The employment report shows more jobs and a higher unemployment rate. That is not necessarily a statistical mistake. Payroll jobs and household unemployment are measured differently, and the labor force itself can grow when people start looking for work. More people may find jobs while an even larger number enter the search. To understand what changed, readers need the numerator, the denominator and the source survey, not just two apparently conflicting headlines.

The two figures can rise together without any contradiction. They describe related but different aspects of the labor market. New people enter the labor force, unemployed workers find positions, employees move between jobs, and statistical surveys use different samples and definitions.

Understanding those mechanics is useful when evaluating economic news. It also prevents the opposite mistake: treating every positive payroll number as proof that finding work has become easier for everyone.

Job creation and unemployment measure different things

In US labor statistics, one headline measures employment reported by businesses through a payroll survey. Another relies on a household survey that classifies people according to their labor-force status.

The US Bureau of Labor Statistics' employment report FAQ explains that the household survey and establishment survey have distinct sample designs, definitions and purposes. A payroll count is a measure of jobs on employers' payrolls within the survey scope. The household survey is designed to assess whether people are employed, unemployed or outside the labor force.

One person may hold more than one payroll job. A self-employed person may be counted as employed in the household survey without appearing as a conventional nonfarm payroll employee. The populations and treatments of certain categories differ.

It follows that a monthly increase in payroll employment need not be mirrored by exactly the same change in the number of employed individuals. The indicators often move in related directions over longer periods, but they are not duplicate measurements of one underlying total.

For news readers, the first question should be which statistic a headline has cited.

The unemployment rate has a denominator

The unemployment rate is the number of unemployed people divided by the labor force. The labor force includes people classified as employed or unemployed. It excludes people who are neither working nor actively seeking work under the survey's rules.

The BLS explanation of how unemployment is calculated describes the requirements for being counted as unemployed. Generally, someone must lack employment, be available for work and have actively looked for work within the relevant period, with certain exceptions such as temporary layoffs.

The practical consequence is easy to miss. An adult who stops searching for work may move out of the measured labor force. Their departure could reduce the unemployment rate even though that person's situation has not improved.

The reverse can happen when people return to job seeking. If more people begin looking for work than employers hire during the period, the number classified as unemployed may rise. That can occur even as jobs are added.

This does not mean that a higher unemployment rate is inherently good. It means that participation changes are necessary context before interpreting the figure.

A numerical example that reconciles the headlines

Consider an invented economy with 1,000 people in its labor force. Of those, 950 are employed and 50 are unemployed. The unemployment rate is 5%.

In the next month, employers create 20 net new jobs. Separately, 40 previously inactive people begin actively searching for work. Suppose the employment gains translate into 20 additional employed people, with no other relevant changes, and that the newly active group does not all find work immediately.

The labor force becomes 1,040. Employment rises to 970. Unemployment rises to 70. The new unemployment rate is approximately 6.7%.

There are more employed people than before, yet the unemployment rate is higher. The increase is not an arithmetic trick. The labor force grew faster than employment.

This is a deliberately simplified illustration, not a statistical reconstruction of an actual monthly report. In real data, household and establishment estimates come from surveys and include sampling uncertainty, seasonal adjustments and additional flows between statuses.

Still, the example identifies the causal mechanism. More job openings and more people seeking work can coexist, and the balance between those movements shapes the unemployment rate.

Participation changes the interpretation

A declining unemployment rate can accompany improving employment opportunities. It can also reflect fewer people looking for work. To distinguish those stories, examine the labor-force participation rate and the employment-to-population ratio.

The participation rate compares the labor force with the relevant civilian noninstitutional population. The employment-to-population ratio compares employed people with that population. Together, they indicate how much of the eligible population is engaged in the labor market and how much is employed.

The BLS household survey definitions spell out these concepts and the population covered. Age structure, schooling, retirement and family obligations can all influence participation, so changes should be interpreted with demographic context.

A rise in participation may reflect people becoming optimistic about finding work. It may also reflect financial pressure that makes job seeking necessary. A fall may reflect retirement, education or discouragement. The aggregate number cannot identify motivations on its own.

For that reason, a careful labor-market assessment looks at multiple indicators over time rather than assigning a single explanation to one monthly movement.

