- Benefits and costs of economic growth
- Measuring development — beyond GDP · coming soon
Two stories the measuring stick gets backwards
An oil tanker runs aground and oil covers a coastline. Clean-up crews are hired. Boats, booms and chemicals are bought. Lawyers are paid. All of this spending is counted in GDP, so GDP goes up.
Now a different story. A woman leaves her paid job as a nurse to care for her father at home. She works just as hard as before, and her care matters just as much. But nobody pays her, so her work is not counted. GDP goes down.
In the first story, the country is clearly worse off, and GDP rises. In the second, nothing valuable has been lost, and GDP falls. Something is wrong with our measuring stick.
The last lecture ended with a promise. Real GDP tells you how much an economy produces, but it does not yet tell you how well its people live. This lecture asks that question. The disruption is this: GDP is a measure of market production. Welfare is how well people live. The two are related, but they are not the same thing.
A warning from the man who built it
The first lecture in this topic met Simon Kuznets, who presented the first official estimates of United States national income to the US Senate in January 1934. They were built during the Great Depression, to measure how far the economy had fallen. During the Second World War, governments used the same accounts to plan how much they could produce for the war.
In that same 1934 report, Kuznets warned against using his number for something else:
"The welfare of a nation can, therefore, scarcely be inferred from a measurement of national income."
National Income, 1929–1932, US Senate Document No. 124, page 7.
GDP is still the most useful single number we have about an economy. Richer countries, on average, have longer lives, more schooling and more choice. So the question is not "Is GDP useless?" The question is: what exactly does it miss? There are five gaps, and each one comes from the definition you already know. GDP adds up the market value of final goods and services produced in a period.
Gap 1: GDP per head says nothing about who gets it
To compare living standards, economists divide GDP by the population. This gives GDP per head, also called GDP per capita. It is an average. And an average can hide almost everything about how income is shared.
Two invented countries each have five people and the same total income: $150 000 a year. So both have an income per head of dollars.
Arland: incomes of 26, 28, 30, 32 and 34 thousand dollars.
Bervia: incomes of 10, 12, 14, 24 and 90 thousand dollars.
The median is the middle income when the incomes are in order. In Arland it is $30 000, the same as the mean. In Bervia it is $14 000, less than half the mean.
The richest person's share. In Bervia, one person receives of all income. Four people out of five earn less than the average.
Check: ✓. The averages really are equal, and the typical person's life is very different.
One number that measures this is the Gini coefficient. It is 0 when everyone has the same income. It rises towards 1 as income becomes more concentrated, and it would be 1 if one person had everything. The World Bank publishes it multiplied by 100, as the "Gini index". Its latest estimates, for different years, are about 27.7 for Norway (2019), 41.5 for the United States (2019), 49.3 for Costa Rica (2020) and 63.0 for South Africa (2014).
Explore how much the average can hide.
An invented country of five people has a total income of $150 000, so its income per head is always $30 000. Move the slider to give the richest person a bigger share. The other four share the rest equally. Watch what happens to the median and the Gini coefficient while the average stays fixed.
Presets:
Five people, total income $150 000, income per head always $30 000. The richest person takes the share shown; the other four share the rest equally.
| Richest share | Richest income | Median income | Gini |
|---|---|---|---|
| 20% | $30 000 | $30 000 | 0 |
| 40% | $60 000 | $22 500 | 0.2 |
| 60% | $90 000 | $15 000 | 0.4 |
| 80% | $120 000 | $7500 | 0.6 |
GDP per head is $30 000 in every row. The median person's income falls to a quarter of that.
What should you have seen? The dashed line never moves, so GDP per head never changes. Yet the median person goes from $30 000 to $7500. A government that watched only GDP per head would see no difference between these countries.
Gap 2: unpaid work is invisible
GDP counts what is sold in markets. Cooking, cleaning, childcare, caring for older relatives and volunteering are real work, but they are usually unpaid. So they are left out.
