Macroeconomics

Zooming out from the individual to the whole

Microeconomics studies individual pieces of the economy - a single buyer, one firm, or one market. Macroeconomics steps back to study the economy as a whole: the total value of everything a nation produces, the average level of prices, the share of workers without jobs, and the pace at which living standards rise. It asks why entire economies boom and slump together, and what governments and central banks can do about it.

A quick way to feel the difference is to compare questions. Microeconomics asks: if the price of coffee beans rises, how many fewer lattes will a cafe sell? Will a tax on sugary drinks change what teenagers buy? Macroeconomics asks: why do millions of workers across hundreds of unrelated industries lose their jobs in the same year? Why do the prices of nearly everything rise together in some decades and barely move in others? Why is the average person in some countries dozens of times richer than in others? Same discipline, different altitude.

The shift in scale matters because the whole is not simply the sum of its parts. A single household spending less is thrift; every household cutting spending at once can shrink the economy and cost people their jobs. This surprising result - that what is sensible for one actor can be harmful for all at once - is called the paradox of thrift, and it is a hint that macroeconomics needs its own tools.

The fallacy of composition

The paradox of thrift is one instance of a broader logical trap called the fallacy of composition: assuming that what is true for a part must be true for the whole. If you stand up at a stadium, you see better. If everyone stands, no one sees better and everyone is less comfortable. If one farmer harvests a record crop, her income rises; if every farmer harvests a record crop, the market price can fall so far that farm incomes drop.

Banking supplies the sharpest macroeconomic example. If one depositor withdraws her cash because she doubts her bank, that is prudence. If every depositor rushes to withdraw at once, even a sound bank fails, because banks keep only a fraction of deposits on hand. Waves of exactly such panics swept the United States in the early 1930s, closing thousands of banks and deepening the Great Depression. Individually reasonable choices added up to collective disaster - the defining signature of a macroeconomic problem.

This is why macroeconomics cannot simply scale up the supply-and-demand diagram of a single market. It needs models built for the whole system, where my spending is your income and your spending is mine. The circular-flow model in the next lesson is the first of these, and the aggregate demand and aggregate supply model arrives in Module 6.

How macroeconomics became its own field

Economists had written about money, trade, and national wealth for centuries, but macroeconomics as a distinct field was born in the Great Depression. Between 1929 and 1933, output in the United States fell by roughly a third, the price level dropped, and about one worker in four was unemployed. The classical economics of the day taught that markets adjust on their own and that prolonged mass unemployment should not happen. Yet it happened, year after year. A theory was needed for an economy stuck far below its capacity.

The British economist John Maynard Keynes supplied one in 1936 with his General Theory of Employment, Interest and Money. Keynes argued that total spending - what we now call aggregate demand - can fall short, and that output and employment then fall with it. Around the same time, Simon Kuznets built the first systematic national income accounts for the United States, giving the new theory something to measure. Modern GDP statistics descend directly from that work, which later earned Kuznets a Nobel Prize.

The decades since have refined rather than discarded these founding ideas. Milton Friedman and the monetarists showed how much money and central banks matter, and the stagflation of the 1970s forced economists to take supply shocks and expectations seriously. The mainstream view taught in this course blends the lessons: spending drives the economy in the short run, productive capacity governs it in the long run, and policy operates in the space between.

The three big questions

Nearly everything in this course serves three goals that economists use to judge how an economy is doing:

1. Output and growth - Is the economy producing more goods and services over time? We measure this with gross domestic product (GDP) and its growth rate.
2. Employment - Are the people who want to work able to find jobs? We measure this with the unemployment rate.
3. Stable prices - Is the overall price level roughly steady, avoiding high inflation or falling prices? We measure this with the inflation rate.

An economy that grows steadily, keeps unemployment low, and holds inflation mild is doing its job. When one of these breaks down - a recession throws people out of work, or prices spiral upward - macroeconomic policy tries to bring it back toward balance.

Each goal earns its place on the list. Growth compounds: an economy growing 2 percent per year doubles its output roughly every 35 years, while one growing 1 percent takes about 70, so small differences in the growth rate separate prosperous generations from stagnant ones. Unemployment wastes the one resource that cannot be stored - an idle hour of labor is gone forever - and long spells without work erode skills, health, and family finances. Unstable prices scramble every contract, wage, and savings plan written in money, making the future harder to plan for everyone at once.

Reading a macro dashboard

Put the three indicators together and you can take an economy's temperature at a glance. Suppose three countries report the following figures for the same year. Country A: real GDP growth of 3 percent, unemployment of 5 percent, inflation of 2 percent. Country B: growth of minus 2 percent, unemployment of 9 percent, inflation of 1 percent. Country C: growth of 6 percent, unemployment of 3 percent, inflation of 9 percent.

Country A looks healthy: output is rising at a sustainable pace, most job seekers are finding work, and prices are nearly steady. Country B is in a recession: output is shrinking, roughly one worker in eleven is unemployed, and weak spending is holding inflation down. Country C is overheating: spending is outrunning productive capacity, labor is scarce, and prices are climbing fast. No single number tells the story; the pattern across all three does. Much of this course is learning to read such patterns and to trace what caused them.

The indicators also interact. Rapid growth tends to pull unemployment down. Deep slumps drag inflation lower. Pushing unemployment very low can push inflation up, a tension captured later by the Phillips curve. Reading one dial in isolation is like diagnosing a patient from temperature alone; the skill is reading the dials together.

Stocks versus flows

A recurring distinction is between a stock, a quantity measured at a single point in time, and a flow, a quantity measured over a period. Your bank balance today is a stock; your monthly income is a flow. In macroeconomics, the national debt is a stock (a total owed on a given date), while the annual budget deficit is a flow (the shortfall added each year). GDP itself is a flow - output produced per year. Confusing the two is a common error, so it is worth fixing the idea early.

A bathtub makes the distinction vivid. The water level in the tub is a stock: it exists at an instant, and you can read it off at any moment. The water pouring from the faucet is a flow: it only makes sense per minute or per hour. The flow changes the stock - every minute of inflow raises the level - but the two are measured in different units and answer different questions.

Try the arithmetic with public finances. Suppose a government starts the year with a national debt of 10,000 (in billions) and runs a deficit of 500 during the year. The debt ends the year at 10,000 + 500 = 10,500. Next year the deficit shrinks to 300. Good news? Partly: the flow is smaller, but it is still positive, so the debt rises again, to 10,800. A headline that celebrates a falling deficit while the debt climbs is describing a slower faucet, not a draining tub.

The pairing shows up everywhere in macroeconomics. Your wealth is a stock; your saving this year is a flow that adds to it. A factory's capital is a stock; this year's investment is the flow that builds it. The money supply on a given day is a stock; the income earned during a year is a flow. Whenever a new measure appears in this course, ask first: point in time, or per period?

Positive and normative, again

As in all economics, we separate positive statements (testable claims about what is - "unemployment rose to 5% last quarter") from normative statements (value judgments about what ought to be - "the government should cut unemployment even at the cost of higher inflation"). Data can settle the first; only values can settle the second. Keeping them apart keeps analysis honest.

Two refinements sharpen the distinction. First, a positive statement can be false and still be positive: "inflation always falls when unemployment rises" is testable, and the stagflation of the 1970s showed it fails, so it is a wrong positive claim rather than a normative one. Second, most policy debates mix the two. "A higher minimum wage reduces employment for some workers by some amount" is positive; "the trade-off is worth accepting" is normative. Evidence can narrow disagreement about the first kind; the second kind ultimately belongs to citizens and their values.

Models: maps, not photographs

Macroeconomists work with models: deliberately simplified descriptions of the economy, usually a handful of relationships among aggregates. The simplification is the point. A map that reproduced every building at full scale would be useless; a subway map that omits almost everything gets you where you are going. A good model keeps only the features needed for the question at hand and holds other influences equal - the familiar assumption of ceteris paribus - so that one cause can be studied at a time.

This course builds a small toolkit of such maps. The circular flow shows how spending, income, and output form a single loop. The GDP accounts measure the loop. Aggregate demand and aggregate supply explain how output and prices move together in booms and slumps. The money and banking module explains the plumbing behind spending, and the policy modules show the levers. None of these is a photograph of reality; each is a map drawn to answer a particular class of questions.

Common misconceptions

Before the tools arrive, it helps to clear away a few ideas that sound plausible and are wrong. Each returns later in the course with a fuller answer.

• "Macro is just micro added up." No. Because of the fallacy of composition, whole-economy outcomes can reverse individual logic, as the paradox of thrift shows. Macroeconomics needs models of the loop, not just models of single markets.
• "Zero unemployment would be ideal." No. A healthy economy always has some people between jobs or searching for a better match. Module 4 defines full employment as a positive natural rate of unemployment, not zero.
• "If inflation falls, prices are falling." No. Falling inflation, called disinflation, means prices are rising more slowly; only deflation means the price level itself is declining.
• "The stock market is the economy." No. Share prices are asset values that swing with profit expectations; GDP, jobs, and inflation measure production and livelihoods. The two can diverge for long stretches.
• "A bigger GDP always means better lives." Only partly. GDP tracks production, not distribution, leisure, unpaid work, or the environment - limits examined directly in Module 2.

Try it

A country begins year 1 with a national debt of 8,000 (in billions of its currency). It runs a budget deficit of 600 in year 1, a deficit of 400 in year 2, and a surplus of 100 in year 3. (a) Classify the debt and the deficit as a stock or a flow. (b) Compute the debt at the end of each year. (c) A commentator says: "the deficit fell from 600 to 400, so the country's debt fell." Evaluate the claim. (d) Is the commentator's claim positive or normative?

Answer: (a) The debt is a stock, measured on a date; the deficit is a flow, measured per year. (b) End of year 1: 8,000 + 600 = 8,600. End of year 2: 8,600 + 400 = 9,000. End of year 3, with the surplus: 9,000 - 100 = 8,900. (c) The claim is false. A smaller deficit still adds to the debt; the debt rose from 8,600 to 9,000 in year 2 and only fell once the budget swung to surplus. (d) It is a positive claim - testable against the numbers - and the test shows it is wrong.

Sources

1. Greenlaw, S. A., Shapiro, D., & MacDonald, D. (2022). Microeconomics and macroeconomics. In Principles of Macroeconomics 3e. OpenStax, Rice University. openstax.org
2. Greenlaw, S. A., Shapiro, D., & MacDonald, D. (2022). How economists use theories and models to understand economic issues. In Principles of Macroeconomics 3e. OpenStax, Rice University. openstax.org
3. Greenlaw, S. A., Shapiro, D., & MacDonald, D. (2022). Measuring the size of the economy: Gross domestic product. In Principles of Macroeconomics 3e. OpenStax, Rice University. openstax.org
4. Richardson, G. (2013). The Great Depression. In Federal Reserve History. Federal Reserve Bank of St. Louis. federalreservehistory.org
5. Henderson, D. R. (Ed.). (2008). Simon Kuznets. In The Concise Encyclopedia of Economics (2nd ed.). Liberty Fund. econlib.org
6. Board of Governors of the Federal Reserve System. (2021). Monetary policy: What are its goals? How does it work? In Monetary Policy Principles and Practice. Federal Reserve Board. federalreserve.gov
7. U.S. Department of the Treasury. (n.d.). Understanding the national debt. In Fiscal Data: America's Finance Guide. Bureau of the Fiscal Service. fiscaldata.treasury.gov

An economy as a loop

The circular-flow model is the simplest picture of how a whole economy fits together. In its basic form, two groups meet in two markets. Households own the resources - labor, land, and capital - and sell them to firms in the resource (factor) markets, earning wages, rent, and profit. Firms use those resources to make goods and services, which they sell back to households in the product markets. Money flows one way around the loop; goods, services, and resources flow the other way.

The key insight is that every dollar of spending by one party is a dollar of income for another. When you buy a coffee, your spending becomes the cafe's revenue, which becomes wages for its workers and profit for its owner. Follow the loop all the way around and you reach a fundamental identity: total income equals total expenditure equals the value of total output. This is why, as the next module shows, GDP can be measured either by adding up all spending or by adding up all income - the two must match.

Economists attach a name to each factor payment. Labor earns wages, land earns rent, capital earns interest, and entrepreneurship earns profit. Notice the deliberate simplifications in the two-sector version: there is no government, no banking system, and no foreign trade yet. That is not carelessness; it is the model-building strategy from Lesson 1. Start with the essential loop, confirm the logic, then bolt on the sectors one at a time and watch what each adds.

A model with a long pedigree

The circular flow is one of the oldest ideas in economics. In 1758 the French physician-economist Francois Quesnay published the Tableau Economique, a table tracing how spending circulates among farmers, landlords, and artisans. As a doctor at the court of Versailles, Quesnay borrowed his central image from the circulation of blood: income must keep moving through the economic body, and a blockage anywhere sickens the whole organism.

Two centuries later the same picture became the backbone of official statistics. The national accounts that Simon Kuznets and others built in the 1930s measure the loop from three directions at once - what is produced, what is earned, and what is spent - precisely because the circular flow implies all three must agree. When you meet GDP in the next module, you are meeting Quesnay's loop with modern plumbing.

Following a dollar around the loop

Trace a concrete purchase to see the identity work. Suppose you spend $100 on a restaurant dinner. The restaurant uses your $100 as follows: $30 buys ingredients from a farmer, $35 pays the cook and the server, $20 pays the landlord for the dining room, and the remaining $15 is the owner's profit. The farmer, in turn, pays $10 in wages to a farmhand and keeps $20 as farm profit. Every cent of your spending has now landed in someone's pocket as income.

Add up the incomes: restaurant wages $35, restaurant rent $20, restaurant profit $15, farm wages $10, farm profit $20. The total is $100 - exactly your expenditure. Equivalently, count value added: the farmer created $30 of value from scratch, and the restaurant added $70 by turning $30 of ingredients into a $100 meal. Value added sums to $100 as well. Expenditure, income, and output are three tallies of the same event.

This is not a coincidence of the example; it is bookkeeping. Money spent must be received. The circular flow simply scales this dinner up to an entire economy, where billions of such transactions chain households and firms into one continuous loop.

Leakages and injections

The real economy is not a perfectly closed loop, because money can leave the flow and re-enter it elsewhere. Money that leaves the domestic spending stream is a leakage; money that enters it is an injection.

Leakages (money leaves the flow)	Injections (money enters the flow)
Saving (S)	Investment (I)
Taxes (T)	Government spending (G)
Imports (M)	Exports (X)

When households save, that income is not spent on domestic goods right away - it leaks out - but banks channel it back as investment when firms borrow to build factories. Taxes leak out to the government, which injects them back through spending. Money spent on imports leaks abroad, while exports inject foreign spending in. When total injections equal total leakages, the flow is balanced and the economy is at rest; when injections exceed leakages, spending and output tend to expand.

Note that "leakage" does not mean the money is destroyed. Each leakage has a matching return channel: saving flows through the financial market and can come back as investment, taxes flow through the government budget and can come back as public spending, and import payments flow through foreign exchange and can come back as export sales. Picture three side loops bolted onto the main loop. The question is never whether the money exists, but whether it returns to the domestic spending stream, and how quickly.

Balancing the flow: a worked example

Suppose an economy records, in billions per year: saving S = 400, taxes T = 600, imports M = 350, investment I = 500, government spending G = 650, and exports X = 150. Total leakages are 400 + 600 + 350 = 1,350. Total injections are 500 + 650 + 150 = 1,300. Injections fall short of leakages by 50, so the spending stream is shrinking: firms see sales soften, trim production, and hire less. Output drifts down until the flows again balance.

Now rerun the year with exports at 200. Injections become 500 + 650 + 200 = 1,350, exactly matching leakages of 1,350, and the loop is at rest: what drains out each period is exactly replaced. Notice the balancing condition is the sum S + T + M = I + G + X. The individual pairs need not match: here saving (400) is below investment (500), and taxes (600) fall short of government spending (650), yet the totals balance because trade runs a deficit of 100.

A third scenario connects back to the paradox of thrift. Return to the original numbers, with exports at 150 and a shortfall of 50, and suppose households respond to hard times by saving less: S falls from 400 to 350. Leakages drop to 1,300, matching injections of 1,300, and the contraction stops. In the short run, less saving means more spending and steadier output. The long-run catch is that saving is also the raw material for investment, so an economy that never saves struggles to build capital - a tension the growth module takes up.

Adding banks, government, and the world

Each side loop has its own story. The financial system gathers household saving and lends it to firms, converting the S leakage into the I injection. The loop only breaks if saving sits idle - for example, cash stuffed in mattresses. That is not a hypothetical worry: in the early 1930s, frightened households and banks hoarded cash on a large scale, the recycling of saving into investment collapsed, and the spending stream contracted with it.

The government loop balances or not as budgets dictate. If G exceeds T, the government runs a deficit and borrows the difference in the financial market, injecting more than it drains. If T exceeds G, the surplus retires debt. The foreign loop works the same way: when imports exceed exports, more money flows out than in through trade, and the difference returns as foreign purchases of domestic assets - a capital inflow. These links between budget deficits, trade deficits, and borrowing reappear in Module 9.

Why the model matters

The circular flow is more than a diagram. It shows that output, income, and spending are three views of the same thing, it names the six forces (S, T, M, I, G, X) that drive the economy up or down, and it previews the accounting behind GDP. Keep the loop in mind and later topics - the spending multiplier, trade balances, and fiscal policy - will feel like natural extensions of a picture you already understand.

The model also disciplines everyday reasoning. When a commentator claims that consumers "need to spend to keep the economy going," the loop shows what is right about that (spending is income) and what is incomplete (saving that banks lend on is also spending, just routed through firms). When another warns that imports "drain money from the country," the loop shows the drain returning through export demand or capital inflows. Accounting is not a verdict, but it rules out stories that do not add up.

From the loop to the multiplier

The loop also explains why shocks to spending amplify. Inject 100 of new spending into the stream and suppose households spend 0.8 of each extra dollar of income, leaking the rest. The first 100 becomes income, of which 80 is re-spent, which becomes income, of which 64 is re-spent, then 51.20, and so on. The rounds sum to 100 x 1 / (1 - 0.8) = 500. Leakages at every round are exactly what keep the sum finite. This multiplier logic is developed fully in the fiscal-policy lesson, but it is born here, in the loop.

The size of the amplification depends on how leaky the loop is. If households re-spend only 0.5 of each extra dollar - because more drains away to saving, taxes, and imports - the rounds are 100, 50, 25, 12.50, and the sum is just 100 x 1 / (1 - 0.5) = 200. A leakier loop damps shocks; a tighter loop amplifies them. The same plumbing that spreads prosperity in a boom spreads weakness in a slump, which is why recessions ripple across industries that never saw the original shock.

Common misconceptions

Four confusions about the circular flow are worth retiring immediately.

• "Money and goods move together." They move in opposite directions around the loop. Your money goes to the cafe; the coffee comes to you. In factor markets, your labor goes to the firm; the wage comes back.
• "Saving disappears from the economy." Saving routed through banks and markets finances investment and stays in the flow. Only saving that sits idle - hoarded cash - truly drains the stream, which is why financial panics are so damaging.
• "Imports make a country poorer dollar for dollar." Imports are a leakage from the domestic spending stream, but the country receives real goods, and the outflow returns as export demand or investment from abroad. The accounting says "routed elsewhere," not "lost."
• "The loop is a fixed pie." The flow can grow or shrink. Injections above leakages expand it, and the same dollar circulates many times per year, so the money stock and the annual flow of income are different things - a stock and a flow, as Lesson 1 warned.

Try it

An economy has saving of 350, taxes of 500, and imports of 250 (all in billions per year). Investment is 400 and government spending is 600. (a) What level of exports balances the circular flow? (b) If exports instead rise to 180, is the economy expanding or contracting, and by how much do injections exceed or fall short of leakages? (c) With exports at 180, does the government budget balance?

Answer: (a) Leakages are 350 + 500 + 250 = 1,100. Injections are 400 + 600 + X, so balance requires X = 1,100 - 1,000 = 100. (b) With X = 180, injections total 1,180 versus leakages of 1,100, an excess of 80, so the spending stream and output tend to expand. (c) No. Taxes are 500 and government spending is 600, a budget deficit of 100 - the loop can balance overall even while individual sectors do not.

Sources

1. Greenlaw, S. A., Shapiro, D., & MacDonald, D. (2022). Measuring the size of the economy: Gross domestic product. In Principles of Macroeconomics 3e. OpenStax, Rice University. openstax.org
2. Greenlaw, S. A., Shapiro, D., & MacDonald, D. (2022). The building blocks of Keynesian analysis. In Principles of Macroeconomics 3e. OpenStax, Rice University. openstax.org
3. Greenlaw, S. A., Shapiro, D., & MacDonald, D. (2022). The national saving and investment identity. In Principles of Macroeconomics 3e. OpenStax, Rice University. openstax.org
4. Henderson, D. R. (Ed.). (2008). Francois Quesnay. In The Concise Encyclopedia of Economics (2nd ed.). Liberty Fund. econlib.org
5. Anderson, L. (2008). Gross domestic product. In The Concise Encyclopedia of Economics (2nd ed.). Liberty Fund. econlib.org
6. U.S. Bureau of Economic Analysis. (n.d.). NIPA handbook: Concepts and methods of the U.S. national income and product accounts. In BEA Methodologies. U.S. Department of Commerce. bea.gov
7. The CORE Team. (2017). Unemployment and fiscal policy. In The Economy 1.0. CORE Economics Education. core-econ.org

What GDP is - word by word

Gross domestic product (GDP) is the market value of all final goods and services produced within a country in a given period. Every phrase does work:

• Market value - we add up dissimilar goods (cars, haircuts, apples) by their prices, giving a common dollar measure.
• Final - we count only goods sold to their end user, not intermediate goods resold or built into something else, to avoid double counting.
• Produced - GDP counts new production, not the resale of used goods or purely financial transactions like buying stocks.
• Within a country - GDP is about location of production, regardless of who owns the firm.
• In a period - usually a year or a quarter; GDP is a flow.

The definition is a checklist. When you wonder whether something belongs in GDP, walk the phrases in order: Is it sold at a market price? Is it a final good or service? Was it produced, and produced now? Was it produced here? Most puzzles dissolve at one of those steps. The rest of this lesson practices exactly that walk, because the accounting habits you build here carry through every later module.

The "within a country" phrase settles a puzzle worth pausing on. A foreign-owned auto plant operating inside the country produces domestic output: its production belongs to the host country's GDP, because GDP cares about where production happens, not who owns the factory. Symmetrically, a factory your country's firms operate abroad counts in the foreign host's GDP. The ownership-based cousin of GDP, called GNP, appears in the next lesson.

What GDP deliberately excludes

Three families of transactions involve money changing hands without any new production, so the accountants leave them out.

• Used goods. A five-year-old car resold for $9,000 adds nothing to this year's GDP; it was counted the year it was built. But the used-car dealer's $700 margin is counted, because the dealer produced a genuine service this year: finding, reconditioning, and matching the car to a buyer.
• Purely financial transactions. Buying $5,000 of shares transfers ownership of existing assets; nothing is produced. The broker's $50 commission, a real service, does count.
• Transfer payments. A government pension or unemployment check moves income between pockets without purchasing any good or service. It enters GDP only later, if and when the recipient spends it on final output.

The same logic handles timing. A house built and sold new this year counts this year. If it resells a decade later for triple the price, the appreciation never enters GDP - only the real estate agent's fee does. GDP measures production, not asset churn.

