Foreword

This Economic and fiscal outlook (EFO) sets out our central forecast for the five years to 2027-28, taking account of recent data and government policies announced up to and including the Spring Budget 2023. The forecasts presented in this document represent our collective view as the three independent members of the OBR’s Budget Responsibility Committee (BRC). We take full responsibility for the judgements that underpin them and for the conclusions we have reached.

As always, we have been greatly supported in our work by the staff of the OBR. We are very grateful for their hard work and expertise. We have also drawn on the work and expertise of numerous officials across government in preparing these forecasts, including in HM Treasury, HM Revenue and Customs, the Department for Work and Pensions, the Department for Levelling Up, Housing and Communities, the Department for Education, the Department for Energy Security and Net Zero, the Ministry of Justice, the Home Office, the Department for Transport, the Department of Health and Social Care, the North Sea Transition Authority, the Office for National Statistics, the UK Debt Management Office, the British Business Bank, the BBC, Homes England, UK Government Investments, the Government Actuary’s Department, the Insolvency Service, the Scottish Government, the Scottish Fiscal Commission, the Welsh Government, the Department for Communities and the Department of Finance in Northern Ireland, Transport for London, and various public service pension schemes. We also held helpful discussions with the Chief Medical Officer and various departmental finance directors. We are grateful for their engagement, expertise, and insights.

Outside government we have held useful discussions with the Bank of England, the Confederation of British Industry, the National Institute of Economic and Social Research, the Institute for Fiscal Studies, the Resolution Foundation, the Institute for Government, the International Monetary Fund, the Health Foundation, Alex Tuckett from the CRU group, and Tony Wilson from the Institute for Employment Studies.

The date for the Budget and this forecast was announced on 19 December, well in advance of the required ten weeks’ notice. This met the agreed process for Budgets and other fiscal events, as outlined in the Memorandum of understanding between the Office for Budget Responsibility, HM Treasury, the Department for Work and Pensions and HM Revenue and Customs (MoU).

We published the timetable of the key stages of the forecast on 19 January, once it had been agreed by signatories of the MoU. That timetable was adhered to at each stage for this EFO and proceeded as follows:

• OBR staff prepared an initial economy forecast, drawing on data released since our previous forecast in November and incorporating our preliminary judgements on the outlook for the economy. This first economy forecast was sent to the Chancellor on 19 January.
• Using the economic determinants from this forecast (such as the components of nominal income and spending, unemployment, inflation, and interest rates), we commissioned updated forecasts from the relevant government departments for the various tax and spending items that in aggregate determine the position of the public finances. We discussed these in detail with the officials producing them, which allowed us to investigate proposed changes in forecasting methodology and to assess the significance of recent tax and spending outturn data. In many cases the BRC requested changes to methodology and/or the interpretation of recent data. This first fiscal forecast was sent to the Chancellor on 1 February.
• As the process continued, we identified further key judgements that we would need to generate our full economy forecast. Where we thought it would be helpful, we commissioned analysis from the relevant teams in the Treasury to inform our views and discussed forecast issues, though not specific judgements, with experts from external organisations. The BRC then agreed further judgements, allowing the production by OBR staff of a second economy forecast, which was sent to the Chancellor on 6 February.
• This second economy forecast provided the basis for the next round of fiscal forecasts. Discussions with HMRC, DWP and other departments gave us the opportunity to follow up our requests for further analysis, methodological changes, and alternative judgements from the previous round. We sent our second fiscal forecast to the Chancellor on 15 February.
• In parallel, as the Budget involved both demand-side and supply-side policy measures, we undertook an intensive process of engagement and analysis to inform the judgements we would need to take about the impact of new policy measures on demand and potential output. This involved several rounds of engagement with the Treasury and other departments as both the specification of policy packages and our assessment of their impact were refined. This was a more resource-intensive process than normal, but allowed us to consider a broad range of evidence across several policy areas, ensuring that the figures included in our forecast would be based on the best possible evidence, and that we could test our interpretation of it.
• We also scrutinised the costing of individual tax and spending measures announced since the November 2022 forecast. As usual, the BRC requested further information and/or changes to almost all the draft costings prepared by departments. We have certified all policy measures in the forecast as reasonable and central.
• We then produced a third and final pre-measures economy forecast, in which we took on the latest data and incorporated judgements embodied in our fiscal forecast. This economy forecast included energy and financial market data based on the average over the five working days to 8 February and was sent to the Treasury on 20 February. The associated fiscal forecast was sent to the Chancellor on 24 February.
• Alongside the development of the final economy forecast we made an initial assessment of the economic and fiscal effects of the emerging policy package. This built on earlier analysis that allowed us to factor in an initial package of measures that was provided by the Treasury on 27 February. We incorporated this package into a preliminary post-measures forecast, in order to provide an early view on the effect of Budget measures on the economy and public finances, which we sent to the Chancellor on 1 March. This forecast round was produced using our internal ready-reckoner models (rather than being sent to departmental forecasters). We are grateful for the input of Treasury officials who have helped develop these models, which have performed well as rapid indicators of changes between fully modelled forecast rounds.
• In line with the agreed timetable, on 3 March the Treasury provided the final package of measures that would cause movements in our economy forecast. We sent the resulting final economy forecast to the Treasury on 7 March and a near-final fiscal forecast on 8 March. All final policy decisions were provided by the Treasury on 9 March and our forecast was then finalised on 10 March and sent to the Treasury on the same day.
• The Treasury made a written request, as provided for in the MoU between us, that we provide the Chancellor and an agreed list of his special advisers and officials with a near-final draft of the EFO on 10 March. This allowed the Treasury to prepare the Chancellor’s statement. We also provided pre-release access to the full and final EFO on 13 March.

During the forecasting period, the BRC held nearly 40 scrutiny and challenge meetings with officials from other departments, in addition to numerous further meetings at staff level and those with external stakeholders. We have been provided with all the information and analysis that we requested and have come under no pressure from Ministers, advisers or officials to change any of our conclusions as the forecast has progressed. The BRC also met with the Chancellor four times to discuss the forecast over the course of its production (on 25 January, 7 and 20 February, and 8 March). A full log of our substantive contact with Ministers, their offices and special advisers can be found on our website. This includes the list of special advisers and officials who received the near-final draft of the EFO on 10 March.

Our non-executive members, Sir Christopher Kelly and Bronwyn Curtis OBE, provide additional assurance over how we engage with the Treasury and other departments. This includes reviewing any correspondence that OBR staff feel either breaches the MoU requirement that it be confined to factual comments only, or could be construed as doing so. That review takes places as soon as practicable after each EFO has been published. Any concerns our non-executive members have will be raised with the Treasury’s Permanent Secretary or the Treasury Select Committee, if they deem that appropriate.

We would be pleased to receive feedback on any aspect of the content or presentation of our analysis. This can be sent to [email protected].

The Budget Responsibility Committee

Richard Hughes, Professor David Miles CBE and Andy King

Chapter: 1 Executive summary

1.1 The economic and fiscal outlook has brightened somewhat since our previous forecast in November. The near-term economic downturn is set to be shorter and shallower; medium-term output to be higher; and the budget deficit and public debt to be lower. But this reverses only part of the costs of the energy crisis, which are being felt on top of larger costs from the pandemic. And persistent supply-side challenges continue to weigh on future growth prospects. Against this backdrop, the Chancellor has spent two-thirds of the improvement in the fiscal outlook on his Budget measures, providing more support with energy bills and business investment in the near term, while boosting labour supply in the medium term. This lowers inflation this year and, more significantly, sustainably raises employment and output in the medium term. But it leaves debt falling by only the narrowest of margins in five years’ time.

Economic outlook

1.2 Developments since our November forecast have been largely positive, but the economy still faces significant structural challenges. Wholesale gas prices have more than halved over the past six months and are expected to fall further over the forecast. At the time we closed our forecast, Bank Rate was expected to peak at 4¼ per cent later this year, rather than the 5 per cent we assumed in November. The economy narrowly avoided contracting in the final quarter of 2022 and the near-term outlook for demand has improved. But gas prices remain more than twice their pre-pandemic level which, when added to the stagnation in business investment since 2016, the recent rise in labour market inactivity, and the slowdown in productivity growth since the financial crisis, means that there remains weak underlying momentum.

1.3 CPI inflation peaked at 11.1 per cent in October and is expected to fall sharply to 2.9 per cent by the end of 2023, a more rapid decline than we expected in November. The drop in wholesale gas prices also means that household energy bills are expected to fall below the energy price guarantee limit from July and to £2,200 by the end of the year. Stronger domestically generated inflation means that inflation oscillates around zero in the middle of the decade rather than falling meaningfully into negative territory as we forecast in November. Inflation returns to target in early 2028, with the offsetting effects of lower gas prices and increased domestically generated inflation leaving the consumer price level at the end of our forecast little changed from November.

Chart 1.1: CPI inflation

1.4 Stronger real wage growth, alongside the drop in interest rate expectations since November, results in a shorter and much shallower economic downturn this year. GDP is expected to contract by 0.4 per cent in the first quarter of 2023 to 0.6 per cent below its recent peak in the second quarter of 2022. Output then flatlines in the second quarter and starts rising again from the third quarter. This means the peak-to-trough fall in GDP is just a quarter of the 2.1 per cent fall assumed in our November forecast and output regains its pre-pandemic peak in the middle of 2024, six months earlier than expected in November. Supported by the fiscal loosening in this Budget, GDP growth gathers pace to reach 2.5 per cent in the middle of the decade. GDP growth then eases back towards its medium-term potential growth rate of 1¾ per cent by the end of the forecast.

1.5 That medium-term potential growth rate is unchanged from November but the level of potential output (and actual GDP) at the end of our forecast is around ½ per cent higher. This reflects modest upward revisions due to higher migration, lower energy prices, and an increase in labour supply from Budget measures. But these are partly offset by a weaker path for the amount of capital available per worker and a lower pre-Budget-measures path for labour market participation.

Chart 1.2: Real GDP

1.6 Despite more positive economic news since November, structural weaknesses remain that have been exacerbated by recent shocks:

• Business investment has stagnated since 2016, with uncertainty surrounding the UK’s future trading relationship with the EU, the pandemic, the energy crisis, and rises in the post-tax cost of capital all weighing on investment decisions since then.
• Labour market participation, having risen since 2010 (despite the ageing of the population), has fallen dramatically in the wake of the pandemic, especially among older workers. This has left the total labour force 520,000 people smaller than we expected prior to the pandemic. And population ageing continues to weigh on participation over the forecast period.
• Productivity has grown at less than half its pre-financial crisis rate since 2010 and has been disrupted more recently by the pandemic and higher cost of energy. This reflects both the stagnation in business investment and weak growth in total factor productivity.

1.7 In terms of the labour market, we have revised up employment since November, leaving it 140,000 (0.4 per cent) higher at the forecast horizon. This is due to: (i) higher net migration adding 160,000 and (ii) the labour supply measures in this Budget adding 110,000, which more than offset (iii) a lower pre-Budget-measures participation rate subtracting 130,000. But as output growth slows, the unemployment rate rises from 3.7 per cent at the end of 2022 to a peak of 4.4 per cent in 2024, before falling back to its unchanged structural rate of 4.1 per cent by the end of the forecast. The participation rate initially falls to a 23-year low of 62.8 per cent in 2024, as demographics weigh on participation, before rising to 63.0 per cent by 2027, due to the increase in the State Pension age by 2028 and the measures announced in the Budget.

Chart 1.3: Employment

1.8 Real household disposable income (RHDI) per person – a measure of real living standards – is expected to fall by a cumulative 5.7 per cent over the two financial years 2022-23 and 2023-24. While this is 1.4 percentage points less than forecast in November, it would still be the largest two-year fall since records began in 1956-57. The fall in RHDI per person mainly reflects the rise in the price of energy and other tradeable goods of which the UK is a net importer, resulting in inflation being above nominal wage growth. This means that real living standards are still 0.4 per cent lower than their pre-pandemic levels in 2027-28. But they are 0.6 per cent higher than we forecast in November thanks to lower market expectations for medium-term gas prices and the upward revision to potential output.

1.9 The squeeze on real household incomes drags down consumption this year, despite a fall in the saving rate from its historic high during the pandemic. Consumption falls by 0.8 per cent in 2023 and then grows by 1.7 per cent on average over the rest of the forecast as real incomes recover. Cumulative business investment over the forecast period is little changed from November, but the profile is volatile due to its highly pro-cyclical nature and the impact of policy measures. Investment falls this year as output stagnates but rises strongly over the subsequent two years as firms bring forward investment to take advantage of the temporary full-expensing capital allowances announced in the Budget. Investment then falls back after the measure expires in April 2026.

1.10 The level of nominal GDP, which is the key driver of our forecast for the public finances, is 0.8 per cent higher at the forecast horizon than in our November forecast. This comprises 0.6 per cent from higher real GDP and 0.2 per cent from a higher GDP deflator, reflecting slightly stronger domestically generated inflation. This upward revision to nominal GDP has fed through to higher forecasts for wages and salaries, nominal consumer spending, and company profits (outside the oil and gas sector). These are the three largest tax bases and underpin around two-thirds of our overall revenue forecast.

Fiscal outlook

1.11 The modest improvement in economic prospects has flowed through to a somewhat brighter outlook for the public finances. Public sector net borrowing in 2022-23 is expected to be £152.4 billion, or 6.1 per cent of GDP. This is down £24.7 billion (14 per cent) relative to our November forecast, which reflects a mix of largely economy-related upward revisions to receipts (£14.8 billion) and largely energy-price-related downward revisions to public spending (£9.9 billion). Headline public sector net debt is expected to finish the year at 100.6 per cent of GDP, 1.2 per cent of GDP lower than forecast in November.

1.12 Before accounting for new policy measures, the outlook for borrowing has improved materially since November, but remains more challenging than a year ago. On average, our pre-measures forecast has been revised down by £24.5 billion (0.9 per cent of GDP) a year from 2023-24 onwards. In 2023-24, the improvement largely reflects the better starting point this year, with energy support measures and debt interest costing less than expected, while receipts are higher across all major taxes other than those related to energy. From 2024-25 onwards, downward revisions to pre-measures borrowing are dominated by upward revisions to receipts (averaging £24.0 billion a year) as stronger outturns are largely assumed to persist, while a shallower economic downturn and modestly higher medium-term output deliver stronger growth in key tax bases. These are only partly offset by upward revisions to spending on welfare, notably another upward revision to spending on health and disability benefits.

1.13 Presented with a £24.6 billion a year improvement in the pre-measures outlook for borrowing, the Chancellor has used two-thirds of this on his Budget measures. Policy measures announced since our November forecast raise borrowing in every year, and by an average of £15.6 billion a year from 2023-24 onwards (Chart 1.4). The policies announced in this Budget fall into five categories:

• Further energy support measures, including the continuation of the energy price guarantee for households at £2,500 for a further three months and extended support for businesses into 2023-24. These cost a total of £4.4 billion in 2023-24.
• A package of measures aimed at increasing labour market participation,whose cost rises to £7.1 billion a year by 2027-28. This includes additional help with childcare costs for working parents of younger children, reforms to working-age benefits, disability employment support, and more generous pensions tax allowances.
• Temporary 100 per cent capital allowances for qualifying business investment undertaken between 2023-24 and 2025-26. This costs an average of £9.1 billion a year during those three years, but it raises money in 2027-28 as capital allowance claims drop back after the temporary effects of the measure unwind.
• Other spending decisions that cost an average of £2.0 billion a year over the forecast period, most importantly an increase in defence spending.
• Other tax decisions that cost an average £3.6 billion a year over the forecast period, principally by freezing fuel duty at its current rate for another year. This involves a one-year extension of the temporary 5p cut coupled with a one-year cancellation of its RPI indexation. It costs £4.8 billion in 2023-24 when both elements apply, and £2.6 billion a year thereafter when only the RPI element has an ongoing cost.

1.14 The indirect effects of these policies, which boost demand in the near term and supply in the medium term, reduce borrowing by £3.0 billion next year (thanks in particular to lower inflation) and then by an average of £1.7 billion a year thereafter (as a boost to receipts that builds to £3.9 billion by 2027-28 is partly offset by the costs of servicing the additional debt issued to finance the measures, which rises to £2.6 billion).

Chart 1.4: Effect of policy measures on borrowing

1.15 The combination of a material improvement in our pre-measures forecast and a relatively large fiscal loosening in this Budget leaves borrowing at £131.6 billion (5.1 per cent of GDP) in 2023-24. It then falls progressively to £49.3 billion (1.7 per cent of GDP) by 2027-28 (Chart 1.5). Borrowing is, on average, around £10 billion a year (0.4 per cent of GDP) lower from 2023-24 onwards than we forecast in November. But it remains £50 billion a year (1.9 per cent of GDP) higher on average than we forecast in March 2022, before the scale of the energy price shock was clear, interest rates tripled, and the Chancellor loosened his fiscal targets to provide more space to accommodate their fiscal impact.

1.16 Our latest forecast continues to see the tax burden (the ratio of National Accounts taxes to GDP) reach a post-war high of 37.7 per cent of GDP at the forecast horizon in 2027-28, including the highest ratio of corporation tax receipts to GDP since the tax was introduced in 1965. We also still expect the ratio of public spending to GDP to settle at 43.4 per cent, its highest sustained level since the 1970s.

Chart 1.5: Public sector net borrowing

1.17 Our forecast for public debt has improved somewhat since November, but is still materially higher than we forecast in March of last year. The headline measure of public sector net debt peaks next year at 103.1 per cent of GDP, then falls steadily to reach 96.9 per cent of GDP in 2027-28. Underlying debt – which excludes the Bank of England and is the measure targeted by the Chancellor – does not peak until 2026-27 (a year later than we forecast in November) at 94.8 per cent of GDP and then falls only marginally (by 0.2 per cent of GDP) in the final year of the forecast. So in 2027-28, underlying debt stands at 94.6 per cent of GDP. That is 2.6 per cent of GDP lower than in our November forecast, thanks to both lower cumulative borrowing and higher nominal GDP. But this downward revision reverses only one-seventh of the upward revision to debt between our March and November forecasts last year. And the tripling of interest rates since this time last year means the share of revenues consumed by servicing that debt rises from 3.1 per cent in 2020-21, to 6.2 per cent in 2021-22, to 7.8 per cent by the end of the forecast period.

Chart 1.6: Public sector net debt excluding the Bank of England

Performance against the Government’s fiscal targets

1.18 The Government’s primary fiscal target – its ‘fiscal mandate’ – is for public sector net debt excluding the Bank of England to fall in the final year of the forecast (2027-28 in this one). On our central forecast it is met by a margin of £6.5 billion (0.2 per cent of GDP), down £2.7 billion (0.1 per cent of GDP) from the margin of £9.2 billion (0.3 per cent of GDP) in our November forecast. That reflects:

• Underlying forecast changes that add £5.3 billion to the Chancellor’s headroom. This is considerably smaller than the effects of the large downward revision to borrowing in 2027-28, which adds £28.4 billion to headroom. Four-fifths of this boost is offset by the combined effects of three headroom-reducing factors. Almost half (£13.8 billion) is lost to slower nominal GDP growth in the final year of the forecast, largely because the shallower near-term economic downturn than we expected in November leaves less scope for above-trend growth in the medium term. A quarter (£7.3 billion) is lost to the fact that part of the improvement in accrued borrowing is not reflected in cash debt in 2027-28 (such as the additions to interest accrued on student loans and due to the lags between the upward revisions to accrued taxes and the associated payments being made). And the remaining £2.0 billion is lost to the debt-to-GDP ratio in 2026-27 being lower than we forecast in November, which means borrowing must be lower to get debt falling in the subsequent year.
• Budget policy measures that reduce the headroom by £8.0 billion. This is more than explained by the direct cost of measures in 2027-28 of £9.9 billion and the higher debt interest spending from the cumulative fiscal loosening over the forecast period that reaches £2.6 billion. These effects are partly offset by a £3.9 billion boost to receipts from indirect effects.^[1]

1.19 £6.5 billion is the smallest amount of headroom any Chancellor has set aside against his primary fiscal target, which has averaged £25.6 billion in today’s terms since the OBR was established in 2010. One evident threat to even that small amount of headroom is the deliverability of the policies in this Budget. Specifically, our revenue forecast includes £4.0 billion in fuel duty receipts in 2027-28 from the Government’s stated, but rarely implemented, policy of raising fuel duty rates in line with RPI inflation every year, plus the reversal of the temporary 5p cut that is now in its second year. Cancelling these planned increases, as every Chancellor has done since 2011, and instead holding fuel duty at its current rate would, in itself, more than halve the Chancellor’s headroom.

1.20 The Government’s supplementary fiscal target – for borrowing to be less than 3 per cent of GDP in the fifth year of the forecast – is met by £39.2 billion (1.3 per cent of GDP). For this target, headroom has doubled since November (up £20.6 billion, or 0.7 per cent of GDP). The improvement reflects lower borrowing (as set out above), combined with the small upward revision to nominal GDP (making 3 per cent a higher target in cash terms).

1.21 History suggests that these margins would be consistent with a 52 and 68 per cent chance of meeting the fiscal mandate and supplementary target, respectively. The former is very low by historical standards, while the latter is relatively high. Chancellors have typically aimed for around a 60 per cent chance of meeting their targets.

1.22 The Government’s ‘welfare cap’ is on course to be breached by £4.1 billion, having been on course to be met by £0.6 billion in November. As we are only tasked by Parliament with formally assessing performance against the welfare cap at the first Budget in each Parliament, this prospective breach carries no formal consequences. And the Government has not fully addressed it through the policies announced in this Budget.

Risks and uncertainties

1.23 Over recent years, large shocks and their aftermath have resulted in significant revisions to our economic and fiscal forecasts from one fiscal event to the next. This has been true again in this forecast, with a material improvement in the pre-measures outlook relative to November – two-thirds of which has been spent via loosening fiscal policy. We therefore continue to emphasise the considerable uncertainty around our central forecast, with the possibility that any of several key judgements could prove too optimistic or pessimistic.

1.24 Three sources of risks to our economic and fiscal forecasts are particularly important at the moment, so we present alternative scenarios around each:

• Labour supply. If participation were to be either 500,000 higher or lower by 2027-28 than in our central forecast, borrowing would be around £11 billion (0.4 per cent of GDP) lower or higher in that year, thanks largely to the impact on receipts. In the upside scenario, the margin against the fiscal mandate would roughly treble, but in the downside scenario it would be gone.
• Energy prices. If gas prices were either to rise back to the average of the past year or to fall back to pre-invasion averages, borrowing would be £5½ billion (0.2 per cent of GDP) higher or lower in 2027-28, thanks largely to inflation-related impacts on debt interest and welfare spending, and wider economy-related effects on receipts. Combined with their effects on the path for nominal GDP, the high energy price scenario would roughly halve the headroom against the fiscal mandate, whereas the low energy price scenario would roughly double headroom.
• Interest rates. If interest rates (both short- and long-term) were to be 1 percentage point higher or lower than assumed across our central forecast, which is around half the rise in 10-year gilt yields over the past year, borrowing in 2027-28 would be higher or lower by around £20 billion (0.7 per cent of GDP), largely due to debt interest spending. That would wipe out headroom against debt falling in the higher rates scenario but would more than treble headroom in the lower rates scenario.

1.25 In addition to the risk from RPI-indexation of fuel duty, there are also risks to our forecast associated with the impact or implementation of other stated policies and aspirations. On the revenue side, these risks include: uncertainties around the yield from the rise in corporation tax, which is estimated to raise receipts by £20 billion a year by 2027-28, and the Chancellor’s aspiration to make the full-expensing capital allowance permanent, which could cost approaching £10 billion a year. On the spending side, risks include: the aspiration to spend 2.5 per cent of GDP on defence by an unspecified date, 0.5 per cent of GDP (£15 billion in 2027-28) higher than the existing NATO commitment; inflationary pressures on departmental budgets over the next two years, including those related to pay (in total worth up to £30 billion); and the squeeze on those budgets beyond the current Spending Review (where history points to more than £30 billion a year being added to totals when the time comes to set detailed plans).

1.26 Even in our central forecast, it is now harder for this Chancellor to deliver a falling path for the debt-to-GDP ratio in the medium term than it has been for any of his predecessors since the OBR was established in 2010. This is due to the combination of: (i) subdued medium-term growth prospects, reflecting post-financial crisis weakness in productivity growth exacerbated by a series of further shocks in the form of the pandemic and rise in energy prices; (ii) a stock of debt that has been pushed to a 60-year high, largely as a result of those shocks; and (iii) interest rates on that higher stock of debt, which have tripled over the past year to their highest level in over a decade. It all adds up to a situation in which for any given debt-to-GDP ratio, less can be borrowed without that ratio rising; and for any given level of borrowing, more must be spent on debt interest, leaving less scope to finance other priorities.

Chapter 2: Economic outlook

Introduction

2.1 This chapter describes our latest economy forecast, including:

• our conditioning assumptions relating to commodity prices, monetary policy and gilt yields, the global economy, and the exchange rate (from paragraph 2.3);
• fiscal policy, including measures announced in this Budget (from paragraph 2.10);
• prospects for inflation (from paragraph 2.12);
• how our forecasts for labour supply, business investment, and total factor productivity determine the path of potential output (from paragraph 2.18);
• prospects for the output gap and real GDP (from paragraph 2.37);
• the outlook for the labour market (from paragraph 2.41);
• our forecasts for household incomes, the saving rate, consumption, and the housing market (from paragraph 2.46);
• the outlook for net trade, the current account and sectoral balances (from paragraph 2.52);
• the path for nominal GDP (from paragraph 2.55); and
• how our economy forecast compares with a range of recent external forecasts (from paragraph 2.58).

Key economy forecast assumptions

2.2 This forecast is based on a set of judgements about the key determinants of economic activity over the next five years, which are summarised in Table 2.1 alongside the changes since our November forecast. We have conditioned this forecast on market expectations for commodity prices, Bank Rate and gilt yields averaged over the five working days to 8 February, a shorter window than normal so that it exclusively covers the period following the Bank of England’s Monetary Policy Committee (MPC) meeting on 2 February. The sensitivity of our forecasts to different interest rate and gas price assumptions is explored in Chapter 5.

Table 2.1: Key economy forecast assumptions and judgements

Commodity prices

2.3 Wholesale gas futures prices have fallen significantly in both the near and medium term since our November 2022 forecast but remain more than twice their pre-pandemic average of around 50p (Chart 2.1). The fall in prices reflects a combination of: (i) Europe increasing its reserves faster than expected over the summer and autumn; (ii) a relatively mild winter; and (iii) rising liquified natural gas deliveries compensating for the loss of Russian imports (see Box 2.1). Having peaked at £4.40 a therm in August 2022, prices are now expected to average £1.50 a therm in 2023, less than half the £3.40 a therm we assumed in November. On our central forecast gas prices hover around the £1.70 mark through 2024 before stepping down again to £1.30 a therm in the second quarter of 2025. In the final two years of our forecast we hold prices constant in real terms. The dollar cost of oil in 2023 has fallen slightly to $81 a barrel compared to our November forecast of $85 a barrel, but the sterling cost has fallen by 10 per cent due to a stronger pound.

2.4 The Government’s energy price guarantee (EPG), which caps the typical annual household energy bill to £2,500 until June 2023 then to £3,000 until March 2024, means near-term falls in gas prices largely reduce the fiscal costs for the Government rather than lower inflation and boost real incomes for households. However, from the third quarter of 2023, prices fall to well below the EPG’s £3,000 cap meaning households directly benefit from lower gas and electricity bills, the average costs of which fall to £2,200 by the end of 2024 (right panel of Chart 2.1). Businesses energy costs are also capped by the energy bill relief scheme and its successor, the energy bills discount scheme, while firms also benefit from lower wholesale energy prices since our last forecast. As in November, gas prices remain volatile and are a major source of uncertainty in our forecast. In Chapter 5 we illustrate the economic and fiscal impact of the higher or lower gas prices shown by the swathe in the left panel of Chart 2.1.

Chart 2.1: Wholesale and retail energy prices

Box 2.1: How gas demand and supply have responded to higher gas prices

In response to the largest rise in energy prices in around 50 years, an immediate reduction in energy demand, and more gradual change in the composition of energy supply, might be expected. Our initial analysis of household gas demand up to February 2023 supports our November forecast assumption that for every 10 per cent rise in households’ energy costs, energy demand would fall by 1 per cent (i.e. a price elasticity of demand of -0.1).^a

After adjusting for weather (removing the effects of the warm winter reducing central heating usage and therefore gas consumption), demand this winter is around 15 per cent lower than the two years preceding the Russian invasion of Ukraine. At the same time, household gas prices were on average around 200 per cent higher (left panel of Chart A). This implies a relationship between prices and demand that is close to our assumed elasticity of -0.1. All else equal, this elasticity would rise over time if consumers gradually purchased more energy-efficient items, such as insulation, although consumers may offset this by increasing energy consumption in other ways, such as by using central heating more often.^b

The full supply response to higher gas prices is likely to take much longer. Averaging roughly 40 per cent of electricity generation, gas is both the baseload and marginal source of electricity for the UK when more intermittent renewable sources (such as wind) are not available. Despite continued investment in renewable energy and renewed interest in alternative sources that can provide a stable baseload such as nuclear,^c the continued shift away from gas as our primary source of electricity is likely to be gradual rather than rapid. Hinkley Point C, the UK’s first new nuclear reactor in 52 years, is due to open in 2027, almost 20 years after its announcement in 2008. Sizewell C, also announced in 2008, is yet to begin construction.

