Scotland's NI Contributions and the State Pension

Summary

Some Scottish nationalists correctly recognise that an independent Scotland would have to fund its own State Pension ... but then incorrectly claim that Scotland's National Insurance contributions are greater than the spending they fund.

The truth is that Scotland accounts for 8.0% of payments into the UK's National Insurance system but receives 8.6% of the spending that system distributes. That translates into a c.£750 million annual benefit to Scotland from pooling and sharing National Insurance contributions with the rest of the UK - a benefit which would be lost were Scotland to separate.

Context

[If you're only interested in understanding how the UK's system of pooling National Insurance works, you can skip this section]

On February 20th in the Sunday National, Scottish nationalist commentator Ruth Wishart berated Ian Blackford MP (the SNP's leader in Westminster) for claiming that, if Scotland became independent, the "commitment to continue to pay [state] pensions rests with the UK Government". The confusion caused by Blackford's intervention has been covered in detail elsewhere on this blog [see Independence by Gaslight: Pensions] but Wishart offers a fair summary when she asserts that Blackford "either hadn't read the rules on pensions or didn't understand them".

If you dip your toes into the turbulent waters of Scottish political Twitter now, you'll find one cohort of nationalists passionately arguing that the UK Government (ie. UK taxpayers) would pay the state pension in an independent Scotland, while another argues with equal passion that "nothing has changed" since 2014 and the Scottish Government (ie. Scottish taxpayers) would foot the bill.

Ruth Wishart is firmly in the latter camp and, after consulting "pensions guru (sic)" Tim Rideout, she offered the following (correct) summary: 

"The short-form version is that, as before, the National Insurance (NI) contributions of today’s workforce pay today’s pensioners."

 Unfortunately she then went on to muddy the waters by adding:

In order to calculate what an independent Scotland could afford, Mr Rideout turned to the GERS figures [..] This, he says, shows that we would collect almost £3 billion more in NI than we would pay out in pensions. Most of the surplus would be used to pay other entitlements derived from national insurance."

We will come on to detail the manifold problems with this claim, but the most obvious is that you can't define "what an independent Scotland could afford" by comparing a single revenue item with a single cost line. The only way to answer that question is to work out what the overall balance of revenue and expenditure would be and take a view as to what level of deficit would be sustainable. We won't go down that particular rabbit-hole here, but in simple summary: the total view of tax and spend as described in GERS reveals Scotland receiving a £12bn annual fiscal transfer from the rest of the UK [see GERS 2021: A Deep Dive]. This means that - before factoring in the other myriad possible economic impacts of independence - Scotland would have to find £12bn of cost savings or tax increases vs the GERS figures and/or be able to sustain a far higher fiscal deficit/GDP ratio than that currently shared with the UK (or than required to comply with the EU's Fiscal Compact). 

That said, because National Insurance (NI) is as close as we get in the UK to a truly hypothecated tax (ie. NI revenue is ring-fenced to be used only for specific spending purposes), this does give us an opportunity to see how pooling and sharing works for this sub-set of total tax and spend.

So let's take a closer look at Rideout's claim, which he first made in tweet form:

This is a textbook example of a common nationalist trope: take two vaguely related numbers, draw a superficial conclusion that makes it seem like they show everything would be fine if Scotland was fiscally autonomous and move on.

Rideout went further on a recent Zoom call with fellow nationalists. After hilariously describing the Scottish Government's GERS figures as "the unionists' numbers" he rattled off a few figures before concluding - quite wrongly - that "there was actually £500 million or so left over from the National Insurance that we paid".  This blog explains in detail why he's wrong (see footnote 14), but the short version is this: he completely missed the £2.5 billion a year Scotland gets back as a direct result of National Insurance Contributions being allocated to the NHS.

To answer the question Rideout is claiming to address we need to put in some proper analytical effort - but Chokkablog has never been shy of doing the hard yards when it comes to the numbers, so let's dive in.

Where National Insurance Contributions Go

We don't have strictly hypothecated taxes in the UK. In layman's terms that means you can't say "this tax pays for this service" because the Treasury has considerable freedom to take money and deploy it where they see fit. National Insurance contributions are, however, a special case where a significant degree of hypothecation is enforced. This allows us to look at National Insurance funds flows as a "closed system" sub-set of the wider fiscal picture.

