Unit 5 · Measuring the Economy

A bakery in Springfield buys $2 of flour, turns it into a loaf of bread and sells the loaf for $5. How much production happened? The tempting answer is $2 + $5 = $7. But the flour is inside the bread. If you count the flour and then the bread, you count the flour twice. Economists call the flour an intermediate good, something bought to be used up in making something else. The loaf is a final good, sold to the person who will actually eat it. Only the $5 counts.

Gross domestic product, or GDP, is the total market value of all final goods and services produced within a country’s borders in a given period, usually a year or a quarter. Every word in that sentence does a job. Market value means each item is measured by its price, so a $40,000 car and 40,000 one-dollar sandwiches add the same amount. Final means intermediate goods are left out to avoid double counting. Produced means the item was made during that period, not just sold. Within the borders means the location of production is what matters, not who owns the business.

The market-value rule is what lets us add cars to haircuts to soybeans. You cannot add 3 cars and 500 haircuts and get a meaningful total. You can add $90,000 of cars and $15,000 of haircuts and get $105,000. Prices do the work of turning different things into one number. That is also GDP’s first weakness: anything without a price is invisible to it, as the next lessons show.

One more example to test the rule. A farmer sells $1,000 of corn to a feed mill, the mill sells $1,500 of feed to a hog farm, and the hog farm sells $4,000 of pork to a grocery store, which sells it to shoppers for $6,000. GDP counts only the $6,000, the final sale to the people who will eat it. Notice that $6,000 also equals the value each business added along the way: $1,000 + $500 + $2,500 + $2,000. Counting only final sales and adding up value added give the same answer.

A Japanese company builds cars at a plant in Indiana. A Chicago company owns a factory in Mexico. Which cars count in U.S. GDP? The Indiana cars do, and the Mexico cars do not. The D in GDP stands for domestic, meaning inside the borders. It does not matter who owns the plant. What matters is where the work happened. The Indiana plant hires Indiana workers, buys from Indiana suppliers and pays Indiana taxes, so its output is U.S. production.

The period matters too. A house built in 2015 and sold again in 2026 is not part of 2026 GDP, because nothing new was produced this year. The house was already counted the year it was built. The same rule applies to a used car, a secondhand bicycle or a vintage guitar. However, the real estate agent’s fee for selling that house does count, because the agent performed a new service this year. Used goods are out; new services attached to them are in.

Some dollar flows look like production but are not. When you buy 10 shares of a company’s stock for $1,500, nothing has been produced. Ownership of an existing asset simply changed hands, so the $1,500 is not in GDP. The broker’s $10 fee is, since that is a service. When the government sends a retiree a Social Security payment or a family an unemployment check, that is a transfer payment: money moved from one group to another with no good or service produced in return. Transfers are not in GDP either, though what the retiree buys with the money will be.

Put the rules together on one case. In one year, a Peoria factory makes $50 million of engines, a local family sells its old boat for $8,000, a worker receives $20,000 in disability payments, and a Peoria firm’s plant in Canada makes $30 million of parts. Only the $50 million of engines counts in U.S. GDP. The boat is used, the disability payment is a transfer, and the Canadian plant is outside the borders.

When you pay $5 for that loaf of bread, where does the money go? Part pays the baker’s wages, part pays rent on the shop, part pays interest on the oven loan, part pays the flour mill, and whatever is left is the owner’s profit. Trace the flour mill’s $2 the same way and it splits into wages, rent, interest and profit too. Every dollar of spending on a final good ends up as somebody’s income. That is not a coincidence; it is accounting. One person’s spending is another person’s earnings.

This gives economists two independent ways to measure the same GDP. The expenditure approach adds up everything spent on final goods and services. The income approach adds up everything earned in producing them: wages, rent, interest and profit. In a perfect set of books the two totals match exactly. In practice the government’s two estimates differ by a small amount, and the gap is reported honestly as a statistical discrepancy. The next section walks through the expenditure approach, which is the one most textbooks and news reports use.

Economists picture this as a circular flow. Households own the labor and the savings that businesses use. Businesses pay households wages, rent, interest and profit. Households spend that income buying goods and services from businesses, and the money flows around again. In a year, the total flow of spending on final goods equals the total flow of income, and both equal GDP. When a news report says GDP grew 3%, it also means total income earned in the country grew about 3%.

Think about who buys final goods in Millbrook. Families buy groceries, haircuts, gas and phones. The parts factory buys a new stamping machine and builds a warehouse. The school district buys buses and pays teachers. And a truck maker in Texas buys the factory’s parts. Economists group all spending into four categories named after these buyers. The formula is GDP = C + I + G + NX. This lesson covers the first three.

