The composition of GDP can be as important to the fiscal forecast as the headline measure, because some components will be more ‘tax rich’ than others. A forecast that alters the composition of GDP can therefore have a significant impact on the public finances, even if the path of GDP itself is unchanged.

The expenditure approach to measuring GDP, known as GDP(E), estimates the money spent on total output in the economy. Expenditure is split into household consumption, private investment, government consumption and investment, as well as exports and imports of goods and services. As with overall GDP, it can be measured in cash or nominal terms and in inflation-adjusted real or volume terms.

  • Household consumption

    Household consumption is the largest component of expenditure, representing 65 per cent of GDP. The consumption forecast is also an important determinant of the projections for VAT receipts in particular. Around 70 per cent of VAT receipts are derived from household consumption. These move largely one-for-one with changes in nominal consumer spending.

    We generate our consumption forecast in two steps:

    • to forecast the next few quarters, we use high-frequency indicators such as consumer confidence surveys. Data from UK retailers are also available on a timely basis and can provide indications of the likely near-term path for private consumption; and
    • the medium-term forecast is informed by implications for the aggregate household financial position and balance sheet. These reflect prospects for real labour income and real financial wealth, which are assumed to represent households’ current and (expected) lifetime future resources. Real labour income in this context can be measured by the real consumption wage, which is given by nominal wages (as measured by compensation of employees) deflated by consumer prices. This gives a measure of wages in terms of the amount that can be purchased with them. In general, we assume that consumption growth will be similar to growth in the real consumption wage, which implies that the proportion of real wages devoted to consumption will remain relatively stable over the forecast. Financial wealth is affected by financial markets through equity and bond prices, which will reflect expectations of future income. Changes in prices will reflect changes in expectations that will cause wealth to be revalued and may also affect behaviour now.

    Once we have produced our forecasts for real labour income and consumption, plus a variety of other sources of household disposable income, we are able to calculate the household saving ratio – in effect, the proportion of total household income that is not consumed. This is a useful diagnostic on the rest of our judgements. If the implied saving ratio is rising or falling over the forecast period, we will look again at the reasons for that and take a judgement as to whether such a path is plausible or whether we need to revise any of our other judgements. We tend to focus on a measure of the saving ratio that excludes elements of pension saving that are a feature of the headline National Accounts measure of the saving ratio. Our latest forecasts for both are shown in the chart below.

    Large but temporary shocks to the economy (such as a fiscal consolidation) may be consistent with households reducing their saving for an extended period, until incomes recover. If we were to judge that an event had a permanent effect on incomes, households would be expected to reduce their consumption plans accordingly. Of course, determining the distinction between an ‘extended but temporary’ effect and a ‘permanent’ effect is easier said than done.

    Our macroeconomic model has an equation that has been estimated to predict private consumption based on its lagged value, real household disposable income, gross physical wealth less the consumer expenditure deflator, claimant count unemployment and interest rates. That equation is used to frame our thinking, but the consumption forecast itself will reflect BRC judgements that take into account the factors described above as well as any wider factors that the BRC considers to be relevant.

    Back to top

  • Investment

    Investment is the most volatile of the main expenditure components of GDP. The data are also subject to large revisions, making it extremely challenging to forecast accurately, even at very short horizons (see Box 2.1 in our March 2017 Economic and fiscal outlook for a discussion of this feature of the data). Our forecast for investment is broken down into its three main components: business investment, residential investment (which includes some transfer costs that are counted as investment, such as estate agent fees associated with transferring the ownership of an asset) and government investment. The most important of these components is business investment, both because it is the largest of the three and because it affects the size of the productive capital stock that contributes to productivity growth.

    Back to top

  • Business investment

    The near-term outlook for business investment is informed by survey indicators and industry reports, such as the CBI Industrial trends survey, the Deloitte Chief Financial Officers survey and the Bank of England Agents’ summary of business conditions, all of which offer information on investment intentions.

