The missing denominator in a wage headline
A business can pay more per hour without experiencing an equally large increase in labor cost per unit of production. If an hour produces more output, higher compensation is spread across more units. That distinction is the reason labor productivity and unit labor costs belong together. Labor productivity compares real output with hours worked; unit labor cost compares labor compensation with real output. Neither is simply a measure of how hard an individual employee works. [1][2]
The denominator also explains why an employment report cannot answer the same question. Ten additional workers may increase production substantially, barely change it during training, or restore capacity after a disruption. Headcount does not reveal hours per worker, and hours do not reveal output per hour. Payroll growth, compensation growth and productivity growth describe different pieces of the same production process.
A worked example of the cost bridge
Consider a hypothetical operation producing 1,000 identical units in 100 hours. Compensation, including employer-paid benefits, is $40 per hour. Productivity is ten units per hour, total compensation is $4,000 and unit labor cost is $4. Assume the output measure is quality-consistent and no intermediate-input complications are present.
In the next period, compensation rises 5% to $42 an hour while productivity rises 3% to 10.3 units an hour. Unit labor cost becomes $42 divided by 10.3, or approximately $4.078. Its increase is 1.94%, not 5%. The exact growth factor is 1.05 divided by 1.03. Subtracting productivity growth from compensation growth gives a useful approximation of 2%, but the ratio provides the exact calculation. [4]
If productivity instead falls 2%, the denominator becomes 9.8 units per hour and unit labor cost reaches approximately $4.286, up 7.14%. The same wage increase now accompanies much stronger unit-cost growth. These examples are identities, not claims that changing pay causes productivity to rise or fall. They show the information needed before interpreting wage pressure.
What output and hours actually mean
BLS measures aggregate business productivity using real value-added output rather than simply adding company revenues. Value added avoids counting intermediate production repeatedly. The business-sector measure also excludes activities for which output or matching labor inputs cannot be measured adequately for this purpose. Nonfarm business is a defined statistical sector, not a synonym for every worker in the economy. [1]
Hours refer to work performed by the relevant population, including employees and appropriate self-employed workers. They are not interchangeable with paid hours or payroll headcount. Changes in average weekly hours can move total labor input even with unchanged employment. Output and hours are assembled from different sources, so revisions to either side can change the measured ratio.
In a hypothetical service business, a shorter average workweek can raise output per hour even while output per employee is unchanged. Conversely, a rush of overtime can raise total production but lower output per hour if the extra shifts face congestion or equipment constraints. Comparing matching sectors, time periods and output definitions is therefore a condition for meaningful interpretation, not a cosmetic statistical preference.
Productivity is broader than worker effort
A worker with better equipment may produce more in an hour while exerting the same effort. Better scheduling, a redesigned process, fewer equipment failures and more reliable materials can have similar effects. Labor productivity captures the combined outcome of these influences because it divides output by only one input. It does not allocate credit among workers, managers, technology and capital. [1]
Total-factor productivity asks a different question: how much output growth remains after accounting for measured combined inputs? In aggregate growth accounting, capital services and composition-adjusted labor input matter. At industry level, intermediate inputs can matter as well. Total-factor productivity is a residual within a measurement framework; it is not a direct meter of invention or software quality.
Suppose a firm doubles its machines per worker and produces 20% more per hour. Labor productivity has risen. Whether total-factor productivity rose requires evidence on the additional capital services and other inputs. Calling the entire 20% a technology breakthrough would skip the cost and quantity of the new equipment. Similarly, an aggregate productivity acceleration cannot by itself identify artificial intelligence as its cause.
How margins and real wages enter
Unit labor cost is measured against real output, while revenue per unit depends on the output price. If the illustrative operation sells each unit for $10 and its unit labor cost rises from $4 to $4.078 while its selling price is unchanged, labor absorbs a slightly larger share of revenue. Whether profit falls depends on materials, depreciation, financing, taxes and other costs too.
If the selling price instead rises to $10.30, labor cost can rise in dollars while falling as a share of sales. This is why a positive unit-labor-cost growth rate is not the same as a shrinking margin. The comparison must include prices, and a whole-economy measure cannot establish the margin outcome at a particular company.
Real compensation introduces another price index. A worker’s purchasing power depends on consumer prices, whereas the employer’s output price reflects what that employer sells. If hourly compensation rises 5% and consumer prices rise 3%, real hourly compensation rises about 1.94%. That calculation does not require employer output prices to rise 3%. Differences between consumer and producer price measures can separate measured productivity gains from real wage gains without implying a statistical contradiction. [2]
Why one quarter can mislead
Quarterly growth rates are often reported at annual rates. A 1% quarter-to-quarter increase annualizes to approximately 4.06% if compounded for four quarters. It does not mean output per hour already rose 4.06% during that quarter, nor that the next three quarters will repeat it. Four-quarter changes and annual-average changes answer different questions. [3]
Timing can produce sharp short-run swings. A business may retain staff as orders fall because replacing trained workers later is expensive. Output then falls faster than hours, weakening productivity. When demand returns, production may rebound before hiring, reversing the pattern. Such changes can be cyclical rather than a lasting change in productive capacity.
Composition also matters. If low-productivity activities lose hours faster than high-productivity activities, the aggregate average can improve even without any establishment becoming more efficient. Conversely, expanding a labor-intensive service can reduce aggregate productivity growth while meeting a genuine demand. An average is informative, but it does not describe every workplace.
Interpreting the release and its revisions
The sector and growth-rate convention determine the scope of a productivity claim. Real output, hours and hourly compensation explain its components, while the exact ratio links compensation and productivity to unit labor cost. Changes between release can reflect revised measurement rather than fresh economic activity.
An accounting observation does not establish a causal explanation. A productivity improvement can make a given compensation path compatible with slower unit-cost growth. That does not guarantee lower consumer inflation, higher profits, a particular wage bargain or a monetary-policy decision. The useful conclusion is narrower: the wage-inflation discussion needs an output denominator, consistent prices and enough history to distinguish a durable change from timing and composition.
Sources
- BLS Handbook of Methods, Productivity Measures: ConceptsOfficial sourceBack to text: ↑1↑2↑3↑4
- BLS Productivity glossaryOfficial sourceBack to text: ↑1↑2
- BLS Productivity and Costs release, March 24, 2026, methodology and annual-rate conventionOfficial releaseBack to text: ↑
- BLS What is unit labor cost?Official sourceBack to text: ↑