Pension Reform in the UK: from contribution to participation
Ashwin Kumar[1]
1Introduction
In 1997, the state pension problem facing the new UK government was not one of funding but of adequacy.
A series of reforms by the administration of 1979-1997 had seen a fall in projected spending on state pensions as a proportion of GDP. Rather than the ‘demographic time-bomb’ of popular perception leading to a funding crisis, the UK was one of few developed countries to face no serious future funding problems in its state pensions. In fact, as a proportion of GDP, state spending on transfers to the elderly was projected to fall by one-third over the next fifty years.
This reduction in future spending arose out of reforms in the 1980s that reduced the future value of both the flat-rate basic state pension and the State Earnings-Related Pension (SERPS). The reforms had also widened the options for individuals making their own private provision.
But those at the bottom of the earnings distribution were least able to deal with the effects of these reforms. Firstly, the basic state pension would no longer guarantee a minimum level of income around the level of Income Support, the main means-tested form of support.
Secondly, the earnings-related pension, SERPS, which was being reduced in value, by its design already provided least support to lower earners. Thirdly, lower earners were least able to take out the new private pensions (known as personal pensions) because their charging structures tended to be designed for those with more to invest.
Short and medium term estimates did foresee the same or a smaller proportion of pensioners reliant on Income Support, probably as a result of cohort effects increasing the average amounts of occupational pension that future pensioners could expect to receive. But the longer-term estimates suggested that the proportion on Income Support would then start to rise as some of the 1980s reforms matured.
One potential effect of a greater proportion of the pensioner population relying in the future on means-tested support might be to discourage voluntary saving on the part of low and moderate earners. Those who perceived, correctly or incorrectly, that they would be unable to save enough to take themselves clear of the thresholds for means-tested support would have a large disincentive to save.
The problem, then, was not how to spend less but how to spend more on the lower paid at the same time as encouraging private saving.
The government’s response, A new contract for welfare: Partnership in Pensions, was published in December 1998.[2] The proposals contained within this document represented a fundamental shift in the concept behind state second tier pension provision in the UK. This shift was from the contributory and earnings-related principle of the 1970s to the wider entitlement of a participatory principle and to the flat-rate benefit principle enshrined in the Beveridge Report of 1942.[3]
This paper seeks to draw out the implications of the reform proposals, over time and across the earnings distribution and to place them in the wider context of the concepts behind pension provision.
AThe Current System
2Overview of the UK state pension system
2.1The pre-1979 position
In the UK, the first tier of state pension provision is the basic state pension, introduced by the post-war government in 1948. Anyone who works and earns above a low earnings threshold[4] for enough years receives a flat-rate pension, currently worth £66.75 per week.
Anyone claiming unemployment benefits, aged 16-18, aged 60 or over or on sickness benefits receives credits automatically.[5] Anyone who cares for a child under 16 or spends at least 35 hours a week caring for a disabled person for a full year also has their basic state pension rights protected.
In 1975, the government committed itself to uprating the basic state pension in line with the higher of average earnings and prices.
The second tier of UK state pension provision is the State Earnings Related Pension Scheme (SERPS), introduced in 1978. This is an earnings-related pension for employees only. Those who earn more than the low earnings threshold pay National Insurance (NI) contributions as a proportion of their earnings up to a higher threshold.[6]
In return for these earnings-related contributions, employees build up rights in SERPS. The SERPS pension at retirement was to be worth 25% of a person’s average earnings between the lower and upper thresholds and the average was to be calculated on the basis of the best 20 years of each person’s working life.[7]
Employees could, however, contract out of SERPS. This meant that employees and their employers paid lower NI contributions and instead the employer guaranteed to provide a pension at least as good as under SERPS. The level of the reduction in NI contributions (known as the contracted-out rebate) was calculated to be ‘actuarially fair’ and to represent the reduced cost to the government of not having to pay that employee their SERPS pension in retirement.[8]
From the state point of view, therefore, contracting-out represented the bringing forward of expenditure: the payment of the second tier pension was made during the employee’s working life (in the form of lower revenue from NI contributions) instead of during the employee’s retirement. From a macroeconomic point of view, if an employee contracted-out of SERPS into a funded pension scheme,[9] contracting-out represented a switch of government expenditure from an unfunded to a funded pension.
In simple terms, the state pension system as it then stood provided a pension broadly equal to the LEL + 25% of 20-best-years-average earnings. But employees could choose to replace the second part of their state pension with at least an equivalent amount of occupational pension.
