Macroeconomics › Section H · Unemployment › Topic H.1
Defining and computing the unemployment rate
The unemployment rate is one of the most watched numbers in economics, and a change of a tenth of a point can move markets. Reading it well means knowing exactly who counts as unemployed, who does not, and how the figure is collected.
Why one number matters
Unemployment is both a human problem and an economic waste. For a family, a lost job can mean lost income, drained savings and real strain on health and relationships. For the economy, every willing worker without a job is output that could have been produced and was not: the opportunity cost of unemployment. The unemployment rate compresses all of this into one percentage. In 2025 the US labour force averaged about 170.8 million people, so a change of just 0.1 percentage point in the rate meant roughly 171,000 people. Before using the number, though, you need its exact definitions.
Three groups, not two
The US survey sorts every civilian aged 16 and over, outside institutions such as prisons and nursing homes, into exactly one of three groups. The employed did any work for pay or profit during the survey week, even one hour, or were temporarily absent from a job, for example on holiday or sick leave. The unemployed had no job, were available to work, and had actively looked for work in the previous four weeks; people on temporary layoff waiting to be recalled also count. Everyone else is not in the labour force: retirees, most full-time students, full-time carers, and people who want a job but have stopped looking.
The key formulas
The labour force is everyone employed plus everyone unemployed: \(LF = E + U\). The unemployment rate is the share of the labour force without a job:
Notice the denominator. The rate is not the share of all adults without jobs; it ignores people outside the labour force. The labour force participation rate measures how many adults are in the labour force at all:
A third ratio, the employment-population ratio, is \(E\) divided by the adult population. It is useful because it cannot be flattered by people dropping out of the labour force.
Worked example: the United States in 2025
BLS annual averages for 2025 (over 11 months, because a federal shutdown stopped the October survey), in millions: adult civilian population 273.65, employed 163.49, unemployed 7.31. 1. Labour force: \(163.49 + 7.31 = 170.80\) million. 2. Unemployment rate: \(7.31 \div 170.80 \approx 0.0428\), about 4.3%. 3. Participation rate: \(170.80 \div 273.65 \approx 0.624\), about 62.4%. 4. Not in the labour force: \(273.65 - 170.80 = 102.85\) million. 5. Employment-population ratio: \(163.49 \div 273.65 \approx 59.7\%\). 6. Check: dividing the unemployed by the whole population instead would give about 2.7%, far below the official rate, which is why the denominator matters.
Worked example: when the rate falls for a bad reason
A hypothetical town has 50,000 adults: 30,000 employed and 2,000 unemployed. 1. Labour force: 32,000. Unemployment rate: \(2{,}000 \div 32{,}000 = 6.25\%\). Participation: \(32{,}000 \div 50{,}000 = 64\%\). 2. After months without success, 500 of the unemployed give up searching. They are now outside the labour force. 3. New labour force: 31,500. New unemployment rate: \(1{,}500 \div 31{,}500 \approx 4.76\%\). New participation rate: 63%. 4. Interpret: no one found a job, yet the unemployment rate fell by about 1.5 points. The employment-population ratio, 60% before and after, shows that nothing improved.
Hidden unemployment
The official count misses several kinds of slack, often called hidden unemployment. Discouraged workers want a job and have looked in the past year but have stopped searching because they believe none is available; they averaged about 509,000 people in 2025. More broadly, marginally attached workers want work and have looked recently, but not in the last four weeks. People working part time for economic reasons want full-time work but can only find part-time hours, about 4.8 million on average in 2025; the survey counts them as employed. Underemployment in a wider sense also includes people working well below their skills, such as an engineering graduate stacking shelves, which no official rate captures.
Broader measures: U-3 and U-6
The BLS publishes six measures of labour underutilisation, labelled U-1 to U-6. The headline rate is U-3, the one defined above. The broadest is U-6, which adds marginally attached workers, including discouraged workers, and people working part time for economic reasons. Its denominator is the labour force plus the marginally attached. In September 2026, seasonally adjusted, U-3 was 4.2% and U-6 was 7.6%. The two measures rise and fall together, but the gap between them widens in recessions, when more people are stuck in part-time jobs or give up searching.
How the numbers are collected
The unemployment rate comes from the Current Population Survey, a monthly household survey run by the Census Bureau for the BLS. It covers about 60,000 eligible households chosen to represent the country. Households stay in the sample for four months, leave for eight, then return for four more, so a quarter of the sample changes each month. Interviewers ask about activity during one reference week. The rate is not based on unemployment insurance claims: many unemployed people never claim benefits. The BLS usually publishes the results in its monthly Employment Situation report, generally on the first Friday of the following month.
The payroll survey and other limits
The same report includes the establishment, or payroll, survey, which asks about 119,000 businesses and government agencies, covering about 622,000 worksites, how many people are on their payrolls. It counts jobs rather than people, so someone with two jobs counts twice, and it leaves out the self-employed. The two surveys can briefly tell different stories. Both have limits. Some people tell the household survey they are looking for work without searching seriously; others work informally and report no job. Both surveys are samples, so small monthly changes may reflect sampling error rather than a real change.
Common mistakes to avoid
First, dividing the unemployed by the whole adult population instead of by the labour force. Second, counting everyone without a job as unemployed: retirees, students and carers who are not looking are outside the labour force. Third, assuming a falling unemployment rate always means more people found work; it can fall because people stop searching. Fourth, thinking the rate counts people claiming unemployment benefits. Fifth, forgetting that part-time workers who want full-time hours count as employed in the headline rate.
Key takeaways
Adults aged 16 and over are employed, unemployed or not in the labour force. To be unemployed you must be without work, available, and have actively searched in the past four weeks, or be on temporary layoff. The labour force is \(E + U\); the unemployment rate is \(U \div (E + U)\); participation is the labour force divided by the adult population. In 2025 the US rate averaged about 4.3% and participation about 62.4%. Discouraged and involuntary part-time workers form hidden unemployment, captured partly by U-6. The rate comes from a survey of about 60,000 households; job counts come from a separate payroll survey.
Flashcards
Flashcard
Did any paid work in the survey week, or temporarily absent from a job
Test your understanding
Question 1 of 5
A country has 240 million adults: 141 million employed, 9 million unemployed and 90 million not in the labour force. What is its unemployment rate?
Question 2 of 5
A retired teacher who has no job and is not looking for one is classified as:
Question 3 of 5
Some unemployed people stop looking for work because they believe no jobs exist. If nothing else changes, what happens?
Question 4 of 5
A woman works 20 hours a week in a shop but wants a full-time job and cannot find one. How does the headline unemployment rate, U-3, treat her?
Question 5 of 5
Where does the official US unemployment rate come from?