The chart below is derived from studying the last five years' National Insurance Fund accounts, mapping the cumulative flows over that period and calculating the related Barnett consequentials. As ever, it takes a lot more effort to understand complex reality than it does to take comfort in a misleading simplification:

This diagram shows that for the five-year period 2017 - 2021:

• Total UK National Insurance (NI) receipts were £678bn¹
  
  
• 19.0% of those receipts (£129bn) never reached the National Insurance Funds because they were allocated directly to NHS England. This is a formulaic allocation defined in the Social Security Administration Act 1992 - the percentage is effectively constant year-on-year²
  
  
• The funds allocated to the English NHS drove increases in the devolved Nations' Block Grants of £24bn (through the application of the Barnett Formula). This means the UK system of pooling National Insurance contributions (in combination with the Barnett Formula) controls the allocation of more than just National Insurance receipts. 
  
  
• There are separate NI Funds for GB and Northern Ireland. Transfers are made from the GB to the NI Fund "in order to maintain parity of balances" (ie. to keep the Northern Ireland NI Fund solvent). This is a direct and explicit fiscal transfer from GB to Northern Ireland of c.£0.6 billion each year.
  
  
• A mechanism exists whereby the Treasury can take funds from elsewhere and inject them into the NI Funds by way of Treasury Grants, but his has not been used in the last five years and is not forecast to be required in the foreseeable future³.
  
  
• 72.5% of UK NI receipts were used to pay State Pensions
   
• 5.6% of UK NI receipts were used to fund a limited subset of other welfare payments and to cover the costs of administering the system
• Other contributory benefits: the contributory components of Employment Support⁴ and Jobseekers' Allowances, Bereavement Benefits, Christmas Bonus⁵, Guardians' Allowance
• Other Payments: primarily redundancy costs (ie. making statutory redundancy payments when employers are unable to, normally due to insolvency)
  
• Administrative Costs: Mainly services provided by DWP and HMRC
  
  
• The remaining 2.9% of UK NI receipts were retained as a surplus, adding to the closing National Insurance Fund balances. Over the last 5 years the NI Funds have operated as a closed system (ie. no Treasury Grants have been required) and NI receipts being retained has led to a £20bn increase in the NI Fund balances. To put that another way: NI receipts into the Funds (net of the NHS allocation) have been greater than NI payments out by c.£20bn over the last five years.

How UK National Insurance Pooling Benefits Scotland

Armed with the above understanding we can immediately see why Rideout's "Our NI Contributions more than fund our State Pension" formulation is, at best, misleading. Apart from anything else, he's completely ignored the sums of money which flow via the allocation of NI receipts to the NHS.

Because we've put in the hard miles to map all this out, it's easy to look at the figures for 2021 alone. So we can see that in the most recent year (the one Rideout uses) 72% of UK NI receipts were needed to fund the State Pension.

The figure Rideout quotes for Scotland's 2021 State Pension costs accounts for 74% of Scottish NI receipts ... but he has omitted Scotland's population share of overseas pensions. Include Scotland's share of overseas pensions payments (as we need to for a like-for-like comparison, see below) and we discover that 77% of Scotland's NI receipts are required to pay Scotland's State Pension compared to 72% for the UK overall.

It's not hard to see where this is leading - but let's not get ahead of ourselves. We need to do some more leg-work to compare what Scotland "puts in" to the UK's pooled NI system with what Scotland actually "gets out".

What Scotland Puts in

National Insurance

This is an easy number - so easy in fact that even Rideout got it right. The UK National Insurance figures in GERS reconcile very closely with those we get from aggregating the GB and Northern Irish NI Funds and with those available directly from HMRC. So we can state with confidence that over this period Scotland's NI contributions accounted for 8.0% of the UK total.  

Other Taxes

This is a bit trickier to follow and frankly would be quite easy to miss - but this is Chokkablog; we do analysis properly here. 

When the devolved nations get their NI-allocation driven Barnett consequentials, that money doesn't come from NI contributions but from other taxes (check the flow charts above). Of course Scotland (and the other devolved nations) contribute to that pool of taxes, so we need to take that into account.

 

What Scotland Gets Out

NHS Funding: 

The question of how much of the allocated NHS funding goes to Scotland is not as easy to answer as one might expect. The relevant legislation merely states that amounts "shall be taken as paid towards the costs of [the English, Scottish and Welsh NHS] in such shares as the Treasury may determine".⁶.

Whilst one might assume this would be pro-rata to the NI contributions made by each nation, according to this Fraser of Allander blog, this Economics Observatory blog and this Institute for Government Explainer, NHS-allocated NI funds all go to the English NHS, with Scotland, Wales & Northern Ireland each then seeing funding increases via the Barnett Formula. This means that Scotland gets back a "Scotland vs England" population share of those NHS-allocated National Insurance funds.  This is to Scotland's net benefit, as both the Fraser Allander commentary⁷ and the Institute for Government analysis⁸ confirm.

