How do economists approach the world?

• Distinguish positive statements from normative statements, and say where value judgements enter policy-making.
• Explain how hypotheses, theories and models are built on assumptions, including ceteris paribus, and tested against empirical evidence.
• Explain what it means to refute a theory, using an example from the course.
• Distinguish equity from equality.
• Outline how economic thinking developed from the 18th century to the 21st, naming for each school its central claim and what it reacted against.
• Explain the reasons given for moving towards a circular economy.

• Positive statement: A claim about what is, was or will be that can be tested with logic and evidence. It may turn out to be false.
• Normative statement: A value judgement about what ought to be or which outcome is better. Evidence can inform it but cannot settle it.
• Hypothesis: A testable statement about how two or more variables are related.
• Theory: A general explanation of why a relationship holds, which has survived repeated testing.
• Economic model: A simplified representation of a theory, usually a diagram or a set of equations, built on stated assumptions.
• Ceteris paribus: 'Other things being equal': the assumption that all influences other than the one being studied are held constant.
• Empirical evidence: Information obtained by observation and measurement: statistics, surveys, experiments and comparisons.
• Refutation: Rejecting a hypothesis or theory because evidence contradicts what it predicts.
• Equality and equity: Equality means the same for everyone and can be measured. Equity means fairness and is a value judgement.

Positive economics describes and explains how the economy works. A positive statement makes a claim that can be tested against evidence. 'The fuel tax reduced fuel purchases by 4%' (an invented example) is positive even if the data are incomplete or disputed. 'The tax caused the fall' is also positive, but it needs a stronger test, because purchases may have fallen for another reason. Positive does not mean true or certain. It means that evidence could, in principle, show the statement to be right or wrong.

Positive economics relies on logic as well as on evidence. It reasons deductively: it starts from stated assumptions and derives the conclusions that must follow if those assumptions hold. 'If the price of a good rises and nothing else changes, the quantity demanded falls' is a conclusion of this kind. Logic makes an argument valid, which means only that the conclusion follows from the assumptions. It does not make the argument true of the world. If an assumption is false, a perfectly logical argument can still give a wrong prediction, which is why positive economics pairs logic with evidence.

Normative economics is concerned with what ought to be. A normative statement expresses a value judgement: it evaluates an outcome or recommends an action. Words such as should, ought, fair, unacceptable and best often signal one, but the wording is not decisive. 'Cleaner air matters more than cheaper fuel' is normative although it contains no 'should', because it ranks two outcomes. Two people can accept the same evidence and still disagree about a normative statement, because they weigh health, cost and freedom of choice differently.

Policy-making needs both. Positive analysis estimates what a policy will do: who pays a tax, how much behaviour changes, what it costs. Value judgements then enter at three points: in choosing the objective (lower inflation or lower unemployment), in weighing the groups who gain against those who lose, and in deciding how much risk to accept. Two economists who agree on every positive estimate can therefore recommend different policies. Keeping the two kinds of claim apart shows where a disagreement lies: in the data, in the predicted mechanism, or in the values used to rank the outcomes.

An economy contains millions of households and firms making decisions at the same time, and many influences change at once. Economists cannot run a controlled experiment on a whole economy, so they build simplified accounts of it. Three terms describe the stages of that work.

• A hypothesis is a testable statement about how two variables are related, for example 'a higher minimum wage reduces the employment of teenagers'.
• A theory is a general explanation of why a relationship holds, which has survived repeated testing. The theory of demand explains why consumers buy less of a good when its price rises.
• An economic model is a simplified representation of a theory, usually a diagram or a set of equations, which leaves out detail in order to show the mechanism. The demand and supply diagram, the production possibilities curve and the circular flow of income are models.

The three are linked. A theory suggests hypotheses. The hypotheses are tested against data. Results that agree with them strengthen the theory, and results that contradict them force it to be revised or dropped. A model is the working form of a theory: it lets an economist change one thing and trace the consequence.