Consumption, written C, is household spending on goods and services: food, rent, clothing, health care, streaming subscriptions, restaurant meals, new cars. It is by far the largest piece of U.S. GDP, roughly two-thirds of the total. Investment, written I, does not mean buying stocks. In GDP it means business spending on new capital: machines, tools, buildings, software, plus new houses and changes in inventories. When a farmer buys a new combine or a company builds a data center, that is investment. When you buy shares of that company, that is not.

Government purchases, written G, are what federal, state and local governments spend on goods and services: teachers’ and soldiers’ pay, highways, fighter jets, school lunches, police cars. Transfer payments are left out here because nothing is produced in exchange. So Social Security and unemployment benefits are not in G, even though they are a large part of the federal budget. When the retiree spends the check on groceries, that shows up in C instead.

A quick sort. A family buys a $30,000 car: C. A delivery company buys the same model for its fleet: I, because it is business capital. The city buys it for the police: G. The distinction is not about the car but about who bought it and why. Common mistake: counting a company’s purchase of another company’s stock as investment. It is a change of ownership, not new capital, and it is not in GDP at all.

Millbrook’s parts go to a truck maker in Texas, and if that truck maker were in Canada instead, the sale would be an export. Exports, written X, are goods and services made here and sold to buyers in other countries. They are U.S. production, so they belong in U.S. GDP. Imports, written M, are goods and services made abroad and bought here. They are not U.S. production, but they are hiding inside C, I and G. When a family buys a phone made in Vietnam, that $800 went into C. To get an honest total of what was made here, we must take it back out. Net exports, NX, equals X minus M.

Now the whole formula: GDP = C + I + G + (X - M). Try it on Millbrook for a year, with invented numbers. Households spend $60 million (C). Businesses spend $15 million on new machines, buildings and inventory (I). Government buys $20 million of goods and services (G). Firms sell $12 million to buyers outside the country (X) and residents buy $17 million of foreign-made goods (M). NX = 12 - 17 = -5. GDP = 60 + 15 + 20 - 5 = $90 million.

Check the logic on the -5. If the class had just added C, I and G, it would have gotten $95 million. But $17 million of that spending bought things made somewhere else. Subtracting imports removes them. Adding exports puts back the $12 million of local production that local people never bought. The result, $90 million, is what Millbrook actually produced. When imports exceed exports, as they have for the United States for decades, net exports are negative and pull the total down slightly.

A common mistake is to say imports hurt GDP. Subtracting M does not punish imports; it just corrects for the fact that they were already counted in C, I and G. If a family spends $800 on an imported phone instead of an $800 domestic one, C is the same, and M goes up by $800, so GDP is $800 lower. But that $800 was never U.S. production in the first place. The subtraction is bookkeeping, not a judgment.

Scale Millbrook up to the whole country and the numbers get very large. U.S. GDP in the mid-2020s is roughly $28 trillion a year. A trillion is a thousand billion, or a million million. Written out, $28 trillion is $28,000,000,000,000. Divided among about 340 million people, that is roughly $80,000 of production per person per year. No other country produces more in total, though a few smaller countries produce more per person.

The shares of that total are steadier than the total itself. In a typical recent year, consumption is about 68% of GDP, investment about 18%, government purchases about 17%, and net exports about -3%. Check that they sum to 100: 68 + 18 + 17 - 3 = 100. The negative net exports reflect a trade deficit that the United States has run since the 1970s. The exact shares move a point or two from year to year and shift more during recessions, when investment usually falls hardest.

The Bureau of Economic Analysis, part of the U.S. Department of Commerce, publishes GDP every quarter. The first estimate for a quarter comes out about a month after the quarter ends and is revised twice as more data arrive. Quarterly figures are usually reported as an annualized growth rate, meaning the rate the economy would grow in a full year if the quarter’s pace continued. A headline like GDP grew at a 2.4% annual rate means the quarter’s growth, stretched out to a year, would be 2.4%.

Illinois is a large piece of the national total. State-level GDP, also published by the BEA, puts Illinois among the top five or six states, with a state economy of roughly a trillion dollars. Its biggest pieces are services such as finance, health care and professional work centered in the Chicago area, along with manufacturing and agriculture across the state.

Picture a tiny economy that makes only two things: lemonade and sandwiches. In Year 1 it makes 100 cups of lemonade at $2 each and 50 sandwiches at $4 each. Total value: 100 x 2 + 50 x 4 = $200 + $200 = $400. In Year 2 it makes exactly the same amounts, 100 cups and 50 sandwiches, but lemonade now costs $2.50 and sandwiches $5. Total value: 100 x 2.50 + 50 x 5 = $250 + $250 = $500. The dollar total rose 25%. Did the economy produce 25% more? No. It produced exactly the same things. Only the prices changed.

