Chapter 2 — How Economists Think: Models, Assumptions, and the Art of Simplification

There is an old joke about economists. Three people — a physicist, a chemist, and an economist — are stranded on a desert island with nothing but a can of beans. The physicist says, "I can compute the trajectory of a falling rock to predict where the can will land if we throw it." The chemist says, "I can analyze the metallurgy of the can to figure out what tool we could shape from it." The economist says, "Assume a can opener."

The joke is funny because it captures something true. Economists assume things — sometimes outrageous things, sometimes obviously false things — for the sake of building tractable models. Then they reason inside the model. Then they hand the conclusions back to the world and act surprised when reality doesn't always cooperate.

The joke is also unfair. Every discipline that tries to think clearly about complicated systems uses models. Physicists assume frictionless surfaces and point masses. Biologists assume well-mixed populations of identical organisms. Doctors assume that the patient in front of them is roughly average for diagnostic purposes and adjust later. Engineers assume that bridges don't experience the once-in-a-thousand-years storm during their first decade of use. None of these assumptions is literally true. All of them are useful, because they let us think carefully about the parts of a problem that matter most without drowning in the parts that don't.

The questions, then, are not: "Should economists use models?" or "Are economic models true?" The right questions are: What are economic models good for? What are they bad at? When do they mislead? How do you know whether to trust the conclusions of a particular model? And why do reasonable, intelligent economists looking at the same models come to different conclusions about the same policy questions?

This chapter answers those questions. By the end, you should understand what a model is, how to read the two simplest economic models (the production possibilities frontier and the circular flow), what positive and normative economics are, and why economic disagreement is a feature rather than a bug. You should also have a better sense of how to read economic claims in the news without either swallowing them whole or dismissing them as "just economists making things up."

2.1 Why models?

Reality is too complicated to think about all at once. The economy of the United States in 2026 involves something like 340 million people, 33 million businesses, federal and state and local governments, hundreds of trillions of dollars of financial assets, complex global supply chains, central banks acting on partial information, weather affecting harvests, geopolitics affecting trade, demographic shifts affecting labor markets, and a billion individual decisions a day. Nobody can hold all of that in their head. Nobody who claimed to could be trusted.

A model is a deliberate simplification that lets you reason about the parts of a system that matter for the question at hand while ignoring the parts that don't. The metaphor most economists use is a map. A map of the New York City subway is not a literal description of the subway: it leaves out the depths of the tunnels, the exact distances between stations, the curves of the tracks, the names of the workers, the soot on the walls. It is wildly simplified. And it is exactly what a tourist needs to find the train that gets her from Times Square to Brooklyn. A more accurate map — say, a satellite photograph of New York City showing every subway tunnel in scale — would be useless for that purpose. The whole point of the simplification is what it leaves out.

Economic models work the same way. The production possibilities frontier (which we'll meet in §2.2) leaves out a thousand things about a real economy. So does the circular flow diagram (§2.3). So does supply and demand (Chapter 5). Each of these models is a map drawn for a specific purpose. None of them is a literal description of reality. Each of them is useful precisely because of what it omits.

A model is a deliberate simplification of reality designed to clarify a particular question.

This definition has consequences. First, the "right" model depends on the question. The production possibilities frontier is the right tool for thinking about tradeoffs and growth at the level of a whole economy. It is the wrong tool for thinking about why your local grocery store stocks the brand of cereal it stocks. Different questions, different tools.

Second, no model can tell you everything. Asking a model the questions it wasn't designed to answer is like asking the subway map how deep the L train tunnels are: the answer isn't on the map because the map wasn't built to provide it. Many bad arguments in economics come from people pushing a model past the limits of what it was designed to do.

Third, a model's assumptions are the price you pay for the model's clarity. When you simplify, you have to leave things out — and the things you leave out are the things you are assuming don't matter for the question. If you choose your assumptions well, the model gives you useful answers. If you choose them badly — or don't notice when reality has changed in a way that makes the assumptions wrong — the model gives you confidently wrong answers.

A famous example: in the run-up to the 2008 financial crisis, many of the models used to price mortgage-backed securities assumed that house prices in different regions of the United States moved roughly independently — so that even if some regions experienced a downturn, others would not, and the overall portfolio would be safe. The assumption had been roughly true for the entire period for which the modelers had data (since the 1950s). It turned out to be wrong in 2007, when housing prices fell across the entire country at once. The models did not predict the crisis because the crisis violated the central simplifying assumption. The lesson is not that the modelers were stupid; the modelers were sometimes very smart. The lesson is that every model has assumptions, and every assumption is a place where reality could prove the model wrong. The work of using a model carefully is being constantly aware of where its assumptions could break down.

