THE ROLES OF THE
PUBLIC AND PRIVATE SECTORS
IN THE UK PENSION SYSTEM

Alan Budd and Nigel Campbell
HM Treasury(Note)



ABSTRACT

The UK is recognised as one of the few major economies which does not face a serious long-term problem of pension (social security) payments. The present value of net public pension liabilities is estimated at 5 per cent of GDP compared with over 100 per cent in Japan, Germany and France. As the authors explain, there are two principal reasons for this. First, the basic state pension has been held constant in real terms since 1980. It is currently about 15 per cent of average earnings. If this policy is continued future reductions in contribution rates will be possible. Second, there is a second-tier State Earnings-Related Pension Scheme but only 17 per cent of employees belong to it. Most employees belong to funded private occupational schemes. Membership has been encouraged by regulation and favourable tax treatment.

Note: The views expressed in this paper are those of the authors and not necessarily those of HM Treasury. Alan Budd is Chief Economic Adviser and heads the Macroeconomic Policy and Prospects Directorate. Nigel Campbell is the Economic Adviser in the Treasury's Social Security team. We are grateful to David Blake, Alexi Chan, Richard Disney, Marty Feldstein, Andrew Young and participants at the National Bureau of Economic Research conference for valuable comments on earlier drafts.

    INTRODUCTION

  1. This paper describes the current arrangements for the provision of pensions in the United Kingdom. In particular it describes the means by which the current and future public sector cost of pensions has been controlled or reduced and private sector provision has been encouraged.

  2. The outstanding feature of the UK pension system is that, under current policies, public expenditure on pension provision will remain modest, compared with other industrial economies. For example Chand and Jaeger (1996) estimate that the present value of the difference between the UK's public pension expenditure and revenue up to 2050 is 5 per cent of GDP with existing policies and current contribution rates. This compares with a ratio of 26 per cent for the US and above 100 per cent in Japan, Germany and France. Contribution rates, on current actuarial projections, will fall by 4 percentage points between now and the middle of the next century, in stark contrast with conditions elsewhere. The projected success in constraining public expenditure on pensions helps to explain, in large part, why the UK avoids the longer-term fiscal crises forecast (on current policies) elsewhere. Shigehara (1995) estimates that the level of General Government Financial Liabilities in the UK will be negative (ie there will be net assets) in 2030 compared with liabilities of 300 per cent of GDP in Japan and 100 per cent or more in the US, Germany, France and Italy.

  3. This can partly be explained by more favourable (or less unfavourable) demographic developments. (The "greying" of the UK population occurs later, and less dramatically, than elsewhere.) But the most important cause is the fixing of the value of the basic state pension in real terms. In addition the future cost of the state's second-tier pension has been significantly cut by reducing the generosity of the benefits and by encouraging people to leave the scheme and move to private provision. Although the ratio of pensioners to the population of working age is forecast to rise from 30 per cent in 1995 to 38 per cent in 2030, the ratio of public expenditure on pensions to GDP is expected to fall over the same period from 4.2 per cent to 3.3 per cent. One result of these changes is that the state pension (and other social security benefits) only accounts for about one-half of average pensioner income. This ratio will continue to fall in the future.

  4. Another feature which distinguishes the UK from many other industrialised countries, particularly those in continental Europe, is that most of its occupational pension liabilities are already funded. UK private sector pension funds have £600 billion worth of investments, more than the rest of the European Union put together.

  5. However current arrangements are not universally accepted as satisfactory. In particular, the current level of the basic state pension is slightly lower than the level of income support (a general means-tested benefit). The final sections of the paper describe some proposed reforms to the system.

  6. Broadly the UK system1 consists of a flat-rate basic state pension based on contributions. There are also means-tested welfare benefits for those deemed to have inadequate income.2 Second-tier pensions for employees are provided either by the State Earnings-related Pension Scheme (SERPS) or by occupational or personal pension schemes. All employees (with incomes above a lower limit) must either belong to SERPS or to an appropriate private scheme. (There are currently about 22 million employees in the UK.) The state system (flat-rate and SERPS) is pay-as-you-go. Private schemes are usually funded, although there are unfunded schemes for some public sector employees. Approximately 73 per cent of all employees are members of occupational pension schemes or have personal pensions. 12 per cent are only in SERPS. (About 5 per cent of employees are in both occupational schemes and SERPS.) The 15 per cent who are not covered consist mainly of those on low incomes. Over their lifetimes such individuals are likely to accrue some entitlements.

    THE STATE PENSION SYSTEM3

    The Basic State Pension

  7. The full basic state pension is currently (1996-97) £61.15 per week. This is about 15 per cent of average full-time male earnings.4 The pension is taxable. Current policy is to raise the pension annually in line with retail price inflation. Between 1975 and 1980 the policy was to uprate the pension in line with prices or average earnings, whichever was the greater. Before 1975 ad hoc increases had kept roughly in line with earnings increases since 1948.

  8. Entitlement is based on a contributions record. Payment of the full level of benefit depends on the payment of contributions, or the receipt of credits, for about 90 per cent of a working life (currently 49 years for a man and 44 years for a woman). Contributions or credits of 44 or more years for men and 39 or more years for women entitle contributors to the full pension. Pensioners over 80 can receive the state pension (at a reduced rate) on a non-contributory basis.

  9. Married women receive pensions on the basis of their own or their husband's contributions, whichever is more favourable. Almost all male pensioners receive the full pension whereas only about 1.7 million out of 6.1 million female pensioners receive the full pension on the basis of their own contributions. This position is steadily being changed by the introduction, in 1978, of home responsibilities protection (HRP) which provides credits for women who are out of work and caring for children or other dependants and, in the same year, by the introduction of a requirement that married women must contribute to the National Insurance System and thus acquire entitlement to full pensions. An increasing proportion of women will thus become entitled to the full basic pension on the basis of their own contributions or credits.

  10. The basic pension is currently about 9 per cent lower than the level of income support for single people and about 6 per cent lower for couples. The gap is even larger for older pensioners because income support increases at age 75 and age 80.

  11. The cost in 1994-95 of the basic state pension was £27 billion (4 per cent of GDP). Means-tested benefits to pensioners cost a further £8.5 billion.

  12. The ratio of the state pension to average earnings has fallen from about 20 per cent in 1977-78 to about 15 per cent now. If real earnings grow by 1 1/2 per cent a year (which is a modest estimate relative to past trends) the ratio will fall to 9 per cent by 2030. For the low paid the basic state pension is likely to remain a substantial proportion of their retirement incomes.

  13. The projected cost of the basic state pension is shown in Table 1. For comparison the table also shows the cost if the basic pension is uprated in line with earnings (at an assumed growth of 1 1/2 per cent a year).

    Table 1: Projected cost of basic pension

    £ billion, 1994-95 prices5

    1994-95 2000-01 2020-11 2020-21 2030-31
    Price Indexed 26.9 29.8 33.6 35.2 41.9
    Earnings Indexed 26.9 32.6 42.6 51.8 71.6

    The State Earnings-Related Pension Scheme

  14. The State Earnings-Related Pension Scheme (SERPS) serves two functions in the UK system. It provides a second-tier pension for its members (about 17 per cent of total employees) and it establishes the minimum (either in terms of benefits or contributions) for those who can be permitted to opt out of the state's second-tier system and to use occupational or personal pension schemes instead.

  15. As subsequent sections will show, SERPS has changed significantly since it was first introduced. It is an earnings-related pay-as-you-go scheme. Benefits are based on earnings between the lower and upper earnings limits (described below). In the long term, on current rules, the pension received by members of SERPS with a full record will be 20 per cent of average reckonable6 earnings over their working life, revalued in line with average earnings to retirement age.

