An example of a popular joke among economists: two economists see a Ferrari. "I want one of those," says the first. "Obviously not," replies the other. To get a smile out of this it is necessary (but not, alas, sufficient) to know about revealed preference. This is the notion that what you want is revealed by what you do, not by what you say. Actions speak louder than words. If the economist had really wanted a Ferrari he would have tried to buy one, if he did not own one already. Economists have three main approaches to modelling demand and how it will change if prices or incomes change. * The cardinal approach involves asking consumers to say how much utility they get from consuming a particular good, aggregating this across all goods and services, and calculating how demand would change on the assumption that people will consume the combination of things that maximises their total utility. * The ordinal approach does not require consumers to say how much utility they get in absolute terms from consuming a particular good. Instead, it asks them to indicate the relative utility they get from consuming one item compared with another, that is, to say if they prefer one basket of goods to another, or are indifferent between them. * The third approach is revealed preference. To model demand it is only necessary to be able to compare an individual's consumption decisions in situations with different prices and/or incomes and to assume that consumers are consistent in their decisions over time (that is, if they prefer wine to beer in one period they will still prefer wine in the next).