Save our Pensions

- Sunday, 19th January 2003




The Foundation?s main activity is to commission and publish original papers by

independent academic and other experts on key topics in the economic and

social fields, with a view to stimulating public discussion on the performance

of markets and the social framework within which they operate.

The Foundation is a registered charity and a company limited by guarantee.

It is independent of any political party or group and is financed by the sales of

publications and by voluntary donations from individuals, organisations and


The views expressed in publications are those of the authors and do not

represent a corporate opinion of the Foundation.


David Lipsey (Lord Lipsey of Tooting Bec)

Members of the Board

Viscount Chandos

Gavyn Davies

David Edmonds

John McFadden

Baroness Noakes

Brian Pomeroy


Philip Collins

Deputy Director

Beth Breeze

SMF Save Our Pensions Text 16/9/03 12:06 am Page 1

First published by The Social Market Foundation, 2003

The Social Market Foundation

11 Tufton Street

London SW1P 3QB

Copyright ? The Social Market Foundation, 2003

The moral right of the authors has been asserted.

All rights reserved. Without limiting the rights under copyright reserved above,

no part of this publication may be reproduced, stored or introduced into a

retrieval system, or transmitted, in any form or by any means (electronic,

mechanical, photocopying, recording, or otherwise), without the prior written

permission of both the copyright owner and the publisher of this book.

ISBN 1-874097-53-4

Cover design and typesetting by Beaufort 5

SMF Save Our Pensions Text 16/9/03 12:06 am Page 2


Peter Lilley


September 2003

SMF Save Our Pensions Text 16/9/03 12:06 am Page 3


I am very grateful to all those who have so kindly commented on successive

drafts of this paper particularly Professor David Blake of The Pensions Institute,

Carl Emmerson of the Institute of Fiscal Studies, Mary Francis of the ABI,

Lorraine Fraser of Winterthur Life UK, David Harris of Watson Wyatt, John Jory

of the Building & Civil Engineering Benefit Schemes, Tim Keogh of William

Mercer Ltd, Andrew Mitchell MP, Ann Robinson of ASSP, Mark Rowlands of

Winterthur Life UK, Andrew Tyrie MP, Mike Wadsworth of Watson Wyatt, David

Willetts MP and Andrew Williams. My thanks are also due to the DWP and

Government Actuary?s Department for providing factual and quantitative


None of them are responsible for its errors or omissions nor necessarily

endorses its analysis and conclusions.

SMF Save Our Pensions Text 16/9/03 12:06 am Page 4


Glossary of Terms 6

Executive Summary 7

Introduction 14

Compulsion: How Much Pension Provision? 16

Funding: Save Now or Tax Later? 24

Pension Age: How Long to Work? 37

How to Cut Pension Costs 44

Conclusion 49

Annex A How the UK State Pension System Works 51

Annex B Basic Pension Plus 55

Annex C Australia?s Experience of Compulsory Personal Pensions 56

SMF Save Our Pensions Text 16/9/03 12:06 am Page 5



(See Annex A for a guide to the complexities of the British pension system)

ABI Association of British Insurers

APP Approved Personal Pension

ASP Additional State Pension (Pension additional to BSP, was SERPS now S2P)

BPP Basic Pension Plus

BSP Basic State Pension

DWP Department for Work and Pensions (formerly DSS)

DSS Department of Social Security

GAD Government Actuary?s Department

HET Higher Earnings Threshold (S2P accrues at 10% of earnings between LET and

HET and at 20% between HET and UEL)

LEL Lower Earnings Limit (Level of earnings - set roughly equal to BSP -below which

employees neither pay NI nor accrue any entitlement to BSP and S2P)

LET Lower Earnings Threshold (Set roughly equal to half national average earnings.

Employees earning between LEL and LET accrue a flat rate element of S2P as if

they earned LET)

MFSP Mandatory Funded Second Pension (The pension generated by NI rebates

invested in a fund to replace the unfunded S2P)

MIG Minimum Income Guarantee (Level to which retirement incomes will be topped

up by Guarantee Credit)

NI National Insurance

NIC National Insurance Contribution

NI Rebate Element of employee?s and employer?s NICs paid into private pension fund of

those who contract out of ASP

ONS Office of National Statistics

Pay-As-You-Go Scheme in which current taxes or NI contributions are used to pay pensions of

people already retired rather than invested to pay future pension liabilities.

PC Pension Credit (composed of two elements: Guarantee Credit which tops up

retirement incomes to the MIG level and Savings Credit which partly

compensates for impact of means testing on income from private saving)

SERPS State Earnings Related Pension Scheme

SPA State Pension Age

S2P State Second Pension

UEL Upper Earnings Limit (Level beyond which higher earnings accrue no extra S2P)

SMF Save Our Pensions Text 16/9/03 12:06 am Page 6


Reform of Britain?s pension system must address four key questions.

They are:

Compulsion how much should people be required to provide for

their old age?

Funding how much of that compulsory provision should be funded by

genuine savings and how much left to future tax payers (by the

state financing on a pay-as-you-go basis)?

Pension age should people be required or encouraged to work longer before

drawing their pension?

Costs how can the portion of savings absorbed by the costs of

running pensions and annuities be reduced?

Compulsion: How Much Pension Provision?

