1The idea in one paragraph

Economics studies people, and people cannot be put in a laboratory and held still. So economists work in two different ways at once. Some of what they say is a claim about how the world is, built by simplifying it into a model and then checking the model against evidence; those claims can be shown to be wrong, and often are. The rest is a claim about how the world ought to be, which no measurement can settle because it rests on what the person making it thinks matters. Knowing which of the two you are doing at any moment is most of what this subtopic teaches, and the history at the end of it is the story of economists arguing about both.

2Two kinds of statement

Positive economics deals with statements about what is, what was, or what will be. A positive statement can in principle be tested against evidence and shown to be false. "A tax of two lira a litre on fuel raises the pump price" is positive. So is "unemployment in Meriç fell last year", and so is "raising the minimum wage by a fifth would cost eight hundred jobs". That last one has not happened yet and might be wrong, which is exactly the point: it is the sort of claim that evidence could knock down.

Normative economics deals with statements about what ought to be. A normative statement rests on a value judgement, a view about what is good, fair or important, and no amount of data settles it. "Fuel ought to be taxed more heavily", "the minimum wage should rise", "unemployment matters more than inflation" are all normative.

Figure 1 sorts any statement you meet with a single question.

Two warnings, because both cost marks.

Positive does not mean true. It means checkable. "The earth's economy will double in size next Tuesday" is a positive statement and a ridiculous one. A positive statement earns its name from the kind of claim it makes, not from its accuracy.

Normative does not mean worthless. Every policy anyone has ever argued for is normative. Governments exist to make normative choices, and the point of the positive work is to make those choices better informed. An economist who refuses to say anything normative has refused to be useful.

A positive statement can be checked against evidence and shown to be wrong. A normative statement cannot, because it says what ought to be rather than what is.

The word "should" is a useful clue but a poor rule. "This tax is too high" contains no "should" and is thoroughly normative. "Most people would be better off under this tax" looks normative and is partly measurable. Read for the claim, not for the word.

3How a positive claim is built: logic, hypotheses, models, theories

Four words, and they come in order.

Logic is reasoning that moves from statements you have accepted to a statement that follows from them. Most economic argument is deductive: if people buy less of a good when it costs more, and a tax raises what the good costs, then a tax reduces the amount bought. The chain is valid, and notice what that does and does not guarantee. A valid chain with a false starting point gives you a false conclusion with total confidence. When an economic argument goes wrong, the fault is usually in an assumption rather than in the reasoning, which is why the assumptions get so much attention.

A hypothesis is a statement put forward to be tested, worded so that it could turn out to be wrong. "Raising the bus fare in Meriç by a tenth will cut the number of trips by more than a tenth" is a hypothesis. "Bus fares matter somewhat" is not, because nothing could contradict it.

A model is a deliberate simplification of part of the economy, built to answer one question. It keeps the few things that drive the answer and throws away everything else. Models can be a diagram, a table, a set of equations or a story told carefully; the demand curve, the production possibilities curve from 1.1 and the circular flow of income are all models.

A theory is an explanation that has survived testing: a hypothesis, or a set of them, that the evidence has not yet knocked down and that can be used to explain and to predict. The law of demand is a theory in this sense.

Figure 2 puts the four in the loop they actually live in.

Work it once with the buses. Meriç's council notices that its buses are emptier than last year. The hypothesis: passengers left because the fare rose. The model: passengers choose between the bus and the alternatives on price alone, with everything else about the town held still. The prediction: towns that raised fares lost passengers, and towns that did not, did not. The evidence: the passenger counts. If the counts fit, the council keeps the model for now. If they do not, the model is mended, because something it threw away was doing real work, or it is dropped.

Nothing in that loop ever finishes. A theory is what is left standing after the evidence has had its turn, and it stays only until the next turn.

4Ceteris paribus

Ceteris paribus is Latin for "other things equal". It is the assumption that while we look at one cause, everything else that could matter is held still.