This creates strange results. If you pay a cleaner, the payment is counted in GDP. If you clean your own home, it is not. The house is equally clean either way. Only the statistic is different.
The amounts are large. The United Kingdom's Office for National Statistics (ONS) estimated the value of unpaid household services in 2016 at about £1.24 trillion. That was equal to about 63% of GDP. The estimate includes childcare, adult care, cooking, laundry, housing upkeep, transport and volunteering, valued at what it would cost to buy the same services.
A farmer in a village grows rice, and her family eats all of it. None of it is sold. Is the rice counted in GDP?
Show the answer
Yes, in principle. The international rules for national accounts count goods that households produce for their own use, such as food from subsistence farming, valued at market prices. In practice the amount has to be estimated, often roughly. But when she cooks that rice into meals for her family, the cooking is a service for her own household, and unpaid services are left out. The rule draws its line between goods and services, not between sold and unsold.
This matters most when countries are compared. Where much food, building and care stays inside families, measured GDP understates how much is really produced.
Gap 3: harm is not subtracted
Go back to the oil spill. The clean-up spending is counted. Some losses do show up: if fishing boats catch less, their sales fall, and GDP falls with them. But the harm itself is never subtracted: dead sea birds, a damaged coastline, years of lost beauty and health. These have no market price, so GDP does not see them.
This is the externalities problem from earlier lectures, seen at the level of a whole country. When a factory produces more, its output is counted at its market price. The external cost, the smoke over the town, is not subtracted. The same happens when a forest is cut down and the timber is sold. The timber is counted as output. The loss of the forest, a stock of natural wealth, is not.
The first lecture in this topic showed that net national income subtracts the wearing out of machines and buildings. Nothing similar is subtracted for using up forests, fish or clean air. So a country can raise its GDP for years by running down its natural wealth.
Gap 4: hidden and illegal activity
Some activity is never recorded. A builder is paid in cash and does not declare it. A street vendor keeps no records. Economists call this the shadow economy, or the informal economy. It is hard to measure by its nature, so every estimate rests on a model.
One widely cited study, published by the International Monetary Fund (Medina and Schneider, 2018), estimated it for 158 countries over 1991–2015. The average was about 31.9% of GDP. The estimates ranged from about 7.2% in Switzerland to about 62.3% in Bolivia. Wherever it is large, official GDP understates real activity.
Illegal activity raises a different question: should it be counted at all? In 2014 the United Kingdom answered yes, as the real case below shows.
The change. New European accounting rules, the European System of Accounts 2010, required every member country to include illegal activities that people take part in willingly. The aim was comparability: some activities are illegal in one country and legal in another. So in its 2014 revisions, the ONS began to include spending on illegal drugs and prostitution in UK GDP.
The prediction from GDP alone. If GDP measured welfare, the United Kingdom became better off overnight, without anyone doing anything differently.
What happened. The ONS reported that these activities added up to about £10 billion a year to the level of GDP. That is less than 1% of the UK's GDP, which was well over £1 trillion. Where full data did not exist, the ONS built estimates from research on numbers of workers and on drug prices.
Verdict. The measurement improved for its own purpose: counting all market production in the same way in every country. But no one in Britain lived better because of the change. GDP rose; welfare did not move. The case shows exactly what GDP is: a careful count of market activity, not a judgement about whether that activity is good.
Source: ONS, "Sex, drugs and GDP", Quarterly blog, 15 October 2014.
Gap 5: time and leisure
GDP measures output. It does not count the hours it took. Imagine two invented countries with the same GDP per head. In one, people work 35 hours a week. In the other, they work 50. The first country has 15 more hours a week for rest, family, health and community. Those hours have real value, but they add nothing to GDP. Measured as output per hour worked, the first country is far more productive, because it produces the same with 30% fewer hours.
The same point works in reverse. If a country's people work longer hours, GDP rises, but part of the rise has been paid for with lost free time.
Now test the five gaps. Each case below shows one of them.
Each card describes a change. Put it in the box for the gap in GDP that it shows.