? Compulsion is undesirable. But it is unavoidable. Any civilised society is

bound to provide at least a means tested safety net for those who reach old

age without providing for their retirement. That creates disincentives to

save which cannot be eliminated, only spread more thinly further up the

income scale. It also means that the prudent, who voluntarily provide for

their own retirement, are compelled through the tax system to pay for

those who improvidently fail to do so. It is better to require everyone who

can do so during their working lives to make provision for their own

retirement. That will not be an extra burden on those who would save

anyway. Indeed, they will then not be compelled to pay so much tax to

support the improvident in the future.

? We already have a compulsory second pension. All employees are required

to pay via their National Insurance Contributions either into the State

Second Pension (previously SERPS) or into an approved occupational or

personal pension scheme offering similar or superior benefits. The key issue

is whether the level of this compulsory provision is sufficient.

? The level of compulsory provision should be sufficient to lift people clear of

the means tested safety net. Means testing discourages people on low

incomes from saving and it provokes resentment among those who


SMF Save Our Pensions Text 16/9/03 12:06 am Page 7

nonetheless do save yet are little or no better off as a result. In Australia

making everyone in work provide adequately for their retirement has been

widely welcomed especially by younger people.

? Coverage of the compulsory second pension should be extended to the

self- employed.

? The government is committed to raising the Minimum Income Guarantee in

line with earnings during this parliament and projects it forward indefinitely

on that basis. Yet state pensions are set to rise only in line with prices. If this

continues the MIG is bound to overtake the level of state and compulsory

pensions. To prevent this happening, the Basic State Pension and the MIG

should be statutorily indexed to prices and, when resources permit, any real

increases made by raising both the BSP and the MIG by the same cash amount.

? Once an adequate Second Pension is in place it will be possible to phase out

the Savings Credit system whose complexities and withdrawal rates act as

disincentives to save over a broad swathe of incomes.

Funding: Save Now or Tax Later?

? The UK has more investments to meet future pension requirements than

the rest of the EU put together. Nonetheless these still only cover 40% of

our future pension liabilities. The government has set a laudable target of

raising that proportion to 60% by the middle of the century.

? This will only be achievable if more people contract out of the unfunded

State Second Pension scheme into genuinely funded pensions.

? Sadly Stakeholder Pensions have had negligible success in persuading more

people to contract out of the State Second Pension.

? Those who contract out receive a rebate from their National Insurance

Contributions payable directly into their personal or occupational pension

fund. The level of the rebate is set on the basis of the Government Actuary?s

estimates of the amount needed to fund a private pension broadly

sufficient to replace the pension entitlement that employees would have

accrued in the State system.


SMF Save Our Pensions Text 16/9/03 12:06 am Page 8

? But contracting out of the unfunded Additional State Pension system into

a personal pension has never been attractive for those on low earnings. This

is because the lower a person?s earnings the smaller their rebate. Charges

do not decline proportionately since there are fixed costs in setting up and

running a fund. So charges absorb a disproportionate share of the small

rebates received by those on low earnings.

? Consequently the option of remaining in the unfunded State Additional

Pension system has been retained for the benefit of those on low and

intermittent earnings.

? That problem was inevitable under SERPS since the accrual of pension and

the corresponding rebate for contracting out were directly proportional to

earnings. However, under S2P, employees with earnings below roughly half

the national average accrue a flat rate entitlement to pension. Yet if they

contract out they still receive a rebate proportionate to their earnings so it

remains unattractive for them to contract out.

? Those on low earnings should be entitled to a flat rate rebate equivalent

to the full value of the pension entitlement they accrue at present under

S2P. That would be sufficient to justify the fixed costs of setting up a

personal pension. It would then be possible gradually to replace the

unfunded pay-as-you-go S2P by a Mandatory Funded Second Pension of

equivalent value. Those on earnings above the Lower Earnings Threshold

would receive a rebate comprising both the flat rate element for earnings

up to the LET and an earnings related element for earnings above the LET

as at present.

? This proposal would come in progressively with each cohort of young

people entering employment. I envisage that everyone in work who was

born, say, thirty years or less before the new system starts would be

required to have a Funded Second Pension. They would receive a NI rebate

equivalent to the value of the old S2P accrual which it would replace.

That payable into their own personal or occupational fund. Those born

more than 30 years before the start date would retain the choice, as at

present, of remaining in the unfunded S2P or opting into a funded

private pension.


SMF Save Our Pensions Text 16/9/03 12:06 am Page 9

? The government originally envisaged compelling some employees to have a

private Funded Second Pension to replace the unfunded S2P. But only

those with earnings above the Lower Earnings Threshold would have been

required to contract out in this way. It now envisages that all those with

earnings below the LET will lose the right to contract out of the S2P. Those

above this level will be encouraged to contract out. If they do they will

receive a rebate equivalent to the flat rate and earnings related elements of

S2P but if not they will only accrue the flat rate of S2P. This would create

an unfair two class system and difficulties for those whose incomes

fluctuate above and below this level.

? My proposal would enable the government to meet its target of 60%

privately funded pensions by mid-century; enormously boost committed

long-term savings and investment; and dramatically reduce the long-term

tax burden. More important, over time it would ensure that everyone had

their own pension fund - the biggest extension of capital ownership since

the growth of home ownership.

