Microeconomics

A hospital has one operating room free tonight and two patients who need it. A city has one budget and both a bridge and a school to repair. None of this is a story about greed or bad management. Each is the same story: something desirable exists in smaller quantity than the uses people would put it to, so somebody must decide. Economics begins exactly there, and its first move is to insist that the cost of the choice you make is the value of the choice you did not.

The fundamental problem

Scarcity is the basic economic condition: human wants are effectively unlimited, but the resources used to satisfy them - time, labor, land, capital - are limited. Because we cannot have everything, every choice to use a resource one way is also a choice not to use it another way. Economics is the study of how people, firms, and societies make these choices under scarcity. Scarcity is not the same as poverty or shortage: even a billionaire faces scarcity, because a day still has only 24 hours and attention, health, and time are finite. Scarcity is permanent and universal, which is exactly why the discipline that studies it applies to everyone.

It helps to separate two ideas that beginners often merge. A need is something required for survival or basic functioning; a want is anything desired beyond that. Economics generally treats both through the same lens, because both compete for the same scarce resources. When wants outrun resources, choices must be made, and the study of those choices - who gets what, and why - is the subject of this course.

Key idea: Scarcity is neither shortage nor poverty - it is the permanent gap between what people would use and what exists, which makes choice unavoidable for everyone.

Opportunity cost: the true cost of anything

The most important idea that follows from scarcity is opportunity cost: the value of the best alternative you give up when you make a choice. The true cost of anything is not just the money you pay - it is everything you sacrifice, including your time and forgone options. A "free" concert ticket is not free if attending means missing a shift that would have paid you; the wage you forgo is a real cost of going.

Opportunity cost has two components worth naming. Explicit costs are out-of-pocket money payments, such as tuition or the price of a ticket. Implicit costs are the value of resources you already own and use up, such as your time or the interest your savings could have earned. A complete opportunity-cost calculation counts both. This is why an entrepreneur who "pays herself nothing" is not actually running a costless business: the salary she could have earned elsewhere is a genuine implicit cost of staying self-employed.

Key idea: A complete cost count includes implicit costs - the time, space, and capital you already own and use up - not only the money that leaves your hands.

Worked example: the cost of a study hour

Suppose on a Saturday you can either work a shift that pays $60, or study for an exam, or hike with friends (which you value at $25 of enjoyment). If you choose to study, your opportunity cost is the single best alternative given up - the $60 shift, not the sum of both alternatives. Opportunity cost is the highest-valued forgone option, never the total of all of them.

The intuition: you could only have done one other thing with that time, so you only "lost" the best one. If the shift had instead paid $20, your best forgone option would have been the $25 hike, and your opportunity cost of studying would be $25. The number depends entirely on what your ranked alternatives are.

Now finish the decision rather than merely costing it. If the extra studying is worth $80 to you, the net gain is 80 - 60 = $20, so studying is the right call. Had the shift paid $95, the opportunity cost would be $95, the net would be 80 - 95 = -$15, and the same hours would be a mistake. The activity did not change; the alternative did.

Key idea: Opportunity cost is the single highest-valued forgone option, never the sum of all of them, and a choice is worth making only when its value exceeds that one number.

Worked example: accounting profit versus economic profit

Maya leaves a $58,000-a-year job to open a bakery, pulling $30,000 from savings that earned 5 percent to buy ovens. Her first year: revenue $210,000; ingredients $74,000; rent $36,000; wages $52,000; utilities $8,000.

Step 1 - add the explicit costs: 74,000 + 36,000 + 52,000 + 8,000 = $170,000. Step 2 - accounting profit = 210,000 - 170,000 = $40,000. An accountant stops here and reports a profitable year. Step 3 - count the implicit costs no accountant records: a forgone $58,000 salary plus forgone interest of 0.05 x 30,000 = $1,500, totalling $59,500. Step 4 - economic profit = 40,000 - 59,500 = -$19,500.

Read that carefully. The bakery is not draining the bank; it is earning $19,500 less than Maya's next-best use of the same year and savings. Economic profit measures performance against the road not taken, which is why it predicts entry and exit while the accountant's measure files taxes.

Key idea: Accounting profit subtracts explicit costs only; economic profit subtracts implicit costs as well, so a business can be accounting-profitable and economically unprofitable at the same time.

Thinking at the margin

Economists rarely ask "all or nothing." They ask about the margin: one more hour, one more unit, one more worker. A rational decision-maker takes an action when its marginal benefit exceeds its marginal cost, and stops when they are equal. This "marginal" lens reappears throughout the course - firms set output where marginal revenue equals marginal cost, consumers buy until marginal benefit equals price, and firms hire until the value of another worker equals the wage.

Marginal thinking resolves a puzzle that stumped early economists: the diamond-water paradox. Water is essential and diamonds are frivolous, yet diamonds command a far higher price. The resolution is marginal, not total. Because water is abundant, the marginal gallon is worth very little even though total water is priceless; because diamonds are scarce, the marginal diamond is worth a great deal. Prices reflect marginal value, not total value - a lesson that recurs whenever we compute costs and benefits one unit at a time.

Worked example: how many hours to study

Marginal analysis is a procedure, not a slogan. Each hour of study costs $15 of forgone leisure; each exam point is worth $2; and later hours add fewer points than earlier ones.

Hour	Extra points	Marginal benefit	Marginal cost	Do it?
1	14	$28	$15	Yes
2	11	$22	$15	Yes
3	8	$16	$15	Yes
4	6	$12	$15	No

Marginal benefit is extra points times $2 per point: 14 x 2 = $28, 11 x 2 = $22, 8 x 2 = $16, 6 x 2 = $12. Take every hour whose marginal benefit is at least its marginal cost and stop at the first one that is not. Hours 1 through 3 clear the $15 bar; hour 4 does not, so the answer is 3 hours, yielding 14 + 11 + 8 = 33 points for $45 of forgone leisure. Total benefit is 33 x $2 = $66, so the net gain is 66 - 45 = $21. Notice the fourth hour still adds points: someone reasoning in totals takes it, someone reasoning at the margin sees it costs more than it adds and stops.

Key idea: Optimal decisions are found by taking each additional unit whose marginal benefit is at least its marginal cost and stopping there, not by comparing totals.

Sunk costs

A sunk cost is a cost already paid and impossible to recover. Because it cannot change no matter what you do next, rational decisions ignore sunk costs and look only at future marginal costs and benefits. If you have paid $12 for a movie ticket and discover after 20 minutes that the film is dreadful, the $12 is gone either way; the only question is whether the next 90 minutes are worth more than what else you could do. Staying "to get your money's worth" is the sunk-cost fallacy: the $12 is equally lost whether you stay or leave, so it should not tip the decision.

Sunk is not the same as fixed, and the two get muddled constantly. A fixed cost does not change with output but may still be avoidable: a bakery's annual lease is fixed this year yet can be escaped by not renewing. A sunk cost can no longer be avoided by any decision still open to you - the non-refundable deposit already wired. Fixed costs belong in the long-run decision about whether to be in this business at all; sunk costs belong in no decision whatsoever.

Key idea: Sunk costs are unrecoverable and so are irrelevant to every future choice, while fixed costs merely do not vary with output and can still matter to longer-run decisions.

Positive vs normative

Economic reasoning separates two kinds of claims. A positive statement describes what is and can in principle be tested against data ("a higher minimum wage reduces teen employment"). A normative statement expresses what ought to be and reflects values ("the government should raise the minimum wage"). Good analysis keeps them distinct: facts inform the debate, but values decide the policy. Two economists can fully agree on a positive prediction - say, that a tax will cut consumption by 8 percent - and still disagree on the normative question of whether the tax is worth imposing, because they weigh the trade-offs differently.

Key idea: Positive claims are about what happens and can be checked against evidence; normative claims are about what should happen and rest on values, and confusing the two turns a settleable question into an argument.

Incentives and why it matters

Because people respond to costs and benefits, incentives - rewards and penalties that change behavior - are central. A tax on sugary drinks raises their cost and tends to reduce consumption; a subsidy lowers a cost and tends to increase an activity. Much of microeconomics is tracing how a change in incentives ripples through choices and markets, sometimes with surprising side effects.

When a policy ignores incentives it often backfires: rent control meant to help tenants can shrink the housing supply, and paying people per bug fixed can encourage writing buggy code to fix later. Mastering opportunity cost, marginal analysis, and incentives gives you a portable toolkit for reasoning about almost any decision, which is why this first lesson underpins everything that follows.

Key idea: People respond to changes in costs and benefits, so a policy is best predicted by asking what it makes cheaper or more expensive rather than by asking what it intends.

Where the standard model simplifies

Everything above assumes a decision-maker who knows the options, values them consistently, and ignores what cannot be recovered. Real people are not reliably like that, and behavioral economics exists because the gaps are systematic rather than random. The sunk-cost fallacy is the clearest case - it is called a fallacy precisely because people commit it constantly. People also weigh losses more heavily than equal gains and choose differently depending on how an identical option is described. None of this makes the model useless: it predicts well in aggregate, over repeated choices, and where the stakes reward care. Treat it as a benchmark for how a fully informed, consistent chooser would act, not a description of how everyone does.

Key idea: The rational-choice model is a benchmark rather than a complete description of behavior, and the documented, systematic ways people depart from it are the subject matter of behavioral economics.

Common wrong turns

• Adding up all the forgone options. Opportunity cost counts only the best one.
• Counting only cash. A choice that spends no money still uses time, space, and capital that had next-best uses.
• Letting past spending steer a current decision. If the right answer would be the same had the money never been spent, that money is sunk.
• Reading "zero economic profit" as failure. It means a firm is doing exactly as well as its best alternative - a normal return, not a collapse.
• Deciding in totals. "Studying is good" does not tell you how many hours; only the marginal comparison gives a quantity.

Recap

• Scarcity - unlimited wants meeting limited resources - is permanent and universal, and it forces every choice economics studies.
• Opportunity cost is the single best forgone alternative, and it includes implicit costs such as your own time and capital.
• Maya's $40,000 accounting profit became a $19,500 economic loss once her forgone salary and interest were subtracted.
• Marginal analysis picks the quantity where marginal benefit last exceeds marginal cost: 3 study hours, a $21 net gain.
• Sunk costs cannot be recovered and belong in no decision; fixed costs merely do not vary with output.
• Positive statements describe and can be tested; normative statements prescribe and rest on values.
• People respond to incentives, and behavioral economics documents where real choosers depart from the benchmark.

Sources

1. Greenlaw, S. A., Shapiro, D., & MacDonald, D. (2022). What is economics, and why is it important? In Principles of Microeconomics 3e. OpenStax, Rice University. openstax.org
2. Greenlaw, S. A., Shapiro, D., & MacDonald, D. (2022). How individuals make choices based on their budget constraint. In Principles of Microeconomics 3e. OpenStax, Rice University. openstax.org
3. Greenlaw, S. A., Shapiro, D., & MacDonald, D. (2022). Explicit and implicit costs, and accounting and economic profit. In Principles of Microeconomics 3e. OpenStax, Rice University. openstax.org
4. Henderson, D. R. (n.d.). Opportunity cost. In The Concise Encyclopedia of Economics. Liberty Fund. econlib.org
5. Rhoads, S. E. (n.d.). Marginalism. In The Concise Encyclopedia of Economics. Liberty Fund. econlib.org
6. Smith, A. (1776). An inquiry into the nature and causes of the wealth of nations. Project Gutenberg. gutenberg.org
7. CORE Econ Team. (2023). Scarcity, work, and choice. In The Economy 2.0: Microeconomics. CORE Econ. core-econ.org

During the Second World War the United States argued in newspapers and cabinet rooms about "guns or butter" - how much of a finite industrial base to point at armaments and how much to leave for civilian life. The phrase stuck because it names the entire problem in three words. There is a frontier of what an economy can make, every point on it costs something, and no amount of wanting moves it. This lesson draws that frontier, reads its slope as an opportunity cost, and then shows the one trick that lets people consume beyond it.

Modeling scarcity: the PPF

The production possibilities frontier (PPF) shows the maximum combinations of two goods an economy can produce when all resources are fully and efficiently employed. It is the simplest model in economics, yet it captures scarcity, choice, opportunity cost, efficiency, and growth all at once. Points on the curve are efficient; points inside are inefficient (resources idle or misused, as in a recession); points outside are currently unattainable given today's resources and technology.

The PPF's slope is opportunity cost. Moving along the curve to make more of one good means making less of the other, and the amount sacrificed is precisely the opportunity cost of the good gained. When resources are not equally suited to both goods, the curve bows outward, giving increasing opportunity cost: each additional unit of a good costs ever more of the other, because we must draw in resources less and less suited to producing it.

Reading a numerical PPF

Suppose a small economy can devote its resources to guns or butter, with these attainable combinations: (0 guns, 15 butter), (1, 14), (2, 12), (3, 9), (4, 5), (5, 0). Lay them out and compute the cost of each successive gun as the butter given up:

Guns	Butter	Butter given up for this gun
0	15	-
1	14	15 - 14 = 1
2	12	14 - 12 = 2
3	9	12 - 9 = 3
4	5	9 - 5 = 4
5	0	5 - 0 = 5

The rising cost per gun - 1, then 2, then 3, then 4, then 5 - is increasing opportunity cost in numbers, and it is exactly what makes the plotted curve bow outward rather than fall in a straight line. Read the table the other way and the same logic runs in reverse: going from 5 guns to 4 buys back 5 butter, while going from 1 gun to 0 buys back only 1. The first resources you move are the ones least suited to gun-making, and every further shift drags in workers and machines that were better at butter.

If instead every gun cost the same amount of butter - say 3 each, giving (0, 15), (1, 12), (2, 9), (3, 6), (4, 3), (5, 0) - the PPF would be a straight line with constant opportunity cost. That is the special case where resources are perfectly adaptable between uses. Real economies almost never look like that, which is why the bowed curve is the standard picture.

Key idea: The slope of the PPF is an opportunity cost, and it bows outward because resources are specialized, so each extra unit of a good costs more of the other than the last one did.

Efficiency, growth, and cost

An economy at a point like A is productively efficient: it cannot make more of one good without making less of the other. A point like B wastes resources - the economy could have more of both goods without any sacrifice, so B can never be a good outcome. Over time, better technology or more resources shift the whole PPF outward - economic growth - letting the economy produce more of both goods.

Crucially, an economy that gives up some current consumption to build capital goods (machines, factories, education) tends to grow its PPF faster, because capital raises future productive capacity. This is the trade-off between consuming now and investing for later, drawn as a choice of where to produce on today's frontier.

Be careful with a distinction that mirrors the demand-curve rule you will meet in the next module. Moving from one point on a given frontier to another is a change in what the economy chooses to make, and it always costs something. Shifting the whole frontier outward is a change in what the economy can make, and it costs nothing at the moment it happens. Growth is the second kind of change; reallocation is the first. Confusing them produces the common political claim that a country can have more of everything simply by rearranging priorities, which is true only if the economy was inside its frontier to begin with.

Key idea: Points on the frontier are productively efficient and points inside waste resources, while an outward shift of the whole frontier - growth - is a different event from moving along it.

Comparative advantage and the gains from trade

Why do individuals and countries specialize and trade? The answer is comparative advantage: the ability to produce a good at a lower opportunity cost than someone else. This differs from absolute advantage, which is simply producing more with the same resources. The single most counterintuitive result in introductory economics is that trade can benefit both parties even when one is better at producing everything.

Consider two people making bread and code in one day:

Worker	Loaves (if only bread)	Features (if only code)
Ana	20	10
Ben	6	6

Ana has the absolute advantage in both. But compare opportunity costs. For Ana, 1 feature costs 20/10 = 2 loaves. For Ben, 1 feature costs 6/6 = 1 loaf. Ben gives up less bread per feature, so Ben has the comparative advantage in coding. Conversely, 1 loaf costs Ana 0.5 features but costs Ben 1 feature, so Ana has the comparative advantage in bread. If each shifts effort toward the good they sacrifice least to make, the pair can end up with more of both goods than they had in isolation. The lesson: gains from trade come from differences in opportunity cost, not from who is "better" overall.

Key idea: Absolute advantage is about who produces more; comparative advantage is about who gives up less, and only the second determines who should specialize in what.

Worked example: does specialization really produce more?

The claim that specialization raises output deserves arithmetic rather than assertion. Suppose each works one day and, in isolation, splits the day evenly between the two tasks.

Situation	Ana	Ben	Combined
No trade (each splits the day 50/50)	10 loaves, 5 features	3 loaves, 3 features	13 loaves, 8 features
Specialize by comparative advantage	70% bread, 30% code: 14 loaves, 3 features	All code: 0 loaves, 6 features	14 loaves, 9 features

Check each cell. Half a day gives Ana 0.5 x 20 = 10 loaves and 0.5 x 10 = 5 features; Ben gets 0.5 x 6 = 3 of each. Under specialization Ana's 70 percent on bread gives 0.7 x 20 = 14 loaves and her 30 percent on code gives 0.3 x 10 = 3 features, while Ben's full day gives 6 features. Combined output rises from 13 to 14 loaves and from 8 to 9 features - more of both, from the same two days of work and no new technology.

One caution the textbooks often skip. Complete specialization is not always the arithmetic winner in a two-person example. If Ana went entirely into bread, the pair would have 20 loaves but only 6 features - more bread than before, but fewer features than the 8 they had in isolation. The reliable statement is that reallocating toward comparative advantage expands the set of combinations available to the pair; how far to push it depends on how much of each good they want.

Worked example: where a mutually beneficial trade price lies

Continue the example. A trade benefits both only if the price sits between the two opportunity costs. Ben will code a feature and trade it away only if he gets more than 1 loaf for it (his cost). Ana will buy a feature only if she pays less than 2 loaves (her cost of making it herself).

So any exchange rate between 1 and 2 loaves per feature makes both better off. Say they settle on 1.5 loaves per feature: Ben gains half a loaf over making bread himself, and Ana saves half a loaf over coding herself. Both come out ahead, and total bread-plus-features produced by the pair rises because each now spends the day doing what they sacrifice the least to do.

Now settle it with numbers. Start from the specialized outputs above - Ana with 14 loaves and 3 features, Ben with 6 features - and let Ben sell Ana 4 features at 1.5 loaves each, so 6 loaves change hands. Ana ends with 14 - 6 = 8 loaves and 7 features; Ben ends with 6 loaves and 2 features. Did each gain? Value the bundles at each person's own opportunity cost. For Ana a feature is worth 2 loaves, so her old bundle (10, 5) was worth 10 + 10 = 20 loaf-equivalents and her new one (8, 7) is worth 8 + 14 = 22. For Ben a feature is worth 1 loaf, so his old bundle (3, 3) was worth 6 and his new one (6, 2) is worth 8. Each is ahead by 2 loaf-equivalents - a genuine gain, not a transfer.

Key idea: Any exchange rate strictly between the two producers' opportunity costs makes both better off, and where inside that range the price lands determines how the gains are split.

What the model leaves out

This is a deliberately bare model, and honesty about its assumptions is part of using it well. It assumes each producer's opportunity cost is constant, which is why complete specialization looks so clean; with rising costs, specialization is usually partial. It assumes resources move costlessly between uses, when in reality a baker retrained as a programmer bears months of lost income. It ignores transport costs, tariffs, and the time it takes to find a trading partner. And it says nothing about who bears the adjustment: the aggregate gain from trade is compatible with real, concentrated losses for workers in a shrinking sector, which is exactly why trade is politically contested even where the economics is uncontroversial.

Key idea: Comparative advantage shows that specialization can enlarge the total, but it does not promise that everyone gains automatically or that adjustment is costless.

Why it matters

Comparative advantage is the intellectual foundation of the case for trade, from two roommates dividing chores to nations negotiating treaties. It explains why a surgeon hires a gardener even if the surgeon could mow faster: the surgeon's opportunity cost of mowing (forgone surgery) is enormous, so specializing and trading raises total output. The same logic warns against the intuitive but mistaken idea that a country should make everything it is "good at." What matters is not absolute skill but relative cost, and recognizing that difference is one of the most practically useful things this course will teach you.

Common wrong turns

• Treating a point inside the frontier as a valid trade-off. It is waste, not choice: more of both goods is available at no cost.
• Assuming the better producer should make everything. Absolute advantage is irrelevant to who should specialize; only relative cost matters.
• Computing opportunity cost with the wrong ratio. The cost of one feature to Ana is loaves per feature, 20/10 = 2 - not 10/20.
• Confusing a move along the frontier with a shift of it. Reallocating costs something; growth does not.
• Reading a straight-line PPF as the normal case. Constant opportunity cost is the special case of perfectly adaptable resources.

Recap

• The PPF shows maximum output combinations; points on it are efficient, inside it wasteful, outside it currently unattainable.
• Its slope is opportunity cost, and it bows outward because specialized resources make each extra gun cost more butter - 1, 2, 3, 4, then 5 in the worked table.
• Outward shifts of the whole curve are growth, driven by more resources, better technology, or investment in capital goods.
• Comparative advantage - lower opportunity cost - determines who should specialize, and it can differ from absolute advantage.
• Reallocating toward comparative advantage raised the pair's output from 13 loaves and 8 features to 14 and 9, using the same two days of work.
• Any exchange rate between the two opportunity costs (here 1 to 2 loaves per feature) leaves both parties better off; at 1.5, each gained 2 loaf-equivalents.
• The model assumes constant costs and costless reallocation, so real-world gains come with adjustment costs borne unevenly.