So, in the near term, the supply response has primarily comprised a change in the sources of the UK’s imported gas. The UK has historically imported a smaller proportion of its gas supply from Russia than other European countries and has therefore had to substitute less of its gas supply for liquified natural gas (LNG) from outside Europe, which is largely replacing pipeline Russian gas in Europe. But LNG as a share of gas supplies has still risen to 47 per cent of UK gas imports in 2022 versus an average of 26 per cent over the previous decade, with supplies from Russia falling to less than 1 per cent (right panel of Chart A). The UK is relatively well positioned for this Europe-wide shift towards LNG imports, holding roughly one-quarter of Europe’s regasification facilities before the pandemic (in part the flipside to the UK’s relatively low gas storage capacity).^d As such, the UK’s gas exports more than tripled in 2022 as some of the LNG imported into the UK was processed and exported to Europe through gas interconnectors. But in the short term, LNG supply to Europe is limited by a lack of regasification capacity and of pipelines from countries such as Spain and the UK that have large regasification capacity. This keeps wholesale gas prices in the region’s highly integrated market high. In the medium term, investment in liquification facilities is expected to increase world LNG supply by 19 per cent between 2021 to 2026.^e Significant investment in new regasification facilities in Europe also contributes to the fall in market expectations for European gas prices over the coming years.

Chart A: UK demand for and supply of gas

Our central forecast is based on futures market prices for gas up to the end of 2025 and then holds them constant in real terms. Evidence supports using futures in the near term (up to one year ahead), with market signals consistent with further demand and supply responses lowering wholesale prices; in the long run a futures-based forecast is not significantly better than a random walk,_fand medium-term futures could also embody a risk premium over future spot prices.^g Given this uncertainty, Chapter 5 explores upside and downside gas price scenarios.

a) DECC, Gas price elasticities: the impact of gas prices on domestic consumption, June 2016.

b) Peñasco, C., and L. Anadón, Assessing the effectiveness of energy efficiency measures in the residential sector gas consumption through dynamic treatment effects: Evidence from England and Wales, January 2023.

c) BEIS, British energy security strategy, April 2022.

d) Yafimava, K., ‘Finding a home’ for global LNG in Europe, 2020. The UK has nine days of gas reserves while Germany has 89.

e) Bloomberg, Global LNG market outlook, June 2022.

f) See Reichsfeld, D., and S. Roache, Do Commodity Futures Help Forecast Spot Prices?, November 2011.

g) This may occur when hedging demand from buyers of futures outstrips hedging from sellers, leading to risk-averse speculators having to take uncovered short positions for which compensation is required.

Monetary policy and gilt yields

2.5 The MPC has raised Bank Rate from 3 per cent at the time of our November forecast to 4 per cent at its February meeting, the highest level since 2008 (left panel of Chart 2.2). The market expectations underpinning our central forecast point to it rising further to a short-lived peak of 4.3 per cent in the third quarter of 2023, 0.7 percentage points below the peak of 5 per cent in our November forecast (Chart 2.2). Bank Rate then falls back to 3.0 per cent by the forecast horizon, 0.6 percentage points below our November forecast.^[2] Since we closed our forecast, market participants’ expectations for Bank Rate have risen to peak at 4.8 per cent, 0.5 percentage points above the peak incorporated in our forecast. (We consider possible implications of this, and of other recent market moves, in Chapter 5.)

2.6 At the time we closed our market interest rate forecasts on 8 February, gilt yields at the short end of the curve had fallen slightly since our November forecast but were little changed at the long end (right panel of Chart 2.2). 10-year spot gilt yields had fallen from 3.7 per cent at the time of our November forecast to 3.3 per cent but were still significantly higher than the post-financial crisis average of around 2 per cent. Gilt yields across the curve were around 1 percentage points lower than the spike following the announcement of the Growth Plan on 23 September, when 10-year spot yields rose to 4.4 per cent. Since we closed our forecast, 10-year spot gilt yields have risen 0.5 percentage points to 3.8 per cent, but are still 0.7 percentage points lower than the 27 September peak.

Chart 2.2: Bank Rate and gilt yields by year of maturity

2.7 The outlook for interest rates remains a key risk to our forecast. Risks to our assumptions for Bank Rate and gilt yields stem both from domestic inflationary pressures and the consequent monetary policy response, as well as a volatile global interest rate environment. In Chapter 5, we look at the impact of higher and lower interest rates on our forecast.

World economy and the exchange rate

2.8 In line with the IMF’s January World Economic Outlook Update, we expect global GDP growth to be 2.9 per cent in 2023, 0.2 percentage points higher than our November forecast as lower energy prices, lower interest rate expectations, and China’s easing of Covid restrictions boost growth. Growth in the euro area in 2023 has also been revised up, largely due to lower gas prices and stronger-than-expected recent GDP outturn, though the bloc remains vulnerable to higher energy prices, and replenishing gas reserves this summer may prove more challenging than in 2022. There is also uncertainty around the extent to which the recent rise in Covid infections in China and escalating global trade policy frictions will affect supply chains, which could hamper global GDP growth in the short and medium term respectively. Global growth rises to 3.7 per cent in 2024.

2.9 The sterling effective exchange rate is little changed since November and in this forecast we assume that the exchange rate remains constant in nominal terms. This reflects two offsetting movements, with sterling having appreciated by 5.5 per cent against the US dollar but having fallen 2.2 per cent against the euro.

Fiscal policy

2.10 Borrowing remains elevated by historical standards in 2023-24 as the economy recovers from the pandemic and the fiscal support for energy bills provided in the wake of the Russian invasion of Ukraine is gradually withdrawn. The cyclically adjusted primary deficit (which is one measure of the ‘fiscal stance’),^[3] falls from 1½ per cent of GDP in 2023-24 to zero in 2024-25, before tightening further in the final two years of the forecast to reach a 1 per cent surplus in 2027-28.

2.11 In this Budget, the Government has spent two-thirds of the improvement in our pre-measures forecast for the public finances since November on new measures. This loosening over the next three years supports households with their energy bills and incentivises business investment. The Government has also loosened fiscal policy in the medium term, by somewhat less, to pay for a set of tax and spending changes, some of which target the supply side of the economy. Their impact on our forecast is discussed in Box 2.2.

Box 2.2: The economic effects of policy measures

Our economic forecasts account for the economic impact of the latest announced government policies. This includes estimates of the direct fiscal costs or savings from all policy measures and their near-term demand-side impacts on the economy. To estimate the effect of discretionary fiscal policy changes on aggregate demand, we use multipliers drawn from the empirical literature. These capture wider effects of fiscal policy measures on output over and above their direct effects on demand, through changes to private incomes and spending. We review these estimates periodically.^a The impact of policies on the supply side of the economy is also accounted for if credible evidence suggests that measures will have a material, additional, and durable impact on potential output.^b

Policies announced in this Budget add around £20 billion a year to public sector net borrowing between 2023-24 and 2025-26, declining to around £10 billion by 2027-28. Of this package, the near-term fiscal loosening boosts aggregate demand relative to supply by 0.3 per cent at the peak of its impact in 2023-24 and 2024-25, narrowing the output gap by the same amount. As usual, we assume the demand impact of these policies tapers to zero over the forecast period, as the Bank of England acts to offset any inflationary pressure by tightening monetary policy to bring aggregate demand in line with potential output.

In 2024-25 in particular, this 0.3 per cent figure largely reflects the impact of the temporary increase in the generosity of capital allowances for businesses, which lets firms reduce their taxable profits by 100 per cent of the cost of their investments in plant and machinery for three years from April 2023. This incentive to accelerate investment plans boosts our business investment forecast by amounts peaking at almost 3½ per cent in 2024-25 and 2025-26. But the policy’s temporary nature leaves the optimal capital stock unchanged in the long run, so in the final year of the forecast business investment is 4 per cent lower than it would otherwise be. We assume that around half of the additional investment is imported, so the initial impact on GDP is smaller than the impact on our business investment forecast would imply.

In addition to these impacts on demand, the Government has announced a £7 billion package of support to target labour supply. We judge that five policies in the Spring Budget will have a material and positive impact on supply, specifically labour supply, and have explicitly incorporated their effects into our forecast for potential output. Of these:

• We expect the new 30 hours a week of free childcare for working parents of nine-month- to two-year-olds to gradually increase labour market participation of parents with young children. By 2027-28, we expect around 60,000 to enter employment and work an average of around 16 hours a week (in line with the average for part-time workers). An equivalent effect on total hours comes from the 1½ million mothers of very young children already in work increasing the hours they work by a much smaller amount. This policy has by far the largest impact on potential output in this Budget.^c
• We have increased employment by around 15,000 at the forecast horizon to account for the impact of changes to childcare support within universal credit, in particular reimbursing parents for the first month’s fees upfront.^d
• Increasing conditionality on parents and carers claiming universal credit increases labour supply marginally at the forecast horizon by encouraging people to increase their work search intensity, raising employment by up to 10,000.
• Changes to the lifetime allowance and annual allowance on pension contributions increase employment by around 15,000 by removing some financial disincentives to continuing in employment for those with large pension pots.^e
• A new disability employment programme will collect referrals from the health-related benefits system and other settings to support inactive disabled people into sustained work. We assume this will increase employment by 10,000, based in part on evaluations of similar programmes.^f

Relative to our pre-measures forecast, our central estimate is that these policies, taken together, increase employment by 0.3 per cent (110,000) by 2027-28. But part of the employment impact on potential output is offset by lower-than-average hours and earnings among many of the new joiners, so the overall impact on GDP is around 0.2 per cent in 2027-28. This is the largest upward revision we have made to potential output within our five-year forecast as a result of fiscal policy decisions taken by a Government in any of our forecasts since 2010.^g

Chart B: Impact of policy measures on real GDP

The bringing forward of investment to benefit from a temporary capital allowance policy will not alter the path of the capital stock or potential output in the long run (i.e., beyond our forecast horizon), but it does temporarily place the capital stock on a higher path. We estimate that it will be higher by 0.2 per cent in 2027-28. This is roughly in line with the increase in total hours from policy interventions, therefore there is no change to the capital intensity of the economy, leaving the level of potential productivity per hour unchanged. The Chancellor has indicated his intention to make the measure permanent when economic and fiscal conditions allow – we discuss the implications this would have for our forecast in paragraph 2.34.

Our central estimate of the increase in labour supply as a result of the policies announced in this Budget is very uncertain, and a plausible range could be as high as 240,000 or as low as 55,000 based on alternative plausible assumptions. The higher estimate might reflect: the increase in mothers’ participation matching historical estimates of changes in participation rates when children go to school; more individuals responding to the changes to the lifetime and annual pension allowances than expected; caseloads for the new disability employment programme being higher than we expect; and changes to universal credit conditionality and childcare payments bringing more people than expected into work. It is also plausible that a lower figure could occur, for instance if parents’ labour market decisions are less responsive to childcare provision or if changes to the pension allowances do not incentivise workers to stay longer in the labour force.

It will be important to monitor the implementation of these policies to ensure that additional resources are being provided to deliver the proposed interventions, rather than them being reprioritised from other programmes, and that programmes are being executed according to the timetables assumed in our forecast. And our ongoing assessment of their economic impact will also be informed by a regular programme of rigorous evaluation of the various interventions which we will draw on to refine our estimates of their supply-side impact over time.

Other policies announced in this Budget directly affect inflation. Taken together, the freezing of fuel duty, changes to alcohol duty and the extension of the EPG lower CPI inflation by 0.7 percentage points in 2023-24. The effects of subsequent increases to fuel and alcohol duties and the EPG measure then add 0.4 percentage points to CPI inflation in 2024-25.

a) See our December 2019 Forecast evaluation report, and November 2020 Economic and fiscal outlook.

b) See our Briefing Paper No. 8: Forecasting potential output – the supply side of the economy, November 2022.

c) Based on our assessment of the literature, e.g., M. Brewer, et al., Does free childcare help parents work more, IFS, March 2020.

d) This assessment is, in part, based on analysis of the impact of similar changes that have already occurred in Northern Ireland.

e) Based on evidence from our NHS pensions forecast and from the medical profession more generally, see NHS England, Submission to the Review Body on Doctors’ and Dentists’ Remuneration, Evidence for the 2023/24 pay round, 2022, which cites pensions tax policy as the most likely reason clinicians will reduce hours or retire early.

f) Our estimates are based on DWP’s unpublished analysis of the Work Choice programme.

g) We can only do this comparison for measures whose impacts we have explicitly quantified. For instance, the November 2020 Economic and fiscal outlook did not explicitly quantify any impact on potential output from the coronavirus job retention and business guarantee schemes but noted that their effects on potential output could be “substantial”.

Inflation

2.12 CPI inflation peaked at 11.1 per cent in October 2022 and we expect it to fall rapidly to around 3 per cent in the final quarter of 2023. The sharp fall occurs as the rapid price increases in 2022 drop out of the calculation of annual inflation, gas and electricity prices fall, and there is some further easing of supply bottlenecks, which flow through to the price of tradable goods. This is only partly offset by continued strength in non-tradable services inflation due to elevated wage growth this year. On a calendar year basis, we expect CPI inflation to be 6.1 per cent in 2023, 1.2 percentage points lower than our November forecast, mainly due to a more rapid decline in energy prices. The EPG stops biting in the third quarter of 2023 – three quarters before the scheme officially ends – so from then on lower wholesale prices feed through to lower bills for households.

2.13 Conditional on market expectations for Bank Rate and gas prices, inflation falls sharply to 0.9 per cent in 2024 as energy and tradables prices fall further and then oscillates around zero through to mid-2026 (Chart 2.3). CPI inflation returns to the 2 per cent target by the forecast horizon. Compared to our November forecast, the level of consumer prices is little changed by the horizon as lower energy prices are offset by more domestically generated inflation from higher demand relative to supply over most of the forecast.

Chart 2.3: Contributions to CPI inflation

2.14 Chart 2.4 illustrates the degree of uncertainty surrounding our central CPI inflation forecast based on the historical distribution of our (and before that the Treasury’s) forecast errors since 2003. The solid black line shows our central forecast with successive pairs of lighter shaded areas around it representing 20 per cent probability bands. The chart implies a roughly 10 per cent probability that the CPI inflation rate will be above 3.6 per cent and a 10 per cent probability that it will be below -1.3 per cent in 2024. But the average of past forecast errors is not always a good indicator of the degree of uncertainty at specific points in time, especially following very large shocks such as last year’s energy price rises. The distribution around our 2023 CPI inflation forecast is therefore likely to understate the degree of uncertainty in the current environment of very high and volatile energy prices.

Chart 2.4: CPI inflation fan chart

2.15 RPI inflation is expected to have peaked in the final quarter of 2022 at 13.9 per cent, 0.2 percentage points lower than the peak assumed in our November forecast. This is largely driven by a lower-than-expected peak for CPI inflation, partly offset by a slightly higher-than-expected wedge between CPI and RPI inflation driven by differences in weights. RPI inflation is also expected to fall rapidly to 4.9 per cent by the final quarter of 2023, as CPI inflation and house prices (which affect the depreciation component of RPI) drop. RPI inflation then eases further in the middle years of the forecast as CPI inflation falls further and mortgage rates (which affect the mortgage interest component of RPI) level off. It then rises in step with CPI inflation to reach 3.1 per cent at the forecast horizon.

2.16 GDP deflator growth – a measure of the change in the price of domestically produced goods and services – was 5.1 per cent in 2022, and our central expectation is for it to slow to 2.9 per cent in 2023, before settling at an average of 1.4 per cent a year for the remainder of our forecast. In 2022 and 2023, the private consumption deflator is the largest contributor to GDP deflator inflation, contributing an average of 4.0 percentage points. GDP deflator growth was significantly lower than CPI inflation in 2022, partly due to a fall in the terms of trade as the price of UK imports rose relative to the price of exports. The level of the GDP deflator at the forecast horizon is little changed from the November 2022 forecast.

2.17 The terms of trade fell because the UK is a net importer of energy, food, and other tradable goods and the prices of these items recently rose relative to the price of goods and services for which the UK is a net exporter. As financial markets currently expect that energy prices will be permanently higher, this negative terms of trade shock is expected to persist, meaning the UK is permanently poorer (albeit less so than was the case based on the market expectations embodied in our November forecast). This is shown in Chart 2.5 where, from a starting point of the final quarter of 2021 (ahead of the Russian invasion of Ukraine), the CPI (the price paid by households for goods and services) rises to 3.8 per cent higher than GDP deflator (the price received for goods and services produced in the UK) by the end of 2022 and remains 2.3 per cent higher at the forecast horizon. The gap between the two indices is one measure of the reduction in the real purchasing power of UK households due to the increase in the price of goods and services that the UK imports relative to those that we produce domestically.

Chart 2.5: The terms of trade shock: GDP deflator versus consumer price index

Labour supply and potential output

2.18 Potential output anchors our forecast for the level of GDP in the medium term. It reflects the quantity of labour and capital available to businesses and the efficiency and intensity with which they are deployed (‘total factor productivity’ (TFP)). While actual GDP can deviate from potential output over the forecast period, we generally assume that by the forecast horizon a combination of monetary policy, fiscal policy, and natural economic adjustment mechanisms bring the two back into line with each other. Were that not to be the case, an excess of demand relative to supply or vice versa would tend to mean inflation permanently deviating from its target.

2.19 Our estimate of the starting level of potential output is 0.4 per cent lower in 2022 than our November forecast (right panel of Chart 2.6). This is largely because of a more pessimistic assessment of labour market participation, reflecting our judgement that more of the recent weakness in participation is structural. We have also lowered our estimate of TFP as indicators of capacity utilisation and supply bottlenecks suggest conditions were tighter than expected, implying lower potential TFP for a given level of actual productivity.

2.20 Over the five years of our forecast, we expect potential output growth to average 1.9 per cent a year and settle at 1¾ per cent at the forecast horizon (left panel of Chart 2.6). Compared to November, our forecast for potential output growth is stronger in the near term because lower energy prices lead to stronger TFP growth (0.2 percentage points in 2023), population growth is higher (adding 0.1 percentage points in each year of the forecast), and policy measures begin to boost the labour force. The final year figure for growth is unchanged from November because the 0.1 percentage point addition from higher population growth is broadly offset by lower investment, which reduces the contribution of capital deepening to potential output. It is a percentage point lower than the average growth rate in the decade prior to the financial crisis and around 0.2 percentage points below the post-financial crisis average.

2.21 As a result, at the end of the forecast in 2027, the final level of potential output is 0.5 per cent higher than our November forecast, reflecting a larger total population and higher TFP which more than offset a slightly lower labour participation rate and less capital deepening. As described in Box 2.2, 0.2 percentage points of this 0.5 per cent upward revision reflects the impact of policies in this Budget. In per-person terms, potential output is broadly in line with our November forecast by the horizon due to the larger population.

Chart 2.6: Potential output

Labour supply

2.22 Relative to our November forecast, annual growth in overall labour supply (the underlying trend in total hours worked across the economy) over the forecast horizon is on average 0.1 percentage points higher due to a larger contribution from net inward migration partly offset by a lower pre-measures participation rate. In 2022, the starting level of overall labour supply is estimated to be 0.4 per cent lower than in November, driven by a downward revision to the trend rate of labour market participation. By the forecast horizon, labour supply is 0.3 per cent higher due to an upward revision to our forecast for net migration and the positive impact of the policy measures in the Budget. Trend average hours and our estimate of the structural unemployment rate are little changed from our November forecast at broadly flat rates, and so have relatively little influence on our revisions to potential GDP growth.

Population

2.23 Population growth over the forecast has been revised up from an average of 0.6 to 0.7 per cent a year. We have taken on the ONS’s latest January 2023 projections for total net inward migration, which we now expect to total 1.6 million cumulatively over the forecast (up from 1.3 million in November and 0.8 million in March 2022). Net migration flows settle at 245,000 a year, rather than the 205,000 assumed in our November forecast and 129,000 in our March 2022 forecast (left panel of Chart 2.7). A larger population, due to increased net migration, adds 0.5 per cent to potential output in 2027.

2.24 Higher forecast levels of inward migration are also consistent with relatively high recent rates of actual net migration, which the ONS estimates reached over half a million in the year to mid-2022. Higher recent rates of inward migration can be attributed to:

• the resumption of international travel following the pandemic, especially among foreign students, with student visas reaching a record high of 490,000 in 2022;
• the post-Brexit immigration regime that began in 2021 and issued 800,000 visas in its first year of operation (only 50,000 of which were for EU citizens who did not require a visa under the previous regime); and
• other changes , including 129,000 British National Overseas visas to Hong Kong nationals and 210,000 visas to Ukrainian nationals.

2.25 Risks around the net migration projection are large, especially as ONS projections simply take averages of past net migration figures and assume flows converge to this level in five years. The latest set of projections use a 22-year average of past migration to give a long-term figure of 245,000. These projections are not forecasts and so do not make any explicit adjustments for global factors (such as the war in Ukraine) or associated policy responses.^[4]

2.26 Given the significant change in the composition of net migration since the new post-Brexit migration regime was introduced in 2021,^[5] there is considerable uncertainty about the participation rate of new migrants. Migrants into the UK have historically been skewed toward those of working age. In 2021, 70 per cent of UK residents born outside the UK were aged 26 to 64 whereas only 48 per cent of those born in the UK were in this age group. Our previous forecasts have therefore typically assumed that migrants have higher average participation rates than the resident population, given their age profile.^[6] However, the latest visa data suggest that work visas made up just 20 per cent of the total visas issued in 2022, study visas 30 per cent, and non-work non-study visas – including dependent, Ukraine and British National Overseas visas – 50 per cent. The latter group is the fastest growing category since the pandemic (right panel of Chart 2.7). It is therefore likely that the participation rate of migrants under the post-Brexit regime will be lower than in the past, so we have assumed they will have the same participation rate as the resident population.

Chart 2.7: Net migration and non-visitor visas

Labour market participation

2.27 We have kept the rate of labour market participation at the forecast horizon broadly unchanged at 63.0 per cent. This is the result of a 0.2 percentage point downward revision from our judgement that more of the recent fall in participation is structural rather than cyclical, offset by a 0.2 percentage point upward revision from new policies.^[7]

2.28 Actual labour market participation rates had been rising steadily over the 2010s from a post-financial crisis low of 63.1 per cent in 2011 to a pre-pandemic high of 64.0 per cent of the adult population in 2019. That reflected increases in the State Pension age, rising female participation, and an inflow of EU migrants who were disproportionately of working age. However, in the wake of the pandemic, the participation rate fell by around a percentage point to 63.2 per cent in the final quarter of 2022, undoing almost all of the increase of the preceding decade. This leaves the overall size of the adult labour force in 2022 around 500,000 lower than assumed in our pre-pandemic March 2020 forecast.

2.29 The UK’s working-age participation rate of 78.6 per cent remains above the OECD average of 73.2 per cent (which the OECD defines as those aged 15 to 64, whereas the above rates relate to the entire adult population aged 16 and over). But the OECD average is back above its pre-pandemic rate whereas the UK’s is still well below, making its post-pandemic performance worse than every other G7 economy (Chart 2.8). Box 2.3 explores the possible factors driving this recent rise in inactivity among working-age UK adults.

Chart 2.8: Participation rates of those aged 15 to 64

Box 2.3: Why has working-age inactivity risen since the pandemic?

Inactivity among over 16s has risen by 830,000 since the pandemic. Over 64s account for 310,000 of this rise, as the ‘baby boom’ population bulge created by the high birth rates in the decades after the Second World War increasingly shifts into retirement age. This demographic effect has long been reflected in our forecasts. But the remaining 520,000 increase among 16- to-64-year-olds has been a more concerning post-pandemic surprise. This box splits inactivity into age groups, and by self-reported main reason, to analyse this development.

Chart C: Change in 16-to-64-year-old inactivity

Inactivity by age

The change in the number of working-age inactive people between the fourth quarter of 2019 and the fourth quarter of 2022 has been largely driven by older age groups,^a with 50-to-64-year-olds accounting for 310,000 of the overall increase in the stock (left panel of Chart C). The number of inactive 16-to-24-year-olds have risen by 30,000, whereas inactivity among 25-to-49-year-olds has only risen by 15,000 (although, as discussed below, there has been a significant change in the composition of inactive people).

Chart D: Inflow and outflow rates to and from inactivity by age group

Looking at the flows into and out from inactivity (shown in Chart D, above) illustrates the drivers of the changes:

• Among those aged 16 to 64 as a whole (top-left panel of Chart D) both inflow rates and outflow rates from inactivity have risen. But rising inflows have more than offset rising outflows to drive higher overall working-age inactivity.
• Among those aged 16 to 24 (top-right panel), the inflow rate into inactivity has increased since 2019 as more young people went into higher education. Outflows subsequently rose as higher education courses ended and this cohort of students entered the labour market. But to a lesser extent than inflows, leaving the inactivity rate slightly higher.
• Among those aged 25 to 49 inflows increased during the pandemic but have since returned to more normal rates (bottom-left panel). Outflow rates have increased and remained higher. This led to a temporary increase in the stock during the pandemic, but a reversal of the flow dynamics has now left the stock little changed.
• Among those aged 50 to 64 there was a big rise in inflows during the pandemic, which have remained elevated (bottom-right panel). The ageing of the ‘second baby boom’ generation (those born during the 1964 birth peak are now in their late 50s) will be a partial driver of this higher flow, as this cohort moves into age brackets where activity rates are typically lower. By contrast with the other age groups, the outflow rate has fallen and remains lower, so both rising inflows and falling outflows explain the increased stock.

Inactivity by reason

We can also look at the changes in inactivity by the main reason survey respondents offer for being inactive (right panel of Chart C). This breakdown is more challenging to interpret because many people have multiple, overlapping reasons for being inactive, and the one they rank as most important can change over time (e.g., as health conditions fluctuate) There is also recent evidence of elevated churn between different reasons for inactivity.^b But it remains important for understanding the reasons behind the sharp rise in inactivity over the past three years:

• The pandemic did not initially interrupt the downward trend in those who cite looking after family or home as a reason for not seeking employment, which has fallen by 140,000 compared to immediately pre-pandemic (three months to February 2020), continuing a 530,000 (22 per cent) fall over the preceding decade. This group now accounts for 19 per cent of the working-age inactive population or 1.7 million people. The recent fall reflects the continuation of two pre-pandemic trends up until early 2021: a declining birth rate reducing the number of mothers of young children and rising maternal employment. Over the past 18 months this latter trend has reversed slightly.
• The pandemic did appear to encourage more people to go into, or stay in, education, with the number of people citing being a student as their main reason for inactivity rising from 2.1 million immediately before the pandemic to 2.3 million in the fourth quarter of 2022, and participation in full-time education among those aged 16 to 24 rising from 43 to 45 per cent. This temporary effect on labour participation is likely to unwind as this cohort graduates and joins the labour force.^c For now, students constitute 26 per cent of the working-age inactive population and 5.5 per cent of the working-age population, compared to 25 per cent and 5.1 per cent respectively immediately before the pandemic.
• The pandemic also seems to have spurred some older workers to take early retirement . Inflows from activity to inactivity due to retirement rose, but were largely offset by an increased outflow from ‘retirement’ to the various other main reasons for inactivity. Retirement inflows are likely to have been facilitated by ‘forced’ savings and the rise in house and other asset prices during the pandemic (a view supported by the increased flow into inactivity from those in better-paid occupations)^d. However, while the average retirement age for men rose steadily for two decades to reach 65.3 in 2020, and then fell to 65.1 in 2021, it then rebounded to a five-decade high of 65.4 in 2022. For women, this metric followed a very similar trajectory, returning to 64.3 – its pre-pandemic high – in 2022, suggesting that the pandemic accelerated retirement for one cohort of better-off older workers, rather than prompting a fundamental turnaround in recent trends.
• Over the past three years, the largest proportion of the increase in the stock of working-age inactivity has been people citing long-term sickness or disability as their main reason for inactivity. This group has increased by 350,000 to 2.5 million in the fourth quarter of 2022 and now accounts for 28 per cent of the total number of working-age inactive and 6 per cent of the working-age population. Reported rates of long-term sickness and disability had been rising among both the active and inactive working-age populations since at least 2010. The proportion of adults recorded as disabled rose from 16 per cent in 2013 to 20 per cent in 2019 and then to 22 per cent in 2022. But since the beginning of the pandemic the steady rise in participation among this group has stalled at around 57 per cent (having risen from 50 per cent in early 2013 to 58 per cent at the end of 2019).^e Flows data suggest much of the increase in long-term sick came from those who were already inactive before the pandemic changing their reason for being inactive during it. Other evidence shows the change in the stock of the inactive due to long-term sickness mainly comprises those who have been out of employment for over three years.^f

a) Here, we define working-age as those aged 16 to 64, to align with inactivity data published in the Labour Force Survey.

b) Data differences also complicate interpretation: unlike the cross-sectional stocks data, the longitudinal flows data exclude non-respondents (rather than simply rolling forward previous quarters’ responses), and do not capture population change.

c) During the pandemic, fewer students working part time and A-level grading changes also had temporary effects.

d) Low Pay Commission, How has the increase in economic inactivity affected low-paying occupations?, January 2023.

e) The rise in participation before the pandemic was largely driven by rising employment within different categories of disability rather than a shift in the disabled population towards higher-employment conditions.

f) Institute for Employment Studies, The Commission on the Future of Employment Support, November 2022.