Each case and the gap it shows:
- Who gets it: profits of a few large firms soar while most wages stay flat; two countries with the same GDP per head, but the median person in one earns half as much.
- Unpaid work: Samira leaves a paid nursing job to care for her father at home; volunteers run a free weekend school for 200 children.
- Harm not subtracted: a factory's output rises, and so does the smoke over the town; a forest is cut down and the timber sold.
- Hidden activity: a builder is paid in cash and never reports the job; street vendors sell food with no licence or records.
- Time and leisure: workers produce the same as another country in 15 fewer hours a week; a firm keeps output the same on a four-day week.
Better measures: adding health and education
If GDP leaves so much out, what can we use alongside it? The best-known answer is the United Nations' Human Development Index (HDI). It combines three things a good life needs: a long and healthy life, knowledge, and a decent standard of living.
Health index = (life expectancy at birth − 20) ÷ (85 − 20).
Education index = the average of (expected years of schooling ÷ 18) and (mean years of schooling ÷ 15).
Income index uses gross national income (GNI) per head at purchasing power parity, on a logarithmic scale between $100 and $75 000.
HDI = the geometric mean of the three: . It runs from 0 to 1.
Two details carry the meaning. Income goes on a logarithmic scale, because an extra $1 000 matters far more to a poor country than to a rich one. And the three indices are combined by a geometric mean, not an ordinary average. The cube root of a product punishes a very low score more. A country cannot make up for poor health with very high income.
The data (UNDP, Human Development Report 2025, for 2023). United States: life expectancy 79.3 years, GNI per head about $73 650 (at 2021 PPP prices). Costa Rica: life expectancy 80.8 years, GNI per head about $23 417.
The prediction from income alone. Income per head in the United States is about times Costa Rica's. If income measured welfare, Americans would be far better off in every way.
Health. Costa Rica: . United States: . Costa Ricans live longer, on average.
Education. Costa Rica: expected 16.3 years, mean 8.8 years, so . United States: 15.9 and 13.9 years, so about 0.905.
Income. On the logarithmic scale, about 0.824 for Costa Rica and 0.997 for the United States.
HDI. Costa Rica: . United States: about 0.937. The report gives 0.833 and 0.938, from unrounded data ✓.
Verdict. The income gap is threefold. The HDI gap is much smaller, and in health Costa Rica is ahead. GDP alone would badly overstate the difference in how well people live. HDI does not settle everything, though: it ignores inequality, unpaid work and the environment.
Other measures go further. The Genuine Progress Indicator, and the older Index of Sustainable Economic Welfare, start from consumer spending, add the value of unpaid work, and subtract the costs of pollution, crime and inequality. Bhutan's Gross National Happiness measures psychological wellbeing, health, culture and the environment alongside living standards.
Does more GDP make people happier?
In 1974 the economist Richard Easterlin reported a puzzle. Within a country, richer people say they are happier than poorer people. But as a whole country got richer over time, average reported happiness did not seem to rise. This became known as the Easterlin paradox.
The size of the growth is not in doubt. Real GDP per head in the United States rose from about $19 400 in 1960 to about $68 800 in 2024, in 2017 prices: about 3.6 times as much. Easterlin's explanation was that happiness depends partly on relative income. If everyone gets richer together, nobody moves up compared with others.
The paradox is still argued about. In 2008, Betsey Stevenson and Justin Wolfers studied many countries and many surveys. They found that people in richer countries report higher life satisfaction, with no point where the link stops. They also found that growth over time goes with rising satisfaction. Easterlin and others have replied with different data. The honest summary is that the evidence is mixed.
Three traps
Higher GDP means more market production. It may come with longer hours, more pollution, or gains that go to a few. Always ask what else changed before you conclude anything about welfare.
This is the opposite mistake. GDP measures short-run economic activity well, and it is the measure governments need for jobs, taxes and recessions. A strong answer says what GDP captures well, what it misses, and which other measures fill the gaps.