The expenditure approach

The most common way to compute GDP is to add up all the spending on final goods and services, sorted into four categories:

GDP = C + I + G + NX

• Consumption (C) - household spending on goods and services; the largest component, often around two-thirds of GDP in rich economies. It spans durables (cars, appliances), nondurables (food, clothing), and services (haircuts, medical care, rent).
• Investment (I) - business spending on capital (machines, buildings) plus new housing and changes in inventories. Note: buying stocks is not investment here.
• Government purchases (G) - federal, state, and local spending on goods and services. It excludes transfer payments like Social Security, which are not payments for production.
• Net exports (NX) - exports minus imports, NX = X - M. We subtract imports because C, I, and G include spending on foreign-made goods that were not produced here.

Two components deserve a closer look. Inventory change belongs in investment because GDP tracks production, and spending is only the measuring stick. A snowblower built in the fall but still unsold on December 31 was produced this year, so it enters this year's GDP as inventory investment; when it sells next year, consumption rises but inventory investment falls by the same amount, and next year's GDP is untouched. Production is counted once, in the year it happens.

The import subtraction is the most misread term in the formula. Imports do not shrink GDP; the subtraction merely cancels imported content that snuck into C, I, and G. If a household buys a foreign-built car for $30,000, consumption rises by 30,000 and imports rise by 30,000, so the net effect on measured GDP is exactly zero - as it should be, since nothing was produced domestically. The minus sign is bookkeeping hygiene, not a penalty on trade.

Worked example

Suppose a country reports, for one year (in billions of dollars): consumption $14,000; investment $3,500; government purchases $4,000; exports $2,100; imports $2,600. Compute GDP.

Net exports NX = X - M = 2,100 - 2,600 = -500
GDP = C + I + G + NX = 14,000 + 3,500 + 4,000 + (-500)
GDP = $21,000 billion (that is, $21 trillion)

The net-export term is negative here because the country imports more than it exports - a trade deficit - which reduces measured GDP relative to what domestic spending alone would suggest.

Reading the shares is as informative as the total. Consumption is 14,000 / 21,000, about 67 percent - the typical two-thirds. Investment is 3,500 / 21,000, roughly 17 percent, and it is the most volatile piece: in recessions, firms can slash equipment orders and builders can halt projects while households merely trim consumption. That volatility, small share notwithstanding, makes investment the usual driver of business-cycle swings, a theme that returns in the aggregate demand module.

A second tally, with classification traps

Practice on a fuller ledger. An economy records, in billions: household purchases of goods and services 6,000; business equipment purchases 1,200; new residential construction 500; inventory change +100; government purchases 1,500; government pension checks 400; exports 900; imports 1,200; used-home sales 700; share purchases 2,000.

C = 6,000. I = 1,200 + 500 + 100 = 1,800. G = 1,500 (the 400 of pensions is a transfer - excluded).
NX = 900 - 1,200 = -300. Used homes and share purchases: excluded entirely.
GDP = 6,000 + 1,800 + 1,500 - 300 = 9,000 billion.

Every excluded item fails the definition somewhere. Pension checks buy no production. Used homes were counted when built. Share purchases shuffle paper claims. Meanwhile the +100 of unsold inventory is production, so it counts even though no final buyer appeared. If you can defend each of those five calls aloud, you understand the definition; if any feels arbitrary, reread the checklist above.

Avoiding double counting

Why count only final goods? Imagine a baker buys $0.40 of flour to make bread sold for $2.00. If we counted both the $0.40 flour and the $2.00 loaf, we would count that flour twice - once on its own, once inside the loaf. Counting only the final $2.00 loaf captures the full value exactly once. An equivalent method sums the value added at each stage (the seller's price minus the cost of its inputs); the total value added always equals the final sale price.

Stretch the chain to three stages to see the equivalence clearly. A farmer grows wheat and sells it to a miller for $0.15. The miller grinds it into flour and sells the flour to the baker for $0.40. The baker sells the loaf for $2.00. Value added: farmer $0.15 (from nothing), miller $0.40 - $0.15 = $0.25, baker $2.00 - $0.40 = $1.60. Sum: 0.15 + 0.25 + 1.60 = $2.00, exactly the final price. Whether statisticians count final sales or sum value added, the loaf contributes two dollars to GDP, once.

The final-versus-intermediate label depends on the buyer, not the good. Flour bought by the baker is intermediate; the same flour bought by a household for weekend baking is a final consumption good. Electricity sold to a factory is intermediate; sold to your apartment, final. This is why statistical agencies track who purchases what - the identical product can sit on either side of the line.

How the statisticians actually do it

No agency observes GDP directly. Statisticians assemble it from thousands of sources: retail sales surveys for consumption, construction permits and equipment shipments for investment, budget records for government purchases, customs data for trade. Early quarterly estimates lean on partial data and are revised, sometimes noticeably, as fuller tax and survey records arrive months later. A first print is a sketch; the settled figure comes with time.

Two conventions help you read reported numbers. Quarterly GDP is usually seasonally adjusted, stripping out predictable swings like holiday shopping so that a fourth-quarter rise means genuine growth rather than the usual December. And quarterly growth is often stated at an annualized rate: the pace the economy would sustain if that quarter's growth continued for a full year. A 0.5 percent quarterly gain annualizes to roughly 2 percent.

Levels, growth rates, and size comparisons

A GDP level on its own says little; use in context is everything. Growth is the percentage change between periods: if GDP rises from 9,000 to 9,270, growth is (9,270 - 9,000) / 9,000 = 270 / 9,000 = 3 percent. Comparing countries requires dividing by population: a country with GDP of 2,000 and 100 million people (20,000 per person) is far richer per person than one with GDP of 3,000 spread over 300 million people (10,000 per person), despite the smaller total.

One more caution before the next lesson formalizes it: GDP measured in current prices can rise merely because prices rose. Separating real production growth from inflation is the whole business of real GDP and the deflator, coming in Module 3. For now, treat every GDP comparison across time as provisional until prices are held fixed.

Common misconceptions

• "Imports subtract from GDP, so buying foreign goods shrinks the economy." The subtraction only cancels imported content already inside C, I, and G. An imported car leaves measured GDP unchanged, not lower.
• "Government checks are government purchases." Transfers move income without buying output; only purchases of goods and services enter G.
• "Investment means buying stocks and bonds." In GDP accounting, investment is new physical capital, new housing, and inventory change. Financial purchases are reshuffled ownership.
• "GDP counts every transaction." It counts production once: final goods only, this period only, this territory only.
• "Unsold goods do not count until they sell." They count now, as inventory investment; the later sale is deliberately neutral.

Try it

An economy reports, in billions: consumption 5,600; investment 1,400; government purchases 1,700; exports 800; imports 1,000. Separately, a dealer sells a used truck for $30,000, of which $2,000 is the dealer's service margin. (a) Compute GDP. (b) How much of the truck sale belongs in GDP? (c) If firms had also accumulated 50 of unsold inventory that was left out of the investment figure, what would corrected GDP be?

Answer: (a) NX = 800 - 1,000 = -200, so GDP = 5,600 + 1,400 + 1,700 - 200 = 8,500 billion. (b) Only the $2,000 margin - a service produced this year. The truck itself was counted when new. (c) Inventory change is investment, so I rises to 1,450 and GDP = 8,550 billion. Production, not the timing of the final sale, is what GDP tracks.

Sources

1. Greenlaw, S. A., Shapiro, D., & MacDonald, D. (2022). Measuring the size of the economy: Gross domestic product. In Principles of Macroeconomics 3e. OpenStax, Rice University. openstax.org
2. Greenlaw, S. A., Shapiro, D., & MacDonald, D. (2022). Comparing GDP among countries. In Principles of Macroeconomics 3e. OpenStax, Rice University. openstax.org
3. U.S. Bureau of Economic Analysis. (n.d.). What to know about GDP. In BEA Learning Center. U.S. Department of Commerce. bea.gov
4. U.S. Bureau of Economic Analysis. (n.d.). Gross domestic product. In National Data: National Income and Product Accounts. U.S. Department of Commerce. bea.gov
5. U.S. Bureau of Economic Analysis. (n.d.). NIPA handbook: Concepts and methods of the U.S. national income and product accounts. In BEA Methodologies. U.S. Department of Commerce. bea.gov
6. Anderson, L. (2008). Gross domestic product. In The Concise Encyclopedia of Economics (2nd ed.). Liberty Fund. econlib.org
7. World Bank. (n.d.). GDP growth (annual %). In World Development Indicators. World Bank Group. data.worldbank.org

The other side of the ledger

Because every dollar spent on final output becomes someone's income, GDP can be measured a second way: the income approach adds up all the income earned in production instead of all the spending on it. The main categories are wages and salaries (compensation of employees), rent, interest, and profits, with a couple of accounting adjustments for depreciation and taxes on production. In principle the income approach and the expenditure approach give the identical total - they are two ways of counting the same circular flow. In practice small measurement gaps appear, reported as a "statistical discrepancy," but the two are conceptually equal.

Walk through a full reconciliation. Suppose the expenditure approach says GDP is 9,000 (in billions). The income side of the same economy reports: compensation of employees 5,400; rents 200; interest 400; proprietors' income (earnings of the self-employed) 700; corporate profits 1,100. Those five sum to 5,400 + 200 + 400 + 700 + 1,100 = 7,800, a total called national income. It falls short of 9,000, and the shortfall is not an error - two wedges remain.

The first wedge is depreciation (formally, consumption of fixed capital): the 900 of machines, vehicles, and buildings worn out producing this year's output. That value was part of what buyers paid for, but it is income to no one; it offsets capital being used up. The second wedge is indirect business taxes, here 300: sales and excise taxes collected inside market prices that go to the government before anyone books them as earnings. Add the wedges: 7,800 + 900 + 300 = 9,000. The ledger closes.

Why bother with a second method when the first works? Cross-checking. Expenditure data come from retail surveys, builders, and customs; income data come from payroll records and tax filings - independent sources, independent errors. When the two tallies land close together, statisticians trust both. The lingering gap, the statistical discrepancy, is printed rather than hidden, an honest confession of measurement in a messy world.

Why profits close the ledger

The identity is guaranteed by the definition of profit. Consider one firm: it sells 100 of output, pays 60 in wages, 10 in rent, and 5 in interest. Whatever remains - 100 - 60 - 10 - 5 = 25 - is profit, by definition. Profit is the residual claim, so wages + rent + interest + profit automatically exhausts the firm's value added. Add this up across every firm and total income must equal total output. Profit is not an awkward leftover in the accounts; it is the item that forces them to balance.

Strictly speaking there is a third road to the same summit: the production (value-added) approach, which sums each firm's output value minus its purchased inputs, as in the bread chain of the previous lesson. Expenditure, income, and production are three censuses of one circular flow, and modern national accounts publish all three. The triple bookkeeping is the practical payoff of the identity you met in the circular-flow lesson: what is produced is sold, and what is sold is earned.

A family of related measures

Several close cousins of GDP appear in the news, and it helps to keep them straight:

• Gross national product (GNP) counts output by a country's residents wherever it is produced, rather than output produced within its borders. GDP is about location; GNP is about ownership.
• Net domestic product (NDP) is GDP minus depreciation (the wearing out of capital), measuring output net of the capital used up making it.
• Real GDP per capita - real GDP divided by population - is the single best rough gauge of average living standards, and the focus of the growth module.

GNP arithmetic is a one-line adjustment: GNP = GDP + factor income residents earn abroad - factor income foreigners earn here. Suppose GDP is 9,000, residents earn 250 from factories and investments overseas, and foreign owners earn 150 from operations here. Then GNP = 9,000 + 250 - 150 = 9,100. For most large economies the two measures sit close together, but for countries hosting heavy foreign investment, or sending many workers abroad whose earnings flow home, the gap can be substantial and the choice of measure genuinely changes the story.

NDP answers a different question: how much output could the economy consume without eating its capital? With GDP of 9,000 and depreciation of 900, NDP is 8,100. An economy that consumed all 9,000 would end the year with a diminished capital stock - like a farmer eating the seed corn. This is also why the "G" in GDP matters: gross means before subtracting depreciation, and gross investment can overstate how much productive capacity is actually growing.

From GDP to take-home income

A chain of subtractions and additions connects headline GDP to what households can actually spend. From national income, subtract earnings that never reach households, such as profits firms retain, and add income received without current production, such as government transfers. The result is personal income. Subtract personal taxes and what remains is disposable income, the budget households divide between consumption and saving.

A compact hypothetical: national income 7,800; firms retain 600 of profits; transfers add 800; personal income is 7,800 - 600 + 800 = 8,000. Personal taxes of 1,500 leave disposable income of 6,500. That 6,500 is the wallet-level number: the link between the production statistics of this module and the consumption behavior that drives aggregate demand later in the course.

What GDP leaves out

GDP is the headline measure of an economy's size, but its designers never intended it as a measure of welfare, and it misses a great deal:

• Non-market production - unpaid housework, childcare, and volunteering produce real value but are not sold, so GDP ignores them.
• The underground economy - legal work paid in cash and illegal activity go largely uncounted.
• Distribution - GDP is a total; it says nothing about whether the income is shared widely or concentrated in a few hands.
• Leisure and non-material well-being - more output produced by working longer hours may not mean people are better off.
• Environmental costs - GDP counts the output from polluting activity but not the damage to air, water, or climate; it can even rise when we spend to clean up disasters.

The non-market omission produces genuinely odd bookkeeping. If two neighbors each watch their own children, GDP records nothing. If they hire each other as babysitters at $200 a week apiece, GDP rises by $400 a week - with not one additional hour of childcare performed. Nothing real changed; production simply crossed the market boundary and became visible. The same applies when a home-cooked dinner is replaced by takeout: measured GDP rises by more than true production does.

Leisure creates a mirror-image blind spot. Imagine two countries with identical real GDP per capita of 20,000. In country A the average worker puts in 1,600 hours a year; in country B, 2,000 hours. B's citizens surrender an extra 400 hours - ten standard workweeks - to produce the same measured income. By any sensible reckoning A's people are better off, yet GDP scores the two as equals. Output statistics count what is made, never the time it cost.

Distribution blindness matters for interpretation. GDP per capita is a mean, and means are pulled up by large values at the top. Two economies can share a GDP per capita of 20,000 while the typical household in one earns close to that figure and the typical household in the other earns far less amid a wealthy few. Whenever the question is how the middle of society lives, median income and distribution data must supplement the aggregate.

Measuring what GDP misses

Statisticians have not shrugged at these gaps. The United Nations' Human Development Index (HDI) combines income with life expectancy and schooling, reranking countries that deliver health and education well or poorly for their income level. Agencies publish satellite accounts that estimate household production and environmental depletion alongside the standard accounts. None of these has displaced GDP, but together they put numbers on several of its silences.

The warnings began at the source. Simon Kuznets, who built the accounts, told Congress in 1934 that "the welfare of a nation can scarcely be inferred from a measurement of national income." The people most fluent in GDP have always been the quickest to say what it is not.

Why we still use it

Despite these gaps, GDP remains indispensable. It correlates strongly with things people plainly value - life expectancy, literacy, and access to goods and services all tend to rise with real GDP per capita. The mature view is to treat GDP as a powerful but partial gauge: excellent for tracking the size and cycles of production, but best paired with other indicators when the question is human well-being rather than economic output.

History explains the design. The accounts were built in the 1930s to answer "how deep is the Depression?" and expanded during the Second World War to answer "how many tanks and planes can the economy produce without starving civilians?" Both are production questions, and GDP answers them superbly. It was drafted as a war-and-slump instrument, not a happiness meter, and it has never pretended otherwise.

GDP also has practical virtues no rival matches. It is available quarterly, computed by a common international rulebook, and comparable across decades and borders, which is what lets us date recessions, compare growth records, and test theories against evidence. A measure that captured well-being perfectly but arrived years late, or meant different things in different countries, could not do this job. The working rule: use GDP for what it measures, production; reach for companions when the question is how life is going.

Common misconceptions

• "The income approach measures a different economy." No - same economy, same total, opposite side of every transaction. Differences between the two tallies are measurement noise, not concept.
• "Depreciation is money set aside somewhere." It is an accounting estimate of capital worn out, not a cash pile; that is why it is income to no one.
• "GDP per capita is what the typical person earns." It is the average, which top incomes can pull far above the median.
• "If GDP omits something, economists ignore it." Household production, environmental damage, and distribution all have dedicated measures; they simply live outside the headline number.

Try it

An economy reports, in billions: compensation of employees 4,200; rents 150; interest 350; proprietors' income 500; corporate profits 800; depreciation 700; indirect business taxes 300. Residents earn 100 more in factor income from abroad than foreigners earn domestically. (a) Compute national income and GDP by the income approach. (b) Compute NDP. (c) Compute GNP.

Answer: (a) National income = 4,200 + 150 + 350 + 500 + 800 = 6,000. GDP = 6,000 + 700 + 300 = 7,000. (b) NDP = GDP - depreciation = 7,000 - 700 = 6,300. (c) GNP = GDP + net factor income from abroad = 7,000 + 100 = 7,100. The expenditure approach, measured well, would also deliver 7,000 - one loop, two ledgers.

Sources

1. Greenlaw, S. A., Shapiro, D., & MacDonald, D. (2022). Measuring the size of the economy: Gross domestic product. In Principles of Macroeconomics 3e. OpenStax, Rice University. openstax.org
2. Greenlaw, S. A., Shapiro, D., & MacDonald, D. (2022). How well GDP measures the well-being of society. In Principles of Macroeconomics 3e. OpenStax, Rice University. openstax.org
3. Kuznets, S. (1934). National income, 1929-1932. In NBER Books and Chapters. National Bureau of Economic Research. nber.org
4. U.S. Bureau of Economic Analysis. (n.d.). NIPA handbook: Concepts and methods of the U.S. national income and product accounts. In BEA Methodologies. U.S. Department of Commerce. bea.gov
5. U.S. Bureau of Economic Analysis. (n.d.). Personal income. In National Data: Income and Saving. U.S. Department of Commerce. bea.gov
6. Anderson, L. (2008). Gross domestic product. In The Concise Encyclopedia of Economics (2nd ed.). Liberty Fund. econlib.org
7. United Nations Development Programme. (n.d.). Human Development Index. In Human Development Reports. United Nations. hdr.undp.org

Why we need two GDPs

Because GDP is measured in money, it can rise for two very different reasons: the economy produces more stuff, or the same stuff costs more. To track genuine changes in output we must separate quantity from price. Nominal GDP values output at current-year prices, so it mixes the two. Real GDP values output at the prices of a fixed base year, holding prices constant so that any change reflects a change in real quantities produced.

Think of money as a measuring tape that stretches. If a tailor's tape lengthened 5 percent a year, every "measurement" would creep upward even if customers never grew. Dollar-denominated statistics have exactly this defect whenever prices are rising. Real GDP fixes the tape: by valuing every year's output at one year's prices, it turns the elastic ruler into a rigid one, so a longer reading finally means a bigger economy.

The distinction has teeth in history. In the United States during the 1970s, nominal GDP roughly tripled from the start of the decade to its end - which sounds like a production miracle. Most of that climb was inflation. Real output grew modestly while the price level surged, and anyone reading nominal figures alone badly misjudged the era. Every serious statement about growth, recession, or living standards is a statement about real GDP.

Worked example: a simple economy

Imagine an economy that produces only bread. Let year 1 be the base year.

Year	Loaves produced	Price per loaf	Nominal GDP	Real GDP (base-year prices)
1 (base)	100	$10	$1,000	$1,000
2	120	$12	$1,440	$1,200

Nominal GDP jumped from $1,000 to $1,440, a rise of 44 percent - but part of that is just higher prices. Real GDP uses the base-year price of $10 for both years: year 2 real GDP is 120 loaves x $10 = $1,200. So real output rose from $1,000 to $1,200, a real gain of (1,200 - 1,000) / 1,000 = 20 percent. In the base year, nominal and real GDP are always equal by construction.

Extend the story one more year. In year 3 the economy bakes 140 loaves at a price of $13. Nominal GDP is 140 x $13 = $1,820; real GDP is 140 x $10 = $1,400. Real growth from year 2 to year 3 is (1,400 - 1,200) / 1,200, about 16.7 percent, even though the nominal jump from 1,440 to 1,820 is over 26 percent. Year after year, the recipe is identical: quantities from the year in question, prices frozen at the base year.

A two-good economy

One good makes the idea clear; two goods make it honest, because real economies change relative prices constantly. Suppose a country produces bread and cloth. In the base year it makes 100 loaves at $10 and 50 bolts of cloth at $20, so nominal GDP = 1,000 + 1,000 = $2,000, and real GDP is the same $2,000 by construction.

In year 2 it produces 120 loaves at $12 and 55 bolts at $22. Nominal GDP = (120 x 12) + (55 x 22) = 1,440 + 1,210 = $2,650. Real GDP at base-year prices = (120 x 10) + (55 x 20) = 1,200 + 1,100 = $2,300. Real growth is (2,300 - 2,000) / 2,000 = 15 percent, a weighted blend of bread's 20 percent quantity growth and cloth's 10 percent.

Notice what happened to prices: bread rose 20 percent and cloth 10 percent, so relative prices shifted. No single number describes "the" price change perfectly, which is why price indexes are always averages with weights. The deflator below weights by what the economy actually produces - and disagreements among indexes, like deflator versus CPI, are mostly disagreements about weights.

The GDP deflator

The gap between nominal and real GDP is itself a useful price measure. The GDP deflator is defined as:

GDP deflator = (Nominal GDP / Real GDP) x 100

For year 2: deflator = (1,440 / 1,200) x 100 = 120. The deflator equals 100 in the base year (since nominal equals real), so a reading of 120 means prices overall are 20 percent higher than in the base year - matching the price rise from $10 to $12. The deflator is a price index: a number that tracks the overall price level relative to a base of 100.

In the two-good economy, the year-2 deflator is (2,650 / 2,300) x 100, about 115.2: an overall price level 15.2 percent above base, sitting between bread's 20 percent rise and cloth's 10 percent because the index blends both goods. The deflator covers everything domestically produced - consumer goods, machine tools, government construction - which makes it the broadest price index the accounts can offer.

The name says what it does: it deflates. Given any nominal figure and the deflator, recover the real figure by dividing: real GDP = nominal GDP / (deflator / 100). If nominal GDP is $21,000 billion and the deflator is 105, real GDP = 21,000 / 1.05 = $20,000 billion. Journalists' habit of calling numbers "inflation-adjusted" refers to exactly this division.

Computing inflation from the deflator

The inflation rate between two years is the percentage change in a price index:

Inflation = ((Index this year - Index last year) / Index last year) x 100

If the deflator rises from 120 to 126, inflation is (126 - 120) / 120 = 6 / 120 = 5 percent. Any price index works the same way, which is exactly how the Consumer Price Index in the next lesson is used.

Two habits keep this calculation honest. First, divide by the earlier index, not the base of 100: from 120 to 126 is 5 percent, not 6. Second, remember inflation compounds like growth: a deflator moving 100, 104, 108.16 records 4 percent inflation in each year, and two years of 4 percent lift prices 8.16 percent, not 8. Small differences, but they accumulate across decades exactly the way growth rates do.

A falling index is also meaningful: a deflator sliding from 100 to 97 records deflation of 3 percent - the price level itself declining, as the United States experienced in the early 1930s. And watch the distinction between the index level and its change: a deflator of 130 with inflation of 1 percent means prices rose a lot in the past but are almost stable now.