2.30 Judging how much of the recent rise in inactivity reflects temporary factors, and so is likely to reverse, and how much reflects more persistent or permanent factors presents a very significant challenge. This is overlaid by uncertainty over the labour supply impact of measures announced in this Budget (see Box 2.2). Given these uncertainties around the outlook for inactivity in our forecast, we have produced the three scenarios for the trend participation rate shown in Chart 2.9, rather than simply focusing on presenting a single central forecast. Compared to the latest actual participation rate outturn of 63.2 per cent in the fourth quarter of 2022:

• In the upside scenario , the trend activity rate recovers quickly and ultimately rises to 63.9 per cent in 2027, 0.9 percentage points (around 500,000 people) above our central forecast, and 0.4 percentage points (around 250,000 people)^[8] above our pre-pandemic forecast. This rise in participation could occur if the number of those classed as long-term sick falls steadily as Covid-related morbidity recedes, as the NHS backlog of elective health treatment is cleared in line with the Government’s ambitions, and, crucially, if this improvement in health outcomes feeds through to a recovery in participation rates in the labour force. But higher participation could also reflect other factors, such as early retirees re-entering the labour force due to the higher cost of living.^[9] I Around 130,000 of the increase in participation in this scenario, relative to our central forecast, could also reflect the policies in this Budget having a larger positive effect on the labour force than our central forecast assumes. As described in Box 2.2, a plausible upside scenario for the impact of policy is that it brings 240,000 back into the workforce.
• In our central forecast the trend participation rate falls slightly in the next two years before rising back to 63.0 per cent in 2027, the same rate as in November. Our pre-measures forecast anticipated a more sustained fall in activity largely attributable to the influences of an ageing population with a partial offset from the rise in the State Pension age (see paragraph 2.32). But our central estimate of the impact of the Budget policy measures is sufficient to arrest that decline, such that the trend participation rate rises from 2025 onwards. Compared to our pre-pandemic forecast, the shortfall in the trend participation rate narrows from 0.7 percentage points (around 380,000 people) in 2022 to 0.4 percentage points (around 250,000 people) in 2027.
• In the downside scenario , the trend participation rate continues to fall to 62.1 per cent in 2027, 0.9 percentage points (around 500,000 people) below our central forecast of 63.0 per cent, and 1.3 percentage points (around 750,000 people) below our pre-pandemic forecast. This would bring the participation rate to its lowest level since 1983. For this to occur, inactivity would have to continue rising at broadly the rate it has since the onset of the pandemic, despite the upward pressure on activity rates from the impending rise in the State Pension age to 67 by 2028. This could reflect demographics continuing to weigh on participation, a continuation of the increases in inactivity due to long-term sickness seen since the start of the pandemic, or the rise in early retirement in the past few years reflecting a structural shift rather than a temporary pandemic-induced phenomenon. Part of the fall could also reflect policies in this Budget having a lower-than-expected impact on participation, which could contribute 55,000 as on outlined in Box 2.2. Overall, we judge this to be less likely than the upside scenario.

Chart 2.9: Trend participation rate of those aged 16 and over

2.31 In the upside scenario, a larger labour force increases average real GDP growth by 0.3 percentage points a year, taking the level 1½ per cent higher by 2027 compared to our central forecast. Shortfalls in the downside scenario are broadly symmetric. These figures reflect stylised assumptions: that the equilibrium unemployment rate, average hours and productivity per hour are unchanged, and that real GDP moves in line with the changes in potential output so there are no cyclical effects of a larger or smaller labour force. The fiscal implications of these scenarios are discussed in Chapter 5.

2.32 Chart 2.10 provides an illustrative decomposition of our central forecast for labour participation by age group that reflects demographic factors, an unwind of some of the recent rise in age-group-specific inactivity rates, and the effects of the labour supply policies in this Budget. The ongoing ageing of the UK population will put significant upward pressure on inactivity over the forecast period.^[10] Between 2022 and 2027, the population aged 16 years and over is expected to increase by close to 1.9 million, with the population aged over 65 rising by 1.2 million. This group has the highest inactivity rate and, while age-specific inactivity rates may fall slightly as the State Pension age rises to 67 by 2028, this will be more than offset by the scale of the demographics-driven rise in number of people aged 65 and over. This group therefore accounts for the bulk of the anticipated increase in the inactive population.

2.33 In contrast, the number of inactive 25-to-49- and 50-to-64-year-olds is set to fall due to weaker population growth in those age groups and some reversion to the longstanding decline in their inactivity rates. For the former group, there may be scope for even further falls if, for example, the childcare measures announced in the Budget have a greater-than-expected impact. For the older group, the fall in inactivity would be consistent with some unwinding of the recent rise in long-term sickness and, from 2025, declines in the size of this cohort as the baby boom generation continues to age into the 65 and over group. A modest rise in inactive 16-to-24-year-olds reflects a gradual rise in students entering higher education, in line with similar assumptions in our student loans forecast, and the effect of demographic factors.

Chart 2.10: Illustrative decomposition of inactivity forecast by age group

Labour productivity

Business investment and capital deepening

2.34 Cumulative business investment from 2022 to the end of the forecast period is little changed from our November forecast (down 0.5 per cent). But as shown in Chart 2.11, there is a very uneven profile of revisions to outturn data and to our forecast which reflects the impact of several new factors:

• Substantial upward revisions to recent outturn data have increased its level in the final quarter of 2022 by 9 per cent compared to our November forecast. We assume that some of this unexpected strength unwinds and business investment falls by 5 per cent in the first quarter of 2023, a 4 percentage point sharper fall than in November.
• We have revised our assessment of the impact of the business cycle on investment. Over the next year, we continue to expect the weak economic outlook to weigh on corporate profitability and investment decisions, with higher interest rates increasing the cost of capital and depressing investment. But a reappraisal of our past forecasts suggested that we had been overestimating the extent to which the recovery in economic activity would support business investment.^[11] Revising our assessment of this effect lowered our pre-policy measures forecast for business investment, which settled at 10 per cent of GDP (its average level over 2012 to 2019), rather than rising more steeply as share of GDP as we expected in November.
• Capital allowances , including new and previously announced measures, have a significant effect on the profile of business investment. We expect the new temporary investment allowances announced in the Budget to incentivise firms to bring forward capital spending, raising investment by over 3 per cent when the scheme’s effect is at its peak (see Box 2.2). This new measure follows on immediately from the withdrawal of the existing temporary super-deduction in April 2023. It should therefore help ensure that firms’ incentives to invest do not fall more significantly in the near term, although investment will nonetheless be lower than it would otherwise have been from April 2026, when the latest temporary incentive is withdrawn.^[12] The Chancellor has indicated his intention to make the measure permanent when economic and fiscal conditions allow. Were this to happen it would remove the incentive to bring forward investment into the next three years, but it would also mean a more sustained higher flow of investment. The net difference in the size of the capital stock within our five-year forecast period might therefore be small. But a permanent measure would provide a sustained boost to investment, the capital stock, and productive capacity, though one that would come at considerable additional fiscal cost.
• The main rate of corporation tax rises from 19 to 25 per cent in April 2023. This measure was announced in the March 2021 Budget (see Box 2.3 of the March 2021 Economic and fiscal outlook (EFO)) and has been reflected in our business investment forecast since then. The economic effect of the tax rise is to raise the cost to firms of acquiring capital, and so weighs on investment as capital stocks adjust to their new desired levels. Our latest forecast assumes that the rise in the headline rate in isolation would reduce the level of business investment in 2027 by around 2 per cent.

2.35 As a result of this broadly similar forecast for cumulative business investment, the size of the capital stock is little different at the forecast horizon compared to November. However, the bringing forward of investment results in faster capital stock growth in the first half of the forecast and then weaker growth for later years. Growth in the capital stock per worker, which matters for our potential output forecast and is also influenced by our judgements about growth in labour supply, is expected to be 0.5 percentage points lower than in November at the forecast horizon. This lowers its contribution to productivity growth (capital deepening) at the forecast horizon, which now settles at ¼ percentage points in the final two years of the forecast compared to 0.4 percentage points in November.

Chart 2.11: Business investment

Total factor productivity

2.36 We have slightly increased our forecast for total factor productivity to reflect lower medium-term gas prices, so TFP now contributes 0.8 percentage points to potential output growth from 2023 onwards, 0.1 percentage points higher than in November. By the forecast horizon, stronger TFP growth adds a third of a per cent to the level of potential output, reflecting our estimate that each 10 per cent increase in medium-term gas prices takes around 0.05 per cent off the level of medium-term potential output.^[13]
Medium-term gas prices have fallen sharply since November and are now 85 per cent higher than the pre-invasion medium-term price expectations of around 70p a therm (embodied in our October 2021 forecast) compared to 170 per cent higher in the assumptions that underpinned our November forecast. This halving in the medium-term energy price shock relative to our November forecast raises productivity by increasing the volume of goods and services that it is profitable for firms to supply when the price of an important, largely imported, input falls.

The output gap and real GDP

Output gap

2.37 We judge there to have been a larger positive starting output gap than assumed in November, at around 1 per cent in the third quarter of 2022 compared to half a per cent in our previous forecast (Chart 2.12). Our suite of output gap models suggests that there was still significant excess demand in the economy, particularly in the labour market where vacancies were over 1.2 million and surveys of recruitment difficulties and capacity utilisation indicated significant tightness. We expect this excess demand to fall away and spare capacity to open up through 2023 as supply bottlenecks continue to ease and consumer demand weakens in the face of falls in real wages and higher interest rates. A higher starting output gap, lower market-implied interest rates, and lower gas prices – and therefore a smaller hit to real wages – mean the output gap falls by less and recovers more quickly than in our November forecast. The negative output gap troughs at 1.6 per cent in the fourth quarter of 2023, compared to a trough of 3.1 per cent in our November forecast, before closing in 2027 as real wages recover and interest rates fall.

Chart 2.12: Output gap

Real GDP

2.38 The first vintage of ONS data suggest that the economy narrowly avoided a technical recession in the second half of 2022 as real GDP fell by 0.2 per cent in the third quarter, but was flat in the fourth quarter.^[14] On our central forecast, the economy contracts again in the first quarter of 2023 by 0.4 per cent and GDP is flat in the second quarter as the rise in interest rates, the elevated cost of energy, and an additional bank holiday for the Coronation in May depress output.^[15] We expect a milder fall in GDP than our November forecast as a result of lower expectations for wholesale gas prices and interest rates. Growth returns in the second half of 2023 due to the bounce back in activity from the bank holiday and as household energy bills fall. In 2023 as a whole, real GDP falls 0.2 per cent, with private consumption, business investment and net trade all dragging on growth, offset by growth in government consumption (Chart 2.13).

2.39 GDP growth picks up to 1.8 per cent in 2024 and 2.5 per cent in 2025 as interest rates start to fall and drops in energy and other tradeable goods prices take inflation below the 2 per cent target. The recovery is driven by private consumption growth as real household incomes rise, with business investment also boosting growth in 2025, supported by the temporary increase in the generosity of capital allowances. Real GDP growth then returns towards the assumed medium-term rate of potential output growth of 1¾ per cent as the output gap closes in the final year of the forecast. Cumulative real GDP growth over the five-year forecast is around half a percentage point higher than in our November forecast, broadly in line with the change in our forecast for potential output growth. The upgrade to real GDP at the forecast horizon is driven by higher real consumption as a result of stronger real household incomes.

Chart 2.13: Contributions to real GDP growth

2.40 Chart 2.14 shows our central forecast for real GDP growth within a fan chart that illustrates the probability of a range of possible outcomes conditioned on the size of past errors in official forecasts. Following large forecast errors during the pandemic, the range of potential outcomes has widened significantly compared to our pre-pandemic fan charts. The chart implies that there is a roughly 50 per cent probability that real GDP growth will be positive in 2023 and around a 30 per cent probability that it will be negative in 2024. But it should be emphasised that past forecast errors may not be a guide for future economic shocks. For this reason, in Chapter 5 we also present three sets of alternative upside and downside scenarios for labour participation, energy prices, and interest rates, in order to illustrate further the uncertainty around our assessment of the Government’s fiscal targets.

Chart 2.14: Real GDP growth forecast

Labour market

Unemployment and employment

2.41 The unemployment rate is expected to rise modestly as growth weakens, to a peak of 4.4 per cent (around 1.5 million people). That is 170,000 people lower than our November forecast for the peak rate of 4.9 per cent (left panel of Chart 2.15). This is due to the improved outlook for real GDP in the near term, partly due to the impact of Budget measures, which reduce the peak in unemployment by 0.1 percentage points (around 25,000 people). The labour market started to show signs of loosening in late 2022, with recruitment difficulties starting to ease and the unemployment rate ticking up to 3.7 per cent in the fourth quarter. Vacancies have also begun to unwind from their record highs in 2022 but remain elevated (at 1.1 million in the three months to January 2023) compared to their pre-pandemic average of around 635,000. We continue to expect the rise in unemployment to lag the fall in GDP as vacancies are likely to fall before firms lay off workers and labour hoarding means firms are likely to reduce hours before cutting staff. By the forecast horizon, unemployment falls back to its unchanged estimated structural rate of 4.1 per cent.

Chart 2.15: Unemployment and employment rates

2.42 As GDP growth slows and unemployment rises, we expect the employment rate to fall from 60.8 per cent in the fourth quarter of 2022 to a trough of 60.0 per cent in the second quarter of 2024, before recovering to 60.4 per cent at the forecast horizon, which is almost the same as in November (right panel of Chart 2.15). Employment was 32.8 million in the fourth quarter of 2022, 120,000 (0.4 per cent) below its pre-pandemic level. It falls a further 75,000 (0.2 per cent) over the next year as the economy contracts and unemployment rises, before recovering to reach 33.8 million at the forecast horizon, around 900,000 higher than the pre-pandemic level.

2.43 Employment growth over the forecast is 120,000 higher than our November forecast as a result of partly offsetting changes: higher net migration adds 160,000; our lower pre-measures participation rate forecast subtracts 150,000; and the measures in this Budget add a further 110,000 workers (as described in Box 2.2).

Earnings

2.44 We expect stronger near-term growth in nominal earnings of 5.0 per cent in 2023, 0.9 percentage points higher than our November forecast, but still a slowdown from the 30-year high of 6.2 per cent in 2022. The strong whole economy pay growth in 2022 largely reflected private sector pay, which rose by close to 7 per cent, while for the public sector the increase was around 2.5 per cent. The latest Bank of England Agents’ pay survey shows private sector pay expectations remain close to 2022 levels at just under 6 per cent, while the labour market remains relatively tight and recruitment difficulties elevated. Another survey reports that 57 per cent of employers struggled to fill vacancies and a third of them planned to address this by raising wages, although this is down from just under half six months earlier.^[16]
Rising unemployment is expected to weigh on pay growth later this year.

2.45 In 2024, we expect nominal pay growth to ease to 1.8 per cent as inflation falls away and unemployment remains above its equilibrium rate, before averaging 2.0 per cent over the remainder of the forecast, broadly in line with our November forecast. Despite strong nominal earnings growth this year, high inflation means real earnings are broadly flat. Real pay growth then rises to 0.7 per cent in 2024 before averaging around 1 per cent a year over the rest of the forecast.

Households

Household income

2.46 Real household disposable income (RHDI) is expected to fall by 2.6 per cent in 2023, following a fall of 2.5 per cent in 2022 (Chart 2.16). Inflation contributes 4.9 percentage points in 2023 (blue bars), continuing to outstrip the nominal earnings contribution of 3.6 percentage points (green bars) while the contribution of other income also falls 1.3 percentage points (purple bars). RHDI would have fallen by a further 1½ per cent in 2023 were it not for the EPG taking 2 percentage points off CPI inflation this year. And a rise in income and wealth tax payments is broadly offset by a rise in social benefits. RHDI returns to growth in 2024, rising by 1.7 per cent as inflation eases significantly and falls below growth in nominal earnings and non-labour income. RHDI continues to recover in the final three years of the forecast, growing by an average of around 2 per cent. Nominal earnings growth and non-labour incomes add around 3.7 percentage points to annual RHDI growth between 2025 and 2027, while net taxes on households subtract 0.6 percentage points as frozen tax thresholds draw more people into tax and more taxpayers into higher bands.

Chart 2.16: Contributions to real household disposable income growth

2.47 On a fiscal-year basis, RHDI per person, a measure of living standards, falls by 6 per cent between 2021-22 and 2023-24 (Chart 2.17). Lower inflation (reflecting lower imported commodity prices which improve the terms of trade) and stronger nominal earnings growth mean that this is a 1 percentage point smaller fall than expected in November, but it would still represent the largest two-year fall in real living standards since ONS records started in the 1950s. The fall between 2021-22 and 2023-24 takes RHDI per person to its lowest level since 2014-15. And in 2027-28, real living standards remain around ½ per cent below pre-pandemic levels. As discussed above, this reflects the fact that the UK has suffered a large terms of trade shock, which increased the prices of the goods and services we consume significantly more than the prices of the goods and services we produce.

Chart 2.17: Real household disposable income per person

Household saving

2.48 Households’ saving (excluding adjustments to net equity in pension funds) is expected to drop to around zero in 2023 and 2024 to support consumption in the face of weak real income growth. Restrictions placed on personal mobility and social consumption during the pandemic led to a build-up in total household savings, boosting household wealth. We expect savings to be drawn upon to stabilise spending while inflation remains high, and, as a result, for wealth relative to household income to return to historical averages (Chart 2.18). As cost-of-living pressures ease, the saving ratio rises to around 1 per cent at the forecast horizon, only marginally higher than our November forecast but still around half the post-financial crisis average. The headline saving ratio (which includes adjustments to net equity in pension funds) falls to 6.3 per cent in 2023 before gradually rising to just below 8 per cent in 2027, 2½ percentage points higher than in November due to larger-than-expected pension equity adjustments in recent outturn data.

Chart 2.18: Household wealth and saving

Private consumption

2.49 Private consumption is forecast to fall by 0.8 per cent in 2023 as the drop in real incomes is only partly offset by lower saving. This fall would be 1½ percentage points larger if the household saving ratio were to remain at 2022 levels, in the absence of stronger incomes. Consumption growth picks up to average 1.7 per cent a year over the rest of the forecast as real income growth turns positive and household saving remains low. Compared to our November assessment, real private consumption is higher over the whole forecast period and by around 1 per cent at the forecast horizon due to lower market expectations for wholesale gas prices and the upgrade in potential output boosting household incomes.

Credit conditions and housing

2.50 Average interest rates on the stock of outstanding mortgages are expected to peak at 4.2 per cent in 2027, two times higher than at their trough at the end of 2021. But the peak is 0.8 percentage points lower than forecast in November given lower market expectations for Bank Rate (left panel Chart 2.19). With more than 80 per cent of mortgages on fixed-term contracts and the prevalence of fixed-rate mortgage contracts with terms of more than two years having risen in recent years, the increase in rates on new mortgages over recent months will take several years to feed through to the average mortgage rate.^[17]

2.51 Our central forecast is that house prices fall by 10 per cent from their high in the fourth quarter of 2022, a 1 percentage point larger fall than in our November forecast (right panel of Chart 2.19). Property transactions are expected to drop by 20 per cent relative to their peak in the same quarter. Leading indicators from Halifax and Nationwide suggest that house prices have already fallen by 3 to 6 per cent between their peak in the middle of 2022 and February 2023. Low consumer confidence, the squeeze on real incomes, and the expectation of mortgage rate rises to come are expected to contribute to continued falls in house prices and a reduction in housing market activity.

Chart 2.19: Mortgage rates and house prices

Trade and the current account

2.52 We expect trade volumes to contract in the near term as the economy stagnates and growth in the UK’s major export markets weakens. In 2023, imports fall by 4 per cent, dragged down by lower consumption and investment, while exports fall by 6.6 per cent. Import and export volumes continue to decline in 2024, by 1.3 and 0.3 per cent respectively. Exports return to growth from 2025 onwards while import volumes continue to fall, partly due to the fall in import-intensive components of business investment as the temporary capital allowance measure ends. Weak growth in imports and exports over the medium term partly reflect the continuing impact of Brexit, which we expect to reduce the overall trade intensity of the UK economy by 15 per cent in the long term. Box 2.4 examines the latest evidence for this assumption – with its subsequent impact on productivity – alongside those made about investment and migration following the 2016 EU referendum.

Box 2.4: How are our Brexit forecasting assumptions performing?

Since the June 2016 EU referendum, our forecasts have incorporated a set of assumptions about the economic impact of Brexit. We have reviewed and refined them in subsequent forecasts as new evidence has arrived. This box assesses our current assumptions against the latest evidence:

• On trade and productivity , we assume that the volume of UK imports and exports will both be 15 per cent lower in the long run than if we had remained in the EU, reducing the overall trade intensity of GDP. And we assume that this leads to a 4 per cent reduction in the potential productivity of the UK economy. The reduction builds over time with the full effect felt after 15 years.
• On investment , we assumed that greater uncertainty from the result of the referendum would see some investment projects postponed or cancelled. In our March 2020 EFO, we estimated that this had lowered productivity by 1½ per cent as a result of a lower capital stock. But we expected that shortfall to fade as uncertainty over the future trading relationship receded and investment recovered.
• On migration , we initially assumed that migration would be lower than otherwise. Once details of the UK’s post-Brexit migration regime became available, we assumed that it would reduce net migration from the EU, but that non-EU flows would be higher. We initially assumed this would reduce net inward migration from 200-300,000 a year in the 2010s to 129,000 over the medium term. Since the new rules came into effect in January 2021, we have revised this estimate of steady-state migration levels up to 205,000 in our November 2022 EFO and to 245,000 in this EFO as the new regime appears to have led to significantly higher levels of non-EU immigration than we originally assumed.

Trade

The Trade and Cooperation Agreement (TCA) was concluded in December 2020 and set the terms of the post-Brexit trading relationship between the UK and EU.^a The current picture is slightly clouded by data measurement issues,^b but two years into the TCA’s implementation, the trends in UK trade volumes remain consistent with our Brexit assumptions.

All major advanced economies experienced a collapse in trade during the pandemic, but the latest (adjusted ONS) data suggest that UK trade volumes remain 3.0 per cent below their 2019 level in the third quarter of 2022, versus an average increase across other G7 countries of 5.5 per cent. Trade intensity (adjusted) is 2.6 per cent lower than its pre-pandemic level in the UK in the third quarter of 2022 but 3.6 per cent above in the rest of the G7 (Chart E).

A recent study that estimates the trade impact of Brexit by comparing UK trade performance to a weighted average of similar countries (a ‘doppelgänger’) suggests that UK goods trade was 7 per cent lower in June 2022 than it would have been had we remained in the EU, but finds less evidence of an impact on trade in services.^c

The Windsor Framework has also now been published and will change the requirements for goods trade between Great Britain and Northern Ireland. As it relates to trade flows within the UK, we do not expect it to affect our forecasts for UK-EU trade flows. But the agreement should help support the effective functioning of the Trade and Cooperation Agreement that is assumed in our central forecast, mitigating a potential source of downside risk.

Chart E: UK trade performance compared to other G7 countries

Investment

Consistent with our initial Brexit assumptions, business investment growth stalled in the years after the EU referendum, which will have partly been due to increased uncertainty about our future trading relationships, as well as diversion of resources from more productive uses towards Brexit preparations. Business investment has stagnated in real terms for much of the period since 2016, such that on the eve of the pandemic it stood 16.2 per cent below our pre-referendum expectations (relative to its level in the second quarter of 2016) (left panel of Chart F). More recently, the pandemic and the increase in global energy prices have also weighed on investment. But in the face of these global shocks, UK non-dwellings investment has continued to underperform relative to other G7 countries (right panel of Chart F).

Chart F: Investment since the EU referendum

Migration

Since the introduction of the new post-Brexit migration regime, net inward migration has significantly exceeded the levels assumed in our forecasts. In the year to June 2022, the ONS estimates that 504,000 long-term net migrants arrived in the UK,^d twice the average numbers seen between 2010 and 2020 and more than triple our initial assumption for the long-run levels of net migration under the new visa regime. While net EU migration is now negative at minus 51,000, this has been more than offset by a large increase in non-EU net migration to 509,000. It is too early to tell how much of these higher net migration flows will be sustained or how many of the new migrants will ultimately enter the workforce. In particular, some of the recent jump is likely to reflect temporary factors (as discussed in paragraph 2.24) that could subsequently be reversed. This EFO incorporates the latest January 2023 ONS population projection variant such that net migration declines and then stabilises at 245,000 a year from 2026-27 onwards.

Conclusion

Overall, while net migration has been higher than we anticipated, investment growth has been significantly weaker than we expected before the referendum, and our assumption about the impact of Brexit on the UK’s trade intensity is broadly on track. As a result, we have not revised our view that productivity will be 4 per cent lower in the long run than if the UK had remained in the EU. And we will continue to monitor our conclusions against the emerging evidence.

a) The EU implemented full customs and regulatory controls on imports from the UK in January 2021, while the UK delayed full customs controls until January 2022 and has announced that some health and security checks will now not come into effect before the end of 2023.

b) Changes in data collection methods in this period may have led to some overstatement of the volume of EU imports in 2022. See ONS, Impact of trade in goods data collection changes on UK trade statistics: update on Staged Customs Controls, February 2023 and pages 85 and 86 of Bank of England, Monetary policy report, February 2023. Based on Bank of England analysis, we estimate that the level of total EU goods trade was overstated by 16 per cent in the first 11 months of 2022 (relative to its late 2020 level).

c) Springford J., The cost of Brexit to June 2022, December 2022.

d) ONS, International migration, provisional: year ending June 2022, November 2022.

2.53 With imports falling by more than exports, we expect the current account deficit to narrow from just above 6 per cent of GDP in 2023, its widest since the 1940s, to reach 2½ per cent of GDP in 2027 (Chart 2.20). The widening of the current account deficit in 2022 was mainly driven by a larger trade deficit due to a rapid deterioration in the terms of trade as the price of energy, food and tradable goods rose significantly. For example, the value of gas imports was £49 billion in 2022, two-and-a-half times higher than in 2021. As energy prices fall, we expect the trade deficit to narrow from 3.9 per cent of GDP in 2023 to 0.9 per cent in 2027. The net investment income and transfers deficits narrow gradually over the forecast.

Chart 2.20: Current account deficit

Sectoral net lending

2.54 The household sector is expected to move from a net lending position in 2022 to balance in 2023 as savings are drawn down to support consumption during the cost-of-living squeeze. During the pandemic and energy crisis, fiscal support to the economy resulted in a large government deficit, the counterpart to which was large surpluses for households and, to a lesser extent, corporations. As fiscal support is withdrawn, the household sector moves from a large surplus to balance before reaching a small surplus at the end of the forecast, supported by a strengthening in incomes. Corporate sector balances have been volatile but turn from a modest surplus in 2023 to balance by the middle of the decade as investment rises gradually over the forecast. Borrowing from the rest of the world, mirrored in a sizeable current account deficit, widens this year before it returns to its longer-run average by the end of the forecast (Chart 2.21).

Chart 2.21: Sectoral net lending

Nominal GDP

2.55 Growth in nominal GDP and its components is the key driver of our forecast of the public finances, which we report on a financial-year basis consistent with fiscal data. We have revised up nominal GDP relative to November, due to higher potential output growth and slightly stronger GDP deflator growth. Within the forecast, nominal GDP growth has been revised up by a cumulative 1½ per cent between 2022-23 and 2025-26 but revised down by 1 per cent in the two years to 2027-28. This leaves the level of nominal GDP in 2027-28 0.8 per cent higher than in November (a difference of £22 billion).

Labour and profit shares of income

2.56 Within nominal GDP, the labour share is expected to rise in the near term as a tight labour market and high inflation drive higher nominal wage settlements and weigh on profits. The share of wages and salaries plus mixed income in nominal GDP rises from 48.0 per cent in 2022-23 to 48.6 per cent in 2023-24 before falling back to around its long-run average of 47 per cent over the remainder of the forecast as firms look to rebuild margins. Mirroring the rise in the labour share, we expect the share of non-financial profits in nominal GDP to fall from above 17 per cent in 2022-23 to its long-run average of just below 17 per cent. Relative to November, we think the labour share will be slightly lower and profit share slightly higher in the near term as weaker headline inflation results in less pressure on profits despite higher wage settlements.

Income and expenditure composition of GDP

2.57 Chart 2.22 shows the changes in both the income and expenditure breakdown of nominal GDP since November:

• On the income side (left panel), both labour income and profits have been revised up materially – with profits (the less tax-rich of the two) revised up proportionately more in the first half of the forecast. By the forecast horizon, both labour income and profits have moved broadly in line with nominal GDP.
• On the expenditure side (right panel), revisions have been dominated by the net trade and other components of GDP that are less important for our fiscal forecasts. But household consumption is higher in every year, boosting several tax bases, whereas the path of corporate investment is dominated by the timing effects of the temporary 100 per cent capital allowances measure announced in the Budget.

Chart 2.22: Nominal GDP: changes since November

Comparison with external forecasters

2.58 Our latest central forecast assumes slightly less inflation and slightly stronger real GDP growth than the average of other external forecasters. In the near term, this difference partly reflects the fall in energy prices since many of these other forecasts were completed. In the medium term, differences relate to our more optimistic assumptions for growth in population, employment, and productivity. In summary:

• Our forecast of 2.9 per cent for CPI inflation in the fourth quarter of 2023 is 0.8 percentage points lower than the latest independent average of 3.7 per cent and 1 percentage point lower than the Bank of England’s forecast of 3.9 per cent. This is likely to reflect the vintage of the gas futures curves used in our forecast as well as the revision to the terms of the EPG announced in the Budget. In 2024, our forecast of 0.5 per cent fourth quarter inflation is 1.9 percentage points below the independent average as energy prices fall. The Bank of England’s latest CPI inflation forecast is higher than ours throughout its forecast period.
• Our forecast for real GDP growth of -0.2 per cent in 2023 is 0.3 percentage points above both the Bank of England and latest independent average (more recently submitted forecasts have slightly stronger growth) (Chart 2.23). In 2024 our forecast of 1.8 per cent is 0.9 percentage points higher than the independent average as we expect a relatively strong recovery in consumption due to falling gas prices, falling inflation and the impact of policy in this Budget. Beyond 2024, our real GDP forecast remains towards the top of the range of external forecasts and is also much more optimistic than the Bank’s, which expects growth of 0.4 per cent in 2025 compared to our forecast of 2.5 per cent. This reflects several factors including: a significant output gap at the end of the Bank’s forecast period, higher household saving, higher gas prices in their earlier forecast, and different assumptions about productivity (we expect growth over the forecast at a slightly faster rate than during the decade following the global financial crisis whereas the Bank assumes that trend productivity growth will be broadly similar to the 2010-2019 average over the 2023 to 2025 period).

Chart 2.23: Comparison of forecasts for real GDP

Chapter 3: Policy measures

Introduction

3.1 This chapter:

• sets out the policy measures announced in this Budget, and in the period since the Autumn Statement in November, including how they have been incorporated in our forecast and the uncertainties around them (from paragraph 3.2);
• provides an update on selected previous measures (from paragraph 3.17); and
• discusses policy risks, which are measures or policy ambitions that are yet to impact our central forecast (from paragraph 3.25).^[18]

Policy announcements in the March 2023 Budget

3.2 Our forecast incorporates the economic and fiscal implications of all policy measures announced since the November 2022 Autumn Statement. Presented with a £24.6 billion-a-year average improvement in the pre-measures outlook for borrowing (described in Chapter 4), the Chancellor uses two-thirds of this on his Budget measures. Policy measures announced since our November forecast raise borrowing in every year, and by an average of £15.6 billion a year from 2023-24 onwards.

