Macroeconomics › Section H · Unemployment › Topic H.2
Patterns of unemployment
The unemployment rate swings with the business cycle, but it never reaches zero and shows no long-run upward trend. Underneath the national figure, the young, the less educated and some minority groups face much higher rates than others.
Looking beneath the headline
The previous topic, "Defining and computing the unemployment rate", explained how the rate is measured. This topic asks what the numbers show: how US unemployment has moved since the survey began in 1948, which groups face the highest rates, how long people stay unemployed, and how the US compares with other countries. The patterns matter for policy: a rise that hits every group in a recession needs a different response from a gap that persists for one group in all years. All US figures here are BLS data, as published by the BLS or the St Louis Fed's FRED database, seasonally adjusted for monthly figures. The 2025 annual averages cover 11 months, because a federal shutdown stopped the October 2025 survey.
Ups and downs since 1948
US unemployment has ranged from a low of 2.5% in May 1953 to a high of 14.8% in April 2020. Severe recessions stand out as peaks: 10.8% in November 1982, 10.0% in October 2009 after the global financial crisis, and the pandemic peak of 2020. Before the survey existed, later estimates put unemployment at about a quarter of the labour force in 1933, during the Great Depression. In good times the rate has dipped to about 3.5% or a little lower, as in 1969, 2019 and 2023. Only in the early 1950s did it stay near 3% for long, and it has never come close to zero.
Unemployment lags the economy
Unemployment rises in recessions and falls in expansions, but its timing trails output. Firms are slow to hire at the start of a recovery, because they first ask existing staff to work longer hours and wait to be sure that demand has really returned. So unemployment often keeps rising after output has started to grow again. Lay-offs can also come late in a downturn, once firms give up hope of a quick rebound. The pandemic was an unusual exception: unemployment jumped by more than ten points in a single month and then fell unusually fast as businesses reopened.
Worked example: two recessions compared
BLS data: in December 2007, 7.65 million people were unemployed and the rate was 5.0%. In October 2009 the figures were 15.35 million and 10.0%. The recession officially ended in June 2009. 1. Rise in unemployed: \(15.35 - 7.65 = 7.70\) million, roughly a doubling over 22 months. 2. The peak came four months after the recession ended, showing the lag. 3. In the pandemic, unemployment went from 5.75 million in February 2020 to 23.08 million in April 2020. 4. Rise: \(23.08 - 5.75 = 17.33\) million in two months, more than twice the 2007 to 2009 increase, in a tenth of the time.
No long-run trend
Over more than seven decades there is no lasting upward or downward trend in US unemployment. That is striking, because the labour force has nearly tripled since 1948, women's participation rose enormously from the 1960s, and wave after wave of automation and trade has reshaped the kinds of jobs available. Each time, the economy eventually created enough new jobs for the larger workforce, and the rate drifted back into its usual range. Fears that technology will cause ever-rising unemployment are two centuries old; so far, the data have not borne them out.
Differences by age and sex
In 2025, BLS annual averages show unemployment of about 14.0% for 16 to 19-year-olds and 8.3% for 20 to 24-year-olds, against 3.6% for people aged 25 to 54 and 3.0% for those 55 and over. Young people change jobs often while they search for a good match, and many are new to the labour market with little experience. Older people who lose jobs often retire instead of searching, which takes them out of the labour force entirely. Men's and women's rates, which once differed noticeably, have been close since about 1980; in 2025 the rates for adults aged 20 and over were 4.0% for men and 3.8% for women.
Differences by race and education
In 2025 the unemployment rate averaged about 6.9% for Black workers, 5.1% for Hispanic workers, and 3.8% for White and 3.6% for Asian workers. For decades the Black rate has run close to twice the White rate, in booms and recessions alike. Economists attribute part of this gap to differences in education, location and job networks, and studies continue to find evidence of discrimination in hiring. Education is just as strong a pattern. Among adults 25 and over in 2025, unemployment was about 6.1% without a high school diploma, 4.3% for high school graduates, 3.4% with some college or an associate degree, and 2.6% with at least a bachelor's degree.
Worked example: how group mix moves the total
A hypothetical labour force has two groups: young workers, with 12% unemployment, and older workers, with 4%. 1. If young workers are 15% of the labour force, the overall rate is the weighted average \(0.15 \times 12 + 0.85 \times 4 = 1.8 + 3.4 = 5.2\%\). 2. If a baby boom raises the young share to 25%, the rate becomes \(0.25 \times 12 + 0.75 \times 4 = 3.0 + 3.0 = 6.0\%\). 3. Interpret: neither group's rate changed, yet the national rate rose 0.8 points purely from the mix. 4. Check: each result lies between 4% and 12%, closer to the larger group, as a weighted average must.
Reasons and duration
The BLS also classifies the unemployed by why they are looking: job losers, who were laid off or fired; job leavers, who quit; re-entrants, returning after time out of the labour force; and new entrants seeking a first job. Job losers make up a much larger share in recessions. Duration matters as well. The long-term unemployed, out of work for 27 weeks or more, were about 27% of the unemployed in September 2026, but 45.5% in April 2010, after the financial crisis. Long spells are especially harmful, because skills fade, contacts weaken and some employers hesitate to hire people with long gaps.
International comparisons
World Bank figures using International Labour Organization modelled estimates put 2024 unemployment at about 2.5% in Japan, 3.4% in Germany, 4.0% in the United States, 4.4% in the United Kingdom, 6.4% in Canada, 7.4% in France, 11.4% in Spain and 32.3% in South Africa. Comparisons need care. Japan's low rate partly reflects firms keeping underused staff and job losers leaving the labour force. In many low-income countries, most people work informally in farming, trading or casual labour, often without benefits, so few can afford to be unemployed at all, and low unemployment does not mean good jobs.
Common mistakes to avoid
First, expecting unemployment to fall as soon as a recession ends; it usually lags. Second, assuming the rate can reach zero in a healthy economy. Third, reading a low rate in a poor country as a sign of a strong labour market, when many people work informally. Fourth, forgetting that a change in the age mix can move the national rate even when no group's rate changes.
Key takeaways
Since 1948, US unemployment has ranged from 2.5% (1953) to 14.8% (April 2020), peaking in recessions and lagging recoveries, with no long-run trend and never near zero. Rates are highest for the young, for people with less education, and for Black workers, whose rate has long been close to twice the White rate; men's and women's rates are now similar. The unemployed include job losers, leavers, re-entrants and new entrants, and the long-term share swells after deep recessions. International rates differ widely, but definitions, labour market institutions and informal work make comparisons difficult.
Flashcards
Flashcard
2.5% in May 1953
Test your understanding
Question 1 of 5
In a labour force, 20% of workers are young with 10% unemployment and 80% are older with 5% unemployment. What is the overall unemployment rate?
Question 2 of 5
A recession officially ends in June. What usually happens to the unemployment rate?
Question 3 of 5
Which statement about US unemployment since 1948 is supported by the data?
Question 4 of 5
Why are unemployment rates for workers aged 55 and over usually low compared with those for teenagers?
Question 5 of 5
A low-income country reports an unemployment rate of 3%. Why should you be cautious about concluding its labour market is strong?