2.2The reforms of the 1979-1997 administration
The 1979-97 administration reformed state pensions on a number of fronts. From 1980, the basic state pension was uprated in line with prices only.[10] Secondly, for those who retired after 1999, the accrual rate in SERPS was to be reduced from 25% to 20%.[11]
Thirdly, the method of calculating a person’s average earnings for SERPS purposes was to change from the best 20 years to all 49 years of the working life. To be phased in one year at a time from 1999, the scheme was to be redesigned so that there was no longer any ‘best years’ calculation. Instead, the SERPS pension would be 20% of lifetime average earnings between the two earnings thresholds.
The combined effect of these reforms was that SERPS had been transformed into a scheme where, in each year, employees built up SERPS rights of 1/245 of earnings between the lower and upper earnings thresholds.[12] Each year’s accruals would be revalued in line with average earnings to the year before reaching State Retirement Age and added together to give the total SERPS pension at retirement.
2.3Means-tested support for pensioners
Underpinning the state pension system is means-tested support for those who had failed to build up enough pensions during their working life. The main benefit is Income Support, which is currently set at £75 per week for a single pensioner and £116.60 for a pensioner couple.[13] The benefit is withdrawn at a rate of 100% for income up to these levels. Each pound of income from state, occupational or personal pensions results in the loss of one pound of Income Support.
On top of Income Support sits help to pay rents and to pay local taxes (council tax). Any pensioner unit[14] with income up to the Income Support level receives, broadly speaking, the entirety of their rent as Housing Benefit and the entirety of their council tax bill as Council Tax Benefit.
However, for those with income above the Income Support level, Housing Benefit is withdrawn at a rate of 65%, i.e. 65 pence of Housing Benefit is lost for each pound of income above that level. Council Tax Benefit is simultaneously withdrawn at a rate of 20% of income above the Income Support level.
Figure 1: Net income of a single pensioner facing rent of £45 per week and council tax of £9 per week[15]
Single pensioners therefore face an effective marginal tax rate of 100% on income up to £75 per week. Those in rented accommodation then face an 85% rate on income above the Income Support level until their Housing Benefit and Council Tax Benefit run out. Those who own their home outright cannot receive Housing Benefit and so face a rate of 20% on income above the Income Support level until their Council Tax Benefit runs out.[16]
3Level of pension provided by the current system
The value of the state pension for future retirees is falling relative to average earnings for a number of reasons. The basic state pension is being uprated only in line with prices. This effect tends to dominate and results in the total pension expected declining over time. Up to 2027, retirees are also affected by the reforms of the 1979-97 administration restricting the future value of SERPS.
For later retirees the primary driver in changes to the value of SERPS is the movement of the earnings thresholds between which NI contributions are paid and upon which SERPS rights are accrued.
Figure 2 shows the expected total state pension under the current system for people who work for 49 years from 16 to 65 and whose earnings go up exactly in line with average earnings throughout their working life.[17] The expected pension is shown in 1999 earnings terms where the expected value of the pension is deflated by the assumed increase in average earnings to show how the pension is expected to vary relative to average earnings.[18]
Where the government achieves its aim of uprating means-tested support for pensioners in line with earnings, the Minimum Income Guarantee (the new name for Income Support for pensioners) is expected to remain constant in 1999 earnings terms. What we see from Figure 2 is that under the present system by 2063 we would expect even average earners who rely entirely on the state system to retire below the level for means-tested support.
Figure 2: Expected pension at retirement under current system, by year of retirement, in 1999 earnings terms[19]
Looking across the earnings distribution for a particular year produces a similar picture. As we see in Figure 3, for the cohort that retires in 2051, a substantial section of the earnings distribution can expect to retire below the level for means-tested support.
Figure 3: Expected pension at retirement in 2051 under current system, by lifetime weekly earnings,[20] in 1999 earnings terms[21]
BThe Pensions Green Paper
4The Government’s proposals
The UK government published its proposals for pension reform on 15 December 1998[22]. The proposals had four main planks:
aThe State Earnings Related Pension Scheme (SERPS) to be reformed and renamed the State Second Pension.
bA commitment to uprate Income Support for pensioners in line with average earnings as funds allow, renaming it the Minimum Income Guarantee.[23]
cA new form of private pension, called Stakeholder Pensions, to be launched with lower charges and easier access for the lower paid.
dAn improvement in awareness of pensions, including the development of annual statements of projected pension income.
4.1The State Second Pension
Expected to commence in 2002, the State Second Pension (S2P) is a reform of SERPS and as such carries over many of its features. For example, the number of working years required to obtain the maximum amount of S2P will be 49 and the rule for inheritance of the rights accrued by a deceased partner are to remain the same as under SERPS.
However, in its purpose the new pension is fundamentally different with its two-stage shift from the earnings-related SERPS to a flat-rate pension. The State Second Pension will differ from SERPS in three key respects:
a£180 per week underpin
Anyone earning between the Lower Earnings Limit (LEL) and £180 per week will be treated for S2P purposes as if they were earning £180 per week. The level of this underpin will be uprated each year in line with earnings.
bNew accrual structure
Earnings between the LEL and the underpin will accrue S2P rights at 40%, i.e. twice the rate under SERPS. This extra generosity for lower earners will be tapered away so that those earning above an upper threshold of around £400 per week will accrue the same rights as they would have done under SERPS.
cCredits for carers and some people with a long-term illness or disability.