Assumptions make the analysis manageable. A simple model of a market assumes that there are many buyers and sellers, that consumers are well informed and that firms aim to maximise profit. These are conditions adopted for the analysis, not claims about every transaction. The test of an assumption is whether it keeps the mechanism that matters for the question, and whether changing it would overturn the conclusion.

Ceteris paribus means 'other things being equal'. It is the assumption that every influence except the one being studied is held constant, so that a single relationship can be isolated. A demand curve is the clearest example. It shows how the quantity demanded responds to the good's own price while income, tastes and the prices of other goods are held constant. When one of those other influences changes, the assumption is relaxed deliberately and the whole curve shifts.

The assumption is a statement about the model, not a claim about the world. 'A higher fuel tax reduces the quantity of fuel demanded, ceteris paribus' does not say that fuel sales must fall in every year in which a tax rises: incomes, public transport and oil prices may change in the same year. It states the direction of the tax's own effect. Measuring the size of that effect is the job of empirical work, which has to separate it from the other changes.

Empirical evidence is information obtained by observation and measurement: official statistics, surveys, the records of firms, and comparisons between places or periods. Because a whole economy cannot be put in a laboratory, economists look for situations in which a change affects one group and not an otherwise similar group, and they increasingly run controlled trials on a small scale.

Suppose an analyst claims that a congestion charge reduces traffic. The claim is positive, so it can be tested. Comparing traffic before and after the charge is a start, but the comparison misleads if fuel prices, working from home or roadworks changed at the same time. A stronger test compares the charged area with a similar area that has no charge, and checks whether traffic changed when the charge began and not before.

Correlation does not establish causation. Two variables can move together because a third influence affects both, because one of them is measured inconsistently, or because causation runs the other way. A model says which mechanism to expect and so which comparison is needed. The evidence then tests the direction and the size of the effect.

Refutation is the rejection of a hypothesis or theory because evidence contradicts what it predicts. A claim belongs to positive economics only if some possible observation could show it to be false. Evidence that agrees with a theory supports it but never proves it, because the next test may fail. Evidence that contradicts it shows that at least one assumption is wrong, and the theory is then revised or replaced.

1. Say's law predicted that general unemployment could not last. The mass unemployment of the Great Depression in the 1930s contradicted the prediction and led to Keynes's theory.
2. In the 1960s policymakers believed in a stable trade-off between inflation and unemployment: a little more of one bought less of the other. In the 1970s both rose together, which the theory said should not happen, and it was replaced by a version in which people's expectations of inflation matter.
3. The assumption that consumers are fully rational predicts that the way a choice is presented should not change what people choose. Experiments show that it does, which led to behavioural economics.

Refutation is harder in economics than in the natural sciences, for four reasons. Whole economies cannot be experimented on. Ceteris paribus never holds exactly, so a failed prediction can be blamed on something else that changed. Economic data are estimates and are often revised. And people change their behaviour when policy changes, so a relationship that held in the past may not hold once a government tries to use it. Competing theories can therefore survive side by side for decades, which is why the course teaches two models of aggregate supply.

Equality means that everyone receives the same amount or is treated in the same way: the same income, the same tax bill, the same funding for each pupil. Whether a distribution is equal is a positive question. It can be settled by measurement.

Equity means fairness. It is a normative idea, so people can agree on every fact about a distribution and still disagree about whether it is equitable. Two standards of fairness are widely used. One holds that people in the same situation should be treated in the same way. The other holds that people with a greater ability to pay should contribute more. Evidence cannot show either standard to be correct.

The two ideas can pull apart, as three examples with invented figures show.

• A tax of $1,000 a year on every adult is equal. Few people would call it equitable, because it takes a much larger share of a low income than of a high one.
• A tax of 20% of income applies the same rate to everyone, but the amounts paid are unequal. Many would judge it more equitable than the equal tax.
• Extra school funding for pupils from low-income families is unequal spending. Its aim is greater equality of opportunity, and those who support it regard it as more equitable.