Real GDP per person: the living-standards yardstick

For living standards, one more division is needed: real GDP per capita, real output divided by population. The subtraction logic of the growth shortcut applies again. If real GDP grows 3 percent while population grows 1 percent, output per person rises roughly 2 percent. If population grows 3 percent while real GDP grows 2, the average person is getting poorer even though "the economy is growing." Fast-growing populations need fast-growing output just to stand still - one reason headline GDP growth and improving daily life can part company.

Extreme price episodes show why the machinery matters rather than breaking it. In Weimar Germany in 1923, prices rose so fast that a price index reset its base repeatedly and wages were spent within hours. The formulas still worked - indexes and percentage changes are just arithmetic - but nominal figures became meaningless within days, and only quantities told anyone what the economy was really producing. Hyperinflation is the reductio ad absurdum of trusting nominal numbers.

A shortcut linking nominal growth, real growth, and inflation

For quick mental work, the three headline rates tie together in one approximation: real growth is roughly nominal growth minus inflation. If nominal GDP grows 5 percent while inflation runs 2 percent, real growth is close to 3 percent (the exact figure, 1.05 / 1.02 - 1, is 2.94 percent). The approximation is excellent for small rates and rough for large ones: in the bread economy, nominal growth of 44 percent minus inflation of 20 percent suggests 24, while the true real growth is 20. Use the shortcut for headlines; use the division for homework.

Choosing, and changing, the base year

The base year itself is an arbitrary anchor. Rebasing from year 1 prices to year 2 prices rescales every real GDP level, but well-measured growth rates barely move - and it is growth, not the level, that carries the economic story. Statistical agencies rebase periodically precisely because old price weights go stale: an economy's mix shifts over decades, and valuing today's output at the prices of a distant year gives odd weight to goods whose relative prices have collapsed, such as computing power.

Modern accounts therefore use chain-weighting: real growth each year is computed using price weights from adjacent years, and the yearly links are chained into a long series. The arithmetic is heavier, but the principle is the one you already know - hold prices fixed, compare quantities - applied with continuously refreshed weights instead of one aging base year.

Reading the news with the distinction in hand

Try the skill on a typical headline: "GDP grew 8 percent this year, the fastest on record, lifting living standards." Three questions expose the claim. Is that nominal or real growth? If inflation ran 6 percent, real growth was about 2 - ordinary, not record-setting. Is it total or per capita? With population up 2 percent, output per person barely moved. And is the base period unusual - a rebound from a deep slump flatters any growth figure. The vocabulary of this lesson is exactly the checklist for such claims.

Common misconceptions

• "Nominal GDP rose, so the economy grew." Not necessarily: nominal GDP can rise on prices alone, with output flat or even falling.
• "Real GDP is physical units." It is still dollars - quantities valued at fixed base-year prices - which is what lets us add bread to cloth at all.
• "Nominal and real being equal in the base year is a coincidence." It is by construction: both use the same prices that year, so the deflator is exactly 100.
• "The deflator and the CPI are the same index." They usually move together but differ by design: the deflator weights all domestic production, while the CPI prices a fixed consumer basket including imports - the next lesson's subject.
• "A falling deflator means output fell." It means the price level fell; real output can rise through a deflation, and the two must be read separately.

Try it

An economy makes one good. Year 1 (the base year): 200 units at $5 each. Year 2: 210 units at $5.50 each. Find year-2 nominal GDP, real GDP, the deflator, inflation since the base year, and the real growth rate.

Answer: Nominal = 210 x $5.50 = $1,155. Real = 210 x $5 = $1,050. Deflator = (1,155 / 1,050) x 100 = 110, so prices are 10 percent above the base year, and inflation since base is 10 percent. Real growth = (1,050 - 1,000) / 1,000 = 5 percent. As a check, the shortcut gives nominal growth 15.5 percent minus 10 percent inflation, approximately 5.5 - close to the exact 5.

Sources

1. Greenlaw, S. A., Shapiro, D., & MacDonald, D. (2022). Adjusting nominal values to real values. In Principles of Macroeconomics 3e. OpenStax, Rice University. openstax.org
2. Greenlaw, S. A., Shapiro, D., & MacDonald, D. (2022). Tracking real GDP over time. In Principles of Macroeconomics 3e. OpenStax, Rice University. openstax.org
3. Greenlaw, S. A., Shapiro, D., & MacDonald, D. (2022). Tracking inflation. In Principles of Macroeconomics 3e. OpenStax, Rice University. openstax.org
4. U.S. Bureau of Economic Analysis. (n.d.). GDP price deflator. In Prices and Inflation. U.S. Department of Commerce. bea.gov
5. World Bank. (n.d.). Inflation, GDP deflator (annual %). In World Development Indicators. World Bank Group. data.worldbank.org
6. Bryan, M. (2013). The Great Inflation. In Federal Reserve History. Federal Reserve Bank of St. Louis. federalreservehistory.org
7. Salemi, M. K. (2008). Hyperinflation. In The Concise Encyclopedia of Economics (2nd ed.). Liberty Fund. econlib.org

Pricing a typical household's basket

The most widely watched price measure is the Consumer Price Index (CPI), which tracks the cost of a fixed market basket of goods and services bought by a typical urban household - food, housing, transportation, medical care, and more. Statistical agencies survey households to fix the basket, then re-price that same basket month after month. Because the basket is held constant, changes in its cost reflect changes in prices, not in what people buy.

The machinery behind the headline is worth picturing. First, an expenditure survey asks thousands of households what they actually buy, which sets the basket's weights: imagine the survey finds a typical household devotes about a third of spending to shelter, a sixth to transportation, an eighth to food, and smaller shares to clothing, medical care, and recreation. Then, every month, price collectors record the prices of specific items - the same brands, sizes, and stores wherever possible - across many cities.

Those weights are the index's steering wheel. A 10 percent jump in a category absorbing a third of the budget moves the index by more than 3 points, while a 10 percent jump in something bought rarely barely registers. When shelter costs accelerate, the headline follows almost mechanically; when an exotic import doubles in price, hardly anyone's basket notices.

Keep the CPI's question distinct from the deflator's. The GDP deflator asks: what happened to the prices of everything we produce? The CPI asks: what happened to the cost of what a typical household buys? A household buys imported goods (excluded from the deflator) and does not buy factory robots or highway construction (included in the deflator). Same toolkit, different basket, different question.

A measure born of wage disputes

The CPI exists because someone needed to settle an argument about pay. During the First World War, prices in the United States surged and shipyard workers demanded raises to match. To arbitrate, the government's statisticians began systematically pricing what working families bought, and in 1919 the Bureau of Labor Statistics started publishing the result - the ancestor of today's index. The origin story is a reminder of what the CPI is for: it converts the vague complaint "everything costs more" into a number precise enough to write into a contract.

Building the index

The CPI expresses the basket's cost relative to a base period set at 100:

CPI = (Cost of basket this year / Cost of basket in base year) x 100

The procedure is always the same four steps. Fix the quantities from the survey. Price the full basket in the base period. Price the identical basket in the period of interest. Divide the two costs and multiply by 100. Everything else in this lesson - inflation rates, dollar conversions, real wages - is arithmetic built on that ratio.

Worked example

Suppose a household's basket contains 10 kilograms of rice and 5 shirts. Prices are:

Item	Quantity	Base-year (2015) price	Current-year (2020) price
Rice	10 kg	$2 / kg	$3 / kg
Shirts	5	$20 each	$24 each

Base-year basket cost = (10 x $2) + (5 x $20) = 20 + 100 = $120.
Current-year basket cost = (10 x $3) + (5 x $24) = 30 + 120 = $150.
CPI for 2020 = (150 / 120) x 100 = 125.

Since the base-year CPI is 100 by definition, a 2020 CPI of 125 means the basket costs 25 percent more than in 2015. The inflation rate from 2015 to 2020 is therefore (125 - 100) / 100 = 25 percent over that span. Year-over-year, if the CPI rises from 125 to 130, inflation for that year is (130 - 125) / 125 = 5 / 125 = 4 percent.

Push the example one period further. Suppose prices later reach $3.30 for rice and $25 for shirts. The basket then costs (10 x 3.30) + (5 x 25) = 33 + 125 = $158, and the CPI is (158 / 120) x 100, about 131.7. Inflation from the 125 reading to the 131.7 reading is (131.7 - 125) / 125, roughly 5.4 percent. Note the pattern of every calculation: quantities never change; only prices do. The moment you let quantities move, you are no longer measuring pure price change.

Notice also that the two goods inflated at very different rates from the base year - rice up 65 percent, shirts up 25 percent - while the index rose 31.7 percent, between the two and much closer to shirts' rate, because shirts absorb the larger share of the budget. An inflation rate is always a weighted average over a basket in which some prices are sprinting and others are standing still.

Using the CPI to compare dollars over time

The CPI lets us convert a past dollar amount into today's dollars to compare purchasing power:

Value in today's dollars = Past amount x (CPI today / CPI then)

A $10 wage when the CPI was 100, compared at a CPI of 125, is worth 10 x (125 / 100) = $12.50 in today's prices - so a nominal wage that stayed at $10 actually lost purchasing power.

The formula settles a whole family of everyday arguments. A grandparent's $0.75 hourly wage earned when the CPI stood at 25 converts, at a CPI of 250, to 0.75 x (250 / 25) = $7.50 in today's money - so a $15 wage today, far from being "the same old pay," buys about twice what the old wage did. Movie tickets, tuition, and house prices all need the same translation before any comparison across decades means anything.

The same arithmetic runs many contracts automatically. A cost-of-living adjustment (COLA) raises indexed payments in step with the CPI: a pension of $1,000 a month, indexed, becomes 1,000 x (126 / 120) = $1,050 when the CPI climbs from 120 to 126. Wage agreements, government benefits, and tax bracket thresholds are commonly tied to the index this way - which is precisely why small measurement biases in the CPI, discussed below, carry large budgetary consequences.

The conversion also runs backward. To express a current amount in past dollars, flip the ratio: value in past dollars = current amount x (CPI then / CPI today). A $25,000 car today, with the CPI at 250 versus 25 back then, corresponds to 25,000 x (25 / 250) = $2,500 in that era's money - the right figure to set beside the sticker prices in an old advertisement.

The real wage

Divide instead of multiply and you get the real wage: the nominal wage divided by (CPI / 100), expressed in base-year dollars. If your wage is $20 and the CPI is 125, your real wage is 20 / 1.25 = $16. Suppose a raise lifts the nominal wage to $21 while the CPI reaches 135. The real wage becomes 21 / 1.35, about $15.56: purchasing power fell despite the raise, because prices rose 8 percent while pay rose 5.

This is the single most practical habit the lesson offers: evaluate every dollar figure that spans time in real terms. Salary offers, loan payments, savings growth, government budgets - each is an elastic-ruler number until the CPI converts it.

Why the CPI overstates inflation

Because the basket is fixed, the CPI tends to overstate the true rise in the cost of living for several reasons:

• Substitution bias - when one good gets pricier, shoppers buy less of it and more of substitutes, but the fixed basket keeps the old quantities, overstating the pain.
• Quality change - if a product improves, part of a price rise pays for more quality, not pure inflation.
• New goods - genuinely new products enter the basket only with a lag, so early benefits are missed.
• Outlet substitution - shoppers migrate to discount stores and online sellers, but price collectors keep visiting the old outlets for consistency, missing part of the savings.

These biases matter because the CPI is used to adjust wages, pensions, and tax brackets. The GDP deflator (previous lesson) and the CPI usually move together but can differ, since the deflator covers all domestic output while the CPI covers a fixed consumer basket that includes imports.

The stakes were quantified in a famous episode. In 1996 a panel of economists chaired by Michael Boskin estimated that the US CPI overstated true cost-of-living inflation by roughly 1.1 percentage points per year. Compounded over decades of indexed pensions and brackets, an error that size sums to enormous amounts, and the report pushed agencies to adopt fixes: more frequent basket updates, quality adjustments, and a chained version of the index that lets weights track actual substitution between adjacent periods.

Core inflation and the index family

Alongside the headline CPI live purpose-built variants. Core inflation strips out food and energy - not because they are unimportant, but because their prices whipsaw with weather and world markets. A frost or an oil spike can dominate one month's headline reading and vanish the next; excluding them reveals the underlying trend policymakers can actually act on. A central bank cannot fix a drought.

Producer price indexes track prices at the factory gate, often foreshadowing consumer inflation months ahead. And each price index answers its own basket's question, so the honest reading is plural: headline CPI for the cost of living now, core for the trend, the deflator for all domestic production.

Common misconceptions

• "The CPI measures all prices." It prices a consumer basket. Machine tools, exports, and government construction belong to other indexes.
• "Inflation fell, so things are cheaper than last year." Lower inflation means prices rose more slowly. Only deflation - a negative rate - means the basket costs less.
• "The CPI is my inflation." It is an average household's. A renter who commutes by car and a homeowner who bikes face different baskets, and their personal inflation rates can differ sharply from the headline.
• "House and stock prices belong in the CPI." Assets are savings vehicles, not current consumption; shelter enters as the cost of housing services, akin to rent, not as the purchase price of the asset.

Try it

A basket holds 20 kilograms of flour and 8 movie tickets. Base year: flour $1.50 per kg, tickets $10. This year: flour $1.80 per kg, tickets $12. (a) Compute the base-year and current basket costs and this year's CPI. (b) What is inflation since the base year? (c) A salary of $2,200 per month in the base year is worth how much per month today, if it kept pace exactly?

Answer: (a) Base cost = (20 x 1.50) + (8 x 10) = 30 + 80 = $110. Current cost = (20 x 1.80) + (8 x 12) = 36 + 96 = $132. CPI = (132 / 110) x 100 = 120. (b) Inflation since base = (120 - 100) / 100 = 20 percent. (c) 2,200 x (120 / 100) = $2,640 per month. Anything less means the salary lost purchasing power even if the number on the paycheck grew.

Sources

1. Greenlaw, S. A., Shapiro, D., & MacDonald, D. (2022). How to measure changes in the cost of living. In Principles of Macroeconomics 3e. OpenStax, Rice University. openstax.org
2. Greenlaw, S. A., Shapiro, D., & MacDonald, D. (2022). Tracking inflation. In Principles of Macroeconomics 3e. OpenStax, Rice University. openstax.org
3. Greenlaw, S. A., Shapiro, D., & MacDonald, D. (2022). Indexing and its limitations. In Principles of Macroeconomics 3e. OpenStax, Rice University. openstax.org
4. Boskin, M. J., Dulberger, E. R., Gordon, R. J., Griliches, Z., & Jorgenson, D. W. (1998). Consumer prices, the Consumer Price Index, and the cost of living. In Journal of Economic Perspectives, 12(1), 3-26. American Economic Association. aeaweb.org
5. Federal Reserve Bank of Cleveland. (n.d.). Center for Inflation Research. In Cleveland Fed Research. Federal Reserve Bank of Cleveland. clevelandfed.org
6. World Bank. (n.d.). Inflation, consumer prices (annual %). In World Development Indicators. World Bank Group. data.worldbank.org
7. Board of Governors of the Federal Reserve System. (2025). Why does the Federal Reserve aim for inflation of 2 percent over the longer run? In FAQs: Economy, Jobs, and Prices. Federal Reserve Board. federalreserve.gov

Sorting the adult population

To measure unemployment, statistical agencies survey households and place every working-age adult into one of three boxes:

• Employed - people who did any paid work in the survey period (or were temporarily away from a job).
• Unemployed - people who have no job, are available to work, and have actively looked for work recently.
• Not in the labor force - everyone else: full-time students, retirees, homemakers, and those who are not looking.

The crucial grouping is the labor force, defined as the employed plus the unemployed - that is, everyone who is either working or actively seeking work. People not in the labor force are simply not counted in the numerator or denominator of the unemployment rate.

The data come from a survey, not from benefit rolls. In the United States, interviewers contact tens of thousands of households every month and ask about each adult's activity in a reference week; other countries run close cousins of the same survey. Because the questions are standardized, the classification follows strict rules rather than anyone's self-description. You are "employed" if you did any paid work at all that week, even one hour, and also if you were merely absent from a job you still hold - on vacation, ill, or on strike.

"Working-age population" itself is a defined term, not a casual phrase. It typically means the civilian, non-institutional population aged 16 and over: children are excluded, as are active-duty military personnel and people living in institutions such as prisons or long-term care facilities. Definitions differ slightly across countries - some cap the age range for participation statistics - which is one reason cross-country comparisons need footnotes read before conclusions.

The edge cases repay attention. A part-time barista who wants full-time work counts as employed. A laid-off worker waiting to be recalled counts as unemployed even without job hunting. A new graduate who has sent applications this month is unemployed; an identical graduate taking the summer off is not in the labor force. The box you land in depends on recent, observable behavior - working, searching, or neither - because behavior is what a survey can measure consistently across millions of people.

The two key formulas

Unemployment rate = (Unemployed / Labor force) x 100
Labor force = Employed + Unemployed
Labor force participation rate = (Labor force / Working-age population) x 100

Both rates are shares, but of different bases. The unemployment rate divides by the labor force: it asks what fraction of active participants cannot find work. The participation rate divides by the whole working-age population: it asks what fraction of adults are in the game at all. Reading them as a pair is the core skill of this lesson, because each can move for reasons the other cannot see.

Households versus payrolls: two surveys, two answers

Most countries actually field two separate employment surveys, and they can disagree. The household survey described above counts people and produces the unemployment and participation rates. A parallel establishment survey asks businesses how many names are on their payrolls, producing the monthly "jobs added" headline. The two measure different things: payrolls count jobs, so one person holding two jobs counts twice, while the self-employed and farm workers are missed entirely. In any given month the surveys can point in opposite directions, and analysts treat agreement between them as the real signal.

Worked example

Suppose a country has, in millions: 152 employed, 8 unemployed, and a working-age (adult) population of 250.

Labor force = 152 + 8 = 160 million
Unemployment rate = 8 / 160 = 0.05 = 5 percent
Labor force participation rate = 160 / 250 = 0.64 = 64 percent

Notice the 90 million people (250 - 160) who are neither employed nor looking do not appear in the unemployment rate at all. The unemployment rate is a share of the labor force, not of the whole population.

Now run a discouragement scenario. Times stay hard, and 2 million of the 8 million unemployed stop searching altogether. They reclassify as not in the labor force. Unemployed falls to 6 million, the labor force shrinks to 158 million, and the unemployment rate becomes 6 / 158, about 3.8 percent. The headline rate just fell from 5 percent to 3.8 percent while not a single job was created. The improvement is a mirage of classification: the participation rate quietly slid from 64 percent to 158 / 250 = 63.2 percent, and that slide is the tell.

The same logic runs in reverse during recoveries. When hiring picks up, people who had given up flood back into the labor force to search. The unemployed count can temporarily rise even as employment grows, nudging the rate up for the most encouraging of reasons. A rising rate with rising participation reads very differently from a rising rate with falling participation - which is why economists never read the unemployment rate alone.

The employment-population ratio

A third gauge dodges the classification problem entirely. The employment-population ratio is simply employed divided by working-age population: here, 152 / 250 = 60.8 percent. Because it never asks who is "searching," discouragement cannot distort it. In the scenario above, where the unemployment rate fell to 3.8 percent on pure discouragement, the employment-population ratio stayed exactly 60.8 percent - flatly reporting that no larger share of adults is working. When the unemployment rate and this ratio tell different stories, believe the ratio.

Why the rate can mislead

The headline number leaves out two important groups. Discouraged workers want a job but have given up actively looking, so they are counted as "not in the labor force" rather than unemployed - which pushes the measured rate down even though joblessness is real. The underemployed - people working part-time who want full-time work, or working far below their skill level - count as fully employed, so the official rate misses the shortfall in hours. For these reasons, agencies publish broader measures alongside the standard rate, and analysts watch the participation rate to see whether people are entering or leaving the labor force entirely.

The broader measures put numbers on the gap. Alongside the official rate, agencies publish alternatives that successively add discouraged workers, the "marginally attached" (people who want work and searched recently, just not recently enough), and involuntary part-timers. Continue the example: to the 8 million officially unemployed, add 2 million marginally attached and 6 million involuntary part-time workers. The broadest rate is then (8 + 2 + 6) / (160 + 2) = 16 / 162, about 9.9 percent - nearly double the 5 percent headline, measuring the same economy with a wider net.

Neither number is "the truth"; they answer different questions. The narrow rate tracks active, fully idle job seekers - the group most sensitive to the business cycle. The broad rate tracks the total shortfall of work people want. In a deep slump the two gap widely apart; near full employment they converge. Reading the family together, plus participation, gives the honest picture no single figure can.

The churn beneath a calm surface

The unemployment rate is a stock, but the labor market underneath is a torrent of flows. Every month, millions are hired while millions separate - quits, layoffs, retirements - and people stream between all three boxes. A steady 5 percent rate does not mean the same people are unemployed all year; it means the pool's inflow and outflow roughly balance, like a bathtub with the faucet and drain both open. This is the stock-flow distinction from Lesson 1 wearing work clothes.

Duration is the hidden variable. Two economies can both post 5 percent unemployment while differing profoundly: in one, spells are short - a churning market where most of the unemployed find work within weeks; in the other, a smaller group is stuck without work for a year or more. Long-duration unemployment does the deep damage - skills fade, networks thin, employers hesitate - so analysts track average spell length and the share unemployed longer than six months, not just the rate.

The mix of flows carries its own information. Separations split into quits and layoffs, and they move oppositely over the cycle: layoffs spike in downturns, while quits surge in strong markets because workers only jump when they expect another offer. A high quit rate is, oddly, a sign of confidence. In a steady labor market, hires roughly balance separations - the level of employment holds while the names constantly change.

Calibrating "high" and "low": some history

Settled history provides the scale markers. At the bottom of the Great Depression in 1933, roughly one American worker in four was unemployed - about 25 percent, the catastrophic benchmark. The severe recession of the early 1980s pushed the US rate above 10 percent, as did the recession that followed the 2008 financial crisis. Healthy stretches have seen rates in the neighborhood of 4 to 5 percent, and never zero: even the strongest labor markets carry frictional search, a fact the next lesson builds into the concept of a natural rate.

One reading habit completes the toolkit: never lean on a single month. Labor data are seasonally adjusted - summer hiring of students and holiday retail surges are stripped out - and still carry sampling noise, then get revised as late responses arrive. The craft is to read three-month averages and watch whether the rate, participation, and the employment-population ratio agree on the direction. One number, one month, proves almost nothing.

Common misconceptions

• "The rate counts everyone without a job." No - only those searching. Retirees, students, and caregivers without jobs are outside the labor force, not unemployed.
• "A falling rate is always good news." Not if it falls because searchers gave up. Check participation and the employment-population ratio before cheering.
• "The rate comes from unemployment benefit rolls." It comes from a household survey; people who never file for benefits are captured, and benefit recipients who stop searching are not counted as unemployed.
• "Part-timers count as partly unemployed." Any paid work makes you employed in the official rate; the shortfall in hours shows up only in the broader measures.

Try it

A country has a working-age population of 200 million, with 118 million employed and 12 million unemployed. (a) Find the labor force, the unemployment rate, the participation rate, and the employment-population ratio. (b) Now 4 million of the unemployed become discouraged and stop searching. Recompute all four measures and interpret.

Answer: (a) Labor force = 118 + 12 = 130 million. Unemployment rate = 12 / 130, about 9.2 percent. Participation = 130 / 200 = 65 percent. Employment-population ratio = 118 / 200 = 59 percent. (b) Unemployed falls to 8 million and the labor force to 126 million: the rate becomes 8 / 126, about 6.3 percent, and participation falls to 63 percent. The employment-population ratio stays 59 percent - employment never changed. The "improvement" from 9.2 to 6.3 percent is discouragement, not hiring, and the unmoved ratio exposes it.