  16. The upper earnings limit is adjusted in line with prices rather than earnings. The value of the SERPS pension will therefore decline relative to average earnings through time.

  17. As mentioned, about 17 per cent of employees are contracted into SERPS, of which about 65 per cent are women. About 70 per cent of those in SERPS earn less than £10,000 a year. (Average full-time male earnings in the UK are £20,000 a year.) The self-employed do not qualify for SERPS.

  18. The projected cost of SERPS is shown in Table 2.

    Table 2: Projected cost of SERPS

    £ billion, 1994-95 prices5

    1994-95 2000-01 2010-11 2020-21 2030-31 2040-41 2050-51
    1.9 4.2 8.4 10.9 12.0 10.2 9.9

    Contributions

  19. Contribution rates are set at the time of the annual Budget. There is a lower earnings limit (LEL) and an upper earnings limit (UEL) although employers pay contributions above the latter. These limits are adjusted annually in line with retail prices and in 1996-97 are £61 and £455 a week. The LEL is set at the rate of the single person's retirement pension rounded down to the nearest pound per week. The UEL is approximately 7.5 times the LEL. No contributions are paid if earnings are below the LEL. The rates for 1996-97 are shown in Table 3.

    Table 3: National Insurance contribution rates, 1996-97

    Percentage NIC rate

    Weekly earnings Employees Employers
    Contracted-in Contracted-out Contracted-in Contracted-out
    1st £61 Rest
    Below £61 0 0 0.0 0.0 -
    £ 61 to £109.99
    2% of £61 plus 10% of earnings between £61 and £455
    2% of £61 plus 8.2% of earnings between £61 and £455
    		 3.0 3.0 0.0
    £110 to £154.99 5.0 5.0 2.0
    £155 to £209.99 7.0 7.0 4.0
    £210 to £455 10.2 10.2 7.2
    Above £455 (UEL) 10.2 10.2 See note

    Note: The contracted-out rebate (described below) is only available in respect of earnings between the LEL and the UEL. Employers' NICs in respect of someone earning over the UEL are therefore 10.2 per cent of their earnings minus 3 per cent of the difference between the LEL and the UEL.

  20. The contracted-in rate is paid by or on behalf of members of SERPS. Those contracted out of SERPS pay lower contributions. This "contracted-out rebate" is currently fixed at 4.8 per cent of earnings between the LEL and the UEL (but will fall to 4.6 per cent from April 1997). 1.8 per cent is deducted from employees' contributions and 3 per cent from employers' contributions.

  21. It can be seen from Table 3 that there is an "entry fee" of £1.22 for employees as soon as their income reaches £61 per week. The marginal contribution rate for employees falls to zero when weekly earnings exceed £455. Employers' contributions are paid on allearnings once they reach £61 per week, ie an increase in earnings from £60 to £61 costs the employer £2.83. There are similar step increases as earnings move through the successive pay bands. An increase from £209 to £210 costs the employer £7.79.

  22. National Insurance contributions (NICs) are also used to finance unemployment benefits and a number of other benefits, including incapacity benefit and statutory maternity pay. In addition about 11 per cent of NIC receipts are paid towards the financing of the National Health Service. The state pension scheme is contributory in the sense that entitlement to the basic state retirement pension depends on a contribution record. However there need be no direct relationship between the payments made by an individual (or his employer) and the basic pension received. This follows automatically from the fact that the pension is paid at a flat rate while contributions depend on earnings. Also contributions are credited for those who are registered as unemployed or who receive certain Social Security benefits. As mentioned above, women can now receive credits for time spent caring for children. Employers' contributions do not gain any entitlement and it is possible to make employee contributions in the course of a year which do not count. Finally those without a full contributions record (who will largely consist of those who have earned below the LEL) can claim means-tested income support which is greater than the basic pension.

  23. The rather tenuous link between contributions and benefits has led Johnson and Stears (1996) to suggest that the contributory requirements should be abolished and replaced by entitlement based on residence. Successive governments have chosen to retain the contributory principle partly on the ground that receipt of the basic pension does not carry the stigma that may be attached to income support and partly because they like to emphasise the link between contributions and benefits. In some cases the contribution requirement saves public expenditure, since there will be some pensioners who receive less than the full basic pension but will not be eligible for income support (eg because they have too much capital). It is also true that the contributory principle means that increases in contribution rates have in the past not been subject to the same opprobrium as increases in income tax rates.

    The National Insurance Fund

  24. The National Insurance Fund (NIF) is one of eight funds and accounts used by central government for its revenue and expenditure transactions.

  25. National Insurance contributions are paid into the NIF, and contributory benefits are paid from it. The main contributory benefits are the basic retirement pension, SERPS, unemployment benefit, incapacity benefit and widow's benefit. Entitlement to all these benefits is based to some extent on contributions made. In addition, some NICs are used to fund part of the National Health Service.

  26. The Fund operates on a pay-as-you-go basis, so contributions by current workers are used to pay current pensions. The NIF has been topped up with a supplementary payment from general taxation in most years since the NIF was set up in 1948-49.

  27. Any pay-as-you-go scheme that pays benefits to all pensioners from the start would initially need significant financial support. This was true of the NIF, whose payments from general taxation averaged 34 per cent of NICs revenue in its first three years. A lower proportion was needed thereafter, but the transfer from the general tax pool to the NIF was still part of the design of the system until 1989. This "Treasury supplement" was worth 18 per cent of NICs revenue up to 1980-81, was reduced steadily to 5 per cent in 1988-89, and then abolished from 1989-90.

  28. As the NIF did not have a huge surplus by 1989, it is clear that the previous forty years of Treasury supplements effectively subsidised the contribution rates (or allowed for more generous contributory benefits) over that period.

  29. There were no payments to the NIF from general taxation in the four years from 1989-90. From 1993-94 however, it became possible for a "Treasury grant" to be payable to the National Insurance Fund if it appeared that the balance on the NIF by the end of the year would be below one-sixth of annual benefit expenditure. This Treasury grant was thus explicitly aimed at maintaining a working balance in the Fund rather than (as with the Treasury supplement) implicitly subsidising the contribution rate. The Treasury grant has been steadily reduced from 24 per cent of NICs revenue in 1993-94 to a forecast 5 per cent in 1996-97.

  30. The NIF ran deficits in three of the four years (1989-90 to 1992-93) in which no Treasury supplements or grants were paid. Only in 1988-89, when the NIF surplus exceeded the Treasury supplement by £1.4 billion, and 1990-91, when there was a NIF surplus of £1.5 billion and no supplement or grant, has NICs revenue exceeded NIF expenditure.

  31. The overall picture, therefore, is of a pay-as-you-go scheme whose revenue has not fully covered its expenditure in all but two of the years since 1948. The transfer from general taxation needed has been significantly reduced over the years, and is no longer built in as an implicit and permanent subsidy to the contribution rate.

    The self-employed

  32. In 1995 about 13 per cent of those in work were self-employed. Unlike employees, the self-employed are not required to make compulsory second-tier pension provision, and self-employment income does not give entitlement to SERPS. The self-employed pay different (and generally lower) National Insurance contributions.

  33. Self-employed people with earnings above £3,430 a year pay a flat rate contribution of £6.05 a week. Those earning below £3,430 can opt to pay the charge. These contributions give entitlement to the basic state retirement pension. There is an additional contribution of 6 per cent of profits between £6,860 and £23,660 a year (the latter equals the Upper Earnings Limit for employees, the former is rather higher than the Lower Earnings Limit).

