Module 12 of 14 · Faster & relative value

Spreads and relative value

Relative-value strategies trade one futures price against another, hoping to profit from the relationship while hedging away the direction of the market:

  • Cross-market spreads — for example 5-year against 10-year Treasury futures, or gold against silver.
  • Triplets — three legs, such as a butterfly on the yield curve or a refining spread (crude against products).
  • Calendar spreads — two delivery months of the same market, possible only where several expiries are liquid (energy, grains, interest-rate and volatility futures).

The mechanics follow the earlier modules: define the spread as a price in its own right (R units of A minus one unit of B), estimate its volatility in price terms, size it with a risk target, and apply trend or carry rules to the spread price. Choose the ratio R so that the legs carry similar risk — roughly the ratio of their volatilities — or use the market's conventional ratio.

The catch is that hedging works too well. Two highly correlated legs leave a tiny residual risk, and hitting a risk target on a tiny residual takes a huge gross position.

🎛 Spread risk lab

4%
7%
0.95
1.75
10%
0.002

Spread volatility

2.21%

per $1 of leg B

Gross leverage needed

12.4×

outright B: 1.4×

Cost per trade

0.0126

6.3× one leg

Risk-neutral ratio

1.75

σB ÷ σA

Spread return per $1 of leg B = R × (return of A) − (return of B), so its volatility is √((RσA)² + σB² − 2ρRσAσB). The more correlated the legs, the smaller that residual — and the more notional you need to reach a risk target. Risk-adjusted cost compares what you pay on both legs with the small risk you actually trade. The chart shows what happens to your realised risk if correlation drops after you have sized the position. Illustrative inputs, e.g. two Treasury futures of different maturity.

Three reasons spreads disappoint

Leverage.The residual is small, so gross notional is many times capital. That is fine while correlation holds — and the chart shows what happens to your real risk when it doesn't. Costs. You pay spread and commission on both legs but earn on a much smaller risk, so the cost per unit of risk is several times that of trading either leg outright. Tails. Spreads often sit near an equilibrium for long periods and then jump to a new one, giving ugly lower-tail ratios that a volatility estimate fitted on the calm period cannot see.

Calendar spreads add their own problems: legs must always be rolled together, and closer expiries make leverage and costs worse. In practice simple trend and carry rules applied to spreads tend to deliver lower risk-adjusted returns than the same rules on outright markets. Relative value is best left to traders who have already exhausted the liquid outright markets and have the capital, systems and nerve for high leverage.

Things to try

  • • Set correlation to 0.95, then 0.99: the leverage needed more than doubles for that small change in correlation.
  • • Size at ρ = 0.95, then read the chart at ρ = 0.85 — realised risk rises by about 70% with no change in your position.
  • • Move the hedge ratio away from σB ÷ σA and watch the spread start to behave like an outright position in one of the legs.

Original interactive lesson. Labs run on a seeded synthetic market in your browser, built to illustrate mechanics, not to predict any real market. Contract figures are illustrative; check exchange specifications. Educational analysis, not investment advice.

How are we doing?