Sources

1. Greenlaw, S. A., Shapiro, D., & MacDonald, D. (2022). How economists define and compute unemployment rate. In Principles of Macroeconomics 3e. OpenStax, Rice University. openstax.org
2. Greenlaw, S. A., Shapiro, D., & MacDonald, D. (2022). Patterns of unemployment. In Principles of Macroeconomics 3e. OpenStax, Rice University. openstax.org
3. Summers, L. H. (2008). Unemployment. In The Concise Encyclopedia of Economics (2nd ed.). Liberty Fund. econlib.org
4. World Bank. (n.d.). Unemployment, total (% of total labor force) (modeled ILO estimate). In World Development Indicators. World Bank Group. data.worldbank.org
5. Richardson, G. (2013). The Great Depression. In Federal Reserve History. Federal Reserve Bank of St. Louis. federalreservehistory.org
6. Sablik, T. (2013). Recession of 1981-82. In Federal Reserve History. Federal Reserve Bank of St. Louis. federalreservehistory.org
7. Rich, R. (2013). The Great Recession. In Federal Reserve History. Federal Reserve Bank of St. Louis. federalreservehistory.org

Three kinds of unemployment

Not all unemployment has the same cause or cure. Economists sort it into three types:

• Frictional unemployment - short-term joblessness as people move between jobs or enter the labor market. A new graduate searching for a first job, or someone who quit to find a better fit, is frictionally unemployed. It reflects the normal time it takes to match workers to jobs and is largely unavoidable, even healthy.
• Structural unemployment - longer-term joblessness from a mismatch between the skills or locations workers have and the jobs available. When technology or trade eliminates an industry, displaced workers may lack the skills the growing sectors need. Retraining and relocation help, but structural unemployment can persist.
• Cyclical unemployment - joblessness caused by the downturn phase of the business cycle. In a recession, total spending falls, firms cut production, and they lay off workers regardless of their skills. This is the type that macroeconomic policy most directly tries to reduce.

Frictional unemployment deserves its reputation as the benign member of the family. Search takes time because information is scarce: workers do not know every vacancy, employers do not know every candidate, and both sides benefit from looking rather than grabbing the first option. A graduate who spends six weeks finding a job that fits her skills produces more over a career than one who takes the first offer in a week. Some friction is the price of good matches - which is why policies that speed matching (job boards, employment services) are more sensible targets than friction itself.

Structural unemployment is the painful one, and history supplies vivid cases. Mechanization emptied agriculture over the twentieth century: a workforce that was heavily farm-based in 1900 became overwhelmingly urban within three generations. Switchboard operators vanished when calls dialed themselves; steel and textile towns hollowed out when production moved or automated in the late twentieth century. In each case the jobs did not come back, and the displaced workers' skills no longer matched the vacancies that existed - often in different industries, cities, or wage brackets entirely.

Cyclical unemployment is different in kind: nothing is wrong with the workers or their skills. The same machinist who was fully employable last year is idle this year because orders collapsed economy-wide. That is precisely why it responds to macroeconomic medicine - restore total spending and the jobs reappear - while structural unemployment shrugs at stimulus and demands retraining, relocation, and time.

Spotting the type: a field guide

Classification is a diagnostic skill. Try four workers. First: a nurse who moved cities with her family and is interviewing at three hospitals - frictional; the skills are wanted, only the match is pending. Second: a coal miner in a region where mines closed for good, with no comparable local employer - structural; the mismatch is between his skills and any surviving vacancy. Third: a hotel clerk laid off when a recession gutted travel spending - cyclical; the job returns when demand does. Fourth: a printer whose trade was automated away and who is retraining in his fifties - structural again, the technology variant.

The test is always the counterfactual: would this person be working if total spending were at capacity? If yes, the unemployment is cyclical. If no - because the search simply takes time, or because the skills and locations do not line up - it is frictional or structural, and it will survive any boom.

Matching the medicine to the diagnosis

The three-way split matters because each type answers to different policy. Frictional unemployment responds to better information and faster matching: public employment services, searchable vacancy databases, and credential transparency all shorten the gap between jobs. Structural unemployment responds to retraining programs, apprenticeships that rebuild skills toward actual vacancies, and help with relocation when the jobs have moved. Cyclical unemployment responds to macroeconomic stimulus - the monetary and fiscal tools of later modules - because its cause is a shortfall of total spending.

Mismatched medicine fails on both sides. Stimulating demand cannot re-employ a workforce whose skills no longer fit the vacancies; it mostly bids up wages and prices in the sectors already at capacity. Conversely, prescribing retraining to workers idled by a collapse in spending wastes resources teaching people skills they already have jobs waiting for, if only demand returns. Diagnosis before treatment is as binding a rule in macroeconomics as in medicine.

The natural rate of unemployment

Because frictional and structural unemployment are always present - there are always people between jobs and always some mismatch - the unemployment rate never falls to zero even in a healthy economy. The natural rate of unemployment is the sum of frictional and structural unemployment: the rate that prevails when the economy is producing at its sustainable capacity and cyclical unemployment is zero.

Natural rate = frictional + structural unemployment
Actual rate = natural rate + cyclical unemployment

When economists say the economy is at full employment, they do not mean zero unemployment; they mean the actual rate equals the natural rate, with no cyclical unemployment. During a boom the actual rate can dip below the natural rate (labor markets are unusually tight); in a recession it rises above it as cyclical unemployment appears.

The term "natural" was coined by Milton Friedman in 1968, with Edmund Phelps developing the same idea independently; both argued that the economy gravitates toward this rate in the long run no matter what monetary policy does, a claim the Phillips curve lesson examines closely. The word is slightly unfortunate: nothing about the rate is natural in the sense of fixed or optimal. It is simply the level of unemployment generated by the economy's real machinery of search and mismatch when spending is neither hot nor cold.

That machinery, and hence the natural rate itself, drifts over time. A younger workforce churns more (new entrants search longer), raising the natural rate; an older one lowers it. Generous unemployment benefits lengthen search; efficient online job matching shortens it. Occupational licensing and housing costs that block moving between cities raise mismatch; portable credentials lower it. Estimates of the natural rate are therefore fuzzy and revised constantly - policymakers steer by a buoy that moves.

Worked reasoning

Suppose an economy's natural rate is 4.5 percent (say 3 percent frictional and 1.5 percent structural) and the actual unemployment rate is 7 percent. Then cyclical unemployment is 7 - 4.5 = 2.5 percent, the part attributable to a weak economy. If a recovery restores full employment, that 2.5 points should disappear while the 4.5 percent natural rate remains. This decomposition tells policymakers how much of a jobs problem is a demand shortfall (treatable with stimulus) versus a structural issue (needing training and time).

Okun's law: putting a price on slack

Cyclical unemployment has a companion cost in lost production, and the economist Arthur Okun gave the relationship a rule of thumb in 1962. In its gap form, Okun's law says each percentage point of cyclical unemployment corresponds to roughly two percentage points of output lost relative to potential. The ratio exceeds one because slack hides in more places than the unemployment count: hours are cut, discouraged workers exit the labor force, and firms hoard underused staff.

Apply it to the example above. With cyclical unemployment of 2.5 points, output runs roughly 2 x 2.5 = 5 percent below potential. If potential GDP is 20,000 (in billions), the economy is producing about 19,000 - a shortfall of 1,000 billion in a single year, goods and services that are never made and never recovered. Multiply across the several years of a slump and the arithmetic explains why recessions dominate policy attention despite being temporary.

The costs of unemployment

Unemployment is costly beyond lost income. For the economy it means output is below potential - goods and services that are never produced. For individuals it can bring loss of skills, health problems, and hardship for families. These human and economic costs are why keeping cyclical unemployment low is a central goal of macroeconomic policy.

The individual costs compound with duration. Long spells erode specific skills and professional networks, and a lengthy gap on a resume itself becomes a barrier, a pattern economists call scarring: workers who lose jobs in deep recessions often earn measurably less for a decade or more afterward. Graduates who enter the labor market during a slump start on lower rungs and can trail their luckier peers for years. The damage, in other words, outlives the recession that caused it.

Public budgets feel the cost from both directions at once. Every unemployed worker is simultaneously a taxpayer who is not paying income and payroll taxes and, often, a recipient of unemployment benefits and assistance. Deficits therefore swell automatically in slumps - the automatic stabilizers of the fiscal module - which is a cushioning feature, but also a reminder that unemployment's bill lands on everyone, employed or not.

Persistence can even convert one type into another, a phenomenon called hysteresis. Western Europe in the 1980s is the classic case: unemployment that rose in the disinflationary recessions of the early decade stayed stubbornly high long after the recessions ended, as the long-term unemployed drifted out of effective competition for jobs. Cyclical unemployment, left standing too long, hardened into structural unemployment - one of the strongest arguments for treating deep slumps quickly rather than waiting out the cycle.

Common misconceptions

• "The natural rate is fixed, or optimal." Neither. It drifts with demographics, institutions, and technology, and an economy can prefer policies that lower it - better matching, portable skills - without any stimulus.
• "Full employment means everyone has a job." It means zero cyclical unemployment; millions remain between jobs or mismatched even then.
• "Structural unemployment reflects weak effort." It reflects a mismatch of skills and locations against vacancies. No amount of searching turns a displaced miner into a nurse without retraining.
• "Each unemployed person wears a type label." The three types are analytical lenses summing to a total, not badges; a given worker's spell can blend demand weakness with mismatch.

Try it

An economy has frictional unemployment of 3.2 percent, structural unemployment of 1.8 percent, and an actual unemployment rate of 8.5 percent. Potential GDP is 12,000 (in billions). (a) Find the natural rate and cyclical unemployment. (b) Using Okun's two-for-one rule, estimate actual output. (c) Classify: a game designer between studio jobs for a month, and a travel agent whose occupation was mostly automated away.

Answer: (a) Natural rate = 3.2 + 1.8 = 5 percent; cyclical = 8.5 - 5 = 3.5 points. (b) Output gap is roughly 2 x 3.5 = 7 percent below potential, so actual output is about 12,000 x 0.93 = 11,160 billion. (c) The designer is frictional - a normal search between matches. The travel agent is structural - the occupation itself shrank, so recovery of spending alone will not restore the old job.

Sources

1. Greenlaw, S. A., Shapiro, D., & MacDonald, D. (2022). What causes changes in unemployment over the short run. In Principles of Macroeconomics 3e. OpenStax, Rice University. openstax.org
2. Greenlaw, S. A., Shapiro, D., & MacDonald, D. (2022). What causes changes in unemployment over the long run. In Principles of Macroeconomics 3e. OpenStax, Rice University. openstax.org
3. Summers, L. H. (2008). Unemployment. In The Concise Encyclopedia of Economics (2nd ed.). Liberty Fund. econlib.org
4. Henderson, D. R. (Ed.). (2008). Arthur M. Okun. In The Concise Encyclopedia of Economics (2nd ed.). Liberty Fund. econlib.org
5. Henderson, D. R. (Ed.). (2008). Milton Friedman. In The Concise Encyclopedia of Economics (2nd ed.). Liberty Fund. econlib.org
6. Nobel Prize Outreach. (2006). Edmund S. Phelps: Facts. In The Sveriges Riksbank Prize in Economic Sciences. NobelPrize.org ↗. nobelprize.org
7. The CORE Team. (2017). Economic fluctuations and unemployment. In The Economy 1.0. CORE Economics Education. core-econ.org

Three words for changing prices

Inflation is a sustained increase in the overall price level - not a one-time jump in a single good, but a broad, ongoing rise measured by an index like the CPI. Two related terms matter: deflation is a sustained fall in the price level (a negative inflation rate), and disinflation is a slowdown in the rate of inflation (prices still rising, but more slowly). Note that a rise in one price - say gasoline - is not inflation unless the general level of prices is rising.

The single-price confusion is worth one more beat, because it dominates everyday conversation. If gasoline jumps 30 percent while eggs fall and electronics drift down, the relative price of gasoline rose: gas now trades for more of everything else. That is a market signal about oil, not inflation. Inflation is the tide that lifts the whole harbor - measured only by a weighted index across the full basket - while relative prices are the boats bobbing against each other. An index near zero can coexist with dramatic relative-price drama underneath.

Two causes: demand-pull and cost-push

Inflation can originate on either side of the market.

• Demand-pull inflation arises when total spending grows faster than the economy's ability to produce - "too much money chasing too few goods." A booming economy, rapid growth in the money supply, or a surge in government spending can pull prices up as buyers compete for limited output.
• Cost-push inflation arises when the cost of producing goods rises across the board, pushing prices up even without extra demand. A sharp jump in oil prices or wages, or a supply shock like a harvest failure, raises firms' costs and forces higher prices while output may actually fall.

History offers clean specimens of each. Demand-pull: the United States in the late 1960s, when heavy government spending on the Vietnam War and new social programs piled onto an economy already near capacity, and inflation ratcheted upward year after year. Cost-push: the oil embargo of 1973-74, when world oil prices roughly quadrupled in months, driving costs up across every industry that shipped, heated, or manufactured anything - prices rose while output fell. The two diagnoses matter because the remedies differ: demand-pull calls for cooling total spending, while cost-push presents the cruel choice explored in the AD-AS module.

A useful lens is the quantity theory of money, summarized by the equation of exchange M x V = P x Y, where M is the money supply, V its velocity (how often a dollar is spent), P the price level, and Y real output. If velocity and real output are roughly stable, then sustained growth in the money supply (M) translates into a higher price level (P). This is the core reason economists link persistent inflation to excessive money growth.

Run the equation with numbers. Let M = 500, V = 4, so total spending M x V = 2,000. If real output Y is 1,000 units, the price level must be P = 2,000 / 1,000 = 2. Now let the money supply grow 10 percent to 550 while V and Y hold still: spending becomes 2,200, and P = 2,200 / 1,000 = 2.20 - exactly 10 percent inflation. The mechanism is that simple in the long run: money growth beyond output growth shows up in prices. Milton Friedman's famous summary was that inflation is "always and everywhere a monetary phenomenon."

In the short run the story is looser - velocity wobbles, and output responds to spending as the AD-AS module shows - so the quantity theory is best treated as a long-run anchor. Sustained double-digit inflation has never happened without rapid money growth, and no economy has stopped such inflation without reining that growth in.

Hyperinflation: the extreme laboratory

Hyperinflation - conventionally, inflation above 50 percent per month - is where the logic shows itself without disguise. Weimar Germany in 1923 is the textbook case: the government, crushed by deficits and reparations, printed money to pay its bills until prices doubled every few days. Workers were paid twice daily and spent within the hour; the famous photographs show banknotes hauled in wheelbarrows and used as wallpaper. Hungary in 1946 recorded the fastest inflation ever measured, with prices doubling in well under a day at the peak. Zimbabwe in 2008 reached the point where prices roughly doubled daily before the currency was abandoned entirely.

Every hyperinflation in history shares one anatomy: a government that cannot tax or borrow enough to cover spending and turns to the printing press. The inflation itself then destroys the tax base - revenues lose value between assessment and payment - forcing even faster printing. And the cures rhyme too: a credible fiscal reform plus a new or re-anchored currency. Germany's Rentenmark stabilization in late 1923 stopped the spiral within weeks once the printing genuinely ended, a speed that stunned observers and still anchors the case that inflation expectations respond to believed policy, not just to slogans.

The laboratory lesson for ordinary times: moderate inflation has many contributing causes, but sustained extreme inflation is a fiscal-monetary choice, and it ends when the choice changes.

Deflation is not the cure

If inflation is costly, falling prices might sound like the remedy. History argues otherwise. Deflation raises the real burden of every debt fixed in dollars: a farmer who borrowed against 100-dollar wheat must repay in dollars that now take twice the wheat to earn. The economist Irving Fisher, writing in 1933, described this debt-deflation spiral at the heart of the Great Depression: falling prices bankrupted borrowers, whose defaults weakened banks, whose failures shrank spending and pushed prices lower still.

Deflation also invites waiting - why buy today what will be cheaper next month? - and it paralyzes interest-rate policy, since rates cannot fall much below zero. This is why most central banks aim for a small positive inflation rate, commonly around 2 percent, rather than zero: a cushion above the deflationary cliff, with room left to cut rates in a slump.

Why inflation is costly

Mild, predictable inflation is manageable, but inflation imposes real costs:

• Reduced purchasing power - each dollar buys less, so anyone whose income does not keep pace falls behind.
• Menu costs - the resources firms spend constantly changing prices, catalogs, and menus.
• Shoe-leather costs - the effort of economizing on cash that is losing value.
• Distorted decisions - when prices are noisy, it is harder to tell real changes from inflation, muddling saving and investment choices.

Two further costs deserve names. The inflation tax falls on everyone holding cash: when the government finances itself by printing, the new money's purchasing power is quietly transferred from existing cash holders to the printer - taxation without legislation, and the engine of every hyperinflation. Bracket creep (or fiscal drag) strikes when tax brackets are fixed in nominal terms: a worker whose wage merely keeps pace with inflation gets pushed into higher brackets, paying a larger share of an unchanged real income. Indexing brackets to the CPI, now common, was adopted precisely to stop this.

The deepest cost is subtler: inflation degrades money's job as a unit of account. Contracts, price comparisons, and long-term plans all assume the measuring stick holds still. At high inflation, firms shorten contracts, lenders refuse long maturities, and energy flows into forecasting prices rather than producing goods. Economies with chronic high inflation visibly invest less in the future, because the future is written in a currency nobody trusts.

Winners and losers

Inflation redistributes wealth, especially when it is unexpected. Borrowers with fixed-rate loans gain, because they repay in dollars worth less than the ones they borrowed, while lenders lose for the same reason. People on fixed incomes or holding cash lose purchasing power. This is why economists distinguish the nominal interest rate (the stated rate) from the real interest rate, which subtracts inflation:

Real interest rate = Nominal interest rate - Inflation rate

If a bank pays 5 percent nominal interest and inflation is 3 percent, the real return is only 2 percent. If inflation unexpectedly hits 6 percent, the real return is negative 1 percent - savers actually lose ground. Because inflation shifts wealth between groups and clouds decisions, keeping it low and stable is a core policy goal.

Lenders defend themselves by pricing inflation in advance, a relationship named the Fisher equation: nominal rate = desired real rate + expected inflation. A lender wanting a 3 percent real return who expects 4 percent inflation quotes 7 percent. The redistribution therefore rides on forecast errors: if inflation lands at 6 percent instead of 4, the realized real return is 7 - 6 = 1 percent, and the surprise transferred wealth to the borrower. If inflation had come in at 2, the lender would have pocketed a 5 percent real return instead. Expected inflation is priced; unexpected inflation redistributes.

One quiet implication: governments are the largest fixed-rate borrowers of all, so surprise inflation erodes public debt in real terms - a temptation policymakers have felt in every indebted era, and a key reason independent central banks exist. Savers can shelter in inflation-indexed bonds, whose payments ride the CPI, converting the inflation gamble back into a guaranteed real return.

Common misconceptions

• "Inflation makes everyone poorer." Inflation redistributes more than it impoverishes: wages and incomes are prices too, and they generally rise as well. The losses concentrate on cash holders, fixed incomes, and surprised lenders.
• "A spike in one vital good is inflation." One price is a relative-price move; inflation is the sustained rise of the weighted basket.
• "Rising wages prove workers cause inflation." Wages chase prices as much as they push them; a wage-price spiral requires accommodating money growth to persist.
• "Zero inflation is the obvious ideal." Zero sits one misstep from deflation, leaves no room to cut real interest rates in a slump, and collides with the CPI's known upward bias - hence the common target of about 2 percent instead.

Try it

An economy has money supply M = 800, velocity V = 5, and real output Y = 2,000 units. (a) Find the price level. (b) The money supply grows 8 percent while V and Y are unchanged; find the new price level and the inflation rate. (c) A lender quoted a 5 percent nominal rate expecting 2 percent inflation. Given the inflation in (b), what real return did the lender expect, and what was realized? Who gained?

Answer: (a) P = (M x V) / Y = 4,000 / 2,000 = 2. (b) New M = 864, so P = (864 x 5) / 2,000 = 4,320 / 2,000 = 2.16, an inflation rate of 8 percent. (c) Expected real return = 5 - 2 = 3 percent; realized = 5 - 8 = minus 3 percent. The borrower gained: repayment dollars lost purchasing power faster than the interest compensated, a transfer caused entirely by the inflation surprise.

Sources

1. Greenlaw, S. A., Shapiro, D., & MacDonald, D. (2022). Tracking inflation. In Principles of Macroeconomics 3e. OpenStax, Rice University. openstax.org
2. Greenlaw, S. A., Shapiro, D., & MacDonald, D. (2022). The confusion over inflation. In Principles of Macroeconomics 3e. OpenStax, Rice University. openstax.org
3. White, L. H. (2008). Inflation. In The Concise Encyclopedia of Economics (2nd ed.). Liberty Fund. econlib.org
4. Salemi, M. K. (2008). Hyperinflation. In The Concise Encyclopedia of Economics (2nd ed.). Liberty Fund. econlib.org
5. Bryan, M. (2013). The Great Inflation. In Federal Reserve History. Federal Reserve Bank of St. Louis. federalreservehistory.org
6. Henderson, D. R. (Ed.). (2008). Irving Fisher. In The Concise Encyclopedia of Economics (2nd ed.). Liberty Fund. econlib.org
7. Board of Governors of the Federal Reserve System. (2025). Why does the Federal Reserve aim for inflation of 2 percent over the longer run? In FAQs: Economy, Jobs, and Prices. Federal Reserve Board. federalreserve.gov

A short-run trade-off

Two of our headline indicators - unemployment and inflation - turn out to be linked in the short run. The Phillips curve describes an inverse relationship between them: periods of low unemployment tend to come with higher inflation, and periods of high unemployment with lower inflation. The intuition connects directly to the aggregate demand and aggregate supply model from the previous topics. When aggregate demand rises, output expands and unemployment falls, but the tighter economy also pushes prices up - so lower unemployment arrives together with higher inflation. A demand-driven boom trades one for the other.

The figure captures the idea in two dots. The red point pairs low unemployment with high inflation - the boom corner. The green point pairs high unemployment with low inflation - the slump corner. A demand expansion slides the economy up and left along the curve; a demand contraction slides it down and right. Everything in this lesson is about when that sliding works, and when the curve itself refuses to sit still.

The mechanism runs through the labor market. When unemployment is low, workers are scarce: firms bid against each other for staff, wages climb, and firms pass the higher costs into prices. When unemployment is high, job seekers outnumber vacancies, wage demands soften, and price pressure fades. The curve is the labor market's thermometer reading translated into inflation.

The relationship carries a discoverer's name. In 1958 the New Zealand economist A. W. Phillips, working in London, plotted nearly a century of British data (1861 to 1957) and found unemployment and wage growth tracing a stable downward-sloping curve. Two years later, Paul Samuelson and Robert Solow reproduced the pattern in US data, restated it with price inflation on the axis, and christened it the Phillips curve. The data fit so well for so long that it entered textbooks as something close to a law.

Reading the curve as a policy menu - and the temptation

Treated as stable, the curve reads like a menu. Suppose the short-run relation in some economy is: inflation = 7 - unemployment (both in percent). Then 6 percent unemployment comes with 1 percent inflation, 5 percent unemployment with 2 percent inflation, and 3 percent unemployment with 4 percent inflation. A government could seemingly choose its favorite combination and buy permanently lower unemployment by accepting permanently higher inflation - just slide along the curve with a bit more stimulus.