Useful simplification vs. dangerous distortion

How do you tell whether a model's assumptions are useful or dangerous? There is no algorithm. But there are habits.

Useful assumption: simplifies a feature of reality that is not central to the question at hand. (Example: assuming that all consumers are identical, when you're trying to figure out how the average market price moves with income. The diversity of individual consumers is irrelevant if you only care about averages.)

Dangerous assumption: simplifies a feature of reality that is central to the question at hand, and you don't realize you've done it. (Example: assuming that all consumers are identical when you're trying to figure out who bears the burden of a tax. The diversity of consumers — some rich, some poor — matters enormously for that question.)

Useful assumption: is widely shared by other models for the same kinds of question, has been tested against data, and survives the testing. (Example: assuming that demand curves slope down. We have approximately a century of empirical work confirming that, almost always, when prices rise, quantity demanded falls. The exceptions are interesting and worth flagging, but as a baseline assumption it's solid.)

Dangerous assumption: is convenient for the modeler but doesn't have empirical support. (Example: assuming that financial markets are perfectly efficient at all times. Some versions of this are reasonable; the strong-form version doesn't survive contact with even basic data, and treating it as bedrock leads to confident wrong conclusions.)

Useful assumption: is honestly stated, so a reader can check whether they agree. (Example: a model that says "we assume households have rational expectations" — meaning households use all available information correctly to forecast the future. You can disagree with this, but at least you know what you're disagreeing with.)

Dangerous assumption: is hidden, so the reader doesn't realize the conclusion depends on it. (Example: a model that quietly assumes that the labor market clears continuously — meaning involuntary unemployment is impossible by assumption. If the reader doesn't notice the assumption, they will believe the model's conclusions about unemployment without realizing the model has assumed unemployment away.)

The skill of reading economic models well — which is one of the skills this book is trying to develop in you — is the skill of identifying the assumptions, asking whether they're plausible for the question being asked, and noticing when a conclusion depends on a particular assumption rather than on any deep feature of the world.

Ceteris paribus

One assumption is so common in economic reasoning that it has its own Latin name. Ceteris paribus means "other things equal" or "holding everything else constant." When an economist says "an increase in the price of coffee causes the quantity demanded to fall, ceteris paribus," they mean: "if nothing else changes about consumers' incomes, tastes, the price of substitutes, the price of complements, and so on, then a higher price of coffee will lead people to buy less of it."

The phrase is a way of isolating one cause from many possible confounders. Real markets always have many things changing at once. To talk usefully about the effect of just one of those changes, you have to imagine the others held still — even though they aren't held still in real life. Ceteris paribus is the verbal flag economists raise to say "I'm doing this isolation trick now."

The trick is useful, but it has costs. The most common mistake students make is forgetting the ceteris paribus clause when they apply a model to reality. "The model says higher coffee prices reduce coffee consumption — but I noticed that coffee prices went up last year and consumption also went up, so the model is wrong." No, the model isn't wrong; the ceteris paribus clause was violated. Income probably rose, or tastes shifted, or the price of substitutes went up faster, or something else changed at the same time. The right interpretation of the data isn't "the model is wrong" — it's "the model shows you what would happen if only the price changed, and you observed a world in which other things also changed, so you can't read off the model's prediction directly."

Reading economic claims in the news requires the same vigilance. "Since the minimum wage went up in Seattle, total restaurant employment has risen, so minimum wages don't kill jobs." Maybe — but Seattle's economy was booming during the period in question, and ceteris paribus was clearly violated. The serious empirical work tries to control for the other changes, often using comparison cities or comparison time periods. We will see this kind of comparative work in Chapter 7 (when we revisit the minimum wage carefully) and Chapter 21 (where we give it deep treatment). For now, the key point is: every claim about the effect of one thing on another is implicitly a ceteris paribus claim, and the harder question is whether the ceteris was actually paribus in the situation being described.

2.2 The Production Possibilities Frontier

The first model we'll meet is the simplest economic model in the textbook: the production possibilities frontier, or PPF. It is a picture of what an economy could produce, given its resources and its technology. The picture has only two goods on it (so it's already wildly simplified — real economies have millions of goods), but the simplification is exactly the thing that lets you see the structure of tradeoffs at a glance.