  34. Although there is no state second-tier provision for the self-employed, they are eligible for tax relief on their personal pension contributions up to a cap. They do not however receive tax relief for pension contributions which are invested in their own business.

    Early retirement

  35. Labour force participation rates in the UK remain among the highest in Europe. But since 1980, there has been a marked fall in the participation rates of men in the last ten years of their working life. In 1980, 90 per cent of men aged 55-59 and 71 per cent aged 60-64 were either working or looking for work. By 1994, those proportions had fallen to 74 per cent and 52 per cent respectively.

  36. Many of those who retire early are eligible for occupational pensions. Those over 50 are able to use their personal pension fund to buy an annuity. They cannot however receive their basic state retirement pension or SERPS until they reach the state pension age (currently 65 for men and 60 for women).

  37. The unemployed receive National Insurance credits, which count towards entitlement to the basic state pension. Early retirees may be eligible for these credits. People aged under 60 would have to show that they are available and looking for work to qualify for credits and (if applicable) unemployment benefit. Men over 60 receive automatic credits independently of availability for work, while women over 60 are already eligible for the state pension.

    SERPS - A history

  38. SERPS was introduced by a Labour Government in the 1975 Social Security Pensions Act, which came into force in 1978.7 It added an earnings-related pension to the existing basic pension.

  39. As now, the scheme was based on contributions on reckonable earnings, ie earnings between the LEL and UEL. It originally provided a pension of a quarter of the individual's average revalued reckonable earnings (as defined at the end of this paragraph). Since one aim of the Act was to make the new pension arrangements mature rapidly, a full earnings-related pension was based on the best 20 years of earnings since SERPS began. Thus anyone who was earning for 20 years or more from 1978 could get a full pension if his earnings were high enough. (By contrast full entitlement to the basic state pension required, for men, contributions of over 90 per cent of a normal working life although again, at its introduction, there was an accelerated accrual rate for those working at that time.) "Average revalued reckonable earnings", for the relevant years, were calculated by revaluing each year's reckonable earnings in line with the change in average earnings for the whole economy between the original date and the retirement date. The pension was based on the average of the 20 best years. After retirement the SERPS pension was to be indexed to prices.

  40. Contracted-out occupational pension schemes had to provide a guaranteed minimum pension (GMP) broadly equivalent to the SERPS pension. They also had to meet a 'requisite benefits test' and therefore had, in effect, to have a higher accrual rate than SERPS. Those retiring from contracted-out schemes also had additional pension rights under SERPS but the GMP was deducted from their SERPS entitlement. Since SERPS was indexed to prices but the GMP was not, SERPS effectively provided some indexation for those in occupational schemes.

    The 1985 Green Paper

  41. In June 1985 the Conservative Government published a Green Paper, Reform of Social Security. By that date the bulk of spending on benefits for pensioners was still in the form of the basic pension. Less than 1 per cent of expenditure resulted from the earnings-related additions, although this was due to rise significantly in later years. The single person's basic pension was worth one-third of average take-home pay for a manual worker, and the pension for a married couple was worth half of average take-home pay for a manual worker.

  42. The Green Paper examined the implications of the basic state pension and SERPS over the following 50 years. It showed that if the basic pension were indexed to prices its cost would increase by 40 per cent over the following 50 years. If it were linked to earnings, its cost would almost treble. Over the same period the ratio of NI contributors to pensioners was expected to fall from 2.3 to 1.6. The Green Paper pointed out that the increased cost of the basic pension would benefit all pensioners equally. However the case was different for recipients of SERPS. Its earnings-related nature meant that the newly-retired would benefit more than older pensioners. Also half the extra cost would result from payments to members of contracted-out schemes (to provide indexation top-up to the GMP). The cost of SERPS (in 1985 prices) was expected to be about £24 billion in 2035, compared with a basic pension cost in 1985 of about £15 billion.

  43. The Green Paper pointed out that the peak of occupational pension scheme membership had been reached in the mid-1960s. There had been growth in the public sector but coverage of the total workforce was still only about half. The earlier decline in private schemes was thought to have been partly caused by uncertainty about the future of pensions policy. But the new system had produced little change. It also believed that the complexities of contracting out discouraged companies from setting up their own schemes, partly because they had to match the standards of a Defined Benefit scheme.

  44. The Government believed that occupational and personal pensions were the right way to provide second-tier pensions. It argued that it was preferable for individuals to make provision for such pensions rather than to leave responsibility wholly to the state and to require taxes to be levied on the next generation. It considered the following possibilities for change:

  45. The Green Paper estimated that these latter changes could cut the eventual cost of the scheme by about half. But, as the Green Paper said, the changes would be essentially negative in their impact.

    The Green Paper therefore proposed "a new partnership" between state provision and occupational and personal provision. Its long-term goal was the position in which the state provided a basic pension for all; and everybody also had an additional pension provided out of individual savings.

  46. The main proposals for reform were:

    47. The 1985 Reforms

  47. In the event the Government decided not to phase out SERPS. In its White Paper, Reform of Social Security. Programme for Action published in December 1985 it reported that a number of organisations, including in particular the Confederation of British Industry and the National Association of Pension Funds, had argued that the scheme should be modified rather than replaced. Opposition by employers partly reflected the perceived difficulties of matching the benefits offered by SERPS for the lower paid. That in turn was due to the fixed cost element of pension arrangements and to the low real returns on pension funds being achieved at the time. The Government accepted their arguments on the basis that it could meet two objectives:

  48. The changes to reduce the cost of SERPS followed those proposed in the Green Paper:

    49. The net result of the changes was expected to nearly halve the cost of SERPS in 2033-34 from £25.5 billion to around £13 billion at 1985 prices.

  49. At the same time the Government announced a strategy to extend individual pension provision, to widen the choice for the individual and to increase competition among pension providers. Seven points of the strategy were emphasised:

  50. These changes were additional to reforms that had already been introduced in the 1985 Social Security Act. The reforms had included:

  51. It can be seen that the proposals were designed to encourage private provision of second-tier pensions and to increase the 'portability' of pensions. The extension of personal pensions to employees would particularly favour those who expected to change jobs frequently and who tended to be penalised by Defined Benefit schemes. (The origin of occupational pensions was of course expressly designed to encourage long-term job tenure in such industries as banking and insurance.)

  52. The Government Actuary estimated that the NI contribution rate to pay for National Insurance benefits by 2033 would be 15 per cent, assuming that the basic pension was uprated in line with prices, compared with 18 per cent if SERPS had been left unchanged. He commented that the new rate was one which future generations should be able to afford. The changes to SERPS (along with other major reforms to the Social Security system) were introduced in the 1986 Social Security Act.

  53. The contracted-out rebate and the additional 2 per cent incentive made personal pensions a very good deal particularly for the young since the rebate was the same for all ages whereas the value of the SERPS benefit given up increases with age. The Department of Social Security's working assumptions were that about 500,000 people would take out personal pensions and the number might ultimately reach 1 3/4 million. In the event take-up reached 4 million by the end of April 1990. By 1993-94 the take-up had risen to 5.7 million, of whom about 4 million had rebates paid into their APPs.

  54. The National Audit Office (1990) commissioned a survey of the cost to the NI Fund of the rebate and the additional incentive for the period to April 1993. It estimated that the gross cost could be £9.3 billion and that the present value of the savings in payments of pensions was about £3.4 billion. Hence there was a net present value cost of £5.9 billion (in 1988 prices). However, this does not take account of the step change in personal pension take-up, which will have reduced state pensions yet further in the next century even after the end of the continuation of the 2 per cent incentive.