Sources

1. Greenlaw, S. A., Shapiro, D., & MacDonald, D. (2022). The production possibilities frontier and social choices. In Principles of Microeconomics 3e. OpenStax, Rice University. openstax.org
2. Greenlaw, S. A., Shapiro, D., & MacDonald, D. (2022). Absolute and comparative advantage. In Principles of Microeconomics 3e. OpenStax, Rice University. openstax.org
3. Greenlaw, S. A., Shapiro, D., & MacDonald, D. (2022). What happens when a country has an absolute advantage in all goods. In Principles of Microeconomics 3e. OpenStax, Rice University. openstax.org
4. Greenlaw, S. A., Shapiro, D., & MacDonald, D. (2022). The benefits of reducing barriers to international trade. In Principles of Microeconomics 3e. OpenStax, Rice University. openstax.org
5. Boudreaux, D. J. (n.d.). Comparative advantage. In The Concise Encyclopedia of Economics. Liberty Fund. econlib.org
6. Ricardo, D. (1817). On the principles of political economy, and taxation. Project Gutenberg. gutenberg.org
7. CORE Econ Team. (2023). Technology and incentives. In The Economy 2.0: Microeconomics. CORE Econ. core-econ.org

Raise the price of coffee by a dollar and some people switch to tea, some brew at home, and some buy the same cup and grumble. Nobody polls them, no committee coordinates them, and yet the total quantity purchased falls in a way regular enough to draw as a line. That line is the demand curve, and learning to read it - especially learning when it moves and when you merely slide along it - is the single most transferable skill in this course.

The law of demand

Demand is the relationship between the price of a good and the quantity buyers are willing and able to purchase, holding other factors constant. The phrase "willing and able" matters: a want backed by no purchasing power is not demand in the economic sense. The law of demand states that, other things equal, as price rises quantity demanded falls, and as price falls quantity demanded rises.

The demand curve therefore slopes downward. By long-standing convention economists put price on the vertical axis and quantity on the horizontal, even though we usually think of quantity as responding to price, so read the curve as "at each price, this is the quantity buyers will take."

Two forces explain the negative slope. The substitution effect: when a good gets more expensive relative to alternatives, buyers switch toward the now-cheaper substitutes. The income effect: a higher price reduces the real purchasing power of a buyer's income, so they can afford less of things in general.

For a normal good the two effects push the same way - higher price, lower quantity demanded - and reinforce the downward slope. For an inferior good they pull against each other: a higher price makes the buyer effectively poorer, and being poorer makes an inferior good more attractive, so the income effect partly offsets the substitution effect. The substitution effect almost always wins, which is why demand curves slope down in practice. The theoretical exception is a Giffen good - an inferior good absorbing so much of a poor household's budget that the income effect dominates and quantity demanded rises with price. Convincing real-world cases are rare and contested, so treat it as a boundary case that shows the law of demand is a strong empirical regularity rather than a logical necessity.

Key idea: Demand curves slope downward because the substitution effect and (for normal goods) the income effect both reduce quantity demanded when price rises - a robust empirical regularity, not a theorem.

Reading a demand schedule

Before drawing curves, work with the numbers behind one. A demand schedule is just a table of price-quantity pairs. Here is one buyer's weekly demand for coffee:

Price per cup	Cups per week
$6	2
$5	4
$4	6
$3	8
$2	10

Every dollar the price falls, this buyer takes 2 more cups, so the schedule obeys the law of demand and traces a straight line. You can write it as an equation. Quantity rises 2 for each $1 fall, so Qd = a - 2P for some constant a; substituting P = 4 and Qd = 6 gives 6 = a - 8, so a = 14 and Qd = 14 - 2P. Check it at another row: at P = 2, Qd = 14 - 4 = 10, which matches the table. Being able to move between a table, a graph, and an equation is exactly the fluency the rest of the module assumes.

Key idea: A demand schedule, a demand curve, and a demand equation are three views of one relationship, and you should be able to convert among them.

Market demand as the sum of individuals

An individual's demand curve shows what one buyer will purchase at each price. The market demand curve is the horizontal sum of all individual demand curves: at each price, add up the quantities every buyer wants. This is why the number of buyers is a demand shifter - more buyers means more quantity demanded at every price, shifting market demand right. Horizontal summation also explains why market demand is typically smoother and flatter than any one jagged individual demand: aggregating many buyers averages out their individual quirks.

Do it once with numbers. Suppose a tiny market has three buyers with demands Qd = 14 - 2P (Ana), Qd = 10 - P (Ben), and Qd = 6 - P (Cara). At a price of $4 they want 6, 6, and 2 cups, so market quantity demanded is 6 + 6 + 2 = 14. At $3 they want 8, 7, and 3, giving 18. Summing the equations gives the market demand curve directly: (14 - 2P) + (10 - P) + (6 - P) = Qd = 30 - 4P. Test it at $4: 30 - 16 = 14, matching the hand count.

One subtlety worth seeing early. That formula holds only while every buyer is still in the market. At $7, Cara's equation would give -1 cups, which really means zero, and Ana's gives exactly zero, so actual market demand is Ben's 3 cups - not the 30 - 28 = 2 the formula predicts. Market demand curves therefore have kinks where buyers drop out, which is one reason real demand curves are flatter and smoother at high quantities than any single buyer's.

Key idea: Market demand is the horizontal sum of individual demands - add quantities at each price, never prices at each quantity - and it kinks where buyers leave the market.

Movement along vs shift of the curve

This distinction is the most common source of confusion in the whole course, so be precise:

• A change in quantity demanded is a movement along a fixed demand curve, caused only by a change in the good's own price.
• A change in demand is a shift of the entire curve, caused by something other than the good's own price. A rightward shift is an increase in demand; a leftward shift is a decrease.

A reliable test: ask what changed. If the good's own price changed, you move along the curve. If anything else changed - income, tastes, a related good's price, expectations, the number of buyers - the whole curve shifts. Getting this right is essential because in the next lesson we combine demand with supply, and mislabeling a shift as a movement (or vice versa) will give the wrong prediction for price and quantity.

Key idea: Only the good's own price moves you along a demand curve; every other cause shifts the whole curve, and the test is simply to ask what changed.

What shifts demand

The determinants of demand are often remembered as TIPES:

1. Tastes and preferences - a good going into fashion raises demand; a health scare lowers it.
2. Income - for a normal good, higher income raises demand; for an inferior good, higher income lowers demand.
3. Prices of related goods - a rise in the price of a substitute raises demand for this good; a rise in the price of a complement lowers it.
4. Expectations - expecting higher future prices, or higher future income, raises demand today.
5. Number of buyers - a larger market raises demand.

Worked example: substitute vs complement

Coffee and tea are substitutes. If the price of tea jumps, some tea drinkers switch to coffee, so the demand for coffee increases (its curve shifts right) even though coffee's own price has not changed. Coffee and cream are complements. If the price of coffee jumps, people buy less coffee and therefore less cream, so the demand for cream decreases.

Notice that a change in a related good's price shifts this good's demand, while a change in the good's own price only moves us along its curve. Trace it slowly: the tea price change never touches coffee's own price, so it cannot be a movement along coffee's curve - it must be a shift.

Worked example: normal vs inferior with a number

Suppose a city gives every resident a raise. Restaurant-meal demand rises (a normal good), while demand for instant ramen falls (an inferior good) because people trade up to better food. Both responses are shifts caused by income, not by the goods' own prices. If instead the price of restaurant meals alone fell, that would be a movement along the restaurant-meal demand curve, an increase in quantity demanded, with no shift at all. Keeping the cause straight - own price versus everything else - is the whole game.

Key idea: Income sorts goods into normal and inferior, and related-good prices sort them into substitutes and complements - four labels that between them explain most demand shifts you will meet.

A drill: shift, or movement?

Run each case through the test. (1) Coffee's own price falls from $5 to $4. Own price changed, so this is a movement along the curve: quantity demanded rises from 4 to 6 cups, and demand itself is unchanged. (2) A study links coffee to better health. Tastes changed, so demand shifts right - at every price, more cups. (3) The price of tea doubles. A substitute's price changed, so coffee demand shifts right. (4) The price of cream doubles. A complement's price changed, so coffee demand shifts left. (5) Buyers expect coffee to be much more expensive next month. Expectations changed, so demand today shifts right as people stock up. (6) A recession cuts incomes. Coffee is a normal good for most, so demand shifts left - though demand for instant coffee, an inferior good for many, may shift right.

Case 6 is the one that separates careful students from careless ones: the same shock moves two related demand curves in opposite directions, and only the normal-versus-inferior classification tells you which is which.

Key idea: Every demand question reduces to naming the cause, classifying it as own-price or not, and then - if not - deciding which direction the whole curve moves.

Common wrong turns

• Calling a price change a "change in demand." A price change moves you along the curve; demand itself has not changed.
• Summing demand curves vertically. Market demand adds quantities at a given price, not prices at a given quantity.
• Assuming higher income always raises demand. True for normal goods, false for inferior ones - the sign flips.
• Mixing up substitutes and complements. A pricier substitute pushes buyers toward this good; a pricier complement pulls them away from it.
• Treating "willing" as enough. Demand requires willing and able - desire backed by purchasing power.
• Reading the axes backwards. Price is on the vertical axis by convention even though quantity is what responds.

Recap

• Demand is the willing-and-able relationship between price and quantity, holding everything else constant.
• The curve slopes down because of the substitution effect and, for normal goods, a reinforcing income effect; Giffen goods are the contested boundary case.
• A schedule, a graph, and an equation are three views of one relationship - the coffee table above is exactly Qd = 14 - 2P.
• Market demand is the horizontal sum of individual demands: 14 - 2P, 10 - P, and 6 - P summed to Qd = 30 - 4P, with a kink once buyers drop out.
• Own price changes move you along the curve; anything else shifts it - a distinction that decides every prediction in the next lesson.
• Tastes, income, related-good prices, expectations, and the number of buyers are the standard shifters.
• The same income shock shifts normal-good demand right and inferior-good demand left, so the classification matters.

Sources

1. Greenlaw, S. A., Shapiro, D., & MacDonald, D. (2022). Demand, supply, and equilibrium in markets for goods and services. In Principles of Microeconomics 3e. OpenStax, Rice University. openstax.org
2. Greenlaw, S. A., Shapiro, D., & MacDonald, D. (2022). Shifts in demand and supply for goods and services. In Principles of Microeconomics 3e. OpenStax, Rice University. openstax.org
3. Greenlaw, S. A., Shapiro, D., & MacDonald, D. (2022). How changes in income and prices affect consumption choices. In Principles of Microeconomics 3e. OpenStax, Rice University. openstax.org
4. Henderson, D. R. (n.d.). Demand. In The Concise Encyclopedia of Economics. Liberty Fund. econlib.org
5. Marshall, A. (1920). Principles of economics (8th ed.). Library of Economics and Liberty. econlib.org
6. Greenlaw, S. A., Shapiro, D., & MacDonald, D. (2022). Elasticity in areas other than price. In Principles of Microeconomics 3e. OpenStax, Rice University. openstax.org
7. CORE Econ Team. (2017). Supply and demand: Price-taking and competitive markets. In The Economy 1.0. CORE Econ. core-econ.org

No one sets the price of wheat. Not the farmers, not the mills, not the government of any single country. Yet on any given morning there is a price, and at that price almost exactly as much wheat is offered as is wanted. The mechanism that produces this coordination without a coordinator is the meeting of supply and demand, and once you can solve for it algebraically and shift it deliberately, you have the tool that carries the rest of this course.

Supply

Supply is the relationship between price and the quantity producers are willing to sell, holding other things constant. The law of supply states that, other things equal, a higher price raises quantity supplied, so the supply curve slopes upward - higher prices cover higher marginal costs and reward more production. The deep reason connects to Module 5: because of diminishing returns, producing more usually costs more per unit at the margin, so sellers need a higher price to justify expanding output. As with demand, a change in the good's own price moves us along the supply curve, while other factors shift it.

The determinants that shift supply are often remembered as input prices, technology, taxes and subsidies, expectations, and the number of sellers. A fall in input prices or an improvement in technology lowers the cost of production and shifts supply right; a new per-unit tax raises cost and shifts supply left. Just like demand, the one thing that does not shift the curve is the good's own price - that produces a movement along it.

Two shifters deserve care because students routinely get their direction backwards. A subsidy per unit lowers the seller's effective cost and shifts supply right, so the curve moves toward more output at every price. An expectation of higher future prices shifts current supply left, because sellers hold inventory back for the better price later - the opposite of what expectations do to demand. Whenever a shifter confuses you, ask what it does to the cost or attractiveness of selling one more unit today.

Key idea: Supply slopes upward because marginal cost rises with output, and everything except the good's own price - input prices, technology, taxes, subsidies, expectations, the number of sellers - shifts the whole curve.

Reading a supply schedule

Numbers make the abstraction usable. Suppose one seller's weekly supply schedule is:

Price	Quantity supplied
$5	0
$10	20
$15	40
$20	60

Quantity rises 20 units for every $5 of price, a slope of 4 units per dollar, so the schedule is Qs = 4P + b. Substituting P = 10 and Qs = 20 gives 20 = 40 + b, so b = -20 and Qs = -20 + 4P. Check at P = 20: -20 + 80 = 60, matching the table. The negative intercept is not a mistake and is not "negative supply" - it simply means this seller offers nothing until the price clears $5, which is where the curve crosses the axis.

Key idea: A supply schedule, curve, and equation are one relationship in three forms, and a negative intercept just marks the price below which sellers do not participate.

Market equilibrium

Equilibrium is the price at which quantity demanded equals quantity supplied. At that price the market clears - there is no shortage or surplus, and no pressure for price to change. If price is above equilibrium, quantity supplied exceeds quantity demanded: a surplus forms and pushes price down as sellers compete to unload unsold goods. If price is below equilibrium, quantity demanded exceeds quantity supplied: a shortage forms and pushes price up as buyers compete for scarce goods. These self-correcting pressures drive the market back to equilibrium, which is why economists call it a stable resting point.

Worked example: solving for equilibrium

Let demand be Qd = 100 - 2P and supply be Qs = -20 + 4P. Set quantity demanded equal to quantity supplied:

100 - 2P = -20 + 4P
120 = 6P
P* = 20

Substitute back: Qd = 100 - 2(20) = 60. So equilibrium price is $20 and equilibrium quantity is 60 units. (Check with supply: Qs = -20 + 4(20) = 60. It matches.) Always verify by plugging the price into both equations; if the two quantities agree, your equilibrium is correct.

Worked example: a surplus at the wrong price

Using the same curves, suppose a seller posts a price of $25. Quantity supplied is Qs = -20 + 4(25) = 80; quantity demanded is Qd = 100 - 2(25) = 50. Supply exceeds demand by 30 units - a surplus. Unsold inventory pushes sellers to cut price, and they keep cutting until the gap closes at P = 20. Now try P = 15: Qs = 40, Qd = 70, a shortage of 30 units that bids the price up to 20. Both experiments land on the same equilibrium, illustrating why $20 is the market's resting point.

Tabulating several prices at once shows the pressure changing sign as you cross equilibrium:

Price	Qd = 100 - 2P	Qs = -20 + 4P	Gap	Pressure on price
$10	80	20	Shortage of 60	Upward
$15	70	40	Shortage of 30	Upward
$20	60	60	None	At rest
$25	50	80	Surplus of 30	Downward
$30	40	100	Surplus of 60	Downward

Notice the gap shrinks steadily as price approaches $20 from either side and reverses sign as it passes through. That sign change is what makes the equilibrium stable: any deviation generates a force pushing back toward it. Note too that the traded quantity away from equilibrium is always the smaller of the two numbers, since no exchange happens without both a willing buyer and a willing seller - at $25 only 50 units actually trade even though 80 are offered.

Key idea: Equilibrium is where quantity demanded equals quantity supplied, and it is stable because surpluses push price down and shortages push it up.

Shocks and comparative statics

To predict how equilibrium moves, shift the right curve and read the new crossing. This method - comparing equilibria before and after a change - is called comparative statics. Useful rules:

• Demand increases (rightward): price up, quantity up.
• Demand decreases: price down, quantity down.
• Supply increases (rightward, e.g. better technology): price down, quantity up.
• Supply decreases (e.g. a bad harvest): price up, quantity down.

When both curves shift, the change in one of price or quantity is determinate and the other is ambiguous without knowing the relative sizes of the shifts. For example, if demand and supply both increase, quantity clearly rises but price could go up, down, or stay the same depending on which shift is larger. A quick way to handle two-shift problems: pin down the variable both shifts push the same way (here, quantity), and label the other ambiguous.

Worked example: when both curves move

Prove the ambiguity to yourself rather than memorising it. Start again from Qd = 100 - 2P and Qs = -20 + 4P, with equilibrium at P = 20, Q = 60. Now let demand rise by 30 units at every price, so Qd = 130 - 2P, and let supply also rise, by 24 units, so Qs = 4 + 4P. Setting them equal: 130 - 2P = 4 + 4P, so 126 = 6P and P = 21, giving Q = 130 - 42 = 88 (check on the supply side: 4 + 84 = 88). Price rose from $20 to $21 and quantity rose from 60 to 88.

Keep the same demand shift but make the supply shift bigger - say supply rises by 60, so Qs = 40 + 4P. Now 130 - 2P = 40 + 4P, so 90 = 6P and P = 15, giving Q = 130 - 30 = 100 (check: 40 + 60 = 100). This time price fell, from $20 to $15, while quantity again rose.

Two shifts, same directions, opposite price outcomes - and quantity up in both. That is precisely what "quantity is determinate, price is ambiguous" means, and it is why a careful analyst refuses to predict the price without knowing the relative size of the shifts.

Key idea: When both curves shift the same way, the variable they both push moves predictably and the other depends entirely on which shift is larger.

What the model assumes

This machinery works because of assumptions worth stating out loud. It treats every buyer and seller as a price taker too small to move the price alone, assumes the good is standardised enough that "the price" is meaningful, and assumes participants know what is on offer. It also assumes price adjusts fast enough that the market reaches its resting point - which is why the model describes commodity exchanges better than it describes housing, where leases and search frictions slow everything down. Where these assumptions fail, later modules supply the repairs: market power in Module 6, and imperfect information and externalities in Module 8. Treat supply and demand as a superb first approximation whose failures are informative rather than embarrassing.

Key idea: The supply-and-demand model assumes many small price-taking participants, a standardised good, decent information, and prices free to adjust - and its predictions weaken exactly where those assumptions do.

Why it matters

The supply-and-demand model is the workhorse of price analysis. It explains why a frost in Brazil raises coffee prices worldwide (supply falls), why a viral product sells out at launch (a shortage at the posted price), and why ride-hailing apps raise fares during a downpour (demand surges against fixed short-run supply). Master the mechanics of solving for and shifting equilibrium here, because every later topic - elasticity, taxes, price controls, and market structure - is built on top of it.

Common wrong turns

• Shifting the wrong curve. A change in production costs moves supply; a change in buyer income moves demand. Ask whose behaviour changed.
• Calling a price change a shift. A higher price moves you along supply - a change in quantity supplied - it does not shift supply.
• Predicting both price and quantity when both curves move. One of the two is always ambiguous without the relative shift sizes.
• Reading a negative intercept as negative supply. Qs = -20 + 4P simply means nothing is offered below $5.
• Assuming the larger quantity trades at a disequilibrium price. Trade equals the smaller of quantity demanded and quantity supplied.
• Getting expectations backwards. Expected higher future prices shift demand right today but supply left today.

Recap

• Supply slopes upward because rising marginal cost means sellers need a higher price to justify more output.
• Input prices, technology, taxes and subsidies, expectations, and the number of sellers shift supply; the good's own price does not.
• The schedule 0, 20, 40, 60 units at $5, $10, $15, $20 is exactly Qs = -20 + 4P.
• Equilibrium solves Qd = Qs: with Qd = 100 - 2P and Qs = -20 + 4P, P* = $20 and Q* = 60, verified in both equations.
• Above equilibrium a surplus pushes price down; below it a shortage pushes price up, and only the smaller quantity actually trades.
• When demand rose 30 and supply rose 24, price went to $21; when supply instead rose 60, price fell to $15 - quantity rose either way.
• The model assumes price-taking participants, a standardised good, and flexible prices, and it is weakest exactly where those fail.

Sources

1. Greenlaw, S. A., Shapiro, D., & MacDonald, D. (2022). Demand, supply, and equilibrium in markets for goods and services. In Principles of Microeconomics 3e. OpenStax, Rice University. openstax.org
2. Greenlaw, S. A., Shapiro, D., & MacDonald, D. (2022). Changes in equilibrium price and quantity: The four-step process. In Principles of Microeconomics 3e. OpenStax, Rice University. openstax.org
3. Greenlaw, S. A., Shapiro, D., & MacDonald, D. (2022). The market system as an efficient mechanism for information. In Principles of Microeconomics 3e. OpenStax, Rice University. openstax.org
4. Ehrbar, A. (n.d.). Supply. In The Concise Encyclopedia of Economics. Liberty Fund. econlib.org
5. Hayek, F. A. (1945). The use of knowledge in society. The American Economic Review, 35(4). Library of Economics and Liberty. econlib.org
6. Greenlaw, S. A., Shapiro, D., & MacDonald, D. (2022). Demand, supply, and efficiency. In Principles of Microeconomics 3e. OpenStax, Rice University. openstax.org
7. CORE Econ Team. (2023). Supply and demand. In The Economy 2.0: Microeconomics. CORE Econ. core-econ.org

Every price control begins as an act of sympathy. Rents are too high for nurses, so cap them. Wages are too low to live on, so raise the floor. Milk prices are ruining farmers, so guarantee a minimum. The intention is never the problem. The problem is that a price is not just a number attached to a good - it is simultaneously a signal to producers, a rationing device among buyers, and a reward for supplying more. Fix the number and all three jobs stop being done, which is why the effects of price controls are so often the opposite of the effects intended.

