Macroeconomics › Section H · Unemployment › Topic H.4

Long-run causes: the natural rate

Even in the best of times, some people are between jobs or lack the skills employers want. That baseline, the natural rate of unemployment, is never zero, and it is shaped by demographics, technology and public policy.

Why unemployment is never zero

The previous topic, "Short-run causes: cyclical unemployment", explained why unemployment rises in recessions. But even at the top of a boom, US unemployment has rarely fallen much below 3.5%, and some European countries have had rates near 10% for long stretches in good years and bad. Economists call the rate of unemployment that would exist with the economy neither booming nor in recession the natural rate of unemployment. "Natural" does not mean fixed or desirable. It reflects the economic, social and political conditions of a particular time and place, and it can change.

Frictional unemployment

A market economy churns constantly. Firms fail, shrink or move while others start or expand. BLS job openings data show about 5.2 million hires and a similar number of separations in an average month of 2025, even though total employment barely changed. Moving between jobs takes time: learning about openings, applying, interviewing, perhaps relocating. Frictional unemployment is the unemployment that occurs while workers and jobs search for each other. Some of it is healthy. A graduate who spends six weeks finding a well-matched job may be far more productive for years than one who grabs the first offer.

What sets frictional unemployment

Frictional unemployment depends on how quickly workers and employers find each other. Better information shortens searches: online job boards and professional networking sites have made it much easier to discover openings than leafing through newspaper advertisements. Willingness to move matters too, and that can depend on housing costs and family ties. So does the age mix of the labour force. Young people try out jobs and change often, so an economy with many young workers has more frictional unemployment. When the large baby boom generation, born from 1946 to 1964, entered the labour force in the 1960s and 1970s, it pushed the natural rate up; its ageing later helped pull it down.

Structural unemployment

Structural unemployment arises when workers' skills or locations do not match the jobs available. It happens when demand shifts away from an industry, as with coal mining, or when technology replaces a task, as with switchboard operators, and the displaced workers lack the skills that growing sectors want. It also affects people who never gained marketable skills. Structural unemployment tends to last longer than frictional unemployment, because fixing it may require retraining or moving. Education helps most: workers with strong general skills can be retrained more easily. New technology often destroys some jobs while creating demand for others, so the problem is usually the transition rather than a lasting lack of work. A smaller type, seasonal unemployment, follows the calendar, as with ski instructors in summer; seasonally adjusted figures remove it.

Full employment and potential GDP

The natural rate links to two other ideas. The economy is at full employment when actual unemployment equals the natural rate; any remaining unemployment is frictional or structural, not cyclical. At full employment, real GDP equals potential GDP, the output the economy can sustain. When unemployment is above the natural rate, cyclical unemployment exists and output is below potential. When unemployment is below the natural rate, output is above potential, which can last only a while, like a factory running on overtime: wages and prices then tend to rise faster. Cyclical unemployment can therefore be negative in a boom.

Worked example: splitting unemployment by type

A hypothetical labour force of 160 million has 2.4 million frictionally unemployed, 4.0 million structurally unemployed, and 10.0 million unemployed in total. 1. Natural rate: \((2.4 + 4.0) \div 160 = 0.04\), or 4.0%. 2. Actual rate: \(10.0 \div 160 = 0.0625\), or 6.25%. 3. Cyclical unemployment: \(6.25 - 4.0 = 2.25\) percentage points, which is \(10.0 - 6.4 = 3.6\) million people. 4. Interpret: the economy is below full employment, so output is below potential. 5. Check: \(3.6 \div 160 = 0.0225\), matching 2.25 points. Only the cyclical 3.6 million would be expected to find work simply because the economy recovers.

Productivity surprises

Over time, wages follow productivity, but pay is usually set from recent experience. When productivity growth slows unexpectedly, as it did after 1973, when the BLS measure fell from about 2.8% a year to about 1.4%, wages keep rising at the old pace. Labour costs then outrun output per worker, firms hire less, and the natural rate rises for a time. The reverse happened after about 1995, when productivity growth jumped to about 3% a year: pay lagged behind, hiring was cheap relative to output, and US unemployment fell to about 4% by 2000 while inflation stayed fairly low. Once expectations catch up, the effect fades, but that can take years.

Public policy and the natural rate

Policy affects both sides of the labour market. On the supply side, generous and especially long-lasting benefits lower the cost of searching longer, which can raise unemployment. US unemployment insurance, created in 1935, is a federal-state programme funded mainly by taxes on employers and run by states, usually paying for up to 26 weeks, fewer in some states. Help with job search or retraining can shorten spells instead. On the demand side, heavy hiring costs, strict rules on dismissals, and minimum wages set high relative to local pay can make firms reluctant to hire. Economists often cite such rules to explain why many European countries have had higher natural rates than the US.

Worked example: is the economy at full employment?

The Congressional Budget Office estimates a noncyclical rate of unemployment, its version of the natural rate, of about 4.4% for the third quarter of 2026. The BLS rate for September 2026 was 4.2%. 1. Gap: \(4.2 - 4.4 = -0.2\) percentage points. 2. Unemployment is slightly below the estimated natural rate, so measured cyclical unemployment is about zero or slightly negative. 3. Interpret: the economy is roughly at full employment. 4. Caution: estimates of the natural rate are uncertain by a point or more, and CBO's estimate was about 6.2% in the late 1970s, so a gap of 0.2 points is well within the margin of error.

Matching policy to the cause

Because different kinds of unemployment have different causes, they need different remedies. Cyclical unemployment comes from weak demand, so policies that support total spending can reduce it. Frictional unemployment falls with better job information and easier moves. Structural unemployment needs training, education and help relocating. Using demand stimulus against structural unemployment mostly raises inflation without matching workers to jobs. Governments cannot stop firms failing or populations ageing, but they can design benefits, rules and training so that workers and employers meet quickly. The topic "Causes of unemployment around the world" compares how countries do this.

Common mistakes to avoid

First, thinking the natural rate is zero, or that it is fixed by nature: it reflects conditions and changes over time. Second, assuming the natural rate includes cyclical unemployment; it is frictional plus structural only. Third, treating all frictional unemployment as waste, when good matches raise productivity. Fourth, assuming that unemployment below the natural rate can last indefinitely. Fifth, using demand stimulus to fix a skills mismatch.

Key takeaways

The natural rate of unemployment is the rate when the economy is neither booming nor in recession: frictional plus structural unemployment. Frictional unemployment comes from job search in a churning economy and depends on information, mobility and age mix. Structural unemployment comes from mismatches of skills or location. At the natural rate the economy is at full employment and real GDP equals potential. Unexpected productivity changes shift the natural rate temporarily, and policies on benefits, hiring costs and dismissals affect it lastingly. CBO's estimate for the US is about 4.4% in 2026, down from about 6.2% in the late 1970s.

Flashcards

Flashcard

Natural rate of unemployment

Unemployment with the economy neither booming nor in recession

Frictional plus structural.

Test your understanding

Question 1 of 5

A labour force of 200 million has 3 million frictionally and 5 million structurally unemployed, and 12 million unemployed in total. What is the natural rate of unemployment?

Question 2 of 5

A newly qualified accountant turns down two poor offers and spends five weeks finding a job that suits her skills. What type of unemployment is this?

Question 3 of 5

Unemployment is 3.5% and the natural rate is estimated at 4.5%. What does this suggest?

Question 4 of 5

Productivity growth suddenly slows, but wage increases continue at their old pace for several years. What happens to the natural rate of unemployment meanwhile?

Question 5 of 5

Which policy is best suited to reducing structural unemployment among former factory workers whose skills are no longer in demand?

← Previous topicNext topic →