The Product Round of a Markets InterviewSome background helps

Five kinds of product question, what a complete answer to each one contains, and the five ordinary ways a well-prepared candidate still loses the round.

What the product round is actually testing

• Not recall. Anybody can memorise that a swap exchanges fixed for floating. The question behind every product question is whether you know what the contract obliges someone to do, and when.
• The tell is always the same: a candidate who has memorised a definition can say what a product is and stops. A candidate who understands it can keep going — who is on the other side, where the money comes from, what has to happen for it to lose, and what it would cost to get out on a bad day.
• So the useful preparation is not more products. It is fewer products, each one you can be pushed on for three minutes without running out. Ten instruments held that firmly beat sixty held loosely, every time.
• Which ten? The ones everything else references: common stock, government bond, corporate bond, T-bill, repo, FX spot, FX forward, equity future, equity option, interest rate swap. Every exotic on this site is a rearrangement of those.

Family 1 — “Tell me what a X is”

• What is being checked: whether you can define an instrument without using the instrument's own name in the definition, and without a diagram.
• What a complete answer contains, in this order: who wants what; what the contract obliges each side to do and when; where the cash comes from; what makes it lose. Four sentences. The template and twelve worked examples are on Explain it out loud.
• The most common incomplete answer is the payoff without the obligation. “A call gives you the right to buy at the strike” is true and describes half the contract — the half where you are long. Say who wrote it and what they are now obliged to do, and the answer is whole.
• Where the follow-up goes: almost always to the other side of the trade. Who takes the opposite position, and why is that rational for them? The site has a whole page on that habit — The other side of the trade.

Family 2 — “What happens to it if Y moves?”

• What is being checked: whether the mechanism is in your head as a machine with inputs, or as a sentence you learned.
• The move is usually one of five: the underlying price, the interest rate, the credit spread, volatility, or time. A product's identity is largely which of those five it is exposed to and with what sign.
• Say the sign and the size separately. “Higher rates hurt it” is the sign. “A one-percent rise costs about eight percent of the price, because modified duration is around eight” is the size, and the size is what separates a definition from an understanding. The arithmetic behind that sentence is on Markets arithmetic without a screen.
• The second-order part is where the interesting answers live: convexity for a bond, gamma for an option, jump-to-default for a CDS. First order says which way; second order says what happens when the move is large, and large moves are the only ones anybody remembers.
• Worth being able to answer cold: why a bond falls when rates rise (the long version), why an option can lose on a day the underlying went your way (three ordinary reasons), and why a floating-rate instrument still has credit duration even with no rate duration (FRN).

Family 3 — “How would you value it?”

• What is being checked: whether you reach for a framework or for a formula. Formulas are looked up; frameworks are what you are being hired for.
• There are only three frameworks on this site, and nearly every instrument is priced by one of them:
  • Discount the cash flows — bonds, loans, annuities, anything with a schedule. IRR and NPV.
  • Replicate it and charge what the replication costs — forwards, futures, swaps, options. This is the one candidates skip, and it is the one that generalises: a forward is priced by borrowing one currency and lending the other, not by forecasting anything.
  • Compare it with something that already has a price — relative value, spread over a benchmark, multiples. Valuation and Reading a market number.
• Then say what the framework assumes, because the assumption is the risk. Replication assumes you can actually borrow, actually short, and actually trade continuously; the case studies on this site are largely a record of what happens when one of those three stops being true.
• If you are asked for a number, give the approximation and say it is one. An at-the-money option worth “about forty percent of spot times vol times the root of time” is a better answer than a wrong Black–Scholes recital, and it can be checked in your head. The derivation is on the arithmetic page.

Family 4 — “What could go wrong with it?”

