Some Forecasts in-depth pages have not been updated for the latest forecast. We will endeavour to update them as soon as possible.

The Treasury manages public spending within two ‘control totals’:

  • departmental expenditure limits (DELs) currently account for around two-fifths of spending – mostly covering spending on public services, grants and administration (collectively termed ‘resource’ spending) and investment (‘capital’ spending). These are items that can be planned over extended periods.
  • annually managed expenditure (AME) currently accounts for around three-fifths of spending – categories of spending less amenable to multi-year planning, such as social security spending and debt interest.

Debt interest spending is one of the bigger items in AME. Most public sector debt interest spending is accounted for by central government. The biggest components are interest paid on government bonds (known as ‘gilts’, of which there are two types: conventional gilts that pay a fixed amount of interest and index-linked gilts that pay an interest rate linked to RPI inflation), to holders of NS&I savings products and on the reserves (in effect electronic money) created by the Bank of England for monetary policy purposes. The Bank holds lots of gilts (mostly held in the Asset Purchase Facility (APF)) and is part of the public sector, so the interest paid by central government on those gilts does not actually leave the public sector. As we outlined in our explainer on the direct fiscal consequences of unconventional monetary policy, the net effect on public sector debt interest of the APF holding gilts is the difference between the (larger) coupon interest earned on the gilts and the (smaller) interest paid on the reserves created to finance the purchase of the gilts – thereby reducing total debt interest spending. We also include estimates of the income from the APF’s holdings of corporate bonds based on the average of eligible bonds.

In our latest forecast, we expect public sector debt interest in 2019-20 to total £41.6 billion (reflecting £52.4 billion of gross debt interest payments and £10.9 billion netted off due to the APF). That would represent 4.9 per cent of total public spending or 1.9 per cent of national income.

  Latest forecast

Our latest fiscal forecast was published in March 2019. Public sector debt interest payments are expected to only decline slightly as a per cent of GDP over the forecast period as the impact of net debt falling as a share of GDP in most years is offset somewhat by projected increases in interest rates.


More detail on our latest forecast and how it was revised relative to our previous forecast in October 2018 was provided in paragraphs 4.106 to 4.108 of our March 2019 EFO.

Expand to read the extract from our March 2019 EFO

Public sector debt interest payments are forecast by applying appropriate interest rates to the corresponding stocks of conventional and index-linked gilts outstanding at different maturities and other debt, such as NS&I products and Treasury bills. Financial market expectations are used to derive relevant interest rates (for example, coupons on newly issued conventional gilts), while our inflation forecast is used for index-linked gilts and other index-linked debt.[1] Flows associated with the Bank of England’s Asset Purchase Facility (APF) similarly apply appropriate market-derived interest rates to the stocks of the APF’s loan liability and to its gilt, corporate bond and loan assets.

Public sector debt interest payments are expected to only decline slightly as a per cent of GDP over the forecast period as the impact of net debt falling as a share of GDP in most years is offset somewhat by projected increases in interest rates.

The table below shows that we have revised our forecast down since October. This reflects:

  • The downward revision to our forecast for RPI inflation reduces spending by substantial amounts in 2018-19 and 2020-21 and more modestly in other years.
  • Market-derived gilt rate expectations have fallen, reducing the costs of financing newly issued debt. This effect builds up over the forecast as more new debt is issued.

Market-derived Bank Rate expectations have also fallen since October, increasing the debt interest saving associated with the APF. In October, Bank Rate expectations moved above 1.5 per cent in 2023-24, so we assumed (in line with MPC guidance) that the Bank would begin to run down the APF’s assets at that point, reducing the associated debt interest saving. But as Bank Rate no longer exceeds 1.5 per cent within the forecast period, we do not assume any asset reductions in this forecast.

 Key changes to central government debt interest since October


[1] Our forecasting approach was explained in Box 4.4 of our March 2015 EFO. We publish a supplementary fiscal table on our website that presents the different stocks, flows and effective interest rates that make up our debt interest forecast.

