Total Factor Productivity

Total Factor Productivity (TFP) captures the part of economic growth that cannot be attributed to increases in capital stock or labor hours. It is a measure of how efficiently an economy uses its inputs—essentially, how much better off it is due to innovation, management, and technological adoption rather than simply more workers or machines.

The Solow residual

In the 1950s, economist Robert Solow examined postwar US growth and asked: How much of the historical GDP growth came from simply having more workers and more machines? Using a production function framework, he attributed growth to three sources: capital accumulation, labor growth, and everything else—the “residual.”

That residual turned out to be huge. Solow found that roughly half of US growth came from this unexplained factor, not from adding more inputs. He concluded that this residual was a measure of technological progress—the economy was becoming better at producing output per unit of input. This became the Solow residual, or Total Factor Productivity.

Mathematically:

$$\Delta \ln Y = \alpha \Delta \ln K + (1-\alpha) \Delta \ln L + \Delta \ln A$$

Where Y is output, K is capital, L is labor, A is TFP, and α is capital’s share of income. Rearranging, TFP growth is:

$$\Delta \ln A = \Delta \ln Y - \alpha \Delta \ln K - (1-\alpha) \Delta \ln L$$

TFP is what’s left after accounting for more capital and more labor.

Why TFP matters for living standards

A worker in 2025 produces far more output per hour than a worker in 1925—not because the modern worker is 20× stronger, but because factories are more efficient, tools are better, logistics are optimized, and knowledge is accumulated. This is TFP in action.

Rising TFP means rising real wages (wages adjusted for inflation). Conversely, if TFP growth stalls—say, from 2% per year to 0.5%—wage growth slows even if labor is more abundant and capital is abundant. This phenomenon, sometimes called “secular stagnation,” was blamed for wage stagnation in developed economies in the 2010s.

Investors care deeply about TFP because it drives long-term earnings power. A firm with rising TFP compounds value through better processes and less waste; one with falling TFP requires continuous cost-cutting or pricing power to sustain margins.

Sources of TFP growth

TFP growth can come from:

• Technological innovation: Better industrial processes, pharmaceuticals, semiconductor designs, AI. These directly raise output per worker.
• Organizational improvements: Lean manufacturing, supply chain optimization, inventory management, corporate restructuring.
• Human capital: Education, training, and experience raise worker productivity even without new machinery.
• Institutional change: Deregulation, trade opening, financial market deepening. These can unlock efficiency.
• Reallocation of labor: Shifting workers from low-productivity sectors (agriculture) to high-productivity sectors (manufacturing, services). This was a major driver of 20th-century growth in developed economies.

Conversely, TFP can fall if:

• Regulatory burden increases (compliance costs eat into output without producing goods).
• Institutional deterioration (corruption, weak rule of law, financial crises).
• Sector shift into low-productivity areas.
• Measurement bias: If unmeasured quality improvements are not captured in official GDP, TFP appears to decline.

TFP growth by country and era

US 1950–1970: TFP growth averaged ~2% per year. Strong innovation (computers, highways, automation), abundant education expansion, post-WWII catch-up effects.

US 1970–1995: TFP growth fell to ~0.5%, despite high R&D investment. This was attributed to stagflation, energy shocks, and measurement issues (computers were hard to price).

US 1995–2005: TFP rebounded to ~1.5% as information technology matured and became productive.

US 2005–2015: TFP fell back to ~0.5-1%, attributed to financial crisis recovery, aging demographics, and slower innovation diffusion.

Emerging markets: When a poor country adopts technology from rich countries (without inventing it), TFP growth can be very high—3-5% per year—because the gap between frontier and local practice is wide. As convergence occurs, TFP growth naturally slows.

Challenges in measuring TFP

TFP is not directly observed; it is a residual—the leftover after subtracting capital and labor. This makes it vulnerable to measurement error:

• Quality adjustment: If a new car is 10% more reliable than last year’s model but costs the same, is output up 10%? Most statistical agencies try to adjust, but quality improvements are hard to quantify.
• Intangibles: Investments in software, R&D, and training are often expensed (counted as costs) rather than capitalized (counted as capital stock), which can understate capital’s contribution and overstate TFP.
• Sector composition: If the economy shifts toward low-measured-productivity services (e.g., healthcare, education), official TFP growth declines even if quality is improving. Healthcare TFP is notoriously hard to measure; how do you price a life-saving drug?
• Utilization rates: In a recession, factories operate below capacity. Measured TFP falls because output is down, but the technology didn’t actually deteriorate—the inputs are just underused.

Endogenous growth and TFP

Early growth models (Solow, Harrod-Domar) treated TFP as exogenous—it falls from the sky like manna. More recent endogenous growth theory models innovation as the result of intentional R&D investment and human capital accumulation. Firms invest in R&D because they expect patents and profit; the economy’s TFP growth endogenously depends on these incentives.

This distinction matters for policy: if TFP is exogenous, there’s little a government can do to speed it up (beyond broad research funding). If it’s endogenous, policies that incentivize innovation and education can boost growth.

TFP and asset markets

Equity valuations reflect expectations of future TFP growth implicitly. In a period of expected high TFP (e.g., the dot-com bubble’s belief in internet transformation), multiples rise. When TFP growth disappoints (e.g., 2000–2010, secular stagnation fears), multiples compress.

Sectors with high TFP potential (tech, biotech, software) trade at premiums; mature, low-TFP sectors (utilities, transportation) at discounts.