Both kinds of index-linked gilt scale their coupon and redemption by the index ratio, and the difference between them is the indexation lag — which decides which month's RPI a payment points at, when that figure is published, and whether the quoted clean price has the inflation uplift inside it.
One mechanism, two vintages
Every index-linked gilt in issue works the same way in one respect: the semi-annual coupon and the principal repaid at maturity are both adjusted in line with the Retail Prices Index, by multiplying each by the index ratio. A real coupon fixed at issue is applied to an uplifted nominal, and the same uplift applies to the £100 of nominal that comes back at maturity. The Debt Management Office describes the adjustment as taking account of accrued inflation since the gilt was first issued.
Where they diverge is in the lag. The index a payment points at is not the index for the month the payment falls in; it is the index for an earlier month, and how much earlier is a property of the gilt's own terms. Two lags are in use, eight months and three, and both groups of gilt are in issue together. The Debt Management Office's index-linked data report groups the gilts by lag, and GiltIQ classifies each gilt the same way so that every quote is read on the basis it was struck on.
What the indexation lag decides
The lag fixes the reference RPI: the index value that goes into the numerator of the index ratio for a given payment. Under the eight-month lag, a coupon paid in a given month is scaled by the RPI for the month eight months earlier. Under the three-month lag, the same payment is scaled by the RPI for the month three months earlier. The base RPI, the denominator, is a different figure entirely: it is set once, at the gilt's first issue, and never changes.
The consequence runs in two directions, and both are worth having straight. A longer lag means the reference RPI was published further back, so the cash amount is knowable well in advance: under the eight-month lag the governing index figure is published roughly seven months before the payment, against roughly two months under the three-month lag. The same longer lag also means the cash is answering to older prices. An eight-month-lag gilt responds to inflation more slowly than a three-month-lag gilt does, in both directions — it lags when prices rise and it lags when they fall.
The lag changes which month's RPI a payment points at; it does not change the mechanism. Both kinds scale the coupon and the redemption by the index ratio, and both can pay less than a previous coupon if the index has fallen.
Two ways of quoting the same gilt
The lag also decides how the gilt is quoted, and this is where a reader meets it without necessarily knowing it. A three-month-lag gilt is quoted at a real clean price: the price of the gilt itself, with the inflation uplift left out. An eight-month-lag gilt is quoted at a nominal clean price: the same price with the uplift left in. The dirty price — the clean price plus the accrued interest — carries the uplift on both.
The practical effect is that an eight-month-lag gilt's clean price and a three-month-lag gilt's clean price are not on the same basis, even though both are printed as a number per £100 nominal. Read one as though it were the other and the error is not a rounding difference: it is the size of the index ratio, which for a gilt issued long ago is a substantial multiple of one. That is why the site deflates an eight-month-lag gilt's quoted clean price by its index ratio before showing a real yield, and takes a three-month-lag gilt's quote as already real.
The same asymmetry explains a detail that otherwise looks arbitrary. Because an eight-month-lag gilt's clean and dirty prices are both uplifted, the difference between them is not the accrued interest in real terms — the uplift sits on both sides of the subtraction. Recovering the real accrued interest means deflating the uplifted dirty price and then subtracting a real clean price, which is the route the site takes. On a three-month-lag gilt the published index ratio does that job directly.
The two styles side by side
The table sets the two styles against each other on the points a reader actually meets. It describes conventions, not a ranking: neither lag is a defect and neither is an advantage in itself.
| Question | Eight-month lag | Three-month lag |
| Reference RPI for a payment | The RPI for the month eight months before it | The RPI for the month three months before it |
| When that figure is published | Roughly seven months before the payment | Roughly two months before the payment |
| Clean price as quoted | Nominal — the uplift is inside it | Real — the uplift is not inside it |
| Dirty price as quoted | Uplifted | Uplifted |
| Accrued interest | Recovered by deflating the dirty price first | Available from the published ratio directly |
| How fast the cash responds to inflation | More slowly, in both directions | More closely, in both directions |
| Base RPI | Fixed at the gilt's first issue | Fixed at the gilt's first issue |
The last row is the one to hold on to. The lag is a difference in the numerator of the index ratio, never in the denominator, and it is a difference in the quotation convention rather than in the promise to index. Two gilts with the same base RPI and the same real coupon but different lags would still pay different cash in the same month, because they are looking at different months of the index.
What neither style does
It is tempting to describe an index-linked gilt as protecting a holding from inflation, and the description does more harm than good. What such a gilt does is scale its cash by a published index, subject to a lag. What it does not do is any of the following.
- It does not fix the cash. The Retail Prices Index can fall as well as rise, and the Debt Management Office says in terms that index-linked cash flows can therefore fall.
- It does not match any particular household's costs. The index measures a basket of prices, not one reader's spending, and a lag moves the cash further away from current prices.
- It does not stop the price moving before maturity. An index-linked gilt's price still responds to real yields, so a holding sold early can return less than it cost.
- It does not promise a positive real return. A real yield can be negative, and the coupon a gilt pays is a real rate that was fixed when it was issued.
- It does not remove the effect of the lag it carries. An eight-month-lag gilt answers to older prices than a three-month-lag one, whichever direction prices have moved.
Where to look
The index-linked gilts in issue is the page that puts the two styles on one basis. Each row carries the index ratio at five decimals and a real yield, so an eight-month-lag gilt's deflated quote and a three-month-lag gilt's real quote are read in the same column rather than in two different conventions. The clean price, the accrued interest and the dirty price sit alongside, which is what lets a reader see the uplift at work rather than take it on trust.
The same index-linked list carries each gilt's real yield after income tax at the band the page's control is set to, which is what makes gilts of different real coupons comparable once tax is in the picture. Opening a single gilt adds its projected cash flows, each with its own index ratio and each marked where it depends on an assumption rather than on a published index value.
The par RPI curve is the neighbouring view: it plots the inflation the market is pricing at each maturity rather than the inflation already recorded. Because the two lags put the cash on different timetables, the curve is where a reader goes to see what the market is charging for inflation at a given point on the term structure, separately from what any single gilt's ratio has already banked.
Background, not advice
This article explains how the two indexation lags differ and how each affects the way a gilt is quoted; it is general background, not investment, tax or legal advice for any reader's circumstances. The gilt's terms, the indexation lag and the two conventions are described by the Debt Management Office, and the Retail Prices Index is published by the Office for National Statistics, which are the primary sources. Nothing here distinguishes one style as preferable, and neither removes risk: the cash can fall with the index and the price moves before maturity.