GDP per head is a mean. One very high income raises it, even when most people earn no more. When you want the typical person's income, use the median.
Work each one on paper first. Open a hint only when you are stuck.
-
An invented country has GDP of $600 billion and a population of 20 million. Its median income is $18 000. Find GDP per head, and explain what the difference between the two numbers suggests.
Hint 1
GDP per head = GDP ÷ population. Be careful with billions and millions.
Hint 2
600 000 million ÷ 20 million = ?
Full answer
GDP per head = $600 000 million ÷ 20 million = $30 000. The median person has $18 000, which is 60% of the mean. So incomes are unequal: a smaller number of high incomes pull the mean well above the typical income. (GDP per head and median income are not measured in exactly the same way, but a gap this large points to inequality.) Check: 20 million × $30 000 = $600 billion ✓.
-
An invented country has a health index of 0.8, an education index of 0.6 and an income index of 0.5. Find its HDI. Then compare two policies: raising the income index by 0.1, or raising the education index by 0.1. Which raises HDI more?
Hint 1
HDI is the cube root of the product of the three indices.
Hint 2
Work out and .
Full answer
HDI = . (An ordinary average would give about 0.633.) Raising income to 0.6: . Raising education to 0.7: . Raising the lowest index, income, helps most. The geometric mean rewards fixing the weakest dimension. Check: both results are above 0.621, as they must be ✓.
-
A storm destroys houses worth $2 billion in an invented country. In the same year, $2 billion is spent rebuilding them. What happens to GDP? Is the country better off than before the storm?
Hint 1
GDP is a flow of production in a year. The houses are a stock of wealth.
Hint 2
Is the building work counted in GDP? Is the destruction subtracted anywhere in GDP?
Full answer
The rebuilding is new production, so it adds about $2 billion to GDP. The destruction is a loss of a stock, and GDP, which measures a flow, does not subtract it. At the end of the year the country has the same houses as before the storm, so it is no better off. It is worse off, because the builders' time and materials could have produced something new instead. GDP rose; wealth and welfare did not.
-
Two invented countries. Country A: output per worker $60 000 a year, with 1900 hours worked per worker. Country B: output per worker $50 000 a year, with 1400 hours worked. Which produces more per hour? If B's workers worked A's hours at B's output per hour, what would B's output per worker be? What does this assume?
Hint 1
Output per hour = output per worker ÷ hours per worker.
Hint 2
Multiply B's output per hour by 1900.
Full answer
A: 60 000 ÷ 1900 ≈ $31.6 per hour. B: 50 000 ÷ 1400 ≈ $35.7 per hour. B is more productive per hour, although its output per worker is lower. At 1900 hours, B would produce about dollars per worker, more than A. This assumes every extra hour is as productive as the earlier ones, which is unlikely, because tired people work more slowly. So B's lower GDP is largely a choice of more leisure, and GDP alone would hide that.
Every measure answers one question well and others badly. GDP answers "How much market production happened here this year?" very well. Real GDP answers "Did that production grow?" PPP answers "How do countries compare?" HDI asks about health and knowledge as well as income. None of them alone answers "How well do people live?"
There is a lesson about measurement here. People tend to improve what is measured. A government judged only on GDP growth may ignore what GDP cannot see. That is why many countries and international bodies now publish wider "dashboards" of wellbeing alongside GDP. The next topic studies economic growth itself: what causes it, and then what it costs as well as what it brings.
- GDP measures the market value of production, not welfare, and its own designer warned in 1934 against reading welfare from it.
- GDP per head is an average, so it hides how income is shared; the median and the Gini coefficient show what the average hides.
- GDP leaves out unpaid household services, does not subtract environmental harm or the using up of natural wealth, misses hidden activity, and ignores leisure.
- The Human Development Index combines health, education and income by a geometric mean, so a weak dimension cannot be fully offset by a strong one.
- Whether rising GDP raises happiness over time, the Easterlin paradox, is still debated, so GDP is best used alongside other measures.