GDP measured in the prices of the year it happened is called nominal GDP. It mixes two things: how much was made and what prices were. To see production alone, economists compute real GDP: output valued at the prices of one fixed base year. Using Year 1 as the base, Year 2 real GDP is 100 x $2 + 50 x $4 = $400, the same as Year 1. Real GDP did not grow at all, which is the truth about this economy.

Now change the story. Suppose in Year 2 the economy made 110 cups and 55 sandwiches at the higher prices. Nominal GDP = 110 x 2.50 + 55 x 5 = 275 + 275 = $550. Real GDP in Year 1 prices = 110 x 2 + 55 x 4 = 220 + 220 = $440. Real GDP grew from 400 to 440, a 10% gain in actual output. Nominal grew 37.5%, but most of that was price. Whenever you see a GDP figure, ask which one it is. Growth over time is almost always reported in real terms.

The common mistake is to compare nominal figures from different years and call the difference growth. If nominal GDP rose from $20 trillion to $22 trillion while prices rose 10%, real output did not grow at all, because 22 divided by 1.10 is 20. Real GDP is the number that tells you whether people actually have more goods and services than before.

Doing the two-price calculation for every good in a country is impossible by hand, so economists build a shortcut called a price index. A price index compares the overall price level in one year to the price level in a base year, with the base year set to 100. If the index is 110, prices overall are 10% higher than in the base year. If it is 95, they are 5% lower. The index used for GDP is called the GDP deflator, and it covers everything in GDP: consumer goods, machines, government purchases and exports.

With an index in hand, the conversion is one step: real GDP = nominal GDP divided by the price index, times 100. Try it. Nominal GDP is $22 trillion and the deflator is 110. Real GDP = 22 divided by 110, times 100 = $20 trillion. The economy produced $20 trillion of goods measured in base-year prices, and the other $2 trillion of the nominal figure was price increase. Do it again with a deflator of 125 and nominal GDP of $25 trillion: 25 divided by 125, times 100 = $20 trillion. Same real output, higher prices.

You can also run the formula backward to find the deflator. Go back to the lemonade economy. Year 2 nominal GDP was $550 and real GDP in Year 1 prices was $440. The deflator = nominal divided by real, times 100 = 550 divided by 440, times 100 = 125. Prices in Year 2 averaged 25% above Year 1. Notice that the base year always has an index of exactly 100, because nominal and real are the same number in that year.

A handy rough rule follows from all this. The growth rate of real GDP is close to the growth rate of nominal GDP minus the inflation rate. If nominal GDP grew 6% and prices rose 4%, real GDP grew about 2%. The rule is approximate, and the exact answer uses division, but for quick reading of the news it works well.

Suppose real GDP was $20.0 trillion last year and $20.5 trillion this year. The growth rate is the percent change: (20.5 - 20.0) divided by 20.0, times 100 = 0.5 divided by 20, times 100 = 2.5%. That is a fairly normal year for the United States. Over the long run, U.S. real GDP has grown around 2% to 3% per year on average, with faster years in booms and shrinking years in recessions. A year when real GDP falls is a serious event; two quarters of falling real GDP is a common rough sign of a recession.

Small rates matter because they compound. An economy growing 2% a year doubles in about 35 years. One growing 3% a year doubles in about 24 years. The quick tool for this is the rule of 70: divide 70 by the growth rate to get the approximate doubling time. 70 divided by 2 is 35; 70 divided by 3 is about 23. Over a working lifetime of 45 years, the difference between 2% and 3% growth is the difference between an economy that is 2.4 times bigger and one that is 3.8 times bigger.

Where does long-run growth come from? Economists point to productivity, output per hour of work. Productivity rises when workers have more and better capital, when they have more education and skill, and when new technology lets the same hour produce more. A farmer in central Illinois today, with a GPS-guided combine, harvests more corn in a day than a crew with horses harvested in a season a century ago. Nearly all of the rise in living standards since then comes from that kind of gain.

One caution when computing. Always divide the change by the starting value, not the ending one. Going from 20.0 to 20.5 is a 2.5% increase. Going from 20.5 back to 20.0 is a 2.4% decrease, because the base is now 20.5. And a 10% fall followed by a 10% rise does not get you back to the start: 100 falls to 90, then rises 10% of 90, which is 9, to 99.

