The Gold Standard of Economics | GDP per Worker versus GDP per Hour Worked

National output can rise because more people work, because workers spend more hours at work, or because the economy creates more output per hour. A headline describing ‘productivity per worker’ mixes the latter two into one denominator. A headline describing ‘productivity per hour worked’ separates the volume of labour time from the value produced during it.

Let Y denote economic output such as appropriately measured GDP, E the number of persons employed and H the total hours actually worked. GDP per person employed is Y/E. GDP per hour worked is Y/H. They are both measures of aggregate labour productivity, but the choice of denominator determines the exact claim.

The two are linked by an accounting identity: Y/E = (Y/H) × (H/E). Output per person employed equals output per hour multiplied by average annual hours worked per employed person. This is an identity when the numerator, workforce and hours refer to consistent definitions and periods, not a behavioural theory that tells us why productivity changed.

An OECD productivity indicator defines GDP per hour as gross domestic product generated per hour of labour input. It also stresses that measured productivity depends on capital, technology, organisation and other factors. A country’s hourly GDP is not a moral rating of its workers’ diligence.

Start with the question the reader wants answered: ‘How much output is associated with each employed person?’ or ‘How much is associated with each hour actually worked?’ When employment patterns differ, those answers can lead to different country rankings.

Contents · Next chapter

Consider two completely imaginary economies in the same fictional year. Express GDP for both in the same comparable output currency and valuation, and suppose each has 10 million employed people. Economy A has 600 billion output units and 18 billion total hours worked. Economy B has 580 billion output units and 14.5 billion hours.

Calculate GDP per employed person first. A: 600 billion / 10 million = 60,000 units per person. B: 580 billion / 10 million = 58,000 units per person. On this measure, A ranks higher.

Now calculate output per hour. A: 600 billion / 18 billion = 33.33 units per hour. B: 580 billion / 14.5 billion = 40.00 units per hour. On the hourly measure, B ranks higher. Nothing in the numerical record is contradictory; the denominators address different aspects of labour input.

Average annual hours per worker are 18 billion / 10 million = 1,800 in A and 14.5 billion / 10 million = 1,450 in B. Multiply hourly output by average hours and you recover the per-worker amount, up to rounding. This is an ideal classroom check for both interpretation and arithmetic.

The comparison is not a report about actual countries. Real international comparisons need national accounts consistency, purchasing power parity adjustments where appropriate, comparable time periods and well-estimated employment and hours. Their economy-wide aggregates also conceal sector composition.

Contents · Previous chapter · Next chapter

Employment headcounts are comparatively intuitive: a person is counted as employed under the relevant statistical definition. Hours worked are less simple. For productivity analysis, the OECD emphasises the number of hours actually worked by people engaged in production, not only the hours written in contracts or paid on a payroll.

Actual hours can include regular work by full-time and part-time workers, paid and unpaid overtime, and hours worked in additional jobs. Hours not actually worked because of leave or several forms of absence should not be treated as productive labour input. The applicable national-accounts production boundary and harmonised methodology determine detailed inclusion rules.

This is why average annual hours are not simply 40 hours multiplied by 52 weeks for every employed person. Part-year work, part-time schedules, sick leave, holidays, self-employment and second jobs all matter. Different surveys and administrative sources can also yield different estimates.

The OECD’s separate Hours Worked indicator warns against simple cross-country level comparisons based on that particular series because statistical sources and methods vary. A harmonised productivity database may make additional adjustments, but its documentation still deserves attention.

One mistake is to compare GDP from a calendar year with hours measured for a nonmatching fiscal or survey period. Another is to count the same second job as a second person in E while counting hours correctly. Numerator and denominator must share a coherent population boundary.

A strong explanation names whether H refers to total actual hours, average hours per person, or some other survey series. The words ‘working hours’ are not enough; they can hide contractual, usual, paid and actual-hour concepts.

Contents · Previous chapter · Next chapter

Take another entirely fictional economy over two periods. Let real output in the first year be 500 billion units, employment 10 million and actual hours 16 billion. GDP per worker is 50,000 units, output per hour is 31.25 and hours per worker are 1,600.

Next year suppose real output is 510 billion, employment remains 10 million and actual hours fall to 15 billion. GDP per worker rises to 51,000, a 2% increase. GDP per hour rises to 34, an 8.8% increase. Average hours per worker fall to 1,500, a 6.25% decline.

Check the identity: 31.25 × 1,600 = 50,000 in the first year, and 34 × 1,500 = 51,000 in the second. Output per hour increased considerably faster than output per worker because the average quantity of labour time fell.

Does this prove that each employee became 8.8% more talented? No. Changes in capital equipment, industry shares, technology, business demand, the composition of employment or measurement could all contribute. GDP per hour is an aggregate ratio, not a score assigned to individuals.

Does shorter working time always raise GDP per hour? Also no. This example holds output high while hours fall. In a different case, output could fall by the same proportion as hours and hourly productivity would remain unchanged; it could even fall more and reduce hourly productivity. Always calculate rather than apply a slogan.

This distinction is useful when discussing four-day working weeks, rising part-time employment, productivity programmes and the relationship between work and leisure. A responsible reader asks what happened to both output and labour time.

Contents · Previous chapter · Next chapter

Comparing the level of GDP per hour across currencies raises a valuation problem. Converting domestic output with a market exchange rate can exaggerate differences in domestic prices. Purchasing power parities are designed to place output into a more comparable common purchasing-power framework. The OECD uses such comparisons for internationally comparable productivity levels.

