The Treasury manages public spending within two ‘control totals’ of about equal size:

  • departmental expenditure limits (DELs) – 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) – categories of spending less amenable to multi-year planning, such as social security spending and debt interest.

Capital spending by public corporations (PCs) is included within AME, and reflects PCs’ gross capital expenditure as recorded in the National Accounts. This includes all capital spending by PCs, regardless of whether it is financed by PCs’ own funds (e.g. retained earnings or borrowing), or is financed by grants or loans from central or local government, but net of asset sales. Our forecasts for central and local government capital spending therefore exclude amounts spent on grants to PCs, in order to avoid double-counting.

As spending by housing associations (HAs) makes up a significant proportion of total PC capital expenditure, and because the HAs forecasting approach is quite different, we forecast HAs and non-HAs capital separately. Our HAs forecast process is more involved than that for most PCs.

This page also details our receipts forecast for PCs’ gross operating surplus (PC GOS) – a National Accounts measure of PCs’ profits before taxation, interest payments or depreciation. PC GOS is included within our current receipts forecast as part of the public sector GOS forecast. The general government element of GOS is equal to general government depreciation, consistent with its measurement in the National Accounts.

In our latest forecast, we expect PC capital spending (excluding HAs) in 2017-18 to amount to £9.2 billion. That would represent 1.2 per cent of total public spending, and is equivalent to £330 per household and 0.5 per cent of national income. We also expect PC GOS (excluding HAs) in 2017-18 to amount to £10.4 billion. That would represent 1.4 per cent of all receipts, and is equivalent to £370 per household and 0.5 per cent of national income.


  • Latest forecast

    Our latest fiscal forecast was published in November 2017. Capital spending by PCs (excluding HAs), was £8.8 billion in 2016-17 and is forecast to increase to £9.0 billion by 2022-23, although the profile is fairly flat over the forecast period (from 2017-18). As the cash values over the forecast period are fairly flat, capital spending as a share of GDP is falling. PC GOS (excluding HAs) was £10.2 billion in 2016-17 and is forecast to increase to £11.1 billion by 2022-23. As a share of GDP, PC GOS falls slightly over the forecast period, as it is forecast to grow at a slower pace than GDP.

    More detail on our latest forecast for public corporations’ capital expenditure and how it was revised relative to our previous forecast in March was provided in paragraphs 4.153 to 4.155 (included alongside the commentary on locally financed capital spending) of our November 2017 EFO. No commentary on PC GOS appeared in the EFO as the relevant section (paragraph 4.96) focused on GOS at the whole public sector level. We have included a short note on PC GOS below.

    Expand to read the extract from our March 2017 EFO

    Back to top

  • Forecast methodology

    Forecast process

    We construct our forecast for PCs’ capital spending and gross operating surplus (GOS) based on the latest outturn data, which are provided to us in detail for each individual PC, under a restrictive data access agreement from the Office for National Statistics (ONS). The data for individual PCs are disclosive, which means that we can only publish outturns and forecasts at an aggregate level.

    The largest elements of the capital spending and gross operating surplus forecasts are for housing associations (HAs), the housing revenue account (HRA) and Transport for London (TfL). The HAs forecast is produced for us by the Department for Communities and Local Government (DCLG) and is scrutinised by the OBR (more detail will be made available in due course on a separate page for the HAs forecast). In the case of the HRA, DCLG produces the forecast for the elements of locally financed capital expenditure (capital LASFE) that are financed from HRA sources. DCLG also forecasts HRA GOS and HRA asset sales. These DCLG forecasts are then scrutinised by the OBR, as part of the process for producing the forecast for capital LASFE. The TfL forecast is based on its latest published business plan. TfL assist us in scrutinising these plans before we produce our own central forecasts for TfL capital spending. For instance, we may include additional timing adjustments on TfL’s capital spending if we feel it would be central to assume some slippage, consistent with the general tendency in spending on large capital projects.

    The remaining, smaller elements of the forecast are produced by us in-house. For different elements of the forecast, different methodologies are used (discussed in more detail in the main forecast judgements section below). These assumptions are reviewed each year to ensure they remain reasonable, although capital spending by public corporations – as with capital spending more generally – is difficult to forecast given volatility in the year-to-year spending profiles.


    Forecasting models

    A number of elements of this forecast are driven by assumptions or business plans. The largest elements that are model-driven (excluding HAs) are:

    • HRA spending and the imputed subsidy for equity injection by local authorities into the HRA (which is neutral for the public finances, being offset in GOS and National Accounts adjustments). The HRA spending model is owned and operated by DCLG and is largely driven by the latest outturn data, our economy forecast and any policy measures determining rents received and future house building programmes. We estimate the imputed subsidy in-house. The forecast is driven by the HRA spending model and a series of simple assumptions about how outturn will evolve for the other items that determine the subsidy.
    • HRA asset sales (which nets off HRA gross capital spending). The model is split into proceeds from Right to Buy (RtB) and non-RtB sales. The model for RtB is again provided by DCLG, with the forecast driven by similar factors to the HRA model. The non-RtB forecast is driven by outturn and a slightly narrower set of determinants.