  55. A further shadow has been cast over this episode by the fact that some of those who joined personal pension schemes left occupational schemes which offered more favourable terms. Inquiries are still continuing into alleged cases of mis-selling of personal pensions during this period.

    56. The 1995 Pensions Act

  56. Further reforms to SERPS were introduced in the 1995 Pensions Act. The main purpose of the Act was to provide additional safeguards for occupational pensions following the Maxwell affair, and to equalise the state retirement age (which mainly affects the basic state pension) by increasing women's retirement age from 60 to 65. But at the same time changes were introduced to SERPS which further reduced its future cost.

  57. The 1986 Social Security Act had, as described above, enabled people to contract out of SERPS and buy so-called "appropriate" personal pensions (APPs) or join a Defined Contribution occupational scheme. The 1995 Pensions Act is expected to reduce further the cost of SERPS in 2030-31 by £6.7 billion in 1994-95 prices. Table 4 shows how the savings arise.

    58. Table 4: Reductions in SERPS expenditure in 2030-31
    as a result of the 1995 Pensions Act

    £ billion, 1994-95 prices
    Equalisation of pension age for men and women 1.9
    Annualisation 2.5
    Abolition of guaranteed minimum pension 2.4

  58. The raising of the state pension age for women to 65 will be phased in over a 10 year period from 2010. It follows the European Court of Justice ruling that men and women must receive equal pension treatment. Following annualisation, there is effectively an annual calculation of entitlement to SERPS, rather than a calculation at the point of retirement. This change prevents earnings below the LEL, in any year, generating SERPS entitlement. Under the old arrangements, earnings in individual years were uprated by the growth in whole-economy average earnings until retirement. The LEL at the retirement date, which had since 1980 been uprated only in line with prices, was then subtracted from these revalued earnings. As a result employees earning below the LEL were able to accumulate entitlement to SERPS. Even contracted-out employees (earning above the LEL) will have acquired around £1.80 a week worth of SERPS entitlement.

  59. The abolition of the GMP breaks the links between SERPS and contracted-out pensions. The state will no longer guarantee price indexation for contracted-out pensions. (The Act has however required private sector schemes to index future pensions by inflation or 5 per cent a year whichever is lower.)

  60. The 1995 Pensions Act also required that age-related rebates for contracted-out Defined Contribution schemes, including personal pensions, should be introduced from April 1997. This was aimed at removing the excessive incentive for younger people to contract out and to encourage older people to do so.

  61. Table 5 shows how the projected future costs of SERPS have changed since it was first introduced in 1975. The changes to the scheme in the 1986 Social Security Act and the 1995 Pensions Act are estimated to reduce the public sector cost of SERPS in 2030 from £41 billion to £12 billion (at 1994-95 prices). These latest estimates imply that the original scheme would have been about £5 billion more expensive (in 1994-95 prices) than expected at the time of the 1986 Act.

    62. Table 5: Effect of reforms to SERPS

    £ billion, 1994-95 prices
    1994-95 2000-01 2010-11 2020-21 2030-31
    Original regime (1975-86) 1.8 4.2 12.0 25.0 41.0
    After 1986 Social Security Act 1.8 4.2 9.2 14.5 18.7
    After 1995 Pensions Act 1.8 4.2 8.4 10.9 12.0

  62. The lower public sector cost of SERPS does not mean that large numbers of people now have only one-third of their previous second-tier pension provision. Instead, much of the reduction comes from more contracting-out, which yields public expenditure savings without necessarily affecting individuals' future pension entitlement.

    63. OCCUPATIONAL AND PERSONAL PENSIONS

  63. In 1991 (the last year for which figures are available), about 68 per cent of employees belonged to non-SERPS pension schemes. About one-third of these belong to personal pension schemes.

  64. All personal pension schemes are Defined Contribution (DC) schemes. Until the 1986 legislation, approved occupational schemes had to be Defined Benefit (DB) schemes. Occupational pension schemes remain predominantly of this type. The position in 1991 is shown in Table 6.

    65. Table 6: Number of employees in occupational pension schemes, 1991

    (Million)
    Type of
    scheme    		 Private Sector
    contracted out    		 Private Sector
    contracted in    		 Public Sector
    contracted out    		 Total
    DB 5.04 0.56 4.20 9.80
    DC 0.43 0.47 0.00 0.90
    Total 5.47 1.03 4.20 10.70

  65. Some employers have expressed interest in shifting from DB to DC schemes now that they are permitted to do so and still be contracted out8. This interest may be partly based on the view that DC schemes are less risky for employers than DB schemes. However, as the table shows 90 per cent of employees in private sector contracted-out occupational schemes were in DB schemes in 1991 and the following account is mainly based on schemes of this type.

  66. DB schemes are typically based on actual final salaries (or some average of final years' salaries). Employee contribution rates payable in 1991 are shown in Table 7.

    67. Table 7: Number of members of DB schemes paying various
    percentages of salary

    (Thousands)
    % of Salary Private sector
    Contracted out     		Public sector
    Under 2% 20 650
    2% and under 3% 150 -
    3% and under 4% 275 -
    4% and under 5% 730 5
    5% and under 6% 1690 2
    6% and under 7% 1125 2805
    7% and over 105 210
    Non-contributory 945 290

  67. It can be seen that a contribution rate of 5 to 7 per cent is typical though a significant number of private sector employees (typically in the financial sector) are in non-contributory schemes. Contributions by employers - for funded schemes - are based on actuarial estimates and can therefore vary from year to year according to the performance of the investment funds. The contracting-out arrangements impose minimum requirements on occupational schemes. The requirements correspond approximately to the benefits provided by SERPS which can in turn be linked to the contracted out rebate (currently 4.8 per cent but to be reduced to 4.6 per cent in 1997). Most occupational schemes are considerably more generous than this and provide combined contribution rates which are equivalent to up to 15 per cent of total salary. It should be emphasised that in the UK, unlike the case in some continental European countries, employers' contributions (as well as those made by employees) are explicit and placed in a separately identifiable fund. They are not simply recorded as a reserve.

  68. Pensions are essentially subject to "expenditure tax" treatment. Contributions by employers and employees are tax free (subject to Inland Revenue limits) and returns on invested funds are also free of tax. Pensions in payment are taxable but a tax-free lump sum may be withdrawn on retirement. The lump sum may be up to 1 1/2 times final salary or 25 per cent of the accumulated fund for a personal pension.9

  69. Benefits vary from scheme to scheme. In the private sector the normal retirement age for men and now for women following the implementation of equal treatment is predominantly 65. In the public sector more than 50 per cent of employees have a normal retirement age of 60 or less. In existing private sector schemes the normal retirement age for 50 per cent of women is 60. Following the 1995 Pensions Act, schemes must treat men and women equally. In private and public sector schemes the accrual rate is typically 1/80th per year for schemes which provide a pension and an additional lump sum and 1/60th for those which provide only a pension. (Since part of a pension-only scheme can be converted to a lump sum, the two systems of benefit are more or less equivalent.) Thus members of a typical DB scheme after 40 years' service in the same scheme retire with a pension of two-thirds of their final salary (part of which may be converted to a lump sum).

  70. Table 8 shows the benefits payable on death in service.

    71. Table 8: DB schemes and benefits payable on death in service Thousands
    Benefit payable Private Sector Public Sector Total
    None 10 - 10
    Lump sum 5540 4050 9590
    Refund of contributions 2870 1020 3890
    Pension to surviving spouse 5450 4200 9650
    Pension to surviving children 4110 4200 8310
    If unmarried, pension to a nominated person 1870 750 2620

    Notes: 1. Lump sum is subject to a maximum of four times annual salary if tax relief is to be provided.
    2. Schemes may appear under more than one category.