This means that Scotland effectively "gets out" 8.2% of the NHS-allocation driven Funds

It's worth keeping a tally here: 22.2% of the Scotland's "money in"⁹ to the UK's NI system comes back via the Block Grant.

State Pension

The State Pension figures we get from the GB NI Fund¹⁰ reconcile well with the figures that come directly from the DWP.  These show us that Scotland (including a population share of overseas pensions) receives 8.6%¹¹ of State Pension payments out of the NI Funds.

For our running tally: 75.2% of Scotland's "money in" to the UK's NI system comes back via State Pension payments

---

Important aside: the "Spending in Scotland: State Pension" line of Box 3.2 in the latest GERS report mistakenly includes Scotland's population share of overseas pensions in the the first two years (but correctly excludes them in the last three). I have checked directly with the office of the Chief Economic Advisor to the Scottish Government and they have confirmed this presentational inconsistency. I share this here partly because it validates the robust audit-trail we're creating here - it's only by continually reconciling between the NI Fund accounts, DWP Data, HMRC data and the GERS report that this issue was spotted. To be absolutely clear: the Total Social Protection in Box 3.2 is correct and the integrity of the overall GERS figures is not being questioned; this is a relatively trivial issue with the breakdown of 2017 and 2018 data in that "for information" box. Caused me quite the headache though, I have to admit¹².

---

*PAUSE*

So through NHS-related Barnett consequentials and direct State Pension Payments, Scotland receives back 22.2% + 75.2% = 97.4% of the funds it contributes into the UK's pooled National Insurance system - and we've yet to cover other contributory benefits, operational costs and Scotland's share of any retained balance.

We're nearly there: one last push ...

Employment Support Allowance (Contributory) [ESAc]

Unfortunately the DWP data we have doesn't split Employment Support Allowance (ESA) between contributory and non-contributory benefits, but we can range the figures. The default assumption we're using is that Scotland's share of contributory ESA is the same as its 10.6% share of overall ESA (as per the DWP data). As a sensitivity, we can assume Scotland merely gets its 8.0% NI contribution share.

Job Seekers' Allowance (Contributory) [JSAc]

The DWP data doesn't split Job Seekers' Allowance (JSA) between contributory and non-contributory benefits either, but again we can range the figures. The default assumption we're using is that Scotland's share of contributory JSA is the same as its 9.8% share of overall JSA (as per the DWP data). As a sensitivity, we can assume Scotland merely gets its 8.0% NI contribution share.

Bereavement Benefits (contributory) [BBc]

The DWP data matches the NI Fund data very closely until the introduction of Statutory Parental Bereavement Pay (which by inference is not paid out of the NI Fund and accounts for less than 15% of the total).  We assume that Scotland receives a 9.0% share of NI funded Bereavement Payments (as per the DWP data).  

Maternity Allowance [MA]

The DWP data matches the NI Fund data, so we can use it directly. Although the numbers are relatively small, what leaps out here is that Scotland's 8.2% of the population receives only 6.4% of Maternity Allowance payments (as per the DWP data).

A quick check of relative fertility rates explains this apparent discrepancy: Scotland's latest recorded fertility rate was 1.29 compared to England's 1.66. Take the relative fertility rate of 1.29/1.66 and apply it to Scotland's 8.2% population share and you get 6.4%, exactly the same as the observed share of Maternity Allowance payments. If you've read this far, you're probably the sort of person who will appreciate the little buzz I got when I realised how neatly that explains what's going on here. 

[I often wonder how many people get this deep into these blogs - if you're one of them, please tweet me @kevverage with "I got the buzz" so I know I'm not wasting my time!] 

*PAUSE*

These four benefits above account for 5.6% of the funds Scotland contributes into the UK's pooled National Insurance system. Add that to the 97.4% we covered with NHS-related Barnett consequentials and direct State Pension Payments and Scotland has received back 103% of what it's paid into the UK's pooled National Insurance System. 

Even if we take the low-case sensitivity (ie. assume Scotland only receives an 8.0% share of UK payments for these benefits), Scotland would still have received back 102% of what it's paid into the UK's pooled National Insurance System.

Scotland is getting out more than it puts in, and we've still to cover the other NI funded costs and Scotland's share of the retained balance. 

So let's tidy up those loose ends now ...