Policymakers of the 1960s drew exactly that conclusion, and for a while the data cooperated: the United States spent the decade's second half riding the curve leftward, with unemployment falling and inflation creeping up. The menu interpretation was about to fail spectacularly, and two economists said so in advance.

Why the trade-off breaks down in the long run

The short-run trade-off tempted policymakers to think they could permanently "buy" lower unemployment by tolerating a little more inflation. Experience, especially the stagflation of the 1970s when high unemployment and high inflation appeared together, showed this does not hold for long.

The modern view is that there is a long-run Phillips curve that is vertical at the natural rate of unemployment. In the long run, the economy returns to its natural rate no matter what the inflation rate is - just as the long-run aggregate supply curve is vertical at potential output. You cannot lower unemployment below its natural rate permanently simply by accepting more inflation.

The prediction came before the failure, which is why it ranks among economics' great forecasting successes. In his 1968 presidential address to the American Economic Association, Milton Friedman argued that the trade-off would evaporate once people adjusted their expectations to persistent inflation; Edmund Phelps reached the same conclusion independently. Within a few years the 1970s delivered the test: oil shocks and years of accommodating policy produced inflation and unemployment rising together - points far off any 1960s-vintage curve. Plotted decade by decade, the tidy line dissolved into loops and clouds, exactly as the expectations argument implied.

The role of expectations

The bridge between the short run and the long run is inflation expectations. Suppose the central bank keeps stimulating demand to hold unemployment below the natural rate. At first, inflation rises and unemployment falls along a short-run Phillips curve. But once workers and firms come to expect higher inflation, they build it into wage demands and price setting.

That shifts the entire short-run Phillips curve upward: the same low unemployment now requires even higher inflation to sustain. Push further and the process repeats, producing accelerating inflation with no lasting gain in employment. This is why credible, stable expectations are so prized - they keep the short-run trade-off favorable and inflation anchored.

The logic compresses into one equation, the expectations-augmented Phillips curve: inflation = expected inflation + a x (natural rate - actual unemployment) + supply shocks. Run it with numbers. Let expected inflation be 2 percent, the natural rate 5 percent, and a = 0.5. Holding unemployment at 3 percent - two points below natural - delivers inflation of 2 + 0.5 x 2 = 3 percent.

Now let time pass. If people come to expect 3, the same policy delivers 4; expect 4, it delivers 5. Inflation ratchets upward period after period while unemployment never durably improves. Economists call this the accelerationist result: holding unemployment below the natural rate buys not permanently high inflation but permanently rising inflation - a treadmill, not a trade.

How fast expectations adjust depends on how they are formed. Under adaptive expectations, people project recent inflation forward, and the ratchet grinds upward gradually. Under rational expectations - associated with Robert Lucas - people anticipate policy itself, and the trade-off can vanish almost immediately once stimulus becomes predictable. Either way the destination is the same vertical long-run curve; only the speed differs.

Supply shocks: when both evils rise together

Expectations are not the only force that moves the curve. A negative supply shock - the oil embargo of 1973-74 is the canonical case - raises firms' costs at every level of unemployment, shifting the entire short-run Phillips curve upward in one jolt. The economy then suffers higher inflation and higher unemployment simultaneously: stagflation, the combination the original curve declared impossible.

This distinction is the key to reading any inflation episode. Demand movements slide the economy along a given short-run curve: inflation and unemployment move in opposite directions. Supply movements and expectation changes shift the curve: the two can rise or fall together. Whenever inflation and unemployment climb in tandem, look for a supply shock or unmoored expectations, not a failure of arithmetic.

The Volcker disinflation: expectations meet resolve

The theory faced its hardest practical test in reverse: how to get high expected inflation back down. By 1980, US inflation ran near 13 percent and was embedded in every wage contract and loan. Federal Reserve chair Paul Volcker, appointed in 1979, tightened money drastically and let interest rates soar. The result was the deep recession of the early 1980s, with unemployment peaking above 10 percent in 1982 - and inflation collapsing to the 3 to 4 percent range by 1983.

The episode put a price tag on lost credibility, which economists measure with the sacrifice ratio: the cumulative extra unemployment (or lost output) needed per point of disinflation. Cutting inflation by roughly 9 points at a cost of several point-years of high unemployment implied a substantial ratio - yet once the new low-inflation regime became believed, unemployment returned to its natural rate while inflation stayed low for decades. The long-run curve was vertical after all, and the 1980s walked down it the hard way. Later disinflations in countries with credible central banks proved cheaper, evidence that believed announcements shrink the sacrifice.

What it means for policy

The Phillips curve reframes the central banker's job. In the short run, there is a genuine trade-off, so policy can lean against a recession or an inflationary boom. In the long run, there is no trade-off to exploit - only a natural rate set by real factors - so the best a central bank can do for prices is keep inflation low and expectations well anchored. This distinction unites the whole course: aggregate demand policy manages the short-run cycle, while long-run outcomes for output and employment depend on growth and the economy's real structure.

Modern practice is built on that division of labor. Central banks announce explicit inflation targets and defend them publicly, precisely to pin expectations in place; with expectations anchored, temporary shocks move inflation briefly without starting a ratchet, and stabilization policy regains room to work. Supply shocks still hurt - they shift the short-run curve up and force a choice between tolerating inflation and tolerating unemployment - but an anchored economy absorbs them with far less drama than the 1970s version did.

A closing irony rewards attention. In the decades after central banks adopted explicit targets and anchoring succeeded, the measured short-run trade-off grew noticeably flatter: unemployment swung while inflation barely budged. Economists debate how much of that flatness is the anchor doing its job - expectations refusing to move - and how much reflects deeper structural change. The debate itself is the lesson: the Phillips curve is not a physical constant but a mirror of how firmly people believe the monetary regime.

Common misconceptions

• "The Phillips curve was proven wrong, full stop." The menu reading died. The short-run relation, augmented with expectations and shocks, remains a workhorse of forecasting and policy.
• "Low unemployment mechanically causes inflation." The pressure runs through wages and expectations, and it operates relative to the natural rate - low unemployment with anchored expectations and adequate capacity need not ignite anything.
• "A vertical long-run curve makes policy pointless." Stabilizing the short run is valuable in itself, and keeping expectations anchored is a policy achievement, not a default.
• "Expectations are mere mood." They are forecasts written into wage contracts, loan rates, and price lists - which is exactly why they move slowly and matter so much.

Try it

An economy has a natural rate of 5 percent, expected inflation of 2 percent, and Phillips coefficient a = 1, so inflation = expected inflation + (5 - u). The central bank holds unemployment at 4 percent for two periods; expectations each period equal the previous period's actual inflation. (a) Find inflation in periods 1 and 2. (b) In period 3 the bank lets unemployment return to 5 percent. What is inflation? (c) What would returning inflation to 2 percent require?

Answer: (a) Period 1: 2 + (5 - 4) = 3 percent. Expectations rise to 3, so period 2: 3 + 1 = 4 percent. (b) With u back at the natural rate, inflation = expected inflation + 0 = 4 percent: the economy keeps the inflation but loses the employment gain. (c) Disinflation requires unemployment above 5 percent for a time - for example, u = 7 gives 4 - 2 = 2 percent in one harsh period - or a credibility gain that lowers expectations directly at a smaller sacrifice.

Sources

1. Greenlaw, S. A., Shapiro, D., & MacDonald, D. (2022). The Phillips curve. In Principles of Macroeconomics 3e. OpenStax, Rice University. openstax.org
2. Hoover, K. D. (2008). Phillips curve. In The Concise Encyclopedia of Economics (2nd ed.). Liberty Fund. econlib.org
3. Bryan, M. (2013). The Great Inflation. In Federal Reserve History. Federal Reserve Bank of St. Louis. federalreservehistory.org
4. Corbett, M. (2013). Oil shock of 1973-74. In Federal Reserve History. Federal Reserve Bank of St. Louis. federalreservehistory.org
5. Medley, B. (2013). Volcker's announcement of anti-inflation measures. In Federal Reserve History. Federal Reserve Bank of St. Louis. federalreservehistory.org
6. Sablik, T. (2013). Recession of 1981-82. In Federal Reserve History. Federal Reserve Bank of St. Louis. federalreservehistory.org
7. The CORE Team. (2017). Inflation, unemployment, and monetary policy. In The Economy 1.0. CORE Economics Education. core-econ.org

From one market to the whole economy

In microeconomics, a demand curve applies to a single good. In macroeconomics, aggregate demand (AD) is the total quantity of all final goods and services that households, firms, government, and foreigners want to buy at each overall price level. Its components are the same four that make up GDP: consumption (C), investment (I), government purchases (G), and net exports (NX). The AD curve slopes downward, plotting the price level on the vertical axis against real GDP demanded on the horizontal axis.

Do not let the familiar shape fool you: this is not a big market-demand curve, and the micro logic does not carry over. A single good's demand slopes down largely because buyers can switch to substitutes when its price rises. But when the whole price level rises, there is no "other economy" to switch to - every good got more expensive together. If the downward slope survives, it must survive for entirely different reasons, and finding them is this lesson's first job.

Keep the axes straight from the start. The vertical axis is the price level - a price index like the deflator - not the inflation rate. The horizontal axis is real GDP demanded, quantities rather than dollars. Each point on the curve answers a question: if the overall price level were this, how much total real output would buyers across the economy choose to purchase?

Why aggregate demand slopes down

The reasons differ from a single market, because when the whole price level changes there is no "other good" to substitute toward. Three effects explain the slope:

• Wealth effect - a lower price level raises the real value of money and savings, so people feel richer and buy more.
• Interest-rate effect - a lower price level reduces the need to hold cash, lowering interest rates, which encourages borrowing and investment.
• Exchange-rate effect - a lower domestic price level (and lower interest rates) makes home-produced goods cheaper relative to foreign goods, raising net exports.

Put small numbers on each. Wealth: a family holding $10,000 in cash and bank balances watches the price level fall 10 percent; the same $10,000 now buys about 11 percent more goods, and feeling wealthier, the family spends part of the gain. Multiply by millions of households and total consumption rises as the price level falls.

Interest rates: at a lower price level, transactions need less cash, so households and firms lend out the excess; more lending pushes interest rates down, and cheaper credit coaxes a builder to start the project that was marginal at the old rate. Exchange rates: with domestic prices (and rates) lower, foreign buyers find our goods relatively cheap, exports rise, and domestic buyers import a bit less - net exports strengthen. All three channels point the same way, which is why the curve leans down and to the right.

Just as important is the wrong reason: "prices fell, so people can afford more." Incomes are prices too - wages fall in the same general deflation - so affordability logic cancels out at the whole-economy level. The slope rests on the three named channels, not on everyone getting a discount.

What shifts aggregate demand

Anything that changes total spending at a given price level shifts the whole AD curve. Because AD = C + I + G + NX, a change in any component shifts it:

• Consumption - rises with higher consumer confidence, wealth, or lower taxes; this shifts AD right.
• Investment - rises with business optimism or lower interest rates; shifts AD right.
• Government purchases - more government spending shifts AD right (a lever of fiscal policy).
• Net exports - a boom abroad or a cheaper currency raises exports, shifting AD right.

Be careful to separate a movement along AD (caused by a change in the overall price level) from a shift of AD (caused by any of the factors above). This mirrors the micro distinction between a change in quantity demanded and a change in demand, applied to the whole economy.

Behind every shifter stands a story about decisions. Consumption leans on confidence and wealth: when home or stock values slump, households feel poorer and retrench even if incomes have not yet moved. Investment is the twitchiest component, driven by profit expectations and financing costs - Keynes famously credited the "animal spirits" of business sentiment. Government purchases move with legislation, the deliberate lever of fiscal policy. Net exports ride two external forces: income abroad (foreign booms buy more of our exports) and the exchange rate (a cheaper currency prices our goods into world markets).

The components are not equally jumpy. Consumption is the giant but a stable one: households smooth their spending, cutting reluctantly and gradually. Investment is a fraction of its size yet swings several times as hard in percentage terms, because expansion plans are postponable - a factory not built this year can wait - and because they lean on credit and confidence, both of which vanish together in a downturn. In most recessions, the fall in investment accounts for a share of the total demand decline far beyond its share of GDP.

Expectations deserve star billing because they move every component at once. A wave of optimism about future incomes lifts consumption today; expected sales growth triggers factory expansions today. Pessimism runs the machine in reverse, which is how a mood - if widely shared - becomes a measurable recession. The AD curve is, in this sense, the economy's confidence made visible.

A shock in numbers

First, see the curve itself in numbers. At a price level of 100, buyers demand C = 6,000, I = 1,500, G = 1,600, NX = 100, for total real demand of 9,200. Raise the price level to 110: the wealth channel trims consumption to 5,900, the interest-rate channel trims investment to 1,420, and the exchange-rate channel cuts net exports to 30, while G holds at 1,600. Total demand falls to 8,950. Those two points - (9,200 at 100) and (8,950 at 110) - are two spots on one downward-sloping AD curve. No shifter moved; only the price level did.

Now make a shift concrete. Suppose business pessimism cuts investment by 200 (in billions) at every price level, and each dollar of reduced spending ultimately removes 2.5 dollars of total demand through re-spending rounds. Then AD shifts left by 200 x 2.5 = 500 billion at each price level: at a price level of 100, demand drops from 9,200 to 8,700; at 110, from 8,950 to 8,450. The whole curve migrates. A tax cut or export boom of similar size would shift it right the same way.

AD is where policy grips the economy

Both arms of macroeconomic policy work by moving this one curve. Monetary policy operates through the interest-rate channel: when the central bank lowers rates, borrowing costs fall, investment and durable consumption rise, and AD shifts right; raising rates does the opposite. Fiscal policy operates directly - government purchases are a component of AD - and indirectly, since tax changes move disposable income and thus consumption.

The two levers differ in first-round strength. A 100 increase in G adds 100 of spending immediately. A 100 tax cut adds only what households spend of it - with an MPC of 0.75, the first round is 75, the rest saved - so a dollar of tax cut shifts AD somewhat less than a dollar of purchases. The full accounting, including multipliers and their limits, is the business of the monetary and fiscal modules; the AD curve is where all of it lands.

The spending multiplier preview

An initial change in spending can shift AD by more than the original amount, because one person's spending is another's income, which is partly spent again. This multiplier effect - explored fully in the fiscal-policy module - is why a change in investment or government spending can have an outsized effect on aggregate demand and output.

Watch one dollar ripple. With a marginal propensity to consume of 0.75, an initial 100 of new spending becomes income to its recipients, who spend 75; that becomes income to others, who spend 56.25; and so on. The geometric total is 100 x 1 / (1 - 0.75) = 400. In practice each round also leaks into saving, taxes, and imports, so real-world multipliers are smaller than the clean formula suggests - but the amplification is real, and it is why investment swings punch above their weight in recessions.

Demand shocks that made history

The great macroeconomic disasters are mostly AD stories. Between 1929 and 1933, US consumption and investment collapsed together: wealth evaporated in the stock crash, waves of bank failures wiped out deposits and credit, and the money supply shrank by roughly a third - a contraction Milton Friedman and Anna Schwartz placed at the center of the Great Depression. Output fell by about a third, and prices fell with it: the signature of demand collapsing along a supply curve.

The pattern repeated, smaller, in 2008-09: falling home prices destroyed household wealth, a financial panic froze credit, and consumption and investment dropped in tandem - AD shifting hard left. The reverse experiment exists too: the enormous military spending of the Second World War drove aggregate demand far rightward and finally absorbed the Depression's idle workers and factories. One curve, three episodes, both directions.

Common misconceptions

• "AD slopes down because cheaper goods are more affordable." Incomes deflate along with prices; the slope comes from the wealth, interest-rate, and exchange-rate channels.
• "A falling price level shifts AD right." A price-level change moves the economy along the curve. Only non-price forces - confidence, policy, wealth, foreign demand - shift it.
• "Imports rising means AD must fall." What matters is net exports and why trade moved: if imports rise because booming income pulls in goods, AD is shifting right for other reasons at the same time.
• "The multiplier makes any spending free." Amplification is not creation from nothing: leakages shrink it, and near capacity, extra demand raises prices more than output - the next lesson's central point.

Try it

Classify each event as a movement along AD or a shift (and its direction), then compute the last one. (a) The price level falls 5 percent. (b) A stock-market boom lifts household wealth. (c) A recession abroad cuts export orders. (d) The central bank raises interest rates sharply. (e) Investment falls by 150 (in billions) with an MPC of 0.8: how far does AD shift?

Answer: (a) Movement along the curve - the price level is the axis variable. (b) Shift right: wealthier households consume more at every price level. (c) Shift left: net exports fall. (d) Shift left: costlier credit cuts investment and durable consumption. (e) Multiplier = 1 / (1 - 0.8) = 5, so AD shifts left by 150 x 5 = 750 billion at each price level.

Sources

1. Greenlaw, S. A., Shapiro, D., & MacDonald, D. (2022). Building a model of aggregate demand and aggregate supply. In Principles of Macroeconomics 3e. OpenStax, Rice University. openstax.org
2. Greenlaw, S. A., Shapiro, D., & MacDonald, D. (2022). Shifts in aggregate demand. In Principles of Macroeconomics 3e. OpenStax, Rice University. openstax.org
3. Greenlaw, S. A., Shapiro, D., & MacDonald, D. (2022). The building blocks of Keynesian analysis. In Principles of Macroeconomics 3e. OpenStax, Rice University. openstax.org
4. Keynes, J. M. (1936). The state of long-term expectation. In The General Theory of Employment, Interest and Money. Macmillan. marxists.org
5. Richardson, G. (2013). The Great Depression. In Federal Reserve History. Federal Reserve Bank of St. Louis. federalreservehistory.org
6. Rich, R. (2013). The Great Recession. In Federal Reserve History. Federal Reserve Bank of St. Louis. federalreservehistory.org
7. Friedman, M., & Schwartz, A. J. (1963). A monetary history of the United States, 1867-1960. In NBER Book Series: Studies in Business Cycles. Princeton University Press. nber.org

Aggregate supply in two time frames

Aggregate supply (AS) is the total quantity of goods and services firms produce at each price level, and its shape depends on the time horizon.

The short-run aggregate supply (SRAS) curve slopes upward: in the short run, some costs (especially wages) are "sticky" and slow to adjust, so a higher price level lets firms sell at higher prices while paying the same wages, raising profits and output. As the price level rises, firms produce more.

The long-run aggregate supply (LRAS) curve is vertical at the economy's potential output (also called full-employment output). In the long run, all prices and wages adjust fully, so output is determined only by real factors - the labor force, the capital stock, and technology - not by the price level. You cannot produce more just because prices are higher once wages catch up.

The figure shows all three curves at once: the vertical LRAS planted at potential output, the upward-sloping SRAS, and the downward-sloping AD, meeting at a single point. That triple crossing is the economy at rest and at capacity. The drama of this lesson consists of knocking one curve out of place and watching where the crossing moves.

Why are wages sticky? Several reasons stack up. Union and employment contracts fix pay for a year or more. Renegotiating everyone's salary constantly carries its own menu costs. Minimum wages set legal floors. And employers themselves resist cutting pay in slumps because pay cuts poison morale and push the best workers out the door - the efficiency wage logic. Whatever the mix, wages move slower than prices, and that lag is the entire reason the short run differs from the long run in macroeconomics.

Sticky wages are one leg of the SRAS story; sticky prices are the other. Many firms print catalogs, sign supply contracts, and post menus that hold for months. When total spending rises, these fixed-price firms simply sell more at the old prices while others raise theirs - so the economy's average price level climbs a little and its output climbs a lot. Both stickiness stories deliver the same upward slope, and both wash out in the long run when everything reprices.

Each curve has its own shifters. SRAS shifts with production costs: a jump in energy or raw-material prices shifts it left, cheaper inputs or better productivity shift it right, and expected inflation shifts it left as workers pre-build raises into contracts. LRAS shifts only with the economy's real capacity - more workers, more capital, better technology - which is to say LRAS shifting right is economic growth, the subject of the next module. A quick test: if a change alters what firms can profitably produce at given wages, it moves SRAS; if it alters what the economy can sustain at full employment, it moves LRAS.

Macroeconomic equilibrium

The economy is in short-run equilibrium where AD crosses SRAS, setting the price level and real GDP. It is in long-run equilibrium when AD, SRAS, and LRAS all meet at potential output. The gap between actual output and potential output is the output gap:

• A recessionary gap exists when actual output is below potential - a slump, with high cyclical unemployment.
• An inflationary gap exists when actual output is above potential - an overheating boom that pushes prices up.

See the equilibrium logic in a schedule. Suppose buyers demand 9,600 of real output at a price level of 100, 9,000 at 110, and 8,400 at 120, while firms supply 8,400 at 100, 9,000 at 110, and 9,600 at 120. At a price level of 100 demand exceeds supply by 1,200: shelves empty, order backlogs build, and prices get bid upward. At 120 supply exceeds demand: inventories pile up and discounting spreads. Only at 110, with output 9,000, do plans mesh - the short-run equilibrium.

Whether 9,000 is good is a separate question, answered by comparing it to potential. If potential output is 9,300, this equilibrium leaves a recessionary gap of 300: the economy is at rest, but resting below capacity with workers idle. Equilibrium means balance of pressures, not health - a patient can be stable in the wrong bed.

From the gap to daily life

Gaps are not abstractions; each has a street-level feel. In a recessionary gap you see layoffs and hiring freezes, showrooms offering discounts, factories running single shifts, and "for lease" signs multiplying - the look of demand falling short of capacity. In an inflationary gap the signs reverse: help-wanted postings everywhere, signing bonuses, overtime, waiting lists for deliveries, and price increases arriving faster than usual. Learning to read a local economy against the diagram is half the value of the model.

Tracing shocks

The model earns its keep by showing what shocks do:

• A negative demand shock (say a collapse in confidence) shifts AD left: in the short run both the price level and output fall, opening a recessionary gap and raising unemployment. This is a typical recession.
• A positive demand shock shifts AD right: output and prices rise, and if pushed past potential, an inflationary gap and demand-pull inflation appear.
• A negative supply shock (an oil-price spike) shifts SRAS left: the price level rises while output falls - the painful combination of stagflation (stagnation plus inflation) that leaves policymakers with no easy fix.

Run a demand shock through the schedule from above. Let confidence collapse cut demand by 600 at every price level: buyers now want 9,000 at a price level of 100, 8,700 at 105, and 8,400 at 110. At the old equilibrium price of 110, firms would supply 9,000 but buyers take only 8,400 - inventories swell, discounting spreads, and the price level slides. The new crossing lands at a price level of 105 with output 8,700: both price and output lower, a recessionary gap of 600 against potential of 9,300, and cyclical unemployment rising in its train.

History files neatly into these three drawers. The Great Depression and the 2008-09 recession were negative demand shocks: spending collapsed, and output and prices fell together - the model's signature for AD moving left. The 1973-74 oil embargo was the negative supply shock: oil prices roughly quadrupled, SRAS lurched left, and the United States got the 1970s combination of rising prices and rising unemployment that demand-only reasoning could not produce.

The fourth drawer is the pleasant one: a positive supply shock. The American late 1990s approximated it, as information technology lifted productivity growth; SRAS pressed rightward, and the economy enjoyed fast growth, falling unemployment, and mild inflation simultaneously. When you see growth without price pressure, think supply improving, not demand surging.

