Multifactor Productivity

Multifactor productivity — also called total-factor productivity (TFP) or the Solow residual — is the portion of productivity growth that cannot be attributed to increases in labor and capital inputs. It captures the improvement in how efficiently an economy combines inputs, typically attributed to technological progress and organizational innovation.

The Solow residual

The concept originates with MIT economist Robert Solow’s 1956 growth accounting. He demonstrated that most long-run growth in developed economies could not be explained by accumulating more labor and capital — there was a large residual. That residual, the story goes, is technological progress.

The calculation is straightforward. Start with the growth accounting identity:

Output Growth = Labor share × Labor growth + Capital share × Capital growth + MFP growth

Rearrange to solve for MFP:

MFP growth = Output growth − (Labor share × Labor growth) − (Capital share × Capital growth)

In the US, labor’s share of income is roughly 70% and capital’s share is 30%. If GDP grows 3%, labor hours grow 1%, and capital stock grows 2%, then:

MFP = 3% − (0.7 × 1%) − (0.3 × 2%) = 3% − 0.7% − 0.6% = 1.7%

That 1.7% is attributed to efficiency improvements — the Solow residual.

What drives multifactor productivity

In theory, MFP captures:

• Technological innovation. New machines, software, techniques — whether the internal combustion engine or artificial intelligence.
• Organizational efficiency. Better supply chains, inventory management, and business practices.
• Knowledge spillovers. When one firm’s innovation becomes available to others.
• Management quality. Better oversight, incentive structures, and decision-making.

In practice, MFP is a catch-all. If you cannot explain growth by labor and capital, whatever is left gets labeled “technology.”

The measurement problem

MFP is measured as a residual, which means it includes all measurement error:

• If capital stock is mismeasured, MFP reflects that error.
• If quality improvements in labor and capital are not captured, they appear as MFP.
• If output is mismeasured — particularly in services — MFP is wrong.

A famous critique: unmeasured quality improvements mean that real output (and thus MFP) may be higher than official statistics show. A smartphone is vastly more powerful than 1980s computers, but how much productivity does that add to GDP? National accounts struggle with this.

MFP and technological epochs

MFP growth is not smooth. It accelerates during periods of major innovation and slows during periods of stagnation or low innovation:

• 1900-1950: High MFP growth as electricity, automobiles, and mass production spread.
• 1950-1973: Continued strong MFP growth as technological momentum carried forward.
• 1973-1995: MFP slowdown — the “productivity paradox” period.
• 1995-2005: MFP acceleration as information technology diffused widely.
• 2005-present: MFP slowdown again, suggesting either genuine innovation stagnation or measurement failure.

The secular stagnation debate

One of the central macroeconomic debates concerns whether MFP growth is permanently low. If true, potential GDP growth is low, real interest rates must be low, and wealth accumulation is slow.

• Pessimists argue that major innovations (electricity, antibiotics, aviation) are exhausted and that recent digital innovations are narrower in impact.
• Optimists argue that MFP growth is being mismeasured and that artificial intelligence will drive a new acceleration.

The truth — whether MFP growth of 0.5% is the new normal or 1.5% is achievable — has enormous implications for policy and investment.

MFP versus labor productivity

The key distinction:

• Labor productivity = Output per hour of labor (includes capital effect)
• Multifactor productivity = Output per all inputs (isolates the technology/efficiency effect)

A factory can raise labor productivity by 10% by adding expensive machinery; multifactor productivity might rise only 2% because the capital is expensive. The first is easier to measure; the second better captures true technological progress.

See also