← Back to the lecture: Factor Markets and Labor, Part A: What an Extra Worker Is Worth
Lecture 17 Recap
Why firms hire workers for what they help make, how an owner decides how many to hire, what a wage change or a price change does to that choice, and why machines and AI do not use up the work there is to do.
Class meeting 23 · November 12, 2026 · 10-minute review
A firm hires a worker for what the worker helps make, so the demand for labor is derived from the demand for the product. What one more worker is worth to the firm is the extra output that worker adds, valued at the price of the output. The firm keeps hiring while one more worker adds at least what that worker must be paid, and the wage it must pay is set by what the worker could earn elsewhere. A higher wage means fewer workers; a higher product price means more. New machines change which work is worth doing, but they do not use up the work there is to do.
How to use this page
This recap covers the November 12 class. Class covered the slides through the college wage premium: derived demand, the extra fish each crew member adds, the hiring rule, a higher wage and a higher fish price, teamwork, the winch and the lump-of-labor mistake, AI, and why pay differs. The last two slides (human capital and harder jobs) and TopHat question 5 were not reached. Read the five-step diagnostic first, then the lecture sections, then the checks. The separate Lecture 17 Transcript has the full explanation.
Use these steps whenever a question gives you a firm that sells its output at a market price and hires workers at a market wage.
What you should be able to do
Class started with the plan for the rest of the semester (see below). The lecture then followed one example most of the way through. Suppose that on an island, other jobs pay $80 a day. Someone gets a fishing boat. The boat is one of many selling into a large fish market at $10 a fish, so its own catch cannot move the price. How many crew should the owner take on?
Derived demand
Labor markets are markets for an input, or a factor of production, and they use the same tools as the rest of the course. A restaurant hires cooks because diners value the meals the cooks help make. The demand for work is derived from the demand for what the work helps make: the demand for doctors comes from the demand for health care, and the demand for cattle from the demand for steak, milk, and leather. Doctors are paid a lot because people will pay a lot to stay healthy, not the reverse.
The instructor contrasted this with the labor theory of value (not on the slides), the older idea that a good’s value comes from the labor that goes into it. It runs the wrong way: what buyers will pay for the output sets what firms will pay for the inputs. Marx built on that idea to argue that profit must mean exploitation. The instructor’s answer: suppose an owner pays a worker $70 and sells the output for $100. The owner carries the risk. If no one buys, the owner earns nothing but still owes the $70. Profit is in part the return for bearing that risk.
So what a worker is worth to a firm is a market value. It is not a fixed personal trait, and it is not a person’s worth as a human being. A skilled mechanic with good tools adds more than the same mechanic with poor ones.
Marginal product · TopHat
With the same boat, the crew catch these totals a day:
| Crew | Total catch | Extra fish |
|---|---|---|
| 1 | 12 | 12 |
| 2 | 30 | 18 |
| 3 | 46 | 16 |
| 4 | 58 | 12 |
| 5 | 66 | 8 |
| 6 | 70 | 4 |
| 7 | 72 | 2 |
| 8 | 64 | −8 |
The extra fish from one more crew member is the marginal product. The second adds 18, more than the first: two can work the boat better than one. From the third on, each adds less: diminishing marginal returns. The boat stays the same size, so the deck gets crowded. The total still rises through seven crew; the eighth cuts the catch by 8 fish, so no owner would take on an eighth even if that worker were free. A bigger boat would let more crew work before diminishing returns set in, but they would set in again.
TopHat check. Another boat catches 10 fish with 1 crew member, 25 with 2, 45 with 3, 60 with 4, and 70 with 5. Which crew member is the first to add fewer fish than the one before? The extra fish are 10, 15, 20, 15, and 10, so the answer is the fourth. Most of the class got it. Quite a few chose the fifth: the fifth does add less, but diminishing returns start with the fourth.