  71. Since there are 10.7 million members of pension schemes it can be seen that almost all members are entitled to a lump sum and/or a pension for the surviving spouse on death in service.

  72. Since the 1986 Social Security Act members of occupational pension schemes have been able to transfer their accrued pension rights to another occupational scheme or to a personal pension.

  73. Large schemes (ie with more than 5000 members) are usually self-administered with the funds invested by their own investment managers or an insurance company. Smaller schemes are usually insured schemes.

  74. By international standards there are few restrictions on the types of investment for funded occupational schemes. There are, for example, no limits on the proportions devoted to investment in equities or in overseas securities. There are limits on self-investment and schemes cannot provide loans to members or on residential property. As Table 9 shows, the UK has a high proportion of private sector pension funds invested in overseas equities.

    75. Table 9: Percentage of pension
    funds held in foreign assets     

                   1993

  Belgium        29
  Canada          9
  France          5
  Germany         3
  Japan          14
  New Zealand    34
  UK             27
  US              4

    Personal pensions

  75. Information on personal pensions is difficult to extract and analyse - see Johnson et al (1996) Chapter 5. It is estimated that in 1991, 24 per cent of employees had contracted-out personal pensions. The only investment into about three-fifths of these personal pensions was the rebate, equal to the contracted-out rebate together with any related incentives, which is paid into the scheme by the Department of Social Security. For young individuals in particular, a rebate-only pension should provide a better pension than SERPS.

  76. Personal pensions are available from a number of providers. They are mainly insurance companies although building societies, unit trusts and other financial organisations are permitted to administer pensions (at least up to retirement). As with occupational schemes there are relatively few restrictions on investments.

  77. Protected rights from appropriate personal pensions are now available from age 50 for both men and women. In general proceeds from personal pension funds must be used to purchase an annuity. Recent changes in legislation have increased the individual's freedom of choice between alternative annuity suppliers. They have also allowed the purchase of an annuity to be deferred until age 75 (with upper and lower limits on the sums that can be withdrawn up to that point).

  78. Personal pensions receive the same tax privileges as occupational pensions.

    79. Annuities

  79. In the UK, people "annuitising" any individual pension savings will normally stay with the insurance provider or, on occasion, take advantage of any "open market option" to withdraw their accumulated fund and purchase the annuity from a different insurer. Any penalty is usually relatively small (and, where there is a change of insurer, is presumably more than offset by the better terms offered by the annuity provider). This market at present is unaffected by contracted-out Appropriate Personal Pensions as these have been concentrated at younger ages. The impact of these on the annuity market will not be evident for about 20 years.

  80. Income withdrawals from a personal pension fund prior to purchasing an annuity has generally been limited to those with relatively large pension savings. One reason for this is that such people are in a better position to understand and accept the risks involved in betting against the investment market and their individual mortality. But it is also true that the expense charges appear to be relatively high and would be an onerous burden on smaller pensions savings.

  81. One particular aspect of the uncertainty over the date of annuity purchase is selection, ie individuals and/or the insurance companies trying to opt for arrangements which prove to be (or at least seem likely to be) most beneficial to them. This can reduce the validity of detailed historical analysis looking solely at what has happened in the annuity market before, although as relevant data become available, proper allowance can be made for such options.

  82. Insurance companies could aim to select people in the hope that they would die quickly. In a perfect market, another company would offer higher pensions at retirement to such "poorer lives". It is worth noting that some insurance companies in the UK offer enhanced pensions to such "impaired" lives. One company - Stalwart Insurance - has recently extended this approach beyond the very ill. They intend to offer annuities which pay more to those with medical conditions which are more likely to cause an early death (obesity, high cholesterol levels, high blood pressure and diabetes) but whose conditions are not life-threatening.

  83. Indexed annuities are widely available in the UK market.

  84. In summary, the UK pension annuity market is well developed and on a sound actuarial basis, so that the scope for selection is very restricted. It is only in newer aspects, such as perhaps income withdrawal, where statistical experience is lacking, although it is possible to use the data available to make reasonable assumptions about the potential impact on both those taking up such options and equally important, those not doing so.

    85. DISTRIBUTION OF PENSIONER INCOMES

  85. Average pensioner incomes have risen by about 50 per cent in real terms since 1979. Higher income from occupational pensions and from investments, both of which have more than doubled, account for most of this increase. Pensioners' earnings from employment have fallen somewhat, while state pensions plus benefits has gone up. Table 10 shows how the mean income from each source, adjusted for the falling proportion of pensioners who are single, has changed since 1979.10 (The raw figures show even higher average real increases in incomes.)

    86. Table 10: Average pensioner incomes
    1979 1993

    (£ per week, July 1993 prices)

    		 % real increase

    1979-1993

    Gross income

    of which:

    		 113.40 166.70 47
    State pension plus benefits 69.20 89.40 29
    Occupational pension 18.20 40.90 125
    Investment income 12.30 26.70 117
    Earnings 13.10 9.20 -30
    Other income 0.70 0.50 -25

  86. The increases in occupational pension and investment income reflect both higher real amounts on average and a bigger proportion of pensioners receiving these types of income. 73 per cent of pensioners had some investment income in 1993, up from 62 per cent in 1979. The average amount increased by 88 per cent over the same period. The proportion receiving occupational pensions rose by nearly one-half (from 43 per cent to 62 per cent), and the average occupational pension paid increased by 63 per cent.

  87. This increase in average incomes has been accompanied by a widening of the distribution of pensioners' incomes, as it has for the population of working age. However, Table 11 below shows that real incomes have increased significantly for all quintiles of the distribution.

    88. Table 11: Distribution of pensioners' incomes
    1979 1993

    (£ per week, July 1993 prices)

    		 % real increase

    1979-1993

    Mean 98.20 148.55 51
    10th percentile 54.40 67.30 24
    30th percentile 63.40 86.10 36
    Median 69.40 98.50 42
    70th percentile 86.30 142.40 65
    90th percentile 165.30 262.40 59

  88. Fewer pensioners are now at the bottom of the distribution of all household incomes. The bottom quintile of the income distribution contained 25 per cent of all pensioners in 1992-93, compared to 47 per cent in 1979.

    89. NOTES TOWARDS AN ASSESSMENT

  89. The main focus of this paper is on description rather than evaluation. As already discussed the main features of the UK system are:

    90. These sections discuss some aspects of the economic consequences of the UK system. Typically the relevant consequences are:

  90. A common element in assessing these consequences is the rate of return on pensions contributions. For example, Feldstein (1996) argues that contributions to the federal system are in effect a tax since the implied returns are lower than those available in private schemes. The following section discusses some calculations in respect of the UK system. We do not discuss labour supply effects, a discussion of which can be found in Dilnot et al (1994).

    91. Rates of return

  91. As Samuelson (1958) has shown, the return on a pay-as-you-go earnings-related scheme depends on the growth of the working population and of average earnings. In the UK the basic state pension, under current policy, is fixed in real terms, whereas contributions are related to earnings. It follows that the return on contributions will be different for different cohorts. In addition, since benefits are flat-rate, whereas contributions are earnings-related, there is a redistributional element in the basic state scheme. (However, as Dilnot et al (1994) point out, the effect is offset by the greater longevity of the higher paid.)

  92. Disney and Whitehouse (1993a) estimated rates of return for the state pension scheme as a whole. Their results - reproduced in Table 12 - show that the real rate of return on state pensions will be negative for men retiring after 2020. (The returns for women will be higher than for men.)

