ECON30019 Chap.3 Revealed Preference and the Choice Axioms
Revealed Preference and the Choice Axioms
The previous chapter built a theory of an object nobody can observe. Revealed preference theory is the bridge to data: it asks how far observed choices pin down an underlying ranking, and it is what turns rationality from a stance into something a table can falsify.
The chapter introduces the choice correspondence, which returns a set rather than a single item so that indifference can be recorded as indifference, and then the two conditions that make it consistent with a ranking. Axiom alpha is a contraction condition running from large menus to small ones; Axiom beta is an expansion condition about what happens to ties when the menu grows.
Together they are exactly the weak axiom of revealed preference, and the representation theorems say a correspondence can be represented by a rational relation, or equivalently by a utility function, if and only if both hold.
What this chapter covers
- 01
Why the preference relation is unobservable and the choice data is not
- 02
The choice correspondence, and why it returns a set
- 03
What a rational chooser takes, written two ways
- 04
Axiom alpha, its statement and its large-to-small intuition
- 05
The pattern Axiom alpha forbids, written as a witnessing triple
- 06
Axiom beta, its statement and its small-to-large intuition
- 07
The two representation theorems, and why both run in both directions
- 08
Testing a table of observed choices, step by step
- 09
Reading a binary menu as strict preference rather than as indifference
- 10
Where every anomaly in the next chapter lands
Test a table of observed choices
- 2Read each binary menu as a revealed strict preference: tram over bike, bike over rideshare, rideshare over tram.
- 1Apply transitivity to the first two: tram over rideshare, which contradicts the third row, so no complete and transitive relation generates the table.
- 1Translate the same three rows into utility inequalities; the identical contradiction reappears, so no utility function exists either, by the representation theorem.
- 1Find the witnessing triple: take the item rideshare, the small menu of tram and rideshare, and the large menu of all three.
- 1State the violation: rideshare is chosen from the small menu, which is a subset of the large one, yet the large menu returns only the tram, which is the pattern Axiom alpha forbids.
Key terms
- Revealed preference
- The study of when and how observed choices are related to underlying preferences, which is what makes an unobservable ranking testable against a table of data.
- Choice correspondence
- A rule assigning to each menu the subset the decision maker takes from it. It returns a set so that ties can be recorded rather than broken by an invented rule.
- Contraction condition
- A requirement about what survives when a menu shrinks. Axiom alpha is the example here: an item chosen from a large menu must still be chosen from a smaller menu that contains it.
- Expansion condition
- A requirement about what survives when a menu grows. Axiom beta is the example here: two items tied in a small menu cannot be separated by adding options.
- Representation theorem
- A statement that a body of data fits a model exactly when it meets a stated set of conditions. The conditions are the examinable part, because they are the only part that can be checked.
Revealed Preference and the Choice Axioms FAQ
Why is it called a correspondence rather than a function?
So that it can return more than one item. Forcing a single answer would require inventing a tie-break the data never supplied, and it would make genuine indifference invisible. The practical consequence is that a violation of Axiom beta can only be detected when the data actually contains a multi-element choice set.
How do I tell which of the two axioms a dataset breaks?
Look at the direction. A winner discarded when options are removed is a failure of Axiom alpha. A tie broken when options are added is a failure of Axiom beta. Naming the direction and the two menus that witness it is what earns the marks, since the axiom name on its own could be a guess.
If someone chooses one item from a two-item menu, might they have been indifferent?
No. Choosing one item from a two-item menu reveals strict preference, because the rejected item was available and was not taken. Indifference is revealed only when the choice set itself contains both items, and treating a singleton choice as a possible tie throws away the strict ranking you need to display a contradiction.
Exam move
Practise this chapter with a pen and a small table, because the assessed task is always the same shape. Draw four or five menus over three or four alternatives, convert every row into a revealed preference, and hunt for a cycle.
Then do it in reverse: construct a table that violates Axiom alpha only, and another that violates Axiom beta only, since being able to build the two cases separately is the fastest way to stop confusing them. Finish by writing an explicit utility function for a table that has no cycle, because the construction is worth marks and takes one line.
Working through Revealed Preference and the Choice Axioms in ECON30019? Sia is AskSia’s AI Economics tutor — ask any ECON30019 Revealed Preference and the Choice Axioms question and get a clear, step-by-step explanation grounded in how ECON30019 is taught and assessed. Read this chapter free, then take your hardest questions to Sia.