• What is being checked: whether your risk vocabulary is a list or a structure.
• A structure that holds for every instrument on this site: market risk (the price moves), credit risk (the other side does not pay), liquidity risk (you cannot get out at anything like the quoted price), funding risk (you are right but the margin call is now), operational and legal risk (the document does not say what everyone assumed it said).
• Name all five, then say which one dominates for this product. For a repo it is funding and collateral; for a high-yield bond it is credit and liquidity together; for a autocallable it is a short volatility position the buyer usually has not noticed.
• One historical example, told in a sentence, is worth more than three adjectives. The case studies here exist for exactly this: the six patterns they fall into are set out in How products fail, and “this is the liquidity-mismatch pattern, the same one as an open-ended fund over illiquid assets” is a complete answer.
• Do not reach for a tail risk while the ordinary one is unmentioned. The commonest way a structured note loses money is not a crash; it is four dull mechanisms written into the term sheet.

Family 5 — “Talk me through a market you follow”

• What is being checked: whether you have a view that survives a follow-up question, and whether you know the difference between a view and a fact.
• A shape that works: here is the thing that changed; here is what the market seems to be pricing as a result; here is the specific figure I would watch to know whether that is right; here is what would make me wrong. Four sentences again.
• The last one is the one that gets remembered. A candidate who cannot say what would falsify their view has not got a view, they have a preference.
• Quote a number you can source and date, or quote none. This site deliberately publishes no live market sizes — the reason is written into its own rules — and the same discipline reads as strength rather than ignorance in a conversation: “as of last Friday's close” is precise, “about seven trillion” with no date is an invitation to be corrected.
• Depth beats breadth here too. One market you have actually followed for a month, with a number you check daily and could explain to a stranger, is worth more than a tour of five.

The trade-pitch question, answered honestly

• “Pitch me something” is not asking for a prediction. It is asking whether you can state a position, its size, its horizon and its exit in one breath — which is the whole job, compressed.
• The four parts, none of them optional: what you would do; how much (and why that much rather than twice that); what has to happen and by when; where you would stop. Sizing a position and the arithmetic of drawdowns cover the two that candidates leave out.
• The dullest strength you can show: knowing what the trade costs to hold. Financing, borrow, roll, carry, spread. A view that is right over six months and negative-carry throughout can still lose. What liquidity costs and Who gets paid are about exactly that.
• And say plainly that it is a view, not a recommendation. That distinction is not a formality — it is the same one this entire site is built on.

Five ordinary ways a prepared candidate loses the round

• Answering a different, easier question. Asked how a product is hedged, describing what it is. If you need a moment, say what the question is asking before you answer it.
• Numbers with no units and no date. “Spreads are wide” — wide against what, measured how, compared with when?
• Confidence past the edge of knowledge. “I don't know, but here is how I would work it out” is a strong answer and is scored as one. An invented number is unrecoverable, because everything else you said now has to be re-checked.
• Jargon used as a substitute for the mechanism. If you say “negative convexity”, be ready to say what it means with no Greek letters at all: the position gets shorter as prices rise and longer as they fall, so it loses on both sides of a round trip.
• Never mentioning the other side. Every product on this site exists because two parties wanted opposite things. A candidate who only ever describes the buyer has understood half of every instrument they named.

A two-week route through this site

• Days 1–3 — the ten instruments, Beginner and Intermediate levels only: Path 1. Aim to explain each in four sentences without notes; the template is on Explain it out loud.
• Days 4–5 — the mechanisms: leverage, volatility, liquidity, margin and collateral, the yield curve. These are the five that turn up inside answers about other things.
• Day 6 — the arithmetic: the rules of thumb, until each one takes under ten seconds.
• Days 7–9 — the desk you are interviewing with: which desk trades what, then that market's asset-class page end to end, including its quiz.
• Day 10 — risk: How products fail, plus two case studies from the desk's own market, well enough to tell each in three sentences.
• Days 11–12 — documents: one playbook in full — a term sheet for a structuring seat, an option chain for a derivatives desk, a factsheet for asset management.
• Day 13 — vocabulary: the words of the first week, so nothing said to you is a surprise.
• Day 14 — test it: the market exam, then re-read only what you got wrong. Everything you can already explain is finished; there is no credit for revising it again.