Economics Oxbridge Interview Questions 2026 — Model Answers

1. Home
2. Resources
3. Oxbridge Interview Questions
4. Economics

A monopolist's cost curve, a consumer's budget constraint, and Bayes' theorem used to work out the real odds a faulty aircraft wing is actually faulty: the Economics I pack turns each into a model you build yourself rather than a diagram to recognise. It opens with the pack's own ice-cream-van location puzzle and carries all ten questions through to a finished argument, not just a stated answer.

Oxford and Cambridge Economics interviews are not tests of A-level Economics knowledge. They assess your ability to think like an economist: to identify economic forces in unfamiliar situations, build a simple model, reason through its implications, and engage critically with your own conclusions. Interviewers are looking for the analytical instinct — the ability to ask "what are the incentives here?" and "what would change at the margin?" — not for definitions or memorised diagrams. The Economics I pack is built for that: ten questions grouped into microeconomics, macroeconomics and economic history, and mathematics and statistics, each one reasoned through in full rather than answered in a line. It leans mathematical — profit-maximisation conditions, partial derivatives, Bayes' theorem — which suits a Cambridge Economics applicant especially well, though the economic-history and current-affairs questions give Oxford E&M candidates useful practice moving between derivation and open discussion.

What Are Oxford and Cambridge Economics Interviews Like?

Oxford Economics and Management (E&M) candidates typically have two 25–30 minute panel interviews at their applied college, covering economic theory and quantitative reasoning. Cambridge Economics candidates have two panel interviews that are particularly mathematically demanding, reflecting the quantitative emphasis of the Cambridge Economics Tripos from the first year. The TSA (Thinking Skills Assessment) is used by Oxford for E&M shortlisting for 2026 entry, transitioning to the TARA from 2027. Cambridge Economics has no pre-interview test. Approximately 100 students are admitted to Oxford E&M annually and 140 to Cambridge Economics.

What Types of Questions Come Up in Economics Interviews?

Consumer and firm microeconomics. Economics I opens with a shopper's budget constraint built up from nothing — no diagram supplied until you set the problem up yourself — and then a monopolist's profit-maximising output, derived algebraically from its own marginal revenue and marginal cost rather than read off a textbook graph. Nothing is quoted back at you; the mark comes from constructing the model, not naming it.

Spatial competition and game theory. Two identical vendors choosing where to set up on a beach turns, a few prompts later, into the model political scientists use to explain why parties converge on the centre ground; two children splitting a pile of sweets turns into a repeated game with a discount rate, and then into a question about how a government enforces a law people would rather break. Marks are won by seeing the structure under an everyday scenario, not by reciting a definition.

Trade, economic history and current affairs. A real-exchange-rate question is solved with partial derivatives — holding one variable constant while another moves — sitting alongside an open essay on how living standards have changed over the course of history, and a question on what actually moved stock markets when COVID-19 hit. Cambridge interviewers in particular expect a candidate to move between calculus and discursive argument inside a single sitting.

Mathematics and statistics. A loan-repayment problem solved as a difference equation, Bayes' theorem applied to a faulty product test, and a counting problem built around ordering a sequence of cards: the pure quantitative reasoning that Cambridge Economics leans on hardest and Oxford E&M tests more lightly.

How to Structure an Economics Interview Answer

Apply the identify-model-imply-evaluate structure. First, identify the economic forces: "This is fundamentally a problem of negative externalities — the private market does not internalise the social cost of carbon emissions." Second, model the mechanism: "Without intervention, output is at the competitive equilibrium Q*, which exceeds the socially optimal quantity Q-star because marginal private cost lies below marginal social cost." Third, reason through the implications: "A Pigouvian tax equal to the external marginal cost shifts the private cost curve up, restoring the efficient outcome." Fourth, evaluate: "The challenge is measuring the externality precisely — if the tax is set incorrectly, it overcorrects or undercorrects." This four-step structure can be applied to any economic question. The step candidates drop is the fourth. An answer that stops at "the tax is justified" invites the interviewer to supply the objection instead, and you then spend the rest of the conversation defending a claim you never qualified; saying what would have to be true for your own conclusion to be wrong takes that move away from them.

A Worked Economics Interview Question

A market described in one sentence, no numbers, nothing to recall. What follows is not advice about how to answer it. It is the answer, at the length you would really be given, including the turn where the first idea failed.

