Budget Deficit Tops £100 Billion In Just Two Months

30th June 2020

From the Office of Budget Responsibility 25 June 2020.

May's public finances data continue to show the budget deficit rising sharply. Two months into 2020-21, tax payments received by HMRC are down 43 per cent on the same period last year, while central government spending is up 48 per cent. Relative to our April scenario, initial estimates of both GDP and tax receipts have fallen slightly less sharply than we assumed. Debt is estimated to have topped 100 per cent of GDP for the first time since the early 1960s.

Today's data highlight the gathering fiscal impact of the coronavirus crisis, but the numbers will be prone to revision. It will be many months before the true scale of the shock becomes clear.

• Public sector net borrowing (PSNB) totalled £55.2 billion in May, £5.2 billion higher than marketexpectations. Borrowing across April and May reached £103.7 billion, despite April's deficit being revised down by £13.6 billion as accrued tax receipts were revised higher.

• HMRC cash receipts in May fell by 43 per cent on a year earlier, with VAT again accounting for the bulk of the deterioration thanks to the Government's deferral scheme.

• Central government spending was up 48 per cent on last year, reflecting the coronavirus job retention scheme, additional grants to local authorities and higher public services spending.

• Net debt rose by 20.5 per cent of GDP on a year earlier to 100.9 per cent in May (the first time it has exceeded 100 per cent of GDP since the early 1960s). This reflects the impact of higher borrowing and Bank of England schemes on cash debt, as well as the falls in nominal GDP over the coming months assumed in the reference scenario we published on 14 April.

• Relative to our reference scenario, initial estimates show that GDP and overall tax receipts have performed a little less badly than assumed, while average earnings have fallen less sharply. The

headline unemployment measure has yet to show much effect from the crisis, but the claimant count measure that relates directly to benefit claims has climbed very sharply.

Monitoring the public finances through the lockdown

1. The Office for National Statistics and HM Treasury published their Statistical Bulletin on the May 2020 Public Sector Finances this morning.

Each month the OBR provides a brief analysis

of the data and a comparison with our most recent forecast. But given the large shock to the economy and public finances caused by the coronavirus lockdown, it is no longer meaningful to compare the latest data with our March 2020 Economic and fiscal outlook (EFO). Instead, this commentary compares the latest numbers to the monthly profiles we published on 14 May consistent with our April coronavirus reference scenario (available on our website).

2. We will update our coronavirus scenario analysis in our annual Fiscal sustainability report (FSR) on 14 July. This will reflect the smaller-than-assumed peak-to-trough fall in GDP indicated by last week's GDP release (discussed below) and the public sector finances data to date. From this updated starting point we will explore three economic scenarios: one that updates our illustrative April reference scenario, with a sharp rebound in activity and no

medium-term economic scarring; one that sees activity recover more slowly and incorporates some scarring to potential GDP; and one where recovery is slower still and scarring is deeper.

The latest GDP and labour market data

3. High frequency indicators suggest that April represented the nadir in economic activity. So the April GDP data that were released last week provide the first estimate of the peak-to-trough fall in output due to the lockdown, although they will be subject to revision. Between February and April, GDP is currently estimated to have fallen by 25 per cent - an unprecedented decline, though somewhat less than the 35 per cent peak-to-trough fall assumed in our April

reference scenario. Table 1.1 compares current estimated outturns with that scenario.

4. The largest differences were in the education and health sectors. The former contracted much less than we expected, whereas the latter contracted while we had assumed output would rise.

In part this reflects how the ONS has taken on board the closure of schools and the reduction in non-coronavirus-related NHS activity, which was not clear when we prepared our reference scenario.2

In terms of market sector output, activity fell less sharply than we expected in manufacturing, construction, finance, business services and real estate. Only the last of these reflected measurement issues, where we had not properly captured the share of real estate

output made up of actual and imputed rents that have been less affected by the lockdown.

5. Our reference scenario assumed for simplicity that full lockdown would be maintained for the whole of the second quarter, but in practice it has begun to ease already and activity indicators have correspondingly begun to improve. Although we will not have GDP data for May until July 14, the ONS Business Impact of Coronavirus Survey (BICS) - also shown in Table 1.1 - provides a timelier source of information, and is used by the ONS to inform its GDP estimate. We estimate that the April BICS results were consistent with private sector turnover 36 per cent below normal - a sharper fall than shown in the GDP data, although the figures for those sectors for which consistent data are available do match up better. The May BICS results were consistent with private sector turnover being 31 per cent below normal, suggesting that GDP is likely to have remained around a fifth below its pre-virus level.