? Once people own a personal pension fund they can readily choose to save

more than the minimum ? and Australian experience suggests many will do

so. Ownership of a pension fund will also give everyone a direct stake in

national prosperity.

? The government should guarantee that if for any reason the element of

anyone?s Funded Second Pension derived from the flat rate rebates is not

sufficient at retirement age to buy an annuity equivalent to the flat rate

element of S2P the state will top up the annuity payments to that level.

Earnings related rebates and voluntary additional contributions will not be

covered by this guarantee.

? To prevent fund managers betting against the guarantee if the fund has

underperformed they would be required to invest on a Prudent Person

basis. These rules could be enshrined in statute as in Australia. Also fund

managers would be required to invest the guaranteed and non-guaranteed

elements of the funds on the same basis.


SMF Save Our Pensions Text 16/9/03 12:06 am Page 10

Pension Age: How Long to Work?

? Unless it moves to a Mandatory Funded Second Pension, the government

could only reach its target of 60% funded pensions by 2050 by raising the

State Pension Age to make people work longer.

? Given that the life expectancy of those aged 65 has increased by roughly one

month every year for several decades there are strong pressures to raise the

SPA annually by that amount simply to contain the cost of state pensions.

? As far as possible the government should avoid telling people when to

retire or forcing them to work longer by raising the SPA. The most it should

do is to raise, in line with average life expectancy, the age at which people

can draw their tax free lump sum other than to buy an annuity.

? The great advantage of everyone having their own pension pot is that with

ownership comes choice. People will be free to choose when to retire once

their fund can provide an income sufficient to keep them from dependence

on means tested benefits. At the same time ownership of a personal

pension fund will give them a double incentive to work beyond the normal

pension age ? to save more for fewer years of retirement.

? Compelling people to work longer would bear hardest on those who work

in stressful, manual and lower paid jobs. These are the very people whom

the present pension system treats most unfairly. They pay in on the same

terms as everyone else. Yet they have shorter life expectancy and so

typically draw their pensions for fewer years. The less well paid subsidise

the retirement of the higher paid and longer lived.

? Government should encourage providers to remedy this by relating

pensions and annuities to the life expectancy of different income groups.

This would mean that the lower people?s earnings during their working life

the less they would pay to buy a given annuity income. So the less well paid

would not need to accumulate so large a pension fund before they can

choose whether to retire or continue working (possibly part-time).

? To enable annuity providers to do this the government could make

available, in a form equivalent to their tax code, a summary of each


SMF Save Our Pensions Text 16/9/03 12:06 am Page 11

person?s life time earnings from the National Insurance Recording System.

The likely growth of affinity group providers once everyone has their own

pension pot will also lead to provision of annuities more fairly reflecting

the life expectancy of lower income groups.

How to Cut Pension Costs

? The costs of running pension funds typically reduce the value of the final

pension pot by over a third. Over half of this is the cost of acquiring


? Requiring everyone to have a pension fund would be the single most

effective way of reducing those costs. Compulsion would eliminate costs of

persuasion, largely remove the need for advice, simplify and standardise the

product, and spread overheads more widely. So costs might be reduced by

as much as a half, increasing the final pension by as much as 13%.

? Insurance is a good way to pool risks which are stable across a population.

But there is no way to hedge the risk that average life expectancy might

suddenly outstrip past trends ? e.g. because of amagic bullet cure for cancer.

? Annuity providers have to make provisions against possible but unlikely

developments in longevity. If those developments fail to occur and

mortality is in line with central estimates pensioners will collectively have

paid for more years of retirement than they receive. On the other hand if

annuity providers fail to make adequate provision for future increases in

longevity they may prove unable to pay pensions they have promised. The

state would have to step in.

? Most of the uncertainty annuity providers must provide against concerns

the number of people who may live well beyond current average life

expectancy. So if the Mandatory Funded Second Pension was only required

to cover the first twenty years of retirement and the State provided an

unfunded S2P for the years beyond age 85 the state would be bearing

much of the unquantifiable longevity risk.

? The cost of twenty year fixed term annuities should be disproportionately

cheaper than full life annuities. Anyone wanting a retirement income


SMF Save Our Pensions Text 16/9/03 12:06 am Page 12

above the Second Pension level would still have to buy an additional full

life annuity


SMF Save Our Pensions Text 16/9/03 12:06 am Page 13


The biggest domestic problem facing every developed country is how to

finance pensions as people live longer and have fewer children.

Britain is better placed to cope with this than most of our European

partners. They rely almost entirely on taxation to pay for pensions. So the

growing burden of their aging populations will mean rising tax burdens on

declining work forces. By contrast, we have encouraged people to contract out

of the unfunded Additional State Pension and save in occupational and

personal pension schemes. As a result we have built up more investments to

meet future pension liabilities than all the rest of the EU put together.

Nonetheless, UK pension funds only account for 40% of future pension

liabilities ? the other 60% falling on future tax payers. The Labour government

set a target to reverse those proportions so that 60% of pensions would come

from savings by the middle of this century. That is an admirable target that a

future Conservative government would, I hope, endorse.

Unfortunately, having set itself that target, the government has actually

moved in the opposite direction. It has increased unfunded state pension

promises enormously while imposing a ?5 billion per annum tax on pension

funds and creating disincentives to save by its Minimum Income Guarantee

and Pension Credits.