By contrast, measuring a country’s productivity growth over time normally uses real or volume measures of output, not raw GDP at current prices. Nominal GDP may rise because prices rise even when the quantity of production is flat. A productivity headline that ignores this difference can turn inflation into imaginary efficiency.

Suppose nominal GDP grows 8% and its broad price measure rises enough that real GDP grows only 2%. Hours worked are unchanged. Approximate real output-per-hour growth is 2%, not 8%. The numerator should match the economic question: real production efficiency rather than current money value.

Cross-country PPP conversions are not a complete solution to measurement differences. Differences in industrial structure, self-employment, the informal economy, imputed output, national accounts revisions, survey coverage and the treatment of multinational activity can complicate interpretation.

Use consistent data vintages and definitions when comparing years. A revised GDP series or revised hours estimate can change measured productivity even though nobody travelled back in time to change their work. The earlier seasonal adjustment and revisions guide explains why published economic histories can change after methodological updates.

More precise-looking figures do not repair inconsistent definitions. The gold standard is to state units, price basis, population, period and source before drawing a comparison.

Contents · Previous chapter · Next chapter

GDP per hour combines output produced with labour and other inputs. Imagine two workers with identical skills and work habits: one uses modern equipment in a well-organised high-value production chain, while another works with old equipment and unreliable electricity. Their observed value added per hour can differ for reasons unrelated to personal effort.

Capital intensity, technology, supply-chain organisation, management quality, resource endowment and market conditions all influence national measured productivity. Sector composition matters too: an economy concentrated in services with difficult-to-measure output should not be interpreted by the same simple physical-units story as a uniform factory.

An economy can increase GDP per worker by moving employees into jobs with longer hours. That is not necessarily an improvement in output per hour, nor does it tell us whether wellbeing rose. Conversely, hourly productivity can rise while per-person output barely changes if annual working hours decline.

Do not use GDP per hour as a direct measure of happiness, job quality, income equality or wage fairness. A nation’s income distribution can change independently from aggregate productivity. GDP includes returns to several factors of production, and output prices differ from individual wage rates.

For policy discussions, supplement productivity figures with employment, actual hours, wages, real household income, working conditions and distributional measures. The objective is not to flatten the economy into one ranking but to explain what each indicator can responsibly establish.

The International Labour Organization emphasises that labour productivity depends on how labour combines with other factors, including machinery and technology. That is a helpful antidote to blaming or praising individual workers based solely on a national ratio.

Contents · Previous chapter · Next chapter

Which is the better labour productivity measure?

GDP per hour is often preferable when the question concerns output relative to labour time, especially across different working patterns. GDP per person employed remains useful when hours are unavailable or the question concerns output per employed person.

Can GDP per worker rise with no improvement per hour?

Yes. Longer average hours with unchanged output per hour make output per worker rise by accounting identity.

Can hourly productivity rise while output per worker falls?

Yes. A decline in average working hours can more than offset an increase in output per hour.

Why not simply divide GDP by the whole population?

GDP per capita addresses output or income per resident, including children, retirees and people not employed. GDP per worker and GDP per hour answer different labour-input questions.

Is overtime part of actual hours?

When actually worked in activities covered by the measurement, both paid and unpaid overtime are generally included under the relevant OECD definitions.

Are actual working hours always internationally comparable?

Not automatically. The OECD identifies measurement-source differences and warns that some hours-worked level series are not suited to straightforward country-level comparisons.

Do higher GDP-per-hour figures imply higher individual wages?

No. The link depends on the distribution of output, compensation practices, capital returns, employment and other economic conditions.

Why do productivity rankings change when an official data release is revised?

GDP, employment or estimated working hours can be revised, as can methodological and price-adjustment assumptions. A different vintage can change the ratios without changing events in the past.

Contents · Previous chapter · Next chapter

The practical lesson is not merely memorising two fractions. It is asking what belongs in the denominator and whether that denominator answers the question you are asking. A change in hours worked can change the GDP-per-worker story without changing hourly efficiency in the same way. That insight helps readers evaluate international rankings, workplace proposals and economic policy.

Reconstruct the two-country example on paper. Economy A produces 60,000 units per employed person but only about 33.33 per hour; Economy B produces 58,000 per employed person and 40 per hour. Explain both rankings, then change average hours while holding hourly productivity fixed. If the ranking adjusts predictably, the identity has become a tool rather than a slogan.

Subject and learning map

The Gold Standard of Economics — parent subject guide and linked specialist material.

The Gold Standard of Macroeconomics — connect output, labour inputs and aggregate growth.

The Gold Standard of Work — human, institutional and employment context.

The Gold Standard of Education — move denominator reasoning into unfamiliar evidence.

How X Works — the wider eduKate explanation ecosystem.

Companion economics readings

How Productivity Works — broad overview of useful output and scarce inputs.

Seasonal adjustment and revisions — why economic time-series comparisons are not fixed forever.

Age composition and labour-force participation — another example of the denominator changing a headline.

Authoritative sources and scope

References reviewed on 22 October 2026. Both sets of numerical cases are entirely fictional. They use internally comparable valuation assumptions and are not official country productivity estimates.

OECD: GDP per hour worked — definition, labour input and the role of other production factors.

OECD Compendium of Productivity Indicators 2025 — international level comparisons and the hours versus employment choice.

OECD: Hours worked — actual hours, included and excluded time, and series-comparability warnings.

ILO: Labour Productivity FAQ — why capital and technology influence labour productivity.

Contents · Previous chapter · Economics hub