3.3 The main policies announced in this Budget fall into five categories (Table 3.1):

• Further energy support measures, including the continuation of the energy price guarantee (EPG) for households at £2,500 for a further three months and extended support for businesses into 2023-24. These cost a total of £4.4 billion in 2023-24.
• A package of measures aimed at increasing labour market participation whose cost rises to £7.1 billion a year by 2027-28. This includes additional help with childcare costs for working parents of younger children, reforms to working-age benefits, disability employment support, and more generous pensions tax allowances.
• Temporary 100 per cent capital allowances for qualifying business investment undertaken between 2023-24 and 2025-26. This costs an average of £9.1 billion a year during those three years, but it raises money in 2027-28 as capital allowance claims drop back as the temporary effect of the measure unwinds.
• Other spending decisions that cost an average of £2.0 billion a year over the forecast period, most importantly an increase in
  defence spending.
• Other tax decisions that cost an average £3.6 billion a year over the forecast period, principally by freezing fuel duty at its current rate for another year. This involves a one-year extension of the temporary 5p cut coupled with a one-year cancellation of its RPI indexation. It costs £4.8 billion in 2023-24 when both elements apply, and £2.6 billion a year thereafter when only the RPI element has an ongoing cost.

3.4 The indirect effects of these policies, which boost demand in the near term and supply in the medium term, reduce borrowing by £3.0 billion next year (thanks in particular to lower inflation) and then by an average of £1.7 billion a year thereafter (as a boost to receipts that builds to £3.9 billion by 2027-28 is partly offset by the costs of servicing the additional debt issued to finance the measures, which rises to £2.6 billion).

Table 3.1: Total effect of Government decisions since November

Energy support measures

3.5 The Chancellor announced further energy bills support to households and businesses for 2023-24. These increase and extend the existing support put in place over the past year (as outlined in Box 3.1). Policy measures announced since November include:

• An increase in the generosity of the energy price guarantee by maintaining the cap at £2,500 for an additional three months. From April 2023, the EPG price cap was due to rise to £3,000. This measure delays that rise until July 2023 and prevents household bills from rising in the first quarter of 2023-24. It costs £2.9 billion, raising the cost of the EPG to £4.0 billion in 2023-24 and £27.0 billion over its lifetime.
• The energy bills discount scheme (EBDS), announced on 9 January 2023 to succeed the energy bill relief scheme (EBRS), which ends on 31 March 2023. The EBDS extends the support available to eligible businesses at a reduced rate. The average modelled unit discount for those on fixed contracts under the EBDS is around 80 per cent less generous than the EBRS. While the use of threshold and reference wholesale prices reduces the volume of energy in scope by around 50 per cent for the universal element of the scheme. For energy- and trade-intensive industries (ETII) there is a more generous unit discount available. The scheme will run for 12 months from this April and is expected to cost £0.5 billion, which is around 90 per cent less than the six-month cost of the EBRS in 2022-23.
• Other energy support measures that cost £0.9 billion and support customers that are not eligible for the EPG or the EBDS. These include fixed payments for those who use alternative fuels for heating and a further discount for domestic customers in properties served by heat networks to achieve the same level of support as the EPG.

Box 3.1: An international comparison of the cost of energy support packages

Most European governments have sought to shield households and businesses from the rises in wholesale and retail energy prices that followed the Russian invasion of Ukraine. This box assesses how the cost of the UK Government’s response compares internationally.

The cost of UK Government support amounts to £78.2 billion across 2022-23 and 2023-24 (1.5 per cent of GDP across the two years or 3.1 per cent of this year’s GDP), of which just over half (£39.9 billion) is recouped via windfall taxes on energy producers over the medium term (Chart A). The main policy components include:

• Energy price subsidies to households and businesses costing £47.0 billion (1.9 per cent of GDP) across this year and next. The energy price guarantee (EPG), announced last September and amended in this Budget, accounts for almost three-fifths of this total (£27.0 billion), while the energy bills support scheme, a £400 discount to households that was announced in two steps in March and May last year, accounts for a quarter (£11.9 billion). The energy bill relief scheme (EBRS), announced in September, performs the same role as the EPG for businesses, costing £7.3 billion. Its successor, the energy bills discount scheme that was announced in January, costs £0.5 billion in 2023-24. The EPG and EBRS both cap retail energy prices, so their costs vary with wholesale energy prices. As those have fallen, the estimated cost of these schemes has fallen too, from a combined £56.0 billion to £34.2 billion.
• Cost-of-living payments to households totalling £25.0 billion (1.0 per cent of GDP). The cost of the initial measures announced last May is £12.0 billion in 2022-23, with £9.0 billion of that coming from one-off payments provided to those on means-tested, pension-age and disability benefits; and £2.9 billion from the £150 council tax rebate that most households received in April 2022. Further one-off payments and other smaller measures announced in November’s Autumn Statement cost £12.3 billion in 2023-24.
• Other energy-related support of £7.8 billion (0.3 of GDP), including the 5p cut in fuel duty announced last March and extended for a second year in this Budget, costing a combined £4.8 billion over 2022-23 and 2023-24. Fuel duty is also frozen in both years rather than rising in line with RPI inflation, but since this has happened every year since 2011 it is not counted as a specific response to the energy price crisis. We expect the bailout of Bulb Energy to cost a further £3.0 billion (see Box 4.1).
• Windfall taxes on energy producers that yield £39.9 billion (1.5 per cent of GDP) across the forecast, with £14.8 billion of that in 2022-23 and 2023-24. The temporary energy profits levy, announced in May last year, then raised and extended in November’s Autumn Statement, raises the tax rate on North Sea profits from 40 to 75 per cent until 2027-28. It yields an average of £4.3 billion a year. The temporary electricity generator levy – a tax on excess revenues until 2027-28 – is expected to raise £2.3 billion a year from non-gas generators that benefit from gas-related rises in electricity prices.

Chart A: The net cost of the UK Government’s energy and cost-of-living support

Analysis by the IMF shows that the gross size of the UK Government’s energy support package relative to GDP is one of the highest in Europe (the columns in Chart B). The IMF figures incorporate some measures that we have not categorised as energy price support,^a and do not net off revenue from windfall taxes, so put the two-year cost at 4.4 per cent of GDP. But this is lower than the 5.2 per cent that this figure would have been in November thanks largely to lower energy prices. The average cost of support in other European countries is significantly lower, at 2.6 per cent of GDP. It includes:

• direct interventions, such as limits on retail energy price increases, the largest component accounting for just under half of the average overall cost;
• support for businesses, which represents almost a third of the average overall cost, though there is significantly variation across countries;
• energy-related tax cuts, which make up around a sixth of the cost across Europe; and
• cash transfers, which make up around an eighth of the cost.

One of the reasons for the UK’s high level of fiscal support is its reliance on natural gas for final energy consumption (the diamonds in Chart B). The share of natural gas in gross inland energy consumption in the UK, at 36.6 per cent, is around 1½ times the European average of 23.9 per cent. Natural gas also largely drives UK wholesale electricity prices, while limited storage capacity leaves the UK more exposed to fluctuations in energy market prices.

Chart B: European countries’ size of energy support packages in 2022 and 2023 and the share of natural gas in final energy consumption

a) The UK estimates in Chart B are based on our own figures but including the same types of measures that the IMF has included for other countries so that they are on a like-for-like basis with other countries. This means that our latest March 2023 estimate is materially higher than that in Chart A, primarily because the latter excludes the cost of abolishing the health and social care levy, that we do not consider to constitute energy price support.

Increasing labour market participation

3.6 The Government is spending amounts rising to £7.1 billion a year by 2027-28 on a set of measures aimed at increasing labour market participation. The majority of this is spent on childcare measures, at a cost of £5.3 billion a year by 2027-28. The package also includes disability and welfare measures (£0.6 billion) and raising various pensions tax allowances (£1.2 billion). We expect these policies, in aggregate, to raise employment and potential output – adding 110,000 to employment and increasing GDP by 0.2 per cent by 2027-28 (as described in Box 2.2 in Chapter 2). This boosts receipts by £1.5 billion and lowers welfare spending by £0.2 billion in 2027-28, but the debt interest paid on the cumulative cost of the measures reaches £0.7 billion by 2027-28.^[19]

3.7 Taking the measures in turn:

• Childcare measures are the largest part of the total package. These include 30 hours a week of free childcare for working parents of nine-month- to two-year-olds, which is estimated to cost £4.9 billion in 2027-28. We expect this policy to raise employment by 60,000 by 2027-28, as well as raising the hours worked by mothers already in work. The package also includes increasing the cost cap and addressing upfront childcare costs in universal credit. These cost £0.1 billion by 2027-28, and the latter increases employment by 15,000 by the end of the forecast period.
• Disability and welfare measures include policies to support employment for people with disabilities, including additional work coach time for incapacity benefit claimants and a new disability employment programme, which provides intensive employment support, placements and training for economically inactive disabled individuals. We estimate that the latter will increase employment by 10,000 by 2027-28. This package also includes a number of health measures, including a digital mental health hub and increasing access to psychological therapies for which we assume no additional employment impact. Finally, the Government is introducing reforms to the welfare system, including changes to the additional earnings threshold and greater conditionality for parents, with the latter increasing employment by up to 10,000.
• Pensions tax allowances measures increase the amount that can be saved tax-free in a pension over an individual’s lifetime and in any given year. We expect these to raise employment by 15,000. These measures are covered in more detail below.

3.8 Other labour market measures include policies aimed at raising skills, such as sector-based work academy programmes and skills bootcamps. We were not able to model the full implications of the Budget childcare measures on welfare spending using DWP’s models due to restrictions on the sharing of Budget-sensitive information between departments. It is therefore likely that we will need to revise some assumptions underpinning this forecast when we produce our next one. This relates in particular to our estimate of the welfare savings from the free childcare hours policy. Parents who move into work due to this offer will include existing benefit recipients, whose benefit awards will fall as household earnings rise. There is also a small saving from those parents that switch from universal credit childcare support. Together, these are estimated to reduce welfare spending by £35 million in 2024-25, when the policy begins, rising to a saving of £100 million in 2027-28, once it has been fully rolled out. Any revisions we need to make to these estimates in our next forecast are therefore unlikely to be fiscally material.

Pensions tax allowances

3.9 The Chancellor has announced a package of measures relaxing the amount of pensions savings that individuals can accumulate before they incur tax charges, largely reversing earlier decisions to restrict those amounts. The main changes are:

• The lifetime allowance (LTA) will be abolished from April 2024, with the LTA charge removed from April 2023. The LTA is the total amount of tax-relieved contributions that an individual can accumulate in their pension pots before incurring a tax charge. It was £1.8 million in 2010-11 before a succession of policy measures brought it down to £1 million in 2016-17. It is set at £1,073,100 for 2022-23. The number of LTA charges paid by schemes through HMRC’s accounting for tax (AfT) return rose from around 1,000 in 2010-11 to nearly 9,000 in 2020-21, with the value of charges rising from around £40 million to close to £400 million.
• The annual allowance (AA) is to be increased from its current £40,000 to £60,000 from April 2023. The AA is the amount an individual can contribute each year to pension pots before contributions are subject to a tax charge. This measure also partly reverses earlier decisions to reduce the AA, which was £255,000 in 2010-11, cut to £50,000 in 2011-12, and then cut again to the current £40,000 in 2014-15. There were a negligible number of AA charges in 2010-11, but by 2020-21 there were just over 17,000 reported through AfT and around 41,000 through self-assessment, with the value of charges estimated at around £300 million.^[20]
  The AA taper, introduced in 2016-17, reduces the amount of AA that is available to higher earners, and has been identified as a key driver behind the rapid recent rise in charges, including from higher-paid practitioners in the NHS. This Budget also relaxes rules around the taper.^[21]
• An increase in the money purchase annual allowance – the annual limit on the amount of tax-free contributions that can be made into a defined contribution pension where amounts have previously been flexibly withdrawn – from £4,000 to £10,000, from April 2023.
• A change in the treatment of negative pension input amount
  (PIA) that enables members of different pension schemes to offset a negative PIA in one scheme against a positive amount in another.^[22]

3.10 The combined cost of the measures rises to an average of £1.1 billion a year during the final three years of the forecast. The main uncertainty in these costings relates to the behavioural response – the degree to which individuals and employers in the private sector increase their pension contributions and the numbers of (especially) public sector workers that choose to opt in to (or not opt out of) pension schemes. A second uncertainty relates to our baseline AfT forecast, particularly the extent to which the rapid growth in receipts since 2016-17 will continue. We have also assumed that the relaxation of these rules will result in higher employment among those currently incentivised to retire because they can no longer save into their pensions without incurring tax charges – another source of uncertainty.

Capital allowances: 100 per cent full-expensing for three years

3.11 The Government has announced a temporary 100 per cent capital allowance (or ‘full-expensing’) regime will be in place for three years from April 2023 – and that it will look to make this permanent when economic and fiscal conditions allow. This three-year boost to the generosity of investment allowances follows on from the even more generous two-year boost from the ‘super-deduction’ that ends this month. It means that all investment in new plant and machinery that qualifies as a ‘main rate’ asset can be written off against taxable profit in the year the cost is incurred. Under the super-deduction, this figure was 130 per cent. The amount of investment that can benefit from both measures is uncapped.

3.12 The measure is designed to promote business investment, and the extent of this dynamic response is the main uncertainty in the costing. As a temporary measure, it provides a strong incentive to bring forward investment that is currently planned for a later date, to benefit from its generosity. We expect it to raise business investment in the three years the measure is in place by around 3 per cent when the scheme’s effect is at its peak, equivalent to around £6 billion a year (Chart 3.1). This is less than the 10 per cent increase we initially assumed for the super-deduction. But as we explained in our January 2023 Forecast evaluation report, we now think that the super-deduction increased business investment by around 5 per cent, and that we overestimated its dynamic benefits. This measure is less generous than its predecessor, and spread across three years rather than two, hence its smaller peak impact on investment.^[23]

Chart 3.1: Three-year full expensing: impact on business investment

3.13 As a temporary measure, we have assumed that the Budget measure has no long-run impact on the capital stock, and that all the additional investment is ultimately displaced from future years (some of which lie beyond our five-year forecast horizon). As described in Chapter 2, this would be different if the Chancellor were able to follow through on his intention to make the measure permanent, in which case investment would be higher for a sustained period until the capital stock had risen to a level consistent with the lower post-tax cost of capital. But this could cost amounts approaching £10 billion a year.

3.14 The temporary measure that has been announced costs £8.0 billion in 2023-24, £10.7 billion in 2024-25 and £8.7 billion in 2025-26. Full expensing also encourages businesses to accelerate their capital allowance claims which, together with its effect on the time profile of investment, means that it actually raises
receipts by £2.2 billion in 2027-28 (and for perhaps a further decade beyond the forecast horizon). This is because the bringing forward of both claims and investment means fewer capital allowances to offset against taxable profits once the temporary measure has ended.

Measures with highly uncertain costings

3.15 As detailed in Annex B, we assign an uncertainty rating to all certified policy costings.^[24] We have already set out the uncertainties relating to the measures described above. The other measures that we have given ‘high’ or ‘very high’ uncertainty ratings are:

• The new retirement flexibilities for NHS pensions will offer 1995 Section members flexibility around how and when they can take pension benefits, with the goal of retaining more experienced NHS doctors by removing incentives to retire and barriers to returning to work. The fiscal impact therefore depends on how affected members respond, which is highly uncertain. The data and modelling on which this is based is also subject to its own uncertainties, adding to the risks around its impact.
• ‘R&D tax reliefs: additional tax relief for R&D intensive SMEs’ introduces, from April 2023, a higher 14.5 per cent rate of relief for qualifying companies, compared to 10 per cent for non-qualifying ones. There is limited data on current users, which makes predicting their behavioural response to the policy more uncertain.
• A package of support to businesses in the creative sectors to claim relief for qualifying expenditure within the corporation tax credit system.^[25] The main uncertainty relates to the behavioural response, with outturn data on claims often surpassing expectations.
• ‘Amending self-assessment forms for crypto-assets’
  introduces a dedicated tax form box to report such assets. It generates yield by prompting some individuals with gains from selling crypto-assets to report them and pay the appropriate tax. There is uncertainty in all aspects of this costing. Most obviously, it relies on a behavioural response for which there is limited evidence. In addition, HMRC has little information on individuals with crypto-assets and very little data on capital gains and losses; the modelling is complex and relies on very small sample sizes, while key assumptions are necessarily supported by limited evidence; and crypto-asset prices can be highly volatile.

Policy measures not on the Treasury scorecard

3.16 Our forecasts include the effect of four policy decisions or related impacts that the Treasury has chosen not to present on its scorecard (see Table B.2 in Annex B). The AME savings from the free childcare hours measure are described above. The Supplementary Estimates for 2022-23 that were laid in February 2023, which include a £2.4 billion increase in departmental resource budgets and a £3.6 billion reduction in departmental capital budgets. The council tax referendum limit waivers that have been granted to three councils and is fiscally neutral, raising both council tax revenue and the local authority spending that it finances by equal amounts. Other non-scorecard spending, including our assumed underspends, is also shown in Table B.2.

Scottish Government decisions

3.17 Our UK public finances forecasts are also affected by decisions taken by the devolved administrations. Since November, the Scottish Government has announced several tax measures (see Table B.3 in Annex B):^[26]

• Scottish income tax. The higher and additional rates of Scottish non-savings, non-dividend income tax have been increased by a penny, to 42p and 47p respectively. The additional-rate threshold has been reduced, from £150,000 to £125,140, in line with the UK Government’s decision at the Autumn Statement, while all other thresholds have been frozen, rather than rising with CPI inflation as assumed in the baseline.^[27] The cumulative yield from these measures rises to £190 million in 2027-28.
• Scottish non-domestic rates. The Scottish Government has maintained the poundage rate at 49.8p for 2023-24, relative to the baseline assumption that it rises in line with CPI inflation, at an average cost of £310 million a year.
• Land and buildings transaction tax. The Scottish Government has increased the additional dwelling supplement rate from 4 to 6 per cent, with effect from 16 December 2022, raising an average of £40 million a year.

Update on previous measures

3.18 We cannot review and re-cost all previous measures at each fiscal event (the volume being too great), but we do look at those where the original (or revised) costings are under- or over-performing, and at costings that were identified as particularly uncertain.

Tax threshold freezes

3.19 Tax thresholds are an essential part of every tax system and help to define precisely what is to be taxed (the tax base). Many taxes contain provisions that allow for a certain amount of income to be earned before tax is due, such as the income tax personal allowance. Some thresholds are always held flat in cash terms, such as the income tax additional rate. But for most tax thresholds the Government’s default policy (documented in its Policy costings document alongside each Budget) is to raise them in line with RPI or CPI inflation. Recently, Governments have raised significant amounts of revenue by choosing to freeze or lower a range of thresholds, both generating new taxpayers and dragging others into higher tax brackets. Those measures relating to income tax and NICs are discussed in Box 3.2.

Box 3.2: The impact of frozen or reduced personal tax thresholds

Many personal tax thresholds have been frozen in cash terms since April 2021, whereas previously most were due to rise in line with CPI inflation. The resulting ‘fiscal drag’ is set to raise significant sums as the average effective tax rate (total tax paid as a share of total income) rises more quickly over time. This happens as nominal earnings rise relative to tax thresholds, so that more of taxpayers’ income is taxed, and more of what is taxed falls into higher tax bands.^a The most important threshold changes and freezes include:

• At March Budget 2021, then Chancellor Sunak announced that the personal allowance (PA) and higher-rate thresholds (HRT) of income tax would be frozen at 2021-22 levels for the four years up to and including 2025-26. At Autumn Statement 2022 Chancellor Hunt extended the freeze by a further two years. The now six-year freeze in the PA takes its real value in 2027-28 back to its 2013-14 level.
• The additional-rate threshold (ART) was lowered from £150,000 to £125,140 from April 2023 (to align with the personal allowance taper) at Autumn Statement 2022. It had been frozen in cash terms at £150,000 since its introduction in April 2010 and will remain frozen in cash terms at its lower level.
• The primary threshold and lower profits limit for NICs were increased to align with the PA at Spring Statement 2022 (with an equivalent rise in Class 2 NICs) but (like the PA) also frozen until 2025-26. The freeze was extended to 2027-28 at Autumn Statement 2022, when the employer NICs secondary threshold was also frozen from 2023-24 to 2027-28. The latter measure means that all the main personal tax thresholds are now frozen in cash terms across our entire forecast period.

Our latest estimate is that these measures will increase receipts by a combined £29.3 billion a year (1.0 per cent of GDP) in 2027-28, as shown in Table A. Based on HMRC ready reckoners, this would be equivalent to a 4p increase in the basic rate of income tax. It is dominated by the yield from freezing the PA and the HRT.

Table A: Latest costings of personal tax threshold measures

As nominal earnings grow, these measures bring more people into income tax and NICs, and pull more taxpayers into higher and additional rates than would have occurred had the thresholds continued to rise with CPI inflation (i.e. there would be fewer taxpayers subject to each marginal tax rate in the counterfactual, assuming the same level of earnings growth). Based on our latest forecasts for earnings growth and CPI inflation, these measures are expected to generate 3.2 million (9 per cent more) new taxpayers, 2.1 million (47 per cent more) new higher-rate taxpayers, and 0.35 million (47 per cent more) additional-rate taxpayers by the end of the forecast than would have been had the thresholds continued to be uprated with inflation or, in the case of the ART, remained flat at a higher level. These figures are a touch lower than our estimates in November thanks largely to the revised profile for CPI inflation.

Table B: Number of individuals moved into paying income tax and into paying higher marginal tax rates due to threshold policies announced since March 2021

a) OBR, Fiscal drag analysis (supplementary release), July 2014.

3.20 The VAT registration threshold is the level beyond which businesses must charge VAT on their sales but can also reclaim VAT on input costs, remitting the difference – their VAT liability – to HMRC. It has been known for some time that the threshold has a “distortionary impact on business growth and activity”,^[28] possibly made worse by initially only mandating those above the threshold to sign up for making tax digital. Box 3.3 discusses the impact of the Government’s decision to freeze the threshold at £85,000 until March 2026.

Box 3.3: The impact of the frozen VAT registration threshold

One of the many aspects of the tax system that is currently subject to a freeze (meaning a threshold that typically rises with inflation is frozen in cash terms) is the turnover threshold at which firms must register for VAT. It reached £85,000 in 2017-18 and on current policy will be frozen at that level for eight years, until March 2026.a Our forecast assumes that this freeze will raise £1.4 billion a year in VAT revenues by 2027-28, by increasing the number of firms within the VAT system by 169,000 compared with indexing the threshold to RPI inflation.

Given the administrative burden and pricing consequences of being subject to the VAT regime, the registration threshold also creates an incentive for firms to cap their annual turnover just below it. And freezing the threshold while firms’ turnover rises due to inflation means that over time, while more firms become subject to VAT and more revenue is raised, there are also more firms that pile up against the threshold by capping their turnover.

Chart C shows the growing extent of the distortion in the distribution of firms by size in the yellow outturn lines. These show that the distortion predates the current freeze, with the bunching of firms shifting rightwards as the threshold was raised each year. But it also shows that since 2017-18 when the threshold first reached £85,000, the scale of the distortion below the threshold has been increasing, with 2019-20 the latest year for which outturn data are available. The blue line shows what our forecast assumes about this distortion by 2025-26 when the freeze ends. Relative to 2017-18, the number of firms capping their turnover is expected to have almost doubled from 23,000 to 44,000. And relative to a smooth distribution of firms by size, the lost turnover associated with this distortion among these traders is expected to have risen from £110 million to £350 million.

Chart C: Bunching in the VAT turnover distribution at the registration threshold

a) The threshold was frozen for two years to 2019-20 in Autumn Budget 2017, then for a further two years to 2021-22 in Budget 2018, then for another two years to 2023-24 in Spring Budget 2021, and most recently for two more years to 2025-26 in Autumn Statement 2022.

Off-payroll working: reforms in the private sector

3.21 This Budget 2018 measure shifted the responsibility for checking the tax status of individuals working (typically) through a personal service company (PSC) from the PSC to its private-sector client or ‘engager’, for those working for medium- and large-sized engagers. Working through a PSC allows individuals to pay less tax and NICs than if they were an employee of the engager, though pre-existing rules state that employment taxes are due in cases where the PSC/engager relationship is effectively that of an employee/employer. This measure was expected to raise revenue by improving compliance with those rules – increasing the amount collected via PAYE income tax and NICs by more than the amount lost through lower corporation tax, dividends tax and VAT. The measure was due to take effect in April 2020, delayed to April 2021 as a result of the pandemic, briefly scrapped in September’s Growth Plan, then reinstated a few weeks later.

3.22 Our latest estimate is that the measure yields an average of £1.5 billion a year over the forecast period, around double our previous estimate. This is largely due to the results of an HMRC evaluation of the initial tax impact of the measure using administrative data largely from 2019-20 to 2021-22. Initial HMRC research suggests that the number of new PSC payrolls may have declined since 2016-17, which in turn might suggest the measure has had a deterrent effect. Chart 3.2 shows the PAYE income tax and NICs liabilities for those in the ‘control’ and ‘treatment groups’ in the HMRC evaluation,^[29] which follow a similar path until April 2020 (the initial start date for the measure), after which the PAYE income tax and NICs liability for those individuals that switched to non-PSC payrolls during both 2020-21 and 2021-22 rose markedly (April 2021 was the delayed start date).^[30] Overall, looking across all tax heads, the HMRC evaluation finds that the 250,000 individuals working through PSCs that changed their employment arrangements had a 2021-22 tax liability that was around £9,800 higher than those in the control group.

Chart 3.2: Off-payroll reforms in the private sector: PAYE income tax and NICs monthly impact

Making tax digital

3.23 HMRC’s making tax digital (MTD) initiative for small businesses was first announced at Autumn Statement 2015, with it beginning for VAT payers from April 2019. It was expected to generate additional tax yield by introducing digital records and software that would lower taxpayers’ error and their ‘failure to take reasonable care’. HMRC has analysed the first year of outturn data, comparing the tax liability of those VAT traders that signed up for MTD and those that did not, split by those above the VAT threshold (where MTD was mandatory at the time) and those below (where it was not). For those above the threshold, the ‘MTD group’ had an aggregate 2019-20 tax liability that was estimated to be £190 million higher than the ‘non-MTD group’ (or around £230 annually per business), with the take-up of MTD within this group now around 95 per cent and expected to rise further. For those below the VAT threshold the difference in aggregate tax liability between the MTD and non-MTD groups, while still positive, was estimated at a more modest £5 million. We now expect MTD for VAT to generate an average of £0.4 billion a year over the forecast, around 40 per cent higher than our previous estimate.

3.24 In this Budget the Government has announced the latest in a series of delays to the introduction of MTD for businesses and landlords that pay self-assessed income tax. MTD for these groups is now scheduled to begin in April 2026, fully eight years later than stated in the original 2015 announcement. This is another example of how the delivery of large-scale reforms to the tax (and indeed welfare) system can, for various reasons, slip well behind schedule, posing a continuing risk to our forecasts.^[31] The 2015 measure also included an April 2020 launch date for MTD for corporation tax, for which there is currently no provision in our forecast.

Pensions flexibility

3.25 Pensions flexibility (or ‘freedoms’) allows individuals with defined contribution pensions to withdraw their funds from age 55, subject to tax paid at their marginal rate, rather than the 55 per cent charge for early withdrawal that was in place before the measure was introduced in April 2015. We have increased our forecast for tax paid on withdrawals as the initial outturn data for 2022-23 suggest that take-up this year will once again exceed our previous expectations (Chart 3.3). We now forecast £2.0 billion of receipts this year, and have revised up our forecast by £0.1 billion a year thereafter. As we discuss in Box 2.3, the pandemic has led some workers to take early retirement, which has contributed to – and perhaps been facilitated by – the rise in flexible drawdowns over the past two years.

Chart 3.3: Successive forecasts of tax receipts from the use of pensions flexibility

UK Infrastructure Bank

3.26 The UK Infrastructure Bank (UKIB) was launched in June 2021 to replace the European Investment Bank (EIB)’s function in the UK and to help deliver the National Infrastructure Strategy. It provides loans, equity financing and guarantees to projects that contribute to the Government’s net zero and ‘levelling-up’ objectives. The Government initially expected the UKIB to lend and invest around £1½ billion a year (net of lending to local authorities that would otherwise have taken place through the Public Works Loans Board), equivalent to around a third of the financing previously provided by the EIB.^[32] Our current view is that cumulative UKIB outlays between 2022-23 and 2025-26 will be 37 per cent lower than the initial estimate (£3.7 billion versus £5.9 billion). This continues to reflect the initial over-optimism in the UKIB’s investment pipeline activities and the fact that new ventures have taken longer than expected to deliver. The new profile still increases gradually over time and reaches the levels assumed in the original costing by the final two years of the forecast.

Policy risks

3.27 Parliament requires that our forecasts only reflect current government policy. As such, when the Government sets out ‘ambitions’ or ‘intentions’, we ask the Treasury to confirm whether they represent firm policy. We use that information to determine what should be reflected in our forecast. Where they are not yet firm policy, we note them as a source of risk to our central forecast – for example, the Chancellor’s ambition to move to a permanent full-expensing regime for business investment when he can afford to do so, as discussed earlier in the chapter. A full database of risks to this forecast and changes from previous updates is available on our website. Here we summarise risks that have changed materially since our November forecast and those that are new.

3.28 Risks that have crystallised and are now reflected in our forecast include:

• The EU-UK legal case on customs fraud has now crystallised, with final payments now made by the UK to the European Commission. This case relates to the undervaluation of imports of textiles and footwear from China between November 2011 and October 2017 (as first noted in our March 2019 Economic and fiscal outlook (EFO)). The UK has now accepted its liability and paid €2.6 billion, which is reflected in our spending forecast.^[33]
• The First Homes scheme has now been reflected in local authority plans, following a successful pilot of the Affordable Homes Programme that was reflected in our October 2021 EFO.
• Freeports in Scotland, Wales and Northern Ireland. In March 2021 the Government declared its ambition to establish freeports in each of Scotland, Wales, and Northern Ireland. This risk has partially crystallised with the Government announcing, in this Budget, two green freeports in Scotland, with the selection of Inverness and Cromarty Firth, and Firth of Forth, announced on 13 January 2023.^[34] An agreement between the UK and Welsh Governments was established on 12 May 2022, with bidding closing in November, while negotiations with the Northern Ireland Executive remain ongoing – these will remain as risks to our forecast until they translate into specific measures.