6. Labour market data released earlier this week are less directly informative about the assumptions in our reference scenario, as some of those effects are likely to build more slowly. The headline measure of unemployment rose much less than assumed, but this may simply reflect the difficulty of ‘actively seeking work' during the lockdown (a requirement of the ILO definition of unemployment). The claimant count has, though, risen sharply and broadly in

line with the unemployment assumptions we made in April. And average hours worked fell by 23 per cent between mid-March and the end of April.

7. The official measure of average weekly earnings (AWE) in April fell 0.9 per cent on a year earlier. That was much less sharp than the fall embodied in our reference scenario. However, to aid fiscal forecasting we use an implied measure obtained by dividing the National Accounts measure of wages and salaries by the number of employees and we had also assumed that CJRS payments would be recorded as transfer payments from government to households. The ONS has subsequently decided they should be recorded as a subsidy to companies with the associated payments to employees entering wages and salaries. That would imply a less steep fall in our measure of average earnings this year and a smaller rebound next year. We will update our earnings assumptions in next month's FSR.

The public finances in May 2020

8. The first estimate of public sector net borrowing (PSNB) in May 2020 was £55.2 billion, up £49.6 billion on May last year and £5.2 billion higher than market expectations. This is the largest deficit in any month on record and was close to the deficit for 2019-20 as a whole. The sharp rise reflected a £30.6 billion rise in central government spending, compounded by a £16.2 billion fall in central government receipts. Local government borrowing was up £3.0

billion on last year while public corporations' net borrowing was little changed.

9. Over the first two months of the year, PSNB totalled £103.7 billion, up £87.0 billion on the same period in last year and higher than full-year borrowing in any year since 2013-14. That

was despite PSNB in April being revised down by £13.6 billion in this month's release, largely thanks to higher accrued tax receipts. Accrued income tax and NICs receipts were revised up by £5.3 billion, in part because the ONS has assumed that more of the weakness in cash receipts reflects delays in companies paying off their employees' liabilities. VAT receipts were revised up by £2.8 billion, reflecting higher cash receipts in May. Central government

spending was revised down by £3.2 billion, largely due the cost of the CJRS being lowered.

10. PSNB over the first two months of 2020-21 is now £13.8 billion (11.7 per cent) below our reference scenario. Higher central government receipts more than explain the difference - in

particular PAYE income tax and NICs have performed less badly than assumed - possibly reflecting more resilience in the underlying tax base, but also because the wedge between accrued and cash receipts associated with non-payment of tax liabilities that will be made up in future is larger in these latest ONS estimates than our scenario assumes. It also reflects our assumption that smaller receipts lines would fall in proportion to the fall in nominal GDP,

which has not yet been fully borne out - although this could reflect information lags.

11. As we set out last month, and as illustrated by the revisions to April data reported today, initial estimates of accrued spending, receipts and borrowing are likely to be particularly prone to revision over the coming months. In many cases, the outturn data are currently based on our reference scenario profiles or other forecasts, reflecting the often significant lag between the underlying economic activity and corresponding cash payments. As more cash data become available over the coming months and these initial assumptions are replaced, large revisions can be expected. These issues are overlaid with other challenges in the data, such as adjusting the accrued tax data for non-payment of tax liabilities. If the economy and cash receipts continue to perform less badly than our reference scenario assumed, we might expect favourable revisions to accrued tax receipts in the coming months. But there is huge uncertainty around that and across other components of the public finances too.

Receipts

Cash receipts collected by HMRC

HMRC collects around 90 per cent of all central

government cash receipts. Receipts in May were

down by £19.4 billion (43 per cent) on last year.

That represented a £7.3 billion surplus relative to

our reference scenario.

Upside surprises were recorded across income tax,

NICs and VAT, whereas corporation tax fell short of

our scenario profile. These surpluses broadly fit with

the fall in GDP in April having been smaller than

expected, but also with the corporate sector being

hit much harder than the household sector.

PAYE income tax and National Insurance contributions (cash basis)

May cash receipts for PAYE income tax and NICs

relate to April liabilities. HMRC's ‘real-time

information' (RTI) shows that total pay fell 1.6 per

cent on a year earlier in April. Cash receipts in May

were down £3.3 billion (13.6 per cent) on last year.