When government sets the price

Sometimes policymakers set a legal limit on price instead of letting the market clear. A price ceiling is a legal maximum price; a price floor is a legal minimum. Whether a control matters depends entirely on where it sits relative to the equilibrium price. A control on the "wrong" side of equilibrium is non-binding and does nothing, because the free-market price is already legal. A control on the "effective" side is binding and changes the outcome. Keeping straight which side binds is the crux of every price-control problem.

The mnemonic that never fails: a ceiling is something you bump your head on, so it has to be below you to matter; a floor is something you stand on, so it has to be above the market price to matter. Draw the equilibrium first, then draw the control as a horizontal line, and the answer becomes visual before it becomes algebraic.

Key idea: A ceiling binds only below equilibrium and a floor only above it; a control on the other side is legal but irrelevant, because the market price is already permitted.

Binding price ceilings and shortages

A price ceiling only "binds" if it is set below the equilibrium price. Rent control is the classic case. At the artificially low legal price, quantity demanded exceeds quantity supplied, producing a persistent shortage. Because price cannot ration the good, other mechanisms appear: long waiting lists, reduced quality (landlords defer maintenance), favoritism, and sometimes black markets where the good trades illegally above the ceiling. A ceiling set above equilibrium is non-binding and has no effect, because the market already clears at a legal price beneath it.

Worked example: a binding ceiling

Let demand be Qd = 100 - 2P and supply be Qs = -20 + 4P, so equilibrium is P = 20, Q = 60 (from the previous lesson). Now impose a rent-style ceiling at P = 14. Quantity demanded becomes Qd = 100 - 2(14) = 72; quantity supplied becomes Qs = -20 + 4(14) = 36. The result is a shortage of 72 - 36 = 36 units. Only 36 units actually trade (you cannot buy what is not supplied), even though 72 are wanted. Notice the ceiling did not just lower the price - it shrank the quantity actually exchanged from 60 to 36.

Now put a number on the waste. Rewrite each curve with price on the left. From Qd = 100 - 2P, the highest price a buyer will pay for the Qth unit is P = 50 - 0.5Q. From Qs = -20 + 4P, the lowest price a seller will accept for the Qth unit is P = 5 + 0.25Q. At the traded quantity of 36, a buyer values the unit at 50 - 18 = $32 while a seller would supply it for 5 + 9 = $14. That $18 gap is value nobody captures. At the free-market quantity of 60 the gap closes to zero (50 - 30 = 20 and 5 + 15 = 20, the equilibrium price).

The lost gains form a triangle with base 60 - 36 = 24 units and height $18, so the deadweight loss is 0.5 x 24 x 18 = $216. Verify it the long way if you like: total surplus before the ceiling is 0.5 x 60 x (50 - 20) = $900 of consumer surplus plus 0.5 x 60 x (20 - 5) = $450 of producer surplus, or $1,350. After the ceiling, consumer surplus is 0.5 x (36 + 18) x 36 = $972 and producer surplus is 0.5 x 36 x (14 - 5) = $162, totalling $1,134. The difference, 1,350 - 1,134, is exactly $216.

That consumer-surplus figure carries a hidden assumption worth naming: it credits the 36 available units to the 36 buyers who value them most. Real ceilings ration by queue, by luck, or by connections, so some units reach buyers who value them less. Actual consumer surplus is therefore usually lower than the calculation suggests, which means $216 understates the true loss.

Key idea: A binding ceiling cuts the quantity traded, and the value of the trades it blocks - a triangle between the demand and supply curves - is the deadweight loss, here $216.

Binding price floors and surpluses

A price floor binds only if set above equilibrium. The minimum wage is a labor-market floor: if the legal wage is above the market-clearing wage, the quantity of labor supplied exceeds the quantity demanded, and the surplus of labor is unemployment. Agricultural price supports work the same way, generating unsold surpluses that governments often buy up or store. A floor set below equilibrium is non-binding, because the market already clears above it.

Worked example: a binding wage floor

Let labor demand be Ld = 100 - 5W and labor supply be Ls = -20 + 5W, where W is the hourly wage. Setting them equal: 100 - 5W = -20 + 5W, so 120 = 10W and the market wage is W = $12 with employment of 100 - 60 = 40 workers (check: -20 + 60 = 40).

Impose a minimum wage of $15. Employers now want Ld = 100 - 75 = 25 workers, while Ls = -20 + 75 = 55 people want jobs. The surplus of labor is 55 - 25 = 30 - that is the unemployment the floor creates in this model. Employment falls from 40 to 25, so 15 workers who had jobs at $12 no longer have them.

The distribution is mixed, not uniformly good or bad. The 25 workers who keep their jobs earn $3 more per hour, a transfer of 25 x $3 = $75 per hour from employers to workers. The 15 displaced workers lose. For the deadweight loss, invert the curves: demand price is W = 20 - 0.2L and supply price is W = 4 + 0.2L. At L = 25 an employer values the worker at 20 - 5 = $15 while the worker would accept 4 + 5 = $9, a gap of $6, closing to zero at L = 40. So deadweight loss = 0.5 x (40 - 25) x 6 = $45 per hour of lost mutually beneficial employment.

State the model's limits honestly. This prediction assumes a competitive labor market with many small employers. Where a few employers dominate hiring - monopsony, covered in the labor module - firms already hold wages below the competitive level, and a moderate minimum wage can raise both wages and employment. Empirical work since Card and Krueger's 1993 study of New Jersey fast-food restaurants has found employment effects that are often small and sometimes statistically indistinguishable from zero, which is why the minimum wage remains actively debated rather than settled. Use the model to see the mechanism; use the evidence to size the effect.

Key idea: A binding wage floor raises pay for those still employed while reducing the number employed in the competitive model - but real labor markets are imperfectly competitive, so the size and even the sign of the employment effect is an empirical question.

Efficiency cost: deadweight loss

A binding control prevents some mutually beneficial trades from happening. The value of those lost trades - trades where a buyer was willing to pay more than a seller's cost, but which the control blocks - is the deadweight loss. It is pure lost gains from trade, benefiting no one.

Controls also redistribute: a binding ceiling helps buyers who can still buy at the low price but hurts sellers and the buyers shut out; a binding floor helps sellers who can still sell but hurts buyers and shut-out sellers. So a price control is not simply "good for buyers" or "good for sellers" - it creates winners, losers, and a slice of value that vanishes for everyone.

Keep two categories separate whenever you evaluate a control. A transfer moves surplus from one group to another - the tenant who keeps a rent-controlled flat gains roughly what the landlord loses - and reasonable people disagree about whether a given transfer is desirable, because that is a normative question. A deadweight loss is different in kind: it is surplus that reaches nobody, destroyed because a trade both sides wanted did not happen. Almost every serious argument about price controls is really an argument about whether the transfer is worth the deadweight loss.

Key idea: Price controls produce both transfers between groups and deadweight loss that benefits no one, and only the second is unambiguously a cost to society.

Policy	Binds when	Result
Price ceiling	Set below equilibrium	Shortage
Price floor	Set above equilibrium	Surplus

The takeaway

Price controls are politically appealing because they seem to help a favored group directly. But by preventing prices from doing their rationing and signaling job, binding controls create shortages or surpluses and destroy some gains from trade. Understanding who is helped, who is hurt, and how large the deadweight loss is lets you evaluate the trade-off rather than assume the policy is simply good or bad. Economists often prefer targeted transfers - a housing voucher or a wage subsidy - that help the intended group without breaking the price mechanism, though those carry budgetary costs of their own.

The empirical record on rent control illustrates both halves of the trade-off. Diamond, McQuade, and Qian's study of San Francisco found that tenants covered by rent control did benefit and were substantially more likely to stay in the city, exactly as intended - while landlords responded by converting and redeveloping buildings, reducing the rental housing supply and pushing market rents up for everyone not covered. The policy worked for its beneficiaries and worsened the underlying shortage at the same time. That is the characteristic shape of a price control, and it is why the honest question is never "does it help someone?" but "whom does it help, whom does it hurt, and how much value disappears in between?"

Key idea: Evaluating a price control means naming the winners, the losers, and the deadweight loss, then asking whether a targeted transfer could achieve the same distributional goal without suppressing the price signal.

Common wrong turns

• Assuming any ceiling lowers prices. A ceiling above equilibrium is non-binding and changes nothing at all.
• Thinking a ceiling helps all buyers. It helps those who still manage to buy and hurts those rationed out entirely.
• Reading the shortage as the quantity traded. Trade equals quantity supplied, 36 units - the shortage of 36 is the unmet want on top of that.
• Confusing transfers with deadweight loss. A transfer changes who holds surplus; deadweight loss destroys it.
• Treating the competitive prediction as the final word on minimum wages. Under monopsony the sign can reverse, and the empirical effects are often small.
• Forgetting non-price rationing. When price cannot ration, queues, quality cuts, and favouritism do - and those are real costs the diagram omits.

Recap

• A ceiling binds only below equilibrium and creates a shortage; a floor binds only above it and creates a surplus.
• With Qd = 100 - 2P and Qs = -20 + 4P, a ceiling at $14 gave Qd = 72, Qs = 36, a shortage of 36, and only 36 units traded.
• The blocked trades cost 0.5 x 24 x 18 = $216 of deadweight loss, confirmed by comparing total surplus of $1,350 before and $1,134 after.
• With Ld = 100 - 5W and Ls = -20 + 5W, a $15 wage floor cut employment from 40 to 25, created a labor surplus of 30, and destroyed $45 of surplus per hour.
• Controls transfer surplus between groups as well as destroying some, and only the destruction is unambiguously a social cost.
• Real minimum-wage effects depend on employer market power and are empirically modest, so the model shows the mechanism rather than settling the debate.
• Targeted transfers such as vouchers or wage subsidies can pursue the same goal without suppressing the price signal, at a budgetary cost.

Sources

1. Greenlaw, S. A., Shapiro, D., & MacDonald, D. (2022). Price ceilings and price floors. In Principles of Microeconomics 3e. OpenStax, Rice University. openstax.org
2. Greenlaw, S. A., Shapiro, D., & MacDonald, D. (2022). Demand, supply, and efficiency. In Principles of Microeconomics 3e. OpenStax, Rice University. openstax.org
3. Block, W. (n.d.). Rent control. In The Concise Encyclopedia of Economics. Liberty Fund. econlib.org
4. Gorman, L. (n.d.). Minimum wages. In The Concise Encyclopedia of Economics. Liberty Fund. econlib.org
5. Rockoff, H. (n.d.). Price controls. In The Concise Encyclopedia of Economics. Liberty Fund. econlib.org
6. Diamond, R., McQuade, T., & Qian, F. (2018). The effects of rent control expansion on tenants, landlords, and inequality: Evidence from San Francisco (NBER Working Paper No. 24181). National Bureau of Economic Research. nber.org
7. Card, D., & Krueger, A. B. (1993). Minimum wages and employment: A case study of the fast food industry in New Jersey and Pennsylvania (NBER Working Paper No. 4509). National Bureau of Economic Research. nber.org

A city council wants to cut smoking and proposes a cigarette tax. A streaming service wonders whether dropping its monthly price would win enough new subscribers to be worth it. A wheat farmer watches a record harvest come in and wonders, uneasily, whether a bumper crop will make him poorer. All three questions have the same shape, and none of them can be answered by the law of demand alone. Direction is not enough; you need magnitude. Elasticity is how economists put a number on it.

Why elasticity?

The law of demand tells us the direction quantity moves when price changes, but not by how much. That magnitude is often what a business or a government most needs to know. Price elasticity of demand (Ed) measures the responsiveness of quantity demanded to a change in price:

Ed = percent change in quantity demanded / percent change in price

Because the two changes move in opposite directions (law of demand), Ed is negative; economists usually discuss its absolute value, written |Ed|, and speak of demand being "more" or "less" elastic. Using percentages rather than raw units makes elasticity unit-free, so you can compare the price sensitivity of gasoline (gallons) and airline seats (tickets) on the same scale.

That unit-free property is worth demonstrating rather than asserting, because it is the whole reason elasticity beats slope. Take a price rise from $4 to $6 with quantity falling from 120 to 80 gallons. Measure the same event in liters instead: 120 gallons is 454.2 liters and 80 gallons is 302.8 liters, so the quantity change is -151.4 on an average of 378.5, which is -0.40 - exactly the same percentage as -40 on an average of 100. Measure price in cents rather than dollars and 400 to 600 gives 200 over an average of 500, again 0.40. Elasticity is unitless: it is a pure number, a ratio of two ratios, and it does not change when you change the yardstick. A slope, by contrast, is measured in gallons per dollar and becomes a completely different number in liters per cent, which is why slope cannot be compared across goods and elasticity can.

Key idea: Elasticity is a unitless ratio of percentage changes, so it can be compared across goods measured in entirely different units - something a slope can never do.

The midpoint (arc) method

Ordinary percent changes give different answers depending on the direction you move: a rise from 100 to 150 is a 50 percent increase, but the reverse fall from 150 to 100 is only a 33 percent decrease. The midpoint method fixes this by dividing each change by the average of the start and end values, so you get the same elasticity whether price rises or falls:

percent change in Q = (Q2 - Q1) / ((Q2 + Q1)/2)
percent change in P = (P2 - P1) / ((P2 + P1)/2)
Ed = (percent change in Q) / (percent change in P)

The reason this works is simple: both directions of travel share the same denominator. Ordinary percentage change divides by the starting value, and the starting value differs depending on which way you go, so the answer differs too. The midpoint formula divides by the average of the two endpoints, which is the same number whichever endpoint you started from. Economists therefore call the result an arc elasticity - one number describing the whole segment between two points, rather than a value at a single point.

Key idea: The midpoint method divides by the average of the two endpoints, so it gives the same elasticity whether price rises or falls - removing the direction-dependence of ordinary percentage change.

Worked example

Price rises from $4 to $6; quantity falls from 120 to 80.

percent change in Q = (80 - 120) / ((80 + 120)/2) = -40 / 100 = -0.40
percent change in P = (6 - 4) / ((6 + 4)/2) = 2 / 5 = 0.40
Ed = -0.40 / 0.40 = -1.0

The absolute value is 1, so demand here is unit elastic: quantity changes in exactly the same proportion as price. Work the reverse to see the midpoint method's payoff: dropping price from $6 to $4 while quantity rises 80 to 120 gives a quantity change of 40/100 = 0.40 and a price change of -2/5 = -0.40, again |Ed| = 1.0. Same answer either direction, which is the whole point.

Compare what the naive method would have produced. Going up, ordinary percentage change divides by the starting values: quantity falls 40/120 = -33.3 percent while price rises 2/4 = 50 percent, giving |Ed| = 0.67 and the label "inelastic." Coming back down, quantity rises 40/80 = 50 percent while price falls 2/6 = -33.3 percent, giving |Ed| = 1.5 and the label "elastic." The same two points, the same market, two contradictory conclusions - which is exactly the disease the midpoint method cures.

Key idea: Without the midpoint correction the same pair of points yields |Ed| = 0.67 one way and 1.5 the other; with it, both directions give 1.0.

Classifying elasticity

|Ed|	Name	Meaning
> 1	Elastic	Quantity responds more than proportionally
= 1	Unit elastic	Proportional response
< 1	Inelastic	Quantity responds less than proportionally
0	Perfectly inelastic	Quantity does not change (vertical curve)
infinite	Perfectly elastic	Any price rise drops quantity to zero (horizontal curve)

Elasticity varies along a straight-line demand curve

A common trap is thinking a demand curve has one elasticity. On a linear demand curve, slope is constant but elasticity is not. Near the top (high price, low quantity) demand is elastic, because a small absolute price change is a small percentage of a big price while the quantity change is a big percentage of a tiny quantity. Near the bottom (low price, high quantity) demand is inelastic, and exactly at the midpoint it is unit elastic. So "elastic" and "inelastic" describe a region of a curve, not the whole curve, unless the curve has a special constant-elasticity shape.

Worked example: three segments of one straight line

Take the demand curve Qd = 10 - P and compute the elasticity of three different segments with the midpoint method.

Segment	Percent change in Q	Percent change in P	Ed	Label
P: $8 to $6 (Q: 2 to 4)	2 / 3 = 0.667	-2 / 7 = -0.286	-2.33	Elastic
P: $6 to $4 (Q: 4 to 6)	2 / 5 = 0.400	-2 / 5 = -0.400	-1.00	Unit elastic
P: $4 to $2 (Q: 6 to 8)	2 / 7 = 0.286	-2 / 3 = -0.667	-0.43	Inelastic

Trace one row to be sure. In the top segment quantity goes 2 to 4, an increase of 2 on an average of 3, so 0.667. Price goes $8 to $6, a fall of 2 on an average of 7, so -0.286. Dividing gives -2.33, comfortably elastic. Every segment sits on the same straight line with the same constant slope of -1, yet the elasticity swings from 2.33 to 0.43. High up, a $2 move is a small percentage of a large price while the quantity move is a large percentage of a tiny quantity; down low, the arithmetic reverses. The unit-elastic point falls at the middle of the line, at P = $5 and Q = 5.

Key idea: Slope and elasticity are different things - a straight-line demand curve has one slope but every elasticity from infinite to zero along its length.

What makes demand elastic?

• Availability of substitutes - more and closer substitutes make demand more elastic, because buyers can easily switch away.
• Necessity vs luxury - necessities tend to be inelastic (insulin); luxuries elastic (cruise vacations).
• Share of budget - goods that eat a large share of income are more elastic, because a price change is felt more.
• Time horizon - demand is more elastic in the long run, as buyers adjust habits and find alternatives. Gasoline demand is inelastic this week but far more elastic over years as people buy efficient cars.
• Definition of the market - "a particular brand of cola" is far more elastic than "beverages" in general, because the narrower the good, the more substitutes it has.

Why it matters

Elasticity is the hinge between the last module and the next lesson. A government taxing cigarettes to cut smoking cares whether teen demand is elastic; a firm deciding whether a price cut will pay off cares whether its demand is elastic; a farmer bracing for a bumper crop cares that food demand is inelastic. The single number Ed converts the vague idea of "sensitivity" into a lever policymakers and managers can actually pull.

Key idea: Substitutes, budget share, necessity, time horizon, and how narrowly the market is defined together determine elasticity - and "more substitutes" is the thread running through all of them.

Elasticity on the supply side

The same construction works for sellers. Price elasticity of supply is the percentage change in quantity supplied divided by the percentage change in price, and because both move the same way it is positive. Its determinants are about flexibility: how easily producers can add capacity, whether inputs are readily available, whether the good can be stored, and above all how much time they have. A parking garage on the night of a concert has a perfectly inelastic supply - the spaces exist or they do not - while over a decade the supply of parking is highly elastic. This is the same time-horizon logic as on the demand side, and it explains why sudden shocks move prices violently while the same shock sustained over years moves quantities instead.

Key idea: Supply elasticity is positive and rises with the time available to adjust, which is why short-run shocks show up mainly in prices and long-run changes show up mainly in quantities.

Common wrong turns

• Confusing slope with elasticity. One straight line has a constant slope and a continuously changing elasticity.
• Using ordinary percentage change. It gives a different answer depending on direction - 0.67 one way, 1.5 the other, for the very same two points.
• Attaching units to Ed. Elasticity is a pure number; "1.8 gallons per dollar" is a slope, not an elasticity.
• Being confused by the negative sign. Ed is negative by the law of demand, so comparisons use the absolute value.
• Defining the market too broadly. "Food" is inelastic; "this brand of frozen pizza" is highly elastic, because substitutes multiply as the definition narrows.
• Treating a short-run estimate as permanent. Almost everything is more elastic once buyers have had time to adjust.

Recap

• Price elasticity of demand is the percentage change in quantity demanded divided by the percentage change in price, and it is unitless.
• Measuring the same gasoline example in liters and cents left Ed at -1.0, which is what "unitless" means in practice.
• The midpoint method divides by the average of the endpoints, so it returns the same elasticity in both directions instead of 0.67 versus 1.5.
• Price $4 to $6 with quantity 120 to 80 gives -0.40 / 0.40 = -1.0, unit elastic.
• Along Qd = 10 - P the elasticity ran 2.33, then 1.00, then 0.43 - elastic at the top, unit elastic at the midpoint, inelastic at the bottom.
• Substitutes, budget share, necessity versus luxury, time, and market definition drive how elastic demand is.
• Supply elasticity is positive and grows with time, so short-run shocks hit prices and long-run adjustments hit quantities.

Sources

1. Greenlaw, S. A., Shapiro, D., & MacDonald, D. (2022). Price elasticity of demand and price elasticity of supply. In Principles of Microeconomics 3e. OpenStax, Rice University. openstax.org
2. Greenlaw, S. A., Shapiro, D., & MacDonald, D. (2022). Polar cases of elasticity and constant elasticity. In Principles of Microeconomics 3e. OpenStax, Rice University. openstax.org
3. Greenlaw, S. A., Shapiro, D., & MacDonald, D. (2022). Elasticity and pricing. In Principles of Microeconomics 3e. OpenStax, Rice University. openstax.org
4. U.S. Department of Agriculture, Economic Research Service. (n.d.). Commodity and food elasticities. USDA ERS. ers.usda.gov
5. Chaloupka, F. J., & Warner, K. E. (1999). The economics of smoking (NBER Working Paper No. 7047). National Bureau of Economic Research. nber.org
6. Henderson, D. R. (n.d.). Demand. In The Concise Encyclopedia of Economics. Liberty Fund. econlib.org
7. CORE Econ Team. (2017). The firm and its customers. In The Economy 1.0. CORE Econ. core-econ.org

A gym owner is about to raise membership fees and cannot decide whether she is about to make more money or drive her members away. A supermarket wants to know whether cutting the price of hot dogs will sell more buns. A city planner forecasting transit ridership needs to know what happens when incomes rise. Each question is answered by an elasticity - but not the same one. This lesson takes the tool from the last lesson and points it in three directions: at revenue, at income, and at the price of a different good.