You have met this geometry before in different clothes. Slide along an SRAS curve and output rises while the price level climbs - which, translated into rates, is lower unemployment arriving with higher inflation: the short-run Phillips curve of Module 5. The vertical LRAS is likewise the vertical long-run Phillips curve wearing output units instead of unemployment units. One economy, one logic, two diagrams.

Note the diagnostic rule the drawers imply: prices and output moving in the same direction point to a demand shock; moving in opposite directions, a supply shock. That single observation lets you classify most macroeconomic episodes before reaching for any further data.

Self-correction in the long run

The model also predicts recovery. In a recessionary gap, high unemployment eventually pushes wages and other costs down, which shifts SRAS right and returns output to potential - a slow, automatic self-correction. Whether to wait for this adjustment or to speed it up with active policy is one of the central debates of macroeconomics, and the subject of the policy modules ahead.

The mechanism runs symmetrically in both directions. Below potential, idle workers eventually accept lower wages, costs fall, and SRAS drifts right until output returns to potential at a lower price level. Above potential, scarce labor wins raises, costs climb, and SRAS drifts left until the boom cools at a higher price level. In both cases the economy ends where real capacity dictates; the detour changed only prices. This is exactly the long-run neutrality that the vertical LRAS asserts.

The catch is speed. Wage cuts are resisted fiercely, so the downward leg can take years, and Keynes aimed his most famous line at economists content to wait: "In the long run we are all dead." Worse, the waiting is not neutral - the Depression showed falling prices squeezing debtors and banks, deepening the very slump the adjustment was meant to cure. How long self-correction takes is the hinge on which the entire activist-versus-hands-off policy debate turns.

Classical and Keynesian readings of one diagram

The AD-AS diagram is a peace treaty between two traditions. The classical tradition, descending from the pre-Depression economists, emphasizes the long run: prices and wages adjust, the economy homes in on potential, and demand policy changes only the price level. The Keynesian tradition emphasizes the short run: wages are sticky, economies can idle below potential for years, and shifting AD is therefore genuinely powerful medicine.

Modern mainstream macroeconomics holds both, sorted by horizon: Keynesian in the short run, classical in the long run. That synthesis is why the same textbook can recommend stimulus in a deep recession and warn that chronic stimulus buys only inflation - the two claims live on different stretches of the same diagram.

Common misconceptions

• "A vertical LRAS means output can never rise." It means the price level cannot raise output permanently. LRAS itself shifts right every year the economy accumulates workers, capital, and ideas.
• "Equilibrium means the economy is fine." Short-run equilibrium can sit well below potential, with mass unemployment - balanced pressures, unhealthy position.
• "Falling prices are the painless cure for a slump." Deflation raises real debt burdens and can deepen the slump before the SRAS shift completes; the self-correction is slow and rough, not free.
• "Any shock can be offset by demand policy." Demand shocks, yes, in principle. A negative supply shock forces a choice: support output and accept more inflation, or fight inflation and accept a deeper slump.

Try it

An economy's potential output is 9,300. After a collapse in business confidence, the new short-run equilibrium is output 8,700 at price level 104. (a) Name the gap and its size. (b) Using Okun's two-for-one rule from Module 4, roughly how much cyclical unemployment does this imply? (c) Describe the self-correction path if policymakers do nothing. (d) What could policy do instead?

Answer: (a) A recessionary gap of 9,300 - 8,700 = 600, about 6.5 percent of potential. (b) Okun's rule attributes roughly half the percentage gap to cyclical unemployment: about 3.2 percentage points above the natural rate. (c) Unemployment gradually pushes wages and costs down, SRAS shifts right, and the economy returns to 9,300 at a price level below 104 - eventually. (d) Shift AD back right with lower interest rates or fiscal stimulus, closing the gap faster at the cost of forgoing the lower price level.

Sources

1. Greenlaw, S. A., Shapiro, D., & MacDonald, D. (2022). Building a model of aggregate demand and aggregate supply. In Principles of Macroeconomics 3e. OpenStax, Rice University. openstax.org
2. Greenlaw, S. A., Shapiro, D., & MacDonald, D. (2022). Shifts in aggregate supply. In Principles of Macroeconomics 3e. OpenStax, Rice University. openstax.org
3. Greenlaw, S. A., Shapiro, D., & MacDonald, D. (2022). How the AD/AS model incorporates growth, unemployment, and inflation. In Principles of Macroeconomics 3e. OpenStax, Rice University. openstax.org
4. Greenlaw, S. A., Shapiro, D., & MacDonald, D. (2022). Balancing Keynesian and neoclassical models. In Principles of Macroeconomics 3e. OpenStax, Rice University. openstax.org
5. Corbett, M. (2013). Oil shock of 1973-74. In Federal Reserve History. Federal Reserve Bank of St. Louis. federalreservehistory.org
6. Henderson, D. R. (Ed.). (2008). John Maynard Keynes. In The Concise Encyclopedia of Economics (2nd ed.). Liberty Fund. econlib.org
7. Field, A. J. (2008). Productivity. In The Concise Encyclopedia of Economics (2nd ed.). Liberty Fund. econlib.org

The most important number in the long run

Over years and decades, nothing matters more for living standards than economic growth - a sustained rise in real GDP, and especially in real GDP per capita. What looks like a small difference in the annual growth rate compounds into an enormous gap over a lifetime, which is why economists take even fractions of a percentage point seriously.

Zoom out far enough and growth is the strangest fact in economics: for most of recorded history there was essentially none. Century after century, average incomes crawled along near subsistence, and a farmer's grandchildren lived roughly as the grandparents had. Sustained growth only ignited with the Industrial Revolution in the early 1800s, first in Britain, then spreading - and in the two centuries since, living standards in the leading economies have multiplied more than twentyfold. Everything modern life takes for granted rides on that ignition never stalling.

Compounding is the engine's secret. Take an income of 10,000 growing 2 percent a year: after 35 years it reaches about 20,000. Lift the growth rate to 3 percent and the same 35 years deliver roughly 28,100 - a 40 percent higher standard of living from a single percentage point. The economist Robert Lucas remarked that once you begin thinking about why growth rates differ, it is hard to think about anything else. This lesson is about where those differences come from.

The rule of 70

A handy shortcut, the rule of 70, estimates how long it takes a growing quantity to double:

Doubling time (years) is approximately 70 / (annual growth rate in percent)

Growth rate	Approximate doubling time
1%	70 years
2%	35 years
3.5%	20 years
7%	10 years

The lesson is dramatic: an economy growing at 7 percent doubles its income in a single decade, while one growing at 1 percent takes a lifetime. Sustained differences in growth are why some nations that were similarly poor a century ago are now vastly richer than others.

The number 70 is not magic; it comes from the mathematics of compounding (the natural logarithm of 2 is about 0.693, and 69.3 rounds to 70 for easy division). What matters is the consequence: doublings stack. Over 70 years, 1 percent growth doubles income once - a factor of 2. Growth of 3.5 percent doubles it every 20 years, three and a half times over, for a factor of about 11. Same span, same planet, elevenfold difference.

History has run this experiment. Around 1960, South Korea and Ghana had roughly comparable incomes per person. Korea then sustained growth rates near 7 percent for a generation - doubling each decade - while Ghana's growth sputtered. Within a single lifetime Korea joined the world's rich economies while Ghana remained poor, a divergence produced not by one dramatic event but by the quiet, relentless arithmetic in the table above.

Where growth comes from

Long-run growth ultimately comes from rising labor productivity - output per hour worked. Producing more per worker is what lets average incomes rise. Productivity, in turn, grows from a few sources:

• Physical capital - more and better tools, machines, and infrastructure make each worker more productive.
• Human capital - education, skills, and health raise what workers can do.
• Natural resources - land, minerals, and energy help, though they are neither necessary nor sufficient for growth.
• Technological progress - better knowledge and methods let us produce more from the same inputs, and it is the main driver of growth in already-rich countries.

A key idea is diminishing returns to capital: piling on more machines while holding other inputs fixed yields ever-smaller gains, so simply accumulating capital cannot sustain growth forever. Lasting growth in rich economies leans on innovation, which does not run into the same wall.

Watch diminishing returns operate on a single farm. With fixed land and workers, the first tractor might add 500 bushels a year; a second adds 300, since the urgent hauling is already covered; a third adds 150; a fourth mostly sits idle. Each machine costs the same, yet buys less output than the last. Poor economies with little capital therefore get spectacular returns from their first roads, ports, and machines - and rich economies, already stuffed with capital, cannot grow much by simply adding more of it.

Human capital deserves equal billing with machines. A workforce that can read manuals, do arithmetic, and stay healthy adopts new technology faster, maintains equipment better, and supplies the stream of small improvements that add up. The East Asian growth stories began with mass basic schooling laid down before the boom; when the factories arrived, people were ready to run them. Health belongs in the same ledger: curing debilitating disease converts sick days into working days and lets school years become skills rather than lost time.

Growth accounting: splitting the credit

Economists decompose growth with this logic, an exercise called growth accounting. Suppose output grows 4 percent in a year: perhaps 1 point traces to a larger workforce, 1.5 points to a bigger capital stock, and the remaining 1.5 points to producing more from the same inputs - the technology residual. That residual, formally total factor productivity, is the piece that carries living standards in the long run.

The formal recipe weights each input by its share of national income. Suppose capital earns about 30 percent of income and labor about 70 percent. If the capital stock grows 4 percent, it contributes roughly 0.3 x 4 = 1.2 points of output growth; if hours worked grow 2 percent, labor contributes 0.7 x 2 = 1.4 points. Now let measured output grow 4.1 percent. Inputs explain 1.2 + 1.4 = 2.6 points, so the remaining 4.1 - 2.6 = 1.5 points must come from squeezing more out of the same inputs. That leftover is the total factor productivity residual.

Economists treat the residual with respect and suspicion at once. Respect, because across long stretches of rich-country history it accounts for the largest share of growth in output per worker. Suspicion, because it is computed as a leftover: measurement errors, unmeasured improvements in worker skill, and better management practices all land in the same bucket, which is why one early researcher called it a measure of our ignorance. Growth accounting names the contributions; explaining the residual is the research program hiding behind the arithmetic.

The Solow model in plain words

The modern theory of these mechanics belongs to Robert Solow, whose 1950s model follows an economy that saves, invests, and accumulates capital. Its central result: capital accumulation alone eventually stalls. As the capital stock grows, diminishing returns shrink each new machine's payoff while depreciation claims a growing bite of investment just for upkeep. The economy glides toward a steady state where saving merely maintains the stock, and growth in income per person stops - unless technology improves.

A small numeric sketch shows the stall. Imagine an economy with a capital stock of 40,000 that saves 20 percent of its income of 10,000, so investment is 2,000 per year. Let 5 percent of capital wear out annually: depreciation is 0.05 x 40,000 = 2,000. Investment exactly covers wear, the stock stops growing, and so does income per person. Raise the saving rate to 25 percent and investment of 2,500 outruns depreciation for a while, but as the stock climbs toward 50,000 the wear bill catches up at 2,500, and growth stalls again at a higher plateau.

The punchline generalizes: a higher saving rate buys a higher level of income per person, not a permanently faster growth rate. The model also predicts catch-up. An economy starting far below its steady state earns large returns on each new machine and grows quickly on the way up - exactly the pattern of postwar Japan, and later South Korea, rebuilding toward the technological frontier.

Technological progress resets the ceiling continuously, which is why it, and not thrift alone, powers long-run growth. When Solow measured US history against the model, he found the majority of growth in output per worker attributable to the technology residual rather than to capital deepening - a finding that redirected growth policy toward research, education, and the conditions that produce ideas.

Institutions and convergence

Good institutions set the stage for all of this: secure property rights, the rule of law, stable government, and openness to trade encourage the investment and risk-taking that growth requires. This framework also suggests the hopeful possibility of convergence - poorer countries can grow faster than rich ones by adopting existing technology and catching up - but only where institutions allow that investment to take root. Where institutions are weak, catch-up stalls.

The cleanest evidence comes from a grim natural experiment. In 1948 the Korean peninsula split into two states sharing one people, one language, and one starting economy. The South built market institutions with secure property and outward trade; the North did not. Two generations later the South stood among the world's richest societies and the North among its poorest - same culture, same geography, different rules. Institutions are not a garnish on growth theory; on the evidence, they are close to the whole meal.

The mechanism behind "institutions matter" is mundane and powerful: people invest when they expect to keep the returns. A farmer with a secure land title plants orchards that pay off in a decade; without the title, she plants only what she can harvest before someone stronger claims the field. Scale that logic up to factories, patents, and thirty-year bonds and it explains why the rules of the game move so much money.

The practical growth-policy menu follows from the whole framework: fund basic education and public health (human capital), maintain infrastructure and financial systems that channel saving into investment (physical capital), support research and protect inventions long enough to reward them while letting ideas eventually spread (technology), keep trade reasonably open (adoption and competition), and hold inflation and public finances stable enough that long horizons feel safe (institutions). No single lever is decisive; growth is the compound interest on doing many ordinary things tolerably well.

Convergence works where those conditions hold. War-shattered Japan and West Germany rebuilt to rich-country status within decades by importing and improving existing technology; South Korea and later coastal China rode the same escalator. The escalator is real but conditional: countries lacking stable law, basic schooling, or open markets have watched it pass them by, which is why economists speak of conditional convergence rather than an automatic law.

Levels versus growth rates in policy arguments

Keep one distinction sharp in every growth debate: a policy that raises the level of output once is worth far less than one that raises the growth rate permanently. A reform that adds a one-time 5 percent to GDP is a pleasant step; a reform that lifts growth from 2 to 2.5 percent compounds forever - over 70 years it turns 4 times richer into about 5.7 times richer, a gap that dwarfs the one-time gain and keeps widening.

Growth also collides with real constraints, and honest accounting includes them. Output that poisons rivers or exhausts soils borrows from future GDP; the growth literature's answer is not to stop growing but to grow through ideas - the input that does not deplete - and to price the resources that do.

Common misconceptions

• "Growth means working ever more hours." Sustained growth comes from output per hour rising; rich economies typically work fewer hours than their ancestors while producing far more.
• "Natural resources guarantee riches." Resource-poor Japan and Switzerland grew wealthy while some resource-rich states stagnated - institutions and technology decide what resources are worth.
• "Poor countries will automatically catch up." Convergence is conditional on institutions, education, and openness; without them the gap persists or widens.
• "Half a percentage point of growth is a rounding error." Compounded over a lifetime it separates prosperity from stagnation - reread the rule of 70 table until this feels visceral.

Try it

Two countries each start at an income of 10,000 per person. Country A grows at 2 percent per year, Country B at 5 percent. (a) Use the rule of 70 to find each doubling time. (b) Roughly what are incomes after 70 years? (c) How many times richer is B than A at that point?

Answer: (a) A doubles every 70 / 2 = 35 years; B every 70 / 5 = 14 years. (b) In 70 years A doubles twice: 10,000 to 20,000 to 40,000. B doubles five times: 10,000 to 320,000 - thirty-two times its start. (c) B ends about 8 times richer than A (320,000 versus 40,000), from nothing but a 3-point growth difference sustained across one long lifetime.

Sources

1. Greenlaw, S. A., Shapiro, D., & MacDonald, D. (2022). The relatively recent arrival of economic growth. In Principles of Macroeconomics 3e. OpenStax, Rice University. openstax.org
2. Greenlaw, S. A., Shapiro, D., & MacDonald, D. (2022). Labor productivity and economic growth. In Principles of Macroeconomics 3e. OpenStax, Rice University. openstax.org
3. Greenlaw, S. A., Shapiro, D., & MacDonald, D. (2022). Components of economic growth. In Principles of Macroeconomics 3e. OpenStax, Rice University. openstax.org
4. Greenlaw, S. A., Shapiro, D., & MacDonald, D. (2022). Economic convergence. In Principles of Macroeconomics 3e. OpenStax, Rice University. openstax.org
5. Romer, P. M. (2008). Economic growth. In The Concise Encyclopedia of Economics (2nd ed.). Liberty Fund. econlib.org
6. Henderson, D. R. (Ed.). (2008). Robert Merton Solow. In The Concise Encyclopedia of Economics (2nd ed.). Liberty Fund. econlib.org
7. Roser, M. (2021). What is economic growth? And why is it so important? In Our World in Data. Global Change Data Lab. ourworldindata.org

Growth is not a straight line

Long-run growth trends upward, but the path is bumpy. The business cycle is the recurring pattern of short-run fluctuations in real GDP around its long-run trend. Every cycle has the same anatomy - expansion, peak, contraction, trough - even though length and depth vary so much from one cycle to the next that economists think of cycles as siblings rather than twins. Some expansions run a decade; some contractions are over in months. The pattern repeats; the timetable does not.

Keep the two time scales separate in your head. The previous lesson explained the trend: capital, skills, and technology push potential output upward decade after decade. This lesson is about the wobble around that trend - the alternating seasons of boom and slump that dominate news coverage and elections. A useful mental picture is a person walking uphill playing with a yo-yo: the hand climbs steadily (the trend) while the yo-yo rises and falls around it (the cycle). Growth policy aims at the hill; stabilization policy, the subject of the coming modules, aims at the yo-yo.

The figure shows the anatomy. The dashed line is the long-run trend of potential output, rising steadily. The solid wave is actual real GDP, running above the trend in booms and below it in slumps. The peak is the moment the wave turns down; the trough is the moment it turns back up. The vertical distance between wave and trend at any date is the output gap from Module 6: positive at the top of a boom, negative in the pit of a recession.

The four phases

• Expansion - real GDP is rising, unemployment falling, and confidence growing. Most of economic history is spent here: hiring spreads from a few industries to most, factories add shifts, and credit flows easily.
• Peak - the top of the cycle, where output is highest and the expansion runs out of steam; the turning point into decline. Labor and materials are scarce and prices press upward, often with the inflationary gap of Module 6 wide open.
• Recession (contraction) - a significant, widespread decline in economic activity lasting more than a few months. A common rule of thumb is two consecutive quarters of falling real GDP, though official bodies weigh a broader set of indicators. Unemployment rises during a recession. An especially deep and prolonged recession is called a depression.
• Trough - the bottom of the cycle, where activity stops falling and recovery begins; the turning point back into expansion.

The phases are not symmetric. Historically, expansions grind on for years while contractions are shorter and steeper: the economy climbs the staircase and takes the elevator down. Nor do the phases announce themselves. A peak becomes visible only in the rearview mirror, months later, once the data confirm that the decline was broad and persistent rather than a one-quarter stumble.

A recession, quarter by quarter

Walk through a hypothetical downturn to see how the pieces move. Start with an economy producing real GDP of 10,000 at an annual rate, growing 3 percent a year, with unemployment at 5 percent. In quarter one, something breaks confidence - suppose a housing bust wipes out wealth and frightens lenders. Households postpone cars and renovations; firms shelve expansion plans. New orders slow first, because an order is the easiest thing in the economy to cancel.

In quarters two and three the slowdown feeds on itself through the circular flow. Factories receiving fewer orders cut overtime, then shifts, then jobs; laid-off workers trim their own spending, which cancels someone else's sale. Real GDP slips from 10,000 to 9,800, then to 9,650. Unemployment climbs from 5 toward 7 percent, and it keeps climbing even as the decline slows, because firms wait for solid evidence before rehiring. By now the two-quarter rule of thumb is satisfied, and commentators say the word recession out loud.

In quarter five output stops falling at 9,600: the trough, exactly 4 percent below the peak. Recovery begins, but the early expansion feels like recession on the ground. Output must retrace lost ground before setting new highs, and unemployment, a lagging indicator, may not crest until the recovery is two or three quarters old. Newspapers coin phrases like jobless recovery for precisely this stage: GDP rising, payrolls flat.

Inventories: the cycle's flywheel

Inventories are a small slice of GDP but a mighty force in the cycle, because they transmit and amplify swings in demand. When sales dip, unsold goods pile up on shelves unplanned. Firms respond by cutting production below sales until the excess is worked off - so a modest drop in demand becomes a sharper drop in output. In recoveries the flywheel spins the other way: restocking adds a burst of production on top of reviving sales.

Numbers make it vivid. Suppose a retailer sells 100 appliances a month and likes to hold two months of sales, 200 units, in stock. Sales slip 10 percent to 90. The target stock falls to 180, so for a month the retailer orders only 70 from the factory: sales fell 10 percent, but orders fell 30 percent. Multiply that arithmetic across every firm in a supply chain and you see why manufacturing swings so much harder than final sales, and why an inventory correction shows up in nearly every recession on record.

Reading the cycle with indicators

Because turning points are hard to see in real time, economists watch economic indicators classified by their timing:

• Leading indicators tend to change before the overall economy does, offering early warning - examples include new building permits, stock prices, new orders for capital goods, and initial claims for unemployment insurance.
• Lagging indicators change after the economy turns, confirming a trend already underway - the unemployment rate is a classic lagging indicator, since firms are slow to hire and slow to fire.
• Coincident indicators move roughly with the economy, such as real GDP, industrial production, and payroll employment.

The logic behind each category is economic, not statistical. Permits lead because a permit today is construction spending months from now. Capital-goods orders lead because they embody firms' bets about future demand. Stock prices lead, imperfectly, because investors price in the profits they expect next year rather than last year. Unemployment lags because hiring and firing are costly decisions that firms delay until the trend is unmistakable. Statistical agencies bundle leading series into composite indexes to smooth the noise in any single line.

Handle leading indicators with care: they warn often and lie sometimes. Stock prices in particular have swooned before recessions that never arrived, which is what the economist Paul Samuelson meant by joking that the market had predicted nine of the last five recessions. The practical craft is to look for many leading series turning together, for several months, rather than reacting to one twitchy chart.

Who decides when a recession happened?

In the United States the semi-official scorekeeper is the National Bureau of Economic Research (NBER), whose dating committee defines a recession as a significant decline in economic activity spread across the economy, lasting more than a few months. The committee dates the specific peak and trough months by weighing employment, income, sales, and production together - not GDP alone, and not the two-quarter shortcut. Its verdicts arrive long after the fact, sometimes a year or more, because the job is to write accurate history rather than fast headlines.

The delay carries a policy lesson worth flagging now: by the time a recession is officially declared, it is well underway and occasionally already over. Policymakers must act on noisy real-time data instead, a difficulty that returns in the fiscal-policy lesson under the name of recognition lag.

Two slumps for the record

Settled history offers a matched pair. The Great Depression remains the benchmark catastrophe: between 1929 and 1933, United States output fell by roughly a third, the price level dropped, thousands of banks failed, and about one worker in four was unemployed. There was no deposit insurance and, for years, little effective policy response; the contraction fed on bank runs and deflation, and full recovery took the better part of a decade.

The 2008-09 recession began the same way - collapsing asset prices and a financial panic - but ran a different course. Output fell about 4 percent, severe by postwar standards yet nowhere near a third, and unemployment peaked near 10 percent rather than 25. Part of the difference was machinery built after the 1930s: deposit insurance kept household bank runs at bay, the central bank acted aggressively as lender of last resort, and fiscal stimulus supported demand. The pair reads like a controlled comparison of a cycle without shock absorbers and one with them.

What causes cycles?

Cycles arise from shocks to aggregate demand or aggregate supply - swings in business and consumer confidence, changes in policy, financial crises, or supply disruptions like an oil shock. The AD-AS model from the previous module is the standard tool for analyzing them: a leftward AD shift produces a demand-driven recession in which output and prices fall together, while a leftward SRAS shift produces a supply-driven downturn in which output falls while prices rise. The diagnostic rule from Module 6 applies directly - prices and output moving together point to demand, moving oppositely to supply.