    Main forecast determinants

    For the HRA forecast, the main economic determinants are GDP growthhouse prices and rental values, which affect income to and spending from the HRA via demand for social housing and the levels of rental income received by the operating local authorities. (The HRA is regarded as a public corporation, meaning this is where the final receipts and spending score in the National Accounts). The main components making up the HRA forecast – and the economic determinants driving them – are as follows:

    • HRA GOS: this forecast is driven by the growth in the council rented housing stock, nominal GDP and social sector rents (with these rents being driven by CPI inflation from 2020).
    • Major repairs reserve (MRR): this forecast is driven by the growth in the council rented housing stock and nominal GDP.
    • HRA capital expenditure from revenue account (CERA): this forecast is driven by past trends, which are used in adjusting the jumping-off point. The forecast rises in line with our forecast for HRA GOS.

    For the asset sales forecast, the main economic determinants are those that affect the demand for housing purchases via the price of the property and the purchasing power of the eligible population. They are:

    Also reflected is the effective level of discount offered on the purchase of social properties by existing tenants.

    Meanwhile, the non-RtB elements are driven by outturns and our forecasts for property transactions and house prices.

    The most common assumptions used in other forecasts for PCs are to start from the latest outturn data on in-year estimates and grow the forecast in line with:

    We utilise quarterly in-year information to test whether these assumptions seem to be producing a sensible view for the current year, and whether they therefore constitute reasonable assumptions for later forecast years. (These data are however subject to limitations and potentially subject to significant revision by the time quarterly outturns are finalised.)


    Main forecast judgements

    Capital spending can be subject to significant year-to-year fluctuations. Since we typically do not know in advance whether such fluctuations are likely to be upward or downward, our forecasts tend to assume the year-to-year pattern is fairly smooth (albeit with occasional step changes when the timing and size of one-off events can be predicted with sufficient confidence). Other year-to-year fluctuations that materialise but were not forecast often reflect unforeseen, and sometimes unforeseeable, events or timing effects. Our central forecasts are largely built on simple judgements about the evolution of certain spending and GOS items, in the knowledge that outturns will typically show much greater year-to-year volatility – in both directions – than our forecasts. (At the latest forecast, around a quarter of total PC GOS in 2016-17 came from forecast elements that are solely assumption-driven.)

    The main forms of assumption we use include:

    • grow in line with recent trends;
    • grow in line with related items (for example, some HRA spending items increasing in line with projected HRA GOS and the council rented housing stock);
    • grow in line with latest business plans;
    • remain flat in cash terms;
    • grow in line with nominal GDP; or
    • grow in line with CPI inflation (thereby remaining flat in real terms).

    A judgement is taken using recent data on which of the above methods seems most appropriate for each individual public corporation capital spending or GOS line that is forecast in this manner. These are reviewed ahead of each forecast.

    Our asset sales forecast assumes recent trends in housing sales serve as a reasonable proxy for future years, although allowance is made for the effect of future interest rates, house price inflation and the general state of the economy (all of which contribute to demand).

    Back to top

  • Previous forecasts

    For public corporations’ (PC) capital expenditure, the most significant source of forecast-to-forecast revision is outturn data, which is used as the jump-off point for forecast years. For example, in the chart below, spending was revised up between our November 2010 and March 2011 forecasts, but then revised down again in November 2011. Some of these revisions come from changes in the composition of the PC sector – for instance, when the Office for National Statistics (ONS) reclassified British Energy and Royal Mail to the private sector (in 2002 and 2013 respectively), following the Government’s sales of shares in those entities. Such changes are explained the relevant EFO. Forecast-to-forecast revisions have generally been smaller between more recent forecasts – but the level and timing of capital spending remains subject to significant uncertainty. Final outturn financial year data are not available until some time after the year’s end. This means potentially significant revisions between forecasts can occur, as outturn data are revised well after the financial year’s end. Most forecasts exhibit a fairly consistent trend across forecast years: the projections being fairly flat in cash terms.

    For PCs’ gross operating surplus (PC GOS), the most significant source of revision is again outturn data – the most notable example was in our December 2013 forecast, where revisions reflected the latest ONS Blue Book (2013) estimates. All forecasts exhibit a fairly consistent trend across forecast years: the projections rising steadily in cash terms. This upward trend is generally driven by profits being expected to rise in cash terms due to inflation. Further, some forecast elements are solely assumption-driven and assumed to grow in line with either GDP or remain flat in real terms (growing in line with CPI inflation).

    Back to top

  • Policy measures

    Since our first forecast in June 2010, governments have announced 21 policy measures affecting our forecast for public corporations’ capital spending and three affecting our forecast for public corporations’ gross operating surplus. (Some of these policy changes, at least in part, cut across our forecasts for housing associations.) 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.

    For capital expenditure, key themes have included policy changes in respect of public corporations’ capital spending financed by:

    • the Housing Revenue Account (HRA);
    • social sector rents; and
    • asset sales (including Right to Buy).

    Changes in any of the above affect the ability of public corporations to self-finance capital expenditure. Other sources of income – i.e. grants from central and local government – are not self-financed elements of spending, but changes in these grants also affect the spending power of public corporations.

    For gross operating surplus, the changes have all been small, with no single-year estimate for the impact of any of the three policies exceeding £0.1 billion. The policy changes all pertain to estimated income from either asset sales or social sector rents.

    Back to top

Take our survey