Elasticity and total revenue

Total revenue (TR) equals price times quantity, TR = P x Q. A price change moves P and Q in opposite directions, so the effect on revenue depends on elasticity - which force wins:

• If demand is elastic (|Ed| > 1), quantity moves more than price. A price cut raises revenue; a price increase lowers it.
• If demand is inelastic (|Ed| < 1), price moves more than quantity. A price increase raises revenue; a price cut lowers it.
• If demand is unit elastic, revenue is unchanged at the margin (it is maximized).

This is the total revenue test, and it runs both ways: if you observe that a price hike raised revenue, you have learned demand was inelastic over that range. It is why a farmer facing inelastic demand can see revenue fall after a bumper harvest pushes prices down, and why a business with elastic demand may boost revenue by discounting.

Key idea: Price and quantity move in opposite directions, so whether revenue rises or falls with a price change depends entirely on which movement is proportionally larger - which is what elasticity measures.

Worked example: revenue along a whole demand curve

Return to the demand curve from the previous lesson, Qd = 10 - P, and compute revenue at every price.

Price	Quantity	Total revenue	Region
$9	1	$9	Elastic
$8	2	$16	Elastic
$7	3	$21	Elastic
$6	4	$24	Elastic
$5	5	$25	Unit elastic
$4	6	$24	Inelastic
$3	7	$21	Inelastic
$2	8	$16	Inelastic
$1	9	$9	Inelastic

Read the revenue column from the top. Cutting the price from $9 to $8 raises revenue from $9 to $16, and every further cut keeps raising it - all the way down to $5, where revenue peaks at $25. Cut past that and revenue falls back: $24, $21, $16, $9. The peak sits exactly at the unit-elastic midpoint identified in the last lesson, P = $5 and Q = 5, which is no coincidence. Above it demand is elastic and price cuts pay; below it demand is inelastic and price cuts destroy revenue.

One more observation, because it returns in the monopoly lesson. Look at what the last unit adds to revenue. Moving from Q = 4 to Q = 5 lifts revenue from $24 to $25, so that unit adds $1. Moving from Q = 5 to Q = 6 drops revenue from $25 to $24, so that unit adds -$1. Marginal revenue passes through zero exactly where total revenue peaks and demand is unit elastic - which is why no profit-maximising seller with positive costs ever operates on the inelastic half of its demand curve.

Key idea: Total revenue rises as price falls through the elastic region, peaks where demand is unit elastic and marginal revenue is zero, and falls through the inelastic region.

Worked example: the total revenue test with numbers

A theater sells 400 tickets at $50, for TR = $20,000. It cuts the price to $40 and sells 600 tickets, for TR = $24,000. Revenue rose after a price cut, so demand must be elastic over this range. Check with the midpoint method: the quantity change is 200/500 = 0.40 and the price change is -10/45 = -0.222, so |Ed| = 1.8 > 1. The elasticity number and the revenue movement agree, as they always must.

Worked example: should the gym raise its price?

A gym charges $40 a month and has 500 members, so total revenue is 40 x 500 = $20,000. Market research puts the elasticity of demand for memberships at about 0.6 in absolute value - inelastic, because the nearest gym is across town. The owner is considering a 10 percent increase to $44.

Work it through. With |Ed| = 0.6, a 10 percent price rise cuts quantity by 0.6 x 10 = 6 percent, so membership falls to 500 x 0.94 = 470. New revenue is 44 x 470 = $20,680, an increase of $680. The revenue test predicted this before any arithmetic: demand is inelastic, so raising price raises revenue.

Then add the caution that separates a good analyst from a careless one. Revenue is not profit. The gym also sheds the cost of serving 30 fewer members - towels, cleaning, staffing at peak - so profit rises by more than $680. Conversely, a business with elastic demand that cuts price to raise revenue may find profit falling, because the extra units brought in still cost something to produce. Elasticity tells you what happens to revenue; only bringing in costs tells you what happens to profit.

Key idea: The total revenue test predicts revenue, not profit, and a complete pricing decision must also account for the cost of serving the units gained or lost.

Income elasticity of demand

Income elasticity (E_I) measures how quantity demanded responds to a change in income:

E_I = percent change in quantity demanded / percent change in income

• E_I > 0: a normal good (demand rises with income). If E_I > 1 it is a luxury (demand grows faster than income); if 0 < E_I < 1 it is a necessity (demand grows, but slower than income).
• E_I < 0: an inferior good (demand falls as income rises).

Worked example: income rises 10 percent and your restaurant meals rise 25 percent. E_I = 25 / 10 = +2.5. Positive and above 1, so restaurant meals are a normal good and specifically a luxury for you. If instead your instant-noodle purchases fell 5 percent, E_I = -5 / 10 = -0.5, marking noodles an inferior good.

Use the midpoint method here too when you are given levels rather than percentages. Suppose a household's income rises from $40,000 to $44,000 and its restaurant meals go from 20 a year to 28. The income change is 4,000 divided by the average of 42,000, which is 0.0952. The quantity change is 8 divided by the average of 24, which is 0.3333. So E_I = 0.3333 / 0.0952 = +3.5: strongly positive and far above 1, so restaurant meals are a normal good and a pronounced luxury for this household. The same discipline as before applies - divide by the average of the endpoints so the answer does not depend on which direction you travelled.

Key idea: Income elasticity classifies goods by sign - positive for normal, negative for inferior - and by size, with values above 1 marking luxuries and values between 0 and 1 marking necessities.

Cross-price elasticity of demand

Cross-price elasticity (E_xy) measures how the quantity demanded of good X responds to a change in the price of good Y:

E_xy = percent change in quantity demanded of X / percent change in price of Y

• E_xy > 0: X and Y are substitutes (Y gets pricier, buyers switch to X, so X's quantity rises).
• E_xy < 0: X and Y are complements (Y gets pricier, people buy less of both).
• E_xy near 0: the goods are unrelated.

Worked example: the price of tea rises 20 percent and coffee purchases rise 8 percent. E_xy = 8 / 20 = +0.4. The positive sign confirms tea and coffee are substitutes. If instead cream purchases had fallen 8 percent when coffee's price rose, E_xy = -8 / 20 = -0.4, marking coffee and cream as complements.

With levels, run the midpoint formula on both goods. Say the price of tea rises from $2.50 to $3.50 while weekly coffee sales rise from 40 to 48. The tea price change is 1 divided by the average of 3.00, or 0.3333. The coffee quantity change is 8 divided by the average of 44, or 0.1818. So E_xy = 0.1818 / 0.3333 = +0.55. Positive confirms substitutes, and the modest size says they are imperfect ones - a tea drinker faced with pricier tea switches to coffee only sometimes. Cross-price elasticities near zero, by contrast, tell antitrust authorities that two products are not really in the same market at all, which is why these numbers appear in merger cases as well as in marketing plans.

Key idea: Cross-price elasticity is positive for substitutes and negative for complements, and its magnitude measures how closely the two goods compete.

Reading the signs at a glance

Two elasticities carry meaning in their sign, unlike price elasticity of demand, whose sign is always negative and therefore uninformative. For income elasticity, the sign separates normal (+) from inferior (-) goods, and the size separates luxuries (>1) from necessities (<1). For cross-price elasticity, the sign separates substitutes (+) from complements (-). Whenever you compute one of these, state the sign first and interpret it, then use the magnitude for the finer classification. Firms use these numbers constantly: a supermarket setting the price of hot dogs will consider the cross-price elasticity with buns (complement) before deciding.

Key idea: Price elasticity carries information only in its size because its sign is always negative, while income and cross-price elasticities carry their most important information in their sign.

Common wrong turns

• Assuming price cuts always raise revenue. They do so only where demand is elastic; below the unit-elastic point they reduce it.
• Treating revenue as profit. The revenue test ignores the cost of serving the units gained or lost.
• Ignoring the sign on income elasticity. The sign, not the size, is what separates normal from inferior goods.
• Calling a negative cross-price elasticity "substitutes." Negative means complements: pricier coffee means less cream.
• Reporting price elasticity as negative when comparing magnitudes. Convention compares |Ed|, so -2.0 is more elastic than -0.5.
• Applying one elasticity everywhere on the curve. Elasticity varies along the curve, so a number estimated near current prices need not survive a large price change.

Recap

• Total revenue is P x Q, and a price change moves the two in opposite directions, so elasticity decides which effect wins.
• Along Qd = 10 - P revenue climbed to a peak of $25 at the unit-elastic point P = $5, then fell - and marginal revenue passed through zero there.
• A theater raising ticket sales from 400 at $50 to 600 at $40 raised revenue to $24,000, consistent with |Ed| = 1.8.
• The gym with |Ed| = 0.6 raising price 10 percent lost 6 percent of members and gained revenue: 44 x 470 = $20,680.
• Income elasticity of +3.5 for a household whose income rose 9.5 percent and restaurant meals 33 percent marks meals a luxury.
• Cross-price elasticity of +0.55 between tea and coffee confirms imperfect substitutes; negative values mark complements.
• Interpret the sign of income and cross-price elasticities first, then use the magnitude for the finer classification.

Sources

1. Greenlaw, S. A., Shapiro, D., & MacDonald, D. (2022). Elasticity and pricing. In Principles of Microeconomics 3e. OpenStax, Rice University. openstax.org
2. Greenlaw, S. A., Shapiro, D., & MacDonald, D. (2022). Elasticity in areas other than price. In Principles of Microeconomics 3e. OpenStax, Rice University. openstax.org
3. Greenlaw, S. A., Shapiro, D., & MacDonald, D. (2022). Price elasticity of demand and price elasticity of supply. In Principles of Microeconomics 3e. OpenStax, Rice University. openstax.org
4. Greenlaw, S. A., Shapiro, D., & MacDonald, D. (2022). How changes in income and prices affect consumption choices. In Principles of Microeconomics 3e. OpenStax, Rice University. openstax.org
5. U.S. Department of Agriculture, Economic Research Service. (n.d.). Commodity and food elasticities. USDA ERS. ers.usda.gov
6. Henderson, D. R. (n.d.). Demand. In The Concise Encyclopedia of Economics. Liberty Fund. econlib.org
7. CORE Econ Team. (2017). The firm and its customers. In The Economy 1.0. CORE Econ. core-econ.org

So far the demand curve has been an assumption. We drew it sloping downward because buyers behave that way, and left it there. That is unsatisfying: a model that assumes its own conclusion explains nothing. This lesson goes underneath the curve and builds it from something more primitive - a person with a fixed budget, a set of things they might buy, and a preference for more satisfaction over less. The downward slope will fall out as a consequence rather than an assumption.

Utility

Utility is the satisfaction a consumer gets from consuming goods and services. We do not need to measure it in real physical units; we only need consumers to compare and rank options consistently. Early economists imagined a cardinal unit called a "util" as a teaching device, but modern theory rests only on the weaker assumption that people can say which bundle they prefer. Total utility is overall satisfaction from a quantity consumed; marginal utility (MU) is the extra utility from one more unit. The marginal concept, as in Module 1, is what drives decisions.

Diminishing marginal utility

The law of diminishing marginal utility says that as you consume more of a good within a given period, the marginal utility of each additional unit eventually falls. The first slice of pizza when hungry is wonderful; the fourth is merely fine; the sixth might make you queasy. Total utility can still rise while marginal utility falls - it just rises more slowly. If a unit ever makes you worse off, its marginal utility is negative and total utility actually declines.

Slices	Total utility	Marginal utility
1	20	20
2	34	14
3	44	10
4	50	6
5	52	2

Notice each MU is the increase in total utility from the previous row, and MU is shrinking - classic diminishing marginal utility. Marginal utility is the "slope" of the total utility curve: where MU is positive but falling, total utility rises at a decreasing rate.

Key idea: Marginal utility is the increment to total utility from one more unit, and it falls as consumption of a good rises within a period - which is why total utility can keep rising even as each extra unit matters less.

The utility-maximizing rule

A consumer with a fixed budget maximizes utility by allocating spending so that the marginal utility per dollar is equal across all goods:

MU_x / P_x = MU_y / P_y for all goods

The logic is marginal: if one good delivered more utility per dollar than another, you could raise total utility by shifting a dollar toward it. You keep shifting until the "bang per buck" is equalized everywhere and the budget is spent. Two conditions must both hold at the optimum: the equal-bang-per-buck condition above, and the budget being fully used. If either fails, a better bundle exists.

Worked example

You have $12. Burritos cost $4 with MU = 40; smoothies cost $2 with MU = 30. Compare bang per buck: burrito MU/P = 40/4 = 10 utils per dollar; smoothie MU/P = 30/2 = 15 utils per dollar. The smoothie gives more utility per dollar, so buy smoothies first. As you consume more smoothies their MU falls (diminishing marginal utility); you keep buying whichever good has the higher MU/P until the two ratios are equal and the $12 is gone. This step-by-step reallocation is exactly how the optimum is reached in practice - you never need to know the total utility number, only the marginal comparison.

Worked example: spending a whole budget

Now run the rule to completion with a real budget of $16. Burritos cost $4, smoothies cost $2, and here are the marginal utilities together with the bang per buck each delivers:

Unit	Burrito MU	Burrito MU per $	Smoothie MU	Smoothie MU per $
1st	40	10	30	15
2nd	32	8	24	12
3rd	24	6	20	10
4th	16	4	16	8
5th	-	-	12	6
6th	-	-	8	4

Every entry in a "per dollar" column is the MU divided by that good's price: the second burrito gives 32/4 = 8 utils per dollar; the third smoothie gives 20/2 = 10. Now buy greedily, always taking the highest available bang per buck, and track what is left of the $16.

The first smoothie leads at 15, costing $2 and leaving $14. The second smoothie at 12 leaves $12. The third smoothie at 10 ties with the first burrito at 10; take the smoothie first, leaving $10, then the burrito for $4, leaving $6. Now the fourth smoothie and the second burrito are tied at 8. Take the smoothie ($2, leaving $4) and then the burrito ($4, leaving $0). The budget is exhausted.

The optimal bundle is 2 burritos and 4 smoothies, costing 2 x $4 + 4 x $2 = $16 exactly. Total utility is (40 + 32) + (30 + 24 + 20 + 16) = 72 + 90 = 162 utils. Check that no rearrangement beats it: 3 burritos and 2 smoothies gives 96 + 54 = 150, one burrito and six smoothies gives 40 + 110 = 150, four burritos alone gives 112, and eight smoothies gives less still. The greedy rule found the true maximum.

Confirm the optimality condition. The last burrito bought delivered 32/4 = 8 utils per dollar and the last smoothie delivered 16/2 = 8 - equal, exactly as the rule requires. Equivalently, MU_burrito / MU_smoothie = 32/16 = 2, which matches the price ratio P_burrito / P_smoothie = 4/2 = 2. Those two statements are the same condition written differently, and either one is a valid test that a bundle is optimal.

Key idea: A bundle is optimal when the budget is fully spent and the last dollar spent on each good buys the same amount of utility - equivalently, when the ratio of marginal utilities equals the ratio of prices.

From utility to the demand curve

Diminishing marginal utility is exactly why demand curves slope downward. Since each extra unit is worth less to you, you will only buy more if the price is lower. More formally, a consumer buys additional units up to the point where marginal utility (in dollar terms, MU/P) equals the marginal utility per dollar available elsewhere. When a good's price falls, its MU/P rises, so the optimum shifts toward buying more of it - a larger quantity demanded at the lower price. The law of demand thus emerges from rational, utility-maximizing choice rather than being assumed.

Watch it happen in the numbers above. At $4 a burrito the optimum held 2 burritos. Suppose burritos fall to $2. Every burrito's bang per buck doubles - the third burrito now yields 24/2 = 12 utils per dollar instead of 6 - so burritos start outranking smoothies further down the list, and the reallocation pulls the bundle toward more burritos. A lower price raised the quantity demanded of burritos without any change in tastes, income, or the price of smoothies. Repeat the exercise at every possible burrito price and you trace out a demand curve, point by point, from nothing but preferences and a budget.

Key idea: Lowering a good's price raises its marginal utility per dollar at every quantity, which pulls the optimal bundle toward that good - and repeating this at every price traces the demand curve.

Where the model simplifies

This apparatus assumes a consumer with stable, consistent preferences who knows the options and computes. Real shoppers are noisier, and the departures are systematic enough to have their own literature. People's choices depend on how options are framed - the same yoghurt described as 90 percent fat-free or 10 percent fat is chosen at different rates. They anchor on irrelevant reference points, weigh a loss more heavily than an equal gain, and often pick the default simply because it is the default. Daniel Kahneman's Nobel lecture surveys the evidence that judgment under uncertainty follows heuristics rather than expected-utility arithmetic.

The right response is calibration, not abandonment. The equal-bang-per-buck rule predicts aggregate market behaviour well, especially for repeated purchases where feedback teaches, and it is the foundation on which the demand curves in every later module rest. But it describes a tendency rather than a mechanism, and where a policy depends on people making a single high-stakes decision correctly - choosing a pension, a mortgage, a health plan - the behavioural qualifications matter a great deal.

Key idea: Utility maximisation is a good aggregate description and a poor literal one, and behavioural economics documents the framing, reference-point, and default effects that the standard model omits.

Why it matters

This lesson closes a loop opened in Module 2. There we simply asserted that demand slopes down; here we derive it from how people actually weigh satisfaction against price. The equal-marginal-utility-per-dollar rule also generalizes far beyond shopping: it is the same "equate marginal returns across uses" logic a firm uses to allocate a budget across inputs, or an investor uses across assets. Whenever a scarce resource must be split among competing uses, the efficient split equalizes the marginal payoff per unit of resource in every use.

Common wrong turns

• Buying the good with the higher marginal utility. Compare MU per dollar; a 40-util burrito at $4 loses to a 30-util smoothie at $2.
• Thinking falling MU means falling total utility. Total utility keeps rising while MU is positive; it falls only once MU turns negative.
• Stopping before the budget is spent. Optimality needs both equal bang per buck and a fully used budget.
• Treating utils as measurable quantities. They are a bookkeeping device; only the rankings and comparisons carry meaning.
• Equating MU across goods instead of MU per dollar. At the optimum MU_x/MU_y equals the price ratio, not 1.
• Reading the model as a literal description of shoppers. It is a benchmark, and framing and default effects are real deviations from it.

Recap

• Utility is satisfaction; total utility is the whole amount and marginal utility the increment from one more unit.
• Marginal utility diminishes, so the pizza table's MU fell 20, 14, 10, 6, 2 while total utility kept climbing to 52.
• A consumer maximises utility by equalising marginal utility per dollar across goods and spending the whole budget.
• With $16, burritos at $4 and smoothies at $2, greedy allocation gave 2 burritos and 4 smoothies for 162 utils.
• At that optimum the last burrito and last smoothie both returned 8 utils per dollar, and MU_burrito/MU_smoothie = 2 matched the price ratio.
• Cutting a good's price raises its bang per buck at every quantity, pulling the bundle toward it - which is the demand curve being derived rather than assumed.
• The model is a benchmark; framing, reference points, and defaults produce documented departures from it.

Sources

1. Greenlaw, S. A., Shapiro, D., & MacDonald, D. (2022). Consumption choices. In Principles of Microeconomics 3e. OpenStax, Rice University. openstax.org
2. Greenlaw, S. A., Shapiro, D., & MacDonald, D. (2022). Behavioral economics: An alternative framework for consumer choice. In Principles of Microeconomics 3e. OpenStax, Rice University. openstax.org
3. Rhoads, S. E. (n.d.). Marginalism. In The Concise Encyclopedia of Economics. Liberty Fund. econlib.org
4. Mullainathan, S., & Thaler, R. H. (n.d.). Behavioral economics. In The Concise Encyclopedia of Economics. Liberty Fund. econlib.org
5. Kahneman, D. (2002). Maps of bounded rationality [Prize lecture]. The Nobel Foundation. nobelprize.org
6. Marshall, A. (1920). Principles of economics (8th ed.). Library of Economics and Liberty. econlib.org
7. CORE Econ Team. (2023). Scarcity, work, and choice. In The Economy 2.0: Microeconomics. CORE Econ. core-econ.org

Wanting something is free; buying it is not. The last lesson gave a consumer preferences and let them rank bundles, but ranking alone would have them consuming everything. What stops them is a budget - a hard line across the space of possible bundles separating the affordable from the merely desirable. This lesson draws that line, and then uses the gap between what buyers would have paid and what they actually paid to build the measuring stick that the rest of the course uses to judge whether a market is doing its job.

The budget constraint

A consumer's choices are limited by income and prices. The budget constraint shows all combinations of two goods that exactly exhaust income I: (P_x times X) + (P_y times Y) = I. Graphed with X on the horizontal axis, its slope is -P_x/P_y, the rate at which the market lets you trade one good for the other. The intercepts are the most of each good you could buy if you spent everything on it: I/P_x and I/P_y. Every point on the line is affordable and spends the whole budget; points below it are affordable but leave money unspent; points above it are unaffordable.

Changes move the line predictably. A rise in income shifts it outward (parallel) so more of both goods is affordable, without changing the slope, because prices are unchanged. A change in one good's price rotates the line, pivoting around the unchanged intercept and changing the slope. For instance, if X gets cheaper, the X-intercept moves further out (you can buy more X with all your income) while the Y-intercept stays put, flattening the line. The consumer's best choice is the affordable bundle that reaches the highest utility - intuitively, where the marginal-utility-per-dollar rule from the last lesson is satisfied.

Worked example: drawing and moving a budget line

Let income be $120, books cost $20, and meals out cost $10. The constraint is 20X + 10Y = 120 with books on the horizontal axis.

