Fundamentals

Total Factor Productivity: What It Measures and Why It Matters

By Daniel Sardá · Published on October 14, 2026

7 min read1,361 words

TFP estimates the share of output not explained by measured quantities of capital and labor. It is useful, but its residual does not by itself identify the causes.

An economy can produce more because it employs more people, accumulates more machines, or makes better use of the resources it already has. Total factor productivity (TFP) seeks to isolate this last dimension: the share of output not explained by the measured quantities of labor and productive capital, such as equipment, machinery, and facilities.

The idea sounds simple—doing more with the same resources—but measuring it requires caution. TFP is not observed directly in the way hours worked or the number of machines are. It is estimated after choosing a model, measuring inputs, and calculating how much growth remains unexplained. It is therefore a very useful indicator for studying economic performance, but not an automatic answer about its causes.

Key idea: TFP is not the sum of each factor's productivity. It is a joint measure of how an economy turns several inputs into output, conditional on the data and model used.

What total factor productivity measures

Suppose a firm increases its output. Part of the change may be due to adding more labor hours; another part may come from installing more equipment. If output rises by more than those changes can explain within the model, the difference is attributed to TFP growth.

This is a residual measure. Operationally, the OECD calculates its change by subtracting the weighted contributions of inputs from output growth. TFP therefore reflects the estimated capacity to generate output from the set of measured resources.

It is useful to distinguish between its level and its growth. The level aims to compare estimated productive capacity at a given point in time. Growth indicates how that capacity changes between two periods. Both calculations are sensitive to data quality and to how capital, labor, and output are measured.

TFP and labor productivity are not the same

Economic productivity can be measured in several ways. One of the best known is labor productivity: how much output is generated per hour worked or per employed person. Its advantage is its intuitive appeal. But it cannot, on its own, show why each hour produces more.

Labor productivity can rise simply because each worker has more or better machines. It can also change because of the training and composition of the workforce, or because TFP rises. In other words, higher output per hour does not demonstrate that the entire organization has become more efficient.

TFP considers several inputs jointly. It does not divide output by hours alone; it attempts to net out the contributions of labor and capital. This distinction helps separate two processes that are often confused: adding resources and combining them more effectively.

An example: two workshops with the same resources

Imagine two furniture workshops. Both have ten people, the same working hours, and comparable machinery. The second produces more usable pieces over the week.

The difference might stem from a clearer division of tasks, less idle time, better designs, more precise coordination with suppliers, or knowledge that reduces errors. If the measured quantities of labor and capital are the same, the model would record the output difference as a difference in TFP.

But the result does not reveal which explanation is correct. Perhaps one machine was measured incompletely, one workshop used higher-quality materials, or the recorded hours do not reflect effective effort. The example illustrates both the usefulness and the limit of TFP: it detects a difference that the recorded inputs do not explain without automatically identifying its source.

Useful distinction: Higher TFP means output grew by more than the measured inputs explain. It does not, by itself, prove which innovation, rule, or decision caused the change.

The formula and the Solow residual

The intuition can be expressed through a production function:

Y = A · F(K, L)

Here, Y represents output; K, capital; L, labor; and A, TFP. The function F represents the assumed relationship between inputs and output. A common version is the Cobb-Douglas production function:

Y = A · K^α · L^(1−α)

The parameter α approximates capital's weight, while 1−α represents labor's weight under certain assumptions. This formula is a widely used illustration, not a universal definition of TFP.

In growth accounting, researchers estimate how much of the rise in Y corresponds to the weighted growth of K and L. What remains is recorded as a change in A. This procedure is associated with Robert Solow's 1957 work, hence the term Solow residual.

Calling the entire residual “technology” can be misleading. Technology is part of the story, but the residual may also capture organization, knowledge, economies of scale, changes in capacity utilization, omitted inputs, and measurement error. Above all, it is an accounting separation built under explicit assumptions.

What may lie behind higher TFP

TFP can improve when a firm adopts a process that reduces waste, when knowledge circulates more quickly, or when workers and capital move toward activities where they generate more value. Management quality, innovation, and resource reallocation are plausible mechanisms.

Relations among firms also matter. In a market economy, prices that convey information, voluntary exchange, and competition can help discover better methods and move resources from less-valued uses toward those in greater demand. Yet these mechanisms do not appear as separate variables in the basic equation. Additional evidence is needed to determine their importance.

The IMF highlights the connection among TFP, resource allocation, and the capacity to generate income. The relationship is economically relevant, but it does not justify attributing a particular estimate to a single institutional cause.

The limits of measurement

Every TFP estimate depends on prior choices. Analysts must measure not only the quantity of capital but also the services it provides; not only the number of people but also hours and, in some exercises, differences in skills. They must also estimate the shares through which each input contributes to output.

These choices create several limits:

• Capital is not homogeneous: a computer, a building, and a machine provide different services.
• Capacity utilization changes: the same equipment can operate for few or many hours.
• Input quality can vary without being fully captured in the quantities measured.
• Output in some sectors, especially certain services, is difficult to measure.
• Results depend on the production function and assumptions chosen.

In addition, correlation is not causation. If an institution and TFP move together, causation may run in one direction, both directions, or neither, because a third variable drives both. The OECD cautions about measurement, specification, and causality problems when studying the drivers of productivity.

Caution: The residual groups together what the model does not explain. Interpreting it requires investigating its components; calling it “efficiency” does not remove the uncertainty.

Why it matters for growth and living standards

Accumulating labor and capital can expand production, but those inputs have costs and cannot grow indefinitely at the same rate. Improving how they are coordinated offers another path: obtaining more value from the available bundle of resources.

This is why TFP has a central place in the analysis of long-run growth and income per person. It does not mean that every improvement in well-being can be reduced to a production figure. Nor does it mean capital accumulation no longer matters. It means that counting workers and machines is not enough to understand a lasting rise in output.

From a classical liberal perspective, institutions such as predictable property rights, general rules, openness to exchange, and competition may encourage experimentation, coordination, and reallocation. This is a hypothesis about economic channels, not a term in the formula or a guaranteed effect. The connection between economic freedom and prosperity must be assessed with evidence and attention to context.

TFP helps frame the right question: after accounting for inputs, what share of productive performance remains unexplained? Its answer is a starting point for investigating how value is created, not a verdict on the cause. Properly interpreted, it makes it possible to distinguish between growth that comes from having more resources and growth that comes from learning to coordinate them better.