Country A has a GDP of $2,000 billion and 50 million people. Country B has a GDP of $3,000 billion and 300 million people. Which country is richer? In total output, B, by a lot. But B’s output is spread over six times as many people. GDP per capita means GDP per person: divide GDP by population. Country A: 2,000 billion divided by 50 million = $40,000 per person. Country B: 3,000 billion divided by 300 million = $10,000 per person. The average person in A produces, and on average earns, four times as much.

Per capita figures also fix a problem with growth over time. If real GDP grows 1% but population grows 2%, output per person actually fell. Economists therefore track real GDP per capita, which combines both adjustments: fixed prices and per-person division. That is the closest thing GDP offers to a measure of the average standard of living. U.S. real GDP per capita has roughly doubled since the early 1980s.

Comparing across countries adds one more wrinkle. A dollar buys different amounts in different places. A haircut might cost $30 in Chicago and the equivalent of $4 in a low-income country. Converting at market exchange rates would make that country look poorer than it really is. So economists adjust for purchasing power parity, or PPP, which values each country’s output at a common set of prices. PPP comparisons narrow the gap between rich and poor countries but do not close it.

Even per capita GDP is an average, and averages hide spread. Two countries with the same GDP per capita can look very different if one spreads income widely and the other concentrates it in a few hands. The last section of this chapter takes up that limit and the others.

Back in Millbrook, Mr. Alvarez fixed his own roof and saved $4,000. Nothing was sold, so GDP did not move. If he had hired a roofer, GDP would have risen by $4,000, and the roof would be exactly the same. This is GDP’s largest blind spot: unpaid work. Cooking, cleaning, child care, elder care, repairs and volunteering produce real value every day, but with no market price they do not count. When a parent leaves a job to care for children, measured GDP falls, though the same care is still being given.

The lawn-mowing kid who never wrote anything down is part of a second gap, the underground economy: cash work that goes unreported, along with illegal trade. It is real production, but the government cannot count what it cannot see. Estimates for the United States put the underground economy at perhaps a tenth of GDP or less; in some countries it is far larger. When a country’s rules push more work off the books, its measured GDP shrinks even if actual output does not.

GDP also ignores leisure. Suppose everyone in the country worked ten more hours a week. GDP would jump. Would people be better off? They would have more goods and far less time. Two countries can have the same GDP per person, one working 30 hours a week and the other 50, and GDP will call them equal. Time off is something people value enormously, and GDP places its value at zero.

These gaps make comparisons across time tricky. Over the past century a great deal of work moved from homes into markets: meals from kitchens to restaurants, child care from relatives to day-care centers. Part of the measured growth in GDP is that shift, not new production. None of this makes GDP useless. It means GDP measures market production, exactly as defined, and nothing more.

Suppose a factory produces $10 million of chemicals and, in doing so, pollutes a river that the town then spends $3 million to clean up. GDP records $10 million of chemicals plus $3 million of cleanup services: $13 million. The pollution made GDP larger, not smaller. GDP counts production and never subtracts the damage that production causes to air, water, forests or climate. A country can log every forest, empty every fishery and post record GDP while doing it. Economists call this using up natural capital without recording the loss.

The same blind spot appears after disasters. A tornado that destroys 200 homes in a town does not lower GDP. The rebuilding raises it. People are clearly worse off, since they had homes and now must spend to get them back, but the accounts see only the new construction. GDP measures the flow of new production, not the stock of things people own or the state of the world they live in.

GDP also says nothing about distribution: who gets the output. A country’s GDP per capita can rise 20% while most families see no change. That happens if all the gains go to a small group at the top. An average of $80,000 per person fits two very different countries. In one, almost everyone earns near $80,000. In the other, a few earn millions while many earn much less. To see how a typical person is doing, economists look at median income. That is the income of the household right in the middle, and it rises only when the middle rises.

Put these limits next to the ones from the last lesson and you have the standard list of what GDP misses. Unpaid work, the underground economy, leisure, the environment and distribution. Add quality of life itself: health, safety and education. GDP touches those only through what is spent on them. The measure was never designed to capture them, as its own inventor pointed out from the very beginning.

Because GDP leaves so much out, economists and statistical agencies have built other yardsticks. None replaces GDP; each answers a different question. The best known is the Human Development Index, published by the United Nations since 1990. It combines three things for each country: life expectancy, years of schooling, and income per person. A country with high income but short lives and little schooling scores lower than its GDP alone would suggest. The index reminds readers that money is a means, and health and knowledge are part of the end.

To see distribution, statisticians report median income alongside the average, and they compute the Gini coefficient, a number from 0 to 1 that summarizes how unequal incomes are. A Gini of 0 would mean everyone has the same income; a Gini of 1 would mean one person has everything. Real countries fall in between, with more equal countries around 0.25 to 0.30 and less equal ones above 0.45. The United States sits in the upper part of the range among wealthy countries.