Other NI Funded Costs

The remaining NI funded costs account for just 1.2% of the total, split as follows 

• 58%: Administration (primarily DWP and HMRC)
• 29%: Redundancy (ie. making statutory redundancy payments when employers are unable to, normally due to insolvency)
• 13%: Other

Frankly the assumption used here doesn't really matter, but it seems reasonable to allocate a simple population share of these costs to Scotland:

****

So now all that remains for us to do is add all of that up and allocate Scotland a share of the retained NI Fund balances:

And there you have it: 

• Scotland pays in 8.0% to the UK's pooled National Insurance system and gets back 8.6% of the spending that pooled system drives.¹³
   
• Scotland net benefits by over £750m a year¹⁴ (assuming Scotland lays claim to a population share of the NI Funds' retained balances¹⁵) 

Now. I recognise this has taken a bit more work than the "pick two numbers and pretend it proves your point" method beloved of Scottish nationalists - but isn't it a tad more satisfying to understand the complex reality?

****************

Footnotes

1. This figure (reverse engineered from the GB and NI National Insurance Fund accounts) reconciles very closely with HMRC's published data as we show later in the blog. The Total NI figure from the NI Fund accounts is 98% accounted for by direct NI Contributions, the balance being "compensation for statutory recoveries" and a few other bits and bobs - the assumption here is that the components used to calculate Total NI Fund Receipts are the same as those included within HMRC's definition of National Insurance Contributions - the reconciliation would suggest this assumption is sound.
   
   
2. To be absolutely correct: different percentages are applied to different classes of contribution (see Social Security Administration Act 1992), so changes in the contribution mix can lead to slight changes in the overall average contribution percentage. There is also a mechanism contained within that Act which would allow this allocation to be varied at the whim of the Treasury:  "The Secretary of State may, with the consent of the Treasury, by order amend any of [the allocation percentages] in relation to any tax year, by substituting for the percentage for the time being specified in that paragraph a different percentage".
   
   
3. As the NI Fund accounts explain: "the Social Security Act 1993 allows for money provided by Parliament to be paid into the NIF via a Treasury Grant if HM Treasury considers it expedient to do so". 
   
   The accounts go on to explain: "The report on the Up-rating Order published by the Government Actuary in January 2021 projected an increase in the balance of the Fund in the year ended 31 March 2022, and also projected that no Treasury Grant is likely to be required in that year in order to maintain the Fund above the targeted minimum balance of 16.7% of benefit expenditure. However, as a contingency, under section 2(2) of the Social Security Act 1993 (c.3), HM Treasury Ministers have made provision for a Treasury Grant of up to 17% of estimated benefit payments. This equates to a provisional facility of £17.5 billion."
   
   The last Treasury Grant was in 2016, when £9.6bn was used to top-up the GB Fund balance
   
   
4. Previously Incapacity Benefit
   
   
5. "Christmas Bonus is a tax-free payment of £10 paid to people in receipt of a qualifying benefit during the relevant week, normally the first full week in December"
   
   
6. From the Social Security Administration Act 1992:
   
   "From the national health service allocation in respect of contributions of any class there shall be deducted such amount as the Secretary of State may estimate to be the portion of the total expenses incurred by him or any other government department in collecting contributions of that class which is fairly attributable to that allocation, and the remainder shall, in the hands of the Secretary of State, be taken as paid towards the cost—
   (a)of the national health service in England;
   (b)of that service in Wales; and
   (c)of that service in Scotland,in such shares as the Treasury may determine.
   
   
7. Fraser of Allander Blog:  "If the UK Government funded its social care plans by increasing NICs, the funding implications are quite straightforward. The tax increase (whether to employee or employer contributions, or to the self-employed) would apply in Scotland. The Scottish budget would benefit from a Barnett consequential as a result of the increased spending in England. A £10bn increase to spending on social care in England would result in approximately a £1bn increase to the Scottish block grant.
   
   As is the case with income tax, the NICs raised in Scotland from an increase in NICs may not fully cover the sum of the additional Barnett consequential that flows to Scotland. This is seen as ‘fair’ in the context of a tax that is ‘pooled and shared’ across the UK."
   
   
8. This Institute for Government example shows £12bn of additional NIC allocation would result in a £1.1bn increase in Scotland's Barnett consequentials, £0.7bn for the Wales and £0.4bn for NI - a rounded total of £2.2bn.
   
   Applying the relevant population share figures to the £12.0bn we get £1.16bn, £0.67bn and £0.40bn respectively - a total of £2.2bn. I think we can safely assume that the minor difference in the presented figure for Scotland is merely a function of rounding by the IoG (this often happens - to keep the total right, people sometimes use a bit of poetic licence when it comes to rounding).
   