At the same time private pensions have been plunged into a crisis. Defined

benefit schemes are closing to new employees at an accelerating rate.

Employers are changing from defined benefit to money purchase and often

contracting their employees en bloc back into the state system. Insurers are

urging many employees to opt back into the unfunded State Second Pension.

The stock market fall has provoked a crisis at Equitable Life, caused other withprofits

pension providers to cut their bonuses and hit even harder those

invested in unsmoothed funds.

In the face of all this some commentators advocate retreat: abandon the

60:40 savings target, leave it to tomorrow?s taxpayers to pay for our pensions

and/or force people to work longer before they receive their state pension.

To retreat from funded pensions would be folly. Britain?s funded pension

provision is an enormous asset. We should see current problems as

opportunities to strengthen it and build on it. But that needs new, imaginative

and radical thinking. Sadly the government?s Green Paper, though it contained


SMF Save Our Pensions Text 16/9/03 12:06 am Page 14

some useful minor reforms, did not remotely match up to the scale of the

pension crisis. It largely sidestepped the four key issues facing our pension

system which are:

compulsion ? how much people should be required to provide for their

old age;

funding ? the extent to which compulsory provision should be funded by

genuine savings or left to future tax payers;

pension age ? whether people should be encouraged or required to work

longer before drawing their pension;

costs ? how the cost of pensions and annuities can be reduced.

These are the issues addressed in the four sections of this report.


SMF Save Our Pensions Text 16/9/03 12:06 am Page 15



Should the state compel people to make adequate provision for their

retirement? And if so, how much?

In this chapter I do not address the issue of the extent to which any such

compulsory provision should be financed by the state on a pay-as-you-go basis

and how much by private savings. That is the question considered in the next


The importance of the issue of compulsion has been accentuated by the

erosion of incentives to save (through the extension of means testing to the

majority of future pensioners) and the limiting of any reward for persuading

people to save extra (through the cap on charges for stakeholder pensions). Yet

the government?s Green Paper largely sidestepped it ? instead setting up a

Commission to monitor whether voluntary provision is sufficient.

Concerns about the adequacy of people?s provision for retirement have been

growing. The Association of British Insurers claims there is a "pension shortfall

of ?27 billion".1 As the trade organisation for the providers of personal pensions,

they have a vested interest in talking up the need for more pension provision.

Nonetheless, since that figure was calculated the stock market has fallen

much further, more pension funds have closed and Equitable Life?s problems

have cast their shadow over other with-profits providers.

As a result, calls for the state to compel people to make greater provision

for their retirement have multiplied. Often such calls take the form of demands

that the state require people (or their employers) to provide for a "second


A surprising number of commentators seem unaware that all employees are

already compelled to provide for a pension in addition to the basic state


Ever since 1978 all employees have been required to make payments via

their national insurance contributions either into the Additional State Pension

(called the State Earnings Related Pension System until April 2002 when it was

restructured and renamed the State Second Pension) ? or into an approved

occupational or personal pension scheme offering equivalent or superior



SMF Save Our Pensions Text 16/9/03 12:06 am Page 16


People may be unaware of this because they never see the contributions

they are making - since they are deducted at source from their pay and go

direct to the National Insurance Fund or onwards to their pension fund.

Is the State Second Pension adequate?

The issue is not, therefore, whether a compulsory second pension needs to be

introduced but whether the new State Second Pension or the rebates for

opting into a private scheme will be adequate.

At first sight the amount of provision people should make for their

retirement is a matter for them. Why should the State compel anyone to make

a minimum level of provision for themselves? I am on the libertarian end of the

political spectrum. So I started from a presumption that compulsion should be

avoided if at all possible. But I concluded that if, as any civilised society must,

we make provision for those who cannot provide for their own retirement, an

element of compulsion is inevitable. Either we compel all those who can do so

to provide sufficiently for themselves; or the prudent, who voluntarily provide

for themselves, will be compelled also to support via the tax system the

imprudent who fail to do so.

Virtually everyone would accept that, at very least, we must help those who

simply could not afford to save for retirement whether through sickness,

disability, caring responsibilities, unemployment or very low income when they

were of working age. Not everyone would be so willing to help those who

could have made provision for retirement but failed to do so. In practice it

would be difficult to distinguish between those who would have saved but

couldn?t and those who could have saved but didn?t. In any case, however

much we may disapprove of the improvidence of the latter, they cannot be left


So, at the very least, any decent society must have some means tested

safety net for those who reach old age with insufficient provision for their

retirement. However, once the state effectively guarantees a means tested

minimum income in old age it inevitably creates two problems. It discourages

people on low incomes from saving and it arouses valid resentment among

those who do nonetheless save yet are little or no better off as a result.

The only logical way to remove both disincentive and resentment is to

require people of working age who can afford to do so, to provide for a pension

at least equal to the minimum retirement income guaranteed by the state. Those

who might otherwise have been improvident would then be forced to make

SMF Save Our Pensions Text 16/9/03 12:06 am Page 17

provision during their working lives. So they would no longer arouse resentment

among the prudent nor burden the taxpayer. Those who would have made

adequate provision for themselves anyway will not be directly affected by

compulsion. Indeed, they will not be compelled to pay so much in taxation to

support the imprudent in future. Those who cannot make such provision in any

year (e.g. because of caring responsibilities) would be given credits towards such

a level of retirement income. Ideally those credits would score as accruals for

pay-as-you-go funds and be paid in cash for funded pensions.

