What makes up the price?

A ten-year gilt pays for two different things. One is the path of short rates being locked in. If the one-year yield were simply going to average some number for a decade, and nobody charged for risk, the ten-year yield would be that average. It is not. The difference is the term premium: the extra yield for fixing a rate for years instead of rolling one-year gilts.

It is not a forecast error, and it is not a credit rating. Inflation may come in higher than expected. Real rates may rise, and the bond's price fall. Or the long gilt may be the bond that dealers and pension funds have to sell at once. Those are different risks. They all sit in the gap between the yield and the expected path of short rates.

Expected real rate. The model's expected path of the one-year gilt yield, after taking out the inflation path below. This is a real return only if inflation arrives as assumed and the one-year gilt is rolled. It is not a Bank Rate forecast.

Expected inflation. An assumed path for RPI, not a survey and not the breakeven itself. Until February 2030, when the Retail Prices Index is due to be aligned with CPIH, the path uses today's 2.5-year breakeven — the shortest maturity on the real curve — as the near-term RPI figure. After that date it fades toward 2%, the inflation target. Setting the short breakeven equal to expected inflation forces the inflation risk premium to about zero there. That is a choice, made so the chart has an anchor.

Inflation risk premium. Today's breakeven, nominal spot minus real spot, minus that path. It is the extra inflation compensation further along the curve. It rises with maturity. Treat the whole breakeven as expected inflation instead, and this slice moves into the one above, with the real-rate term premium rising by the same amount. The prices alone do not say which labelling is right.

Real-rate term premium. What is left of the model's term premium once inflation risk and the gilt-over-swap slice are taken out. Compensation for uncertainty about real rates. It is the residual of this allocation, so it inherits every assumption above.

Fiscal, credit, liquidity and technical. The nominal gilt spot yield minus the sterling overnight-index swap at the same maturity. One bundle: the public finances, how readily a dealer will hold the bond, and the balance-sheet and pension flows that show up in a squeeze. It is wider at long maturities. A trading desk sometimes calls the flat two-year gap "liquidity" and the slope beyond it "credit". That is a label, not something this curve identifies, so the slices stay together. The same gap, in par yields, is on the premium page, which then compares Britain with the United States and the safest euro governments. That is a different question.

Joslin, Singleton and Zhu (2011) write this same family of Gaussian arbitrage-free models in a form that is estimated by maximum likelihood. Adrian, Crump and Moench (2013) showed the prices can instead be recovered with three linear regressions: one for how the factors move, one for how bond returns load on those moves, and one cross-section that forces a single price of risk to fit every maturity. The single price of risk is what "arbitrage-free" means here. The same risk is not allowed one price at five years and another at twenty. This page uses that regression estimator. Five factors, the specification they preferred. The factors are fit on the whole sample, so the path is not a number a trader could have published on the day without later data.