3.29 Risks that have evolved or that are new since November include:

• Bulb energy bailout. As described in Box 4.1, our forecast reflects £3.0 billion of spending in relation to this bailout, but much of that cost is ultimately likely to be recouped to a timetable and scale that is insufficiently certain to incorporate.
• The ‘Windsor Framework’ was agreed on 27 February and replaces the Northern Ireland Protocol. It covers issues that range from the terms of the separate deals in respect of customs, agrifoods and VAT, to governance and the role of institutions. The new framework reduces the risks posed to our forecast from frictions or disputes over the original protocol. But it entails ongoing policy risks since much of the detail has yet to be agreed. For example, the new trusted trader scheme is not expected to be in place until September 2023 and the business-to-customer parcel scheme a year later.^[35] Political uncertainty remains a factor too.
• In a speech on 27 January 2023, the Chancellor spoke of investment zones that “…will be focused on our research strengths and executed in partnership with local government, with advantageous fiscal treatment to attract new investment”. In this Budget he has provided further policy detail, but not enough for us to estimate the impacts that these investment zones might have on our forecast at this time.
• Home responsibilities protection (HRP) state pensions correction. DWP has identified underpayments of the state pension in respect of the HRP (available from 1978 to 2010 to some people with caring responsibilities, reducing the number of years required to get a full state pension). The scale of associated arrears payments and when they will be made is currently insufficiently certain to include in our forecast. It is anticipated that the scale of the costs and next steps in remedying them will be announced in the first half of 2023.
• The energy market financing scheme was announced on 8 September to support energy firms that were facing short-term liquidity challenges as a result of price volatility in energy markets but were otherwise in sound financial health. It launched on 17 October and the application window ran until 27 January 2023. It has subsequently been closed due to a lack of applications, so no longer represents a risk.

Costs of the Government not implementing its stated indexation policies

3.30 Parliament requires us to base our forecasts on stated Government policy, which for many of the rates and thresholds within the tax and benefit system is to increase them over time by a ‘default indexation’ parameter, generally CPI or RPI inflation – in the absence of specific decisions to the contrary like the threshold freezes explored in Boxes 3.2 and 3.3.^[36]

3.31 In some cases, despite Governments restating these policies every year, they are rarely implemented. The fuel duty and aggregates levy rates have been frozen for more than a decade, while rates have not risen for more than two decades for VED paid in respect of heavy goods vehicles. The biggest revenue effects from these decisions relate to fuel and alcohol duties, but a similar pattern has been seen across several smaller taxes. We estimate that the cumulative cost of freezing fuel duty rates between 2010-11 and 2023-24 relative to increasing them in line with RPI inflation has risen to around £80 billion, after factoring in the expected negative impact on demand for fuel from higher duty rates.^[37]

3.32 The freezing of rates for several taxes in this Budget comes at a cumulative cost of £16.0 billion across the forecast (Table 3.2) – a figure that equates to the direct contribution of these decisions to the level of public debt in 2027-28. Double-digit RPI inflation means that the cost of these decisions is much higher than it has been in the past – for example, it is almost 50 per cent higher than the equivalent decisions at Autumn Statement 2021, and is four times higher than those in the March 2020 Budget.

Table 3.2: Costs of not following the Government’s stated indexation policy

3.33 Non-implementation of stated indexation policies is not limited to tax cuts. For example, in this Budget the ISA subscription limit is once again frozen at £20,000, where it has stood since 2017-18, despite the Government’s stated policy throughout the period being to increase it every year by CPI inflation. Given the high rate of CPI inflation in September 2022, this year’s freeze raises amounts rising to £140 million a year by 2027-28.

Chapter 4: Fiscal outlook

Introduction

4.1 This chapter:

• notes classification issues affecting our forecast (from paragraph 4.4);
• describes the outlook for public sector receipts (from paragraph 4.6) and public sector expenditure (from paragraph 4.36);
• presents forecasts for borrowing and other deficit aggregates, including measures of the overall, current, and primary fiscal balances (from paragraph 4.69);
• describes the outlook for financial transactions (such as government lending to the private sector) and the public sector balance sheet (from paragraph 4.78); and
• summarises key uncertainties and risks to the fiscal outlook (paragraph 4.86).

4.2 The forecasts in this chapter start from the estimates of 2021-22 outturn data published by the Office for National Statistics (ONS) on 21 February. We then present an in-year estimate for 2022-23 that makes use of ONS outturn data for April 2022 to January 2023 (but not the February outturn data that will be released on 21 March). Finally, we present forecasts for 2023-24 to 2027-28.^[38] We compare our latest forecasts with those from our most recent Economic and fiscal outlook (EFO) published in November 2022.

4.3 In recent forecasts, the course of the pandemic, the path of Brexit, and the impact of the Russian invasion of Ukraine have been highly uncertain, necessitating a range of assumptions and the presentation of alternative scenarios around our central forecasts. Uncertainty remains elevated, particularly around interest rates and energy prices, and also in relation to recent increases in labour market inactivity. Throughout this chapter we discuss how changes in these factors have affected our forecasts – for example those for North Sea taxes; spending on energy-related schemes and debt interest; and working-age welfare expenditure. Chapter 5 then presents both upside and downside scenarios in relation to each of these three key sources of uncertainty.

Classification and other statistical changes

4.4 No ONS classification decisions since our November forecast have affected this forecast. But the ONS has reclassified Bulb, for the period in which it was under the special administration regime (SAR), to the public sector. We have not incorporated this into our forecast yet, but will do so when the ONS releases estimates of its impact on the public sector finances. It seems likely that, because the Government covered Bulb’s losses during this period, the additional impact of incorporating Bulb on borrowing will be small.^[39] Box 4.1 sets out how Bulb affects different aspects of this forecast.

Box 4.1: How transactions relating to Bulb feature in this forecast

Following sharp rises in wholesale gas prices in the second half of 2021, a total of 30 energy supply companies have failed. The customers of 29 of these were transferred to other suppliers under OFGEM’s supplier of last resort (SoLR) process. The largest company, Bulb Energy Limited, instead entered the special administration regime (termed the Bulb SAR) in November 2021. The Government provided the Bulb SAR administrators access to a financing facility, largely to cover the operating losses of the Bulb SAR. The ONS has classified the Bulb SAR as a public corporation, but has not yet incorporated it into the public sector finances statistics. It has also classified payments under the financing facility as capital transfers (treated as transfers from government to the private sector during its period in administration, pending Bulb’s reclassification in the statistics).

In December 2022 the administrators reached a deal with Octopus Energy Limited to transfer Bulb’s 1.5 million customers. A second financing facility has been put in place to provide cover for Bulb’s obligations under the sales process. This will mainly cover the costs of purchasing energy on the wholesale markets and runs to the end of March 2023. Pending an ONS decision on transactions under this second facility, we have treated them as capital transfers too.

In our November forecast we included a total of £6.5 billion in capital transfers from government under these two facilities: £2.0 billion for the first facility (split roughly evenly across 2021-22 and 2022-23) and £4.5 billion for the second facility. In this forecast we have revised both estimates down. For the first facility, we have reduced transfers to £1.1 billion up to March 2023. There will be further transfers to and from the Bulb SAR before the SAR order ends, but we do not expect these to be material. For the second facility, the Government has reduced the allotted sum for the facility to purchase energy from wholesale markets from £4.5 to £2.9 billion as wholesale energy prices have dropped. This represents an upper limit for these payments, and we have assumed an underspend of £1.0 billion, meaning we expect payments under this facility to total £1.9 billion and total gross payments to the Bulb SAR to be £3.0 billion.

The money made available under the Government’s second facility is used to purchase energy on the wholesale markets. This energy will then be sold to the former Bulb customers at prices close to the Ofgem price cap. The proceeds from these sales will ultimately be returned to government at some point between September 2024 and September 2025, and the amount to be returned is uncertain. So, as we only record transactions when there is sufficient certainty about their size and timing, we have not included any estimates in this forecast. More details are likely to be forthcoming before our next forecast so we will revisit these assumptions again then.

It is possible that government will recoup its outgoings via this payment – indeed Octopus is quoted as stating this will result in a profit for government of around £1.2 billion.^a This would mean the government essentially breaks even across both financing facilities. And the Government has stated its intention that the Bulb SAR will ultimately be fiscally neutral.^b This would involve recouping any remaining shortfall by issuing a ‘shortfall direction’ under the Energy Act 2011, which would place a levy on the energy supply industry. Any such decision will only be reflected in our forecasts after it has been made.

a) Financial Times, Rival UK power groups slam interest-free loans in Bulb rescue, March 2023.

b) UK Parliament, Bulb Energy, Question for Department of Business, Energy and Industrial Strategy, November 2022.

4.5 We have, on the advice of Treasury classification experts and pending an ONS view, made the following assumptions about new policies in this Budget: that payments under the energy bills discount scheme are subsidies in line with those under its predecessor energy bill relief scheme; and that the 30 hours a week of free childcare payments are also subsidies.

Public sector receipts

Summary of the receipts forecast

4.6 Continuing their strong rebound after the pandemic, we expect receipts to rise in the coming years, outstripping growth in nominal GDP and raising receipts as a share of GDP by 2.5 percentage points between 2021-22 and the forecast horizon. Receipts are forecast to rise by 11.1 per cent (£102 billion) this year, faster than nominal GDP which rises by 7.1 per cent. This means that both receipts and National Accounts taxes (a slightly narrower measure that is more comparable over longer historical periods) continue to rise as a share of GDP, the latter by 1.4 percentage points this year (Chart 4.1). The tax burden (i.e. the ratio of National Accounts taxes to GDP) now rises to 37.7 per cent of GDP in 2027-28, which would be a post-war high and is 4.7 percentage points above where it stood before the pandemic. Box 4.2 puts the evolution of the UK tax burden into an international context.

4.7 Relative to our November forecast, the tax-to-GDP ratio has been revised up by 0.4 percentage points this year. This is driven by stronger outturn receipts from income tax and NICs, VAT, corporation tax, and capital gains tax in particular, partly reflecting stronger-than-expected growth in wages and nominal consumer spending. The tax-to-GDP ratio then falls below the November forecast between 2023-24 and 2025-26, by 0.2 percentage points on average. This is more than explained by measures announced in this Budget, mainly the three-year increase in the generosity of capital allowances and the latest fuel duty freeze, which offset the stronger pre-measures position in those years (which reflects our assumption that most of the strength in recent outturn will persist). The tax burden rises above our November forecast again in 2026-27 and 2027-28, as the temporary rise in capital allowances ends, leaving the forecast higher due to pre-measures revisions.

Chart 4.1: National Accounts taxes as a share of GDP

Box 4.2: The UK’s tax burden in historical and international context

While the UK tax burden is currently high by historical standards, it has remained below the average across other advanced economies. In 2021, the most recent year for which there are internationally comparable outturn data, the UK’s tax-to-GDP ratio was 33.5 per cent of GDP on the OECD’s measure (which is slightly lower than the ONS definitions on which our forecast is based).^a That is 2.2 per cent of GDP below the average of other advanced economies,^b 3.3 per cent of GDP below the average of other G7 economies, and 6.4 per cent of GDP below the average of 14 other western European countries (the ‘EU14’). As Chart A shows, there is considerable variation across countries, with tax-to-GDP ratios among the advanced economies ranging from as low as 21.1 per cent of GDP in Ireland (where GDP is inflated by the recorded profits of foreign multinationals) to as high as 46.9 per cent of GDP in Denmark.

Chart A: Tax burdens in the advanced economies in 2021

Tax-to-GDP ratios have also been rising in other advanced economies over the past half-century. Chart B plots the evolution of the UK’s tax-to-GDP ratio since 1965 against the average of other G7 and western European (EU14) economies. It shows that:

• In 1965, at 29.7 per cent of GDP, the UK’s tax burden was relatively high (3.0 and 3.8 per cent of GDP above the other EU14 and G7 averages, respectively). This was due in part to the need to finance the UK’s larger stock of debt left by the Second World War (94.6 per cent of GDP,^c compared to an average debt stock of 28.7 per cent of GDP for the other G7 economies), as well as its higher level of defence spending (6.4 per cent of GDP)^d than that of many other advanced economies.
• Between 1965 and 1980, the UK’s tax burden was volatile (reflecting the ‘stop-go’ fiscal policies of the era and large fluctuations in nominal GDP). Nonetheless, the UK ended the 1970s with a lower tax-to-GDP ratio than in other EU14 countries (but similar to other G7 economies), whose tax burdens had risen more quickly to pay for expanding welfare states.
• Between 1981 and 1995, the UK tax burden fell from a high of 33.9 per cent of GDP in 1982 to a low of 27.4 per cent in 1993, 6.8 and 10.6 per cent of GDP below the G7 and EU14 averages, respectively. This largely reflected a fall in taxes on income and profits (of 3.0 per cent of GDP), with successive cuts to the top and basic rates of income tax during the period and a sharp fall in oil and gas revenues (as outlined in Box 4.3). By contrast, the tax burdens in other G7 and western European economies continued to rise (by 2.3 and 3.2 per cent of GDP respectively) over the same period.
• Between 1996 to 2005, the UK tax burden rose back towards the G7 and EU14 averages. This was primarily due to growth in income tax and NICs receipts, which increased by 2.2 per cent of GDP due to fiscal drag as earnings rose faster than inflation, and policy decisions to increase NICs rates.
• Between 2005 to 2020, the UK tax burden remained relatively stable, as the rate of VAT was raised but the income tax personal allowance was made more generous.^e Other high-income economies, particularly Japan, Canada and France, saw growth in their tax burdens. This widened the gap between the tax burdens of the UK and rest of the G7 to 4.4 per cent of GDP and the EU14 to 7.2 per cent of GDP.

While tax-to-GDP ratios remained remarkably stable across high-income economies during the pandemic year of 2020 (as tax revenues and GDP fell to similar degrees), the average tax burden rose sharply in 2021. This was in large part thanks to taxes on corporate incomes as profits recovered more quickly than GDP in the wake of the pandemic.^f Growth in the UK’s tax burden in 2021 (1.4 per cent of GDP) was above the advanced-economy average (0.6 per cent of GDP). In the G7, only Germany saw a larger rise than the UK in 2021 at 1.6 per cent of GDP.

Chart B: Tax burdens in the UK, G7 and EU14 since 1965

As described elsewhere in this chapter, and also illustrated in Chart A based on the OECD’s measure, the UK’s tax burden is set to rise to a post-war high of 37.7 per cent of GDP in 2027-28 (on a UK ONS National Accounts basis). What that would imply for the UK tax burden relative to its peers in five years’ time is unclear, because we cannot anticipate how fiscal policy will evolve in other advanced economies. But, like the UK, many of those economies also face growing fiscal pressures associated with ageing populations, higher stocks of debt, higher interest rates, energy insecurity and climate change, and growing geopolitical threats. Meeting these pressures while also respecting their own fiscal objectives may require further increases in tax burdens in these countries over the remainder of this decade, unless they are prepared to significantly scale back spending in other areas.

a) There is a 1.5 per cent of GDP difference between the ONS and OECD measures of the tax-to-GDP ratio in 2021. This relates to differences in both measures of tax revenues and of nominal GDP. See OECD, Annex A the OECD classification of taxes and interpretive guide for further information, 2021.

b) Based on the average of high-income countries excluding the UK, as defined by the World Bank, where OECD data are available.

c) IMF, Global Debt Database, December 2022. This is based on general government gross debt as opposed to public sector net debt, which is the headline measure in the UK and the focus of our analysis in this chapter.

d) Stockholm International Peace Research Institute, SIPRI Military Expenditure Database, accessed 1 March 2023.

e) Mee, J., OBR working paper No.15: The evolution of public sector receipts over the past decade, April 2020.

f) OECD, Revenue Statistics 2022, November 2022.

4.8 The tax-to-GDP ratio at the forecast horizon in 2027-28 is expected to be 4.7 percentage points higher than it was in 2019-20 (Chart 4.2). That reflects both pre-existing trends and subsequent changes:

• The tax system inherited by Chancellor Sunak in March 2020 would have increased the tax-to-GDP ratio by 0.7 percentage points (thanks largely to fiscal drag).
• Underlying forecast changes since then raise the ratio by 1.8 percentage points over the eight-year period, reflecting tax-rich changes in the composition of economic activity.
• Chancellor Sunak’s net tax increases, including raising the rate of corporation tax from 19 to 25 per cent, introducing the health and social care levy, and freezing income tax thresholds, would have added 2.0 percentage points to the tax burden in 2027-28.
• The net impact of November 2022 Autumn Statement measures increases the tax-to-GDP ratio by a further 0.3 percentage points. This includes tax rises announced by Chancellor Hunt in the Autumn Statement, including various threshold freezes, reversing Chancellor Sunak’s 1p cut to the basic rate of income tax, council tax rises and increasing and extending various energy-related taxes. These more than offset the tax cuts retained from Chancellor Kwarteng’s Growth Plan, notably the abolition of the health and social care levy before it had taken effect.
• Measures announced in this Budget reduce the tax-to-GDP ratio by 0.1 percentage points in 2027-28, mainly driven by the fuel duty freeze and more generous pensions tax allowances. The temporary rise in capital allowances announced in this Budget reduces the tax burden on average by 0.3 percentage points between 2023-24 and 2025-26. But after it ends in 2025-26, it temporarily raises the tax-to-GDP ratio in 2027-28, due to the lower use of capital allowances in that year after investment has been brought forward to benefit from the temporary measure.

Chart 4.2: The rise in the tax-to-GDP ratio between 2019-20 and 2027-28

Change in receipts since our November 2022 forecast

4.9 Relative to our November forecast, we have revised receipts up by £14.8 billion (1.5 per cent) in 2022-23, down slightly by £1.1 billion next year, and then up by an average of £18.8 billion a year (1.6 per cent) between 2024-25 and 2027-28. The six main factors explaining the revisions comprise three that raise receipts and three that lower them:

• Raising receipts relative to our November forecast are: (i) higher-than-expected outturn this year, driven by a combination of a modestly stronger economy (higher wage growth and consumption) and greater strength in receipts relative to their tax bases; (ii) a stronger economy forecast for future years boosting key tax bases; and (iii) higher equity prices boosting capital tax receipts.
• Lowering receipts are: (iv) lower energy prices reducing North Sea revenues, by £5.0 billion a year on average and by a peak of £10.3 billion in 2023-24; (v) lower interest rates reducing receipts on households’ interest income and the government’s financial assets; and (vi) the effects of policies announced in this Budget, which reduce receipts by £1.3 billion in 2022-23, by an average of £13.5 billion between 2023-24 and 2025-26 while the capital allowances measure is in place, and then by an average of £4.1 billion in the final two years of the forecast. The indirect effect of the near-term fiscal loosening and the medium-term supply-side boost to employment partly offset the direct costs of Budget tax measures up to 2026-27, and then more than offset them in the final year of the forecast.

Table 4.1: Receipts: changes since November

Tax-by-tax analysis

Income tax and NICs (excluding self-assessment)

4.10 PAYE income tax and NICs receipts are expected to rise by £37.6 billion (10.7 per cent) in 2022-23, down only slightly from the 12.8 per cent year-on-year rise in 2021-22. This year’s rise reflects continued strong nominal earnings growth of 5.8 per cent and the ongoing freezes to income tax thresholds that generate powerful fiscal drag.^[40] From 2023-24 onwards, we expect receipts to rise more gradually, with growth averaging 3.3 per cent a year to reach £458.9 billion (15.6 per cent of GDP) by the end of the forecast period.^[41] This slowdown in receipts growth reflects more modest rises in employment and wages, with receipts boosted by the lowering of the additional-rate threshold from April 2023 and by fiscal drag from other threshold freezes continuing through to 2027-28. The contribution of various threshold freezes to receipts is discussed in Box 3.2 in the previous chapter.

4.11 Relative to our November forecast, all non-SA income tax and NICs receipts have been revised up by £3.3 billion (0.9 per cent) in 2022-23 thanks to stronger outturns, and then by an average of £8.2 billion (2.0 per cent) a year between 2023-24 and 2027-28. The main sources of these changes are:

• On a pre-measures basis, the tax base (income from wages and salaries) has been revised up by an average of 0.7 per cent each year, raising receipts by an average of £2.8 billion. This reflects stronger pay growth in the near term and upward revisions to broader economic growth in the medium term. We have also revised up our assumptions about growth in bonus pay this year, particularly in the professional services sector (where we now expect a year-on-year rise of 20 per cent).
• The effective tax rate (ETR) has been revised up by an average of 0.1 percentage points a year, raising receipts by an average of £3.4 billion.^[42] On a pre-measures basis, this largely reflects the stronger outturn in 2022-23, which picks up further in 2023-24 before easing back in the medium term. The latter reflects a modest downward revision to the extent to which we expect the top of the earnings distribution to outperform the average over the medium term. The combined effect of a lower additional-rate threshold from April 2023 and frozen tax thresholds across the forecast means we expect the ETR to continue rising, from 36.9 per cent in 2022-23 to 37.9 per cent in 2027-28.
• Budget measures cost amounts rising to £1.2 billion a year by 2027-28, with the abolition of the lifetime allowance on pensions savings the largest single cost. But this is outweighed by the 0.5 per cent increase to the level of wages and salaries by 2027-28 from the labour supply and other measures announced in the Budget, which together raise receipts by £2.4 billion a year by 2027-28. (We have assumed that the ETR on this additional income is somewhat lower than the economy-wide average since the boost to employment is likely to be somewhat skewed to part-time work and to lower-than-average hourly pay, particularly among newly working mothers.)

Table 4.2: Non-SA income tax and NICs: changes since November

Self-assessment (SA) income tax

4.12 Self-assessed income tax receipts (which relate to income from sources including dividends, savings, property, and self-employment) are forecast to reach £42.0 billion in 2022-23 (up £5.0 billion or 13.4 per cent on 2021-22 receipts). They then rise by an average of £4.1 billion (8.3 per cent) a year, reaching £62.5 billion by 2027-28. Most SA payments are made after the tax year has ended, with SA income tax, Class 2 and 4 NICs, and capital gains tax (CGT) paid together, so this year’s receipts largely relate to tax liabilities in 2021-22. For the second year in succession, SA payments have exceeded our forecasts by a significant margin. Last year, the surplus was £10.2 billion (split roughly half income tax and NICs and half CGT). This year, the surplus is £3.3 billion (split roughly one-third income tax and NICs and two-thirds CGT, with the CGT surprise explained later).

4.13 The upward revision to SA income tax this year reflects a combination of stronger growth in self-employment income than would be consistent with ONS outturn data on such incomes in 2021-22 (repeating last year’s experience, albeit to a lesser extent), and a greater degree of forestalling of dividend income ahead of this year’s 1¼ percentage points rise in dividend tax rates (the only part of the health and social care levy reforms to survive Chancellor Kwarteng’s September Growth Plan). Tax paid on other sources of income was also modestly higher than expected. And across the board, growth in tax payments was strongest among those with higher incomes, echoing what has happened in PAYE income tax.

4.14 From 2023-24 onwards, we have revised receipts up by an average of £0.8 billion a year. This reflects the stronger starting point for most income streams, with the exception of dividend income where the strength reflects the shifting of income between years. Lower interest rates also weigh on interest income over the forecast period, more than offsetting upward revisions to income from dividends, self-employment, and property. And the latest delay in introducing elements of the ‘making tax digital’ programme takes around £0.4 billion a year off the forecast from 2025-26 onwards.

4.15 With SA income tax on self-employment incomes higher than would be consistent with outturn economic data for those incomes for the second year, Table 4.3 shows the extent to which provisional SA tax returns data (available in February for liabilities in the year 2021-22) differ from information captured in the Labour Force Survey (LFS) and National Accounts.^[43] Looking at the single-year growth rate from 2020-21 to 2021-22 – the driver of the upside surprise in 2022-23 receipts described above – self-employment incomes have risen 5.4 percentage points faster than recorded in the National Accounts, while self-employment numbers have risen 6.5 percentage points faster than recorded in the LFS. Cumulatively over the two years of the pandemic, those differences rise to 14.0 and 14.6 percentage points, respectively. These are historically large differences and help to explain the historically large upside surprises in SA receipts in this and last March’s forecasts. This higher implicit starting point for self-employment incomes consistent with the latest tax returns rather than in the National Accounts is largely carried through to later years of the forecast.

Table 4.3: Measures of self-assessed income and self-employed numbers compared

VAT

4.16 VAT is expected to raise £159.6 billion in 2022-23, up £16.3 billion (11.4 per cent) on last year, reflecting strong growth in nominal consumer spending and the end of some pandemic-era temporary tax cuts. Receipts rise more gradually over the rest of the forecast, by an average of 2.7 per cent a year from 2023-24 onwards. (The freeze to the VAT registration threshold at £85,000 until 2025-26 adds £1.4 billion to receipts by the final year of the forecast – as described in Box 3.3 in the previous chapter.)

4.17 Relative to our November 2022 forecast, receipts have been revised up by £2.9 billion (1.8 per cent) in 2022-23 due to stronger-than-expected outturns and by £4.9 billion a year on average over the remainder of the forecast. The additional medium-term upward revision on top of a higher starting point reflects the combination of higher nominal consumption (adding £0.6 billion a year on average from 2023-24 onwards) and a more tax-rich composition of spending due to lower energy prices (adding £1.0 billion a year on average). Lower energy prices boost VAT receipts by reducing the share of consumer spending on reduced-rate energy bills and increasing it on standard-rate items.

Onshore corporation tax

4.18 Onshore corporation tax (CT) receipts have performed strongly – and much more strongly than expected – through the pandemic and more recently. The path of receipts has been heavily influenced by policy changes and will continue to be over the forecast period. Over the past two years, the super-deduction capital allowance and anticipation of the rise in the main rate from 19 to 25 per cent in April 2023 have been key influences.^[44] From next year, the implementation of that rate rise and a new temporary increase in the generosity of capital allowances – this time a three-year ‘full-expensing’ regime announced in the Budget – will be the main factors. Combined with the surprising resilience of underlying profits and the average tax rate paid on those profits, this has resulted in repeated upside surprises relative to our forecasts, and greater uncertainty around future receipts than normal. By 2027-28, after the full-expensing measure has ended, the combined effect of the main rate rise and the unwinding of investment-shifting due to the temporary capital allowance measures adds £23.2 billion (21.4 per cent) to our forecast of onshore CT receipts.

4.19 Our in-year estimates for onshore CT have been revised up at each forecast since July 2020. Our latest forecast for 2022-23 (prior to measures) has been revised up by £3.5 billion relative to our November forecast. We had assumed in November that the strength in receipts seen in the first half of 2022-23 would gradually diminish over the rest of the financial year as higher wage and energy costs squeezed profit margins, but there has so far been little evidence that this has been the case. From 2023-24, receipts have been revised down in the near term and up in the medium term due to:

• An upward revision to profits in 2023, which are now expected to fall less steeply than we assumed in November and to recover more strongly in 2024. This boosts receipts relative to November by an average of £1.4 billion a year.
• Larger upward revisions to the pre-measures effective tax rate. Receipts in 2021-22 were concentrated in some relatively tax-rich sectors of the economy and among very large companies,^[45] a trend that has persisted into 2022-23. The upward revision to receipts in 2022-23 is more than accounted for by very large companies (those with profits greater than £20 million), with receipts from sectors such as financial services, retail, and professional services all notably strong. This has lifted receipts relative to profits as measured in the National Accounts (the tax base in our forecasts), which we have assumed will largely persist across the forecast period.
• The direct effects of Budget measures reduce receipts by £9.0 billion a year from 2023-24 to 2025-26, almost entirely due to the cost of the temporary full-expensing measure. By 2027-28, Budget measures raise receipts by £2.3 billion, as lower use of capital allowances beyond the temporary measure raises taxable profits. (This effect from the bringing forward of investment to benefit from full-expensing and the subsequent unwinding of that investment-shifting is included in the costing.)
• Other indirect effects of Budget measures raise receipts by an average of £0.5 billion a year over the forecast period from the effect of the near-term boost to aggregate demand and labour supply measures on profits.

Table 4.4: Onshore corporation tax: changes since November

4.20 Onshore CT receipts are forecast to reach 3.7 per cent of GDP by 2027-28, their highest level since corporation tax was introduced in 1965 (Chart 4.3).^[46] The rise in the ratio of onshore CT to GDP during the 2010s happened despite the main rate falling from 28 per cent in 2010-11 to 19 per cent in 2017-18. Our previous analysis suggests that this was mainly due to the cost of those rate cuts being largely offset by base-broadening measures restricting allowances and reliefs, for example on interest and loss relief, as well as the recovery in profits after the financial crisis.^[47] Over the forecast period, the 6 percentage points rise in the main corporation tax rate boosts receipts from this broader tax base significantly. The generosity of capital allowances has been boosted in recent years, and has been again over the next three, by the temporary super-deduction and full-expensing measures. But by 2027-28 the corporation tax system will include a higher main rate than seen during most of the 2010s, the temporary increase in the generosity of capital allowances will have ended, and the restrictions on tax reliefs introduced in the 2010s will remain in place. These factors combine to deliver the highest level of onshore CT receipts on record.

Chart 4.3: Onshore corporation tax as share of GDP

Oil and gas receipts

4.21 Oil and gas receipts include offshore corporation tax (‘ring fence’ corporation tax and the supplementary charge), petroleum revenue tax (PRT), and the temporary energy profits levy (EPL) announced last May and extended in the Autumn Statement (which raises the total rate of tax on North Sea profits to 75 per cent until the end of March 2028). We expect these taxes to raise £11.0 billion in 2022-23 – four times higher than receipts in 2021-22 – thanks to higher energy prices and the introduction of the EPL in May 2022. We expect receipts to fall relatively steadily from the peak in 2022-23 as energy prices decline, with receipts roughly halving by 2027-28 to reach £5.4 billion. As Box 4.3 describes, oil and gas receipts return close to their all-time high in cash terms in the near term, although as a share of GDP they remain far below their mid-1980s peak.