An economist cannot do what a chemist does. You cannot run Meriç for a year with the fare rise, rewind it, and run the same year again without it. Everything that could affect bus use moves at once: the weather, fuel prices, a new tram line, wages, the school timetable. So instead of holding the world still, the economist holds it still on paper, which is what ceteris paribus does. Figure 3 shows the difference it makes.

The left panel is the model: one cause, one effect, a clean line. The right panel is what the data actually looks like, because in the world nothing waited its turn. The line in the left panel is not a lie about the right panel. It is the claim that if everything else had waited its turn, this is the relationship that would have shown.

What this costs you is that every conclusion drawn this way is conditional. "A fare rise cuts passenger numbers, ceteris paribus" means: unless something else moved at the same time. In 1.1 the production possibilities curve was drawn with the quantity of resources and the state of technology assumed fixed; in 2.1 the demand curve is drawn with income, tastes and the prices of other goods assumed fixed. Same device, different subject.

It also makes testing harder, which matters in the next section. When a prediction fails, there are always two possible reasons: the theory is wrong, or something else moved. An economist who always reaches for the second reason has made their theory impossible to disprove, and that is the fault the next section names.

5Empirical evidence and refutation

Empirical evidence is evidence from observation and measurement rather than from reasoning alone: official statistics on prices and employment, household surveys, firms' records, and experiments run on volunteers or on real customers. A positive claim is worth only as much as the evidence that has been thrown at it.

Refutation is the test that matters. A claim belongs in positive economics only if you can say, in advance, what you would have to observe for it to be false. The work is then to look for that observation, not to collect examples that agree. One clean contrary case does more to the claim than a hundred friendly ones.

Figure 4 shows the three fates. A claim stands, for now, because nothing has yet contradicted it. A claim is refuted, and the honest response is to mend it or drop it. Or the claim is normative, in which case refutation does not apply and the argument is a different sort of argument.

Economics finds refutation hard, and the guide expects you to know why rather than to pretend otherwise.

• The experiment cannot be repeated. There is one 2024, it happened once, and it is over.
• Nothing is held still, so a failed prediction can always be blamed on something else that moved.
• The subject matter answers back. People change what they do when they are taxed, watched, or told what economists have predicted, so a good prediction can undo itself.
• The measurements are rough. National figures are estimates, revised months later, and countries count things differently.

The result is that refuted theories in economics are more often patched than buried, and that two economists can look at the same figures and keep their different views. That is a fair criticism of the subject. It is also a reason to write "the evidence suggests" rather than "this proves", and examiners notice which one you wrote.

6One model read properly: the circular flow of income

1.1 taught the circular flow of income as content: the five decision-makers and the six flows between them. Here it is the worked example of everything in sections 3 to 5, because it shows what a model is for and what a model costs. Figure 5 is the full version.

Read the ring first. Households own the factors of production and sell their services to firms; firms pay for them, and that payment is household income. Households spend that income on the output firms make. Money goes round, and one side's spending is the other side's income.

Then read the three doors in the side of the ring, because they are what the model was built to show.

• The financial sector takes in money households do not spend and lends it to firms.
• The government takes money out in tax and puts money back in through spending.
• The foreign sector takes money out when we buy imports and puts it back when foreigners buy our exports.

Money leaving the flow is a leakage: saving (S), taxes (T) and spending on imports (M). Money entering it is an injection: investment (I), government spending (G) and export revenue (X). They pair up, which is how you keep all six: households save and the banks lend it back as investment; households pay tax and the government spends it back; households buy imports and foreigners buy our exports. If injections exceed leakages the flow of income grows; if leakages exceed injections it shrinks; if they match it holds its size.

Now treat it as a model rather than as a picture. Its assumptions are worth listing, because a question can ask what a model assumes.

1. Every household and every firm can be added together into one of each, so the differences between them do not matter for this question.
2. Money that leaves through one door comes back only through the matching door.
3. Nothing is produced outside the money economy: unpaid work at home counts for nothing here.
4. The environment is not in the diagram at all. Nothing enters the flow from it, and nothing is dumped back into it.