    93. Table 12: Intergenerational Real Rates of Return

    (% per annum)
    Cohort retiring in Mean 2nd decile Median 8th decile
    2000 1.0 0.5 0.0 2.2
    2010 0.2 -0.1 0.1 0.5
    2020 -0.9 -0.1 -1.2 -0.5
    2025 -0.7 -0.5 -0.9 -0.8

  93. These results precede the 1995 Pensions Act, which reduced both future benefits and contributions, and whose net effect on the rate of return was therefore unclear.

  94. Disney and Whitehouse (1993b) also attempted to estimate the effect of contracting out of SERPS on the above analysis. They believe that those contracted out may have a 0.5-1 per cent higher intergenerational rate of return than those who remain contracted in.

  95. The reason for these low estimated rates of return is that the basic state pension is assumed to be fixed in real terms. Current employees are paying for pensions which are higher (relative to average real earnings) than those that they will receive when they in turn retire. This effect adds to the normal effect whereby rates of return for successive cohorts fall as state pension schemes mature.

  96. Blake (1994) has estimated the rate of return from SERPS and from a typical contracted-out Defined Benefit occupational scheme. For the first 20 years, the real rate of return from SERPS was 6.7 per cent a year, about twice the 3.3 per cent from the occupational scheme. This reflected the generosity of SERPS when it was initially set up, not because SERPS paid a very high pension - it paid 25 per cent of the best twenty years' earnings, usually well below the occupational equivalent - but because contributions into SERPS were comparatively low and particularly benefited those close to retirement.

  97. After the 1985 reforms, the real rate of return on SERPS fell from 6.7 per cent per annum to 1.7 per cent. The 1995 Pensions Act reduced it yet further, to 1.2 per cent per annum. Thus the real return on SERPS varies between 6.7 per cent (for those who retire before 1999) and 1.2 per cent (for those who will retire after 2030).

  98. The rate of return on SERPS will not only affect savings (for those who remain members). It will also affect the decision on whether or not to join the scheme. As described earlier, the choice is complicated by the special temporary incentives to leave the scheme and by the fact that the contracted-out rebate was the same for all ages. That particularly encouraged the younger employees to leave. Optimal strategies which involve leaving and then re-joining SERPS, are described in Dilnot et al (1994). The contracted-out rebate will, from April 1997, be age-related, drastically reducing this incentive to switch into and out of SERPS.

  99. It is clear that contributions to the basic state pension, particularly for the higher paid, are in effect a form of taxation and can therefore give rise to deadweight costs of the type described in Feldstein (1996). The position on SERPS is rather more complicated since employees are able to contract out into a private scheme, including personal pensions. Those who remain members must, if they are acting rationally, believe that the returns on contributions to SERPS are at least as good as those in a private pension scheme (for an equal contribution). Since most of those who have remained in SERPS are among the lower paid, it is reasonable to believe that they are indeed acting rationally. It is, however, possible that the element of compulsion causes them to contribute more to pension schemes than they would choose freely.

    100. Macroeconomic effects

  100. This section provides a preliminary assessment of the extent to which the current system and changes to it over the past 20 years have affected private and public savings.

  101. The introduction of SERPS in 1978 extended earnings-related pensions, on a generous basis for older employees, to a wider section of the population than had previously had access to them. Since it was a pay-as-you-go system and was additional to the basic state pension, one might have expected it to replace, at least at the margin, individual savings for retirement and thereby to have reduced personal sector savings. As Feldstein (1996) has pointed out, the extent of displacement will depend, among other things, on the relative returns on the state scheme and on private sector schemes. As already described, Blake's calculations suggest that the return on SERPS, particularly for those approaching retirement, was significantly higher than those available on occupational pension schemes. This reinforces the view that personal savings are likely to have fallen as a result of the introduction of SERPS.

  102. The same arguments would suggest that the subsequent reductions in the benefits available from SERPS would have raised personal sector savings for those who remained members of the scheme. This effect should have been reinforced by those who left SERPS and moved to personal pensions, which are funded DC schemes.

  103. Blake (forthcoming) has made a preliminary study of whether changes in SERPS wealth can explain movements in the personal sector savings ratio. Using data from 1949 to 1993, he finds no significant effect from the ratio of SERPS wealth to income. In other words the introduction of SERPS does not seem to have lowered the personal sector savings ratio and the subsequent reforms do not seem to have raised it. He similarly finds no effects on savings from changes in basic state pension wealth, whereas one might have expected the freezing of the real value of pensions (which was a change from the previous policy) to have raised the personal sector's savings ratio. This apparent lack of response may reflect uncertainty about future pensions policy.

  104. If private savings were unaffected we can ask whether the changes in policy had any effect on public sector savings. At the time of the introduction of SERPS the total NI contribution rate, for contracted-in employees, was raised from 14 1/2 per cent to 16 1/2 per cent. There was also a 2 per cent surcharge, introduced as a fiscal measure. Thus the increased flow of revenue could have been used to raise public sector savings. However there are reasons for believing that this is unlikely to be the case, although it is difficult to define the counterfactual. The UK sets its fiscal policy in relation to the Public Sector Borrowing Requirement which is a cash flow measure. Its policy does not take explicit account of contingent liabilities such as future pension payments. Thus it will not have reduced its budget deficit (or increased its surplus) to allow for higher future payments under SERPS. In practice the increased current revenue from higher NI contributions will have been used to finance higher public spending or cuts in other taxes.

  105. Since 1988 the Government's fiscal objective has been a zero PSBR when the economy is at its trend level. Thus the reforms to SERPS, which reduced revenue flows, will not have led to a fall in public savings. Thus Blake's evidence plus the Government's fiscal objective suggest that total savings have been unaffected by the changes to the public sector pension scheme.

  106. Blake does however find a large and significant positive effect from the occupational pension wealth-income ratio. Thus it appears that the favourable tax treatment provided to occupational schemes has raised private savings and, given the exogenous nature of public sector savings, will have raised total savings.

  107. Finally Blake finds a small but significant negative effect from the personal pension wealth-income ratio. He suggests that a possible explanation for this result may be the use of personal pensions as collateral for loans. Thus the move from SERPS to a personal pension eases credit restraints and allows a higher level of consumption. It is, of course, possible that the credit effect is temporary and that in the long run the availability of personal pensions to employees will raise personal savings.

  108. The fact that income support, which is a means-tested benefit, is higher than the basic state pension is likely to cause some discouragement of savings for the lower paid. There is in effect a 100 per cent marginal tax rate on income from savings between the basic state pension and income support. Since the gap between the two is between 6 and 9 per cent, the disincentive effects could be significant11, although this will be offset by the requirement for all employees earning over £61 a week to make second-tier provision. Approximately 17 per cent of pensioners receive income support.

    109. PROPOSALS FOR REFORM

  109. If the present system is maintained the real value of the basic state pension will remain constant and hence fall relative to average earnings. At the same time the value of the SERPS pension and of the compulsory element of contracted-out pensions will also fall relative to average earnings. If the real value of income support is also held constant, the increased level of entitlement to earnings-related schemes should reduce the need for payments of income support to pensioners. However those pensioners with no or only a small amount of earnings-related pension will have a low level of entitlement relative to average earnings in the longer term.

  110. This expected outcome is consistent with the Government's approach to public expenditure and to individual choice and is consistent with the Beveridge approach to pensions. An adequate level of income (but low relative to earnings) will be available to pensioners, mainly through the contributory basic state pension. Compulsory second-tier provision provided either by the state or by the private sector will provide a fairly modest earnings-related pension. Beyond that, the choice will be up to individuals and employers, assisted by tax incentives. The public sector cost of pensions will fall relative to GDP and, for given demographics, NI contribution rates will be reduced.