Anyone caring for a full year for a child of up to five years of age[24] or caring for a disabled person for at least 35 hours a week will receive a credit at the underpin level. Those of the long-term disabled who, by the time they retire, have worked for at least 10% of their working life (usually five years), will receive a credit for each year that they were in receipt of long-term disability benefits.
Figure 4a compares the proposed accrual structure with that of SERPS by showing how much an employee would accrue under each system if S2P were introduced in 2000.
Figure 4a: Accrual structure of S2P Phase One compared with SERPS
Phase Two of the State Second Pension is expected to be introduced in 2006. Under Phase Two, the benefit paid to those contracted-out of S2P and those who remain in the state scheme will diverge.
Those who do not contract out (referred to as being ‘contracted-in’) will face a flat-rate accrual structure. Earnings above the underpin will not accrue any pension and everyone will receive the same pension as those at the underpin level.
Figure 4b shows the accrual structure if Phase Two were introduced in 2000.
Figure 4b: Accrual structure of S2P Phase Two
5Long-term effects of the State Second Pension
5.1Accrual structure
Whilst the accrual structure looks relatively simple when shown for a single year, the dynamic picture is more complex.
With the LEL and UEL rising with prices, they are falling relative to earnings. This means that the band of earnings with a 40% accrual rate is widening and, the 10% band of earnings is shrinking.
Figure 5: NI contribution and S2P thresholds in 1999 earnings terms
The effect of these movements in the bands is that all individuals, across the earnings distribution will do better under S2P Phase One (or S2P Phase Two contracted-out) than under SERPS. Further, whether or not a person does worse contracted-in under S2P Phase Two than under SERPS depends on the width of the relevant bands in the years that they work.
5.2Distributional effects
With the State Second Pension to be introduced in 2002, the first cohort, therefore, to have spent a full working life under S2P will retire in 2051. Figure 6 shows the lifetime effects across the earnings distribution of S2P for this cohort.
Figure 6: Expected pension in 2051 under S2P, by lifetime weekly earnings in 1999 earnings terms[25] (See Figure 3)
As we might expect from the design of the scheme, S2P will provide substantially higher levels of pension than SERPS at the bottom of the earnings distribution. The extra benefit will be greatest for those earning just above the Lower Earnings Limit and will remain substantial further up the distribution.
For the cohort that retires in 2051, the switch to the flat-rate Phase Two will not leave higher earners who remain contracted-in much worse off than under SERPS. Those who earn on average more than £370 per week are likely to be no more than £2.50 per week (in 1999 earnings terms) worse off. Those who contract-out will be better off than under SERPS all the way up the earnings distribution.
We can also see that anyone who works (or receives S2P credits) for the entirety of their working life will retire with a state pension above the level of the Minimum Income Guarantee, putting themselves above the level of the 100% effective marginal tax rate.
5.3Costs and financing
State support for the elderly in the form of state pensions and Income Support for pensioners currently costs 5.4% of GDP. Under current policies, this proportion is projected to fall to 3.6% by 2050. The effect of the proposals in the Pensions Green Paper is to halve this fall so that, by 2050, the government can expect to be spending around 4.5% of GDP on pensioners.
If the extra cost of the State Second Pension were raised entirely from NI contributions, the potential reduction in contribution rates by 2050 would be reduced. Phil Agulnik[26] has estimated that, instead of a fall of 4.25 percentage points in contribution rates, financing the State Second Pension would require the fall to be limited to only one percentage point. However, lower expenditure on means-tested benefits would also allow a 0.6 percentage point fall in income tax rates.
6Adequacy of the State Second Pension
The State Second Pension is heavily redistributive. It provides a very large boost to the pensions of the lowest earners and, if financed through NI contributions, will tax those with higher earnings to pay for it. It also ensures that those who work for a full working life will retire above the level of the Minimum Income Guarantee. Has the state pension problem been solved?
The key question to ask is how successful will the new system be at keeping people above the Minimum Income Guarantee level? The Department of Social Security estimated that, had SERPS remained but an earnings-uprated Minimum Income Guarantee been introduced, approximately 1 in 3 pensioners would have had pension income lower than the Minimum Income Guarantee.[27] The State Second Pension was estimated to reduce this proportion to 1 in 4.
Although these estimates are subject to a high degree of uncertainty, it is apparent that, despite the extra generosity of the State Second Pension, substantial numbers of pensioners will still find themselves below the level of the Minimum Income Guarantee.