Building the PPF

Imagine a small economy that produces only two goods: bread and books. The economy has a fixed amount of resources (workers, land, machinery) and a fixed level of technology. We're going to ask: given those resources and that technology, what combinations of bread and books can the economy produce?

If the economy puts all of its resources into making bread, suppose it can make 100 loaves per day and zero books. If it puts all of its resources into making books, it can make 30 books per day and zero loaves. In between those extremes, it can produce some combination of both — but as it shifts resources from bread to books, it loses bread and gains books.

                 PRODUCTION POSSIBILITIES FRONTIER
   Books
   per day
    30 |●A     (all resources to books)
       |
       |  ●B      (some bread, mostly books)
       |
    20 |    ●C    (intermediate)
       |
       |       ●D   (some books, mostly bread)
    10 |
       |          ●E
       |
     0 |_____________●F______________________
       0       50       100       150         Bread per day
                                              (all resources to bread)

Figure 2.1 — The production possibilities frontier for an economy that produces bread and books. Each labeled point represents a feasible combination of the two goods. Point A is all books, no bread. Point F is all bread, no books. Points B–E are intermediate. The line connecting them traces all the maximally efficient production combinations the economy could achieve. Any point inside the frontier is feasible but inefficient. Any point outside the frontier is currently infeasible.

What does this picture tell us?

First, it makes the idea of opportunity cost concrete. To produce more books, the economy must produce less bread. The slope of the PPF is the opportunity cost of one good in terms of the other. If you read off the chart that moving from point B to point C costs you 10 loaves of bread and gives you 5 more books, then the opportunity cost of one additional book at that point is roughly 2 loaves of bread.

Second, it lets you distinguish between efficient and inefficient outcomes. Any point on the frontier itself is efficient in the sense that you can't get more of one good without giving up some of the other. A point inside the frontier is inefficient: there exist combinations that have more of both goods. Inefficiency means resources are being wasted somehow — workers idle, machines underused, talent misallocated.

Third, it shows that the frontier itself can move. If technology improves (better ovens, better printing presses), or if resources expand (more workers, more land), the frontier shifts outward. The economy can now produce more of both goods than before. This is economic growth in its simplest visual form.

Fourth, the shape of the frontier matters. In our simple drawing, the line is roughly straight. But in real economies, the PPF tends to be curved — bowed out from the origin — because resources are not perfectly substitutable across uses. The first workers shifted from bread-making to book-making are probably the ones who are best suited to making books; later workers are people who were really good at making bread, so the bread cost of getting one more book rises. This is increasing opportunity cost, and it's the realistic shape for almost every PPF you'll see.

                 INCREASING OPPORTUNITY COST PPF
   Books
   per day                ╭─ ─ ─ ─
    30 |●A             ╱
       |            ╱
       |          ╱
    20 |       ╱
       |     ╱           (bowed out frontier — typical real-world shape)
       |    ╱
    10 |  ╱
       | ╱
       |╱
     0 |_____________●F________________
       0       50       100       150       Bread per day

Figure 2.2 — The increasing-cost PPF. The frontier is bowed outward because moving resources from one industry to another becomes harder as you go. The first units shifted are the ones that were least suited to their original use; later units are increasingly costly.

You will see PPFs again in Chapter 3 (where they help us think about trade and specialization between two countries) and Chapter 25 (where the outward shift over time becomes a model of long-run economic growth). For now, what matters is that you can read one and that you understand what it represents: a deliberate, useful simplification of an economy down to two goods so that the structure of tradeoffs becomes visible.

What the PPF leaves out

Notice what the PPF doesn't show. It doesn't show prices. It doesn't show who gets the bread and who gets the books. It doesn't show the labor market. It doesn't show the financial system. It doesn't show how the economy actually makes its production decisions (central planning? markets? mixed?). It doesn't show whether the resources are fairly distributed among the people. It doesn't show happiness or wellbeing or anything else that might matter to the people living in the economy.

That's all fine. The PPF is built to answer one question — what are the feasible combinations of two goods this economy can produce, and at what tradeoff? — and it answers that question well. To answer different questions, we'll need different models. This is the nature of model-building. No single model does everything. Trying to make one model do everything is a recipe for confusing people and getting wrong answers.

2.3 The Circular Flow Diagram

The second simple model is the circular flow diagram, which gives you a bird's-eye picture of how money and goods move through an economy. Like the PPF, it leaves out almost everything in order to make the central structure visible.

The basic version has two kinds of actors — households and firms — and two kinds of markets — the market for goods and services and the market for factors of production (labor, land, capital). The arrows trace what flows where.