Figure 6 sets what it keeps against what it drops. The model answers one question extremely well: is the flow of income growing, shrinking or holding steady, and through which door. It says nothing about which goods are made, who ends up rich, what anything costs, or what the whole business does to the land it runs on. Unit 3 is built on the first question. The fourth assumption is exactly what the 21st-century economists in section 11 objected to.

Neither of those is a mistake in the model. Everything in the right-hand box of Figure 6 was left out on purpose, so that the one question could be answered clearly. Evaluating a model means naming what it left out and asking whether that thing was doing work in the case in front of you.

7Normative economics and value judgements in policy

Every policy argument has two layers, and mixing them is the commonest fault in a part (b) answer.

Meriç is considering a charge for driving into the centre. The positive layer asks what the charge would do: how much traffic would fall, how much money would be raised, how many drivers would move to the buses, what would happen to the shops on the square, how much cleaner the air would be. Every one of those is a question evidence can attack, and economists will still disagree about the numbers.

The normative layer asks whether it is worth doing. That means weighing the driver who now pays against the child who breathes, the shopkeeper against the bus passenger, and today's convenience against tomorrow's air. There is no measurement that performs that weighing. It is a value judgement, and the judgement is where policy actually gets made.

Value judgements hide in three places a careful reader learns to look.

In what gets measured. Judge a policy by national output and you have already decided that output is what matters; judge it by economic well-being and you get a different answer with the same data.

In who gets counted. A gain of ten to a wealthy household and a gain of ten to a poor one are the same ten in the arithmetic. Whether they are the same in the decision is a judgement about equity.

In how the future is weighed. Costs that fall in forty years can be treated as nearly weightless or as fully real. That choice decides most arguments about sustainability before any evidence is heard.

Write it in the exam as two moves: first what the evidence supports, then what the recommendation assumes about what matters. An answer that says "on this evidence the charge cuts traffic, but whether it should be introduced depends on how much weight is given to the drivers who can least afford it" is doing exactly what the evaluation marks are for.

8Equity and equality

The guide asks for these two separately because students use them as if they were one word. They are not.

Equality means sameness of amount. Two households have equal incomes when the two numbers match. It is a measurement, and anyone with the figures can check it.

Equity means fairness in the way something is shared. It is a judgement about what each party is owed, and you cannot settle it by measuring, because two people with the same figures in front of them can hold different views about what is fair.

Figure 7 shows why the two come apart. On the left every household gets a hundred: perfectly equal. On the right the poorest gets a hundred and fifty and the richest fifty: unequal, and many people would call it the fairer of the two. Someone else will say the first is fairer, because it treats everyone alike. Both are arguing about equity, and neither is producing a fact.

The same split runs through the tax system, which is where the exam usually meets it. A tax that takes the same percentage from every income treats everyone equally by one measure and takes far more from the poorest household's ability to pay by another. Whether that is equitable is a value judgement, which is why tax is argued about at elections rather than settled in a spreadsheet.

Hold this beside the pair from 1.1: efficiency and equity are separate questions, and a market can be efficient while producing an outcome almost nobody calls fair.

9The nine key concepts, read as questions

The nine concepts were introduced in 1.1. Section 2 gives you a second way to hold them, which is more useful in an exam than a list: each concept is a question, and the questions are not all of the same kind.

Scarcity, choice, efficiency, change and interdependence ask positive questions. How much is there? What was given up? Is anything being wasted? What moved, and what followed? Who else is affected? Every one of those can be attacked with evidence.

Equity, economic well-being, sustainability and intervention cannot be answered without a value judgement as well. Is the share fair? Are people actually better off, and by whose reckoning? Can this go on without cost to people who are not yet born? Should the government step in here?

Intervention shows the pattern clearly. What a minimum wage does to employment is positive, argued with evidence, and economists still differ about the size of the effect. Whether the government ought to set one is normative, and no amount of evidence closes it. An answer that keeps those two apart reads like an economist's answer.