This was roughly the structure envisaged by Beveridge when he designed

the welfare state. People were initially required to contribute during their

working lives towards a basic state pension that was then set somewhat higher

than the minimum level of income guaranteed by National Assistance. Over

time, however, the minimum means tested income for the elderly (provided

first by National Assistance and latterly by Income Support) has been increased

to a level above the Basic State Pension.

So long as the gap between the two was small comparatively few people

suffered means testing of their savings. This government has greatly

exacerbated the problem by setting the Minimum Income Guarantee nearly

one-third higher than the Basic State Pension. This has substantially increased

the number of people facing disincentives to save and feeling resentment if

they do. The Minimum Income Guarantee in 2003/4 is ?102.10 pw against the

Basic State Pension of ?77.45 pw for a single person.

So someone with just the BSP of ?77.45 will be entitled to claim Guarantee

Credit of ?24.65 to bring their total income up to the MIG of ?102.10. A

pensioner with a small private pension of, say, ?10 pw on top of their BSP will

currently get ?10 less Guarantee Credit ? just ?14.65 ? to top them up to the

MIG level.

The MIG on its own thus renders voluntary saving towards a modest second

pension of up to ?24.65 pw completely pointless. Even a private pension

somewhat higher than this represents a very poor net return for saving. For

example, a pension of ?34.65 on top of the BSP giving a total income of

?112.10 leaves the pensioner only ?10 pw better off than if they had not saved

a penny. Most occupational pensions are quite modest. About half are less than

?40 pw. They will barely lift pensioners above the level they could have got

relying solely on the MIG.

Moreover, the government has promised that the Minimum Income

Guarantee will rise in line with earnings. Yet the Basic State Pension is still only


SMF Save Our Pensions Text 16/9/03 12:06 am Page 18

indexed to prices. So the gap between the two is set to widen making the

problems of disincentives and resentment progressively worse. Even if someone

has a pension sufficient to lift them clear of the means tested zone at the

point of retirement they may well find themselves subject to means testing in

the course of their retirement as the MIG increases in line with earnings and

overtakes their pension which is indexed to prices.

The government hopes to mitigate the disincentive effect of means testing

by introducing the Pensions Credit. This comes in from October 2003.

The effect of this is to restore some, but not all, of the reward for saving to

those who previously would have done just as well to rely on the MIG.

Unfortunately this comes at a cost. Those whose income takes them above the

MIG level lose Pension Credit as it is phased out the higher a pensioner?s

income. This introduces a new disincentive to save for a far wider range of

pensioners. And the very complexity of the new system is so bewildering that it

is bound to undermine the savings culture. The summary of the state pension

system in Annex A will give the reader a taste of this complexity.

The Pension Credit works roughly as follows. It will have two components.

The first is the Guarantee Credit which is simply the new name for the benefit

needed to bring a pensioner?s income up to the MIG level. The second component

is called the Savings Credit. Those pensioners whose private pension is less then

the gap between the BSP and the MIG will get a Savings Credit of 60p for every

?1 of their private pension. They will continue to forego ?1 of Guarantee Credit

for every ?1 of pension income. So they face a net loss of benefit of ?only? 40p in

the ?1 for their saving income. In other words they still suffer the same penalty

on marginal savings as someone paying the top rate of income tax.

Take a worked example: George has a Basic State Pension of ?77.45 plus a

private pension of ?10 pw. He therefore foregoes ?10 of Guarantee Credit and

is entitled to the ?14.65 pw to make his income up to the MIG level of ?102.10.

But from October 2003 he also gets ?6 of ?Savings Credit?. So his take home

income will be ?108.10 pw. His ?10 pension will make him only ?6 a week

better off than someone who has not saved a penny.

If George had a pension of ?24.65 pw - equal to the difference between the

BSP and the MIG - he would not be entitled to any Guarantee Credit but could

claim the maximum Savings Credit. That would be worth ?14.79 which is 60%

of his additional pension.

To avoid giving that much Saving Credit to everyone with a pension higher

than this the Credit is tapered out over higher incomes. For each ?1 of income


SMF Save Our Pensions Text 16/9/03 12:06 am Page 19

above that level the amount of Savings Credit to which a pensioner is entitled

is reduced by 40p.

Suppose George had a pension of ?34.65 pw - ?10 more than the gap

between the BSP and the MIG. His entitlement to Savings Credit would then be

?4 less than the maximum. So he would receive a Credit income of just ?10.79.

Entitlement to Savings Credit disappears entirely for those with additional

pensions in excess of ?61.62 pw. But anyone with a pension up to that level

finds that every ?1 of pension income that they have saved for has only made

them 60p better off than if they had saved nothing.

The disincentive is even worse for those with incomes between ?127 and

?140 per week in 2003/04. They face a loss of both Saving Credit and 10 per

cent tax amounting to a combined withdrawal rate of 46p in the ?2.