The question

"Suppose the law changed so that anyone selling a second-hand car had to disclose every defect they knew about. What happens to that market?"

What the interviewer asks next

Not one of them arrives at the start. Each lands after you have committed to something.

• "You have assumed the buyer is risk-neutral. Make her risk-averse instead. Does your story get stronger or weaker?"
• "Suppose only one car in ten is a bad one, rather than one in two. Does your argument still go through?"
• "The law says known defects. What about the ones the seller has not noticed?"
• "If disclosure helps honest sellers this much, why did they need a law? Why not simply tell the buyer?"
• "Name me somebody this law makes worse off."
• "Does anything you have said depend on it being cars?"

Working it through

The attempt that did not survive. My first move was a supply and demand diagram: disclosure means better-informed buyers, better-informed buyers are more confident, confidence shifts demand right, so price and quantity both rise. I said it out loud and was asked which good the diagram was for. That is where it fell over. There is no single good here. There are two — a sound car and a defective one — and the whole problem is that they change hands at one price because the buyer cannot tell them apart. One demand curve had assumed away the thing the question was about, and "confidence" was standing in for a mechanism I had not found.

Starting again, with numbers I choose. These are invented to make the arithmetic visible, not because I know what used cars sell for. Half the cars on the forecourt are sound and half defective. A buyer would pay up to £8,000 for a sound one and £2,000 for a defective one; an owner will not part with a sound car below £6,000, or a defective one below £1,000. If both sides could see which was which, every car trades. Hide the type from the buyer alone and she faces a coin flip, so the most she will offer is what the average car is worth to her: 0.5 × £8,000 + 0.5 × £2,000 = £5,000.

And £5,000 is below £6,000. The owner of a sound car will not sell at the only price on offer, so sound cars leave. The buyer works that out too: if what remains is the defective ones, she will pay at most £2,000. What is left is a market in defective cars trading between £1,000 and £2,000, and every sound car that never sells destroys £8,000 − £6,000 = £2,000 of surplus that was sitting there. The question as asked is that story in reverse: compulsory disclosure removes the asymmetry, so the sound cars have no reason to withdraw.

The one-in-ten prompt is where the argument nearly breaks. If nine cars in ten are sound, the buyer will pay up to 0.9 × £8,000 + 0.1 × £2,000 = £7,400, which clears the £6,000 the owner needs. Nobody withdraws; the market survives with no law at all. So the collapse is not automatic — it turns on how bad the mix is. Write the condition down rather than hand-wave it: sound cars stay while p × £8,000 + (1 − p) × £2,000 ≥ £6,000, which rearranges to 6,000p ≥ 4,000, so p ≥ 2/3. Below two-thirds sound the market unravels; above it, it does not. Volunteering that inequality unprompted is worth more than the paragraph before it.

Why a law, if honesty pays? The prompt I fumbled. An honest seller can already say the car is sound; the difficulty is that a dishonest one says the same words, and a statement that costs nothing to make carries no information. What separates them is one the liar cannot afford — a warranty, an independent inspection, a returns policy only somebody with a sound car would offer. The law is not conjuring information out of nothing. It puts a cheap public rule in place of a private signal honest sellers were otherwise buying for themselves.

Somebody loses. Owners of defective cars had been selling into a pool that priced their car as an average car; disclosure prices it as what it is. The people a rule hurts by compelling them to say what they know are exactly the people who knew it, which is why it has to be a rule. Risk aversion runs with the main argument, not against it: a buyer who dislikes the gamble values the coin flip below £5,000, so the two-thirds threshold rises and sound cars withdraw sooner. And the wording sets its own limit — disclosing known defects leaves the unknown ones where they were, so the asymmetry shrinks rather than disappears, which is why inspections do not vanish the day it passes.

What is being tested here

Not whether you have read Akerlof; naming the paper takes two seconds and buys nothing. Three things: whether you notice that one price is being asked to clear two different goods; whether you will put numbers on a problem that arrived without any, and use them to check whether a seller actually participates, rather than staying at "buyers are more confident"; and whether, having reached a conclusion you like, you can name the person it hurts. The candidate who reaches p ≥ 2/3 was not cleverer. They were concrete earlier.

What Students Say

Frequently Asked Questions