10Where the ideas came from: the 18th and 19th centuries

The guide asks you to know where economic ideas came from and what problem each was answering. None of these people was simply right or wrong, and an answer that hands out marks to them is not doing what the guide asked. Figure 9 is the whole story on one line.

18th century: Adam Smith and laissez faire. Smith published An Inquiry into the Nature and Causes of the Wealth of Nations in 1776, and the problem he attacked was the ruling idea of his day: that a nation grew rich by piling up gold, selling much abroad and buying little. His answer was that a nation's wealth is the goods and services it produces, that the division of labour raises how much it can produce, and that trade leaves both sides better off. His most quoted claim is that a butcher or a baker serves you not out of kindness but out of self-interest, and that in a competitive market that self-interest is guided, as if by an invisible hand, to an outcome nobody intended and most people gain from. Laissez faire, French for "let it be", is the policy that follows: leave markets alone, because direction from above does worse than the market's own coordination.

Two things are worth getting right, because they are commonly overstated. Smith did not argue that government should do nothing; he wanted it to provide defence, justice and the public works that no private buyer would pay for. And he was a moral philosopher who wrote sharply about merchants combining against the public. The invisible hand is an argument about competitive markets, not about business in general.

19th century: utility and the margin. For a long time economists explained value by the labour that went into a thing, which left an awkward puzzle: water keeps you alive and costs almost nothing, while diamonds are useless and cost a fortune. In the 1870s a set of economists, among them Jevons, Menger and Walras, answered it by looking at the last unit rather than the whole. Utility is the satisfaction a good gives; what a buyer will pay depends on the utility of one more unit, and because water is plentiful, one more litre is worth very little, while one more diamond is worth a great deal. That is the margin, and the habit of thinking at the margin, one more unit at a time, went on to organise the whole of microeconomics. Marshall later set demand and supply together as the two blades of a pair of scissors. You will meet the margin again in 2.1.

19th century: Say's law. Jean-Baptiste Say argued that producing output pays out incomes exactly equal to the value of that output, so the means to buy everything made is created by the making of it. Supply, in the slogan, creates its own demand. It follows that a general glut, everything unsold at once, cannot happen, and that unemployment can only be local and temporary. Classical macroeconomics rested on this, and so did the view that a slump needs no policy because the economy rights itself.

19th century: the Marxist critique. Marx accepted that value came from labour and drew a different conclusion. In his account the worker produces more value than the wage returns to them, and the difference, which he called surplus value, is where profit comes from; so profit is not a reward for organising but a transfer from the people who did the work. He argued that competition drives firms to accumulate and concentrate, that crises of overproduction are built into the system rather than accidental, and that the conflict between the owners of capital and the people who work for them would eventually break it.

Say plainly what happened next, because it is neither a victory nor a defeat. The labour theory of value that Marx shared with the classicals was displaced by the marginal account above, and the collapse he expected did not arrive in the way he described. His questions did not go away: who gets the output, who holds the power in the bargain between employer and employee, and whether instability is built into a market economy are all still live, and economists still disagree about the answers.

11The 20th and 21st centuries

The Keynesian revolution. In the 1930s output fell across the industrial world and stayed down for a decade, with millions unemployed year after year. Say's law could not explain a fact that lasted that long. John Maynard Keynes published The General Theory of Employment, Interest and Money in 1936 and argued that total spending, not the capacity to produce, is what sets output and employment. The hinge of the argument is a leakage you already know from Figure 5. Say's law assumes saving comes back automatically as investment. Keynes said it need not: households save for their own reasons and firms invest for theirs, and when firms are frightened they do not invest whatever the banks offer. Figure 10 sets the two accounts side by side.

If leakages stay above injections, the flow of income settles at a smaller size and can stay there, with willing workers unemployed and machines idle, for as long as nothing lifts spending. The policy follows: if private spending will not fill the gap, government spending can.