                     SIMPLE CIRCULAR FLOW DIAGRAM

                ┌─────────────────────────────┐
                │   Market for Goods and      │
                │         Services            │
                │                             │
                │  Households buy →           │
                │  Firms sell ←               │
                └─────────────────────────────┘
                  ▲                         ▲
                  │ Spending                │ Goods and
                  │ ($)                     │ services
                  │                         │
                  │                         │
                  │                         ▼
              ┌──────┐                 ┌──────┐
              │HOUSE-│                 │ FIRMS│
              │HOLDS │                 │      │
              └──────┘                 └──────┘
                  │                         │
                  │                         │
                  │ Labor, land,            │ Wages, rent,
                  │ capital                 │ profit ($)
                  ▼                         │
                                            ▼
                ┌─────────────────────────────┐
                │   Market for Factors of     │
                │         Production          │
                │                             │
                │  Households sell →          │
                │  Firms buy ←                │
                └─────────────────────────────┘

Figure 2.3 — The simple circular flow. Households sell labor (and other factors) to firms in the market for factors of production. Firms use those factors to produce goods and services, which they sell back to households in the market for goods and services. Money flows in the opposite direction around the loop. Each transaction is one half of a two-sided exchange.

What does the diagram show?

It shows that the economy is a circular system. Money doesn't disappear or appear from nowhere; it circulates. Households earn money by selling their labor (and other resources) to firms, then spend that money buying things from those firms, who then use the revenue to pay wages to households, who then spend the money buying things, and so on. The two markets — for goods and for factors — are mirror images of each other, with households and firms playing opposite roles in each.

It also shows that there are no leaks. In this simplified version of the economy, every dollar a household earns gets spent and every dollar a firm earns gets paid out as wages, rent, or profit. The total income of households equals the total revenue of firms equals the total value of goods and services produced. This identity — income = output = expenditure — is the foundation of how we measure GDP, which we'll see in Chapter 22.

What does the diagram leave out?

A lot. The simple version doesn't include:

• Government, which collects taxes from both households and firms and spends money on goods, services, and transfer payments
• The financial sector, which channels household savings into firm investment
• The foreign sector, which adds imports and exports to the picture
• Within-household and within-firm production, like cooking dinner at home or training employees inside a company
• Time, which is utterly absent — the diagram is a snapshot, not a movie

A more complete circular flow diagram would add boxes for the government, the financial sector, and the rest of the world, and additional arrows showing taxes, transfers, savings, investment, imports, and exports. We'll see the expanded version in Chapter 22 (when we use it to derive the GDP identity). For now, the simple version is enough to make the basic point: the economy is a system in which everyone's spending is someone else's income, and every transaction has two sides.

This insight matters. It tells you that you can measure the size of an economy three different ways — by adding up everyone's income, by adding up everyone's spending, or by adding up the value of everything produced — and you should get the same answer. We will do exactly that in Chapter 22.

2.4 Positive vs. Normative Economics

We previewed this distinction in Chapter 1. Now we'll look at it more carefully, because it shapes how the rest of the book reads.

Positive economics makes claims about what is. Positive claims can in principle be checked against evidence.

Normative economics makes claims about what should be. Normative claims depend on values and judgments about goals.

Let's look at three pairs of statements to make the distinction concrete.

Pair 1 — Minimum wage - Positive: "When Seattle raised its minimum wage from $9.47 to $13 in 2015, the average earnings of low-wage workers in Seattle rose by [X]% over the next year, while their total hours worked changed by [Y]%." - Normative: "Seattle was right to raise its minimum wage to $13."

The positive claim is — in principle — checkable against the data. (In practice, measuring it well is hard, and economists have argued about exactly how hard. We'll revisit this in Chapter 21.) The normative claim depends on how you weight the wage gains for some workers against the hours losses for others, on what you think the purpose of minimum wage policy is, and on how you compare these effects to other things the policy might achieve or prevent. Two people who agree on the positive claim can still reasonably disagree on the normative claim.

Pair 2 — Tax policy - Positive: "A 1% reduction in the top marginal income tax rate is associated with a [Z]% increase in declared taxable income for top earners over the following five years." - Normative: "The top marginal income tax rate should be lowered."

The positive claim is checkable against tax-return data. The normative claim depends on what you think the purpose of tax policy is (revenue? fairness? incentives for investment? deterring inequality?) and on how you weight competing objectives.

Pair 3 — Trade - Positive: "Free trade increases total economic output across countries by allowing each to specialize in goods where it has comparative advantage." - Normative: "Free trade is good policy."