As a result of Savings Credit some 3.8 million pensioners will face a penalty

on their pension or other income of 40% (plus any tax) in 2003/4. Some 1.3

million of them will have incomes above the MIG level. The other 2.5 million

will have pre-benefit incomes below the MIG.3 The number facing this

disincentive will rise steadily as the MIG is up rated faster than the basic

pension. The government?s own projections show that by the middle of this

century two thirds of pensioners will be eligible for Pension Credit.4

The only way to remove in the longer term the disincentives and

resentments inherent in this system is to set the new State Second Pension,

together with the Basic State Pension, at a level at least sufficient to lift people

above the Minimum Income Guarantee. The need for a system of Saving

Credits would then disappear. The ABI calculate5 that the basic element of S2P

will need to be over a third higher than it is currently set if the combined value

of the Basic and Second State Pensions is to exceed the Minimum Income

Guarantee. That means that the Lower Earnings Threshold, which is currently

set at ?11,200 per annum, will need to be set at ?15,000.

Even this only ensures that people with BSP and the flat rate of S2P at

retirement will have an income in excess of the MIG. As MIG is up rated in line

with earnings it could soon exceed their compulsory pension income plunging

them back into the means tested zone.

Should pensioners? incomes rise in real terms?

The government is committed to raising the Guaranteed Minimum Income

(MIG) in line with average earnings during this parliament. It has indicated its

intention to do so indefinitely by projecting the MIG forward on that basis.


SMF Save Our Pensions Text 16/9/03 12:06 am Page 20

By contrast, the government is only obliged to raise the Basic State Pension

in line with the Retail Price Index (though it has chosen to make discretionary

increases in excess of that in recent years). The value of each person?s State

Second Pension will reflect the average level of earnings during their working

life. So the value of the S2P for each successive year cohort of people reaching

retirement age will tend to be higher than those retiring the previous year.6

But once individuals retire their S2P (like their BSP) will only be up rated in

line with prices. So the Minimum Income Guarantee will steadily overtake the

combined value of the two compulsory pensions for those already retired even

if they exceeded the MIG at the date of retirement. This will reintroduce the

disincentive and resentment effects. The only way to avoid means tested

benefits overtaking contributory benefits is to up rate both by the same


The populist arguments ? that pensioners should not be left behind and

that they have a right to share in rising prosperity - naturally favour up rating

both means tested and contributory pensions in line with earnings. However,

since the early 80s governments of both parties have shied away from any such

commitment on the grounds that:

? it would be hugely expensive and would therefore impose a significant

tax burden on the productive economy,

? each generation is entitled to the standard of living that they have

earned, they do not have a ?right? to the wealth created by the next


? if people wished to have incomes rising in real terms during their

retirement they could tailor their pensions to achieve this but nobody

does so7, and

? even if it were possible to up rate state pay-as-you-go pensions to

reflect rising earnings it would not be feasible to up rate the supposedly

equivalent private pensions of those who have opted out of the S2P.

It is therefore better that both compulsory pensions and state retirement

benefits should be statutorily indexed to prices and any increases in excess of

that should be made on a discretionary basis if and when government finances

permit. In practice the Basic State Pension should be increased by the same

cash amount as the MIG. Thus if the MIG is raised by ?5 more than inflation


SMF Save Our Pensions Text 16/9/03 12:06 am Page 21

then the BSP would also be raised by ?5 and not by the same percentage. That

would avoid the MIG getting out of line with the BSP plus S2P without

needing to raise the S2P in real terms. It would be undesirable to make

discretionary increases in S2P since that would be unfair to those who had

contracted out into private pensions.

Should the coverage of S2P be extended?

At present S2P does not cover those earning below the Lower Earnings Limit

(LEL) and the self-employed.

The LEL is set roughly equal to the Basic State Pension (BSP). It would be

strange to force people earning less than the BSP during their working life to

save in order to have a higher income than the BSP in retirement.

The self-employed have been excluded for two reasons. First, they often see

their business as their provision for retirement and want to use any spare

money they have to invest in it. In practice very few people fund their business

out of current savings. Moreover only a minority of the self-employed do end

up with a saleable business capable of funding their retirement. Second,

reported incomes of self-employed people are a notoriously inadequate guide

to their real disposable incomes. If true that merely means that the level of

contributions they would be required to make would be too low to provide a

commensurate standard of living in retirement. But an inadequate Second

Pension would be better than no provision at all. At present only a minority of

self-employed people do make sufficient provision to avoid reliance on means

tested benefits.

So the time has come to require self-employed people to make provision

for a second pension on the same basis as employees.

Should the second pension be earnings related?

It is easy to justify the state requiring people to contribute towards a pension

sufficient to float them clear of means tested benefits. But why should people

be compelled to provide for additional pension income proportionate to their

earnings? The original Beverage system did not relate pensions (or any other

benefits) to the level of earnings.

This came in with the Graduated Pension, then the State Earnings Related

Pension Scheme and persists in the State Second Pension. Presumably the

rationale is: the higher the standard of living people are used to during their

working lives, the harder they will find it to make do on the Basic State


SMF Save Our Pensions Text 16/9/03 12:06 am Page 22

Pension. Yet most people are short sighted. They do not realise that until it is

too late. So they must be compelled to contribute towards a pension

proportionate to their earnings during their working life.