Other efforts try to correct GDP itself. So-called green accounting subtracts the estimated cost of pollution and the depletion of natural resources. Some researchers add estimates of the value of unpaid household work, which can be a large fraction of measured GDP. Surveys of well-being ask people directly how satisfied they are with their lives and compare the answers across countries and over time. Each of these involves judgment calls that GDP avoids, which is why GDP remains the common starting point.

The practical lesson for a reader of the news is to match the number to the question. Asking whether the economy grew last quarter? Real GDP. Asking whether the typical family is better off than ten years ago? Real median income. Asking whether a country’s people are healthy and educated? HDI. Asking who gained from growth? Median income and the Gini. GDP is the wrong answer to a lot of questions only when it is the only number anyone looks at.

Suppose a student named Jordan buys the same things every month: 10 gallons of gas, 4 pizzas and 2 movie tickets. In the base year, gas is $3 a gallon, pizza $10 and a ticket $15. The basket costs 10 x $3 + 4 x $10 + 2 x $15 = $30 + $40 + $30 = $100. A year later gas is $3.60, pizza $10.50 and a ticket $16. The same basket now costs $36 + $42 + $32 = $110. Jordan buys exactly the same things and pays $10 more.

A market basket is a fixed list of goods and services in fixed amounts, chosen to stand for what a typical buyer purchases. A price index turns the basket’s cost into one simple number. The rule: index = cost of the basket this year divided by its cost in the base year, times 100. For Jordan, 110 divided by 100, times 100 = 110. The base year always equals 100, so an index of 110 means prices are 10% above the base year.

The Consumer Price Index, or CPI, uses the same method for the whole country. The Bureau of Labor Statistics, part of the U.S. Department of Labor, collects tens of thousands of prices each month in cities across the country. It prices food, rent, gas, clothing, doctor visits, car insurance and much more. Each item is weighted by the share of household budgets spent on it. So a rise in rent moves the index far more than a rise in the price of salt. The BLS publishes the CPI every month.

The official CPI uses the years 1982 to 1984 as its base, set to 100. In 2024 the index stood a little above 310. That means a basket that cost $100 in the early 1980s cost a little over $310 in 2024. A common mistake is to read an index number as a price. A CPI of 310 does not mean anything costs $310. It means prices are about 3.1 times their base-period level.

Jordan’s basket index went from 100 to 110 in one year. The inflation rate is the percent change in the price level: (110 - 100) divided by 100, times 100 = 10%. Now go one more year. The index rises from 110 to 121. The change is 11 points, but the inflation rate is 11 divided by 110, times 100 = 10%, not 11%. Always divide the change by the starting index. Subtracting index points works only when the starting index is exactly 100.

Real figures work the same way. If the CPI was 300 last June and 312 this June, inflation over those twelve months was 12 divided by 300, times 100 = 4%. News reports usually give this 12-month rate. They also report core inflation, which leaves out food and energy. Those two prices jump around with weather, wars and oil supplies. Leaving them out shows the steadier trend underneath.

In economics, inflation is a rise in the general level of prices, not a rise in one price. Suppose avocados jump after a bad harvest while most other prices hold still. That is a change in a relative price, not inflation. Inflation means the whole basket costs more, so each dollar buys less. The Federal Reserve aims for inflation of 2% a year over time. It treats a small, steady rate as healthier than either high inflation or falling prices.

U.S. inflation has swung widely. It reached about 13.5% for the year 1980, then fell during the 1980s. From the mid-1990s through 2020 it mostly stayed between about 1% and 4%. It jumped again in 2021 and 2022, and the 12-month rate peaked near 9% in mid-2022 before coming down. Knowing these rough figures helps you judge whether a year’s number is ordinary or alarming.

Suppose beef jumps from $6 to $9 a pound while chicken stays at $4. Many families switch from beef to chicken. A fixed basket still assumes they buy the old amount of beef. So it shows their costs rising more than they really did. This is substitution bias. The BLS reduces it by updating the basket’s weights regularly and by allowing for some switching within categories, such as from one kind of apple to another. It cannot remove the problem completely.

Quality is the second problem. A new phone may cost $800, the same as a phone five years ago, but it has a better camera, more storage and a faster chip. If the price is the same and the phone is better, the true price of what you get has fallen. The BLS makes a quality adjustment for some goods, but valuing an improvement is hard. New goods raise a third problem. When a new kind of product appears, it takes time to enter the basket, so its arrival does not register as a gain.