In economics 'greater equity' usually refers to a less unequal distribution of income, wealth or opportunity, but the two words are not interchangeable. A clear answer says which kind of equity is meant: equity of opportunity, in which people have the same chances, or equity of outcome, in which results such as incomes are closer together.

The distinction maps onto the difference between positive and normative economics. How unequal a distribution is can be measured, for example with the Lorenz curve of topic 3.4: that is positive. Whether the measured inequality is unfair, and what should be done about it, is normative. Equity is one of the nine central concepts of the course, and many policy questions come down to a conflict between equity and efficiency. A tax that redistributes income may be judged more equitable and may also weaken the incentive to work or to invest.

Economic ideas are answers to the problems of their time, and each major school began as a reaction to an earlier one. Knowing the sequence explains why the course teaches competing models side by side.

In the 18th century the dominant view, mercantilism, held that a nation's wealth was its stock of gold and silver, so the state should restrict imports and direct trade. Adam Smith, in The Wealth of Nations, argued the opposite on each point.

• A nation's wealth is what it can produce, not the gold and silver it holds.
• The division of labour, in which each worker specialises in one task, raises output per worker.
• Individuals who pursue their own interest in competitive markets are led, 'as if by an invisible hand', to serve the interests of others, because they can profit only by supplying what buyers want at a price buyers will pay.
• Markets therefore coordinate production without anyone directing them.

The policy conclusion is laissez-faire ('leave alone'): the government should interfere with markets as little as possible, because competition already steers self-interest towards the general good. Smith did not oppose all government. He gave the state the tasks of defence, justice and those public works that no private firm would find profitable. His ideas return in the price mechanism (topic 2.3), the gains from specialisation and trade (4.1) and market-based supply-side policies (3.7).

In the 19th century classical microeconomics asked what gives a good its value. The early classical economists, Smith and David Ricardo among them, explained value by the cost of production, chiefly the labour needed. Jeremy Bentham and John Stuart Mill brought in utility: the satisfaction that a good gives its consumer. Later in the century William Stanley Jevons, Carl Menger and Léon Walras, working separately, argued that value depends on marginal utility, the satisfaction from one more unit, and not on total usefulness.

The change resolved an old puzzle. Water is essential to life and cheap; diamonds are inessential and expensive. Water is so plentiful that one more litre adds very little satisfaction, so its marginal utility is low and so is the price people will pay. Diamonds are rare, so one more adds a great deal. Alfred Marshall then joined the two halves of the argument: marginal utility lies behind demand, cost of production lies behind supply, and price is set where the two meet. That is the demand and supply diagram of Unit 2.

The concept of the margin is the general idea that came out of this work: decisions are made by comparing the extra benefit of one more unit with its extra cost, not by comparing totals. A consumer deciding whether to buy another coffee, a firm deciding whether to produce another unit and a government deciding whether to remove another tonne of pollution all reason at the margin. The same rule, continue until the marginal benefit equals the marginal cost, defines the efficient quantity in a market (2.3) and the socially optimal quantity when there are externalities (2.8).

Classical macroeconomics is summed up in Say's law, named after the French economist Jean-Baptiste Say and usually stated as 'supply creates its own demand'. Producing goods pays out wages, rent, interest and profit equal to the value of the goods, so the income needed to buy the economy's output is created by producing it. If households save part of that income, the saving lowers the rate of interest and firms borrow it to invest, so spending is maintained.

• A general glut, with unsold goods across the whole economy, cannot last.
• Unemployment is temporary.
• Flexible wages, prices and interest rates return the economy to full employment without help: if workers are unemployed, wages fall until firms hire them.
• The government therefore has no need to manage total demand.

The law is about the economy as a whole. It does not deny that a single market can have a surplus. It returns in the monetarist/new classical view of aggregate supply in topic 3.2.