On the other hand all sorts of people face sharp falls in their standard of

living during the course of their lives. Yet the state simply ensures a basic

minimum level of income regardless of their previous income. It does not

require anyone to insure for an extra cushion related to their previous

earnings. By definition the better off people are the more they can afford to

save. Arguably,8 in a free country it is up to the individual to choose how much

pension they provide beyond the mandatory minimum.

The Labour government has talked of the State Second Pension becoming

flat rate once it has settled in. On closer reading, however, they envisage that

those with earnings greater than the Lower Earnings Threshold (currently

?11,200 pa) will continue to be required to provide for an earnings related

second pension (financed by rebates from their national insurance


My proposals assume that the Mandatory Funded Second Pension includes

an earnings related element based on the State Second Pension. However, it is

not an essential part of the scheme. The key component is the flat rate element

of the State Second Pension. This must be set at a level which, together with

the Basic State Pension, is at least sufficient to avoid reliance on means tested

retirement benefits.

1 "The Future Regulation of UK Savings & Investment", September 2001, Oliver Wyman & Co/ABI.

2 Department of Work and Pensions reply to PQ from Peter Lilley MP 24th February 2003 Hansard.

3 Figures provided to the author by DWP.

4 "Pension Credit: long-term projections" DWP January 2001.

5 "Adequacy, affordability and incentives: a better future for state pensions" ABI March 2003.

6 This is quite distinct from the fact that during the introductory generation successive year

cohorts will have accrued additional years of entitlement.

7 To buy an annuity equal to the Basic State Pension increasing in line with prices would cost a 65

year old man ?70,000. If it was increased by 1.5% per annum more than prices it would cost

about ?80,000. Estimates provided to the author by GAD.

8 Nigel Lawson did argue this when SERPS was reviewed. See "The view from No 11" by Nigel



SMF Save Our Pensions Text 16/9/03 12:06 am Page 23



The previous chapter discussed the minimum pension provision people should

be required to make during their working lives. An equally important question

is how much of that provision should be funded by saving and investment and

how much should be left to be paid by future taxpayers?

The government has recently reaffirmed its commitment that by the

middle of this century 60% of all pension liabilities should be funded? against

40% at present. They stand no hope of reaching that target without decisive

action to shift people from reliance on the state pay-as-you-go pension

system to private fully funded pensions.

When the state provides pensions it invariably funds them on a pay-asyou-

go basis. This year?s taxes are used to pay for this year?s pensions. Nothing

is saved or invested for the future. That is the method adopted by most of our

continental partners to pay for the bulk of their pensions. So as their

populations live longer they face the nightmare of higher taxes falling on a

declining number of people of working age to pay for their pensions.

By contrast, in this country we have encouraged people to save and invest

for pensions beyond the basic minimum. We have enabled employees to

contract out of the Additional State Pension into occupational or personal

pension schemes. A majority of eligible employees have done so. When

employees contract out, they receive rebates from their National Insurance

Contributions which are saved and invested in industry to earn the profits that

will eventually pay their pensions - without any further burden on taxpayers.

Meanwhile this generates an enormous flow of committed long-term

investment. As a result the total amount invested to meet Britain?s future

pension liabilities is greater - not just than that invested by any other EU

country - but than all the rest of the EU put together.

Nonetheless those funds still only represent 40% of Britain?s future pension

liabilities. The other 60% will fall on future taxpayers. To meet the

government?s target of reversing those proportions by the middle of the

century will require a substantial shift in the number of people contracting out

of the Additional State Pension system.

At present employees are given the choice of whether to remain in the


SMF Save Our Pensions Text 16/9/03 12:06 am Page 24

unfunded Additional State Pension or to contract out into a funded private

pension scheme. Those who remain in the state system accrue rights each year

to an additional pension when they reach the State Pension Age. Those who

contract out receive a rebate from their National Insurance Contributions

payable directly into their personal or occupational pension fund. The size of

the rebate is set on the basis of the Government Actuary?s estimates of the

amount needed to fund a private pension broadly sufficient to replace the

pension entitlement that employees would have accrued in the State system.

The most straightforward way to shift the proportion of employees with

funded second pensions would be simply to withdraw, at least for new entrants

to the labour market, the option of an unfunded State Second Pension. All new

workers would then be required to join a Funded Second Pension scheme of

equivalent value to the present S2P, financed by rebates from their National

Insurance Contributions.

Encouraging more people to opt for a funded second pension by increasing

the value of rebates relative to accrued rights would have two serious

disadvantages. First, it would increase the cost to the exchequer. Second, it

would put the government in the dubious position of offering what would be

an inferior alternative bearing its imprimatur. If anyone else did this they

would be guilty of misselling.

Given that the government wants more people to have funded private

pensions it is odd to offer an alternative ? an unfunded State Second Pension

? which it no longer wants them to join.

The reason both this government and its predecessor have nonetheless

continued to offer employees the option of an unfunded Additional State

Pension is that contracting out into a personal pension has never been

attractive for those on low and fluctuating earnings. This is because the lower

an employee?s earnings the smaller the rebate they receive. But the costs of

running a fund do not decline proportionately since there are fixed costs in

setting up and running a fund. So charges tend to absorb a disproportionate

share of the small rebates received by those on low earnings. To avoid this

unfairness to those on low earnings employees have been given the option of

remaining within the unfunded Additional State Pension scheme.