  111. A more radical application of this approach might raise questions as to why there is any compulsory second-tier provision at all and why such provision should still involve, at least to some extent, the public sector. However most comment on pensions focusses on rather different questions. In particular it draws attention to the continued reliance on income support of a significant proportion of pensioners and argues that the compulsory element of the second-tier pension is inadequate.

  112. Suggested changes to the basic state pension include restoring the basic state pension to 20 per cent of average male earnings and uprating it thereafter in line with average earnings. Such a move would be costly and would bring little benefit to the poorest pensioners, who are already entitled to income support. An alternative proposal is to concentrate the benefit of the basic pension on the least well-off by increasing the benefit earned by lower-paid contributors and reducing the benefit earned by higher-paid contributors. Schemes of this type would only achieve a significant effect after many years. They have been criticised on the grounds that they would be highly complex, undermine the contributory principle and could have unfavourable effects on incentives.

    113. An Assured Pension

  113. The Retirement Income Inquiry (RII)12 proposed a new minimum pension guarantee, provided to all people over state pension age as of right. It would be specifically designed to provide an adequate level of income for those who do not have sufficient income from other sources. It proposed the name "Assured Pension".

  114. The assessed income would disregard all assets but would take into account other pension income and other income from savings (with a small disregard). The withdrawal of the Assured Pension would be tapered. The RII did not propose a specific rate of taper but argued that some degree was essential. All those receiving an income related pension under the Assured Pension would qualify automatically for means-tested benefits such as housing benefit.

  115. The RII did not make a specific proposal about the level of the Assured Pension which would be critical in determining both the cost and the fundamental nature of the overall scheme. It illustrated the costs of schemes which, at age 65, would provide a weekly pension for a single person of:

  116. In its least costly form the Assured Pension (with no taper) is much the same as the present system except for increasing the take-up of income support and allowing extra disregards. If (as the RII proposes) there are changes to the second tier system which improve private provision of pensions, the introduction of an Assured Pension should not involve any increase in costs beyond those in Table 13.

    117. Table 13: Additional costs of an Assured Pension

    Amount per week single (couple in brackets)
    £65
    (£104)    		 £73.50
    (£117.60)    		 £80
    (£128)    		 £120
    (£192)
    Withdrawal rate
    (£bn)
    100% (no taper) 1.0 2.6 4.6 17.2
    75% 1.2 3.0 5.4 19.7
    50% 1.8 4.0 6.8 22.9

  117. The RII proposed that the Assured Pension should be financed from general taxation, apart from the basic pension which would continue to be met from the National Insurance Fund. If the basic pension continues to be indexed to prices rather than earnings it will provide a decreasing proportion of the Assured Pension (and of pensioners' incomes in general).

  118. The RII argues that the Assured Pension is a more effective way of using money to provide an adequate minimum retirement income than proposals to increase the value of all pensions. This is part of the general debate about means-testing versus universal provision.

    119. Second-tier provision

  119. Compulsory second-tier provision for employees is currently either provided through the SERPS or requires, in effect, contributions of 4.8 per cent of earnings between the LEL and UEL to be paid into an approved occupational or personal pension scheme.

  120. The argument for compulsory second-tier provision may rest partly on the view that individuals are not the best judges of their own savings decisions. A more compelling argument may be that it reduces the need for individuals to rely on (non-contributory) state support. Frank Field MP in particular has argued for the extension of compulsory earnings-related schemes - see Field (1995); Field and Owen (1994). He is strongly opposed to means-tested benefits which, he argues, "penalise all those values which make strong, vibrant communities" (Field (1995) p10). He proposes a two-part social security system. Public welfare provision would be administered by an independent stakeholders' insurance corporation which would, in due course, set contribution and benefit rates subject to a veto by the Chancellor of the Exchequer. Contributions would be earnings-related for all employees including part-time workers and others below the LEL. The Government would make contributions for the lowest-paid workers, but there would be a clear distinction between the re-distributive function of the scheme (based on taxation) and its savings function (based on graduated contributions).

  121. There would also be a compulsory second-tier system operating through a stakeholders' private pension corporation. All employees and self-employed persons would be compelled to become members of a private pension scheme. The board of the corporation would set the rate of contributions for employee and employer. Contributions would start at a modest level and would be progressively raised.

  122. Many of Mr Field's proposals were incorporated in the Report of the Commission on Wealth Creation and Social Cohesion in a Free Society (Dahrendorf et al (1995)).

  123. The RII concluded that a compulsory second-tier is needed in order to avoid excessive costs to the taxpayer of providing an adequate minimum retirement income. It proposed that the minimum level of contributions to the second tier should be sufficient and its coverage wide enough to ensure that in future almost all the retired population will have adequate incomes and will not require additional state support to achieve the minimum level of income guaranteed by the Assured Pension. If the Assured Pension is related to average earnings and if the basic state pension remains fixed in real terms, the level of contributions to the second tier will need to rise through time.

  124. Beyond that the RII argued that the level of contributions to pensions should be a matter of personal choice. It proposes that SERPS should be phased out and replaced by a funded National Pension Scheme. It would be a Defined Contribution scheme, with funds invested in a well-balanced portfolio of investments. The scheme would be compulsory for those not in a scheme meeting minimum requirements, but it would be open for anyone to join and members could contribute above the minimum required level. Tax relief would be given to contributions and to income earned within the scheme, as is currently the case with occupational and personal pensions. On retirement the proceeds will be used to purchase annuities. The Committee argues that the National Pension Scheme would be particularly suitable for low-paid and mobile workers and those in small firms. They also believe it would keep administrative costs low. They assume that the National Pension Scheme would not be identified as part of public expenditure.

  125. Employers and employees would be required to provide between them at least a minimum contribution to an approved scheme or to the National Pension Scheme. The employee would exercise the choice. They propose that the minimum combined contribution should be set at the current contracted out rebate of 4.8 per cent. Compulsory contributions would be subject to an upper earnings limit, set at about the same level as the current UEL. There would also be a lower limit, which could be fixed at about the current LEL.

  126. They propose that rights already accrued under SERPS would be preserved and the eventual SERPS pensions financed on a pay-as-you-go basis from the NI Fund. They also recommend that anyone within 15 years of pension age should be allowed to remain in SERPS if they wish. Their compulsory contributions would be paid into the National Insurance Fund. Contracted out rebates would cease. The new NIC rate would be set at 0.5-0.7 per cent higher than the current contracted-out rate; 4.3-4.1 per cent lower than the contracted-in rate.

  127. If the minimum contribution remains at 4.8 per cent, the effect on contributions for employees and employers would be:

  128. Compulsory contributions to the second tier should be increased through time to constrain the public sector cost of paying the Assured Pension. This could presumably be done through increases in the compulsory contribution rate. They recommend that pensions should be required to provide at least limited price indexation and should include survivors' benefits. They also recommend that men and women should get equal pension rates for equal contributions within the compulsory element.

    129. PRIVATISATION AND THE UK EXPERIENCE

  129. In this final section we consider briefly, in the context of this collection of papers, the extent to which privatisation has played a part in the development of the UK pensions system.

  130. It is clear, from the other studies, that the key elements in the reforms elsewhere are privatisation and funding, with the funds being invested in productive capital. Although the UK played a leading part in the privatisation of activities formerly run by the public sector, it has not generally counted its changes to the pension system as part of this process. Richard Disney has commented that the first major change in the UK pensions system in the post-War period went in the opposite direction. The introduction of SERPS, and its predecessors, was an attempt to supplement the private sector's occupational pension schemes by extending second-tier pensions to a wider range of employees. It did not involve nationalisation of the existing private schemes but at the time of its introduction it provided a generous alternative, particularly for those close to retirement. In addition, it provided a potential supplement to all private sector schemes by providing post-retirement indexation at a time when most private schemes did not do so.