4.22 Relative to our November forecast, we have revised receipts down by £4.0 billion (26.6 per cent) this year, by £10.3 billion (49.6 per cent) in 2023-24, and by diminishing amounts thereafter (with the downward revision averaging £3.9 billion a year between 2024-25 and 2027-28). The downward revisions are driven for the most part by lower wholesale gas prices across the forecast. Weaker-than-expected receipts so far in 2022-23 also appear to reflect, in part, that gas prices achieved by firms have been lower than the market prices assumed in our November forecast because of hedging and forward sales.

Box 4.3: The evolution of North Sea oil and gas receipts

Between 2022-23 and 2027-28, North Sea oil and gas receipts are forecast to average £8.6 billion, up from an average of just £0.8 billion over the six years to 2021-22. They reach close to their all-time high in cash terms in 2023-24.^a But while large relative to recent years, relative to the size of the economy these figures remain modest by historical standards.

North Sea exploration started following the Continental Shelf Act 1964. Gas was first discovered in 1965, with production starting the following year.^b However, it was not until 1970 that commercial oil was first found and not until 1975 that production began. Since 1975, there have been several dramatic swings in receipts from oil and gas (Chart C):

• Between 1975-76 and 1984-85, receipts increased rapidly reaching a then record high of £12.0 billion (3.1 per cent of GDP). This reflected growth in production, high oil prices, and the introduction, then raising of the rate, of petroleum revenue tax (PRT).
• Between 1984-85 and 1991-92, receipts fell by 91.9 per cent, to £1.0 billion. This reflected a combination of falling oil prices, rising operating and capital expenditure, and the cut in the main rate of corporation tax from 45 to 33 per cent.
• Between 1991-92 and 2008-09, receipts rose more than twelve-fold to a record high in cash terms of £12.4 billion (but only 0.8 per cent of GDP given the larger economy by this stage). This reflected higher oil prices, which boosted sales, and the introduction of the supplementary charge (initially at 10 per cent and later raised to 20 per cent).
• Between 2008-09 and 2021-22, receipts fell by 88.4 per cent to £1.4 billion. This was largely thanks to a fall in the effective tax rate due to a rise in tax-deductible expenditure as well as policy changes. The supplementary charge was reduced to 10 per cent from 2016 and PRT was set permanently to zero in 2016 (but not abolished, in order that losses could be carried back against past PRT payments).
• Between 2021-22 and 2027-28, cash receipts are forecast to rise to £10.6 billion in 2023-24. This is 14.6 per cent shy of the all-time high in cash terms reached in 2008-09 but is only an eighth as high as a share of GDP relative to the peak in 1984-85. This rise reflects a combination of high gas prices – with gas sales expected to exceed oil sales throughout the forecast – and the rise in the effective tax rate following the introduction of the energy profits levy. Receipts then fall back steadily as gas prices ease.

Chart C: Oil and gas receipts and commercial revenues since 1970

a) All figures in this box relate to receipts scored on a cash rather than a National Accounts accruals basis to enable historical comparisons. Over the forecast period, receipts peak on a cash basis in 2023-24 but, as described in paragraphs 4.21 and 4.22, on a National Accounts basis they peak in 2022-23.

b) National Archives, North Sea oil and gas, accessed 1 March 2023.

Fuel duties

4.23 Fuel duty is expected to raise £24.8 billion this year, a £1.1 billion (4.4 per cent) drop from last year reflecting the temporary 5p cut in fuel duty rates. Receipts fall by a further £0.5 billion (1.8 per cent) in 2023-24 as the temporary duty cut has been temporarily extended in the Budget, and RPI indexation has once again been cancelled for a year, while the tax base shrinks partly due to the continuing rise in electric vehicles. Based on announced policy, receipts are expected to increase by £2.3 billion to £26.6 billion in 2024-25, due to the reversal of the 5p cut and the RPI uprating of the duty rate. Receipts then gradually decline to £25.9 billion in 2027-28 as the subsequent RPI increases in duty rate are outweighed by the continued decline in the volumes of fuel consumed.

4.24 Relative to November, receipts have been revised down by an average of £3.0 billion a year, largely reflecting the cost of cancelling the 10.1 per cent RPI indexation that was also due to take effect. Lowering the duty rate relative to going ahead with policy as stated in November is expected to increase the volume of petrol and diesel consumed by 0.3 per cent from 2024-25 onwards (raising a modest £0.1 billion a year).^[48] Beyond policy changes, lower-than-expected diesel receipts and stronger electric vehicle sales have also led us to revise down receipts in future years (by an average of £0.7 billion a year).

4.25 Fuel duty has nearly halved as a share of GDP between 1998-99 and 2021-22 (from 2.1 to 1.1 per cent of GDP) and is expected to fall further to 0.9 per cent of GDP by 2027-28. This is due to a shrinking tax base, initially as a result of fuel efficiency improvements and, over the forecast period, driven by electric vehicles, as new petrol and diesel car sales will be banned from 2030, in line with the Government’s commitment to net zero. In our 2022 Fiscal risks and sustainability report we projected fuel duty receipts to halve again as a share of GDP between 2023-24 and 2035-36 and to fall to zero by 2050-51 as older petrol and diesel vehicles are ultimately scrapped and the tax base therefore falls to zero.

4.26 As set out in the previous chapter, absent policy announcements our forecasts assume that the fuel duty rate will increase each year in line with RPI inflation – in accordance with the Government’s stated, but rarely implemented, indexation policy. To illustrate the risk associated with not implementing this policy assumption, Chart 4.4 shows our latest and past forecasts for fuel duty revenues, and the extent to which our latest forecast is boosted by the stated reversal of the temporary 5p cut and the resumption of RPI indexation of the duty rate. On average, these add £3.3 billion (0.1 per cent of GDP) a year to receipts between 2024-25 and 2027-28, and £4.0 billion in 2027-28 alone, relative to the duty rate remaining constant in cash terms. In Chapter 5 we show how this would affect headroom against the Chancellor’s fiscal targets.

Chart 4.4: Fuel duty: forecasts versus outturns

Other receipts

4.27 Capital gains tax (CGT) is forecast to raise £18.1 billion this year, up £2.8 billion (18.4 per cent) from last year. Receipts are then expected to fall very slightly (by £0.3 billion) next year before rising in each year thereafter to reach £26.1 billion (0.9 per cent of GDP) in 2027-28. This profile reflects the near-term downturn in the housing market, with rising equity prices then boosting receipts from 2024-25 onwards.

4.28 Relative to November, CGT receipts have been revised up by £2.2 billion (13.9 per cent) in 2022-23 due to stronger-than-expected self-assessment receipts (paid on 2021-22 liabilities) and higher payments on property disposals. Receipts have then been revised up by increasing amounts reaching £8.2 billion in 2027-28, as higher equity prices boost receipts. As has been the case in recent years, the £2.8 billion (18.4 per cent) year-on-year rise in CGT receipts this year is dominated by payments from those making large gains – roughly half the total growth comes from the top 1 per cent of taxpayers.

4.29 Stamp duty land tax receipts are expected to rise by £1.9 billion (12.0 per cent) this year to £17.3 billion. Receipts then fall by £4.7 billion (27.0 per cent) in 2023-24, reflecting the downturn in the housing market coupled with the September policy measure to increase the nil-rate band thresholds. Receipts then rise from 2025-26 onwards (by an average of 16.8 per cent a year) thanks to the recovery in house prices and property transactions, plus the nil-rate band measure coming to an end in March 2025 (as was announced in November). The outlook for receipts is little changed from our November forecast.

4.30 Tobacco duty receipts are expected to raise £10.0 billion this year, down £0.2 billion (1.8 per cent) on 2021-22. Receipts are expected to increase to £10.4 billion in 2023-24, driven by higher duty rates, and fall slightly in cash terms in each year thereafter as volumes consumed are expected to continue their downward trend, more than offsetting the impact from increases in the duty rate. Relative to our November forecast, receipts have been revised down by £0.7 billion this year and by an average of £0.8 billion a year from 2023-24 to 2027-28. This downward revision reflects lower outturn receipts, lower inflation (affecting duty rates), and the ongoing shift towards new tobacco categories, such as heated tobacco, that are subject to a lower tax rate.

4.31 Vehicle excise duty (VED) is expected to raise £7.4 billion this year, up by £0.3 billion (3.7 per cent) on 2021-22. Receipts have been revised up by an average of £0.3 billion a year from 2023-24 onwards, largely due to recent Department for Transport outturn data showing that car survival rates have been higher than we had previously assumed.

4.32 The electricity generator levy is expected to raise £0.8 billion this year before rising to a peak of £3.3 billion in 2023-24. Receipts are then expected to fall to £2.4 billion by 2027-28 due to the declining path of wholesale energy prices. Relative to November, receipts have been revised down by an average £0.5 billion a year between 2022-23 and 2025-26, reflecting lower gas and wholesale electricity prices. But they have been revised up by an average of £0.8 billion a year in 2026-27 and 2027-28 due to a methodological change that raises the wholesale electricity path used in this forecast.^[49]

4.33 Receipts from environmental levies have been revised up by an average of £1.2 billion a year across the forecast. In the near term, this is largely due to lower energy prices affecting the transfers to electricity producers covered by the contracts for difference scheme. In the final years, the changes are largely due to the increase in the projected costs of the capacity market due to higher than projected clearing prices. (These levies are recorded equally in our tax and spending forecasts, so are neutral for public sector net borrowing.)

4.34 Air passenger duty is expected to raise £3.3 billion this year, up by £2.1 billion (177.8 per cent) on 2021-22. Relative to our November forecast, receipts have been revised up by an average of £0.1 billion (2.8 per cent) a year from 2023-24 onwards, due to less permanent scarring to business-related travel following the pandemic than previously thought.

4.35 Interest and dividend receipts include income from the government’s financial assets such as student loans and bank deposits. These are expected to rise by £8.3 billion (34.4 per cent) this year to £32.5 billion thanks to higher interest rates and higher RPI-linked accrued interest on student loans. Receipts then reach a high of £40.5 billion in 2023-24 before edging down to £39.5 billion by 2027-28. The path of receipts reflects the profile of interest rates, which are forecast to peak next year before falling. Relative to November, we have revised receipts down by an average of £1.4 billion a year between 2023-24 and 2025-26 primarily due to the lower path of interest rates, but have revised them up in 2026-27 and 2027-28 as higher RPI inflation increases accrued student loan interest.

Public sector expenditure

Summary of the expenditure forecast

4.36 Having fallen back from its pandemic-induced peak of 53.0 per cent of GDP in 2020-21 to 44.5 per cent in 2021-22, energy support measures, higher inflation, higher interest costs, and weaker nominal GDP growth raise total managed expenditure (TME) to 46.8 per cent of GDP in 2022-23 and then 46.2 per cent next year (Chart 4.5). Spending falls back sharply to 44.6 per cent of GDP in 2024-25 as energy support measures end and inflation-related interest costs subside. Spending-to-GDP then declines more gradually from the middle of the decade to reach 43.4 per cent of GDP in 2027-28, unchanged from our November forecast. But total spending is still 4.1 per cent of GDP higher at the forecast horizon than in our pre-pandemic March 2020 forecast, and at levels not seen on a sustained basis (outside shocks and their aftermath) since the 1970s.

Chart 4.5: Public spending as a share of GDP

4.37 The declining path of public spending as a share of GDP across the forecast reflects the downward paths for both department spending (‘DEL’), which in normal times can be planned over multiple years, and annually managed expenditure (‘AME’), which is less amenable to such multi-year planning (Table 4.5). We expect DEL spending to rise slightly as a share of GDP next year due to higher investment (‘capital’) spending, before falling the year after. Both capital and current (‘resource’) DEL spending decline steadily as a share of GDP over the medium term. AME spending is expected to decline by 1.8 per cent of GDP between 2022-23 and 2024-25 as energy support schemes end and debt interest spending eases. In the final three years of the forecast, both DEL and AME spending decline gradually as a share of GDP, such that they constitute roughly the same proportions of total spending at the forecast horizon as they do this year (DEL a little less than half and AME a little more).

Table 4.5: Total managed expenditure (TME) split between departmental expenditure limits (DEL) and annually managed expenditure (AME)

4.38 The size of the state (measured by the ratio of public spending to GDP) at the forecast horizon in 2027-28 is 4.1 per cent of GDP higher than it was in 2019-20, prior to the pandemic (Chart 4.6). That reflects both pre-existing trends and subsequent changes:

• The spending plans inherited by Chancellor Sunak in March 2020 would have resulted in the spending-to-GDP ratio falling by 1.4 per cent of GDP over this period.
• Underlying forecast revisions raise spending by 3.5 per cent of GDP relative to our pre-pandemic expectations. Within that: debt interest spending contributes 2.0 percentage points, due to upward revisions to both the amount of debt and the cost of servicing it (via higher inflation and higher interest rates); revisions to welfare spending raise the spending-to-GDP ratio by 1.6 percentage points, due to the higher path of inflation and increases in health- and disability-related benefit caseloads; and other factors add 1.3 per cent of GDP, reflecting other inflation-linked spending and items where cash amounts are relatively insensitive to economic developments, so rise as a share of GDP as a consequence of the post-pandemic downward revisions to GDP.
• Spending increases announced by Chancellor Sunak would have raised the spending-to-GDP ratio by 1.4 percentage points, largely thanks to departmental spending increases in the March 2020 Budget and again at the October 2021 Spending Review.
• Spending cuts announced by Chancellor Hunt in the 2022 Autumn Statement lower spending by 1.0 per cent of GDP in 2027-28, largely due to cuts to departmental resource and capital spending (down 0.6 and 0.4 percentage points, respectively).
• Spending increases announced by Chancellor Hunt in this Budget raise spending by 0.3 per cent of GDP in 2027-28. Three-fifths of this reflects more generous provision of free childcare hours, with another quarter explained by higher defence spending. This therefore reverses around a quarter of the cuts announced in November.

Chart 4.6: The rise in the spending-to-GDP ratio between 2019-20 and 2027-28

Changes in spending since our November 2022 forecast

4.39 Relative to our November 2022 forecast, spending has been revised down by £9.9 billion this year and £9.6 billion next year, largely as a result of energy support measures costing considerably less than expected due to lower energy prices, and fewer businesses than expected qualifying for support. From 2024-25 onwards, pre-measures spending is little changed on average (higher by an average of £0.3 billion a year), albeit uneven across the period, as lower debt interest spending is offset by a range of increases across other spending lines, notably welfare spending (Table 4.6). Spending measures announced in this Budget cost an average of £6.4 billion a year, with additional energy support the largest near-term cost and the provision of free childcare hours in respect of younger children the largest medium-term cost. And the indirect effects of the policy package raise spending by a further £0.9 billion a year on average, which is more than explained by the costs of servicing the additional debt issued to finance the overall fiscal loosening.

Table 4.6: Public spending: changes since November

Spending within departmental expenditure limits

4.40 Spending subject to departmental expenditure limits (DELs) makes up just under half of all public spending and declines as a share of GDP over the next five years. In this section, ‘RDEL spending’ refers to departmental resource, or day-to-day, spending, and ‘CDEL spending’ refers to departmental capital, or investment, spending.^[50] Our latest forecasts (reported in Tables 4.7 and 4.8) reflect:

• Departments’ latest ‘forecast outturns’ for 2022-23 that were sent to the Treasury in February, the latest local government finance settlement, this year’s Supplementary Estimates, plus our assumptions regarding any further underspending relative to them.
• Departments’ plans for 2023-24 and 2024-25, as announced in the October 2021 Spending Review, including the effects of policy announcements since then.
• The Government’s post-Spending Review spending assumptions, which set overall spending totals, but not detailed plans, for 2025-26 onwards. These have been increased to reflect the cost of new measures, notably the provision of free childcare and an increased defence budget, but otherwise remain consistent with the plans set out at the November 2022 Autumn Statement – for RDEL spending to grow by 1 per cent a year in real terms and capital spending to remain flat in cash terms (with both categories therefore falling as a share of GDP).

Departmental spending in 2022-23

4.41 Departmental resource spending in 2022-23 is expected to be £415.5 billion, up £1.7 billion (0.4 per cent) from 2021-22, and unchanged from our forecast in November. Our RDEL spending estimate for 2022-23 reflects: a bottom-up assessment of likely full-year actual spending across departments; a top-down assessment of likely underspending in the final few months of the year relative to Supplementary Estimates; year-to-date spending as a share of full-year spending; and cash spending to date. These suggested little change in the extent to which the significant cost pressures in the current year are affecting departments’ ability to keep to the spending limits set by the Treasury. Given the provision for underspending in our November forecast was already modest, we have left our forecast for pre-measures RDEL spending largely unchanged (although this reflects a £2.3 billion increase in limits, offset by an increase in the assumed amount of underspending).

4.42 Departmental capital spending in 2022-23 is expected to be £86.3 billion, up £9.3 billion (12.1 per cent) on 2021-22, but £2.5 billion lower than our November forecast. Departments agreed to £3.6 billion in voluntary reductions in capital limits as part of the Supplementary Estimates process, offset by the allocation of reserve funding to meet new pressures. Our broader engagement suggests that risks to capital spending in the current year are to the downside, with supply chain pressures continuing and spending being slowed down by the need to reprioritise in light of cost increases. Overall, given our already high provision for underspending on capital in 2022-23, and the significant end-loading of capital spending seen in 2021-22, we have revised capital spending in 2022-23 down by £2.5 billion. This includes our expectation of lower spending on Bulb than assumed in our November forecast (see Box 4.1). In combination with lower departmental capital spending limits, this implies a reduction in assumed underspending of £1.1 billion.

Departmental spending from 2023-24 onwards

4.43 Departmental resource spending is expected to be £421.7 billion in 2023-24, an increase of £6.2 billion (1.5 per cent) on 2022-23 as cost-of-living and energy support measures wind down, before rising gradually in cash terms to reach £458.9 billion in 2027-28. Relative to our November forecast, RDEL spending has been revised up by increasing amounts (£2.0 billion higher in 2023-24, increasing to £6.3 billion higher in 2027-28), reflecting policy decisions taken at this Budget:

• The Government’s new childcare policies, notably extending the provision of 30 hours a week of free childcare to working parents of nine-month- to two-year-olds, cost £2.5 billion in 2024-25 and then amounts rising to £5.3 billion in 2027-28 (following the free childcare policy’s full rollout in 2025-26).
• Additional funding to meet defence pressures increases resource spending by an average of £0.6 billion from 2023-24 onwards (with larger amounts having been added to capital spending, as described below).
• The net effect of other policy decisions increases spending by an average of £1.1 billion a year, including increases to meet the resource DEL cost of delivering several elements of the Government’s labour supply package.

4.44 Beyond the costs of new policy decisions, our forecast for departmental resource spending is little changed since November. We made no changes to our judgement on pre-measures underspending, and after considering the elements of the policy package have applied our usual assumption that 5 per cent of additions to budgets will be underspent.

4.45 After accounting for policy changes announced in the Budget, the largest pressures on RDEL spending in the near term relate to inflation (and the associated pressures on pay). At the time of finalising our forecast, the Government intends to await the findings of the pay review bodies before making pay offers for 2023-24, so our forecast does not assume any RDEL additions relating to pay or inflation. The potential for significantly higher pay deals than budgeted for in departmental settlements poses a material risk to our forecast. To understand this risk more fully, in addition to our usual discussions with the Treasury, we have expanded our engagement to include several finance directors across Whitehall.

4.46 Around half of RDEL in 2022-23 is spent on pay (either via central departments or local authorities). Departments have stated that pay increases of up to 3.5 per cent next year are affordable within current settlements, lower than our forecasts for either CPI inflation (4.1 per cent) or private sector pay growth (4.5 per cent). Funding pay increases of this magnitude would imply a pressure of up to £2 billion on RDEL spending, while external estimates suggest a further pressure of up to £9 billion to close the pay gap that opened up between the public and private sectors in 2022-23.^[51] If absorbed within current budgets, the implied total pressure of between £2 and £11 billion would either exhaust some or almost all of the Treasury’s roughly £12 billion reserve or significantly reduce non-pay budgets. If added to budgets, it would represent an increase of between 0.5 and 2.6 per cent.

4.47 Departmental capital spending is expected to be £96.1 billion in 2023-24, an increase of £9.8 billion (11.4 per cent) from 2022-23. This is a significant rise between years, with continued supply chain issues suggesting that there are some downside risks to this. But our already high assumed level of underspending in 2023-24 (£8.9 billion) means that we still consider this forecast to be central. Beyond 2023-24, departmental capital spending increases modestly in cash terms to £97.6 billion in 2027-28 (but falls as a share of GDP from 3.7 per cent in 2023-24 to 3.3 per cent in 2027-28).

4.48 Relative to our November forecast, departmental capital spending has been revised up by an average of £1.6 billion a year between 2023-24 and 2027-28, more than explained by the additional capital funding for defence in this Budget (adding an average of £1.7 billion a year from 2023-24 onwards). We have applied an adjustment for underspending of 5 per cent to this, lower than our standard 20 per cent assumption for capital spending, reflecting indications that longer-term capital plans and resilient supply networks have historically resulted in less underspending in defence than is seen in other departments.^[52]

Table 4.7: Departmental resource spending: changes since November

Table 4.8: Departmental capital spending: changes since November

Welfare spending

4.49 Total welfare spending in our forecast refers to AME spending on social security and tax credits. Around half is subject to the Government’s ‘welfare cap’, which excludes the state pension and those payments most sensitive to the economic cycle (we discuss performance against the cap in Chapter 5). Welfare spending rises sharply (by £33.0 billion, or 12.6 per cent) next year, owing to the September 2022 CPI uprating of most benefits by 10.1 per cent that takes effect in April (raising welfare spending by £26.0 billion) and the second tranche of the ‘one-off’ cost-of-living payments (costing £10.3 billion, £2.0 billion more than those in 2022-23). This is a larger year-on-year percentage increase in conventional welfare spending than during the height of the pandemic (in 2020-21, albeit excluding the furlough and self-employment income schemes that are treated as subsidies rather than welfare spending in the public sector finances statistics), and the largest year-on-year percentage increase since the recession-induced rise in 1992-93.

4.50 After the inflation-driven spike in 2023-24, welfare spending rises more smoothly, by an average of £9.0 billion (2.9 per cent) a year over the rest of the forecast period. The main components of this increase are pensioner spending (which grows steadily due to the ageing population) and spending on health- and disability-related benefits (where the share of the population in receipt of these benefits continues to rise). Taken together, spending on pensions and health- and disability-related benefits rises from two-thirds of total welfare spending prior to the pandemic in 2019-20 to just under three-quarters of it in 2027-28.

Table 4.9: Total welfare spending

4.51 Relative to our November forecast, welfare spending is higher in all years (by an average of £1.9 billion) due to upward revisions to health- and disability-related benefits and to fraud and error, which more than offset smaller downward economy-related revisions. Table 4.10 shows that the overall welfare spending revision is driven by:

• Downward revisions averaging £3.7 billion a year from 2024-25 onwards due to lower inflation and unemployment forecasts, which reduce spending across the forecast by an average of £2.6 billion and £1.0 billion respectively.
• Upward revisions averaging £2.5 billion a year to health- and disability-related benefits. We have revised up spending across: personal independence payment (PIP), by an average of £0.8 billion; health-related universal credit (UC) and employment and support allowance (ESA), by £0.9 billion; and disability living allowance (DLA) by £0.5 billion. These revisions reflect higher-than-expected recent caseload outturns that we largely assume will persist, and come on top of the large upward revisions in our November forecast. They result in the combined PIP, DLA and health-related UC caseload being revised up by 0.4 million (4.5 per cent) in 2027-28, on top of the 0.5 million revision to the working-age health- and disability-related caseload for 2026-27 we made in November. The factors behind this increase are explored further below.
• An upward revision averaging £0.8 billion due to higher prevalence of fraud and error, particularly in UC. Our forecast now assumes a slower decline in the rate of fraud and error in UC from its 2021-22 peak of 14.7 per cent, as less favourable underlying trends offset some of the savings from the large increase in resources to tackle it that were announced in the Spring and Autumn Statements last year.
• Other pre-measures changes raise spending by £1.1 billion a year on average. Notably, the volume of UC ‘deductions’ – reductions in claimants’ payments to recover benefit overpayments or pay off debts, including UC advances – have yet to recover to their pre-Covid levels. Reflecting this we have revised up spending by an average of £0.6 billion over the forecast period. Other changes are small.
• Small changes due to policy measures. These include the introduction of a new disability employment programme and changes to conditionality, which we expect to move some claimants into employment and thereby reduce their benefit awards; and modest reductions to welfare spending from the employment impacts of extending the provision of 30 hours free childcare to working parents of younger children.

Table 4.10: Welfare spending: changes since November

4.52 In our latest forecast, health- and disability-related welfare spending costs £8.3 billion more in 2026-27 than we forecast it would in March 2022 – an upward revision of 11.7 per cent. This is more than explained by revised assumptions about caseloads, with lower inflation reducing spending relative to November. Our judgements about caseloads largely reflect new claims for both health- and disability-related benefits rising further above pre-pandemic averages (Chart 4.7). The latest data on ‘health-related UC’ – broadly equivalent to incapacity benefits in the pre-UC legacy system – and PIP show that the number of monthly new claims reached new highs in June and October 2022, respectively.

Chart 4.7: Monthly new claims for PIP and for health-related UC and ESA

4.53 The rise in new claims for PIP and health-related UC has been compounded by an increase in the success rate of those claims, lifting flows onto these benefits further:

• For PIP, the share of new claims that receive an award increased from 37.4 per cent in 2019-20 to 43.8 per cent in 2022-23. This is due largely to a fall in the dropout rate after a claim has been initiated. As a result, monthly PIP onflows have increased by 103 per cent since 2019-20, exceeding the 73 per cent rise in new claims.
• For health-related UC, the share of ‘decisions’ that have resulted in a successful health claim increased from 68.3 per cent in 2019-20 to 78.2 per cent in 2022-23.^[53] This increase is more than explained by flows into the ‘limited capability for work-related activity’ (LCWRA) group for people with more severe health conditions, with the share of decisions flowing into this group rising from 49.2 per cent in 2019-20 to 59.5 per cent in 2022-23.

4.54 This post-pandemic increase in the PIP caseload is largely driven by a rise in the number of people claiming for mental health and musculoskeletal conditions. Average monthly new claims that receive an award for mental health conditions have increased by 6,700 (146 per cent) since the pandemic, accounting for 41 per cent of the total rise in monthly onflows. The 4,500 increase in average monthly new claims that receive an award for musculoskeletal conditions represents a further 28 per cent of the total. There is no available data on health-related UC onflows by condition.

4.55 The result of our latest judgements about trends in working-age PIP and health-related UC caseloads has been to assume a continuously rising path for the share of working-age adults in receipt of at least one of health-related UC, ESA, PIP, or DLA. By 2027-28, in our central forecast 12.5 per cent of working-age adults receive at least one of these benefits, up by more than half from 7.8 per cent in 2019-20 and 1.9 percentage points higher than in the final year of our March 2022 forecast. This would be consistent with a continuation of recent trends for both the rising prevalence of disabilities in the population and rises in the prevalence of disability benefit receipt among that group – but it would be a significant rise in absolute terms and is subject to considerable uncertainty. Over the past decade, the share of the working-age population reporting a disability has risen by 36 per cent, while the share in receipt of a health- or disability-related benefit has risen by 40 per cent.

Chart 4.8: Prevalence of disability and health- and disability-related benefit receipt

Household energy support schemes

4.56 Energy support measures for households are expected to cost £35.8 billion in 2022-23 and £4.8 billion in 2023-24 (across our DEL and AME forecasts). They are comprised of:

• The energy price guarantee (EPG) announced in September 2022, which costs £23.0 billion in 2022-23 and £4.0 billion in 2023-24 and provides support for gas and electricity costs for domestic consumers by capping the unit price. Energy suppliers are compensated by the Government for the difference between the actual cost of supplying energy and Ofgem’s price cap. The cost of the EPG therefore depends on the level of the EPG cap, the volume of gas and electricity consumed, wholesale prices, and other factors that feature in the Ofgem price cap methodology.
• The energy bills support scheme (EBSS), a £400 discount on energy bills between October 2022 and March 2023 (announced in two tranches in February and May 2022). This is expected to cost £11.9 billion, £0.1 billion less than we estimated in November.
• The domestic alternative fuels payment announced in November 2022, which costs £0.6 billion in 2022-23. It provides a £200 payment for roughly 460,000 households in which the main form of heating is alternative fuels and that are not connected to the mains gas grid.

4.57 Spending on the EPG, by far the largest of the three programmes, has been revised down £1.7 billion in 2022-23 and £8.4 billion in 2023-24 due to a combination of forecast and policy changes (Table 4.11):

• Forecast changes reduce EPG spending in 2022-23 as a result of modestly lower-than-expected energy use (reflecting outturn spending up to 1 February). In 2023-24, lower energy prices reduce spending on a pre-measures basis, in which the EPG was £3,000 throughout the fiscal year, by £11.7 billion. This reflects much lower wholesale gas prices lowering our forecast for the Ofgem price cap, which meant the EPG would only bind in the first quarter of 2023-24, and only provide a modest subsidy in that quarter, with costs in the subsequent three quarters of the year falling to zero.
• Policy changes partially offset this downward revision, thanks to the Budget announcement that the EPG will continue at £2,500 for the first quarter of 2023-24 before rising to £3,000 as previously planned. This £500 increase in the generosity of the annualised cap is expected to cost £2.9 billion. Support for domestic customers on heat networks increases spending in 2023-24 by a further £0.4 billion.