The intercepts come straight from division: spending everything on books buys 120/20 = 6 books, while spending everything on meals buys 120/10 = 12 meals. The slope is -P_x/P_y = -20/10 = -2, meaning the market obliges you to surrender 2 meals for every additional book, which is simply the relative price restated as a trade-off.

Testing three bundles establishes the geography. Four books and four meals costs 80 + 40 = $120, so it sits exactly on the line and exhausts the budget; three books and five meals costs 60 + 50 = $110, which is affordable but leaves $10 unspent and therefore lies inside the line; five books and three meals costs 100 + 30 = $130, exceeding the budget entirely and lying outside it.

Now move it. Raise income to $150 with prices unchanged: the intercepts become 150/20 = 7.5 books and 150/10 = 15 meals, while the slope stays at -2. The line shifts outward parallel, because relative prices did not change. Instead hold income at $120 and cut the book price to $12: the book intercept moves out to 120/12 = 10 while the meal intercept stays at 12, and the slope flattens to -12/10 = -1.2. The line rotates. That difference matters: an income change alters what you can afford, while a price change alters both what you can afford and the terms on which you trade one good for the other.

Key idea: Income changes shift the budget line parallel because they leave relative prices alone, while a change in one price rotates the line and changes its slope.

Consumer surplus

The demand curve reveals what buyers are willing to pay for each unit - their reservation price, reflecting marginal benefit. Consumer surplus is the difference between what buyers are willing to pay and what they actually pay, summed across all units. On a graph it is the area below the demand curve and above the price. It measures the net gain buyers capture from being able to buy at a single market price rather than at their own maximum willingness to pay.

Worked example

Four buyers each want one concert ticket. Their willingness to pay is $50, $40, $30, and $20. The price is $25. Who buys, and what is total consumer surplus?

• The $50 buyer buys; surplus = 50 - 25 = $25.
• The $40 buyer buys; surplus = 40 - 25 = $15.
• The $30 buyer buys; surplus = 30 - 25 = $5.
• The $20 buyer does not buy (willingness to pay is below price), surplus = $0.

Total consumer surplus = 25 + 15 + 5 = $45. Each buyer captures the gap between personal value and the market price; the buyer who valued the ticket below its price simply stays out and loses nothing. Now suppose the price fell to $20: the fourth buyer enters with $0 surplus, and each of the other three gains an extra $5, so total surplus rises to (30 + 20 + 10 + 0) = $60. Lower prices raise consumer surplus both by helping existing buyers and by bringing new ones into the market.

Key idea: Consumer surplus is willingness to pay minus price, summed over all units bought - the area below the demand curve and above the price.

Worked example: surplus as an area

Four buyers made the idea concrete; a curve makes it general. Return to the market from Module 2, where demand was Qd = 100 - 2P and supply was Qs = -20 + 4P, meeting at P = $20 and Q = 60. Rewrite each with price on the left: inverse demand is P = 50 - 0.5Q and inverse supply is P = 5 + 0.25Q.

Consumer surplus is the triangle between the demand curve and the $20 price line. Its height is the distance from the choke price to the market price, 50 - 20 = $30, and its base is the quantity traded, 60. So consumer surplus = 0.5 x 60 x 30 = $900. Producer surplus is the triangle between the price line and the supply curve: height 20 - 5 = $15, base 60, so producer surplus = 0.5 x 60 x 15 = $450. Total surplus is 900 + 450 = $1,350.

Checking the components against intuition confirms the arithmetic: the very first unit is worth $50 to some buyer while costing some seller only $5 to supply, a gain of $45 on that single unit, whereas the sixtieth unit is worth $20 and costs $20, generating nothing. Because every unit in between contributes something, the triangles are simply the accumulation of those diminishing gains, and the final unit traded is always the one whose gain has dwindled to zero - which is precisely why the equilibrium quantity is the efficient quantity.

Key idea: Surplus is measured as an area, and total surplus of $1,350 in this market is the sum of every gain from every mutually beneficial trade that actually happened.

Producer surplus and total surplus

Sellers have an analog. Producer surplus is the price a seller receives minus the lowest price they would have accepted (their cost), summed over units. On a graph it is the area above the supply curve and below the price. Adding the two gives total surplus = consumer surplus + producer surplus, the standard yardstick for how much value a market creates. At the competitive equilibrium, total surplus is maximized: every trade whose benefit to a buyer exceeds its cost to a seller actually happens, and no wasteful trades occur.

Two honest caveats belong with the claim, because "maximises total surplus" is a narrower statement than it sounds. First, the result holds only when the assumptions hold: no market power, no externalities, no missing information. Each later module removes one of those and watches surplus fall. Second, total surplus is silent about distribution. A market can maximise the total while allocating almost all of it to one side, and moving a dollar of surplus from a wealthy seller to a struggling buyer leaves the total unchanged while plainly mattering to the people involved. Efficiency and fairness are different questions, and total surplus answers only the first.

Key idea: Competitive equilibrium maximises total surplus under its assumptions, but total surplus measures the size of the pie and says nothing about how it is divided.

Why surplus matters

Consumer surplus, producer surplus, and their sum are the tools economists use to judge whether a policy helps or hurts. Policies that shrink total surplus - such as binding price controls (Module 2), monopoly pricing (Module 6), or unaddressed externalities (Module 8) - are inefficient precisely because they leave gains from trade on the table, creating the deadweight loss you have already met.

Conversely, removing a barrier that had blocked beneficial trades raises total surplus. This welfare framework is the through-line that ties the whole second half of the course together: nearly every "market failure" is diagnosed as a loss of total surplus, and nearly every remedy is judged by whether it recovers more surplus than it costs.

Common wrong turns

• Reading the budget line's slope as a preference. It is a market trade-off set by relative prices, not a statement about what the consumer likes.
• Shifting the line when one price changes. A single price change rotates the line around the other intercept; only income changes shift it parallel.
• Counting a non-buyer's surplus as negative. A buyer whose willingness to pay is below the price simply stays out with zero surplus.
• Measuring surplus above the demand curve. Consumer surplus is below demand and above price; producer surplus is above supply and below price.
• Using the choke price as the height for producer surplus. Producer surplus uses the gap between price and the supply curve.
• Treating maximum total surplus as maximum fairness. Efficiency is about the size of the gains, not their distribution.

Recap

• The budget constraint (P_x times X) + (P_y times Y) = I has intercepts I/P_x and I/P_y and slope -P_x/P_y.
• With $120, books at $20 and meals at $10, the line ran from 6 books to 12 meals with slope -2; higher income shifted it parallel, a cheaper book rotated it to slope -1.2.
• Consumer surplus is willingness to pay minus price: the four ticket buyers at a $25 price captured $25 + $15 + $5 = $45.
• Cutting the price to $20 raised that to $60, both by helping existing buyers and by drawing in a new one.
• As areas, the running market gave consumer surplus of 0.5 x 60 x 30 = $900 and producer surplus of 0.5 x 60 x 15 = $450.
• Total surplus of $1,350 is the sum of every gain from every trade, and the last unit traded adds nothing - which is what makes the quantity efficient.
• Total surplus measures efficiency only; it is deliberately silent about who ends up with the gains.

Sources

1. Greenlaw, S. A., Shapiro, D., & MacDonald, D. (2022). How individuals make choices based on their budget constraint. In Principles of Microeconomics 3e. OpenStax, Rice University. openstax.org
2. Greenlaw, S. A., Shapiro, D., & MacDonald, D. (2022). How changes in income and prices affect consumption choices. In Principles of Microeconomics 3e. OpenStax, Rice University. openstax.org
3. Greenlaw, S. A., Shapiro, D., & MacDonald, D. (2022). Demand, supply, and efficiency. In Principles of Microeconomics 3e. OpenStax, Rice University. openstax.org
4. Greenlaw, S. A., Shapiro, D., & MacDonald, D. (2022). Indifference curves. In Principles of Microeconomics 3e. OpenStax, Rice University. openstax.org
5. Heyne, P. (n.d.). Efficiency. In The Concise Encyclopedia of Economics. Liberty Fund. econlib.org
6. Marshall, A. (1920). Principles of economics (8th ed.). Library of Economics and Liberty. econlib.org
7. CORE Econ Team. (2017). Supply and demand: Price-taking and competitive markets. In The Economy 1.0. CORE Econ. core-econ.org

Put one cook in a kitchen and the meals come slowly. Add a second and output more than doubles, because now one person works the grill while the other plates. Add a tenth and they are queueing for the same two burners, apologising as they squeeze past each other. Nothing about the tenth cook is worse than the first; what changed is that the kitchen did not grow. That single observation - a variable input pressed against a fixed one - generates the shape of every cost curve in the next lesson and, through them, the supply curve of the whole economy.

Firms as transformers of inputs

A firm combines inputs - labor, capital, materials - to produce output. The production function describes the maximum output achievable from given quantities of inputs. It is a statement about technology and engineering, not yet about cost or profit: it tells you what is physically possible before money enters the picture. Economists split time into two planning horizons. In the short run, at least one input (typically capital, such as the factory or equipment) is fixed. In the long run, all inputs are variable and the firm can change its entire scale, including building a new plant or exiting the industry.

The short-run/long-run split is defined by flexibility, not by a fixed number of months. For a food truck the "long run" might be weeks; for a nuclear plant it may be a decade. What matters is whether every input can be adjusted. This distinction organizes the whole cost analysis: fixed inputs create fixed costs in the short run, and the inability to adjust them is what produces diminishing returns.

Key idea: The short run is defined by having at least one input you cannot change, not by a length of time, and that fixed input is the source of everything that follows.

Total, marginal, and average product

Holding capital fixed and adding labor, three related measures describe output:

• Total product (TP): total output produced.
• Marginal product of labor (MPL): the extra output from one more worker, MPL = change in TP / change in labor.
• Average product (APL): output per worker, APL = TP / labor.

Workers	Total product	Marginal product	Average product
0	0	-	-
1	10	10	10.00
2	24	14	12.00
3	33	9	11.00
4	39	6	9.75
5	42	3	8.40

Each marginal product is the jump in total product from the row above: the third worker takes output from 24 to 33, so MPL = 9. Each average product is total product divided by the number of workers: at four workers, 39/4 = 9.75. Notice that average product peaks at two workers (12.00) and falls thereafter, while marginal product peaks at the same place and falls faster. That is not a coincidence, and the next section explains why.

Key idea: Marginal product is the extra output from one more worker and average product is output per worker, and the two are computed from the same total-product column in different ways.

The law of diminishing marginal returns

Reading the marginal-product column, output first rises quickly, then rises more slowly. Marginal product increased from worker 1 to worker 2 (10 to 14) - early workers benefit from specialization and teamwork - but from worker 3 onward MPL falls (9, 6, 3). This is the law of diminishing marginal returns: as more of a variable input is added to a fixed input, the marginal product of the variable input eventually declines.

With a fixed amount of capital, each additional worker has less equipment and space to work with, so each adds less than the last. Note the word "eventually": returns can rise at first (increasing marginal returns) before the inevitable decline sets in.

Two things diminishing returns is not. It is not a claim that workers get lazier or less skilled - every worker in the table is identical, and the decline comes entirely from the fixed capital they must share. And it is not a claim that total output falls: total product rises through the whole table, from 10 to 42. Only the increments shrink. Total output would fall only if marginal product turned negative, which happens in genuinely overcrowded workplaces where an extra body gets in everyone's way.

Key idea: Diminishing marginal returns comes from a fixed input being shared ever more thinly, not from declining worker quality, and it shrinks the increments to output rather than output itself.

The relationship between marginal and average product

Marginal and average product move together in a predictable way, and the pattern reappears for costs in the next lesson. When MPL is above APL, it pulls the average up; when MPL is below APL, it pulls the average down; so MPL crosses APL at APL's maximum. The reasoning is the same as a test-score average: a new score above your average raises it, and a score below it lowers it. This "marginal pulls the average" logic is worth internalizing now, because the identical relationship governs marginal cost and average cost, only flipped upside down.

Check it against the table. At two workers, average product is 12.00 and the second worker's marginal product is 14 - above the average, and indeed the average rose from 10.00 to 12.00. At three workers, marginal product is 9 while average product is 11.00 - below the average, and the average duly fell. So marginal product crosses average product somewhere between the second and third worker, which is exactly where average product peaks.

Key idea: Marginal product above average product pulls the average up and below it pulls the average down, so the two curves must cross at the maximum of average product.

Worked example: from product to cost

Here is where this lesson pays for itself. Suppose each worker costs a wage of $60 and labor is the only variable input. Then variable cost is simply 60 times the number of workers, and we can compute cost per unit directly from the product table.

Workers	Output	Variable cost	Marginal cost of the added output	Average variable cost
1	10	$60	60 / 10 = $6.00	60 / 10 = $6.00
2	24	$120	60 / 14 = $4.29	120 / 24 = $5.00
3	33	$180	60 / 9 = $6.67	180 / 33 = $5.45
4	39	$240	60 / 6 = $10.00	240 / 39 = $6.15
5	42	$300	60 / 3 = $20.00	300 / 42 = $7.14

Every marginal cost entry is the wage divided by that worker's marginal product, because hiring one more worker always costs $60 and always buys MPL extra units. The relationship is exact: MC = wage / MPL. Similarly AVC = wage / APL, which you can verify - at three workers, 60 / 11.00 = $5.45, matching 180/33.

Now read the columns as mirrors. Marginal product rose from 10 to 14 and marginal cost fell from $6.00 to $4.29. Marginal product then fell 14, 9, 6, 3 and marginal cost climbed $4.29, $6.67, $10.00, $20.00. Average variable cost bottoms out at $5.00 with two workers - exactly where average product peaks at 12.00. Diminishing returns and rising marginal cost are not two facts; they are one fact seen from opposite sides.

Key idea: With one variable input, MC = wage/MPL and AVC = wage/APL, so rising marginal cost is simply falling marginal product expressed in dollars.

Short-run returns versus long-run scale

One distinction prevents a persistent muddle. Diminishing marginal returns is a short-run idea: it requires at least one input to be fixed while another varies. Returns to scale is a long-run idea: it asks what happens to output when every input rises in proportion. Doubling labor alone in a fixed kitchen runs into diminishing returns; doubling labor, kitchen, and equipment together might double output (constant returns), more than double it (increasing returns), or less than double it (decreasing returns). A firm can have increasing returns to scale in the long run and diminishing marginal returns to labor in the short run at the same time, with no contradiction, because the two statements hold different things constant.

Key idea: Diminishing marginal returns varies one input against a fixed one in the short run; returns to scale varies all inputs together in the long run, and a firm can exhibit both at once.

Why it matters for costs

Diminishing marginal returns is the engine behind the shape of cost curves you will study next. When each extra worker produces less additional output, each extra unit of output requires more additional labor - so marginal cost rises. The falling productivity of inputs in the short run translates directly into rising marginal cost, and that in turn shapes the firm's supply decision.

Note the parallel to consumer theory: just as marginal utility diminishes for consumers, marginal product diminishes for producers, and both drive the key curves of the model. Recognizing this symmetry - diminishing marginal benefit on the demand side, diminishing marginal product on the supply side - is one of the unifying insights of microeconomics.

Common wrong turns

• Reading diminishing returns as falling output. Total product kept rising to 42; only the increments shrank.
• Blaming the workers. Every worker in the table is identical; the fixed capital is what thins out.
• Confusing marginal and average product. The third worker's MPL is 9 while APL at three workers is 11.00.
• Defining the short run in months. It is defined by having a fixed input, which is weeks for a food truck and years for a refinery.
• Mixing diminishing returns with diseconomies of scale. One holds an input fixed; the other scales everything up together.
• Forgetting that MC = wage/MPL. Marginal cost is not independent of productivity - it is productivity inverted and priced.

Recap

• The production function states what output is technologically possible from given inputs, before cost enters.
• The short run has at least one fixed input; in the long run every input, including plant size, can change.
• From total product 0, 10, 24, 33, 39, 42 the marginal products are 10, 14, 9, 6, 3 and average products 10.00, 12.00, 11.00, 9.75, 8.40.
• Marginal product rises at first through specialisation, then falls as workers share a fixed capital stock - diminishing marginal returns.
• Marginal product above average pulls it up and below pulls it down, so MPL crosses APL at APL's maximum of 12.00.
• At a $60 wage, MC = 60/MPL gave $6.00, $4.29, $6.67, $10.00, $20.00 and AVC = 60/APL bottomed at $5.00 where APL peaked.
• Diminishing returns is short-run and varies one input; returns to scale is long-run and varies all of them.

Sources

1. Greenlaw, S. A., Shapiro, D., & MacDonald, D. (2022). Production in the short run. In Principles of Microeconomics 3e. OpenStax, Rice University. openstax.org
2. Greenlaw, S. A., Shapiro, D., & MacDonald, D. (2022). Production in the long run. In Principles of Microeconomics 3e. OpenStax, Rice University. openstax.org
3. Greenlaw, S. A., Shapiro, D., & MacDonald, D. (2022). Costs in the short run. In Principles of Microeconomics 3e. OpenStax, Rice University. openstax.org
4. Munger, M. (n.d.). Division of labor. In The Concise Encyclopedia of Economics. Liberty Fund. econlib.org
5. Smith, A. (1776). An inquiry into the nature and causes of the wealth of nations. Project Gutenberg. gutenberg.org
6. Greenlaw, S. A., Shapiro, D., & MacDonald, D. (2022). The theory of labor markets. In Principles of Microeconomics 3e. OpenStax, Rice University. openstax.org
7. CORE Econ Team. (2023). The firm and its employees. In The Economy 2.0: Microeconomics. CORE Econ. core-econ.org

Ask a business owner what a unit costs and you will usually get one number. Ask an economist and you will get four, because "cost per unit" hides several different questions. What did the last unit cost? What does the average unit cost? How much of that is overhead that would have been paid anyway? Which of these should influence the decision to produce one more? The four curves in this lesson answer those questions separately, and keeping them apart is what makes the profit-maximising rules in the next module work.

The building blocks of cost

In the short run a firm's costs split into two parts. Fixed cost (FC) does not vary with output - rent on the factory, insurance, a loan payment - and must be paid even at zero output. Variable cost (VC) rises with output - labor, materials, power. Total cost is their sum: TC = FC + VC. Fixed cost exists only in the short run; in the long run every cost is variable because the firm can shed or expand any input, including the factory itself.

Distinguish fixed from sunk once more, because the two are confused constantly and the confusion is expensive. A fixed cost does not vary with output but may still be avoidable by a big enough decision - not renewing the lease, selling the machine. A sunk cost has already been spent and can be recovered by no decision at all. The custom-built oven that no other bakery would buy is sunk the moment it is installed; the lease payment due next year is fixed but avoidable. Fixed costs belong in the decision about whether to operate at all; sunk costs belong in no decision, which is why they never appear in the shutdown rule you will meet next.

Note too that these are economic costs, so they include the implicit ones. If the owner uses a building she already owns, the rent she could have charged is a real fixed cost of the business even though no cheque is written. Cost curves built from accounting records alone systematically understate cost and overstate profit.

Key idea: Fixed costs do not vary with output but can be avoidable, sunk costs are unrecoverable and irrelevant to any decision, and all of these are economic costs that include implicit opportunity costs.

Average and marginal cost

Dividing by output Q gives per-unit costs, and comparing successive totals gives marginal cost:

• Average fixed cost: AFC = FC / Q (always falls as Q rises, since fixed cost spreads over more units - firms call this "spreading the overhead").
• Average variable cost: AVC = VC / Q.
• Average total cost: ATC = TC / Q = AFC + AVC.
• Marginal cost: MC = change in TC / change in Q, the cost of producing one more unit. Because fixed cost does not change with output, MC also equals the change in variable cost - fixed cost never affects marginal cost.

Worked example: a cost schedule

Q	FC	VC	TC	MC	AFC	AVC	ATC
0	60	0	60	-	-	-	-
1	60	30	90	30	60	30	90
2	60	50	110	20	30	25	55
3	60	78	138	28	20	26	46
4	60	120	180	42	15	30	45
5	60	180	240	60	12	36	48

Each MC is the rise in TC from the previous row (for example, from Q=3 to Q=4, TC goes 138 to 180, so MC = 42). Each ATC is TC/Q (at Q=4, ATC = 180/4 = 45). Marginal cost first falls (30, then 20), then rises (28, 42, 60) - the mirror image of marginal product first rising then falling under diminishing returns. When MPL is rising, each unit is getting cheaper to make; once diminishing returns bite, MC turns upward.

Check the other columns the same way. AFC is 60 divided by Q, so it falls relentlessly: 60, 30, 20, 15, 12. AVC is VC divided by Q: at Q = 3 that is 78/3 = 26. And ATC is TC divided by Q, which must equal AFC + AVC - verify at Q = 3, where 20 + 26 = 46, and at Q = 5, where 12 + 36 = 48. That identity is worth memorising because it explains the shape of the graph: the vertical gap between the ATC and AVC curves is average fixed cost, so the two curves converge as output rises and overhead thins out. At Q = 1 they are $60 apart; at Q = 5 only $12.

Key idea: ATC = AFC + AVC, so the gap between the average total and average variable cost curves is average fixed cost, which shrinks toward zero as output grows.

The U-shaped average cost and the MC crossing

ATC is typically U-shaped. At low output, falling average fixed cost dominates and pulls ATC down; at high output, rising marginal cost from diminishing returns pushes ATC up. There is a crucial relationship between marginal and average cost:

• When MC is below ATC, ATC is falling (each cheaper unit pulls the average down).
• When MC is above ATC, ATC is rising (each pricier unit pulls the average up).
• Therefore MC crosses ATC at ATC's minimum.

In the table, ATC bottoms out near Q=4 (ATC = 45), which is exactly where MC (42) has caught up to and is about to exceed ATC. This "marginal pulls the average" logic is the same reason a student's next test score below their average lowers the average, and above it raises the average. The same relationship holds between MC and AVC: MC crosses AVC at AVC's minimum too, a point that becomes the firm's shutdown price in the next module.