The surplus of £4.9 billion relative to our reference

scenario probably reflects earnings holding up

better than we had assumed. The faster fall in cash

receipts than in RTI total pay points to some

liabilities not being paid by employers.

Provisional RTI for May points to a 1.7 per cent fall

in employee numbers on last year. Information on

total pay in May is not yet available, but median

pay was down 1.8 per cent. The associated drop in

May PAYE liabilities will hit June cash receipts.

VAT receipts (cash basis)

In cash terms, VAT repayments again exceeded

payments in May, so receipts were negative at

minus £0.6 billion. This largely reflects the deferral

of most VAT payments since the end of March

thanks to the deferral policy.

Our scenario assumed minus £3.1 billion for VAT

receipts in May. The smaller fall in outturn could

reflect lower-than-assumed take-up of the deferral

scheme, weaker VAT repayments or higher-thanexpected consumer spending boosting liabilities.

Onshore corporation tax receipts (cash basis)

On a cash basis, onshore corporation tax receipts

were down £1.4 billion (62 per cent) on last May, a

shortfall of £1.0 billion relative to our reference

scenario. This could reflect liabilities being hit

sooner than the scenario assumes or cash flow

problems causing firms not to pay tax that is due.

The fall relative to last May will in part reflect

companies' revised estimates of their total liabilities

for the year, but this effect should be clearer next

month when larger companies with calendar-year

accounting periods make quarterly instalments.

Fuel duty receipts (cash basis)

Fuel duty receipts in May were down £1.3 billion

(57 per cent) on last year, but they exceeded our

reference scenario by £0.3 billion. The Department

for Transport's daily data suggest that motor vehicle

use was down to around half its normal level in

May, so the fall in receipts was broadly in line with

what this might suggest.

The surplus relative to our scenario is likely to reflect

both the easing of the lockdown and the

composition of vehicle use, with use of HGVs and

LGVs having fallen less than use of cars.

Selected other HMRC cash receipts

Other notable movements in cash receipts include:

• Alcohol duty fell just 13 per cent on last year,

which is less than our scenario assumes.

Relative to our assumptions, the boost to

receipts from higher supermarket sales has

been greater than the loss from pubs closing.

• Stamp duty land tax fell 56 per cent on last

year - much less than our scenario assumed,

indicating that transactions held up more than

we had anticipated. Stamp duty on shares rose

15 per cent on a year earlier, well above the

scenario, due to higher turnover in equity

markets (not captured in the scenario).

• Air passenger duty was extremely weak again

(down 98 per cent on last year) as the number

of flights in May remained very low. The fall is

even larger than we assumed.

Public spending

Total CG spending in May was up £30.6 billion (48

per cent) on last year, but very close to our

reference scenario. So far this year it is up £65.8

billion on last year, but only £1.8 billion (0.9 per

cent) higher than our scenario. This reflects

offsetting differences, with CJRS spending revised

down to match our most latest estimate, but other

spending coming in higher than assumed (in part

due to a timing effect related to SEISS payments).

At this stage of the year, even in normal

circumstances, expenditure data are highly

provisional. The current situation has further

increased the potential for substantial revisions.

‘Other current expenditure' (largely departmental spending)

‘Other current expenditure' includes departmental

spending and grants to local authorities, but also

subsidies like the CJRS and SEISS.

Spending in May is up £27.5 billion (74 per cent)

on last year and £2.9 billion higher than our

scenario assumes. The jump relative to last year is

largely explained by the CJRS (with £10.5 billion

assumed to accrue to May) and the SEISS (with £6.8

billion accrued to May); current grants to local

authorities being brought forward, amounting to

around £1.6 billion; and around £4 billion in

additional public services spending, most of which is

likely to be coronavirus-related.

The difference relative to our scenario is largely

related to SEISS payments, which have been scored

on a cash basis whereas we assumed they would be

accrued equally across three months.

Net social benefits spending

This category includes both welfare spending and

net public service pension payments. CG net social

benefits spending in May was up £1.6 billion (8.7

per cent) on last year, but was £1.5 billion lower

than our scenario assumes. Despite the sharp rise in

the claimant count measure of unemployment, it is

possible that the scenario overstates the extent or

pace of the rise in unemployment taking place, or

the speed with which that will hit benefits spending.