    Beyond the near term, we start with an assumption that firms will invest with the aim of achieving their optimal capital stock and we make a judgement about how quickly firms will adjust their investment plans to achieve that. When considering investment prospects we also consider the prevailing conditions in financial markets and the structure of firms’ balance sheets. Corporate profitability and retained earnings will influence the ability of credit constrained firms to finance investment. We also consider other factors such as whether there are events expected during the forecast period that would raise or suppress business investment, for example by generating uncertainty that might lead investment projects to be put on hold of even cancelled.

    As with other components of GDP, we make adjustments to our forecast for business investment to reflect policy decisions that are expected to have a meaningful impact on firms’ capital expenditure. For measures that are expected to influence the cost of capital (like changes in the headline rate of corporation tax), we estimate the effect on the optimal capital stock and then determine a revised path for business investment to reach that new optimum. Other measures could affect business investment directly.

    Our macroeconomic model allows us to derive the optimal capital stock based on an estimate of the tax-adjusted cost of capital. The model’s business investment equation implies that in the short run business investment is positively influenced by current output, but negatively influenced by greater uncertainty, while in the medium-term, business investment is assumed to adjust so that the capital stock is returned to its optimal level over time. The output of the model is used to frame our thinking around business investment, but the forecast itself relies on judgements made by the BRC. Given that business investment is a volatile component of GDP and the data are also subject to significant revisions, judgement can play a particularly significant role in this case.

    Our forecast for business investment has a significant effect on projected corporation tax receipts because higher nominal business investment would directly reduce corporation tax payments as capital allowances rise.

    Back to top

  • Residential investment

    Residential investment is composed of investment in new houses, investment in improvements to existing homes and transfer costs associated with moving between homes.

    In our forecast, the first two components are grouped together as a forecast for investment in new builds and improvements. The new builds element of this series is derived from a model that informs our house price forecasts. That model contains an equation for new builds (known as ‘new starts’ in that context). We then apply a degree of top-down judgement about the prospects for the housing market and household consumption to arrive at the overall forecast for improvements. The forecast for transfer costs is closely linked to the overall outlook for property transactions.

    Back to top

  • Government investment

    Similar to government consumption, our forecast for government investment is largely determined by the Government’s nominal spending plans. In particular, the forecasts for nominal spending on gross fixed capital formation by central government and local authorities are estimated based on the Government’s latest policy decisions and any adjustments we make to reflect the extent to which we expect plans to be met in specific years. We generally assume that half of any change in nominal government investment will be reflected in real government investment and half in the government investment deflator.

    As with other parts of the forecast, if we were to judge that applying the methodology above to the Government’s fiscal plans implied an implausible path for the government investment deflator, with unwanted consequences for the GDP deflator, we could move away from the above assumption.

    Back to top

  • Government consumption

    Our forecast for government consumption is largely determined by the Treasury’s public spending plans, which we adjust to reflect our own judgements about the extent to which plans will be met. Government consumption is broadly, but not entirely, equivalent to central government expenditure on public services and administration – known as ‘Resource Departmental Expenditure Limits’ (RDELs) in the Government’s spending framework. These plans are typically set in nominal terms, so to produce a forecast for real government consumption we also need to forecast a path for the government consumption deflator. We generally assume that half of any change in nominal government consumption will be reflected in real government consumption and half in the government consumption deflator.

    Total government spending is the sum of government consumption and government investment. Our forecast for government investment is also largely driven by the Treasury’s spending plans. It is discussed further in the investment section.

    Back to top

  • Exports

    The judgements which determine our exports forecast can be split into two categories:

    • our expectations for UK export markets growth, which shows the amount that trade among the UK’s export markets is expected to grow over the forecast period; and
    • the share of UK export markets growth that we expect to be satisfied by UK exporters. In recent years, the UK’s export market share has fallen and our recent forecasts for total exports have all been based on an assumption that that trend will continue. We also adjust our forecast to reflect exchange rate movements, which would be expected to alter the UK’s export market share as exporting from the UK becomes more or less profitable relative to supplying the domestic market.