Why the crossing must happen at the minimum

The crossing is not a drawing convention but an arithmetic necessity, because an average can only fall while the next item added lies below it and can only rise while the next item lies above it. An average that falls and then rises must therefore reach its lowest value exactly where the marginal item equals the average, since one instant earlier the marginal value was below and pulling downward while one instant later it is above and pulling upward. There is simply no other location at which an average can reverse direction.

Read it off the table. AVC runs 30, 25, 26, 30, 36, so it bottoms at $25 when Q = 2. The marginal cost of the second unit is $20, below that average, and the average duly fell from 30 to 25. The marginal cost of the third unit is $28, above the new average of 26, and the average duly rose. MC therefore crosses AVC between Q = 2 and Q = 3, right at the bottom of AVC. Now ATC: it runs 90, 55, 46, 45, 48, bottoming at $45 when Q = 4. MC at Q = 4 is $42, below ATC, so ATC was still falling; MC at Q = 5 is $60, well above ATC, so ATC rises. The crossing sits at the ATC minimum.

Two consequences follow immediately, and both matter in the next module. Because MC cuts AVC at its lowest point, the minimum of AVC is the lowest price at which producing beats shutting down - the shutdown price. And because MC cuts ATC at its lowest point, the minimum of ATC is the lowest price at which a firm can survive in the long run - the break-even price. Note also that MC crosses AVC (at Q = 2) before it crosses ATC (at Q = 4), because ATC carries falling average fixed cost that keeps pulling it down after AVC has already turned up.

Key idea: Marginal cost must intersect both average curves at their minimum points, because an average turns from falling to rising exactly where the marginal value equals it - and those two minima are the shutdown and break-even prices.

Economies of scale in the long run

In the long run all inputs vary and the firm chooses its scale. The long-run average cost curve is the lower envelope of all the short-run ATC curves the firm could build, and it can show economies of scale (average cost falls as output grows, from specialization, bulk buying, and spreading large fixed investments), constant returns (flat, where doubling all inputs doubles output), or diseconomies of scale (average cost rises as the firm becomes too large to coordinate and manage).

The output level where long-run average cost first stops falling is the minimum efficient scale, and it helps explain industry structure: where minimum efficient scale is large relative to the market, only a few firms survive.

Why it matters

Cost curves are the bridge from production technology to the supply decisions that fill the rest of the course. The marginal cost curve, as you will see next, is the competitive firm's supply curve above a certain point, and the shape of long-run average cost determines whether an industry tends toward many small firms or a few giants. Every pricing and output decision a firm makes traces back to the curves derived here.

Common wrong turns

• Letting fixed cost into marginal cost. MC is the change in total cost, and fixed cost does not change, so MC equals the change in variable cost.
• Computing MC as TC divided by Q. That is average total cost; marginal cost is the difference between successive totals.
• Expecting MC to cross the averages anywhere but their minima. The crossing point is forced by arithmetic, not chosen by the artist.
• Treating fixed and sunk as synonyms. A lease is fixed and avoidable; a custom installation is sunk and is not.
• Assuming average fixed cost eventually rises. AFC = FC/Q falls forever, approaching but never reaching zero.
• Building cost curves from accounting records alone. Omitting implicit costs understates cost and overstates profit.

Recap

• Total cost is fixed plus variable cost; fixed cost exists only in the short run, and sunk cost is a separate idea entirely.
• All of these are economic costs, so the forgone rent on an owner-occupied building counts even though no money moves.
• From the schedule, MC ran 30, 20, 28, 42, 60 - the mirror of marginal product rising then falling.
• AFC fell 60, 30, 20, 15, 12; AVC ran 30, 25, 26, 30, 36; and ATC = AFC + AVC held at every row.
• The gap between ATC and AVC is average fixed cost, so the curves converge as overhead spreads.
• AVC bottomed at $25 (Q = 2) and ATC at $45 (Q = 4), and marginal cost cut each at its minimum because that is where a falling average must turn.
• Those two minima become the shutdown price and the break-even price in the next module.

Sources

1. Greenlaw, S. A., Shapiro, D., & MacDonald, D. (2022). Costs in the short run. In Principles of Microeconomics 3e. OpenStax, Rice University. openstax.org
2. Greenlaw, S. A., Shapiro, D., & MacDonald, D. (2022). Costs in the long run. In Principles of Microeconomics 3e. OpenStax, Rice University. openstax.org
3. Greenlaw, S. A., Shapiro, D., & MacDonald, D. (2022). Explicit and implicit costs, and accounting and economic profit. In Principles of Microeconomics 3e. OpenStax, Rice University. openstax.org
4. Thurow, L. C. (n.d.). Profits. In The Concise Encyclopedia of Economics. Liberty Fund. econlib.org
5. Coase, R. H. (1991). The institutional structure of production [Prize lecture]. The Nobel Foundation. nobelprize.org
6. Greenlaw, S. A., Shapiro, D., & MacDonald, D. (2022). Regulating natural monopolies. In Principles of Microeconomics 3e. OpenStax, Rice University. openstax.org
7. CORE Econ Team. (2017). The firm and its customers. In The Economy 1.0. CORE Econ. core-econ.org

A wheat farmer with a thousand acres cannot raise her price by a cent. If she asks more than the going rate, buyers simply go next door, because her wheat is indistinguishable from everyone else's. She has no pricing decision at all - only a quantity decision. That extreme powerlessness sounds like a bad deal for the farmer, and in some ways it is, but it produces an outcome so efficient that economists use it as the standard against which every other market is judged. This lesson works out what such a firm does and why the result is the benchmark.

The benchmark model

Perfect competition is an idealized market with four features: (1) many small buyers and sellers, (2) a homogeneous (identical) product, (3) free entry and exit, and (4) perfect information. No single real market meets all four exactly, but many come close (agricultural commodities, foreign exchange), and the model is the yardstick against which every other market structure is judged.

Because each firm is tiny relative to the market, it is a price taker: it must accept the market price and cannot influence it. The firm's demand curve is therefore horizontal at the market price, which means price equals marginal revenue (P = MR): selling one more unit adds exactly the market price to revenue, because the firm can sell all it wants without moving the price.

Two of those four assumptions deserve flagging as simplifications rather than facts. Perfect information - every participant knowing prices and product quality - is never literally true, and Module 8 shows what asymmetric information does to a market. And a truly homogeneous product is rare outside commodities; most sellers differentiate, which is the subject of the next lesson. The model is a limiting case deliberately built to isolate what competition does when nothing obstructs it, which is exactly what makes it useful as a yardstick.

Key idea: A perfectly competitive firm is a price taker facing a horizontal demand curve, so price equals marginal revenue - and the model's strong assumptions make it a benchmark rather than a description.

Profit maximization: P = MC

Every firm, in any market structure, maximizes profit by producing where marginal revenue equals marginal cost. For a competitive firm, since P = MR, this becomes the famous rule P = MC. The logic is marginal: if P > MC, the next unit adds more to revenue than to cost, so produce it; if P < MC, the last unit lost money, so cut back. Profit is maximized where they are equal, and specifically on the rising portion of MC (on the falling portion, P = MC would be a profit minimum).

Worked example

A wheat farm's marginal cost is MC = 2q, and the market price is P = $40. Set P = MC:

40 = 2q, so q = 20 units

The firm should produce 20 units. Whether it earns a profit depends on average total cost at q = 20. If ATC there is $30, profit per unit is 40 - 30 = $10 and total profit is 10 x 20 = $200. If ATC there is $45, the firm makes a loss of $5 per unit, or $100 total, but may still produce in the short run (see the shutdown rule). The key discipline: find quantity from P = MC first, then read profit from ATC at that quantity - never the other way around.

Prove that q = 20 really is the maximum by computing profit directly at several quantities. With MC = 2q the variable cost is q squared, and taking fixed cost as $200 gives total cost of q squared plus 200 - which is exactly the cost structure that makes ATC equal $30 at q = 20.

Quantity	Total revenue (40 x q)	Total cost	Profit
10	$400	$300	$100
15	$600	$425	$175
18	$720	$524	$196
20	$800	$600	$200
22	$880	$684	$196
25	$1,000	$825	$175

Profit climbs to $200 at q = 20 and falls away on both sides - symmetrically, since the profit function is smooth around its peak. The algebra and the arithmetic agree, as they must. Notice why the peak is where it is: below 20 units, price of $40 exceeds marginal cost (at q = 18, MC = 2 x 18 = $36), so each extra unit adds $4 or more to profit; above 20, marginal cost exceeds $40 (at q = 22, MC = $44), so each extra unit subtracts. The peak is exactly where the two are equal.

Key idea: Producing where P = MC maximises profit because every unit with P above MC adds to profit and every unit with P below MC subtracts, so the total peaks where they meet.

The shutdown rule

In the short run a firm keeps producing as long as price covers average variable cost. If P < AVC, the firm cannot even cover its variable costs, so it does better to shut down and lose only its fixed cost. If P is at least AVC, producing helps pay down part of the unavoidable fixed cost even at a loss, so it is better to operate than to close.

The short-run supply curve of a competitive firm is therefore its marginal cost curve above the minimum of AVC. Below that price, the firm supplies zero. Fixed costs, being sunk in the short run, correctly play no role in this decision - only variable costs and revenue matter.

Worked example: to shut down or not

Suppose price falls to $18. A firm has AVC = $20 and ATC = $32 at its best output. Since P ($18) is below AVC ($20), each unit sold fails to cover even its variable cost, so producing loses more than the fixed cost alone. The firm should shut down in the short run, producing nothing and losing only its fixed cost. If instead price were $24 (above AVC $20 but below ATC $32), the firm would keep operating at a loss, because the $4 above variable cost on each unit helps offset the fixed cost it must pay regardless.

Test the rule against the cost schedule from the last lesson, where fixed cost was $60 and marginal costs ran 30, 20, 28, 42, 60 with average variable costs of 30, 25, 26, 30, 36. At a price of $28, setting P = MC gives 3 units. Revenue is 3 x 28 = $84 against variable cost of $78 and total cost of $138, so the firm loses 84 - 138 = -$54. Shutting down instead would mean losing the whole $60 of fixed cost. Operating at a loss of $54 beats closing at a loss of $60, and the shortcut agrees: P of $28 exceeds AVC of $26.

Now drop the price to $20. P = MC gives 2 units, revenue of $40 against variable cost of $50 and total cost of $110, so the loss is -$70 - worse than the $60 lost by shutting down. Again the shortcut agrees: $20 is below the AVC of $25 at that output. Producing here would mean paying $50 of wages and materials to generate $40 of revenue, digging the hole deeper than the fixed cost alone.

Key idea: Operate whenever price covers average variable cost, because the excess over variable cost pays down unavoidable fixed cost; shut down when it does not, because each unit then adds to the loss.

Long-run equilibrium and zero economic profit

Free entry and exit drive the long-run outcome. If firms earn positive economic profit, new firms enter, supply rises, and price falls. If firms suffer losses, some exit, supply falls, and price rises. This continues until price is driven to the minimum of average total cost and firms earn zero economic profit.

Zero economic profit does not mean zero accounting profit - it means firms earn exactly their opportunity cost, a normal return, with no incentive to enter or leave. At this point the competitive market is allocatively efficient: price equals marginal cost, so all mutually beneficial trades occur, and it is productively efficient: output is produced at the minimum of average total cost. This double efficiency is why perfect competition is the welfare benchmark for the rest of the module.

What zero economic profit actually means

This phrase misleads more students than any other in the course, so state it plainly: zero economic profit does not mean the firm makes no money. It means the firm earns exactly enough to cover every cost including the implicit ones - the salary the owner could have earned elsewhere, the return her capital could have made in another business. The owner is paid. The workers are paid. The lenders are paid. What is absent is a surplus above the next-best alternative, and that absence is precisely what removes the incentive for anyone new to enter.

Take the cost schedule again, where ATC bottomed at $45 with 4 units. In long-run equilibrium the market price is driven to exactly that $45. The firm produces 4 units, earns 4 x 45 = $180 in revenue against $180 in total cost, and books zero economic profit - while its accounting profit is comfortably positive, because the accountant never subtracted the owner's forgone salary. In this state the full chain P = MR = MC = minimum ATC holds, which is a compact way of saying the firm is simultaneously charging the socially correct price and producing at the lowest possible average cost.

One honest caveat: this is a long-run tendency in a model with free entry and identical firms, not a prediction that real industries settle there. Firms differ in cost, entry takes time and money, and technology keeps moving the target. Perfect competition describes a destination that real markets move toward without arriving.

Key idea: Zero economic profit means earning exactly the opportunity cost of every resource - a normal return, not a failure - and it occurs where P = MR = MC = minimum ATC.

Common wrong turns

• Reading zero economic profit as "no money." Everyone is paid; only the surplus above the next-best alternative is gone.
• Choosing quantity from ATC. Quantity comes from P = MC; ATC only tells you whether that quantity is profitable.
• Using ATC in the shutdown rule. The short-run test is against AVC, because fixed cost is paid either way.
• Shutting down whenever there is a loss. A firm losing $54 by producing but $60 by closing should keep producing.
• Solving P = MC on the falling part of MC. That intersection is a profit minimum; the maximum is on the rising portion.
• Treating the model as a description of real markets. Perfect information and homogeneous products are deliberate idealisations.

Recap

• Perfect competition assumes many small firms, a homogeneous product, free entry and exit, and perfect information - a benchmark, not a description.
• Each firm is a price taker facing horizontal demand, so P = MR.
• Profit is maximised where P = MC on the rising portion: with MC = 2q and P = $40, q = 20, confirmed by a profit peak of $200.
• Whether that quantity is profitable depends on ATC: $30 gives $200 of profit, $45 gives a $100 loss.
• The short-run supply curve is MC above minimum AVC; at P = $28 producing lost $54 versus $60 from closing, but at P = $20 producing lost $70.
• Free entry and exit drive price to minimum ATC - $45 in the running example - where economic profit is zero.
• Zero economic profit means a normal return covering all implicit costs, with P = MR = MC = minimum ATC and both allocative and productive efficiency.

Sources

1. Greenlaw, S. A., Shapiro, D., & MacDonald, D. (2022). Perfect competition and why it matters. In Principles of Microeconomics 3e. OpenStax, Rice University. openstax.org
2. Greenlaw, S. A., Shapiro, D., & MacDonald, D. (2022). How perfectly competitive firms make output decisions. In Principles of Microeconomics 3e. OpenStax, Rice University. openstax.org
3. Greenlaw, S. A., Shapiro, D., & MacDonald, D. (2022). Entry and exit decisions in the long run. In Principles of Microeconomics 3e. OpenStax, Rice University. openstax.org
4. Greenlaw, S. A., Shapiro, D., & MacDonald, D. (2022). Efficiency in perfectly competitive markets. In Principles of Microeconomics 3e. OpenStax, Rice University. openstax.org
5. Kasper, W. (n.d.). Competition. In The Concise Encyclopedia of Economics. Liberty Fund. econlib.org
6. Hayek, F. A. (1945). The use of knowledge in society. The American Economic Review, 35(4). Library of Economics and Liberty. econlib.org
7. Stigler, G. J. (1982). The process and progress of economics [Prize lecture]. The Nobel Foundation. nobelprize.org

The competitive farmer of the last lesson could not raise her price at all. Now consider the opposite extreme: the only water utility in a city, the holder of a drug patent, the operator of the single rail line into a port. Such a seller can choose its price - but not freely, because at a higher price it sells less. The whole theory of monopoly follows from working out what that trade-off implies, and the answer explains both why monopolies are profitable and why every industrial economy has laws against them.

A single seller

A monopoly is a market with one seller of a product that has no close substitutes, protected by barriers to entry. Unlike a price taker, a monopolist is a price maker: it faces the entire downward-sloping market demand curve and chooses the point on it that maximizes profit. Barriers to entry can arise from control of a key resource, government-granted rights such as patents or licenses, or a natural monopoly where a single large firm can supply the whole market at lower average cost than several small firms (as with water pipes or an electric grid, where duplicating the network would be wasteful).

Why marginal revenue is below price

Because the monopolist faces a downward-sloping demand curve, to sell one more unit it must lower the price - and it lowers the price on all units sold, not just the last one. So the marginal revenue from an extra unit is the new price minus the revenue lost by cutting price on the units it already sold.

This makes MR < P at every quantity beyond the first. For a linear demand P = a - bQ, marginal revenue is MR = a - 2bQ: same intercept, twice the slope. This single fact - that expanding output claws back revenue on existing sales - is the entire reason a monopolist restricts output and charges more than a competitive industry would.

Watch the two forces separate in a table. With demand P = 100 - 2Q:

Q	Price	Total revenue	Marginal revenue
1	$98	$98	$98
2	$96	$192	$94
3	$94	$282	$90
4	$92	$368	$86

Decompose the second unit. Selling it brings in its price of $96. But to sell two units the firm must charge $96 rather than $98, so it gives up $2 on the unit it was already selling. Net addition to revenue: 96 - 2 = $94, exactly the MR in the table. The third unit brings in $94 and costs $2 each on two earlier units, so 94 - 4 = $90. The fourth brings $92 and gives back $6, so $86. Price falls by $2 a step while marginal revenue falls by $4 - twice as fast - which is precisely what MR = 100 - 4Q says.

Key idea: Marginal revenue falls twice as fast as price on a linear demand curve because each extra unit both earns its own price and forces a price cut on every unit already being sold.

Worked example

Demand is P = 100 - 2Q and the firm's total cost is TC = 20Q + 100, so marginal cost is MC = 20. Find output, price, and profit.

1. Total revenue TR = P x Q = (100 - 2Q)Q = 100Q - 2Q^2.
2. Marginal revenue MR = 100 - 4Q.
3. Set MR = MC: 100 - 4Q = 20, so 4Q = 80 and Q = 20.
4. Price from demand: P = 100 - 2(20) = $60.
5. Profit = TR - TC = (60 x 20) - (20 x 20 + 100) = 1200 - 500 = $700.

Notice the key difference from competition: the monopolist finds Q where MR = MC, then charges the higher price the demand curve allows at that Q. Price ($60) exceeds marginal cost ($20). A competitive industry with the same cost would produce where P = MC, at 100 - 2Q = 20, giving Q = 40 - twice the monopoly output at a lower price.

Key idea: A monopolist finds quantity where MR = MC and then charges what the demand curve will bear at that quantity, so price ends up above marginal cost.

Worked example: measuring the damage

Put the inefficiency in dollars using the same numbers. Under competition the market would settle at P = MC = $20 with Q = 40. Under monopoly it settles at P = $60 with Q = 20. The 20 units between them are the trades that do not happen even though buyers value each of them above the $20 it costs to make.

The deadweight loss is the triangle over those missing units. Its base is 40 - 20 = 20 units and its height is the gap between price and marginal cost, 60 - 20 = $40, so the loss is 0.5 x 20 x 40 = $400.

Confirm it by adding up surplus both ways. Under competition, consumer surplus is 0.5 x 40 x (100 - 20) = $1,600 and producer surplus is zero, since marginal cost is constant at $20. Total surplus: $1,600. Under monopoly, consumer surplus shrinks to 0.5 x 20 x (100 - 60) = $400, while producer surplus becomes the rectangle (60 - 20) x 20 = $800. Total: $1,200. The difference is exactly $400.

Now separate the two effects, because they are morally different. Consumers lost 1,600 - 400 = $1,200. Of that, $800 was transferred to the monopolist as profit, and $400 was destroyed and reached nobody. Economists object to monopoly primarily on account of the $400. The $800 is a distributional matter, and reasonable people weigh it differently - though note that the prospect of capturing it is also what motivates the innovation that patents deliberately reward.

Key idea: Monopoly both transfers surplus from consumers to the firm and destroys some entirely; here $800 was transferred and $400 vanished as deadweight loss.

Inefficiency and deadweight loss

Because the monopolist sets P > MC, it produces less than the efficient quantity (where price would equal marginal cost). Some consumers who value the good above its marginal cost do not get it, so mutually beneficial trades are lost. That lost total surplus is deadweight loss - the efficiency cost of monopoly. The monopolist also captures surplus that would have gone to consumers, transferring it into profit.

This combination of a transfer (from consumers to the firm) plus a deadweight loss (destroyed for everyone) is why monopoly is generally considered inefficient and why governments use antitrust law, price regulation, and patent-term limits to check market power. Note that the transfer is not a social loss - it is a distributional change - whereas the deadweight loss is pure waste.

Price discrimination

A monopolist that can charge different prices to different buyers based on willingness to pay is engaging in price discrimination (for example, student and senior discounts, airline fares that depend on booking timing, or software priced differently for firms and individuals). It requires some market power, the ability to segment buyers by willingness to pay, and prevention of resale (so cheap buyers cannot resell to expensive ones).

Price discrimination lets the firm capture more surplus and can actually raise output relative to single pricing, because the firm will serve lower-value customers it would otherwise ignore. Under perfect (first-degree) price discrimination, output reaches the efficient level and deadweight loss vanishes, but all the surplus goes to the firm - efficient yet highly unequal.

Key idea: Price discrimination requires market power, separable buyers, and no resale, and it can raise output while shifting surplus from consumers to the firm.

Policy responses to market power

Because monopoly power is a diagnosable, measurable problem, governments respond to it with specific tools. Antitrust law is the main instrument in the United States: the Sherman and Clayton Acts, enforced by the Department of Justice's Antitrust Division and the Federal Trade Commission, prohibit monopolisation and block mergers likely to reduce competition substantially. The FTC is careful to note that simply being large or successful is not illegal - what the law targets is acquiring or maintaining monopoly power through exclusionary conduct rather than through a better product.