Central government debt interest spending

Debt interest spending in May was up £0.2 billion

on last year and £1.1 billion higher than our

scenario assumes. This was mostly due to higherthan-expected RPI inflation, which raised spending

on index-linked gilts relative to our scenario.

Central government net investment

CG net investment in May was up £1.2 billion on

last year and £1.6 billion above our scenario

profile. Year-on-year growth reflects some

coronavirus-related capital spending, as well as

higher spending on High Speed 2 and previously

planned increases in health and social care. Capital

grants to local authorities were higher than last May

as a result of some delayed payments from last

month. Our scenario assumes that the disruption to

construction activity will weigh on capital spending

in the initial months of 2020-21, but to date that is

not apparent in the provisional data, although

April's initial estimate was revised down this month.

Central government net cash requirement (CGNCR)

Central government cash borrowing reached

£62.7 billion in May, up £46.1 billion on last May

and exceeding the total in 2019-20 as a whole.

To a large extent this reflects a sharp rise in outlays,

including the initial CJRS and SEISS cash outlays, as

well as spending on grants and public services. The

sharp fall in HMRC cash receipts also played a part,

with other receipts also down on last year

Public sector net borrowing (PSNB)

Public sector net borrowing in April totalled

£55.2 billion, up £49.6 billion on last year and a

little higher than our reference scenario assumes,

reflecting slightly higher CG spending. But PSNB in

April 2020 was revised down by £13.6 billion, with

the bulk of the revision explained by tax receipts,

taking it below our scenario. Despite that,

borrowing topped £100 billion in just two months.

Public sector net debt (PSND)

PSND in May rose 20.5 per cent of GDP on a year

earlier to reach 100.9 per cent - the first time it has

exceeded 100 per cent of GDP since the early

1960s.

Cash debt rose by £173.2 billion relative to last

May, largely due to higher central government cash

borrowing but also in part thanks to the extension to

the Bank of England's Term Funding Scheme and

the expansion of its gilt purchases.

By ONS convention, the debt-to-GDP ratio in any

given month is measured as the stock of cash debt

in that month relative to the sum of GDP over the

twelve months that straddle the end of the current

month. In this release, the ONS has again used the

assumed path of nominal GDP in our reference

scenario to fill in the period for which GDP data are

not yet available. This assumed sharp fall in

nominal GDP contributed more to the recorded rise

in the debt-to-GDP ratio than the rise in cash debt.

The ONS has not adjusted the denominator for the

smaller-than-assumed fall in April GDP, which all

else equal would lower the May debt-to-GDP ratio.

Financing

Gilt issuance, QE purchases and use of the Ways & Means Facility

Between 1 April and 16 June, the DMO issued

£146 billion of gilts - 65 per cent of the £225

billion it plans to issue between April and July (net

of redemptions, gilt issuance totalled £125 billion).

In the period since 18 March, when the Bank of

England commenced purchases under the initial

£200 billion extension to quantitative easing, it has

purchased £174 billion of gilts from the market. So

in effect the Bank has purchased £28 billion more

gilts from the private and overseas sectors than the

DMO has issued so far in 2020-21 (and £49 billion

more than net DMO issuance). To date, the

Treasury has not made use of the ‘Ways & Means

Facility' - its overdraft at the Bank.

Issues for next month's release and beyond

12. We can expect significant data revisions over the coming months, although it is difficult to estimate the potential size and direction of these changes. Some key issues include:

• Accrued spending data on the CJRS payments are currently aligned to our costings. In our 4 June update to our policy costings database, we revised the cost of the CJRS scheme down by around 30 per cent, reflecting evidence that average claims are lower

than assumed despite take-up being higher. This has been incorporated into the ONS figures, but they may be revised again as more data become available. SEISS payments have been scored on a cash basis in this release, but the ONS plans to revisit this.

• The ONS is continuing to consider how to reflect non-payment of liabilities in the accruals-based receipts figures used when estimating PSNB. This month it revised its approach to non-payment of PAYE liabilities, assuming that 93 per cent of estimated

non-payments will eventually be repaid and so score as accrued receipts in the month of the liability arising despite no cash payment having yet been made. The cash receipts data suggest this has been a material issue across many taxes.

• There is currently limited information about taxes not collected by HMRC, such asbusiness rates, which the ONS notes are currently likely to be overestimated.

See this release with graphs at

https://obr.uk/docs/June-2020-Commentary-on-the-public-sector-finances.pdf