Table 4.11: Household energy support schemes: changes since November

Non-domestic energy support schemes

4.58 Spending to support businesses with the cost of energy in 2022-23 has largely been provided via the energy bill relief scheme (EBRS), which was announced in September 2022 and provides support for non-domestic consumers of gas and electricity. It is a discount in price per unit on wholesale gas and electricity prices. Spending on the EBRS depends on the pricing and renewal of business contracts eligible for support. Eligibility is therefore less straightforward to estimate than for the EPG, where the unit subsidy is simply the difference between the actual price and the Ofgem price cap for all households. One element of the EBRS that has proved more important than expected is the price floor on energy bills below which businesses are not eligible for support.

4.59 The expected cost of the EBRS in 2022-23 has been revised down significantly – to £6.7 billion from £18.4 billion in our November forecast (Chart 4.9). Just over one-third of this huge £11.7 billion (64 per cent) downward revision is due to sharp falls in wholesale energy prices (explaining £2.8 billion) and lower discounts than we expected in November (even after accounting for lower wholesale prices) (£1.6 billion). But almost two-thirds is the result of lower volumes of eligible energy use (which explains £7.3 billion of the revision). This will reflect both lower overall energy use (due to the warm autumn and early winter) and a lower share of energy use being eligible for the scheme. It is the latter that dominates, although it is not possible to split its effect out precisely. This relates to the impact of the price floor, with a far smaller proportion of fixed price contracts proving to be eligible for support than we assumed in our November forecast.^[54]

4.60 The EBRS ends on 31 March 2023, and the Government has announced a new scheme for non-domestic customers, the energy bills discount scheme, for 2023-24. This operates in a similar manner to the EBRS, but rather than capping prices paid, it provides a discount should prices rise above certain thresholds. With wholesale prices materially lower than in the autumn, we expect this to cost just £0.5 billion in 2023-24.

Chart 4.9: Energy bill relief scheme spending: changes since November

Locally financed expenditure and public corporations’ expenditure

4.61 Locally financed current expenditure rises from £56.6 billion in 2022-23 to £70.0 billion in 2027-28.^[55] Relative to our November forecast, it has been revised down by £2.1 billion in 2022-23, revised up by £0.6 billion in 2023-24, and revised down by an average of £0.5 billion in subsequent years. As we noted in our November 2022 EFO, the in-house ready-reckoning approach used in that forecast provided more limited ability to scrutinise and report diagnostics. This was particularly the case for locally financed expenditure, which is less suited to a ready-reckoning approach than other tax and spending forecasts. As a result, we can only provide a high-level breakdown of the revisions since our previous forecast. These include:

• Business rates retained by local authorities increase spending in 2023-24 by £1.1 billion as a result of a methodological change to the treatment of the transitional reliefs measures introduced in the Autumn Statement.
• Updates for outturn and judgements on likely use of reserves in 2022-23 reduce spending by £1.1 billion this year, and by a further £0.6 billion in 2023-24 and £0.3 billion in 2024-25.
• Council tax measures announced since November result in increases in current local authority expenditure averaging £0.2 billion a year.
• Other changes reduce spending by an average of £0.5 billion a year.

4.62 Locally financed capital expenditure and public corporations’ capital expenditure increase from £21.2 billion in 2022-23 to £21.4 billion in 2027-28.^[56] Spending has been revised up by £0.1 billion in 2022-23, down by £0.8 billion in 2023-24, and up by an average of £0.2 billion in subsequent years. These revisions reflect the combination of several partly offsetting and uneven revisions to capital spending by Transport for London, local authorities’ housing revenue accounts, and the sale and acquisition of assets by funded pension schemes.

Debt interest spending

4.63 Debt interest spending (net of Asset Purchase Facility, or APF, flows) reaches 4.5 per cent of GDP in 2022-23, more than double the 2.0 per cent spent last year and the highest since immediately following the Second World War, both as a share of GDP and as a share of revenue (11.2 per cent) (Chart 4.10). This reflects the sharp rise in RPI inflation from 7.8 per cent in January 2022 to 13.4 per cent in January 2023.^[57] Spending then falls as a share of both revenue and GDP in each year until 2025-26, but most materially in 2023-24, before rising in the final two years of the forecast. This leaves spending at a similar level to the late 1990s relative to revenues, but at a level not seen since the late 1980s as a share of GDP. The effective interest rate on debt (total interest payments divided by total debt stock) follows a similar path to spending in the forecast, rising sharply in line with the spike in RPI inflation, then falling back sharply as inflation and Bank Rate decline towards the middle of the decade, before rising again to 3.0 per cent in 2027-28. While the effective interest rate at the end of the period is still lower than the level seen for much of the past 50 years, the much higher stock of debt means interest payments are consuming a historically high proportion of public revenues.

Chart 4.10: Debt interest spending relative to GDP and revenues

4.64 Debt interest spending peaks at £114.7 billion in 2022-23, £5.7 billion lower than forecast in November, but up £58.3 billion relative to 2021-22 and the highest level of nominal spending on record (Table 4.12). (£32.1 billion of the rise relative to 2021-22 reflects higher accrued spending on index-linked debt due to higher RPI inflation.) Spending then falls in each year to reach £76.9 billion in 2025-26 before rising in the final two years to reach £96.5 billion in 2027-28.

4.65 Debt interest spending has been revised down, albeit unevenly, in each year of the forecast relative to November, and by an average of £6.3 billion. This reflects several factors:

• Revisions to our pre-measures forecast for RPI inflation, which is lower by 1.0 percentage points this year and 0.8 percentage points in 2023-24, but is higher thereafter and by a maximum of 1.5 percentage points in 2025-26. This reduces spending by £5.2 billion in 2022-23 and £5.6 billion in 2023-24, but raises it in each year thereafter, most notably by £9.1 billion in 2025-26.
• Lower interest rates, with Bank Rate and gilt rates lower by 0.7 and 0.2 percentage points respectively on average over the forecast. This reduces spending by an average of £6.8 billion a year across the forecast. (As discussed in Chapter 5, interest rates have risen since our forecast assumption was settled – by enough to reverse around four-fifths of this downward revision and add £6.5 billion to spending in 2027-28.)
• Revisions to our pre-measures forecast for the Government’s net financing requirement which, alongside other minor modelling changes, increase spending by £1.2 billion this year and reduce it by an average of £2.1 billion in future years, in line with the improved pre-measures outlook for borrowing.
• Debt interest costs associated with financing the net cost of this Budget’s policies add £2.6 billion a year to spending by 2027-28, reflecting the £78.2 billion cumulative addition to borrowing by then from the direct and indirect effects of Budget measures.

Table 4.12: Central government debt interest (net of APF): changes since November

Public service pensions

4.66 Spending on public service pensions reflects outlays on pensions net of contribution income. This net cost is expected to rise from £4.2 billion in 2022-23 to £9.9 billion in 2025-26, before declining to £7.9 billion in 2027-28. Relative to our November forecast, spending has been revised up by an average of £1.0 billion a year, peaking at £1.7 billion in 2023-24. This is largely due to lower scheme receipts relating to a downward revision to pensionable earnings, which raises net spending. This partly reflects an underestimate in our November forecast of the extent to which departmental spending cuts announced in the Autumn Statement would reduce contributions income. The policy changes announced in the Budget increase spending in 2023-24 and 2024-25 by an average of £0.3 billion (as NHS pension flexibilities bring forward lump-sum payments), before reducing spending by an average of £0.5 billion a year from 2025-26 onwards (as changes to the lifetime and annual allowances reduce expenditure through scheme pays).

4.67 The NHS pension scheme is the largest scheme-level driver of changes in spending since our November forecast, contributing £0.7 billion (around three-quarters) of the average increase in overall expenditure across the forecast. The Government’s policy to offer new retirement flexibilities for NHS pensions is expected to increase spending by around £0.3 billion over the forecast horizon (with lump-sums brought forward in early years but greater contributions from members continuing to work in later years).^[58]

Other annually managed expenditure

4.68 As regards our other AME spending forecasts:

• Scottish Government current expenditure increases from £39.9 billion in 2022-23 to £44.9 billion in 2027-28. It has been revised up by an average of £1.3 billion a year relative to November, reflecting the knock-ons of policy decisions in this Budget, stated plans to spend revenues received from Crown Estate Scotland over the early years of the forecast, and higher devolved tax revenues from 2024-25 onwards.
• Scottish Government capital expenditure falls slightly from £5.4 billion in 2022-23 to £5.1 billion in 2027-28, and is little changed from our November forecast.
• When student loans are issued, the public finances record an amount of spending equal to the expected portion of the loan that will not be repaid. Spending has been revised down relative to our November forecast due to a mixture of modelling and data updates – including a modestly lower path for student entrant numbers.
• Current National Accounts adjustments have been revised up by £5.7 billion in 2022-23 and by an average of £1.8 billion higher from 2023-24 onwards, largely reflecting a correction to fully incorporate adjustments relating to local authority spending and knock-ons of updates to local authorities’ housing revenue accounts.

Borrowing and other deficit aggregates

Borrowing

4.69 Public sector net borrowing (PSNB) soared to a pandemic-induced high of 15.0 per cent of GDP in 2020-21 before dropping sharply to 5.2 per cent of GDP in 2021-22. We now expect it to rise again by 0.8 per cent of GDP to 6.1 per cent of GDP (£152.4 billion) in 2022-23, due largely to the impact of higher inflation and a range of temporary energy-related spending measures announced last year. Borrowing resumes its downward trajectory falling to 5.1 per cent of GDP in 2023-24 and 3.2 per cent of GDP in 2024-25 as energy support ends and inflation drops to close to zero. PSNB declines more gently in the second half of the forecast period, reaching 1.7 per cent of GDP (£49.3 billion) in 2027-28.

Chart 4.11: Public sector net borrowing

4.70 Relative to our November 2022 forecast, we have revised borrowing in 2022-23 down by £24.7 billion (1.0 per cent of GDP), due to a mix of higher receipts and lower spending, as detailed below. Before accounting for the impact of new policy measures, borrowing is then down by an average of £24.5 billion a year, reflecting the improvement in the outlook for receipts that builds up, while the one-off savings from lower energy prices abate (Chart 4.12). The Chancellor has used two-thirds of this improvement in pre-measures borrowing to finance his Budget measures, which raise borrowing by over £20 billion a year between 2023-24 and 2025-26, dropping to £9.9 billion in 2027-28. Revisions to post-measures borrowing are therefore uneven across the forecast period, with borrowing down in 2023-24, little changed in 2024-25 and 2025-26, and then down by larger amounts in the final two years of the forecast. The revision reaches £19.9 billion (0.7 per cent of GDP) at the forecast horizon as the cost of new temporary energy support and capital allowance measures falls away.

Chart 4.12: Public sector net borrowing: changes since November

4.71 Looking in more detail at the profile of the £11.5 billion downward revision to borrowing on average over the forecast, Table 4.13 shows that much of this reflects a more favourable starting point for the forecast, with borrowing £24.7 billion lower than expected in 2022-23. This is smaller than one might expect given the £30.6 billion like-for-like shortfall in borrowing over the first 10 months of the year relative to the monthly profiles consistent with our November forecast.^[59] In part that reflects a timing effect within our tax forecast: self-assessment receipts were more concentrated in January than we had assumed, so while year-to-date receipts were up £5.8 billion on profile, full-year receipts have been revised up a more modest £3.3 billion.

4.72 In 2022-23, the £24.8 billion pre-measures downward revision to borrowing (excluding the modest cost of new policy measures) is comprised of:

• A £12.1 billion upward revision to receipts.^[60] This largely reflects greater strength across the major tax heads, as detailed earlier in the chapter. The largest contribution comes from income tax and NICs (£4.4 billion higher), of which around three-quarters is PAYE and linked to stronger nominal wage growth, and a quarter is SA and partly explained by forestalling ahead of last year’s dividend tax rise. Next come onshore corporation tax (£3.5 billion higher), probably due to both higher profits and fewer deductions, and VAT (£2.9 billion higher) as higher wage growth supported higher consumer spending. Capital gains tax paid this year on 2021-22 liabilities also surprised to the upside (£2.2 billion higher). Against this, North Sea revenues are down £4.0 billion due to lower gas prices.
• A £12.7 billion downward revision to spending.^[61] This is more than explained by two large downward revisions to the cost of energy-related support for households and businesses (£13.5 billion) and to debt interest spending (£5.7 billion). While the cost of the household-focused EPG has been revised down somewhat (£1.7 billion or 7 per cent), the cost of the business-focused EBRS has been revised down a lot (£11.7 billion or 64 per cent), as many more firms than expected had secured energy deals that placed them below the price floor in the novel scheme (see paragraph 4.58). The downward revision to debt interest spending is largely the consequence of lower-than-forecast RPI inflation. These large downward revisions are only partly offset by a range of smaller upward revisions, plus a larger upward revision to the adjustments we make to bring our spending-control-based bottom-up forecasts in line with National Accounts categories (as noted in paragraph 4.68).

4.73 Revisions to pre-measures borrowing in future years partly reflect this better starting point. In 2023-24, downward revisions to spending on energy support measures and debt interest persist, while upward revisions to receipts remain broad-based. We have therefore revised down borrowing by £27.0 billion relative to November on a pre-measures basis. From 2024-25 onwards downward revisions to borrowing average £24.0 billion a year. They reflect the upward revisions to receipts that primarily result from an improved economic outlook, thanks to materially lower energy prices and upward revisions to net migration. These are only partly offset by upward revisions to spending on welfare, which reach £3.0 billion by 2027-28, and smaller upward revisions across a range of other spending lines, which when combined average £4.3 billion a year.

4.74 The direct effect of policy decisions increases borrowing in every year. From 2023-24 onwards, Budget measures cost £21.2 billion a year on average for the first three years, before falling to £14.4 billion in 2026-27 and £9.9 billion in 2027-28. This reflects tax and spending measures that fall broadly into five areas and that are detailed in Chapter 3. In summary, near-term energy support measures cost £4.4 billion in 2023-24; measures to boost labour supply cost amounts that rise to £7.1 billion a year by 2027-28; the three-year full-expensing capital allowances measure costs £9.1 billion a year on average from 2023-24 to 2025-26 while it is in force, but the unwinding of its impact on business investment means it raises £2.2 billion in 2027-28; other spending measures (notably more defence spending) cost £2.7 billion a year on average over five years; and other tax measures (notably extending the 5p fuel duty cut and cancelling this year’s RPI rise) cost another £3.6 billion a year on average over five years.

4.75 The indirect effects of these policies boost demand in the near term and supply in the medium term, reducing borrowing by £3.0 billion next year (thanks in particular to lower inflation) and then by an average of £1.7 billion a year thereafter (as a boost to receipts that builds to £3.9 billion by 2027-28 is partly offset by the costs of servicing the additional debt issued to finance the measures, which rises to £2.6 billion).

Table 4.13: Public sector net borrowing: changes since November

Other deficit aggregates

4.76 Beyond headline PSNB, several other deficit aggregates provide insights into the state of the public finances. The primary deficit, which excludes net interest spending, is a useful proxy of the extent to which discretionary spending is covered by revenues and is sometimes referred to as a measure of ‘fiscal effort’. The current deficit, which excludes net investment spending, is a useful proxy for the extent to which spending that mostly benefits today’s population is met by the taxes they pay. And all measures of the deficit can be presented in cyclically adjusted terms, correcting for the position in the economic cycle, which provides an indication of the underlying or structural deficit.

4.77 In this forecast, these alternative measures of the deficit show that:

• The primary deficit is 2.0 per cent of GDP this year but moves to a primary surplus of 1.1 per cent of GDP by the forecast horizon. The primary balance has not been in surplus since 2001-02. The reasons why so much more ‘fiscal effort’ is required to get debt falling in this forecast relative to recent history are explored in Box 5.1.
• The current deficit falls from 3.7 per cent of GDP this year to close to balance by 2026-27 and then moves to a surplus of 0.4 per cent of GDP in 2027-28. Net investment spending declines steadily from a peak of 2.9 per cent of GDP in 2023-24 to 2.1 per cent of GDP by 2027-28.
• Cyclically adjusted measures of the deficit are higher than unadjusted metrics in 2022-23, but decline faster over the next two years as cyclical borrowing rises while spare capacity in the economy opens up. The decline in headline borrowing beyond 2024-25 then reflects both that cyclical borrowing easing as the economy recovers, and cyclically adjusted borrowing declining as the Budget fiscal loosening gets smaller.

Financial transactions

4.78 Financial transactions and valuation effects have an uneven impact on the path of debt over the forecast, adding £24 billion in 2023-24 but reducing debt by £83 billion in 2025-26. This uneven path is driven largely by the timing of repayments of Bank of England Term Funding Scheme (TFS) loans. Absent the TFS, the impact of financial transactions is much smoother and adds to headline debt in all years by an average of £28 billion (1.0 per cent of GDP) and by slightly more (£29 billion) to underlying debt. The largest net contributors are transactions relating to student loans (£15 billion on average) and payment lags in the tax system (£9 billion). Revisions since our November forecast are uneven, reducing debt by £16 billion in 2022-23 but increasing it by an average of £8 billion a year thereafter.^[62]
Notable changes include:

• The Term Funding Scheme, where early repayments of TFS loans totalling £12 billion in 2022-23 mean fewer repayments later in the forecast than assumed in November.
• A near-term reduction of accruals adjustments relating to energy support measures which have declined as their expected cost has fallen. The net effect is to move £3 billion from 2022-23 to 2023-24.
• Changes to timing effects in respect of corporation tax due to the temporary full-expensing capital allowances measure, which reduces debt by £6 billion in the first three years of the forecast but raises it by £7 billion in the final three – over and above the measure’s effect on debt via its impact on accrued borrowing.

4.79 An important component of both our PSNB and financial transactions forecasts is the activities of the Bank of England’s Asset Purchase Facility (APF). These increase debt by £40 billion) from 2022-23 to 2027-28, which is more than explained by £46 billion in net interest losses (which affect PSNB too) due to the interest rate at which the Bank borrows to finance the APF (Bank Rate) now being above the average interest paid on the APF’s holdings of gilts. This debt impact is £21 billion smaller than in our November forecast as market expectations of Bank Rate have fallen somewhat. We have not changed our assumptions about the pace at which the APF will reduce its gilt holdings via quantitative tightening. Box 4.4 explains how the fiscal impact of the APF is reflected in our forecast.

Box 4.4: The fiscal impact of quantitative easing and quantitative tightening

Since quantitative easing (QE, the purchasing of government debt and other assets financed by the issuance of central bank reserves) was introduced in the wake of the 2008 financial crisis, central banks around the world have made large profits on these interventions. This is because the interest they pay on the reserves that finance QE asset purchases has been lower than the interest received on those assets for much of this period. In the UK, by March 2022 this had allowed the Bank of England to transfer £120 billion of cash profits on the APF to the Treasury (which both receives the profits from, and indemnifies the Bank against any losses on, the APF).

In recent months, as Bank Rate and other market interest rates have risen and so market prices for government debt held by the APF fallen, these profits have turned to losses from two sources:

• Interest losses, as the variable rate paid on central bank reserves exceeds the fixed rates paid on the assets purchased over the past 14 years (in the UK mainly gilts).
• Valuation losses as the market value of the roughly £820 billion in gilts still held in the APF has fallen below the purchase value by about £165 billion. These ‘mark-to-market’ losses only crystallise under the Treasury indemnity as and when gilts are sold.

Different central banks have different arrangements for sharing both past QE-related profits and future losses with their national treasuries and will account for these losses in a variety of ways and over different time periods. But, ultimately, the losses will end up on the books of their national governments who are the banks’ beneficial owners. In the UK, where we are used to accounting for the whole public sector including the central bank, both past profits and current and future losses feed into our forecasts for public borrowing and debt as they crystallise. The current losses on the APF affect fiscal aggregates in the following ways:

• Public sector net borrowing records the interest losses that occur when interest received on holdings of gilts and other assets is less than interest paid on central bank reserves.
• In public sector net debt, as stocks of APF assets are run down (either actively or passively), any differences between purchase price and sale price (for active sales) or redemption price (for passive run-offs) result in a realised cash loss. In debt, this is partially offset as any difference between the purchase and redemption price was recorded as public debt when the gilts were purchased. So, any further loss recorded in debt is any difference between redemption value and sale value. In addition, the net interest losses that increase annual borrowing also add to debt.
• Public sector net debt excluding the Bank of England is affected by APF losses at the point that the Bank’s APF calls on its Treasury indemnity, which (with lags) reflects the cash loss.

Table A summarises the starting and ending financial position of the APF over the current and next five years. It shows:

• Interest gains/losses. In the 13 years to the end of 2021-22, the APF had received £175 billion in interest on its assets and paid £22 billion in interest on the reserves issued to finance those assets, yielding a net profit of £153 billion for the public sector as a whole. From 2022-23 onwards, this reverses as the interest rate on reserves rises above that on gilts and other assets – yielding a net loss that peaks on an annual basis at £17 billion in 2023-24. The net loss then declines across the forecast as Bank Rate falls back and the volume of assets held in the APF decreases. Total interest losses, and so borrowing, sum to £46 billion between 2022-23 and 2027-28.
• Crystalised valuation gains/losses. In the 13 years to the end of 2021-22, £29 billion of valuation losses had crystalised in the APF. These losses increase across the forecast to a total of £61 billion, as gilt sales at prices below those at which those gilts were purchased add to losses crystallised on gilts that are redeemed at maturity.

Combining interest and valuation losses, total cash losses amount to £108 billion over the forecast, almost reversing the £124 billion cash profits to date. The APF is assumed to call on the Treasury for most of this loss within the forecast period, increasing debt excluding the Bank of England by £103 billion.

By the forecast horizon public sector net debt excluding the Bank of England (PSND ex BoE) remains £16 billion lower than it would have been in the absence of QE (in a narrow sense that ignores the support it provided to the economy). Headline PSND is £49 billion higher than it would otherwise have been, with the difference explained by £65 billion of valuation losses recorded within Bank of England debt as a result of gilts still held in the APF that were purchased at prices above their redemption value.

Table A: Forecast of APF stocks and flows

While this year has seen a reversal of net flows under the APF – from benefiting to costing the Exchequer – that has been dramatic in its speed and scale, it has always been expected that one day the direction of these flows would reverse. For example, when opting to withdraw the cash held by the APF, rather than leave it to build up in the facility, in November 2012,^a the Government foresaw the likelihood of losses as Bank Rate rose or the scheme wound down – both of the mechanisms by which losses are now occurring. And we have been forecasting future losses (dependent on market expectations and QE policy assumptions) since our first forecast with APF related flows in December 2012. Indeed, in a world where the yields and prices of longer-dated government bonds reflect expectations of the future path of Bank Rate, it should be expected that the gain from building up and running down a portfolio of gilts financed at Bank Rate would, on average, net out to zero.

a) HM Treasury, Changes to cash management operations, November 2012. The Treasury stated that “At some point in the future, as monetary conditions normalise, it is likely that the cash flows will need to be reversed.” And that “For this reason, any net coupon income transferred from the APF to the Exchequer should be used solely to pay down government debt.”

4.80 The central government net cash requirement (CGNCR, the amount of cash the government needs to raise each year), rises from £116 billion in 2022-23 to £160 billion next year before declining to £84 billion by 2027-28, generally following the path for the deficit. The CGNCR will largely be financed by the issuance of gilts to the private sector. The total supply of gilts that the private sector must absorb is especially high over the forecast period from a high CGNCR, large rollovers of maturing debt, and the sale of gilts that are currently held in the APF under the Bank of England’s quantitative tightening programme.

4.81 The volume of government debt that private sector holders will need to absorb in the coming years is likely to reach the levels last seen during the financial crisis. As shown in Chart 4.13, in the years immediately preceding the financial crisis and the advent of QE, the Government issued around 4 per cent of GDP in new gilts each year. With much of this issuance replacing (or ‘rolling over’) maturing debt, net issuance averaged 2.3 per cent of GDP. Cash borrowing via gilts shot up to 14.6 per cent of GDP in 2008-09, remained elevated over the ensuing decade, and peaked at 23.3 per cent of GDP in gross terms and 18.6 per cent of GDP in net terms at the height of the pandemic in 2020-21. However, much of this issuance has been absorbed by the APF through QE operations as monetary policy was eased, and especially so in the peak issuance years. This has meant that while net issuance of gilts by the Debt Management Office (DMO) on behalf of the Government has averaged 5.9 per cent of GDP a year between 2009-10 and 2022-23, well above pre-financial crisis levels, net private sector holdings of gilts have only increased by an average of 3.2 per cent of GDP a year. From 2023-24 onwards with borrowing subsiding, net gilt issuance by the DMO falls to 3.7 per cent of GDP a year, levels not sustained since before the financial crisis. But, with the APF now running down its own gilt holdings, the private sector needs to absorb 6.5 per cent of GDP in additional gilts each year over the next five years, the highest sustained levels this century.

Chart 4.13: UK gilt issuance since 1999-00

Debt and other balance sheet aggregates

Public sector net debt

4.82 Public sector net debt (PSND), on the headline measure that includes the Bank of England, rises from 100.6 per cent of GDP in 2022-23 to a peak of 103.1 per cent of GDP next year. It then declines in the final four years of the forecast, falling to 96.9 per cent of GDP in 2027-28, its lowest level since 2020-21 (Chart 4.14). This fall is flattered by a 5.9 per cent of GDP repayment of loans under the Bank of England Term Funding Scheme. At the forecast horizon, PSND is 2.3 per cent of GDP lower than we forecast in November, three-quarters of which is thanks to lower cumulative borrowing and a quarter of which is due to a higher nominal GDP forecast. The overall revision is made up of a 4.5 per cent of GDP downward revision to our pre-measures forecast that is partly offset by the 2.1 per cent of GDP increase in debt that results from the fiscal loosening announced in the Budget, net of its positive impact on GDP.

Chart 4.14: Public sector net debt

4.83 Public sector net debt excluding the Bank of England (PSND ex BoE) – the underlying measure that is targeted by the fiscal mandate and whose yearly path is not affected by the Term Funding Scheme – rises from 88.9 per cent of GDP this year to a peak of 94.8 per cent of GDP in 2026-27. It then falls to 94.6 per cent of GDP by 2027-28 (Table 4.14). This is despite losses made from the APF and passed onto the Treasury, which increase this measure of debt by 3.4 per cent of GDP from 2022-23 to 2027-28. Relative to November, underlying debt has been revised down by an average of 2.7 per cent of GDP over the forecast period. This is, in part, due to upward revisions to our nominal GDP forecast, which reduce the debt-to-GDP ratio by an average of 1.0 per cent of GDP relative to November, but it is largely due to a lower cash level of debt thanks to lower borrowing, which reduces debt by 1.7 per cent of GDP on average.

4.84 Looking at the effect of policy, over two-fifths of the downward revision to PSND ex BoE on a pre-measures basis (4.8 per cent of GDP in 2027-28) is offset by the 2.1 per cent of GDP added to debt by the Budget by 2027-28. That reflects just over half of the cumulative downward revision to pre-measures borrowing by 2027-28 being used to finance the fiscal loosening in the Budget, which is partly offset by the boost to nominal GDP by the forecast horizon as a result of the measures that boost labour supply and potential output.

Table 4.14: Public sector net debt (ex Bank of England): changes since November

Other balance sheet aggregates

4.85 Alongside PSND and PSND ex BoE, we also produce forecasts for public sector net financial liabilities (PSNFL) and public sector net worth (PSNW) – broader balance sheet measures shown in Chart 4.15.^[63] These provide a more comprehensive picture of the public sector balance sheet than PSND. In contrast to both measures of debt, wider measures of net liabilities peak earlier and then fall through to the forecast horizon. While underlying debt rises by 5.7 per cent of GDP from 2022-23 to 2027-28, PSNFL falls by 2.2 per cent of GDP. This is largely explained by the net acquisition of 4.3 per cent of GDP in financial assets and by a 2.7 per cent of GDP change in asset valuations. The costs of acquiring financial assets (largely student loans and equity assets of funded pension schemes) increase PSND, but as both the assets and liabilities are included in PSNFL they have little net impact. (Inverted) PSNW improves faster still, falling by 8.5 per cent of GDP between 2023-24 and 2027-28. In addition to the build-up of financial assets also recorded in PSNFL, large net worth changes include a 2.6 per cent of GDP increase in non-financial assets and a 1.4 per cent of GDP fall in unfunded pension liabilities.

Chart 4.15: Four measures of the public sector balance sheet

Risks and uncertainties

4.86 Over recent years, large shocks and their aftermath have resulted in significant revisions to our economic and fiscal forecasts from one Budget to the next. This has been true again in this forecast, with material improvement in the pre-measures outlook relative to November. We therefore continue to emphasise and illustrate the considerable uncertainty around our central forecast, with the possibility that any of several key judgements could prove too optimistic or pessimistic relative to how the economy and public finances ultimately unfold.

4.87 In Chapter 5 we illustrate upside and downside scenarios in respect of three sources of risk and uncertainty that are key to the UK’s economic and fiscal prospects, namely: the impact of labour market inactivity, the effect of energy prices, and the effect of interest rates.