Shocks start cycles; propagation stretches them. The multiplier passes one sector's cutbacks along as another sector's lost income, inventory corrections amplify the swing, and credit tightens exactly when borrowers look riskiest. That is how a shock lasting weeks produces a slump lasting years. Understanding the cycle sets up the central policy question of the remaining modules: what, if anything, should government and the central bank do to smooth these fluctuations?

Common misconceptions

• "Two negative quarters is the official definition of a recession." It is a journalistic shortcut. The official call weighs depth, diffusion, and duration across many indicators, so a shallow two-quarter dip may not qualify while a savage shorter collapse might.
• "Cycles run on a schedule." They are recurrent, not periodic. Expansions do not die of old age on a timetable; something has to hit demand or supply.
• "Rising unemployment means the recovery has failed." Unemployment lags. It routinely keeps rising after output turns up, because firms rehire only once the recovery looks durable.
• "In a recession everything declines everywhere." The decline must be widespread, not universal; some industries grow through almost any downturn while others shrink even in mild ones.

Try it

An economy reports real GDP (at an annual rate) over seven quarters: 10,000; 10,050; 10,020; 9,900; 9,780; 9,760; 9,850. (a) In which quarter is the peak, and in which the trough? (b) Compute the peak-to-trough decline in percent. (c) Does the two-quarter rule of thumb signal a recession, and after which quarter's data? (d) Unemployment is still rising in quarter seven while GDP grows. Is that consistent with this lesson?

Answer: (a) The peak is quarter two (10,050), the last quarter before sustained decline; the trough is quarter six (9,760), after which output rises. (b) (10,050 - 9,760) / 10,050 = 290 / 10,050, or about 2.9 percent. (c) Yes. Quarters three and four are consecutive declines, so the rule flashes once quarter four's data arrive - well after the downturn actually began at the peak. (d) Yes. Unemployment is a lagging indicator, so it commonly keeps climbing through the first quarters of recovery while output rebounds.

Sources

1. National Bureau of Economic Research. (n.d.). Business cycle dating. In NBER Research. National Bureau of Economic Research. nber.org
2. National Bureau of Economic Research. (n.d.). US business cycle expansions and contractions. In NBER Data. National Bureau of Economic Research. nber.org
3. Romer, C. D. (2008). Business cycles. In The Concise Encyclopedia of Economics (2nd ed.). Liberty Fund. econlib.org
4. Greenlaw, S. A., Shapiro, D., & MacDonald, D. (2022). How the AD/AS model incorporates growth, unemployment, and inflation. In Principles of Macroeconomics 3e. OpenStax, Rice University. openstax.org
5. Richardson, G. (2013). The Great Depression. In Federal Reserve History. Federal Reserve Bank of St. Louis. federalreservehistory.org
6. Rich, R. (2013). The Great Recession. In Federal Reserve History. Federal Reserve Bank of St. Louis. federalreservehistory.org
7. The CORE Team. (2017). Economic fluctuations and unemployment. In The Economy 1.0. CORE Economics Education. core-econ.org

What money does

Money is anything widely accepted in exchange, and economists define it by three functions rather than by its physical form:

• Medium of exchange - you trade it for goods and services, avoiding the awkward "double coincidence of wants" that barter requires.
• Unit of account - prices are quoted in it, giving a common yardstick for value.
• Store of value - it holds purchasing power over time, though inflation erodes this.

To feel why the first function matters, imagine an island economy without money. A fisher who wants sandals must find a sandal maker who happens to want fish - today, before the fish spoils. Fail, and she must arrange a chain of swaps: fish for rope, rope for leather, leather for sandals. Every trade requires a matching pair of wants. Money dissolves the problem: sell fish to anyone for money, buy sandals from anyone with it. Trade that was a scavenger hunt becomes two quick transactions, and specialists can spend their days producing instead of matchmaking.

The unit-of-account function does quieter work. With 100 goods and no money, there are 4,950 barter exchange ratios to track; with money, just 100 prices. And the store-of-value function makes saving practical: a carpenter can convert this month's work into money and spend it next year, something impossible with perishable output. The three functions travel together - and they fail together. Whatever settles debts and buys groceries in an economy is its money, whether that is gold coin, paper, or an entry in a bank's computer.

Through history the form has migrated. Commodity money - gold, silver, salt, even cigarettes in prison camps - has value of its own. Modern money is fiat money: it has value by government decree and common trust, not because it is backed by gold or any commodity. A paper note is intrinsically worthless; it works because everyone expects everyone else to accept it, because taxes must be paid in it, and because its supply is kept scarce. That last condition is the load-bearing one.

When the scarcity condition fails, the functions collapse in order. In Germany's hyperinflation of 1923, prices doubled every few days; workers were paid twice daily and spent within hours as the store-of-value function died, shops reposted prices constantly as the unit-of-account function followed, and by the end people bartered or used foreign notes as the medium-of-exchange function gave way. Zimbabwe ran the same grim script in the late 2000s before abandoning its currency altogether. Money is a social technology, and hyperinflation is what its breakdown looks like.

Measuring money: M1 and M2

Because "widely accepted in exchange" is a matter of degree, statisticians define the money supply in nested layers ordered by liquidity, the ease of converting an asset into spending. M1 is the narrow layer: currency in circulation plus checkable deposits - money you can spend this afternoon. M2 adds near-money: savings deposits, small time deposits, and similar balances that take a step or a small penalty to spend. Every dollar of M1 is inside M2; the difference is how instantly spendable the layer is.

Just as important is what does not count. A credit card is not money; it is access to a loan that must later be repaid with money. Stocks, bonds, and houses are valuable but must first be sold - at an uncertain price - before they can buy groceries. And reserves held by banks inside the central bank are not counted, because the public cannot spend them. The boundary between M1 and M2 blurs as banking technology makes savings balances easier to tap, which is why economists watch several measures rather than swearing loyalty to one.

How banks create money

Most money in a modern economy is not printed by the government but created by ordinary banks through fractional-reserve banking. A bank keeps only a fraction of its deposits on hand as reserves and lends the rest out. Those loans are spent and redeposited in other banks, which again keep a fraction and lend the rest, and so on. Each round creates new deposits, expanding the money supply far beyond the original cash.

A bank's balance sheet makes the mechanics concrete. On the liability side sit deposits: what the bank owes its customers on demand. On the asset side sit reserves (vault cash and balances at the central bank) and loans (promises of future repayment from borrowers). The business model is to pay depositors a low rate, charge borrowers a higher one, and live on the spread. A T-account is the economist's shorthand sketch of this balance sheet, with assets on the left and liabilities on the right - and it is the ideal tool for watching money creation happen.

The story in T-accounts, step by step

Suppose every bank targets a reserve ratio of 10 percent. Ana deposits 1,000 in cash at First Bank. First Bank's T-account changes on both sides: reserves (an asset) rise 1,000, and Ana's deposit (a liability) rises 1,000. So far the money supply has not changed - currency worth 1,000 left circulation and a deposit worth 1,000 appeared, a swap of form, not size.

First Bank now needs only 100 in reserves against Ana's deposit, so it lends the idle 900 to Ben. The moment Ben spends the 900 at a hardware store and the store banks it, Second Bank's T-account shows reserves +900 and deposits +900. Count the money: Ana still holds a 1,000 deposit she can spend at will, and the store holds 900. The money supply has grown from 1,000 to 1,900. The extra 900 was created by a bookkeeping entry - a loan - not by any printing press.

The process repeats and shrinks. Second Bank holds 90, lends 810; the borrower's spending lands in Third Bank as a deposit of 810, which keeps 81 and lends 729. After four rounds, deposits total 1,000 + 900 + 810 + 729 = 3,439, and every bank in the chain holds exactly a tenth of its deposits in reserve. Each round passes along 90 percent of the last, so the rounds form a geometric series that converges: total deposits approach 10,000, at which point all 1,000 of original cash sits as required reserves and the expansion stops on its own.

The money multiplier

The total expansion is captured by the money multiplier, which in the simplest model is:

Money multiplier = 1 / reserve ratio

where the reserve ratio is the fraction of deposits banks hold as reserves. The maximum new money the banking system can create from a fresh deposit is:

Maximum change in deposits = initial deposit x money multiplier

Worked example

Suppose the reserve ratio is 10 percent (0.10) and someone deposits $2,000 in fresh cash into a bank.

Money multiplier = 1 / 0.10 = 10
Maximum new deposits = $2,000 x 10 = $20,000

The initial $2,000 can support up to $20,000 in total deposits across the banking system. Now raise the reserve ratio to 20 percent (0.20): the multiplier falls to 1 / 0.20 = 5, so the same $2,000 supports only $2,000 x 5 = $10,000. A higher reserve ratio means banks lend less of each deposit, so the money supply expands by less. This mechanism is exactly what the central bank influences when it conducts monetary policy.

Bank runs and deposit insurance

Fractional-reserve banking is productive and inherently fragile, because it commits a maturity mismatch on purpose: deposits are payable on demand, while the loans they fund are locked up in mortgages and business projects for years. On any normal day only a trickle of depositors wants cash, so a 10 percent reserve is ample. But if depositors believe the bank may fail, each has reason to withdraw first, and the belief fulfills itself: no bank, however sound its loans, can pay out everyone at once. This is the fallacy-of-composition trap from Lesson 1 wearing a banker's suit.

The early 1930s showed the full cost. Waves of runs swept the United States, thousands of banks closed, and as deposits were yanked out and loans called in, the money-creation process of this lesson ran in reverse: the money supply collapsed by roughly a third between 1929 and 1933, strangling spending and deepening the Great Depression. The policy answer arrived in 1933: federal deposit insurance, which guarantees deposits up to a limit even if the bank fails. Once depositors knew their money was safe either way, the reason to run evaporated, and runs on insured deposits all but vanished.

Insurance carries its own side effect, moral hazard: depositors who cannot lose stop scrutinizing their banks, and a bank funded by insured deposits may be tempted toward riskier loans. That is why insurance travels with supervision, capital requirements, and limits on what banks may do - the safety net and the leash come as a set.

A note on the real world

The simple multiplier assumes banks lend out every spare dollar and the public redeposits all of it. In practice banks may hold extra reserves and people hold some cash, so the actual multiplier is smaller than 1 / reserve ratio. If households keep a slice of each loan as pocket cash, that slice never returns to a bank to back further lending, and the chain shortens. If bankers turn cautious - as they famously did in the 1930s and again after 2008 - they hold reserves beyond any requirement, and the chain shortens further.

The principle still holds: fractional-reserve banking lets deposits multiply, and the size of that expansion depends on the reserve ratio, the public's taste for cash, and banks' willingness to lend. The multiplier is best read as an upper bound on a process whose true strength varies with confidence - which is precisely why the next lesson's central bank watches it so closely.

Common misconceptions

• "When a bank lends my deposit, my balance goes down." No. Your deposit remains yours to spend; the borrower gets newly created purchasing power. That doubling of spendable claims is exactly how the money supply grows.
• "Only the government creates money." The state prints currency, but most of the money supply is bank deposits created by lending. The printing press is the smaller part of the story.
• "Reserves are the bank's cushion against bad loans." Reserves are liquidity for meeting withdrawals. Losses are absorbed by the bank's capital - its owners' stake - a different line on the balance sheet.
• "A credit card is money." It is a standing offer of a loan. The money appears only when the loan is drawn, and settling the bill requires actual money later.
• "The multiplier guarantees a 10x expansion." It is a maximum. Cash holdings and cautious banks routinely hold the real expansion below the formula's ceiling.

Try it

An economy's banks target a reserve ratio of 25 percent, and a saver deposits 4,000 of cash into Harbor Bank. (a) Compute the money multiplier and the maximum total deposits the system can support from this deposit. (b) How much new money can be created beyond the original 4,000? (c) Trace the first two lending rounds explicitly. (d) If banks instead hold 30 percent of deposits from caution, what happens to the maximum expansion?

Answer: (a) Multiplier = 1 / 0.25 = 4, so maximum deposits = 4,000 x 4 = 16,000. (b) New money is 16,000 - 4,000 = 12,000, all created by lending. (c) Harbor Bank keeps 1,000 and lends 3,000; once spent and redeposited, the next bank keeps 750 and lends 2,250. Deposits so far: 4,000 + 3,000 + 2,250 = 9,250. (d) The effective multiplier falls to 1 / 0.30, about 3.33, so maximum deposits shrink to about 13,333 - caution at the banks shrinks the money supply with no change in policy.

Sources

1. Greenlaw, S. A., Shapiro, D., & MacDonald, D. (2022). Defining money by its functions. In Principles of Macroeconomics 3e. OpenStax, Rice University. openstax.org
2. Greenlaw, S. A., Shapiro, D., & MacDonald, D. (2022). Measuring money: Currency, M1, and M2. In Principles of Macroeconomics 3e. OpenStax, Rice University. openstax.org
3. Greenlaw, S. A., Shapiro, D., & MacDonald, D. (2022). How banks create money. In Principles of Macroeconomics 3e. OpenStax, Rice University. openstax.org
4. Board of Governors of the Federal Reserve System. (2026). Money stock measures - H.6. In Statistical Releases. Federal Reserve Board. federalreserve.gov
5. Schwartz, A. J. (2008). Money supply. In The Concise Encyclopedia of Economics (2nd ed.). Liberty Fund. econlib.org
6. Kaufman, G. G. (2008). Bank runs. In The Concise Encyclopedia of Economics (2nd ed.). Liberty Fund. econlib.org
7. Maues, J. (2013). Banking Act of 1933 (Glass-Steagall). In Federal Reserve History. Federal Reserve Bank of St. Louis. federalreservehistory.org

The central bank

Every modern economy puts one institution in charge of its money, and in the United States that institution is the Federal Reserve ("the Fed"). Congress created it in 1913, after the Panic of 1907 showed what a banking system without a backstop looks like: a run on a few trust companies cascaded through the country until a private financier, J. P. Morgan, personally organized the rescue. Lawmakers concluded that the job of stabilizing money and banking was too important to depend on one rich man's judgment, and built a permanent public institution to do it.

The Fed is deliberately layered. A seven-member Board of Governors in Washington, appointed to staggered fourteen-year terms, oversees the system; twelve regional Federal Reserve Banks keep it rooted in the country's different economies; and the Federal Open Market Committee (FOMC) - the seven governors plus the New York Fed president and four other regional presidents in rotation - makes the interest-rate decisions. Congress assigned the Fed a dual mandate: promote maximum employment and stable prices. Beyond steering the economy, the Fed serves as lender of last resort during panics and supervises much of the banking system.

The long terms and self-funded budget are not accidents; they are insulation. Raising interest rates is medicine that hurts before it helps, and an official who must face voters next year has every incentive to delay it. Economies whose central banks answer directly to politicians have historically tolerated more inflation for the same growth. Independence does not mean isolation - the Fed reports to Congress and its chair testifies publicly - but it means the people adjusting the money supply are not the people running for reelection.

The Fed's main tools

The Fed influences the economy by changing the supply of money and the level of interest rates, using three principal tools:

• Open-market operations - buying and selling government bonds. This is the primary day-to-day tool. When the Fed buys bonds, it pays with newly created reserves, increasing the money supply and lowering interest rates; when it sells bonds, it drains reserves, shrinking the money supply and raising rates.
• The discount rate and interest on reserves - the interest rate the Fed charges banks to borrow, and the rate it pays banks on reserves. Adjusting these rates changes banks' incentives to lend and helps steer the key short-term interest rate (the federal funds rate) toward the Fed's target.
• Reserve requirements - the required reserve ratio itself. Lowering it lets banks lend more (raising the money multiplier); raising it does the reverse. This blunt tool is used rarely.

Open-market operations in T-accounts

Follow the bookkeeping of a purchase, because the whole mechanism lives there. Suppose the Fed buys 100 million of government bonds from Coastal Bank. On the Fed's T-account, assets rise by 100 million of bonds, and liabilities rise by 100 million of bank reserves - the Fed pays by crediting Coastal's reserve account with a keystroke. On Coastal's T-account, one asset is swapped for another: bonds down 100 million, reserves up 100 million. No depositor is involved, and nothing physical moves.

The reserves did not come from anywhere; they were created in the act of payment. That is the unique power of a central bank, and it is why the purchase loosens credit. Coastal now holds idle reserves earning little. It can lend them overnight to other banks in the federal funds market, and as every bank in the same position does likewise, the extra supply of overnight funds pushes the federal funds rate down. Or Coastal can expand customer lending, setting off exactly the deposit-multiplication process of the previous lesson. A sale runs the film backward: the Fed takes reserves out, overnight money tightens, and the funds rate rises.

One subtlety keeps students honest: reserves themselves are not part of M1 or M2, since the public cannot spend them. The money supply grows only as banks turn the new reserves into loans and deposits. Open-market operations load the spring; bank lending releases it.

Expansionary versus contractionary policy

Monetary policy comes in two directions, chosen to fit where the economy sits in the business cycle:

	Expansionary (easy) policy	Contractionary (tight) policy
Used when	Recession, high unemployment	Overheating, high inflation
Fed action	Buy bonds, lower rates	Sell bonds, raise rates
Money supply	Increases	Decreases
Effect on AD	Shifts right (more spending)	Shifts left (less spending)

The transmission mechanism works through interest rates. In a recession the Fed lowers rates, which makes borrowing cheaper, encouraging firms to invest and households to buy homes and cars. That extra spending shifts aggregate demand right, raising output and employment. To fight inflation the Fed does the opposite: raising rates cools borrowing and spending, shifting AD left and easing price pressures - at the risk of slowing growth.

Trace the expansionary case with hypothetical numbers. Suppose output sits 300 below potential and the funds rate stands at 5 percent. The Fed buys bonds until the funds rate falls to 3 percent. Banks flush with reserves compete to lend, so mortgage, auto, and business loan rates follow the overnight rate down. A factory expansion that was unprofitable at the old loan rate clears the hurdle at the new one; a family priced out of a mortgage payment is priced back in. Component by component, interest-sensitive spending revives, the multiplier passes each new dollar around the circular flow, and AD shifts right toward potential.

The contractionary case is the mirror. With an inflationary gap open and prices accelerating, the Fed sells bonds, reserves drain, and the funds rate climbs. Construction cools first - housing is the most rate-sensitive sector in the economy - followed by business equipment and big-ticket consumer purchases. AD shifts left, the gap closes, and inflation pressure eases. The arithmetic is symmetric; the politics never is, because this direction of the medicine raises unemployment on the way to price stability.

The Volcker disinflation

The classic demonstration that the tools work - and what they cost - is the disinflation engineered by Fed chair Paul Volcker. By 1979, after a decade of oil shocks and accommodating policy, US inflation was running near 13 percent and expectations of more were built into every wage bargain and loan. Volcker tightened relentlessly: the federal funds rate was pushed toward 20 percent by 1981, credit-financed spending collapsed, and the economy went through back-to-back recessions, with unemployment peaking near 10.8 percent in late 1982 - the worst since the Depression.

The payoff arrived on the other side. Inflation fell to about 3 percent by 1983 and stayed low for a generation, and the episode rewrote the textbook in two ways. First, it settled the argument over whether monetary policy could break an entrenched inflation: it can. Second, it demonstrated the value of credibility. Once firms and workers believed the Fed would not relent, they stopped building double-digit inflation into contracts, and the economy could grow again without reigniting prices. Central banks worldwide drew the same conclusion, which is why so many now guard their inflation-fighting reputations so jealously.

Lender of last resort

The Fed's oldest job is not steering the cycle but stopping panics. The classic prescription was written by the British journalist Walter Bagehot in 1873: in a panic, the central bank should lend freely, at a penalty rate, against good collateral. The logic targets the self-fulfilling run from the previous lesson - a solvent bank facing a stampede needs cash now, and a central bank that can create reserves at will is the one institution that can always supply it. Knowing the backstop exists is often enough to stop the stampede from starting.

History graded the Fed on this duty twice. In the early 1930s it largely stood aside as thousands of banks failed and the money supply collapsed by a third - the failure Milton Friedman and Anna Schwartz later placed at the center of the Great Depression. In 2008 the lesson had been absorbed: the Fed lent aggressively through the panic, and the financial system, though battered, did not repeat the 1930s cascade. The enduring difficulty is telling an illiquid bank (sound but short of cash, deserving help) from an insolvent one (truly broke, deserving orderly failure) mid-crisis, with billions riding on the diagnosis.

Strengths and limits

Monetary policy is flexible and fast to enact - the FOMC can change rates at a scheduled meeting without waiting for legislation. But it faces limits. Its effects arrive with a lag of many months. It can lose traction at the zero lower bound, when interest rates are already near zero and cannot be cut much further, as in a severe slump. And pushing too hard on either side risks overshooting - too much stimulus can fuel inflation, too much tightening can trigger a recession. These trade-offs are why central banking is often described as an art as much as a science.

The lag deserves respect: a rate cut today filters into mortgages, construction schedules, and hiring over a year or more, so the FOMC must aim at where the economy will be, not where it is - steering a supertanker, turning the wheel well before the bend. The zero bound bites hardest after financial crises; when short rates hit zero after 2008, the Fed turned to large-scale bond purchases (quantitative easing) to press down longer-term rates directly. And stimulus works by invitation, not command: the Fed can make borrowing cheap but cannot force a frightened firm to borrow - easing, in the old complaint, can be like pushing on a string.

Common misconceptions

• "The Fed sets interest rates, full stop." It targets one overnight rate and influences the rest. Mortgage and bond rates are set in markets, which respond to expected inflation and growth as well as to policy.
• "Open-market purchases mean printing paper money." The Fed pays with electronic reserve credits. Paper currency is supplied passively, whenever the public wants to hold more cash instead of deposits.
• "Low interest rates prove money is easy." Rates can be low because the economy is weak and expected inflation is low - a point Milton Friedman stressed. Judge policy by conditions, not by the level of the rate alone.
• "The Fed is just another government department." It operates inside government but with unusual independence - long terms, its own budget - precisely so that monetary decisions are not synchronized to elections.
• "Rate changes work within weeks." The main effects on output and inflation take many months to arrive, which is why policy must be forward-looking.

Try it

An economy has a recessionary gap, and its banks hold reserves at a 10 percent ratio. (a) Should the central bank buy or sell bonds? (b) Show the T-account effects of a 50 million purchase from Bank One. (c) What is the maximum deposit expansion the new reserves could ultimately support? (d) Trace the chain from the operation to aggregate demand in one sentence.

Answer: (a) Buy bonds - expansionary policy. (b) Central bank: bonds +50 million (assets), bank reserves +50 million (liabilities). Bank One: bonds -50 million, reserves +50 million - an asset swap that leaves it holding lendable reserves. (c) With a 10 percent ratio the multiplier is 1 / 0.10 = 10, so deposits could expand by up to 50 million x 10 = 500 million if fully lent and redeposited. (d) The purchase raises reserves, pushing the overnight rate down; cheaper credit spurs interest-sensitive spending, and the multiplier carries that spending around the circular flow, shifting AD right toward potential.