The positive claim is among the most well-supported in economics — there is overwhelming theoretical and empirical support for the idea that trade increases the total size of the economic pie, in the sense of total goods and services produced and consumed. The normative claim is genuinely contested, because the distribution of the gains from trade is uneven, the losers are concentrated and visible, and the gains for the broad consumer base are diffuse and invisible. Two people who agree completely on the positive claim can disagree on whether free trade should be the policy a particular country pursues. (We will see this in much more depth in Chapter 9.)

The pattern: agreeing on the positive does not require agreeing on the normative. Conversely, you cannot make a normative claim follow logically from a positive claim alone. If someone tries to derive an "ought" purely from an "is," they have smuggled in a normative assumption somewhere. (Philosophers call this the "is-ought problem" or "Hume's guillotine," after David Hume's 1738 statement of it.) Watch for the smuggling.

The honest version of "economics says..."

People sometimes write "economics says we should do X" or "economists agree that we should do Y." These claims should make you suspicious. Economics, the discipline, is good at making positive claims (with appropriate uncertainty). It is not good at — and is not authorized to — make normative claims by itself. When you see "economists say we should...," ask: which normative goal are they assuming the policy is trying to achieve? Which values are they weighting? Which tradeoffs are they accepting? Sometimes the answers are clear and reasonable. Sometimes they aren't. Either way, it's a good habit to check.

There are versions of normative claims that economics genuinely can support — claims of the form "if you value X, then policy Y is more efficient at achieving X than policy Z." Those claims are conditionally normative: their normative weight depends on accepting the value premise. That's a legitimate kind of statement and economists make it often. The illegitimate kind is the unconditional "economists agree that policy Y is correct," without specifying what value framework "correct" is being measured against.

2.5 Why Economists Disagree

Here is one of the most useful things this book can teach you: economists disagree, and the disagreements are not signs that the field is broken. They are signs that the field is taking hard questions seriously.

There are roughly four reasons honest economists disagree about important questions. Sorting any particular disagreement into these four buckets is one of the most practical analytical skills you can develop.

1. Disagreement about the facts

Sometimes, two economists looking at the same question are working from different data, or from data they read differently. The minimum wage debate has a lot of this. Card and Krueger's famous 1994 study of fast-food employment in New Jersey and Pennsylvania found that the minimum wage increase did not seem to reduce employment. Neumark and Wascher, using a different data source for the same period, found that it did. The two teams of economists were not really disagreeing about theory; they were disagreeing about which empirical measurement was right. Resolving this kind of disagreement requires more data, better data, or methodological debate about how to handle the existing data.

Factual disagreements are not always resolvable, but they are the kind of thing that data could in principle resolve. That's a useful property.

2. Disagreement about which model to use

Sometimes, two economists agree on the data but use different models to interpret it. The 2021–23 inflation surge is a current example. Some economists analyzed it primarily through the lens of a demand model: stimulus spending was too large, the economy overheated, prices rose. Others analyzed it primarily through a supply model: pandemic-driven supply chain disruptions and labor force exits caused costs to rise, and the resulting price increases were largely driven by costs rather than demand. A third group (the eventual partial winners of the debate) said it was both: large demand stimulus into a constrained-supply economy made the cost-push effects much bigger than they would otherwise have been.

Each of these views uses a different simplifying model. Each can be defended honestly. The disagreement isn't about whether prices rose (everyone agrees they did) or about whether stimulus was large (everyone agrees it was) or about whether supply was disrupted (everyone agrees it was). It's about which model to foreground and which to treat as background.

These kinds of disagreements often turn out to be at least partly resolvable as more data accumulates and as different models are tested against each other. But they can persist for a long time, because the question of which model fits a particular situation is often genuinely hard.

3. Disagreement about values

This is the kind we have already encountered in §2.4. Two economists who agree on the data and on the model can still disagree about what should be done, because they value different outcomes. One might prioritize efficiency (the size of the total pie). Another might prioritize equity (who gets which slice). A third might prioritize freedom (people's right to choose without interference). A fourth might prioritize stability (avoiding disruption). When the same policy has different effects on these different objectives, economists who weight the objectives differently will recommend different policies, even with identical positive analyses.

These disagreements are not resolvable by more data, because they aren't about the data. They're about what we should care about. They are real disagreements, and they are honest. The right response is not to pretend they aren't there but to be transparent about which values each economist is weighting, so the reader can decide which value framework they share.

