to avoid treating an early decimal place as more stable than it is.
A revision is not automatically evidence that the earlier estimate was careless. It is a consequence of constructing an economic measure from data that arrive at different times and are themselves updated. The interpretive question is not “Which number is the real one?” as if one release must be fraudulent. It is “Which estimate is being quoted, what information did it use, and how much precision is appropriate for the decision at hand?” That question is useful in a newsroom, a policy discussion, and a workplace conversation alike.
Compare sectors without flattening them
The same reading method changes the meaning of the sector figures. In manufacturing, labor productivity increased 2.4 percent in the second quarter of 2026 at a seasonally adjusted annual rate. Manufacturing output increased 5.4 percent and hours worked increased 2.9 percent. Durable manufacturing productivity increased 3.6 percent, while nondurable manufacturing productivity increased 2.1 percent. These figures describe a strong quarterly output–hours relationship within manufacturing, but they should not be used as a shorthand for the entire economy.
The longer manufacturing comparison points in a different direction. From the fourth quarter of 2019 through the second quarter of 2026, manufacturing productivity grew at an annualized rate of only 0.5 percent. Manufacturing output grew at a 0.2 percent rate while hours worked declined at a 0.3 percent rate. The current-cycle rate was above the previous cycle's 0.1 percent but below the manufacturing long-term rate of 2.1 percent since the first quarter of 1987. A single quarter can therefore be stronger than a multi-year average without contradicting it. The time horizon is doing real analytical work.
There is also a measurement boundary. BLS says that the concepts, sources, and methods used for manufacturing output differ from those used in the business and nonfarm-business output series. Those output measures are not directly comparable. The warning is easy to overlook because the labels look similar and the tables place the percentages near one another. But a manufacturing output index is constructed for a different production boundary. A comparison can reveal sector differences, yet it cannot serve as a clean ranking of the people working in those sectors.
The durable and nondurable figures illustrate why even one sector contains different production stories. Durable manufacturing recorded output growth of 8.9 percent and hours growth of 5.1 percent, producing a 3.6 percent productivity increase. Nondurable manufacturing recorded output growth of 1.4 percent and a 0.6 percent decline in hours, producing a 2.1 percent productivity increase. The durable result was driven by much faster output growth; the nondurable result combined modestly higher output with fewer hours. The rates alone conceal that difference.
The release also reports a preliminary measure for nonfinancial corporations. Productivity increased 2.2 percent in the second quarter as output increased 3.9 percent and hours increased 1.7 percent. Unit profits of nonfinancial corporations increased at an annualized rate of 43.0 percent. This is an important companion observation, but “preliminary” is part of its meaning. It should not be treated as a settled description of all corporations, nor should a unit-profit movement be translated into a claim about the productivity or compensation of an individual employee.
Sector comparison is most useful when it asks a bounded question. Did output grow faster than hours in the sectors being measured? Was the result a recent-quarter movement or a longer trend? Are the output concepts constructed in comparable ways? Are the estimates revised or preliminary? Those questions can reveal heterogeneity without turning a table into a league ranking.
The current release shows why a national productivity headline can be both informative and incomplete. Nonfarm-business productivity increased 1.4 percent in the latest quarterly annualized comparison and 2.2 percent year over year. Manufacturing increased 2.4 percent in the quarter but had a much weaker current-cycle rate. Nonfinancial corporations recorded a preliminary 2.2 percent productivity increase alongside a large unit-profit movement. The figures are not competing answers to one question. They are answers to related questions about different production boundaries and time windows.
A useful report of the evidence would therefore include the components and the limit: “Nonfarm-business labor productivity increased in the revised second-quarter estimate because real output grew faster than hours worked; manufacturing also increased in that quarter, but its output measure is constructed differently and its current-cycle trend is weaker. These figures show sector-level relationships, not the cause of any individual performance.” That formulation is less sensational than a claim that workers have become more or less productive, but it preserves more of the truth.
The final test is to notice what the release cannot establish. It cannot identify whether technology, capital, organization, demand, energy, management, or worker effort contributed most to the quarter's movement. It cannot show whether the output was safer, better, or more fairly distributed. It cannot turn a sector average into a performance review. It can show a carefully defined relationship, provide the components needed to inspect it, and offer comparisons across periods and sectors—subject to the construction and revision limits stated by BLS.
That is enough to make the release valuable. A statistic becomes more useful, not less, when its scope is explicit. Once productivity has been separated from personal judgment and read alongside its output, hours, sector, and revision boundaries, the next question is what it has to do with compensation, unit labor costs, labor share, and profits. Those measures are connected to the production relationship, but they answer different questions about its economic consequences.