Karl Marx, writing later in the 19th century in his work Capital, accepted the classical idea that labour creates value and turned it against the classical conclusion that free markets serve everyone. The Marxist critique is a chain of reasoning.

1. Workers own only their labour. The owners of capital own the means of production: the factories, machines and land.
2. Workers are paid less than the value of what they produce.
3. The difference, which Marx called surplus value, is the source of profit. He described this as exploitation.
4. Competition forces the owners of capital to reinvest profit and to replace workers with machines.
5. Capital becomes concentrated in fewer hands and inequality widens.
6. Workers cannot afford to buy all that is produced, so crises of overproduction recur. This is a direct rejection of Say's law.
7. Conflict between the two classes would, Marx predicted, end capitalism and bring collective ownership of the means of production.

Marx's ideas shaped the planned economies of the 20th century (the economic systems of topic 1.1), the labour movement and the growth of the welfare state.

Two things should be kept apart when the critique is explained. The collapse that Marx predicted did not happen in the industrial economies, where real wages rose over time. The questions he raised remain central to economics: how income is divided between labour and the owners of capital, which bears on the study of inequality in topic 3.4, and whether market economies are prone to recurring crises.

The 20th century began with classical macroeconomics in command and ended with two rival schools. The turning point was the Great Depression of the 1930s, when mass unemployment lasted for years. Say's law said that this could not happen.

John Maynard Keynes, in The General Theory of Employment, Interest and Money, explained why it could. His argument, the Keynesian revolution, is the origin of Keynesian economics.

• Output and employment are determined by aggregate demand, the total planned spending in the economy. Firms produce only what they expect to sell.
• Wages and prices do not fall easily, and a general fall in wages cuts households' spending. The economy can therefore settle at an equilibrium below full employment and stay there.
• Investment depends on firms' expectations, which are volatile. When firms are pessimistic they do not invest, however low the rate of interest, so private spending can stay depressed.
• The government should therefore raise demand itself, by spending more or taxing less.

The rise of macroeconomic policy followed. After the Second World War governments accepted responsibility for the level of employment. National income accounts were developed to measure total output and spending (topic 3.1), and fiscal policy and monetary policy were used to manage aggregate demand: to expand it in a recession and restrain it in a boom. Policymakers also came to believe that they faced a stable trade-off and could choose a little more inflation in exchange for lower unemployment.

The 1970s broke that belief. Inflation and unemployment rose together, a combination called stagflation. The simple Keynesian model of demand management had not predicted it and could not cure it: raising demand would worsen inflation, and reducing demand would worsen unemployment.

The monetarist/new classical counter-revolution was the response. Monetarism is associated with Milton Friedman, who argued that sustained inflation is caused by the money supply growing faster than output, and that there is a natural rate of unemployment set by the supply side of the economy. Holding unemployment below that rate by raising demand works only while people underestimate inflation. Once they expect higher inflation and demand higher wages, unemployment returns to the natural rate and inflation is permanently higher. New classical economists, Robert Lucas prominent among them, added rational expectations: people anticipate the effects of policy, so systematic demand management cannot raise output even for a short time.

The policy conclusions were to aim monetary policy at stable prices, to make central banks independent of government, and to raise output through the supply side of the economy. They return in the vertical long-run aggregate supply curve (3.2), inflation targeting (3.5) and market-based supply-side policies (3.7).

The two schools differ on one assumption: how quickly wages and prices adjust. If they are slow to fall, as Keynesians hold, a fall in demand leaves the economy below full employment for a long time and government action can help. If they adjust quickly, as monetarist and new classical economists hold, the economy returns to full employment output by itself and extra demand ends up in higher prices. The Keynesian aggregate supply curve and the vertical long-run aggregate supply curve of topic 3.2 are these two views drawn as diagrams. Neither school has displaced the other: governments still turn to fiscal and monetary stimulus in deep recessions, and to independent central banks with inflation targets in normal times.