   
9. Scotland's "money in" = £54,401m of National Insurance payments and £1,906m of contribution to the Barnett consequential (via other taxes) = £56,307.    
   
   
10. Confusion can arise from the fact that the DWP figures cover GB only (because the Northern Ireland Executive administer and report pensions separately) and also because the NI Funds report the (State pension) Christmas Bonus as a separate line-item. Note also the DWP figures and the National Insurance Fund figures do include overseas pension payments.
    
    
11. The 8.6% figure is sometimes (including by me in the past) mistakenly quoted as 8.8%.  This is because when filtering the supplementary-expenditure-database which accompanies GERS for "State pension", the figures that appear in the "UK" column relate to GB only, they do not include Northern Ireland (it appears the Northern Ireland State Pension figures are not split out from other Northern Ireland welfare payments)
    
    
    
    
12. The supplementary-expenditure-database provided with GERS (which reconciles with the overall GERS figures) does include a consistent population share of overseas pensions, so this issue is limited to the "for information" Box 3.2 within the GERS report (ie. it doesn't affect the integrity of the total government spending figure used in the GERS deficit calculation). I have contacted the office of the Scottish Government's Chief Economist and they have confirmed that Box 3.2 does include this error (the State Pension line is intended to be just the "in Scotland" figure, so Scotland's share of overseas pension payments should not be included). For our purposes, we are interested in Scotland's total pension payments which should include a population share of overseas payments, as included within the "Share of benefit spending outside UK and corporate spend" line in that table. The fact the State Pension figure drops between 2018 and 2019 really should have been a warning that something was up:
    
    
    
    
    

    Our reconcilaitions also highlight a couple of relatively small differences in stated NI figures - but these are not material to our conclusions

    
    
    
    
    
    
13. This is not a surprising finding. By looking at a ring-fenced sub-set of tax and spend (covering c.15% of Scotland's total Public Sector Expenditure) we are seeing a similar picture to that we see when we look at the total tax and spend numbers: Scotland benefits directly from the UK's system of pooling and sharing because Scotland receives a higher percentage of spending than its contribution to revenue.

    Over the last five years: 

    • This analysis shows Scotland accounted for 8.6% of UK public spending driven by the pooling of National insurance while generating just 8.0% of National Insurance receipts (and related other tax contributions) - it put in less than it got out.
      
      
    • GERS figures show Scotland accounted for 9.2% of UK public spending while generating just 7.9% of tax receipts - it put in less than it got out.
      
    
    

    The fact that the relative scale of Scotland's benefit is lower for this ring-fenced sub-set of all tax and spend is not surprising. The majority of spending considered here relates to the state pension and other contributory benefits where the same policy is applied UK-wide - Scotland's greater share of spending is therefore entirely explained by greater like-for-like need; there is no policy difference driving the higher spend (as there will be in devolved expenditure areas). 
    
    
    
    
14. We can now fully explain why Rideout's interpretation of the 2021 figures he's used is just wrong:
    
    He is correct to use £11.5bn for Scotland's NI contributions, but he missed £1.3bn of benefits Scotland gets back - so rather than the £0.5bn surplus he claims, Scotland benefitted from sharing in the UK's pooled NI system by £0.9bn in 2021 alone:
    
    
    
    
    
1. When he quotes £8.5bn for State Pensions he is only quoting the in Scotland figure, so he's missing £0.3bn for Scotland's population share of overseas pensions (see table in note 12).
   
   
2. What he claims on the Zoom call is £2.5bn of "other benefits" is in fact £3.0bn, so he's missing £0.5bn of benefits. I honestly have no idea where he gets his £2.5bn from: the lowest figure I can imagine him getting would be if he took a population share of all non State Pension NI fund payouts, but that would be only £0.7bn -- then even assuming he made the "low-side" mistake of taking the same share of Scotland's NI payments to NHS Scotland as for the GB NI fund, that would be £2.2bn, giving an "explainable mistake" total of £2.9bn
   
   
3. I imagine he has not noticed the £5.7bn surplus retained in the NI Funds in 2021, so he will have missed Scotland's £0.5bn population share of the movement in the NI Funds
   
   
   
15. This analysis suggests that Scotland might struggle to justify that claim given it has historically been a consistent net beneficiary of the system - but there would be a lot of noise if we ever came to splitting the UK's assets on separation, and a blanket "population share" assumption seems likely