4. Disagreement about what counts as "the long run"

This one is subtle but important. Many economic models give different answers in the short run and the long run. Trade lowers prices for consumers and creates more total wealth — in the long run. Trade also displaces specific workers in specific industries — in the short run. Both are true. Two economists looking at the same trade policy can end up disagreeing because one is focusing on the long-run consumer benefit and the other on the short-run displacement.

John Maynard Keynes famously wrote: "In the long run, we are all dead." His point was not that long-run analysis was useless, but that some things matter very much in the short run even if they wash out in the long run. For some questions, the right time horizon is decades; for others, it is months. Choosing the time horizon is partly a value judgment and partly an empirical claim about how long the relevant adjustments actually take.

Disagreements about time horizon often look like disagreements about facts or values, but they have their own structure. When you see two economists disagreeing, ask: what time horizon is each one using? You will sometimes find that they would agree if forced to specify the same window.

Putting the four together

Most real economic disagreements are mixtures. The minimum wage debate is partly about the data (factual), partly about which model best describes the labor market (model selection), partly about how to weight wage gains versus possible employment losses (values), and partly about whether short-run effects or long-run effects are more relevant (time horizon). Sorting through which kind of disagreement is at work in any particular debate is one of the most useful things you can do as a reader of economic news. It tells you what kind of evidence would change your mind, what kind would not, and where reasonable people might still reasonably disagree.

Disagreement is a feature, not a bug. Throughout this book, we will be honest about which questions have genuine economic consensus (free trade increases total wealth, rent control as written reduces housing supply, a price on carbon is the most efficient response to climate change — all of these have ~85–95% support in expert surveys) and which questions are genuinely contested (the right minimum wage, the size of fiscal multipliers, the optimal top tax rate, the role of government in the economy). Where there is consensus, we will say so. Where there isn't, we will present multiple perspectives and let you think for yourself.

2.6 Reading Economic Claims Like an Economist

Let's put what we have learned into a checklist. When you encounter an economic claim — in a textbook, a news article, a politician's speech, a friend's social media post — ask yourself:

1. Is this a positive claim or a normative claim? If positive, what evidence would check it? If normative, what values is it assuming?
2. What model is the speaker using, and what does the model assume? Are the assumptions plausible for the specific situation?
3. Is the ceteris paribus clause being honored? Are the "other things" that the model holds equal actually equal in the real situation?
4. What time horizon is the speaker assuming? Are they reporting a short-run effect, a long-run effect, or something in between?
5. If economists disagree, what kind of disagreement is it? Factual? Model-selection? Values? Time horizon?
6. Whose values does the speaker's framing favor? Is there a hidden weighting of objectives that the speaker is not flagging?

You will not always be able to answer all six of these questions. You will sometimes have to act on incomplete information. But asking the questions — getting into the habit of asking them automatically — will make you a much sharper consumer of economic claims than you were before. It will also make you slower to declare confidently that you know what should be done. That slowness is a virtue. The people who are most certain about complicated economic questions are usually the ones who have thought about them least.

2.7 Where this is going

In Chapter 3, we'll meet trade and specialization — and the deeply counterintuitive insight that two parties can both gain from trade even when one is better at everything. The chapter is short, and the result is one of the most powerful in all of economics. It will set up Chapter 9 (international trade) and the long thread about why economists support free trade in principle while recognizing that its losers are real.

In Chapter 4, we'll learn how to read economic data — the chapter that makes you data-literate before you encounter the more elaborate models of Part II.

In Chapter 5, we'll meet supply and demand — the central model of microeconomics and the foundation that the rest of Part II builds on.

For now, you have two new tools: the production possibilities frontier and the circular flow diagram. You also have a way of thinking about economic models in general — what they're for, what they leave out, and how to read them charitably without being credulous. And you have the four reasons economists disagree, which is the framework you should use whenever you read about a policy debate.

Take a breath. Do at least two of the exercises. Then come back for Chapter 3.

Key terms recap: model — a deliberate simplification of reality assumption — a feature of reality the model treats as fixed for tractability production possibilities frontier (PPF) — a graph of feasible production combinations circular flow — a model of how money and goods circulate between households and firms ceteris paribus — "other things equal"; the holding-constant device that lets you isolate one cause from many positive economics — claims about what is (in principle checkable against evidence) normative economics — claims about what should be (depends on values)

Themes touched in this chapter: Disagreement (foundational), Tradeoffs (in PPF), Markets power+imperfect (in circular flow), Data tells stories (positive vs. normative).