A store is preparing to hire one more person, while the job seeker is weighing work, commuting, and care arrangements. Wages may look like a price, but labor transactions also involve skill matching, working conditions, long-term relationships, and incomplete contracts. We start with a pair of concrete choices, then examine why the same decline in the unemployment rate may represent entirely different changes.
The first half of this article explains the conditions for hiring and wages; the second half places individual mobility into employment statistics. The numerical values are set for teaching purposes; statistical definitions use the US Bureau of Labor Statistics (BLS) Current Population Survey (CPS) instructions as a clear example, and age ranges and survey rules in other countries should be checked separately.
What does it add for a firm to hire one more person?
Consider a small price-taking store where revenue per unit of output is 5, and equipment and store space are fixed for the day. Output for different numbers of workers is as follows, where the number of workers represents equivalent labor units in the same shift:
Number of Workers
Total Output
Marginal Output
Marginal Revenue
0
0
—
—
1
10
10
50
2
18
8
40
3
24
6
30
4
28
4
20
If the wage per shift is 35, the first worker adds 50 in revenue and the second adds 40, both higher than the marginal wage; the third adds only 30, so under these conditions, the firm chooses to hire two people. Directly checking the balance excluding common fixed costs: one worker yields 15, two yield 20, and three yield 15.
The extra revenue from an additional unit of labor is called its marginal revenue product (MRP). It is determined jointly by the marginal physical output and sales conditions, not by the "intrinsic value" of a person after leaving the team, equipment, or customers. An increase in product demand, new tools, or organizational improvements can all change a firm's willingness to hire.
Here we assume that additional products can be sold at a fixed price; with market power, increased sales volume may lower the price of other sales, requiring the use of marginal revenue rather than product price for calculation. It is also not stated that each position can precisely separate individual contributions as shown in the table.
Job seekers compare more than just the sticker wage
A job that pays 30 more per day but adds two hours of commuting and 20 in transportation costs has a net monetary improvement of only 10, and the time cost has not yet been counted. Another job with a lower wage may align better with care arrangements or provide training, and may also be more worth accepting.
Labor supply is therefore affected by alternative jobs, family arrangements, health, after-tax income, and non-monetary conditions. People who lose jobs may not immediately fill any vacancy: occupation, location, work hours, and qualifications may not match. The simultaneous existence of recruitment and job seeking does not alone prove that people are unwilling to work or that firms lack demand.
Higher wages may also reduce turnover, improve recruitment, and enhance work effort; when supervision is incomplete, firms do not always want to push wages down to the point where someone is willing to come. This continues the discussion from the information and contracts section: labor contracts cannot exhaust all future behaviors, and real wage formation requires institutional and organizational context.
Changing one condition: Does the employer have monopsony power?
In models where many independent employers compete for similar labor, raising a binding minimum wage may reduce the number of people firms are willing to hire. But if job seekers find it difficult to change jobs and there are few local employers, a single firm may face an upward-sloping labor supply curve: to hire one more person, it must raise wages, possibly even raising wages for existing employees together.
In this case, the marginal cost of labor may be higher than the wage of the last worker, leading the employer to limit employment. A well-designed minimum wage may simultaneously raise wages and employment within a certain range; if too high, it may still reduce employment. The assumptions behind these two mechanisms are different, and one cannot look at a single supply-demand graph and conclude that any real-world minimum wage adjustment will necessarily have the same directional effect.
To judge actual policy, one needs to compare affected positions, employer concentration, original wage distribution, working hours, benefits, prices, and entry/exit, rather than looking only at one employment figure. Wages are both a cost for firms and income for households; how this income transmits to other firms will be discussed further in the macroeconomic fluctuations section.
From job search experience to three statistical states
Given a reference period and statistical population, first divide into employed (E), unemployed (U), and not in the labor force (N). Using the US CPS as an example, the main statistics cover the civilian non-institutional population aged 16 and older; employment includes those who worked during the reference week or had a job but met the conditions for temporary absence, such as vacation. Unemployment usually requires no job, active job search during the preceding four weeks, and availability for work; people on temporary layoff awaiting recall are excepted from the active-search requirement.
Those who want to work but have stopped actively searching may no longer be counted in the main unemployment indicator U; retirees, students, or caregivers may also be in N, depending on survey behavior and conditions, not just identity labels. Unpaid housework is important for life, but it cannot be directly counted as paid positions in this market employment statistic.
For ease of calculation, fix a group of 100 people, with initial employment 60, unemployment 10, and not in the labor force 30. The labor force is E+U=70; the unemployment rate is U/(E+U)=14.29%, the participation rate is (E+U)/100=70%, and the employment-to-population ratio is E/100=60%.
Why do the same "five fewer unemployed" yield different results?
Independent change relative to initial state
E
U
N
Unemployment Rate
Participation Rate
Employment-to-Population Ratio
Initial State
60
10
30
14.29%
70%
60%
Five unemployed find jobs
65
5
30
7.14%
70%
65%
Five unemployed stop searching
60
5
35
7.69%
65%
60%
Five not in labor force start searching
60
15
25
20%
75%
60%
The first and second changes both cause the unemployment rate to fall, but only the first increases employment. The third causes the unemployment rate to rise, yet it may include people who previously left the labor force seeing opportunities again. Single ratios must be read together with numerators, denominators, and population flows.
Preparing the visual
Change conditions, check results
Out of 100, baseline employment is 60, unemployment 10, and nonparticipation 30. Hiring, stopping search, and starting search change counts and denominators differently.
Each switch starts from the same initial group, not sequentially accumulating three events. First predict which denominator will change, then check the colored blocks and ratios. Do not misread an unemployment rate of 20% as 20 out of all 100 people being unemployed; here it is 15/75.
What else needs to be seen for aggregate improvement?
If the number of employed increases but hours worked per person drop significantly, total labor input may not rise; finding a job may also mean taking a position with income or skill mismatches. Further looking at working hours, wages, long-term unemployment, underemployment, and group differences can supplement the blind spots of a single employment rate.
Conversely, a company laying off workers does not mean national net employment falls by the same amount; other firms may absorb the labor; national employment increasing does not mean the original unemployed all found better positions. Individual experiences, firm recruitment, and overall statistics are at different levels, and need to be connected using the same group of people or clearly comparable cohorts.
Reading a news report
The report says the unemployment rate fell, but the number of employed did not change and the participation rate fell. Can you directly interpret this as "firms expanding recruitment"? What observation needs to be supplemented?
Expand reasoning
No. This is compatible with some unemployed leaving the labor force. One should check flows between employment, unemployment, and not in the labor force, population changes, job search status, and statistical reference periods; if possible, look at transfers within the same cohort. Checking numerators and denominators first is more informative than simply labeling a falling unemployment rate as good news.
In the store table, if the wage changes from 35 to 45, and product prices and output relationships remain unchanged, how many people will the firm choose to hire? Can this result directly represent the effect of wage policy on the entire economy?
Expand reasoning
The first person adds 50 in revenue, and the second adds only 40, so the firm chooses one person. This should not be generalized directly: this is a local choice of a single firm under conditions of fixed demand, equipment, and competition, without incorporating household income feedback, different employers, entry/exit, or labor market power. To expand the conclusion, the model and evidence scope must be correspondingly expanded.
Sources and Further Reading
These references support concepts and statistical definitions; the numerical cases and diagrams are original synthetic teaching examples.