  Back to top

  Latest monthly data

Public sector gross debt interest: ONS (ID: JW2P)

Asset Purchase Facility debt interest netted off: ONS (ID: MU74)

Debt interest spending is volatile on a monthly basis. While there is a pattern that repeats through the year – e.g. with lower spending in March and higher spending in April and October – the accrued interest on index-linked gilts is affected by the monthly path of the Retail Prices Index (RPI), which itself is uneven from month to month due to things like oil price movements affecting petrol and diesel prices. This means that care must be taken when interpreting year-on-year changes in monthly debt interest spending. The focus of our forecast is the full-year position, with in-year data used to inform the starting point for the forecast.

  Back to top

  Forecast methodology

Forecast process

The Treasury coordinates the production of our debt interest forecast based on our assumptions about interest rates and forecast for RPI inflation. We discuss these forecasts at challenge meetings where we scrutinise the latest data and the effects of our assumptions and judgements.

The mechanics of the debt interest forecast are somewhat different to other AME forecasts because of the interactions between the central government net cash requirement (CGNCR) – which we derive from our full fiscal forecast – and our central government gross debt interest forecast. This means that we need to iterate the forecast until the CGNCR and the debt interest forecast converge on their final values. This is known as the ‘debt interest loop’ and is run in the OBR at the very end of our fiscal forecast process. The following diagram summarises this process:



Forecasting model

Public sector debt interest payments are forecast by applying interest rates to the stocks of different liabilities.

Our debt interest forecast process starts with estimates of the stock of different types of debt on which government must pay some form of debt interest. These include:

  • conventional gilts;
  • index-linked gilts;
  • the assets and liabilities of the APF; and
  • other financing products (such as Treasury bills and NS&I products).

The largest component of debt interest is the interest paid on conventional gilts. For the existing stock of conventional gilts, payments are known and fixed. By contrast, estimates for interest payments on existing index-linked gilts and for all new debt issuance can be subject to larger revisions due to changes in our economic and fiscal forecasts (including changes in RPI inflation, interest rates and the net cash requirement – a measure of the deficit).

We forecast changes to the various stocks of debt by considering the gilts and other liabilities that will become due to be redeemed, and the additional debt issuance that will be required to cover the net cash requirement for each year, plus the redemptions of the existing stock.

We then have to make assumptions about the composition of gross financing each year to cover the cash deficit and redemptions, and the interest rates that will apply to the new debt issuance. These interest rates are derived from financial market expectations and our RPI inflation forecast (for index-linked gilts).

The debt interest forecast is expressed as net of the effect of gilts held by the APF. The APF receives coupon income on the gilts it holds, while the Bank pays Bank Rate on the reserves it created to finance the asset purchases made by the APF. The coupon payments cancel out within the public sector, so this debt is in effect financed at Bank Rate. Consistent with statements from the Bank, we assume that gilts held by the APF will not be sold actively during the forecast period, and will only be run down through redemptions once Bank Rate rises to at least 2 per cent (which also was not expected to happen during the period covered by our latest forecast).

Our central government debt interest forecast includes interest payments made by UK Asset Resolution (UKAR) and Network Rail, which are both classified as parts of central government by the ONS, as well as other smaller payments, such as interest on finance leases.

Box 4.4 of our March 2015 Economic and fiscal outlook contains further explanation of how our central government debt interest forecast is constructed.

Debt interest spending by local authorities and public corporations, including the non-APF parts of the Bank of England, are produced in a similar way by applying appropriate interest rates to projected stocks of debt liabilities.

Main forecast determinants

The main determinants driving our debt interest forecast are:

  • Our forecast for the Retail Prices Index that determines the accrued debt interest payments on index-linked gilts.
  • Our market-derived assumptions for conventional and index-linked gilts rates that are used to forecast debt interest payments on conventional and index-linked gilts. We use an average conventional gilt rate obtained combining the yield on a short, a medium and a long-term government bond with the proportions of each in total issuance consistent with recent outturns and government debt management policy. Our real index-linked gilt rate is from a long dated index linked gilt.
  • Market expectations for Bank Rate that are used to forecast the debt interest payments on the APF’s loan.
  • Our market-derived assumption for short-term market interest rates that is used to project forward payments on Treasury bills and other short-term debt instruments. This is typically very close to Bank Rate.

The forecast is also driven by the central government net cash requirement (CGNCR) that determines the net amount of new debt that must be issued on which additional interest payments must be paid. This is added to the new debt issuance associated with the known amounts of gilts that are due to be redeemed to get to gross debt issuance.