Economists have long argued that together these effects make the CPI overstate the true rise in the cost of living, by a fraction of a percentage point a year. That matters because the CPI adjusts real payments. Social Security benefits rise each year with a version of the CPI. Federal income tax brackets are adjusted with a version of it too. A CPI that runs a little high raises those payments a little faster than living costs.

Finally, the CPI is an average, and no family is average. A renter in Chicago whose rent jumps 8% faces higher inflation than a homeowner with a paid-off house. A long commute makes gas prices matter more. An older person spends more on health care. When your own costs rise faster or slower than the CPI, the index is not wrong. Your basket is simply different from the national one.

A dollar’s purchasing power is how much a dollar can buy. When prices rise 25%, a dollar buys only about four-fifths of what it used to, because 1 divided by 1.25 = 0.80. Inflation does not take dollars out of your wallet. It shrinks what each dollar will get you. That is why comparing a wage from 1968 with a wage today is like comparing two different currencies. You have to convert first.

The conversion uses the CPI: value in today’s dollars = old amount x (CPI today divided by CPI then). The CPI averaged about 35 in 1968 and about 313 in 2024. So prices rose about 313 divided by 35, or roughly 8.9 times. Grandpa’s $1.60 an hour x 8.9 = about $14.24 in 2024 dollars. The federal minimum wage of $7.25, unchanged since 2009, buys only about half of what the 1968 minimum bought. Illinois set its own minimum at $15 an hour in 2025.

A nominal wage is pay in the dollars of the day. A real wage is pay adjusted for prices, which shows what the pay can buy. Suppose Ana’s wage rises from $15.00 to $15.60 an hour, a 4% raise, while prices rise 6%. Her nominal wage went up, but her real wage fell about 2%. In last year’s dollars, $15.60 divided by 1.06 is about $14.72, less than the $15.00 she earned before. The quick rule: real wage growth is about nominal wage growth minus inflation.

The common mistake is to call any raise a gain. A raise smaller than inflation is a pay cut in what matters: goods and services. That is why many workers push for a cost-of-living adjustment, a raise tied to the CPI. It is also why economists compare wages over time only in real terms. In 2021 and 2022 many workers got their biggest raises in years. Yet for much of that time prices rose even faster, so real wages fell for many of them.

Suppose you lend a friend $1,000 for one year at 5% interest. You expect to get back $1,050. If prices rise 2% that year, your gain in buying power is about 3%. If prices unexpectedly rise 7%, the $1,050 you get back buys less than the $1,000 did when you lent it. The nominal interest rate was 5%. The real interest rate, the nominal rate minus inflation, was about 5 - 7 = -2%. The borrower won and the lender lost.

That is the core pattern: unexpected inflation helps people who owe fixed dollar amounts and hurts people who are owed them. A family with a 30-year fixed-rate mortgage repays with dollars that are worth less each year, which helps them. The bank that made the loan loses. Savers lose too. Put $10,000 in an account paying 1% while inflation runs 4%. It earns $100 in interest while prices rise about $400 worth, a real loss of about $300.

People on a fixed income also lose, such as a retiree whose pension pays the same $2,000 a month every year. Some incomes are protected. Social Security has a yearly cost-of-living adjustment tied to the CPI, and some union contracts include one. Workers whose raises keep pace with prices come out about even. The key word is unexpected. If everyone expects 7% inflation, lenders charge higher rates in advance, and the transfer mostly disappears.

Inflation also has costs that fall on everyone. Stores must reprint prices and menus, called menu costs. People spend time and effort guarding their cash. Businesses and families find it harder to plan when they cannot tell what a dollar will buy next year. High and changeable inflation makes long-term contracts risky. That is one reason central banks aim for low, steady inflation.

A fall in the general price level is called deflation, a negative inflation rate. At first it sounds like a gift: the same $100 buys more. But look at a family that owes $100,000 on a house. If prices and wages both fall 10%, the family’s income drops from, say, $50,000 to $45,000, while the debt stays at $100,000. The debt is now harder to carry. Across a whole economy, falling prices make debts heavier, and more people fall behind on loans.

Deflation also changes behavior. If you expect a $1,000 refrigerator to cost $950 next year, you may wait. When millions of people wait, stores sell less, cut jobs and cut prices again. Economists call this a deflationary spiral. It happened in the Great Depression, when U.S. consumer prices fell by roughly a quarter between 1929 and 1933. Do not confuse deflation with disinflation, which means inflation is slowing, say from 6% to 3%. Prices are still rising, only more slowly.