  131. One can say that SERPS added an unfunded earnings-related public scheme to the funded private sector pension schemes.

  132. The reforms since 1985 certainly qualify as a move to privatisation and funding. Those who have left SERPS have moved from public sector schemes to private sector schemes. They have also moved from unfunded schemes to funded schemes which are largely invested in equities. About one-half of employees who were members of SERPS in 1985 have left the scheme. In addition, new entrants to the labour force may join occupational pension schemes, if they are available, or can start personal pension schemes, using the contracted-out rebate plus possible supplements from employers or employees. These schemes are funded.

  133. It remains true that all employees and the self-employed (with earnings above a lower limit) must belong to the basic state contributory pension scheme, which is pay-as-you-go. However, under current policy, the value of the basic pension is fixed in real terms. The public sector component of pension incomes is therefore withering away relative to private, funded provision. There is clearly an issue of whether this system too should be privatised and funded. That would in turn raise questions about whether it should become earnings-related or whether there should continue to be at least a flat-rate element.

  134. There is also the question of whether SERPS should be privatised and funded. That would be a rather simpler matter than privatising and funding the basic pension scheme. That was, in effect, the original proposal in 1985 but at that time the private sector was reluctant to accept responsibility for providing earnings-related pensions for the low-paid and workers who moved often into and out of employment.

  135. Like other moves towards funded schemes, the UK's reforms have had a transitional cost. Disney estimates that the switch to personal pensions may currently be reducing National Insurance Fund revenue by about £3 billion a year (about 0.5 per cent of GDP), and that offsetting this would need an increase in contribution rates of up to 2 percentage points. It is clear that current pensioners are not bearing this cost, nor are those who have contracted out of SERPS (except insofar as they share the general burden through higher NI contributions or taxes). The £3 billion is being met by NI contributors and/or taxpayers.

  136. In general one can say that the modest public sector cost of pension provision in the UK does reflect a policy of privatisation. On current policies, funded private sector schemes will provide an increasing proportion of retirement income. In the UK, by contrast with some other cases, this private provision - assisted by favourable tax treatment - will be largely voluntary rather than compulsory. Although employees are compelled to make pension provision (either in the state scheme or by investing in private pensions), there is less compulsion than there was before the reforms, and significantly less than in many other industrial and developing countries.

References

Blake D (1994) Pension choices and pensions policy in the United Kingdom Paper presented at a conference, "Pensions: Funding, Privatisation and Macroeconomic Policy", held in Chile, January 1994

Blake D (1995) Pension Schemes and Pension Funds in the United Kingdom Clarendon Press, Oxford

Blake D (forthcoming) The effect of different wealth components on personal savings in the United Kingdom Birkbeck College Discussion Paper

Chand S K and Jaeger A (1996) Aging Populations and Public Pension Schemes International Monetary Fund Occasional Paper 147. IMF, Washington DC

Dahrendorf R, Field F, Hayman C, Hutton W, Hutcheson I, Marquand D, Sentance A and Wrigglesworth I (1995) Report on Wealth Creation and Social Cohesion London

Department of Social Security (1995) "The Pensioners' Income Series 1993" London

Dilnot A, Disney R, Johnson P and Whitehouse E (1994) Pensions Policy in the UK. An Economic Analysis Institute for Fiscal Studies, London

Disney R (1995) "Occupational Pension Schemes: Prospects and Reforms in the UK" Fiscal Studies Vol 16 No 3 pp 19-39

Disney R and Stears G (1996) "Why is there a decline in defined benefit pension plan membership in Britain?" Institute for Fiscal Studies Working paper W96/4

Disney R and Whitehouse E (1993a) "Will the younger cohorts obtain a worse return on their pension contributions?"in J Creedy and M Casson (eds), Industrial Concentration and Economic Inequality: Essays in Honour of P.E.Hart Aldershot: Edward Elgar

Disney R and Whitehouse E (1993b) "Contracting Out and Lifetime Redistribution in the UK State Pension System" Oxford Bulletin of Economics and Statistics Vol 55 No 1 pp 25-42

Feldstein M (1974) "Social security, induced retirement, and aggregate capital accumulation", Journal of Political Economy Vol 82 pp 905-926

Feldstein M (1996) The Missing Piece in Policy Analysis: Social Security Reform National Bureau of Economic Research Working Paper 5413

Field F (1995) Making Welfare Work Institute of Community Studies, London

Field F and Owen M (1994) National Pensions Savings Plan Fabian Society, London

Hubbard R G, Skinner J and Zeldes S (1995) "Precautionary Saving and Social Insurance" Journal of Political Economy Vol 103, No 2, pp 360-399

Johnson P, Disney R and Stears G (1996) Pensions: 2000 and Beyond. Analysis of Trends and Options Retirement Income Inquiry, Volume 2. London

Johnson P and Stears G (1996) "Should the basic state pension be a contributory benefit?" Fiscal Studies Vol 17 No 1 pp 105-112

National Audit Office (1990) The Elderly: Information Requirements for Supporting the Elderly and Implications of Personal Pensions for the National Insurance Fund. HMSO London

Retirement Income Inquiry (1996) Pensions 2000 and Beyond Volume 1, Report. London

Samuelson P (1958) "An exact consumption loan model of interest with or without the social contrivance of money" Journal of Political Economy Vol 66 pp 467-482

Secretary of State for Social Services (1985) Reform of Social Security Vols 1 and 2. London Cmnd 9518

Secretary of State for Social Services (1985) Reform of Social Security. Programme for Action London Cmnd 9601

Shigehara K (1995) Commentary on "Long-term tendencies in budget deficits and debt" by P Masson and M Mussa Budget deficits and debt: issues and options Proceedings of a symposium sponsored by the Federal Reserve Bank of Kansas City pp 57-87

FOOTNOTES

1. We employ UK usage in this paper. American readers may find the following glossary helpful: 

            UK                    US

            flat rate             lump sum
            occupational          private
            pensioner             retiree
            pensions              social security
            personal pensions     individual retirement accounts, 401(k) plans
            preserved benefits    deferred benefits
            social security       welfare plus social security
            state                 federal

2. Welfare benefits available to UK pensioners also include:

3. Much of the material in this paper is based on Johnson, Disney and Stears (1996). Their comprehensive report on the UK pensions system formed Volume 2 of the Report of the Retirement Income Inquiry. Descriptions of the system and discussions of the main policy issues can be found in Blake (1995) and Dilnot et al (1994).

4. Although the basic pension has declined relative to average earnings it has stayed broadly constant at 20 per cent of average manual male earnings throughout its 90 years' existence. In the UK, as in the US, there has been a significant increase in the dispersion of incomes in recent years, with manual earnings falling relative to non-manual earnings.

5. As a guide to the scale of these costs, GDP in 1994-95 was £680 billion.

6. "Reckonable earnings" are earnings between the lower and upper earnings limits.

7. An account of earlier earnings-related state schemes is provided in Blake (1995).

8. Disney and Stears (1996) discuss the moves between DB and DC schemes.

9. The pension fund industry argues that lump sum treatment is not as generous as it seems since fund accruals are given tax relief only at the lower rate of income tax rather than members' marginal tax rate.

10. The material on the distribution of pensioner incomes comes from Department of Social Security (1995).

11. See Hubbard et al (1995).

12. The Retirement Income Inquiry, under the Chairmanship of Sir John Anson, was set up by the National Association of Pension Funds. The Inquiry was completely independent and members served in an individual capacity.



© Crown Copyright