Payroll survey and household survey differences matter

The establishment survey gathers information from employers about jobs, earnings and hours. The household survey asks individuals about employment status and other characteristics. Each has advantages and sources of error.

The BLS comparison of payroll and household survey trends explains why the employment levels reported by the surveys are different, and why their short-term movements can diverge.

Payroll data are well suited to examining jobs by industry and changes in total payroll positions. Household data capture broader forms of work and allow analysis by demographic groups. Neither should be discarded merely because it tells a more complicated story.

Survey estimates are also revised or updated. Payroll data are often revised as additional employer information becomes available and benchmarked against administrative records. A later release may change the initial estimate. That is a normal feature of a statistical process, though revisions can affect how historians judge a particular month.

Readers should avoid treating preliminary estimates as exact counts of every person hired or dismissed. Their purpose is to describe economic patterns using a consistent methodology.

A positive jobs report can conceal weak job quality

Counting jobs alone does not show whether wages cover living costs, whether hours are stable or whether workers have the benefits and security they need. A rise in payroll employment may be concentrated in sectors with lower compensation or in positions offering fewer hours than workers want.

A labor market can also have substantial underemployment. Some people working part time would prefer more hours; others may accept positions that do not use their skills because alternatives are limited. Official statistical agencies may publish measures of underutilization that complement the headline unemployment rate.

The question of job quality deserves its own data. Look at average hours, wage growth relative to prices, full-time and part-time employment, sector-level patterns and measures of involuntary part-time work where available.

Be careful with averages. Rising average hourly earnings can sometimes reflect changes in the composition of employment rather than equal pay increases for every worker. If lower-paid jobs disappear, the average among remaining workers might rise without anyone receiving a raise.

Labor statistics are rich precisely because employment is more than a binary condition.

Population growth changes the scale

An economy serving a growing working-age population needs job creation just to maintain the same employment proportions. A net increase of 50,000 jobs could be strong progress for one labor market and insufficient for another, depending on the size and growth of its workforce.

Comparing job growth with working-age population and participation trends can reveal whether opportunities are expanding fast enough. The relevant demographic group and definition should be specified; figures may differ among countries.

International comparisons require added care. Official unemployment definitions generally follow international principles, but surveys, coverage, informal employment, eligibility rules and data quality vary. A rate in one economy cannot be interpreted solely by comparing it with another country's number.

A national statistic can also hide substantial regional differences. A major city might experience strong hiring while industrial towns lose employers. Industry-level and regional data help explain why people's experiences diverge from a national headline.

The best labor-market commentary acknowledges those differences instead of presenting one number as the whole economy.

Four questions for the next employment report

When a headline says jobs rose or unemployment increased, begin by checking the source. Is it the payroll survey, the household survey or a different data provider? Determine the reference month, estimate status and whether figures are seasonally adjusted.

Next, examine labor-force participation and the number of employed and unemployed people. If unemployment rises, did employment fall, did the labor force expand or did both occur? The components usually explain much of the apparent contradiction.

Third, inspect wages, hours and industry or regional patterns to assess how broadly the changes are distributed. A gain concentrated in one sector may not signal improving prospects for all workers.

Finally, check revisions and compare several months. One survey release can be noisy. A persistent trend supported by multiple measures is more informative than a dramatic one-month change.

A good labor-market report can contain both encouraging and troubling evidence. More jobs are valuable. So is knowing whether workers can find them, whether pay keeps pace with costs and whether people are entering or leaving the labor force. Reading the indicators together is the only way to avoid choosing a conclusion before examining the data.

SOURCES & FURTHER READING

References and background

Our explainers link to source documents so readers can explore the evidence and interpret it in context. Websites and policies may change after publication.

  1. Bureau of Labor Statistics, *Employment Situation Frequently Asked Questions*www.bls.gov
  2. BLS, *Comparing Employment From the BLS Household and Payroll Surveys*www.bls.gov
  3. BLS, *How the Government Measures Unemployment*www.bls.gov
  4. BLS, *Concepts and Definitions*www.bls.gov
Spot a factual issue? Request a correction ↗Read our editorial standards ↗