The hiring rule · TopHat
A count of fish does not say what the fish are worth. (If fish were poisonous, a big catch would be worth nothing.) Multiply each crew member’s extra fish by the $10 price: the value of marginal product. The crew add $120, $180, $160, $120, $80, $40, and $20 a day. A study saying fish is healthy would raise the price of fish, and so the value of each crew member’s catch.
Why $80? No one would work on the boat for $79 a day when other island jobs pay $80. Employers compete with employers, and that bids wages up; workers compete with workers, and that bids wages down. The instructor’s examples (not on the slides): his own pay depends on what other universities and private firms would pay him; and during COVID, when fast-food restaurants in San Angelo offered $20 an hour and $2,000 signing bonuses, his university’s janitorial staff, paid about $13 an hour, left, and the university had to raise their pay. Its rivals for workers were McDonald’s and Home Depot, not other universities.
A job can also pay less if people like it more: someone who loves being out on the water might fish for less than $80. For now the two jobs are assumed equally pleasant.
The first crew member adds $120 and costs $80, so the owner nets $40. The next ones net $100, $80, and $40. The fifth adds exactly $80, the wage; we count the fifth in. The sixth would add $40 and cost $80, so the owner stops at five. The catch is worth 66 × $10 = $660, wages are 5 × $80 = $400, and the owner keeps $40 + $100 + $80 + $40 = $260 a day. The fifth crew member is just as well off on the boat as in an island job. (A student asked about this case. The instructor noted that real firms hire hours, not whole people, so the choice is smoother than in the example.)
The minimum wage (not on the slides). Suppose other jobs paid $40 a day: the owner would take on six. Now suppose a law says no one may be paid less than $80 a day. The sixth crew member adds only $40, so that worker is let go and earns nothing from the boat. The workers who keep their jobs gain; the sixth, whom the law was meant to help, loses. In the instructor’s view, the evidence shows that minimum wages raise unemployment among low-skilled workers; he added that a minimum wage can help in some high-skilled markets with few employers (monopsony).
TopHat check. The wage that other island jobs pay rises from $80 to $100 a day. How many crew does the owner take on now? Four. The third adds $160 and the fourth $120, both more than $100; the fifth adds only $80, so the owner would lose $20 a day on the fifth. The majority got it; three students said five. (A student asked during the poll how wages are measured. The instructor’s answer: a wage is paid for a worker’s time, so think of dollars per worker-hour, or here per day.)
A higher wage, a higher fish price · TopHat
At $100 the owner keeps 58 × $10 − 4 × $100 = $180 a day, down from $260. The fifth crew member goes back to an island job. That is the right outcome for everyone: on the boat that worker would add $80 of fish while giving up $100 elsewhere, so the island would be $20 a day poorer. The higher wage moved the worker, and no one had to plan it. The instructor tied this to Adam Smith: each person acts in self-interest, and the wage does the job a wise planner would do.
The bars give the boat’s demand for crew: at $140 a day the owner takes on 3, at $100 4, at $60 5, and at $30 6. A higher wage moves along this demand. A student asked why the owner hires three at $140 when the first crew member adds only $120. The answer: the decision is made at the margin. The first three together add $460, more than the $420 they cost, and the owner stops at the third, who adds $160.
Suppose more people want fish, and its price rises to $20. Other jobs still pay $80. Each crew member’s extra fish is worth twice as much: $240, $360, and so on. The owner now takes on six (the sixth adds 4 × $20 = $80, and we count the sixth in) and keeps 70 × $20 − 6 × $80 = $920 a day. The whole demand for crew shifted up because the fish became worth more: the market is telling the owner to catch more fish.
The instructor added (not on the slides) that an employer gets a kind of consumer surplus from hiring: at $20 a fish the owner would pay up to $240 for the first crew member but pays only $80, a gain of $160. Workers’ gains in a labor market are the producer surplus.