At the other extreme is hyperinflation, inflation so fast that prices climb by huge amounts each month or even each day. Germany in 1923 and Zimbabwe in the late 2000s are standard examples. In both cases the government paid its bills by printing money far faster than the economy grew. Workers rushed to spend their pay before it lost value. People turned to barter or foreign currency, and saving became pointless.

These two dangers help explain the Federal Reserve’s 2% target. A small positive rate leaves some room above zero, so a slump is less likely to tip into deflation. It is also low enough that people can plan without constantly thinking about prices. The lesson from both extremes is the same. A stable, predictable price level is part of what lets a market economy work.

Picture four people on one block in Aurora. Dana works 20 hours a week at a hardware store. Marcus lost his warehouse job and has sent out ten applications this month. Priya is a full-time college student who is not looking for work. Mr. Lopez retired last year. Only Marcus is unemployed in the official sense. Dana has a job, even though it is part-time. Priya and Mr. Lopez have no job, but they are not looking, so they are not in the labor force at all.

The rules come from the Current Population Survey, which the Census Bureau conducts for the Bureau of Labor Statistics. Each month it interviews about 60,000 households. People 16 and older, not in the military or in places such as prisons, are sorted into three groups. The employed did any work for pay in the survey week, or had a job but were out sick or on vacation. The unemployed had no job, were available to work, and actively looked in the past four weeks, or were waiting to be called back from a layoff.

The labor force is the employed plus the unemployed: everyone who has a job or is looking for one. Try the numbers for a town with 10,000 adults. If 6,000 are employed and 400 are unemployed, the labor force is 6,000 + 400 = 6,400. The other 3,600 are not in the labor force. They include students, retirees, stay-at-home parents, people too ill to work, and people who want a job but have stopped looking.

A common mistake is to think unemployed means anyone without a job. By that count Priya, Mr. Lopez and a stay-at-home father would all be unemployed, and the number would be huge and meaningless. The official definition asks two questions. Do you have a job? If not, are you looking? The survey counts actions, such as sending applications or contacting employers. Wishing for work is not enough.

Use the town from the last lesson: 6,000 employed, 400 unemployed, a labor force of 6,400. The unemployment rate = unemployed divided by labor force, times 100 = 400 divided by 6,400, times 100 = 6.25%. About 6 of every 100 people who want work and are looking cannot find it. The most common mistake is dividing by all 10,000 adults, which gives 4%. That understates the problem, because it treats retirees and students who are not looking as if they had jobs.

The BLS reports the national rate early each month, usually on the first Friday, in a release called the Employment Situation. The survey is a sample, so small moves of 0.1 or 0.2 points may be noise. Illinois and its counties and cities have their own rates, estimated each month by the Illinois Department of Employment Security working with the BLS. Local rates can differ a lot from the national one. A factory closing hits one county much harder than the whole country.

History gives a sense of scale. Unemployment reached about 25% in 1933, the worst of the Great Depression. It rose to nearly 11% in late 1982, and to about 10% in late 2009 after the financial crisis. In April 2020, as businesses shut down in the pandemic, it jumped to 14.7% in a single month. Then it fell faster than almost anyone expected. In strong years such as 2019 and 2023 the rate stayed below 4%.

Always check what a change is measuring. Suppose the number unemployed rises from 400 to 480 while the labor force stays at 6,400. The rate goes from 6.25% to 7.5%. That is a rise of 1.25 percentage points, but a 20% increase in the number of unemployed people. News stories sometimes mix those up. A percentage point is the difference between two rates. A percent change compares a new number with the old one.

Back in the town of 10,000 adults, 6,400 are in the labor force. The labor force participation rate = labor force divided by adult population, times 100 = 6,400 divided by 10,000, times 100 = 64%. It answers a different question from the unemployment rate: how many adults are taking part in the job market at all? In the United States the participation rate peaked a little above 67% around 2000. In recent years it has been near 62% to 63%.

Now suppose 100 of the 400 unemployed people give up. They have looked for months and believe no jobs are out there, so they stop searching. They are now discouraged workers, and they leave the labor force. The number unemployed falls to 300 and the labor force to 6,300. The unemployment rate drops to 300 divided by 6,300, times 100 = about 4.8%. Nobody found a job, yet the headline rate fell from 6.25% to 4.8%. The participation rate fell to 63%, and that is the clue.

The unemployment rate also misses underemployment. A worker who wants full-time hours but can find only 15 hours a week counts as employed. So does a college graduate in a job that needs no degree. The BLS publishes a broader measure, called U-6. It adds discouraged workers, others loosely attached to the job market, and people working part-time who want full-time work. U-6 always runs well above the official rate.