4.88 Our forecasts are also subject to risks related to the impact and implementation of the Government’s stated policies and aspirations. These include, but are not limited to:

• Changes to corporation tax rates and capital allowances. Raising the main rate of corporation tax from 19 to 25 per cent from April is expected to yield £14 billion next year and £21 billion in 2027-28. In the near term, the more that the recent upside surprises relate to firms hoarding losses to deduct against future profits after the rate rise, the lower the yield will be next year. More generally, the yield reflects several uncertain assumptions about behavioural responses (from cross-border profit shifting to increased non-compliance) and broader dynamic effects (as the higher rate weighs on prospects for business investment). In addition, the Chancellor has indicated his intention to make the new temporary full-expensing regime permanent when conditions allow. That would boost prospects for business investment (as discussed in Chapter 2), but at a medium-term fiscal cost that could approach £10 billion a year.
• The planned 6p increase in the fuel duty rate in March 2024 after the one-year extension of last year’s 5p cut and latest one-year cancellation of RPI indexation announced at this Budget, which adds £2.9 billion to receipts in 2024-25. This would be the first time any Government has raised fuel duty rates in cash terms since 2011 and would increase the price of petrol and diesel by around 7 pence a litre. We have also calculated the contribution of RPI indexation of fuel duty rates to the Chancellor’s headroom against his fiscal targets in Chapter 5 (see Table 5.1).
• The impact of pay and inflationary pressures on departmental budgets over the next two years. With almost half of DEL spent on public sector pay, pay increases to match CPI inflation or private sector pay growth in 2023-24, or to close the pay gap that opened between the public and private sectors in 2022-23, imply a pressure of between £2 billion and £11 billion on departmental budgets (see paragraph 4.46). And the broader inflationary squeeze on real-terms budgets for day-to-day spending by 2024-25 (relative to when these budgets were set in the 2021 Spending Review, plus the additions to them since then) is between £13 billion and £29 billion depending on the measure of inflation used.^[64]
• In the context of the cuts to departmental spending totals beyond the current Spending Review period in November’s Autumn Statement, the tendency of governments to top up spending envelopes as the difficult moment of allocating them out in Spending Reviews approaches. Between December 2014 and November 2015, day-to-day spending totals across the 2015 Spending Review period were raised by £37 billion a year on average. The same happened between March 2021 and October 2021, with totals raised by £32 billion a year on average over the 2021 Spending Review period.
• The Government’s newly stated aspiration to spend 2.5 per cent of GDP on defence at an unspecified point in the future, as fiscal and economic circumstances allow. Relative to the NATO minimum of 2 per cent of GDP, the cost of meeting this commitment if it were met in 2027-28 would be around £15 billion.

4.89 These risks must be managed in the context of long-term pressures from an ageing population and the requirements of net zero, alongside a higher stock of public debt and a higher interest burden than the UK has borne for several decades. This combines with the historically short average maturity of this debt to leave the UK public finances more exposed to movements in interest rates than they have been in recent decades. This means both that lower interest rates can generate significant fiscal space (as illustrated by the £7.5 billion a year average downward revision to pre-measures debt interest spending in this forecast), but also that higher interest rates on a high debt stock squeeze out space for discretionary spending (as illustrated by the fact that the forecast shows a primary surplus in 2027-28 at a 26-year high, but underlying debt only falls by 0.2 per cent of GDP in that year). The challenges of getting debt onto a falling trajectory are explored in the next chapter, alongside our assessment of prospects for meeting the Government’s fiscal targets.

Chapter 5: Performance against the Government’s fiscal targets

Introduction

5.1
This chapter:

• sets out the Government’s fiscal targets and assesses their likelihood of being met on current policy under our central forecast (from paragraph 5.2); and
• considers uncertainty around our fiscal forecast and the risks to the Government meeting its fiscal targets based on historical patterns of shocks of different types, variations in key macroeconomic and fiscal determinants, and alternative scenarios for key forecast judgements (from paragraph 5.18).

The fiscal targets

5.2 The Charter for Budget Responsibility requires the OBR to judge whether the Government has a greater than 50 per cent chance of meeting its fiscal targets under current policy. The first Charter was set in 2011 and it has been updated six times as governments have revised their fiscal targets. The latest version was approved by Parliament on 6 February 2023.

5.3 This seventh edition of the Charter sets out three fiscal targets:

• A ‘fiscal mandate’ that requires public sector net debt excluding the Bank of England as a percentage of GDP to be falling by the fifth year of the rolling forecast period (currently 2027-28).
• A supplementary target that requires public sector net borrowing not to exceed 3 per cent of GDP, also by the fifth year of the rolling forecast period.
• An expenditure cap that requires welfare spending
  (excluding the state pension and payments most closely linked to the economic cycle) to be contained within a predetermined cap and margin set by the Treasury.

5.4 The Charter also includes a clause stating that, in the event of a significant adverse shock to the UK economy, the Treasury will temporarily suspend all the fiscal targets and make a statement to Parliament. At each subsequent Budget, the Chancellor would be required to update Parliament on the Government’s plan for lifting the temporary suspension of the fiscal rules. The Treasury has not invoked this escape clause at this Budget.

5.5 The Charter also identifies a broader set of indicators that the Treasury will consider in its management of fiscal policy but for which no explicit targets are set. These include:

• Wider public sector balance sheet metrics beyond the narrower debt measure that is targeted in the fiscal mandate, including public sector net financial liabilities and overall public sector net worth.
• Debt affordability metrics that look at the cost of servicing the public debt and its sensitivity to changes in the economic outlook.

The implications of our central forecast

5.6 In our central forecast, the fiscal mandate and the Government’s supplementary target for borrowing to be below 3 per cent of GDP are set to be met by modest margins, while the welfare cap is on course to be missed (Table 5.1):

• The fiscal mandate to have underlying debt (excluding the Bank of England) as a share of GDP falling in 2027-28 is met by a margin of 0.2 per cent of GDP (£6.5 billion), a 0.1 per cent of GDP (£2.7 billion) deterioration relative to the 0.3 per cent of GDP (£9.2 billion) of headroom in our November 2022 forecast.
• The borrowing target is met by a margin of 1.3 per cent of GDP (£39.2 billion in 2027-28), a 0.7 per cent of GDP (£20.6 billion) improvement relative to our November forecast for headroom of 0.6 per cent of GDP (£18.6 billion).
• The welfare cap in 2024-25 is on course to be exceeded by £4.1 billion, a £4.7 billion deterioration relative to the £0.6 billion by which it was met in November.

5.7 Given the Treasury Select Committee’s concerns about our forecasts including rarely implemented RPI indexation of fuel duty rates,^[65] we have also calculated margins against the debt falling and borrowing fiscal targets if fuel duty were to not to be raised in cash terms from its current rate, which includes the temporary 5p cut that has already been extended once. On this basis, revenue would be £4.0 billion lower than our central forecast in 2027-28, so the fiscal mandate to get debt falling would still be met but debt would only fall by 0.1 per cent of GDP (£2.8 billion), wiping out over half the Chancellor’s headroom, while the borrowing target would still be met but by £35.2 billion (1.5 per cent of GDP).^[66]

Table 5.1: Performance against the Government’s fiscal targets

Change in headroom against fiscal targets

Debt falling and public sector net borrowing targets

5.8 Improvements in our pre-measures forecasts for tax and spending increase headroom against both the debt falling and borrowing targets, while policy measures announced in the Budget reduce headroom. As shown in Chart 5.1:

• Underlying forecast changes add £5.3 billion and £29.1 billion respectively to the headroom against the debt falling and borrowing targets. This reflects both higher receipts and lower spending for the borrowing target. But these effects are largely offset in terms of the debt target by three factors: (i) slower nominal GDP growth in the final year of the forecast, thanks to the shallower near-term economic downturn than we expected in November leaving less space for above-trend growth in the medium term; (ii) the fact that part of the improvement in accrued borrowing is not reflected in cash debt in 2027-28; and (iii) because the debt-to-GDP ratio in 2026-27 is lower than we forecast in November, which means borrowing must be lower to get debt falling in the subsequent year.
• Budget policy measures reduce the headroom against both targets – by £8.0 billion and £8.5 billion respectively, thereby outweighing the pre-measures improvement relative to debt falling but only partly offsetting the improvement relative to the borrowing target. That reflects the fiscal loosening in 2027-28 of £9.9 billion, less the indirect effects of the overall fiscal loosening across the five-year forecast period that reduce borrowing by £1.4 billion. The latter reflects the boost to receipts of £3.9 billion that is partly offset by the interest paid on the debt issued to finance the Budget measures, which reaches £2.6 billion. For the debt falling target, policy effects also include the modest changes in nominal GDP growth in 2027-28 and the level of debt in 2026-27 that influence the year-on-year change in the debt-to-GDP ratio.
• This leaves overall headroom at £6.5 billion (0.2 per cent of GDP) against the debt falling target and £39.2 billion (1.3 per cent of GDP) against the target for borrowing to not exceed 3 per cent of GDP.

Chart 5.1: Fiscal target headrooms in 2027-28: changes since November

5.9 One striking feature of the changes in headroom against the two fiscal targets since November is that the margin in respect of debt falling in 2027-28 has narrowed despite the £19.9 billion downward revision to borrowing in that year. As noted above, this reflects slower nominal GDP growth at the end of the forecast period, as well as factors that have improved accrued borrowing in 2027-28 but not the level of cash debt.

5.10 These dynamics aside, borrowing reaches a historically moderate 1.7 per cent of GDP in 2027-28 – and there is a substantial primary surplus, i.e. the deficit excluding net debt interest spending, of 1.1 per cent of GDP – but yet debt only falls by 0.2 per cent of GDP in that year. This illustrates how challenging it has become to get debt falling. Indeed, the task of getting debt on a sustained downward path as a share of GDP is one that has eluded successive Chancellors since the 2008 financial crisis. Box 5.1 explores why this has been the case over the recent past and why recent developments have made this task harder.

Box 5.1: The challenges of getting debt to fall as a share of GDP

Successive Chancellors have committed, through their fiscal rules, to reduce the stock of debt as a share of GDP. All seven Charters since 2011 included an objective to put debt on a falling path. The first set a target year of 2015-16. Over time that target has been moved back several times, with the latest Charter
aiming for five years ahead, which in this forecast is 2027-28.^a Despite these targets, falls in outturn have been rare: underlying debt (excluding the Bank of England) fell from 2015-16 to 2018-19 and again in 2021-22 but rose in all other years. Underlying debt is now 23.5per cent of GDP higher than it was in 2009-10, just before the first Charter was passed; and we expect it to be 7.4 per cent of GDP higher in 2027-28 than it is this year. Why has it proven so difficult to sustainably reduce debt as a share of GDP over the past decade and a half? There are five factors at play:

The first is the greater variety, severity, and frequency of shocks. Of the nearly 50 per cent of GDP increase in debt over the past 15 years, almost nine-tenths of it occurred in just four years – the two years after the financial crisis and the pandemic respectively. (Chart A, top-left panel).

The second is the small and diminishing headroom that successive Chancellors have given themselves against debt falling at the forecast horizon. In forecasts since November 2017, debt has typically been on track to fall by less than 1 per cent of GDP in the final year (and just 0.2 per cent in this forecast). This is a very small amount relative to the impact of the shocks we have faced, each of which have added over 10 per cent of GDP to debt. (Chart A, top-right panel).

The third is growth in items other than borrowing that add to debt. In 2027-28, these financial transactions add £27 billion (0.9 per cent of GDP) to debt (excluding the Bank of England). They include the share of student loan outlays and interest that is expected to be repaid, which adds £15.8 billion (0.5 per cent of GDP). They also include losses on Asset Purchase Facility gilt sales and redemptions that are met by the Treasury, which add £2.8 billion (0.1 per cent of GDP). The remaining £8.5 billion (0.3 per cent of GDP) comprises lags in the timing of tax payments (which mean cash tax receipts are recorded later than in accrued borrowing) and other factors (such as removing the interest that accrues on student loan balances).

The fourth is weaker medium-term growth prospects, which mean that Chancellors get less help from the denominator in the debt-to-GDP ratio.As a result of post-financial crisis weakness in productivity growth – and the consequences of Brexit, the pandemic and higher energy prices – we have progressively revised down our forecasts for nominal GDP growth in the medium term. This lowers the extent to which the debt-to-GDP ratio falls in the final year of the forecast due to the size of the economy rising relative to the stock of debt (Chart A, bottom-left panel).

The fifth is higher interest rates,which have tripled over the past year across advanced economies. The effective interest rate on UK public debt in the final year of our forecast has risen nearly 2.0 per cent in our March 2021 forecast to 3.1 per cent in this one (Chart A, bottom-right panel). That has pushed debt interest spending up from 1.2 per cent of GDP in 2020-21 to 3.3 per cent of GDP in 2027-28, adding to borrowing and debt.

Chart A: Challenges in getting debt to fall as a share of GDP

The combined effect of these factors has made it more difficult to get debt falling. Lower growth means that for any given level of debt, less can be borrowed without placing debt on a rising path. And higher interest rates acting on a higher debt stock means that for any given level of borrowing, more must be spent on debt interest and less on other priorities. Indeed, the debt-stabilising primary surplus (the measure of deficit or surplus that excludes net debt interest spending) in the final year of our latest forecast is 0.9 per cent of GDP. This is the highest level in any of the forecasts that we have produced since being established in 2010 (Chart B). On this basis, the Chancellor faces a tougher challenge than any of his post-financial crisis predecessors in turning around the sustained rise in the debt-to-GDP ratio.

Chart B: Debt-stabilising primary balance in the final year of successive forecasts

a) The target was initially expressed in terms of overall public sector net debt (PSND), but due to the uneven effect of the Bank of England’s schemes on that metric, the latest mandate has been set in terms of underlying PSND excluding the Bank.

Headroom against successive fiscal mandates

5.11 Since the OBR was established in 2010, the Government has had six different fiscal mandates. Chart 5.2 shows the headrooms against those mandates. In most cases, Chancellors have chosen to aim for headroom of between £10 billion and £30 billion (expressed in terms of in today’s GDP), with some aiming even higher. In several cases where rules were on course to be broken, new and less stringent ones were adopted. This was the case in November 2022 where the horizon for getting debt falling was extended from three to five years, which turned a prospective miss of £11.4 billion into headroom of £9.2 billion at the time. The £6.5 billion of headroom against the fiscal mandate in this forecast is the smallest, excluding the pandemic years, any Chancellor has had against his main fiscal rule.

Chart 5.2: Successive forecasts for headrooms against fiscal targets

Welfare cap

5.12 The welfare cap was first introduced in 2014 and has been revised many times since, both substantively and to reflect fiscally neutral reclassifications of spending. In this Budget, the Government has once again adjusted the welfare cap – this time down by £0.6 billion compared to November 2022, reflecting the fiscally neutral changes that result from raising the universal credit administrative earnings threshold in Great Britain and Northern Ireland, which moves £0.6 billion of spending from inside to outside the cap.

5.13 The welfare cap and margin is on course to be missed by £4.1 billion in our latest forecast, a material deterioration from November when it was on course to be met by £0.6 billion. Spending subject to the welfare cap has been revised up by £1.8 billion relative to our November forecast, primarily reflecting higher spending on health and disability benefits, as discussed in Chapter 4. But a lower forecast for CPI inflation reduces the inflation adjustment by £2.1 billion relative to November, while modestly higher spending on benefits devolved to Scotland increases the Scottish welfare block grant adjustment by £0.1 billion. These factors combine to generate the prospective breach of the cap.

5.14 The Charter stipulates that we must only make a “formal assessment” of performance against the welfare cap in the first Budget of a Parliament and that instead we should “monitor” progress between those assessments. Only a breach of the cap at the point of a formal assessment triggers Parliamentary processes.^[67]

Table 5.2: The welfare cap and margin

Broader fiscal indicators

5.15 The Charter commits the Treasury to monitoring a broader set of indicators that provide a more comprehensive picture of the public sector balance sheet and consider the affordability of the stock of public debt. This is “with the aim of supporting the achievement of the fiscal objectives”. In addition to public sector net debt (PSND), both including and excluding the Bank of England (ex BoE), the broader balance sheet metrics we forecast include public sector net financial liabilities (PSNFL) and public sector net worth (PSNW). To assist in the assessment of the affordability of debt, we produce forecasts of debt interest costs both as a share of GDP and of government revenue.

5.16 The broader fiscal indicators identified in the Charter
are not formal targets, so we do not assess performance against them. Chart 4.15 in the previous chapter showed that all balance sheet metrics are on improving paths by the end of the forecast period, with the wider balance sheet measures (inverted, in the case of PSNW) peaking earlier and falling more quickly. To facilitate further monitoring, Table 5.3 presents those that feature in our forecast in a dashboard that shows: first, their levels and how these compare with the median that prevailed from 1967-68 to 2006-07 (the four decades preceding the financial crisis, before debt ratcheted higher as a result of it); and second, whether they are improving or deteriorating in each year of the forecast.

5.17 The dashboard shows that the:

• Balance sheet (stock) measures are all currently in a much worse position across the forecast than the pre-2007 median, which can be seen in the sea of red in the top part of the top panel. However, the paces at which the balance sheet metrics improve quicken, or deteriorate more slowly, over the forecast period as we can see in the shift from red/orange to orange/green in the top part of the bottom panel. And generally, the more comprehensive the metric, the faster the pace of improvement, with PSNW improving by 2.6 per cent of GDP in the final year whereas PSND excluding the Bank of England falls by just 0.2 per cent of GDP in 2027-28.
• Debt affordability (flow) measures also peak at historically high levels early in the forecast – indicated by red cells – with net interest costs peaking at 4.1 per cent of GDP and 10.1 per cent of revenue in 2022-23. This is mainly due to higher debt interest spending on index-linked gilts as RPI inflation peaks this year, while interest rates have also risen (although are lower than we forecast in November). Thereafter, net interest costs decrease at a historically fast pace – shown by the green cells at the bottom of the bottom panel – as RPI inflation and interest rates fall. But the improvements to affordability metrics slow over time with the measures increasing by the end of our forecast, as shown in the bottom part of the bottom panel.

Table 5.3: Dashboard of balance sheet and fiscal affordability indicators

Recognising uncertainty

5.18 The headroom that the Government retains against its fiscal rules, and its overall fiscal stance in the medium term, needs to be seen in the light of the potential risks to the fiscal outlook. The succession of shocks that the UK and global economies have faced since the start of the century has underscored the importance of understanding the uncertainties around a central forecast. The OBR is required to assess whether the Government has a better-than-evens chance of meeting its fiscal targets, which we do by producing a median forecast relative to which the outturn is equally likely to be higher or lower than predicted.

5.19 We use several analytical tools to illustrate the risks around our central forecast, including:

• fan charts that reflect the chances of shocks of different sizes (through stochastic simulations drawing on historical experience^[68]) to illustrate the uncertainty around our assessment of the probability of the Government meeting its fiscal targets;
• sensitivity analysis that illustrates the vulnerability of the Government’s debt and borrowing targets to changes in key forecast outcomes including growth, inflation, and interest rates; and
• alternative scenarios, which explore the economic and fiscal implications of an adverse shock to one or more of our central forecast assumptions, in this case concerning the paths for future energy prices, interest rates, and labour market participation.

Fan charts

5.20 Our fan charts are based on stochastic simulations and allow us to assess the probability of the Government meeting its fiscal targets. Chart 5.3 shows the probability distribution around our forecast of PSND excluding the Bank of England and public sector net borrowing (PSNB). It shows just a 52 per cent chance of underlying debt falling as a share of GDP in 2027-28, the lowest probability that any mandate has been met by, but a healthier 68 per cent chance of PSNB being less than 3 per cent of GDP in 2027-28. Relative to our November forecast, the chance of debt falling is slightly lower (down from 53 per cent) whereas the chance of meeting the borrowing target has improved materially (up from 59 per cent).

Chart 5.3: Fan charts for PSND (excluding Bank of England) and PSNB

Sensitivities

5.21 Our sensitivity analysis tests what would need to happen to key forecast parameters and judgements to reduce the headroom against different targets to zero (a ‘test to failure’ or ‘reverse stress test’). In the context of the current rules, we consider:

• The sensitivity of the change in the debt-to-GDP ratio
  (excluding the Bank of England) to changes in the economy-wide effective tax rate, the effective interest rate on government debt, nominal GDP growth, inflation, and departmental spending.
• The sensitivity of PSNB to changes in the effective tax rate, the effective interest rate, the level of GDP, inflation, and departmental spending.

The change in the debt-to-GDP ratio

5.22 We use our fiscal ready-reckoners to calibrate several possible adverse surprises relative to our central forecast that would be sufficient to negate the 0.2 per cent of GDP year-on-year fall in debt (excluding the Bank of England) in 2027-28.^[69] It could fall to zero if:

• The economy-wide effective tax rate (i.e. the tax-to-GDP ratio or ‘tax burden’) were just 0.2 percentage points lower, reducing tax receipts and increasing borrowing by 0.2 per cent of GDP. As we describe in Chapter 4, the tax-to-GDP ratio has risen by 4.0 per cent of GDP between 2019-20 and 2022-23 and is expected to rise by a further 1.0 per cent of GDP between 2022-23 and 2027-28, so a 0.2 percentage point shortfall in five years’ time would represent only a small downside surprise.
• The effective interest rate on government debt were 0.2 percentage points higher, at 3.2 per cent. On several occasions over the past year, the yield on UK government bonds has moved by more than 0.2 percentage points in a single day. (We present a scenario on the effect of a 1 percentage point change in both short- and long-term interest rates either side of our central forecast in the next section, as well as discussing the implications of changes in interest rates since we closed our forecast to new data.)
• Nominal GDP growth in 2027-28 were 0.2 percentage points lower. A surprise of this size would be one-seventh of the size of our average five-year ahead forecast difference for nominal GDP growth.^[70]
• RPI inflation were 1.0 percentage points higher, via its effect on accrued interest on index-linked gilts alone (since this only acts on around a quarter of outstanding debt that is inflation-linked). This margin would be very small relative to the 2.4 percentage point upside surprise in RPI inflation in 2022-23 relative to the forecast we produced just a year ago in March 2022.
• Total departmental spending were 1.2 per cent higher. This would be equivalent to reversing just under one-quarter of the departmental spending cuts announced for 2027-28 in last year’s Autumn Statement.

Public sector net borrowing

5.23 Our central forecast is for PSNB to be 1.7 per cent of GDP in 2027-28, giving headroom of 1.3 per cent of GDP against the target for PSNB to not exceed 3 per cent of GDP. This is large relative to the headroom the Chancellor has against his debt falling target, but it could still fall to zero if:

• The effective tax rate were 1.3 percentage points lower. This would be equivalent to the tax-to-GDP ratio falling by 0.3 percentage points between 2022-23 and 2027-28 rather than rising by 1.0 percentage point in our central forecast.
• Effective interest rates on central government debt were 1.3 percentage points higher at 4.4 per cent, which would be around two-thirds of the upward revision to the effective interest rate in the final year of our forecast between March 2021 and now.
• Nominal GDP were 3.2 per cent lower, via its effects across multiple tax bases. This could result from average GDP growth being 0.7 percentage points a year lower. The overall shortfall would be similar in size to the trade-intensity-related hit to potential productivity that our forecast assumes as a result of Brexit.
• RPI inflation were 5.9 percentage points higher than expected in 2027-28, causing an increase in accrued interest on index-linked gilts. This would be four-fifths of the rise in RPI inflation over the past year.
• Departmental spending were £39.2 billion higher. This would be the equivalent of reversing the Autumn Statement 2022 departmental spending cuts planned for 2027-28 in their entirety, with an additional £10 billion of spending on top.

Scenarios

5.24 In Chapters 2 and 4 we outline an array of global and domestic risks – including several policy risks associated with decisions in this Budget – that could push the public finances off course. Here we set out three alternative scenarios for key economic drivers of our fiscal forecast, maintaining our focus on energy prices and interest rates as in recent Economic and fiscal outlooks, as well as further investigating variations in labour market participation (building on our analysis in Chapter 2). We therefore consider:

• First, a labour market scenario in which labour supply is 500,000 higher or lower than assumed in our central forecast by the forecast horizon (as shown by the swathe on Chart 2.9 in Chapter 2). These scenarios would reflect a combination of labour market inactivity either continuing to rise or falling more rapidly than implied by our central forecast. Labour supply could also differ from the levels assumed in our central forecast if the policies in this Budget had a smaller or larger impact on employment than we expect. The fiscal implications of the resulting changes in employment levels come mainly via receipts (particularly taxes on labour income and the consumer spending that it finances) and welfare expenditure.
• Second, an energy price scenario that assumes gas prices rise 75 per cent – to around the post-invasion 2022 average – or fall back to pre-invasion averages (as shown by the swathe on Chart 2.1 in Chapter 2). The fiscal implications of these scenarios come via the effects of energy prices on GDP and inflation in the near and medium term, and their direct impact on taxes on North Sea gas production and electricity generators, alongside the cost of the various temporary schemes to support households and businesses with their energy bills in 2023-24. We do not assume that those schemes are extended in the high price scenario.
• Third, an interest rate scenario in which short-term and long-term rates are 1 percentage point higher or lower than assumed in our central forecast, which is around half the rise in 10-year gilt yields over the past year. The fiscal effects of these scenarios are dominated by their effect on debt interest spending and interest received on the Government’s stock of financial assets, supplemented by their indirect effects via nominal GDP, which feed through to the major tax bases.

5.25 The swathes in Chart 5.4 show how each of the three scenarios affects borrowing relative to our central forecast, while Table 5.4 shows how they affect the headroom against the Chancellor’s main fiscal targets for debt and borrowing. Together, they show:

• In our labour market scenarios, borrowing is around 0.1 per cent of GDP (£1.9 billion) higher or lower in each scenario in 2023-24, rising to 0.4 per cent of GDP (£11.2 billion) by 2027-28. Around four-fifths of these effects relate to receipts (in particular from income tax and NICs), while the remainder reflects welfare spending. In the upside scenario, underlying debt falls by 0.7 per cent of GDP in 2027-28 thanks to lower borrowing and slightly faster nominal GDP growth in that year, increasing the margin by which the fiscal mandate is met to £22.6 billion (£16.1 billion higher than in our central forecast). In the downside scenario, higher borrowing and slightly weaker nominal GDP growth in 2027-28 combine to mean that underlying debt rises by 0.3 per cent of GDP in 2027-28 and the fiscal mandate is missed by £9.9 billion (a deterioration of £16.4 billion relative to our central forecast). The borrowing target is met by significant margins in both scenarios.
• In our energy price scenarios, borrowing is £21.2 billion higher than in our central forecast in 2023-24 in the high scenario (at 5.9 per cent of GDP), reflecting an £18.7 billion increase in spending on energy support for households and businesses. In our low gas price scenario, spending on these schemes is unchanged, so the £4.8 billion reduction in borrowing in 2023-24 relative to our central forecast reflects the net effect of lower debt interest costs and lower energy-related receipts. The impact of these scenarios on borrowing is more modest from 2024-25 onwards when the support schemes end, but builds over the forecast, with borrowing around 0.2 per cent of GDP (£5½ billion) higher or lower in 2027-28. This reflects the effects of higher or lower inflation on index-linked debt and welfare uprating, alongside the effects of lower or higher nominal GDP on receipts, which are largely offset by higher or lower revenues from taxes on the North Sea and electricity generators. The fiscal mandate is met in both scenarios: in the high energy price scenario, the margin against debt falling in 2027-28 falls to just £2.9 billion (0.1 per cent of GDP), £3.6 billion lower than in our central forecast; while in the low energy price scenario the margin rises by a similar amount to stand at a still modest £10.6 billion (0.4 per cent of GDP). The borrowing target is again met by significant margins in both scenarios.
• In our interest rate scenarios, borrowing is around £20 billion higher or lower in 2027-28, meaning that the fiscal mandate would be missed by £12 billion (0.4 per cent of GDP) in the high interest rate scenario (a deterioration of £18½ billion relative to our central forecast), but met by £23 billion (0.8 per cent of GDP) in the low interest rate scenario. The high interest rate scenario is also sufficient to wipe out around half the headroom against the borrowing target shown in our central forecast. Short- and long-term interest rates have risen somewhat since we closed our central forecast, the implications of which are discussed below.

Chart 5.4: Impact of scenarios on public sector net borrowing

Table 5.4: Impact of scenarios on performance against fiscal targets

Market developments since we closed our forecast

5.26 The salience of these scenarios is demonstrated by developments in the weeks since we closed our forecast to new market data on 8 February, with interest rates higher and gas futures lower over the five days to 9 March than those on which our forecast is conditioned.^[71] The effect of these developments on our fiscal forecast and performance against the fiscal targets is likely to be mixed.

5.27 Interest rates have risen. Bank Rate expectations were an average of 0.6 percentage points higher in the five days to 9 March, and 10-year gilt rates an average of 0.4 percentage points higher, than those on which our forecast is based. At face value these developments would raise borrowing by around £6.5 billion in 2027-28 relative to our central forecast, with higher debt interest spending only partly offset by higher interest and self-assessment receipts. On its own, this would precisely remove the Chancellor’s £6.5 billion headroom against his fiscal mandate. But other developments have gone in the opposite direction, cautioning against viewing interest rate movements in isolation.

5.28 In particular, gas prices have fallen, with gas futures around 40p a therm lower over 2023-24 and 20p a therm lower in the medium term than assumed in our central forecast. This is around a third of the way between the central and low energy price scenarios described above, implying around a £1½ billion reduction in borrowing in 2027-28, and a similar increase in fiscal mandate headroom.

5.29 More generally, recent movements in interest rates may reflect changes in wider economic prospects that are fiscally favourable. To the extent that recent rises in interest rate expectations are consistent with market participants’ expectations about future economic prospects improving (which would be in line with some other recent economic data, including the first estimate of monthly GDP growth in January coming in above consensus expectations, at 0.3 per cent), they could also signal a higher path for nominal GDP than implied by our central forecast conditioned on lower rates. Nominal rates have risen somewhat more than real rates – suggesting inflation expectations have moved up. But real yields have also risen and that is consistent with (though does not unambiguously point to) slightly stronger real outcomes.^[72] Were these to be reflected in improved prospects for tax receipts, the overall impact of market news since we closed our forecast to new data could be negligible or even positive, rather than the negative implications that one would take from movements in interest rates alone.

Annex A: Detailed tables

A.1 This annex contains summary tables providing a detailed breakdown of the economy and fiscal forecasts described in this Economic and fiscal outlook, and changes since our November 2022 forecast. These tables include:

• a detailed summary of our economy forecast and key determinants of the fiscal forecast (Tables A.1 to A.4);
• public sector current receipts and individual taxes (Tables A.5 to A.6);
• total managed expenditure and its components (Tables A.7 to A.8);
• the main fiscal aggregates (Tables A.9 to A.10); and
• sources of year-on-year changes in public sector net debt (Tables A.11 to A.12).

A.2 Further breakdowns of our economy and fiscal forecasts can be found in the supplementary tables on our website, alongside all the data behind the charts and tables in this report.

Table A.1: Economy forecast

Table A.2: Economy forecast: changes since November

Table A.3: Determinants of the fiscal forecast

Table A.4: Determinants of the fiscal forecast: changes since November

Table A.5: Current receipts

Table A.6: Current receipts: changes since November

Table A.7: Total managed expenditure

Table A.8: Total managed expenditure: changes since November