TopHat check. The crew add 12, 18, 16, 12, 8, 4, and 2 fish, other jobs still pay $80, and the price of fish falls from $10 to $6. How many crew? Three. At $6 the crew add $72, $108, $96, and $72. The first adds less than $80, but no owner stops at one: the second and third add $108 and $96. The fourth adds $72, less than $80, so the owner stops at three and keeps 46 × $6 − 3 × $80 = $36 a day. The boat and the crew did not change; the fish are worth less, so the demand for crew fell.
An aside on pay at the top (not on the slides). Some students had been talking about news reports about NVIDIA’s chief executive, and the stock fell sharply. The instructor’s point: if a firm’s owners could easily hire someone just as good for the same pay, the loss of a chief executive would not move the stock much. A big drop suggests he was worth more to the firm than he was paid.
Teamwork
Alone, either of the first two crew members catches 12 fish; together, 30. (The table’s 12 and 18 come only from the order we counted them.) Take away either one and the catch falls by 18, so each “added” 18. But 18 + 18 = 36, more than 30. Both comparisons are right, and both cannot be paid. With five crew, take any one away and the catch falls by 8 fish, or $80; five times $80 is $400, but the catch is worth $660. The owner, who supplies the boat, keeps what is left after wages.
So competition pushes each person’s pay toward what one more such worker adds. It does not measure what each person “really” produced: in a team there is no single answer. The instructor’s examples: the “winning” field goal or basket needed every earlier score, and a running back’s touchdown needed the offensive line’s blocks.
Machines, jobs, and AI · TopHat
A higher wage does more than shrink the crew: over time it makes labor-saving gear worth buying, as commercial boats use nets and pots, not fishing poles. Suppose every boat gets a new powered winch, so each boat lands the same catch with fewer deckhands. The winch also creates work: building, running, and fixing winches, and handling extra fish. A deckhand might become a winch mechanic.
Do fishing jobs rise or fall? Two effects pull in opposite directions. Catching fish costs less, so the boats catch more, which can call for more hands. But with a winch on every boat, the fleet catches more and the market price of fish falls, which lowers the value of each crew member’s catch. Crew jobs = boats × crew per boat: they rise only if the number of boats grows enough to make up for the smaller crews.
The lump-of-labor mistake is the view that there is a fixed amount of work, so every task a machine takes over is a job gone for good. It is wrong for two reasons the instructor gave. First, scarcity: our wants are never used up, so as people get richer they want more, and the demand for workers, derived from those wants, grows. The question is never whether there is work to do, but which work is worth doing. Second, when a firm’s costs fall, it expands output and hires more inputs, labor included.
What is true: a change in technology can hurt the people whose skills, homes, or towns are tied to the old work, and moving to new work can be slow.
Many students worry that AI will take everyone’s job. The instructor called predictions of mass unemployment from AI a case of the lump-of-labor mistake: if AI makes us much richer, we can afford things we could not before, and someone will produce them. Some people do lose out; students who chose to “learn to code” a few years ago face AI that writes code. But, he said, the demand for software engineers has not yet fallen much, partly because someone must check what the AI produces. Ask of AI what we asked of the winch: which tasks it does more cheaply (those workers can lose out), whom it makes more productive (a worker who adds more is worth more), and what buyers do with what they save (new work can follow). How AI will change jobs and pay is a question for the evidence.
His historical examples (not on the slides): when the Model T made cars affordable, the economy needed fewer farriers, blacksmiths, horse ranchers, and buggy makers, and those people moved to other work; companies once had rooms of typists and no IT departments; and before electronic computers, “computers” were people who did calculations by hand, as in the film Hidden Figures.
TopHat check. Winches let every boat land its catch with half as many crew members. Cheaper fish then lead the fleet to run 50% more boats, each with the smaller crew. What happens to the total number of crew jobs? 25% fewer. Each boat needs half the crew, and there are 1.5 times as many boats: 0.5 × 1.5 = 0.75. The fall is real but much smaller than the “cut in half” most people expect, because the number of boats grew. With twice as many boats, crew jobs would not change at all.