Why has participation fallen since 2000? The biggest single reason is age. The large baby boom generation has been retiring, and retirees leave the labor force. Young people staying in school longer also play a part. Economists debate how much of the rest reflects weak job prospects, health problems or other causes. The lesson for reading the news: look at unemployment and participation together, because either one alone can mislead.

Tanya finishes a nursing degree at Northern Illinois University in May. She applies to four hospitals and starts her first job in August. For those months she is unemployed, but nothing is wrong with the economy. She is searching for a good match, and hospitals are checking whether she fits. This is frictional unemployment: the time people spend moving between jobs or entering the workforce. It includes a worker who quits to find a better position and a parent returning to work after years at home.

Frictional unemployment is not all bad. A worker who takes the first offer may end up in a job that wastes her skills. Searching a few extra weeks can raise both her pay and her productivity. Job websites have made searching faster. Payments called unemployment insurance, made weekly to workers who lose a job through no fault of their own, give a laid-off worker time to find a better fit. Economists note that the same payments can also stretch out a search, and they disagree about how large that effect is.

Some joblessness follows the calendar; this is seasonal unemployment. Lifeguards at Lake Michigan beaches have no work in January. Construction slows in an Illinois winter. Farm workers are busiest at planting and harvest. Stores hire extra help for the holidays and let much of it go in January. The pattern repeats every year, so by itself it says little about the health of the economy.

That is why the BLS publishes a seasonally adjusted unemployment rate. Suppose, in an example, the raw rate usually rises from 3.8% in December to 4.4% in January as holiday jobs end. If this January’s raw rate is 4.4%, the adjusted rate shows no change, because the rise was the normal seasonal one. If the raw rate jumped to 5.0%, the extra 0.6 points would show up as real weakness. Headline numbers are almost always seasonally adjusted.

A steel mill closes and 800 workers lose their jobs. Across town, hospitals cannot fill openings for medical technicians, and software firms are hiring. There are jobs, just not jobs these workers are trained for or live near. This is structural unemployment: a lasting mismatch between the skills or locations of workers and the jobs available. It comes from changes in technology, in what consumers buy, and in where production happens. It can last for years, because retraining and moving take time and money.

Structural unemployment can be painful even when the overall economy is healthy. When many factories in the Midwest, including in Illinois, cut jobs in the 1980s, workers in their forties and fifties faced whole new careers. Responses include job-training programs, community college courses and help moving to where the jobs are. Economists debate how well such programs work. Some studies show solid gains in pay; others show little.

Another kind, cyclical unemployment, is different. It rises when the whole economy slows in a recession and falls when it recovers. Businesses sell less, so they need fewer workers in almost every industry at once. In 2008 and 2009, home building collapsed and then the slowdown spread. Carpenters, factory workers, salespeople and office staff lost jobs whose skills were still useful. The problem was not a mismatch but too little spending in the economy overall.

Telling the two apart matters for policy. If unemployment is cyclical, the fix is a stronger economy, through more spending, lower interest rates or simply time. Those tools of fiscal and monetary policy come in later units. If it is structural, a stronger economy helps less, and training or moving matters more. A common mistake is to blame every job loss on technology, or every one on recession. The same laid-off worker can face both at once.

If unemployment could fall to zero, every graduate would start work the day after graduation. No one would ever quit to look for something better, and no factory would ever close. That is not how a changing economy works. So economists define full employment as the level at which there is no cyclical unemployment. Frictional and structural unemployment still remain. The unemployment rate at full employment is called the natural rate of unemployment.

The natural rate cannot be measured directly; it has to be estimated. Many estimates for the United States in recent years fall roughly between 4% and 5%. If the actual rate is 7% and the natural rate is 4.5%, cyclical unemployment is about 7 - 4.5 = 2.5 percentage points. If the actual rate falls well below the natural rate, say to 3%, employers compete hard for scarce workers. They raise wages, and prices can start rising faster. That is one link between jobs and inflation.

Congress has made high employment a national goal. The Employment Act of 1946 committed the federal government to promote maximum employment. A 1977 amendment to the Federal Reserve Act directs the Fed to promote maximum employment and stable prices. This pair of goals is known as the Fed’s dual mandate. The two halves of this chapter, prices and jobs, are exactly the two things the Fed is told to watch.

The natural rate can change over time. Better job-matching websites may shorten frictional searches and lower it. A wave of factory closings may raise structural unemployment and push it up for a while. Economists disagree about how low unemployment can go before inflation speeds up. The late 2010s surprised many of them, when unemployment fell below 4% without much rise in inflation. Estimates of the natural rate are best read as ranges, not exact targets.