Sources

1. Greenlaw, S. A., Shapiro, D., & MacDonald, D. (2022). The Federal Reserve banking system and central banks. In Principles of Macroeconomics 3e. OpenStax, Rice University. openstax.org
2. Greenlaw, S. A., Shapiro, D., & MacDonald, D. (2022). How a central bank executes monetary policy. In Principles of Macroeconomics 3e. OpenStax, Rice University. openstax.org
3. Greenlaw, S. A., Shapiro, D., & MacDonald, D. (2022). Pitfalls for monetary policy. In Principles of Macroeconomics 3e. OpenStax, Rice University. openstax.org
4. Board of Governors of the Federal Reserve System. (2021). Monetary policy: What are its goals? How does it work? In Monetary Policy Principles and Practice. Federal Reserve Board. federalreserve.gov
5. Board of Governors of the Federal Reserve System. (2025). Open market operations. In Monetary Policy Tools. Federal Reserve Board. federalreserve.gov
6. Moen, J. R., & Tallman, E. W. (2015). The Panic of 1907. In Federal Reserve History. Federal Reserve Bank of St. Louis. federalreservehistory.org
7. Medley, B. (2013). Volcker's announcement of anti-inflation measures. In Federal Reserve History. Federal Reserve Bank of St. Louis. federalreservehistory.org

The government's spending-and-tax lever

Fiscal policy is the government's use of spending and taxes to influence the economy. It is set by the legislature and executive - not the central bank - which makes it distinct from monetary policy. Like monetary policy, it works mainly by shifting aggregate demand.

• Expansionary fiscal policy - increasing government spending or cutting taxes to boost aggregate demand during a recession. More spending adds directly to AD; tax cuts leave households and firms more to spend.
• Contractionary fiscal policy - cutting spending or raising taxes to cool an overheating economy and restrain inflation.

Fiscal stabilization as deliberate doctrine dates to the Great Depression, when Keynes argued that an economy stuck below capacity needed someone to spend, and that a government able to borrow could be the spender of last resort. The two levers differ in directness. A dollar of government purchases lands in aggregate demand immediately and in full; a dollar of tax cuts becomes demand only if the recipient spends it. That distinction drives the multiplier arithmetic below.

It also helps to sort fiscal actions by trigger. Discretionary policy requires a fresh decision - a stimulus bill, a public works program, a tax rebate. Automatic policy is built into existing law and fires without anyone voting, a category important enough to get its own section shortly. Monetary policy's comparative advantages are speed and fine-tuning; fiscal policy's are raw power and the ability to aim - at infrastructure, at low-income households, at a specific region - which no interest rate can do.

The spending multiplier

A change in government spending can raise total output by more than the initial amount, because the money is re-spent again and again around the circular flow. The size of this spending multiplier depends on the marginal propensity to consume (MPC) - the fraction of each extra dollar of income that people spend rather than save:

Spending multiplier = 1 / (1 - MPC)

The formula is a geometric series in disguise. Round one: the government spends 1. Round two: recipients spend MPC of it. Round three: MPC of that, and so on. Total = 1 + MPC + MPC squared + ..., which sums to 1 / (1 - MPC). Each round leaks some income into saving, and in a fuller model into taxes and imports as well, which is why real-world multipliers are smaller than the simple formula suggests. The saving leak alone is what makes the series converge instead of exploding.

Worked example

Suppose the MPC is 0.75, meaning people spend 75 cents of every extra dollar. The government increases spending by $100 billion.

Spending multiplier = 1 / (1 - 0.75) = 1 / 0.25 = 4
Maximum increase in GDP = $100 billion x 4 = $400 billion

The first $100 billion is spent by the government; the recipients spend 75 percent of it ($75 billion), whose recipients spend 75 percent of that ($56.25 billion), and so on. The rounds sum to a total of $400 billion. A higher MPC means a bigger multiplier: at an MPC of 0.8 the multiplier is 1 / 0.20 = 5, so the same $100 billion could raise GDP by up to $500 billion. (The saved fraction leaks out each round, which is why a lower MPC gives a smaller multiplier.)

The tax multiplier

Tax changes move demand too, but with one less punch. A $100 billion tax cut does not add $100 billion of spending in round one; households first save a quarter of it (at an MPC of 0.75), spending only $75 billion. Every later round is identical to the spending case, so the whole chain is smaller by that first-round difference:

Tax multiplier = MPC / (1 - MPC), here 0.75 / 0.25 = 3, against 4 for spending.

The gap has a practical edge: dollar for dollar, direct purchases stimulate more than tax cuts, while tax cuts act faster and let households choose. Note also that who gets a tax cut matters, because MPCs differ - a dollar to a family living paycheck to paycheck is mostly spent, a dollar to a wealthy saver mostly is not. Targeting is part of multiplier arithmetic, not a separate topic.

Automatic stabilizers

Automatic stabilizers are the fiscal reflexes wired into existing law: progressive income taxes, unemployment insurance, and income-support programs. In a downturn, tax bills shrink as incomes fall and benefit payments swell as layoffs spread - expansionary policy that arrives without a single vote, in the same months the recession is happening. In a boom the reflex reverses, quietly withdrawing stimulus as tax receipts climb and benefits roll off.

Follow a household through it. Suppose a worker's income falls by 1,000 in a recession. Her income tax falls by 250, and unemployment benefits add 150, so disposable income drops only 600. With an MPC of 0.75 her spending falls by 450 rather than the 750 a cushion-free system would produce. Multiply across millions of households and the recession transmits itself at reduced strength. The visible symptom is a budget deficit that swells automatically in slumps - which is why economists distinguish the cyclical deficit (the recession's fingerprint) from the structural deficit (what would remain at full employment).

Deficits versus debt - a stock and a flow

Two terms that are constantly confused: the budget deficit is a flow - the amount by which government spending exceeds its tax revenue in a single year (a surplus is the reverse). The national debt is a stock - the total accumulated amount the government owes, the sum of all past deficits minus surpluses. Running a deficit adds to the debt; running a surplus reduces it. A government can shrink its deficit while its debt still grows, as long as it is still spending more than it collects.

Raw debt numbers mean little without scale, so economists divide by GDP. Suppose a country owes 6,000 with GDP of 10,000: a debt-to-GDP ratio of 60 percent. It runs a deficit of 400 while nominal GDP grows 5 percent. Next year the debt is 6,400 and GDP is 10,500, so the ratio is 6,400 / 10,500, about 61 percent - up slightly. Rerun the year with a deficit of 200: debt 6,200 against GDP 10,500 is about 59 percent. The debt rose in both cases, but in the second the burden fell, because the economy outgrew its borrowing.

That race between growth and borrowing is the whole game of debt sustainability. The United States ended the Second World War with debt briefly larger than a full year's GDP, then spent three decades walking the ratio down - not by repaying the debt, which barely shrank, but by growing the denominator. A country with a growing economy can carry a stable debt forever; what it cannot do is let the debt persistently outgrow the income that services it.

Crowding out

Deficits must be financed by borrowing, and government borrowing competes with everyone else's in the market for loanable funds. Suppose national saving is 1,500 and firms wish to borrow 1,500 for investment at the going interest rate; now the treasury enters needing 300. Total demand for funds exceeds supply, the interest rate is bid up, and some private projects that penciled out at the old rate are shelved. The stimulus crowds out part of the investment that would have happened - offsetting some of the multiplier and, over decades, leaving a smaller capital stock.

How much crowding out occurs depends on conditions. At full employment, with saving fully employed, it can be nearly dollar for dollar. In a deep slump the story reverses: households are saving heavily, firms are not borrowing, idle funds abound, and the central bank is holding rates down - so the treasury's borrowing displaces little and the multiplier keeps most of its force. This is why economists who argue fiercely about deficits in normal times often agree about deficit spending in a severe recession.

The Ricardian caveat

A further caveat carries David Ricardo's name. Government borrowing is deferred taxation: today's deficit is tomorrow's tax bill, with interest. If households see through the veil - the argument revived by economist Robert Barro - a debt-financed tax cut should not change their behavior at all. They would save the entire cut to cover the future taxes it implies, the MPC out of the windfall would be roughly zero, and the tax multiplier would collapse toward nothing. This is Ricardian equivalence: deficits merely relabel taxes across time.

The evidence treats it as a caveat, not a law. Many households spend tax cuts because they are borrowing-constrained, expect the future taxes to fall on someone else, or simply do not run fifty-year budgets. But the logic is not empty either: heavily indebted governments do find stimulus less potent as citizens start behaving Ricardianly. Treat the theorem as a thumb on the scale that trims multipliers, growing heavier as debt grows larger.

Limits and side effects

Fiscal policy's deepest practical weakness is timing. The recognition lag: recessions are dated by committees a year later, so policymakers act on foggy data. The decision lag: legislation must pass two chambers and survive negotiation, which can take quarters. The implementation lag: even approved projects need contracts and shovels. Stack the three and discretionary stimulus can arrive after the trough, pushing an economy that has already turned - one reason automatic stabilizers, which skip all three lags, earn such affection from economists.

Political economy adds a ratchet: expansionary policy is popular and contractionary policy is not, so deficits accumulate across cycles rather than averaging out. And a rising debt eventually claims a growing slice of the budget as interest, squeezing everything else. None of this makes fiscal policy useless - it remains the strongest tool available at the zero lower bound, when monetary policy runs out of room. It makes fiscal policy a powerful instrument with a slow trigger and a long bill.

Common misconceptions

• "The deficit and the debt are the same thing." The deficit is this year's flow of new borrowing; the debt is the accumulated stock of all past borrowing. Lesson 1's bathtub: faucet and water level.
• "A shrinking deficit means the debt is falling." Any deficit, however reduced, adds to the debt. Only a surplus subtracts from it.
• "Government budgets work exactly like household budgets." Governments differ: they can roll debt over indefinitely, their spending moves the economy that generates their revenue, and what matters is the debt-to-GDP race, not a payoff date. The differences are real, and so is the constraint at the end of them.
• "The multiplier is a constant of nature." It shrinks with leakages, crowding out, Ricardian saving, and a central bank leaning against the stimulus; it is largest in a slack economy with accommodative money.
• "Fighting recessions requires new legislation." Much of the response is automatic - taxes fall and benefits rise the moment incomes do.

Try it

An economy has an MPC of 0.8. (a) Compute the spending and tax multipliers. (b) A package raises government spending by 40 and cuts taxes by 30; what is the maximum combined boost to GDP? (c) The country's debt is 5,000 and GDP is 8,000. This year it runs a deficit of 240 while nominal GDP grows 4 percent. Compute the debt-to-GDP ratio before and after. (d) Rerun (c) with a deficit of 80. What changed, and why?

Answer: (a) Spending multiplier = 1 / 0.2 = 5; tax multiplier = 0.8 / 0.2 = 4. (b) 40 x 5 = 200 from spending, 30 x 4 = 120 from the tax cut: up to 320 in total. (c) Before: 5,000 / 8,000 = 62.5 percent. After: debt 5,240, GDP 8,320, ratio about 63 percent - the burden edged up. (d) Debt becomes 5,080 against GDP 8,320, about 61 percent. The debt still grew, but growth outpaced borrowing, so the burden fell - the stock rose while the ratio declined.

Sources

1. Greenlaw, S. A., Shapiro, D., & MacDonald, D. (2022). Federal deficits and the national debt. In Principles of Macroeconomics 3e. OpenStax, Rice University. openstax.org
2. Greenlaw, S. A., Shapiro, D., & MacDonald, D. (2022). Using fiscal policy to fight recession, unemployment, and inflation. In Principles of Macroeconomics 3e. OpenStax, Rice University. openstax.org
3. Greenlaw, S. A., Shapiro, D., & MacDonald, D. (2022). Automatic stabilizers. In Principles of Macroeconomics 3e. OpenStax, Rice University. openstax.org
4. Greenlaw, S. A., Shapiro, D., & MacDonald, D. (2022). Practical problems with discretionary fiscal policy. In Principles of Macroeconomics 3e. OpenStax, Rice University. openstax.org
5. Weil, D. N. (2008). Fiscal policy. In The Concise Encyclopedia of Economics (2nd ed.). Liberty Fund. econlib.org
6. U.S. Department of the Treasury. (n.d.). Understanding the national debt. In Fiscal Data: America's Finance Guide. Bureau of the Fiscal Service. fiscaldata.treasury.gov
7. The CORE Team. (2017). Unemployment and fiscal policy. In The Economy 1.0. CORE Economics Education. core-econ.org

Why nations trade

Countries trade for the same reason individuals do: comparative advantage. By specializing in what it produces at the lowest opportunity cost and trading for the rest, each nation can consume more than it could in isolation. Trade expands the total economic pie, even when one country is more productive at everything. The gains are real, though they are unevenly distributed, which fuels political debate.

The idea dates to David Ricardo in 1817, and its punchline still surprises: absolute ability is beside the point. A country that is worse at producing everything still has a comparative advantage in whatever it is least bad at, because comparative advantage is about opportunity cost - what must be given up at home to make one more unit - and opportunity costs cannot be lowest everywhere at once. The same logic explains why a surgeon who types faster than her assistant still hires the assistant: an hour of her typing costs an hour of surgery.

Comparative advantage, worked through

Run the argument with two countries and two goods. Aland has 80 hours of labor; a ton of wheat takes 2 hours and a bolt of cloth takes 4. Bexley has 60 hours; wheat takes 6 hours and cloth takes 3. In isolation, suppose each splits its hours evenly. Aland makes 20 wheat and 10 cloth; Bexley makes 5 wheat and 10 cloth. World output: 25 wheat and 20 cloth.

Now compute opportunity costs. In Aland a bolt of cloth takes twice the hours of a ton of wheat, so one cloth costs 2 wheat, and one wheat costs 0.5 cloth. In Bexley the ratios flip: one cloth costs only 0.5 wheat, and one wheat costs 2 cloth. Aland is the cheap place to make wheat; Bexley is the cheap place to make cloth. Let each specialize fully: Aland's 80 hours yield 40 wheat, Bexley's 60 hours yield 20 cloth. World output is now 40 wheat and 20 cloth - fifteen more wheat than before, from the same total labor, with no new technology.

Trade divides the winnings. Suppose they settle on a price of 1.2 wheat per bolt - between the two opportunity costs, so both sides gain - and Aland ships 12 wheat for 10 cloth. Aland ends with 28 wheat and 10 cloth, against 20 and 10 in isolation. Bexley ends with 12 wheat and 10 cloth, against 5 and 10. Each country consumes strictly more than it could produce alone. And the logic survives lopsided productivity: make Bexley worse at both goods and its opportunity cost of cloth can still be the lower one, so the same trade still pays.

The theory promises bigger totals, not painless transitions. Aland's cloth weavers and Bexley's wheat farmers must switch industries, and the burden of that switch falls on specific towns and careers while the gains spread thinly over every consumer. Concentrated losses and diffuse gains are the standing political arithmetic of trade - visible in every era's tariff debates - and honest economics keeps both sides of it on the page.

Trade barriers

Governments often limit trade despite these gains. A tariff is a tax on imports; a quota is a legal limit on the quantity imported. A tariff raises the price of imported goods, which helps protected domestic producers and raises revenue for the government, but forces consumers to pay more and typically shrinks total surplus. Trade barriers are usually defended on grounds such as protecting jobs or infant industries, but economists generally find their costs to consumers exceed their benefits to protected groups.

Follow a tariff's arithmetic. Suppose imported shirts cost 10 and a tariff of 2 lifts the domestic price toward 12. Domestic shirt makers who could produce at 11 now enter, selling at the higher price; the treasury collects 2 on each shirt still imported; and every consumer pays more for every shirt. Some purchases stop happening at all - value destroyed rather than transferred, the deadweight loss. The gains to producers and the treasury are real, but they sum to less than what buyers lose.

History supplies the cautionary tale. The Smoot-Hawley tariff of 1930 raised US import duties across the board just as the Depression began; trading partners retaliated in kind, and world trade spiraled downward - shrinking to a fraction of its former value as the contraction and the tariff war fed each other. The lesson stuck: after the Second World War the major economies built the GATT, later the World Trade Organization, to negotiate tariffs downward by mutual agreement, and world trade grew for decades faster than world output.

Measuring trade: the balance of trade

The balance of trade is the value of a country's exports minus its imports - the same net-exports term from GDP.

• A trade surplus occurs when exports exceed imports (net exports positive).
• A trade deficit occurs when imports exceed exports (net exports negative).

The broader current account adds income and transfers to the trade balance. A trade deficit is not automatically "bad" - it can reflect strong domestic demand or foreign investment flowing in - but persistent large imbalances are watched closely.

The other side of the ledger: capital flows

Trade flows and capital flows are one transaction seen from two sides. A country that imports more than it exports must hand over something for the difference, and that something is claims: its currency, its bonds, its stocks, its real estate. So a current account deficit is matched, by accounting necessity, by a net inflow of capital - foreigners accumulating assets in the deficit country. Run a surplus and the mirror holds: the surplus country accumulates claims on the rest of the world.

The same identity wears a second outfit: net exports equal national saving minus domestic investment. An economy that invests more than it saves must borrow the difference abroad, which appears as a trade deficit; one that saves more than it invests lends the excess abroad, which appears as a surplus. This is why economists shrug at bilateral trade targets: unless a policy changes a nation's saving or its investment, the overall balance has nowhere to move. The identity does not say deficits are good or bad - it says they are financing arrangements, to be judged by what the borrowed resources build.

Exchange rates

Trade across borders requires exchanging currencies, and the exchange rate is the price of one currency in terms of another. In a floating system, exchange rates are set by supply and demand for currencies in the foreign-exchange market: exporters and inbound investors demand the home currency, importers and outbound investors supply it. Two key terms:

• Appreciation - a currency gains value (buys more foreign currency). A stronger currency makes imports cheaper for domestic buyers but makes the country's exports more expensive abroad.
• Depreciation - a currency loses value. A weaker currency makes exports cheaper and more competitive abroad but raises the cost of imports.

Nominal versus real exchange rates

The market quote - so many yen per dollar - is the nominal exchange rate. Competitiveness, though, depends on the real exchange rate: how many foreign baskets of goods one home basket exchanges for, once prices on both sides are counted. The recipe: convert the home basket into foreign currency at the nominal rate, then divide by the foreign basket's price.

Work it once. A home basket costs 200 dollars; a comparable foreign basket costs 24,000 yen; the nominal rate is 100 yen per dollar. The home basket converts to 20,000 yen, so the real rate is 20,000 / 24,000, about 0.83 foreign baskets per home basket - home goods are modestly cheaper than their foreign rivals.

Now let home prices rise 10 percent while the nominal rate and foreign prices sit still. The basket costs 220 dollars, or 22,000 yen, and the real rate climbs to about 0.92. Nothing moved in the currency market, yet home goods became dearer abroad: a real appreciation through inflation alone. Domestic inflation can price exporters out of world markets as surely as a rising currency can.

Worked reasoning

Suppose the exchange rate moves from 1 dollar = 100 yen to 1 dollar = 120 yen. The dollar now buys more yen, so the dollar has appreciated (and the yen has depreciated). A Japanese car priced at 2,400,000 yen falls in dollar terms from 2,400,000 / 100 = $24,000 to 2,400,000 / 120 = $20,000, so it is now cheaper for Americans - imports become more attractive.

At the same time, a US product priced at $1,000 rises for Japanese buyers from 100,000 yen to 120,000 yen, making US exports harder to sell. This is the core link between currency values and trade flows: a stronger currency tends to widen a trade deficit, a weaker one tends to narrow it. Exchange rates thus tie together everything in this course - trade, interest rates, inflation, and policy - into a single interconnected global economy.

Fixed versus floating: the Bretton Woods story

Should a currency's price float like any other, or be fixed by the state? The twentieth century tried both at scale. In July 1944, with the war still on, delegates from 44 nations met at Bretton Woods, New Hampshire, to design a postwar order that would avoid the competitive devaluations of the 1930s. The system they built pegged the dollar to gold at 35 dollars an ounce, pegged every other member currency to the dollar, and created the International Monetary Fund to lend to countries defending their pegs.

The arrangement anchored a generation of reconstruction and booming trade, but its core tension grew: pegged countries had to aim monetary policy at the peg rather than at their own economies, and the United States, printing the anchor currency, gradually issued more dollars than its gold could honor. As foreign claims on that gold mounted, President Nixon suspended convertibility in August 1971, and by 1973 the major currencies floated. The float never spread everywhere - many countries still peg to an anchor currency, run currency boards, or share the euro - so the design question remains live.

The modern verdict is a trade-off, not a winner. Floating rates give a country its own monetary policy and let the currency absorb shocks, at the price of day-to-day volatility. Fixed rates deliver stability and imported credibility - valuable for small, open, or inflation-scarred economies - at the price of surrendering monetary independence, and pegs that lose credibility invite speculative attack. A country can hold a fixed rate, free capital flows, and an independent monetary policy two at a time, never all three; every currency regime is a choice of which one to give up.

Common misconceptions

• "Exports are winning; imports are losing." Imports are the goods and services a country actually gets to consume; exports are what it gives up to pay for them. The gains from trade arrive largely through imports.
• "A trade deficit means the country is losing money." The deficit is matched by capital inflow - foreigners investing in the deficit country. Whether that is worrying depends on what the inflow finances.
• "A strong currency is always good news." Appreciation helps consumers, travelers, and importers while squeezing exporters. Strong and weak are directions, not grades.
• "Tariffs are paid by the foreign country." The tax is collected at the border, but the burden lands mostly on domestic buyers through higher prices.
• "You need to be the best at something to trade." Comparative advantage runs on opportunity cost, so even a country worse at everything has something it is relatively least bad at producing.

Try it

(a) Norland makes a ton of grain in 3 hours and a machine in 6; Sudland needs 4 hours for grain and 12 for a machine. Find each country's opportunity costs, assign comparative advantage, and propose a mutually beneficial price for a machine. (b) The exchange rate moves from 8 kronor per dollar to 10. Which currency appreciated? What happens to the kronor price of a $40 gadget exported from the dollar country, and the dollar price of a 200-kronor sweater?

Answer: (a) In Norland a machine costs 6 / 3 = 2 grain; in Sudland, 12 / 4 = 3 grain. Grain costs 0.5 machine in Norland and about 0.33 machine in Sudland. Norland has the cheaper machines, Sudland the cheaper grain - even though Norland is absolutely faster at both goods. Any machine price between 2 and 3 grain, say 2.5, leaves both better off. (b) The dollar appreciated (each buys 10 kronor instead of 8); the krona depreciated. The gadget rises from 320 to 400 kronor abroad, so dollar-country exports suffer; the sweater falls from $25 to $20, so its imports get cheaper.

Sources

1. Greenlaw, S. A., Shapiro, D., & MacDonald, D. (2022). Absolute and comparative advantage. In Principles of Macroeconomics 3e. OpenStax, Rice University. openstax.org
2. Greenlaw, S. A., Shapiro, D., & MacDonald, D. (2022). What happens when a country has an absolute advantage in all goods. In Principles of Macroeconomics 3e. OpenStax, Rice University. openstax.org
3. Greenlaw, S. A., Shapiro, D., & MacDonald, D. (2022). How the foreign exchange market works. In Principles of Macroeconomics 3e. OpenStax, Rice University. openstax.org
4. Greenlaw, S. A., Shapiro, D., & MacDonald, D. (2022). Measuring trade balances. In Principles of Macroeconomics 3e. OpenStax, Rice University. openstax.org
5. Ricardo, D. (1817). On the principles of political economy, and taxation. In Project Gutenberg. Project Gutenberg Literary Archive Foundation. gutenberg.org
6. Bhagwati, J. (2008). Protectionism. In The Concise Encyclopedia of Economics (2nd ed.). Liberty Fund. econlib.org
7. Ghizoni, S. K. (2013). Creation of the Bretton Woods system. In Federal Reserve History. Federal Reserve Bank of St. Louis. federalreservehistory.org