Macroeconomics › Section G · Economic growth › Topic G.2
Labor productivity and growth
A country cannot keep raising living standards by working ever longer hours. Over the long run, incomes rise because each hour of work produces more, and small gains in that rate compound into very large differences.
Output per hour is the engine
US pay per hour, adjusted for inflation, is almost three times its 1950 level. The reason is not that people work longer: average hours have, if anything, fallen. The difference is that each hour of work now produces far more goods and services. Economists call this labour productivity, and it is the most important driver of long-run living standards. This topic shows how to measure it, what raises it, and why small differences in its growth rate matter so much.
Measuring labour productivity
Labour productivity is output per unit of labour input:
Output per worker is a cruder version that ignores differences in hours. In the United States, the Bureau of Labor Statistics (BLS) publishes real output per hour in the nonfarm business sector. That sector leaves out general government, nonprofit institutions, private households and farms. Government output is valued mainly by staff pay, so its measured productivity barely moves, and farm output swings with the weather. Even so, the sector covers about three-quarters of US GDP, by BLS estimates.
What makes workers more productive
Four broad sources raise output per hour. Human capital is the knowledge, skills and experience workers carry, built through schooling, training and practice. Physical capital is the machinery, buildings, vehicles, software and infrastructure workers use; a road crew with an excavator outdigs one with shovels. Technology means better ways of producing, combining invention, a new idea, with innovation, putting it to practical use. Economies of scale arise when larger operations cut cost per unit, as when a large bakery spreads one industrial oven across thousands of loaves. Most productivity gains combine several sources at once.
The aggregate production function
Economists summarise these sources with an aggregate production function: the relationship between an entire economy's inputs and its total output. In words, real GDP depends on the quantity of labour, the stock of physical capital, the level of human capital and the state of technology. Dividing everything by population gives a per-person version, in which GDP per person depends on capital per person, human capital per person and technology. No one observes "technology" directly, so the function organises thinking rather than giving a formula to look up. The next topic, "The components of economic growth", uses it to sort growth into its sources.
Productivity and GDP per person
Productivity and GDP per person are related but not identical. An identity links them:
where \(Y\) is real output, \(H\) is hours worked, \(E\) is the number employed and \(N\) is population. GDP per person rises if output per hour rises, if each worker works more hours, or if a larger share of the population holds jobs. The last two can only rise so far: hours cannot exceed the day, and employment cannot exceed the population. Only output per hour can keep rising decade after decade, which is why it governs long-run growth.
Worked example: splitting growth per person
In a hypothetical economy over one year, output per hour rises 1.8%, average hours per worker fall 0.3%, and the share of the population employed rises 0.5%. 1. Use the identity: GDP per person grows by the factor \(1.018 \times 0.997 \times 1.005\). 2. Multiply: \(1.018 \times 0.997 = 1.01495\), and \(1.01495 \times 1.005 \approx 1.0200\). 3. So GDP per person grows about 2.0%. 4. Check with the shortcut: adding the rates gives \(1.8 - 0.3 + 0.5 = 2.0\) percentage points. Adding works well for small rates. Note that about nine-tenths of the gain came from productivity.
The US productivity record
In the BLS annual index of nonfarm business output per hour, growth was about 2.8% a year from 1948 to 1973, fell to about 1.4% from 1973 to 1995, revived to about 3.1% from 1995 to 2004, slowed to about 1.5% from 2004 to 2019, and averaged about 2.2% from 2019 to 2025. By 2025, output per hour was about 5.2 times its 1947 level. The 1970s slowdown is still not fully explained. The late-1990s revival, linked to computers and the internet, led some to predict a lasting "new economy"; the slowdown after 2004 showed how risky it is to read a few strong years as a new trend.
Productivity and wages
Over long periods, the average wage in an economy tracks the value of what an hour of work produces. A firm that pays well below a worker's output invites rivals to hire that worker away; a firm that pays well above it makes losses. So competition ties average pay to productivity, though adjustment is slow and not exact. Economists debate how closely the typical worker's pay has followed productivity since the 1970s, since benefits, price measures and labour's share of income all affect the comparison. The core point is widely accepted: no economy pays high wages for long without high productivity.
Compounding and the rule of 70
Growth compounds like interest: each year's gain applies to a base that already includes past gains. After \(n\) years at rate \(g\), a starting value \(Y_0\) becomes
A quick way to see the power of this is the rule of 70: a quantity growing at \(g\)% a year doubles in about \(70 \div g\) years. It works because \(\ln 2 \approx 0.693\). At 1% a year the doubling time is about 70 years, at 2% about 35, at 3% about 23, and at 7% about 10. The exact figures are 69.7, 35.0, 23.4 and 10.2 years, so the rule is accurate for everyday growth rates.
Worked example: the gap two points make
Two hypothetical economies start with real GDP per person of $50,000. One grows at 1% a year, the other at 3%, for 40 years. 1. Slow economy: \(50{,}000 \times 1.01^{40} \approx 74{,}400\) dollars. 2. Fast economy: \(50{,}000 \times 1.03^{40} \approx 163{,}100\) dollars. 3. Ratio: \(163{,}100 \div 74{,}400 \approx 2.19\). 4. Check with the rule of 70: at 3%, income doubles about every 23 years, so 40 years gives a little under two doublings, about 3.3 times the start, matching \(163{,}100 \div 50{,}000 \approx 3.26\). Two extra points of growth, over one working life, more than doubles living standards.
Common mistakes to avoid
First, treating productivity as how hard people work. It is output per hour, and it rises mostly through skills, equipment and better methods, not effort. Second, equating productivity growth with GDP-per-person growth: more hours or a higher employment share can raise GDP per person for a while without any productivity gain. Third, adding growth over many years instead of compounding: 3% a year for 40 years is a 226% rise, not 120%. Fourth, using the rule of 70 with a decimal rate: at 2% growth divide 70 by 2, not by 0.02.
Key takeaways
Labour productivity is real output per hour worked, and the BLS measures it for the US nonfarm business sector. It rises with human capital, physical capital, technology and economies of scale, all summarised by the aggregate production function. GDP per person equals output per hour times hours per worker times the employment share, and only productivity can rise without limit. US productivity growth was fast to 1973, slow until 1995, fast again until 2004, then slower. Over time, average wages follow productivity. Growth compounds as \(Y_0(1 + g)^n\), and the rule of 70 gives a doubling time of about \(70 \div g\) years.
Flashcards
Flashcard
Real output ÷ hours worked
Test your understanding
Question 1 of 5
A factory produces $4.8 million of real output using 60,000 hours of labour. What is its labour productivity?
Question 2 of 5
Using the rule of 70, about how long does an economy growing at 2.5% a year take to double its real GDP per person?
Question 3 of 5
In one year, output per hour rises 1%, hours per worker are unchanged, and the share of the population employed rises 2%. Roughly how fast does GDP per person grow?
Question 4 of 5
Why can rising productivity, but not rising hours or employment, sustain growth in GDP per person for centuries?
Question 5 of 5
Real GDP per person of $40,000 grows at 2% a year for 35 years. Which is closest to its final value?
