Relative-ValueHard
4 min read · 721 words
What the seat actually does
A relative value seat is not trying to say where a market goes. It looks for two instruments whose payoffs are nearly the same and whose prices are not, buys the cheap one, sells the dear one, and waits for the difference to close.
The direction is deliberately removed. If both legs rise or both fall, the position barely notices — what is left is the gap between them, which is usually small. A gap worth basis points only becomes a return worth having when it is held in size, and size here means borrowed money. That trade is the whole of the seat: give up the large moves, keep a small one, and finance it.
- On the curve — one maturity against another, when the shape between them has moved away from where the rest of the curve says it should be. See curve construction.
- Cash against derivative — a bond against the swap that reproduces it, which is what an asset swap makes visible as one number.
- One capital structure against itself — a convertible bond against the shares it converts into, or a bond against the credit default swap on the same borrower.
- Two listings of one company — an ADR against the local line, where the same claim trades in two places and two currencies.
A day, and where it goes
- The spreads themselves — where each pair stands against its own history, and which ones moved without either leg doing anything interesting.
- Financing. What the repo rate is on every bond held, what the borrow costs on every one sold short, and whether either changed overnight. On this seat that is not back-office work; it is the position's profit and loss.
- Margin. What the clearing house and the prime broker are asking for today, and what they would ask for if the spread went the wrong way. See margin and collateral.
- The reason. Why the gap exists at all — a rule, a forced seller, a settlement difference — because a gap with no explanation is usually one leg not being what it looks like.
What it is measured on
- Return against capital at risk, not against notional. The notional on a relative value book says almost nothing about it.
- The worst drawdown, because the losses arrive compressed into short periods rather than spread evenly. See the arithmetic of drawdowns.
- Whether the leverage was known. A book can look calm and carry a multiple of its capital in gross positions — the question is whether that was measured or discovered.
- Financing spread earned or paid across the whole book, which on many pairs is a larger number than the convergence being waited for.
What it touches on this site
- The financing — repo and securities lending, the two markets this seat lives inside.
- Where the leverage hides — the analysis of exactly that, which is written around positions of this shape.
- The instruments — asset swaps, convertibles and basis swaps.
- When it goes wrong — 1998, where the spreads were right about the world and wrong about the calendar.
How it goes wrong
- The spread widens before it closes. Being right eventually is not a defence against a margin call today.
- Financing is withdrawn. The position was never sized against the market; it was sized against a lender's willingness, and that can change overnight.
- The two legs are not the same thing. There is almost always a difference — a delivery option, a coupon convention, a tax treatment — and the question is whether it was priced or ignored.
- Everybody holds the same pair. A crowded convergence trade unwinds as a divergence, because the exits are the same door.
Concepts to master
- Leverage turns a small edge into a real return and a small move into a real loss, symmetrically. See leverage.
- The carry of the position — what it earns or pays for simply existing, before the spread does anything.
- Liquidity is a condition, not a property. Both legs are liquid until the day the trade needs them to be. See what liquidity costs.
- A basis is information. When two prices for the same exposure refuse to converge, that is usually somebody's constraint rather than somebody's mistake.