Main forecast judgements

Much of our debt interest forecast reflects the consequences of judgements we make about other parts of the forecast – about inflation, what source to use for interest rate assumptions, and the overall fiscal forecast that drives the cash requirement. The remaining judgements that are specific to our debt interest forecast are:

  • the proportions of different types of gilts that will be issued and whether this is skewed towards shorter- or longer-term debt, index-linked gilts and so on; and
  • the size and composition of the gilts and other assets held in the APF and whether or how they will be replaced when gilts reach maturity and are redeemed.

  Back to top

  Previous forecasts

Debt interest payments have been significantly lower than expected across our forecasts, much of which can be explained by assumptions about the key underlying determinants. Interest rates – both short-term rates and longer-term gilt rates – have been lower than market expectations (on which we base our assumptions) at the time of almost every forecast. Lower RPI inflation over the recent past has also contributed to the errors by reducing the effective rate on index-linked gilts. We have also changed our assumptions on the behaviour of the APF over time as Bank of England’s guidance about how the APF will be operated has evolved.

Other changes to our forecasts include:

  • The composition of gilts issued, which was more initially skewed towards relatively cheaper short-term debt and index-linked gilts than we had assumed in some earlier forecasts, leading to lower spending. (However, since March 2018, the Government has begun to reduce the proportion of index-linked gilts in its financing plans.)
  • Revising our assumption that the split of issuance would converge towards historical patterns to one where we now assume that it remains in line with the latest year’s financing remit.
  • A correction to the way we predicted the stock of debt, which related to modelling the refinancing of gilts at redemption and had caused us to over-estimate the stock. We corrected this in our December 2014 forecast.

The ONS treatment of the APF was changed between our forecasts in March and December 2014, which means our earlier forecasts are not directly comparable with those since the change. Originally its effects were excluded completely, in line with the approach at the time of excluding the effects of things judged to be related to the financial crisis. It is now included, because only the effects of the public sector banks (e.g. the Royal Bank of Scotland) are excluded from the headline public sector finances statistics.

  Back to top

  Policy measures

All policy measures indirectly affect debt interest spending because they change the net cash requirement. But since our first forecast in June 2010, governments have also announced seven policy measures that have directly affected our forecast for central government debt interest payments, all of which have been relatively small. Recent examples include changes to the Government’s financing remit.

The original costings for these measures are contained in our policy measures database and were described briefly in the Treasury’s relevant Policy costings document. For measures announced since December 2014, the uncertainty ranking that we assigned to each is set out in a separate database. For those deemed ‘high’ or ‘very high’ uncertainty, the rationale for that ranking was set out in Annex A of the relevant Economic and fiscal outlook.

  Back to top

  Ready reckoners

‘Ready reckoners’ show how our fiscal forecasts could be affected by changes in selected economic determinants. They are stylised quantifications that reflect the typical impact of changes in economic variables on receipts and spending. These estimates are specific to our March 2018 forecast and we would expect them to become outdated over time, as the economy and public finances, and the policy setting, continue to evolve. They are subject to uncertainty because they are based on models that draw on historical relationships or simulations of policy settings. They also do not necessarily capture all the impact of a change; for example, the ready reckoner for gilt rates does not include possible interactions with the APF. The table below shows that:

  • a 1 percentage point increase in gilt rates in the base year (2017-18) increases debt interest payments by rising amounts over the forecast period, reaching over £4 billion by the end of the forecast. This increases steadily over time as the rate increase only affects new gilt issuance;
  • a 1 percentage point increase in short-term interest rates in the base year has a sizeable impact in the same year, with the impact remaining fairly stable thereafter. Unlike the gilt rate effect, short rate changes are almost instant and affect the entire stock of relevant debt;
  • a 1 percentage point increase in RPI inflation in the base year has a sizeable impact in the same year as it instantly impacts on accrued payments on index-linked gilts, before increasing debt interest payments by rising amounts over the forecast period; and
  • a £5 billion increase in the central government net cash requirement in the base year has a relatively small impact on the forecast, pushing debt interest spending up by only £0.4 billion by the final year of the forecast. This small impact is due to low interest rates.

  Back to top

Other expediture