In the 21st century economics has drawn increasingly on other disciplines. The most developed dialogue is with psychology. Standard models assume that consumers are rational: they have consistent preferences, use the information available and choose what is best for themselves. The psychologists Daniel Kahneman and Amos Tversky showed in experiments that people often rely on mental shortcuts, respond to the way in which a choice is presented and feel a loss more sharply than an equal gain. Behavioural economics builds these findings into economic analysis. It replaces the fully rational consumer with one who has limited attention, limited self-control and a concern for fairness.

Two things changed as a result. The method changed: economists now test their assumptions about behaviour directly, in laboratory and field experiments, where earlier they judged a model only by its predictions. Policy changed as well: governments design choices so that the easiest option is the one that most people would want on reflection, for example by enrolling workers in a pension scheme automatically while leaving them free to opt out.

The second development is a growing awareness that the economy, society and the natural environment depend on one another. Earlier models, including the circular flow of income, treat the environment as outside the economy: resources arrive and waste disappears at no cost. The current view places the economy inside society, and society inside the environment. Production draws on finite resources and on the capacity of air, water and soil to absorb waste, so growth that depends on using these up cannot continue indefinitely. This is the link between scarcity and sustainability made in topic 1.1. It returns in common pool resources (2.8), the environmental cost of growth (3.3) and sustainable development (4.7).

A circular economy is the response proposed. The existing pattern of production is linear: take resources, make products, dispose of them. A circular economy rests on three principles.

1. Design out waste and pollution, so that products are made to last and to be taken apart.
2. Keep products and materials in use, through reuse, repair, remanufacture and recycling.
3. Regenerate natural systems, returning nutrients to the soil and using renewable energy and materials.

Four reasons are usually given for moving towards a circular economy.

• Resources are finite. A linear economy must keep extracting new materials, so scarcity becomes more severe for future generations.
• Waste and pollution impose costs on third parties and on the ecosystems that absorb them. These costs are not paid by the firms and consumers who cause them.
• Extracting and processing materials is a large source of greenhouse gas emissions, so using fewer new materials slows climate change.
• Firms that reuse materials spend less on inputs and are less exposed to swings in raw material prices.

The change is slow because of three obstacles. Products have to be redesigned, which is costly. Networks for collection and repair do not yet exist at the scale needed. And market prices leave out environmental costs, a negative externality in the terms of topic 2.8, so the linear model looks cheaper than it is.

Each row of the table is a reaction to an earlier one. The last column shows where the idea is used again in the course.

A model is judged against its purpose. The production possibilities curve explains opportunity cost and productive potential, but a single curve shows nothing about prices or about financial instability. The circular flow shows how sectors depend on one another, but not how income is shared between households. A model is not weak merely because it omits something. It is weak when the omission changes the answer to the question being asked.

Assumptions should be visible. If a model assumes rational, fully informed consumers, evidence of habit and incomplete information limits what it can predict about behaviour. Adding assumptions does not make a model better or worse in itself: more realism can improve relevance while making the causal chain harder to follow. The history of economic thought shows the test at work. Say's law, the stable trade-off between inflation and unemployment and the rational consumer were each kept while their predictions held and revised when the evidence turned against them.

• A positive statement can be tested; a normative statement rests on a value judgement.
• Hypothesis: a testable statement. Theory: a tested general explanation. Model: a simplified representation of a theory.
• Ceteris paribus isolates one relationship inside a model. It is not a claim that nothing else changes in reality.
• Evidence can support a theory or refute it; it cannot prove it. Correlation is not causation.
• Equality is sameness and can be measured. Equity is fairness and is a value judgement.
• 18th century: Smith and laissez-faire. 19th century: utility and the margin, Say's law, Marx's critique.
• 20th century: the Keynesian revolution, the rise of macroeconomic policy, the monetarist/new classical counter-revolution.
• 21st century: behavioural economics, and the case for a circular economy.