Why pay differs
What employers will pay depends on what one more such worker adds, valued at the price of what is made, so skill, effort, tools, and the demand for the product all matter. Who can and will do the work depends on the cost of getting the skill (including pay given up while training), on how pleasant, safe, or steady the job is, and on barriers such as licenses, discrimination, or limits on who may enter. A money wage is not a full measure of a job, nor a verdict on anyone’s merit.
The slide’s figures, from the Bureau of Labor Statistics: in 2025 the median full-time worker age 25 or older earned $1,578 a week with a bachelor’s degree and $966 a week with only a high school diploma. (In class the instructor called these averages; they are medians.) Draw two labor markets. In the college market, demand is higher, because these skills add more; and supply is smaller, because college costs tuition and lost pay. Part of the gap is who goes to college, not only what college teaches.
At these wages no one gains by switching. If the college wage were higher still, more people would go to college until the extra supply pushed it down. For the worker on the fence, the extra pay over a career just covers the cost of the degree. A student asked whether the high-school group includes people with more training. The instructor’s answer: no, it is only people whose highest schooling is a high school diploma; the trades likely pay a bit more.
His application (not on the slides): subsidizing college moves workers out of the high-school market and into the college market, which pushes college wages down and high-school wages up. So, holding demand fixed, the trades pay well partly because so many people are pushed toward college; with less of that push, more people would enter the trades and their pay would fall.
Not reached in class
The deck’s last two slides were not shown. Human capital: skills from school, training, and experience are wealth you carry with you; like a machine, they take resources to build and pay off for years, but they cannot be taken from you, and a lender cannot repossess them. Harder jobs must pay more to fill: if deckhands can work on day boats, home each night, or on boats out for weeks, fewer will spend weeks away at any wage, so those boats must pay more ($100 a day against $80 on the slide). The $20 is a compensating differential. The instructor made the same point early in class: a job you enjoy can pay less, as his own academic job, with summers off, pays less than he might earn at a large tech firm. TopHat question 5 was not asked.
Announcements (as given in class)
Three class sessions were left, counting November 12: class next week, then no class in Thanksgiving week. After Thanksgiving, Anna holds an exam review session, as for the other two exams, and then comes the final exam. The instructor believes it is Thursday, December 3; check the course schedule.
The final exam has 30 questions instead of 40: 10 on the first part of the semester, 10 on the second part, and 10 on the material since the second exam. It is open note, like the other exams. A practice exam and a study guide will be posted, as before.
Extra credit for Tuesday’s event. The instructor has the check-in list from Professor Garrison. If you attended but did not check in, send him proof, such as a picture or a video. The credit is applied to the final exam, about half a letter grade (an 85 becomes a 90; a 75 becomes an 80).
Do not memorize an example as a story. Use it to recover the economic principle.
| Example from class | Economic lesson |
|---|---|
| Restaurants hire cooks; doctors and health care | The demand for workers is derived from the demand for what they make. |
| The crew add 12, 18, 16, 12, 8, 4, 2, then −8 fish | With the tools fixed, the extra output from one more worker eventually falls: diminishing marginal returns. |
| Five crew at $80 a day; the owner keeps $260 | Hire each worker who adds at least the wage, valued at the price of the output. |
| San Angelo janitors leave for $20 fast-food jobs | The wage a firm must pay is set by what workers can earn elsewhere. |
| Wage $100: four crew; fish $20: six crew | A higher wage moves along the demand for workers; a higher output price shifts it. |
| A floor of $80 when the market wage is $40 | A wage above what the last worker adds costs that worker the job. |
| 18 + 18 = 36 fish, but the two together catch 30 | In a team, pay follows what one more worker adds, not what each person “really” produced. |
| Half the crew per boat, 50% more boats: 25% fewer jobs | Jobs = firms × workers per firm; cheaper output can bring more firms and more work. |
| The Model T and farriers; AI and coding | The amount of work worth doing is not fixed: technology changes which work is worth doing. |
| $1,578 a week with a degree, $966 without | Higher demand and a smaller, costlier supply of skilled workers raise their pay. |
These use new numbers. Questions 1 to 6 use one example: a bakery sells bread in a large market at $2 a loaf, so it cannot move the price. Other jobs in town pay $45 a day. With the same ovens, its bakers turn out 30 loaves a day with 1 baker, 70 with 2, 105 with 3, 130 with 4, 145 with 5, and 150 with 6. A baker who adds exactly the wage is counted in. Try each question before you open the answer.