Natural monopolies need different treatment, because breaking one into competing firms would duplicate an expensive network and raise average costs for everyone. Here regulators typically permit the single firm and constrain its pricing instead, often toward average cost so the firm covers its costs without extracting monopoly profit. And where the monopoly was deliberately created - a patent - the policy is a time limit: exclusive rights for a fixed term to reward the invention, followed by open competition. Each remedy trades some deadweight loss against some other cost, which is why none of them is applied universally.

Key idea: Antitrust targets conduct that creates or preserves monopoly power, natural monopolies are usually regulated rather than broken up, and patents grant temporary monopoly on purpose to reward innovation.

Common wrong turns

• Setting price where MR = MC. That intersection gives the quantity; the price is read up on the demand curve, here $60 rather than $20.
• Thinking a monopolist can charge anything. It is bound by demand: a higher price always means fewer units.
• Treating the whole consumer loss as deadweight loss. Of the $1,200 consumers lost, $800 was a transfer and only $400 was destroyed.
• Assuming monopolies always earn profits. A monopolist whose ATC lies above demand at every quantity loses money and exits.
• Expecting production on the inelastic part of demand. There MR is negative, so a firm with positive marginal cost never operates there.
• Calling all price discrimination harmful. It can raise output and serve buyers who would otherwise be priced out entirely.

Recap

• A monopoly is a single seller protected by barriers to entry, and it is a price maker facing the whole market demand curve.
• Marginal revenue lies below price because selling one more unit forces a price cut on all earlier units: $96 earned minus $2 given back equals MR of $94.
• For linear demand P = a - bQ, marginal revenue is a - 2bQ - the same intercept and twice the slope.
• With P = 100 - 2Q and MC = 20, MR = MC gives Q = 20, the demand curve gives P = $60, and profit is $700.
• A competitive industry with the same costs would produce 40 units at $20, so monopoly halves output and triples price here.
• Deadweight loss is 0.5 x 20 x 40 = $400, alongside an $800 transfer from consumers to the firm.
• Policy responds with antitrust enforcement, price regulation of natural monopolies, and time-limited patents.

Sources

1. Greenlaw, S. A., Shapiro, D., & MacDonald, D. (2022). How monopolies form: Barriers to entry. In Principles of Microeconomics 3e. OpenStax, Rice University. openstax.org
2. Greenlaw, S. A., Shapiro, D., & MacDonald, D. (2022). How a profit-maximizing monopoly chooses output and price. In Principles of Microeconomics 3e. OpenStax, Rice University. openstax.org
3. Greenlaw, S. A., Shapiro, D., & MacDonald, D. (2022). Regulating natural monopolies. In Principles of Microeconomics 3e. OpenStax, Rice University. openstax.org
4. Stigler, G. J. (n.d.). Monopoly. In The Concise Encyclopedia of Economics. Liberty Fund. econlib.org
5. Federal Trade Commission. (n.d.). Monopolization defined. In Guide to Antitrust Laws. ftc.gov
6. McChesney, F. S. (n.d.). Antitrust. In The Concise Encyclopedia of Economics. Liberty Fund. econlib.org
7. Tirole, J. (2014). Market failures and public policy [Prize lecture]. The Nobel Foundation. nobelprize.org

Neither of the last two lessons describes the market you actually shop in. The coffee shop on your corner is not a price taker selling an identical commodity, and it is not the only coffee shop in the world either. It has a little pricing power and a lot of rivals. Meanwhile the airline you fly has perhaps three serious competitors, each watching the others' fares hour by hour. Most real markets live in this middle ground, and analysing them requires one genuinely new tool: a way to reason about decisions that depend on what someone else decides.

The middle ground

Most real markets lie between perfect competition and monopoly. Two important structures fill this middle.

Monopolistic competition has many firms selling differentiated products - think restaurants, clothing brands, hair salons, or coffee shops. Product differentiation gives each firm a little pricing power (a downward-sloping demand curve), but free entry means economic profits are competed away in the long run, just as under perfect competition. Firms spend heavily on branding, advertising, and variety, and they typically produce with some excess capacity, charging a price above marginal cost. The result is a familiar trade-off: consumers pay a bit more than the competitive ideal but enjoy genuine variety and choice.

Make "excess capacity" concrete. Suppose a coffee shop in long-run equilibrium faces demand P = 10 - 0.05Q, so its marginal revenue is MR = 10 - 0.1Q. Its marginal cost is $4. Setting MR = MC gives 10 - 0.1Q = 4, so Q = 60 cups a day, and the demand curve gives P = 10 - 0.05(60) = $7. Free entry has competed profits away, which means average total cost at 60 cups is also $7 - the demand curve is tangent to ATC there. But this shop's ATC would bottom out at, say, $6 if it sold 100 cups. It is therefore operating 40 cups below its efficient scale, with a price of $7 sitting $3 above marginal cost.

That gap is the standing indictment and the standing defence of monopolistic competition in one number. The indictment: price exceeds marginal cost, so some customers who value a cup above $4 do not get one, and every shop runs below the output that would minimise its average cost. The defence: what buyers get in exchange is a coffee shop on their own street, in their preferred style, rather than one enormous efficient plant serving a standardised product. Whether the variety is worth the markup is not a question the model answers.

Key idea: In long-run monopolistic competition entry drives economic profit to zero so price equals average cost, but price still exceeds marginal cost and firms produce below minimum-average-cost scale - the excess-capacity result.

Oligopoly has a few large firms that dominate a market - think airlines, telecom carriers, or game-console makers. The defining feature is interdependence: each firm's best action depends on what its rivals do, so firms must think strategically rather than simply reading a price off a curve. Oligopolists may compete fiercely on price and features, or may try to collude (openly or tacitly) to act like a monopoly and raise prices, though such agreements are unstable and, when explicit, usually illegal under antitrust law.

How concentrated is an oligopoly? Competition authorities put a number on it with the Herfindahl-Hirschman Index, the sum of the squared market shares of every firm. A market split 30, 30, 20, 20 gives 900 + 900 + 400 + 400 = 2,600. Four equal firms at 25 percent each would give 4 x 625 = 2,500, and a monopoly gives 100 squared = 10,000. Agencies use index levels and the change a merger would cause to decide which deals warrant investigation, which is why a proposed merger's effect on the HHI is often the first thing reported about it.

Key idea: Oligopoly is defined by strategic interdependence rather than by a firm count, and concentration is measured with the Herfindahl-Hirschman Index, the sum of squared market shares.

Game theory: modeling strategy

Game theory is the study of strategic interaction, where each player's payoff depends on the choices of all. A key solution concept is the Nash equilibrium: a set of strategies where no player can do better by unilaterally changing their own choice, given what everyone else is doing. It captures a stable outcome from which no one wants to deviate. A related idea is a dominant strategy, a choice that is best for a player no matter what rivals do; when every player has one, the dominant-strategy outcome is automatically a Nash equilibrium.

The prisoner's dilemma

The most famous game explains why cooperation is hard. Two firms each choose to keep prices High or cut to Low. Each cell shows (Firm A profit, Firm B profit) in millions:

	B: High	B: Low
A: High	(10, 10)	(2, 12)
A: Low	(12, 2)	(5, 5)

Analyze A's best response. If B plays High, A earns 10 by matching High but 12 by playing Low - so A prefers Low. If B plays Low, A earns 2 by High but 5 by Low - so A prefers Low again. Cutting price is a dominant strategy for A: it is best no matter what B does. By symmetry, Low is also dominant for B. So both choose Low and earn (5, 5) - even though both would be better off at (10, 10) if they could cooperate. The (Low, Low) cell is the Nash equilibrium: given the other is playing Low, neither can gain by switching to High alone.

Worked reasoning: verifying the equilibrium

To confirm (Low, Low) is a Nash equilibrium, check each player's incentive to deviate. From (5, 5), if A switches to High while B stays Low, A moves to the (High, Low) cell and earns 2 - worse than 5, so A will not deviate. The same holds for B by symmetry. Now check the tempting (High, High) cell paying (10, 10): from there A can deviate to Low and jump to 12, so (High, High) is not stable - it fails the no-deviation test. Only (Low, Low) survives, which is exactly why the dilemma is a dilemma: the stable outcome is the jointly worse one.

Key idea: A Nash equilibrium is a profile of strategies from which no player can gain by changing alone, and in the prisoner's dilemma it is the outcome both players like less.

Worked example: a game with no dominant strategy

The prisoner's dilemma is famous but atypical, because both players have a dominant strategy. Most strategic problems do not. Suppose two firms must each pick a technical standard, A or B. Customers want compatibility, so mismatching is a disaster for both. Payoffs in millions, (Firm 1, Firm 2):

	Firm 2: A	Firm 2: B
Firm 1: A	(8, 6)	(0, 0)
Firm 1: B	(0, 0)	(5, 9)

Look for a dominant strategy for Firm 1. If Firm 2 plays A, Firm 1 earns 8 from A and 0 from B, so it prefers A. If Firm 2 plays B, Firm 1 earns 0 from A and 5 from B, so it prefers B. Firm 1's best move depends on Firm 2's - it has no dominant strategy, and by the same reasoning neither does Firm 2.

Now test each cell for the no-deviation property. At (A, A), Firm 1 switching to B falls from 8 to 0 and Firm 2 switching falls from 6 to 0, so neither moves: this is a Nash equilibrium. At (B, B), Firm 1 switching falls from 5 to 0 and Firm 2 from 9 to 0, so this is a Nash equilibrium too. The mismatched cells are not: at (A, B) Firm 1 earns 0 and could get 5 by switching to B.

So the game has two Nash equilibria, and the players disagree about which they prefer - Firm 1 wants (A, A), Firm 2 wants (B, B). Nothing in the concept of Nash equilibrium picks between them. That is not a defect of the tool but a genuine feature of standard-setting fights, and it explains why firms invest so heavily in moving first, announcing early, and building industry consortia: the aim is to make one equilibrium the obvious one before rivals can establish the other.

Key idea: Every dominant-strategy outcome is a Nash equilibrium, but not every Nash equilibrium comes from dominant strategies - and a game can have several, with players preferring different ones.

Why it matters

The prisoner's dilemma explains why cartels break down (each member is tempted to secretly undercut the agreed price), why price wars erupt, and why firms crave enforceable contracts and industry standards. It also shows that individually rational choices can produce a collectively worse outcome - a theme that recurs directly in the study of externalities and common resources in the next module, where each actor's self-interest degrades a shared outcome.

Crucially, repeated interaction changes the game: when firms expect to meet again and again, strategies like "cooperate until the other cheats, then punish" can sustain the cooperative (High, High) outcome that a one-shot game destroys. This is why long-lived rivals sometimes maintain high prices without any formal agreement, and why the number of interactions, not just the payoffs, shapes real-world behavior.

Antitrust law exists partly because firms know all of this. Because a cartel's members always have the (Low, Low) temptation, they look for ways to make cooperation stable - agreeing prices, dividing territories, or rigging bids. The Department of Justice treats price fixing, bid rigging, and market allocation as per se illegal, meaning no efficiency justification is entertained, and prosecutes them criminally. Read alongside the payoff matrix, the law is doing something specific: it removes the enforcement mechanisms that would let firms escape the dilemma, deliberately keeping them trapped in the outcome that happens to be good for consumers.

Key idea: Repetition and enforceable agreements can sustain cooperation, which is precisely why antitrust law treats price fixing and bid rigging as criminal offences rather than ordinary contracts.

Common wrong turns

• Equating Nash equilibrium with the best joint outcome. In the prisoner's dilemma the equilibrium pays (5, 5) while both would prefer (10, 10).
• Assuming every game has a dominant strategy. In the standards game each firm's best move depends on the other's.
• Assuming a unique equilibrium. The standards game has two, and the players rank them differently.
• Counting firms to identify oligopoly. The defining feature is interdependence - each firm's payoff depending on rivals' choices.
• Expecting monopolistic competitors to earn long-run profits. Free entry drives economic profit to zero, just as under perfect competition.
• Reading excess capacity as mismanagement. It is the structural price of product variety, not a failure of the individual firm.

Recap

• Monopolistic competition has many firms, differentiated products, and free entry, so long-run economic profit is zero.
• The coffee shop produced 60 cups where MR = MC = $4 and charged $7, sitting 40 cups short of its $6 minimum-cost scale - excess capacity.
• Oligopoly is defined by interdependence, and concentration is measured by the HHI: shares of 30, 30, 20, 20 give 2,600.
• A Nash equilibrium is a strategy profile from which no player gains by deviating alone.
• In the pricing dilemma, Low is dominant for both, giving the Nash outcome (5, 5) even though (10, 10) was available.
• The standards game has no dominant strategy and two Nash equilibria, (A, A) and (B, B), which the two firms rank oppositely.
• Repeated play can sustain cooperation, which is why price fixing and bid rigging are prosecuted as criminal offences.

Sources

1. Greenlaw, S. A., Shapiro, D., & MacDonald, D. (2022). Monopolistic competition. In Principles of Microeconomics 3e. OpenStax, Rice University. openstax.org
2. Greenlaw, S. A., Shapiro, D., & MacDonald, D. (2022). Oligopoly. In Principles of Microeconomics 3e. OpenStax, Rice University. openstax.org
3. Dixit, A., & Nalebuff, B. (n.d.). Game theory. In The Concise Encyclopedia of Economics. Liberty Fund. econlib.org
4. Dick, A. R. (n.d.). Cartels. In The Concise Encyclopedia of Economics. Liberty Fund. econlib.org
5. The Royal Swedish Academy of Sciences. (1994). The Sveriges Riksbank Prize in Economic Sciences in Memory of Alfred Nobel 1994: Nash, Harsanyi and Selten [Press release]. The Nobel Foundation. nobelprize.org
6. U.S. Department of Justice, Antitrust Division. (n.d.). Price fixing, bid rigging, and market allocation schemes: What they are and what to look for. justice.gov
7. Federal Trade Commission. (n.d.). Mergers. In Guide to Antitrust Laws. ftc.gov

Why does a radiologist earn more than a childcare worker, when almost everyone would agree the second job matters at least as much? Why did typesetters' wages collapse within a decade while plumbers' held? Questions about pay feel like questions about justice, and partly they are. But before you can argue about whether a wage is fair, it helps to understand what actually sets it - and the answer turns out to be the same supply-and-demand machinery you already own, pointed at a market for people's time.

Factor markets

So far we studied markets for goods. Factor markets are the markets for the inputs to production - labor, land, capital, and entrepreneurship. Their prices are wages, rent, interest, and profit, and together they determine how the economy's income is divided. The same supply-and-demand tools apply, with one crucial twist: the demand for a factor is a derived demand - it comes from the demand for the goods the factor helps produce. A firm hires workers not for their own sake but because their output can be sold. When demand for a product collapses, so does demand for the workers who make it, no matter how skilled they are.

The marginal revenue product of labor

How much is one more worker worth to a firm? The answer is the marginal revenue product of labor (MRPL): the extra revenue generated by hiring one more worker. It equals the worker's marginal product times the marginal revenue from selling that output:

MRPL = MPL x MR (and in a competitive product market, MR = P, so MRPL = MPL x P)

This formula fuses two earlier ideas: the marginal product from Module 5 (how much extra output a worker makes) and marginal revenue from Module 3 (how much that output sells for). A worker is valuable only if they are both productive and making something the market values.

Key idea: Labor demand is derived from product demand, and a worker's value to a firm is marginal product multiplied by what that output sells for.

The hiring rule

A profit-maximizing firm hires more workers as long as each adds at least as much to revenue as to cost. In a competitive labor market the cost of one more worker is the wage (W). So the firm hires up to the point where:

MRPL = W

This is the labor-market twin of the P = MC rule for goods: keep doing the activity until the marginal benefit equals the marginal cost. Because marginal product falls with more workers (diminishing returns), MRPL slopes downward - and the downward-sloping MRPL curve is the firm's labor demand curve. If the wage is below MRPL, hire more; if above, hire fewer; equality is the optimum.

Worked example

A worker's marginal product is 6 units and the firm sells each unit at $10 in a competitive market. Then MRPL = 6 x $10 = $60. If the market wage is $60, hiring this worker exactly breaks even at the margin; if the wage is $45, this worker adds $60 in revenue for $45 in cost, so hiring is profitable and the firm should hire more; if the wage is $70, this worker costs more than they bring in, so the firm should not hire them.

Suppose the product price now rises to $15: MRPL jumps to 6 x $15 = $90, and the firm that stopped hiring at a $70 wage would now happily hire this worker, illustrating how a rise in output price pulls up labor demand.

Worked example: how many workers to hire

Take the production schedule from Module 5, where marginal products were 10, 14, 9, 6, and 3 for the first through fifth worker. Sell the output at $10 a unit in a competitive market, so MRPL = MPL x $10.

Worker	Marginal product	MRPL at P = $10	Hire at a wage of $60?	Hire at a wage of $80?
1st	10	$100	Yes	Yes
2nd	14	$140	Yes	Yes
3rd	9	$90	Yes	Yes
4th	6	$60	Breaks even	No
5th	3	$30	No	No

At a wage of $60 the firm hires through the fourth worker, whose $60 of added revenue exactly matches her $60 cost - the firm is genuinely indifferent about that last hire. Confirm it: three workers produce 33 units for $330 of revenue against $180 of wages, leaving $150; four workers produce 39 units for $390 against $240, also leaving $150. Five workers give 420 - 300 = $120, which is worse.

Raise the wage to $80 and the fourth worker no longer pays for herself, so employment falls to 3. Check: three workers leave 330 - 240 = $90, four leave 390 - 320 = $70, two leave 240 - 160 = $80. Three is best. A higher wage bought less labor - the labor demand curve slopes downward, and it does so because marginal product diminishes.

Now hold the wage at $80 and let the product price rise to $15. Every MRPL rescales to 150, 210, 135, 90, and 45, so the fourth worker is now worth $90 against her $80 cost and employment climbs back to 4. Nothing about the workers changed; the value of what they make did. This is derived demand in one move, and it is why a collapse in an industry's product price shows up as layoffs among people whose skills are exactly what they were the week before.

Key idea: A firm hires until MRPL equals the wage, so a higher wage reduces employment along a fixed labor demand curve while a higher output price shifts that whole curve to the right.

When one employer dominates: monopsony

The MRPL = wage rule assumes the firm is a wage taker, able to hire as many workers as it wants at the going rate. That fails where one employer dominates local hiring - a monopsony, such as a single large plant in a small town or a hospital system in a rural county. Such an employer faces an upward-sloping labor supply curve: to attract one more worker it must raise the wage, and if it pays everyone the same rate, it must raise it for all existing staff too. The marginal cost of labor therefore exceeds the wage, exactly mirroring how a monopolist's marginal revenue falls below price.

The consequence is that a monopsonist hires where MRPL equals that higher marginal cost of labor, which means fewer workers at a lower wage than a competitive labor market would deliver. This is the case that overturns the simple minimum-wage prediction from Module 2: a floor set above the monopsony wage but below the competitive one can raise wages and employment at the same time. It is also part of why the empirical minimum-wage literature, from Card and Krueger onward, finds employment effects far smaller than the competitive model predicts.

Key idea: Where employers have wage-setting power the marginal cost of labor exceeds the wage, producing lower pay and employment than competition - and making a moderate minimum wage potentially beneficial to both.

What shifts labor demand and supply

• Labor demand shifts with the price of the output (higher output price raises MRPL), with worker productivity (better technology, tools, or training raises MPL), and with the prices of other inputs that are substitutes or complements for labor.
• Labor supply shifts with the size and skills of the workforce, the appeal of the job (wages, conditions, prestige), the value of leisure and non-work options, and immigration or demographic change.

The intersection of labor supply and labor demand sets the equilibrium wage and employment. Wage differences across jobs then reflect differences in productivity (MRPL), skills and human capital, working conditions (compensating differentials - dangerous or unpleasant jobs must pay more to attract workers), and barriers or bargaining power in particular markets, such as unions, licensing, or discrimination.

Key idea: Equilibrium wages reflect productivity, the value of output, working conditions, and bargaining power, so wage gaps have several distinct sources that policy must tell apart.

Reading the theory honestly

Marginal productivity theory explains a great deal, and it is important to be clear about what it does not claim. It is a positive account of what firms will pay, not a normative claim that people deserve what they earn. A childcare worker's marginal revenue product is low largely because the families who value the work most cannot pay much for it, not because the work is unimportant - the theory prices the output, not the moral worth.

Nor does the theory require that markets be frictionless. Measuring an individual's marginal product is often impossible in team production. Search takes time, information about outside offers is poor, and workers rarely move costlessly between employers - all of which give firms some wage-setting latitude even outside pure monopsony. Discrimination persists in ways the simple model would predict away. And Gary Becker's human-capital framework adds an important layer: much of the wage distribution reflects accumulated investment in education, training, and experience, which is why the returns to schooling are among the most studied quantities in economics. The framework is a starting point for diagnosis, and each of its failures points to a specific policy question.

Key idea: Marginal productivity theory describes what firms will pay rather than what work is worth, and real labor markets add search frictions, imperfect information, and discrimination that the bare model omits.

Why it matters

This module reframes the entire supply-and-demand apparatus to answer a question people care about intensely: why do incomes differ? The MRPL framework says wages ultimately track the value a worker adds, which is why investments in education and skills (human capital) tend to raise pay, and why automation that raises some workers' productivity while replacing others reshapes the wage distribution. It also clarifies debates over minimum wages, immigration, and unions by locating each as a shift or constraint in a specific labor market, connecting the abstract model directly to questions of fairness and policy.