    As well as judgements in these areas, which have been common to all our forecasts, since our November 2016 forecast we have also had to make assumptions about the effects on trade of the UK leaving the EU. As there is no meaningful basis for predicting the precise end-point of the negotiations that will determine the UK’s future trading arrangements, we have made a broad-brush assumption that exports growth would be lower over a roughly ten-year horizon rather than making any specific assumptions about the rules that will be put in place after the UK leaves the EU. The size of this adjustment was calibrated to match the average revision that had been assumed in a number of external studies that considered the impact of a vote to leave the EU on the UK economy, including trade. We made a similar revision to imports growth to reflect the adjustment to a new trading regime, so the downward revisions to gross trade flows were broadly neutral in their effect on GDP growth. This contributes to our overall assumptions that the UK export market share will continue to fall over the forecast horizon.

    Our forecast for exports is published at an aggregate level but when constructing the forecast we also consider the paths for goods, services and oil exports. We produce a forecast for oil output because the production of this sector affects our North Sea revenues forecast. Oil exports are then estimated as a proportion of that forecast for oil output. Given that we take a top-down approach to forecasting total exports as described above and that the forecast for oil exports is derived from our oil output forecast, the remaining non-oil element of exports growth is derived by residual and apportioned across goods and services.

    Back to top

  • Imports

    In general, our forecast for imports growth is based on the expected path of domestic demand, where we assume that a proportion of that demand will be met by imported goods and services. A key diagnostic for our imports forecast is therefore the implied import intensity – the ratio between total imports and import-weighted domestic demand – over the forecast period. The judgements that determine our imports forecast can be split into two categories:

    • our expectation for import-weighted domestic demand, which is calculated by weighting the expenditure components of domestic demand according to their respective import contents. These weights are derived from detailed ONS data that estimate the inputs and outputs of all sectors of the economy. The forecast for import-weighted domestic demand is therefore derived mechanically from our forecasts for the other components of GDP; and
    • the likely path for the import intensity of that demand over the forecast period. This could be affected by factors such as exchange rate movements, which will affect the relative prices of domestic goods and services compared with imported alternatives. Domestic demand in the UK has become more import intensive over time, consistent with world trade having grown more quickly than world GDP over a number of years, although that process has been less evident more recently.

    Many of our recent forecasts for total imports have been based on an assumption that the import intensity of domestic demand would continue to rise, but following the vote to leave the EU, our November 2016 forecast assumed a temporary, but protracted, reversal of that trend. We retained this assumption in our March 2017 forecast. The size of the adjustment to import intensity was calibrated in the same way as the export market share adjustment described in the exports section.  Since the revisions to imports and exports growth were similar, the downward revisions to gross trade flows were broadly neutral in their effect on GDP growth.

    As with exports, the forecast for imports is published at an aggregated level but when constructing the forecast we also consider the paths for goods, services and oil imports. Oil imports are obtained as the residual between total demand (domestic oil demand and exports) and domestic supply (North Sea production). Again, stripping out this forecast for oil imports from total imports that are forecast using the top-down approach described above, the remaining non-oil imports growth is derived by residual and apportioned across goods and services imports.

    Back to top

  • Stocks

    Stocks, or ‘inventories’, are goods held by firms in reserve, rather than being immediately sold or used in the production process. These could include raw materials that are awaiting use by firms, or unsold goods that are held by wholesalers or retailers. The stock of inventories is an equivalent concept to the capital stock, so changes in inventories (stockbuilding) affect GDP in the same way that changes in the capital stock – i.e. business investment – affect GDP. An increase in the stock of inventories implies that firms have purchased items and this expenditure is therefore recorded as an expenditure component of GDP.

    We look at the most recent outturn data to make a judgement about the most likely path of stockbuilding in the near term. Over the rest of the forecast period, we generally assume that inventories make no contribution to GDP growth, unless we expect some temporary macroeconomic conditions to cause firms to either build up or reduce their inventories in an unusual way.

    Back to top