Subtract each total from the next: 30, 40, 35, 25, 15, and 5 loaves. The second adds more than the first, so returns are still rising there. The third adds 35, fewer than the second’s 40, so diminishing marginal returns start with the third. The total keeps rising through the sixth baker.
At $2 a loaf the bakers add $60, $80, $70, $50, $30, and $10 a day. The fourth adds $50, more than the $45 wage; the fifth adds only $30. So the owner hires four and keeps 130 × $2 − 4 × $45 = $260 − $180 = $80 a day. Check: with three the owner keeps $75, and with five, $65.
The third adds $70, more than $65; the fourth adds only $50. So the owner hires three and keeps 105 × $2 − 3 × $65 = $210 − $195 = $15 a day (with two, only $10). The fourth baker moves to a job that pays $65, where that time is worth more than the $50 of bread it added. A higher wage moves the bakery along its demand for bakers.
Each baker’s loaves are now worth twice as much: $120, $160, $140, $100, $60, and $20. The fifth adds $60, more than $45; the sixth adds only $20. The owner hires five and keeps 145 × $4 − 5 × $45 = $580 − $225 = $355 a day. The bakers did not change; the demand for bread rose, so the demand for bakers shifted up. That is derived demand.
At $60 the third baker still adds $70, but the fourth adds only $50. The owner hires three and keeps 105 × $2 − 3 × $60 = $30 a day. The three who keep their jobs earn $15 a day more; the fourth, who would have earned $45, loses the job at the bakery.
Take either away and output falls from 70 to 30, so each “added” 40 loaves. But 40 + 40 = 80, more than the 70 they make together. Both comparisons are right, and both cannot be paid. In a team there is no single answer to who produced what; competition pushes pay toward what one more such worker adds.
Jobs = bakeries × bakers per bakery. With 80% more bakeries: 0.5 × 1.8 = 0.9, so baking jobs fall by 10%, not by half. With 2.5 times as many bakeries: 0.5 × 2.5 = 1.25, so baking jobs rise by 25%. Which happens depends on how much buyers want once bread is cheaper. That is a question for the evidence, not one you can answer by assuming the amount of work is fixed.
If your friend assumes there is a fixed list of tasks, that is the lump-of-labor mistake: wants are never used up, so work freed from one task can move to others. Ask: (1) which tasks AI does more cheaply, since those workers can lose out and moving on can be slow; (2) whom it makes more productive, since a lawyer who adds more is worth more; and (3) what buyers do with what they save, since cheaper contracts and the money saved can lead to new work.
More people go to college, so the supply of college graduates rises and their wage falls. Fewer people stay in the high-school market, so that supply falls and the high-school wage rises. The gap shrinks. It stops shrinking when, once again, no one gains by switching: for the worker on the fence, the extra pay over a career just covers what college now costs, which is mostly the pay given up while in school.
Bottom line
A firm hires a worker for what the worker adds, valued at the price of what is made. It keeps hiring while one more worker adds at least the wage, and the wage is what workers can earn in their next best job. A higher wage means fewer workers; a higher price for the output means more, because the demand for labor is derived from the demand for the product. In a team, pay follows what one more worker adds. Machines and AI replace some tasks, help with others, and make goods cheaper; the amount of work worth doing is not fixed. Pay differs across jobs because what workers add, what skills cost, and what jobs are like all differ.
← Back to the lecture: Factor Markets and Labor, Part A: What an Extra Worker Is Worth