Common wrong turns

• Assuming labor demand depends only on skill. It is skill multiplied by the value of the output; a collapsing product price cuts MRPL directly.
• Using marginal product instead of marginal revenue product. Units of output must be converted into dollars before comparing with a wage.
• Hiring while MRPL exceeds zero. The rule is MRPL at least equal to the wage, not merely positive.
• Applying MRPL = wage under monopsony. There the relevant cost is the marginal cost of labor, which exceeds the wage.
• Reading the theory as a claim about desert. It predicts what firms will pay, not what a job is worth.
• Confusing a compensating differential with a productivity gap. Extra pay for unpleasant or risky work is not extra output.

Recap

• Factor markets price inputs, and labor demand is derived from demand for the product labor makes.
• MRPL = MPL x MR, which in a competitive product market is MPL x price.
• With marginal products 10, 14, 9, 6, 3 and a $10 price, MRPL ran $100, $140, $90, $60, $30.
• At a $60 wage the firm hired 4 workers (the fourth exactly breaking even); at $80 it hired 3, tracing a downward-sloping labor demand curve.
• Raising the product price to $15 with the wage still $80 restored the fourth hire - a rightward shift in labor demand.
• Under monopsony the marginal cost of labor exceeds the wage, so pay and employment fall below competitive levels and a moderate wage floor can raise both.
• Wage differences reflect productivity, human capital, compensating differentials, and bargaining power - and the theory is positive, not a statement about desert.

Sources

1. Greenlaw, S. A., Shapiro, D., & MacDonald, D. (2022). The theory of labor markets. In Principles of Microeconomics 3e. OpenStax, Rice University. openstax.org
2. Greenlaw, S. A., Shapiro, D., & MacDonald, D. (2022). Wages and employment in an imperfectly competitive labor market. In Principles of Microeconomics 3e. OpenStax, Rice University. openstax.org
3. Greenlaw, S. A., Shapiro, D., & MacDonald, D. (2022). Market power on the supply side of labor markets: Unions. In Principles of Microeconomics 3e. OpenStax, Rice University. openstax.org
4. Greenlaw, S. A., Shapiro, D., & MacDonald, D. (2022). Employment discrimination. In Principles of Microeconomics 3e. OpenStax, Rice University. openstax.org
5. Becker, G. S. (n.d.). Human capital. In The Concise Encyclopedia of Economics. Liberty Fund. econlib.org
6. Becker, G. S. (1992). The economic way of looking at life [Prize lecture]. The Nobel Foundation. nobelprize.org
7. The Nobel Foundation. (2021). David Card - Facts. NobelPrize.org ↗. nobelprize.org

Everything the course has claimed about the efficiency of markets rested on a quiet assumption: that the people making a deal are the only people affected by it. Drop that assumption and the argument collapses. A factory that pays for its coal and its labor but not for the asthma downwind is not facing the full cost of what it does, and a market where costs go unpaid will produce too much of whatever generates them. This is the most consequential market failure in the book, and the framework for fixing it is the same marginal reasoning you have used since the first lesson.

When private and social costs diverge

A market is efficient when those who make a decision bear all its costs and reap all its benefits, so that private incentives line up with social welfare. An externality occurs when a transaction imposes a cost or confers a benefit on a third party not involved in the deal. Because the decision-maker ignores these spillover effects, the market outcome is inefficient - it violates the very assumption (all costs and benefits fall on the transacting parties) that made competition efficient in Module 6. Externalities are among the most important and common sources of market failure, at the heart of debates over pollution, vaccines, education, and climate.

Negative externalities

A negative externality imposes an uncompensated cost on others - factory pollution harming nearby residents, or traffic congestion where each driver slows everyone else. Here the social cost (private cost plus external cost) exceeds the private cost the producer considers. Because the firm looks only at its private cost, the market overproduces relative to the efficient level, and the value of the excess harm is a deadweight loss. The efficient quantity is where social marginal benefit equals social marginal cost, which is less than the free-market quantity. The gap between the market quantity and the efficient quantity is the problem policy tries to close.

Key idea: A negative externality makes social cost exceed private cost, so the free market produces more than the efficient quantity and the excess destroys surplus.

Positive externalities

A positive externality confers an uncompensated benefit on others - vaccination protecting the wider community by reducing disease spread, or education raising civic participation and the productivity of coworkers. Now the social benefit exceeds the private benefit the buyer considers, so the market underproduces relative to the efficient level. Too little of a good thing is also inefficient, and the remedy is to encourage more of the activity rather than less.

Key idea: A positive externality makes social benefit exceed private benefit, so the market underproduces and the efficient policy encourages rather than restricts the activity.

Policy remedies

The general fix is to make decision-makers face the full social cost or benefit - to internalize the externality:

• A corrective (Pigouvian) tax on a negative externality raises the private cost to match the social cost, cutting output toward the efficient level. A tax per ton of emissions is the classic example, and the ideal tax equals the external cost at the efficient quantity.
• A subsidy for a positive externality lowers the private cost (or raises the private benefit) and encourages more of the activity, as with subsidized vaccines, tuition support, or basic research grants.
• Cap-and-trade sets a total quantity of allowable pollution and lets firms buy and sell permits, achieving a target at least cost because firms that can cut cheaply do so and sell their permits to firms that cannot.
• Assigning property rights can let private parties bargain to an efficient outcome. The Coase theorem holds that if property rights are clear and bargaining is costless, private parties can negotiate an efficient result regardless of who holds the right - though in practice high transaction costs, many affected parties, and holdout problems often block this.

Worked reasoning

Suppose a chemical plant's private marginal cost of a unit is $8 but it also imposes $4 of pollution damage on neighbors, so the social marginal cost is $12. If the good sells for $10, the plant produces because $10 > $8 privately - yet from society's view the last unit costs $12 to make only $10 of value, a $2 loss on each such unit.

A $4 corrective tax raises the plant's cost to $12, so it stops producing the units whose social cost exceeds their value, restoring efficiency. Notice the tax does not aim to eliminate all pollution - it aims to cut it to the point where the marginal benefit of production just equals the true marginal social cost, which is the efficient, not the zero, level.

Worked example: finding the efficient quantity and the right tax

Now do it with curves rather than a single unit. Let the marginal benefit to buyers be MB = 40 - 0.5Q and the plant's private marginal cost be PMC = 8 + 0.5Q. Each unit imposes $4 of pollution damage on neighbours, so social marginal cost is SMC = 12 + 0.5Q.

The unregulated market equates marginal benefit with private cost: 40 - 0.5Q = 8 + 0.5Q, so Q = 32 and the price is 40 - 16 = $24. The efficient outcome equates marginal benefit with social cost: 40 - 0.5Q = 12 + 0.5Q, so Q = 28. The market therefore overproduces by 4 units.

Measure the damage those 4 units do. At Q = 32 the social cost of the last unit is 12 + 16 = $28 while its benefit is only $24, a gap of $4; at Q = 28 the gap has closed to zero, since 12 + 14 = 26 equals 40 - 14 = 26. The deadweight loss is the triangle 0.5 x 4 x 4 = $8.

Finally, set the tax. A per-unit tax of $4 - exactly the external damage - raises the firm's private marginal cost to 12 + 0.5Q, which is the social marginal cost. The firm, still maximising its own profit and thinking about nobody else, now produces 28 units on its own. Tax revenue is 4 x 28 = $112. Notice what the tax did and did not do: it did not order anyone to pollute less, and it did not abolish pollution. It changed one number in the firm's own calculation so that self-interest and social interest pointed the same way.

Key idea: The efficient quantity equates marginal benefit with social marginal cost, and a Pigouvian tax equal to the external damage per unit moves the market to it without anyone needing to change their motives.

Worked example: why cap-and-trade is cheap

Suppose a regulator needs total emissions cut by 100 tons across two plants. Plant A can abate at $20 a ton; Plant B, with older equipment, needs $50 a ton.

A simple command-and-control rule splitting the burden equally makes each cut 50 tons. Cost: (50 x 20) + (50 x 50) = 1,000 + 2,500 = $3,500. Now issue tradable permits for the same total. Plant A abates all 100 tons at a cost of 100 x 20 = $2,000 and sells its unused permits to Plant B at any price between $20 and $50 - above A's cost so A profits, below B's cost so B saves. The same environmental result costs $1,500 less, a saving of over 40 percent, and the permit price falls between the two abatement costs exactly as the trade price fell between the two opportunity costs back in Module 1.

That is the entire case for market-based environmental policy: a regulator does not know which firms can cut cheaply, but the firms do, and a permit market extracts that knowledge without anyone reporting it. The United States acid rain program, which cut sulphur dioxide emissions from power plants using tradable allowances, is the standard demonstration that this works at scale.

Key idea: Tradable permits achieve a given emissions target at lower total cost than uniform mandates, because abatement concentrates where it is cheapest without the regulator needing to know where that is.

Why it matters

Externalities are the economic core of environmental policy and public health. The same logic that says a carbon tax can align private incentives with the planet's welfare also says subsidizing vaccination or education raises total surplus. The framework does not tell you the externality is "bad" and must be banned; it tells you to find the quantity where social marginal benefit meets social marginal cost, and to choose the least-cost instrument to get there. That disciplined, marginal way of thinking about spillovers is one of economics' most influential contributions to public debate.

Two honest limits belong with the enthusiasm. First, the framework requires a number for the external damage, and estimating that number - the health cost of a microgram of particulates, the damage from a ton of carbon - is genuinely hard and contested, which is where much of the real policy argument happens. William Nordhaus's work on integrated climate-economy models exists precisely because putting a defensible figure on carbon damage is the binding constraint on carbon policy. Second, the Coase theorem's promise of private bargaining rarely survives contact with large numbers: two neighbours can negotiate over a noisy generator, but eight billion people cannot negotiate over the atmosphere, which is why climate policy needs institutions rather than contracts.

Key idea: The externality framework is only as good as the damage estimate it uses, and private bargaining scales badly, so real remedies depend on measurement and on institutions rather than on contracts alone.

Common wrong turns

• Aiming for zero pollution. The efficient level equates marginal benefit with social marginal cost; here that meant 28 units, not zero.
• Setting the tax equal to total damage. A Pigouvian tax is per unit and equals the external cost of one unit - $4 here, not $112.
• Treating positive externalities as harmless. Underproduction is inefficient too, and the remedy is a subsidy rather than a tax.
• Assuming the Coase theorem makes policy unnecessary. It requires clear rights and near-zero transaction costs, which large-numbers problems never satisfy.
• Confusing the tax revenue with the efficiency gain. Revenue is a transfer; the gain is the $8 of recovered deadweight loss.
• Assuming uniform mandates and permit markets cost the same. Here permits delivered identical abatement for $1,500 less.

Recap

• An externality is a cost or benefit falling on a third party, so private incentives stop matching social welfare.
• Negative externalities cause overproduction; positive externalities cause underproduction; both waste surplus.
• With MB = 40 - 0.5Q, PMC = 8 + 0.5Q and $4 of damage, the market made 32 units while the efficient quantity was 28.
• The 4 excess units created deadweight loss of 0.5 x 4 x 4 = $8, and a $4 per-unit tax moved the market to the efficient 28.
• Subsidies do the mirror job for positive externalities, raising activity toward the efficient level.
• Tradable permits cut 100 tons for $2,000 rather than the $3,500 a uniform mandate would have cost, because abatement went where it was cheapest.
• The Coase theorem shows clear property rights can suffice when bargaining is cheap, but transaction costs and large numbers usually block it.

Sources

1. Greenlaw, S. A., Shapiro, D., & MacDonald, D. (2022). The economics of pollution. In Principles of Microeconomics 3e. OpenStax, Rice University. openstax.org
2. Greenlaw, S. A., Shapiro, D., & MacDonald, D. (2022). Market-oriented environmental tools. In Principles of Microeconomics 3e. OpenStax, Rice University. openstax.org
3. Greenlaw, S. A., Shapiro, D., & MacDonald, D. (2022). Investments in innovation. In Principles of Microeconomics 3e. OpenStax, Rice University. openstax.org
4. Caplan, B. (n.d.). Externalities. In The Concise Encyclopedia of Economics. Liberty Fund. econlib.org
5. Coase, R. H. (1991). The institutional structure of production [Prize lecture]. The Nobel Foundation. nobelprize.org
6. U.S. Environmental Protection Agency. (n.d.). Acid Rain Program. epa.gov
7. Nordhaus, W. D. (2018). Climate change: The ultimate challenge for economics [Prize lecture]. The Nobel Foundation. nobelprize.org

A lighthouse keeper cannot switch off the beam for the ship that did not pay. A fisher who leaves a cod in the water has no assurance the next boat will do the same. Neither problem is about greed or bad character, and neither is fixed by exhorting people to behave better. Both are structural: something about the good itself makes the price mechanism unable to do its job. This final lesson names those structural properties, shows what goes wrong in each case, and then pulls the whole course together around the single yardstick it has been using all along.

Two properties that classify goods

Whether markets can supply a good well depends on two characteristics. A good is rival if one person's use reduces what is available to others (a sandwich - if I eat it, you cannot). It is excludable if people can be prevented from using it unless they pay (a movie ticket gates entry). Crossing these two yes/no properties gives four categories, and the category determines whether private markets, government, or community management is the natural provider:

	Excludable	Non-excludable
Rival	Private good (food, clothing)	Common resource (ocean fish)
Non-rival	Club good (streaming, toll road)	Public good (national defense)

Private goods are the case the whole course assumed until now, and markets handle them well. The other three cells are where trouble arises, because either non-excludability or non-rivalry breaks the price mechanism.

Be careful with two things the classification does not mean. "Public good" is a technical term about rivalry and excludability, not a synonym for "government-provided" or "socially beneficial" - a public park with a fence is a club good, and government-run postal delivery is a private good with a public supplier. And the properties are matters of degree, not switches: a road is non-rival at 3 a.m. and intensely rival at rush hour, which is exactly why congestion pricing exists.

Key idea: Rivalry and excludability, not who provides the good, determine which of the four categories it falls into - and both properties vary by degree and by circumstance.

Public goods and the free-rider problem

A public good is non-rival and non-excludable - national defense, a lighthouse, clean air, basic research, a mosquito-control program. Because you cannot exclude non-payers and one person's use does not diminish another's, private markets struggle to provide it. Each person hopes to enjoy the good while letting others pay - the free-rider problem.

If a neighborhood tries to fund street lighting by voluntary contribution, everyone benefits whether they chip in or not, so each is tempted to contribute nothing, and too little lighting is provided. As a result public goods are typically underprovided by markets, which is a standard rationale for government provision funded by compulsory taxes - taxation solves the free-rider problem by removing the option to opt out.

Worked example: the free-rider problem in numbers

Put the street-lighting story on a spreadsheet. Ten neighbours share a street. Lighting it costs $800, and each neighbour values having it at $150. Total social benefit is 10 x 150 = $1,500 against a cost of $800, so the lamp is clearly worth installing - it would create $700 of net value.

Now watch voluntary funding fail. Suppose the nine others have already contributed. Should the tenth chip in $80? If she pays, she gets the light and is out $80, netting 150 - 80 = $70. If she refuses, the lamp goes up anyway when someone covers the gap - and she gets the full $150 with no payment, because the light cannot be switched off for her alone. Free riding beats contributing by $80, and every one of the ten faces exactly the same arithmetic. Each reasoning individually rationally, all ten decline, and a project worth $700 to the group does not happen.

This is the prisoner's dilemma from Module 6 wearing different clothes: the individually dominant strategy produces the collectively worse outcome. Compulsory taxation solves it not by changing anyone's preferences but by removing the option to opt out - each pays $80, each gets $150, and everyone ends up $70 ahead of where free riding would have left them.

Key idea: Non-excludability makes free riding individually rational even when a public good is worth far more than it costs, which is why voluntary funding underprovides it and compulsory funding can make everyone better off.

Common resources and the tragedy of the commons

A common resource is rival but non-excludable - ocean fisheries, public grazing land, groundwater, a congested road. Since no one can be excluded, each user has an incentive to consume as much as possible before others do, ignoring the cost imposed on everyone else. This is a negative externality in action: each fisher's catch depletes the stock available to all.

The predictable result is overuse and depletion - the tragedy of the commons. Remedies mirror those for externalities: assigning property rights (individual fishing quotas), setting quotas or licenses, or charging for use (congestion pricing on roads) to align private incentives with the shared interest. The economist Elinor Ostrom documented that communities can also self-govern commons through local rules and monitoring, without either privatization or central control.

Ostrom's finding deserves more than a footnote, because it corrects a widespread misreading of Hardin's original essay. Studying irrigation systems, alpine pastures, and inshore fisheries that had been managed sustainably for centuries, she showed that the tragedy is a prediction about unmanaged open access rather than about common ownership as such. Communities that succeeded tended to share a set of features: clearly defined boundaries around the resource and its users, rules matched to local conditions, participation by users in making those rules, monitoring by people accountable to the users, graduated sanctions for violations, and cheap conflict resolution. The policy menu is therefore three items, not two - privatise, regulate centrally, or support community governance - and which works best depends on the resource and the community.

Modern fisheries policy leans on the first of these. Catch shares, which assign fishers a secure right to a defined portion of the allowable catch, convert a race to fish into an asset worth protecting. NOAA Fisheries manages a substantial share of United States catch this way, and the mechanism is precisely the one from the externality lesson: give the decision-maker a stake in the consequence and the incentive to overuse disappears.

Key idea: The tragedy of the commons describes unmanaged open access rather than shared ownership, and property rights, regulation, and community self-governance are three genuine remedies rather than two.

Efficiency and total surplus: the course in one idea

Throughout this course, the yardstick for judging outcomes has been economic efficiency: an allocation is efficient when it maximizes total surplus, the sum of consumer surplus and producer surplus. A competitive market with no failures reaches this ideal because price equals marginal cost and every trade whose benefit exceeds its cost occurs. This is the standard against which every deviation is measured.

Each topic in the final modules is a way efficiency can break down:

• Market power (monopoly) sets price above marginal cost, restricting output and creating deadweight loss.
• Externalities drive a wedge between private and social costs or benefits, causing over- or under-production.
• Public goods and common resources fail because of non-excludability, producing free riding and overuse.

In every case the diagnosis is the same - some mutually beneficial trades are missed or some harmful activity is overdone - and the policy question is whether an intervention can raise total surplus by more than it costs. That balance of benefits against costs, applied at the margin, is the enduring lesson of microeconomics.

Two cautions keep the yardstick honest, and both have appeared before. First, efficiency is silent on distribution: an allocation can maximise total surplus while leaving the gains very unevenly split, and choosing between efficient allocations is a normative question that economics informs but does not settle. Second, identifying a market failure does not automatically justify intervention. Governments face their own information problems, their own incentive problems, and their own capture by concentrated interests, so the honest comparison is between an imperfect market and an imperfect remedy - not between a flawed market and a perfect planner. The question is always whether a specific policy recovers more surplus than it destroys.

Key idea: Total surplus measures efficiency and not fairness, and demonstrating a market failure is the beginning of a policy argument rather than the end of one.

Why it matters and where to go next

This final lesson unifies everything. The tools you built - opportunity cost, supply and demand, elasticity, utility, cost curves, market structure, and factor markets - all feed into a single question: does this allocation maximize total surplus, and if not, why not and what would help? That question is the working core of applied economics, from designing a carbon market to regulating a utility to deciding how to fund a lighthouse. Mastering it means you can reason about almost any resource-allocation problem, which is exactly what an introductory course in microeconomics sets out to teach.

Common wrong turns

• Calling anything the government supplies a public good. The test is rivalry and excludability, not the identity of the provider.
• Confusing public goods with common resources. Both are non-excludable, but public goods are non-rival while common resources are rival - which is why one is underprovided and the other overused.
• Treating the tragedy of the commons as inevitable. Ostrom documented communities that governed shared resources sustainably for centuries.
• Blaming free riding on selfishness. It is a structural consequence of non-excludability, and it appears even among people who want the good provided.
• Reading maximum total surplus as maximum fairness. Efficiency says nothing about who gets the surplus.
• Assuming a market failure justifies any intervention. The remedy must recover more surplus than it costs, and government has failure modes of its own.

Recap

• Rivalry and excludability sort goods into private goods, club goods, common resources, and public goods.
• Public goods are non-rival and non-excludable, so free riding leaves them underprovided by voluntary funding.
• Ten neighbours each valuing a $800 lamp at $150 would gain $700 collectively, yet each individually gains $80 by not paying - so nobody does.
• Common resources are rival but non-excludable, so each user ignores the cost imposed on others and the resource is overused.
• Remedies include property rights such as catch shares, quotas and congestion pricing, and Ostrom's community self-governance.
• Efficiency means maximising total surplus, and market power, externalities, and non-excludability are the three ways this course has seen it fail.
• Efficiency is silent about distribution, and a market failure begins rather than ends the argument for intervention.

Sources

1. Greenlaw, S. A., Shapiro, D., & MacDonald, D. (2022). Public goods. In Principles of Microeconomics 3e. OpenStax, Rice University. openstax.org
2. Greenlaw, S. A., Shapiro, D., & MacDonald, D. (2022). Efficiency in perfectly competitive markets. In Principles of Microeconomics 3e. OpenStax, Rice University. openstax.org
3. Cowen, T. (n.d.). Public goods. In The Concise Encyclopedia of Economics. Liberty Fund. econlib.org
4. Hardin, G. (n.d.). Tragedy of the commons. In The Concise Encyclopedia of Economics. Liberty Fund. econlib.org
5. Ostrom, E. (2009). Beyond markets and states: Polycentric governance of complex economic systems [Prize lecture]. The Nobel Foundation. nobelprize.org
6. National Oceanic and Atmospheric Administration Fisheries. (n.d.). Catch shares. U.S. Department of Commerce. fisheries.noaa.gov
7. Alchian, A. A. (n.d.). Property rights. In The Concise Encyclopedia of Economics. Liberty Fund. econlib.org