The rise of macroeconomic policy. The consequence was a change in what governments were for. After the Second World War they took on responsibility for employment and output as a whole, built the national statistics needed to watch them, and used taxation and public spending to steer total demand. The whole apparatus of Unit 3 dates from here.

The monetarist and new classical counter-revolution. In the 1970s, inflation and unemployment rose together, which the simpler demand-management account had not expected. Milton Friedman argued that inflation follows from the quantity of money growing faster than output, that there is a rate of unemployment set by how labour markets actually work which demand policy cannot beat for long, and that trying only raises prices once people come to expect it. New classical economists, Robert Lucas among them, pressed further: if people form expectations sensibly, they anticipate predictable policy and it loses much of its effect. What came out of it was a preference for rules over improvisation, independent central banks with inflation targets, and attention to the supply side of the economy.

Where does that leave you? With an argument, not a verdict. Most economists today use pieces of both accounts, and the question of how much a government should do in a slump is genuinely open. Write both sides in an evaluation and say which assumptions each rests on. Writing that one of them was proved wrong is the mistake section 13 names.

The 21st century: dialogue with psychology. Economics has borrowed increasingly from other disciplines, psychology most of all. The models assume a chooser who knows what they want, weighs the options and picks the best; put real people in tests and they are swayed by how the choice is worded, stay with whatever the default option is, feel a loss more sharply than an equal gain, and put off what they intend to do. Behavioural economics is the field that grew out of that, through work by Daniel Kahneman and Amos Tversky and later Richard Thaler, and it now shapes policy: how a pension scheme is defaulted changes how many people save in it.

Notice what has changed and what has not. The method in section 3 is the same: hypothesis, model, evidence, refutation. What has changed is one assumption, and it changed because psychology gave economists a way to test it rather than to argue about it.

The 21st century: economy, society and environment. The other shift is a matter of where the boundary of the model is drawn. The circular flow in Figure 5 has no environment in it. An economy that takes materials from the earth, makes things, and throws them away leaves the taking and the throwing outside the diagram entirely, and for two centuries that was treated as somebody else's subject. It is not treated that way now: materials are finite, waste and emissions have costs that nobody was charged for, and an economy sits inside a society which sits inside an environment it cannot leave.

The circular economy is the response. Instead of the linear take, make, use, throw away, it designs products to last and to be repaired, keeps materials in use through reuse and recycling, and tries to leave the natural systems it draws on in a state to keep supplying. The reasons economists now find compelling are plain ones: the resources are limited, the damage is measurable, and the cost falls on people who never took part in the transaction. That is the key concept sustainability from 1.1, arriving as a research programme rather than as a word on a list.

12Theory of knowledge

The guide sets five questions beside this subtopic. They are not the kind of thing a data-response question asks in those words, but thinking them through is how you get the evaluation marks, so here is a start on each.

How realistic are economic models, and how do we know what to leave out? A model is not meant to be realistic; it is meant to be useful for one question. What to leave out is decided by whether including it would change the answer, and that judgement is itself testable: leave something out, predict, and see whether the prediction fails.

What assumptions do economists make when they apply theory to the world? That the relationship found under ceteris paribus still runs when other things move; that people and firms behave in the ways the model assumed; that the past behaves like the future. Every one of those can fail, and the answers in Units 2 to 4 are only as good as they are.

Will economics become more reliable as data and testing improve? Partly. Better data has already settled arguments that were once opinion. But some of the difficulty is in the subject rather than in the instruments: people respond to predictions, and each economy happens only once.

Does the positive and normative split exist in other subjects? It does, and it is just as contested. History decides what is worth recording, medicine decides whose health matters and at what cost. Economics is unusual mainly in how often its positive work is used to settle arguments that are actually normative.

Have individuals shifted the paradigms of economics? Smith, Marx, Keynes and Friedman all changed what the next generation argued about. Each also arrived when a fact the existing account could not explain was sitting in plain view, which suggests the individual and the moment are hard to separate.