The government sought to overcome the problem of costs absorbing too

great a proportion of the rebates of low earners by introducing Stakeholder

Pensions. Stakeholder charges are limited to a maximum of 1% p.a. of funds

invested. So charges on small funds can no longer be increased beyond that


SMF Save Our Pensions Text 16/9/03 12:06 am Page 25

1% cap to reflect the higher proportionate costs of handling them.

Unfortunately that means that there is little or no incentive for Stakeholder

providers to encourage low earners, who would only have small rebates to

invest, to contract out of the unfunded State scheme. Consequently the

introduction of Stakeholder pensions has done little to encourage more people

to contract out of SERPS/S2P. Instead there is a worrying trend for individuals

to opt back into the State Second Pension. Moreover, companies are

contracting their employees back in wholesale when they switch from direct

benefit to money purchase schemes. This reflects the widespread perception

that the value of the rebate is no longer adequate to fund a pension as

attractive as the unfunded State Second Pension.

The problem of low earners? rebates being below the threshold necessary to

justify the fixed costs of setting up a personal pension scheme was inherent in

SERPS. As its name indicates the amount of pension accrued was directly

proportionate to earnings and so therefore was the rebate for contracting out.

The Conservative government consulted on withdrawing the option of

remaining in the unfunded SERPS. But it concluded that there was no way

round the problem of earnings related rebates and fixed costs. So it decided to

retain the unfunded SERPS option for the benefit of those on low and

fluctuating earnings. Most of those who have not opted out are indeed on low

or fluctuating earnings.

As Secretary of State I reconsidered this issue. I concluded that SERPS could

not be replaced by a Mandatory Funded Pension in a way that was fair to those

on low earnings unless the unfunded Basic State Pension was also being

replaced by a funded equivalent at the same time. Hence the radical plan

called Basic Pension Plus ? (see Annex B). This was based on the recognition

that everyone paying National Insurance Contributions is accruing the same

flat rate amount of Basic State Pension regardless of their earnings (unlike

SERPS). To enable people to build up a funded pension equal to their Basic

State Pension everyone would be entitled to the same flat rate rebate. This

would need to be quite substantial ? roughly ?12 p.w. at today?s levels ? well

above the threshold needed to justify the fixed element of the costs of setting

up and running a personal pension fund. The same fund could also receive and

invest, with no need for additional charges, employee?s earnings related

rebates - however small - to provide a funded equivalent of SERPS. So Basic

Pension Plus involved giving everyone, as they reached working age, their own

fund into which would be paid rebates from their National Insurance


SMF Save Our Pensions Text 16/9/03 12:06 am Page 26

contributions sufficient to provide a pension at least equal to their Basic State

Pension plus their State Earnings Related Pension.

This proposal was traduced during the election campaign which made it

difficult for the incoming government to develop the idea despite adopting

some of its rhetoric. However, when the Labour government replaced SERPS

with the State Second Pension it changed the structure in a way that means

the problem of low earners receiving rebates inadequate to justify setting up a

personal pension is no longer inherent to the system. (See Annex A for a fuller

explanation of the structure of S2P.) It is true that at present the rebates for

employees contracting out of the new State Second Pension are proportional

to earnings just as in SERPS. But this feature seems to have been carried over

from SERPS even though the amount of pension accrued under S2P is no

longer proportional to earnings for those on low pay. Under the S2P the

amount of pension rights accrued does not decline if earnings fall below a

certain threshold (the Lower Earnings Threshold which is set at roughly half

national average earnings). Employees earning say ?5,000 or ?7,000 will accrue

the same amount of S2P as someone earning ?11,200 which is the current

Lower Earnings Threshold. Yet if those with earnings below the LET contract

out of S2P into a personal pension they still receive a rebate proportionate to

their earnings. (They continue to accrue an element of S2P based on the

difference between their actual earnings and the Lower Earnings Threshold.)

The earnings related rebates mean that it remains unattractive for those on

low earnings to contract out of S2P.

The Mandatory Funded Second Pension

The government could equally well have allowed employees earning below the

LET to contract out of all the pension rights they accrue under S2P and receive a

corresponding rebate. Since they accrue a flat rate amount of pension

entitlement within S2P they should be entitled to a flat rate rebate in respect of

earnings up to the LET. Such a rebate would be above the threshold level necessary

to make it worthwhile to set up and handle a personal pension. The problem of

minimal rebates being absorbed by the disproportionate costs of running small

funds would therefore disappear. So it would no longer be necessary to provide

the option of the unfunded S2P for those on low earnings. That option could then

be withdrawn at least from new entrants to the labour market.

All new entrants would instead be required to join a Mandatory Funded

Pension scheme. Those earning less than the LET would receive a flat rate


SMF Save Our Pensions Text 16/9/03 12:06 am Page 27

rebate. Those earning more than the LET would receive the same rebate as

employees who contract out get at present - comprising that flat rate element

plus an element related to their earnings above the LET.

The MFSP compared with the government?s plan.

The government does not appear to have considered this option. It has,

however, considered ways of restricting employee?s rights to choose between

the unfunded state system and a funded alternative in Phase Two of S2P. This

will come in once S2P has bedded down. The government did originally

consider removing the right to remain in the unfunded S2P from all those

earning above the LET and requiring them to join a funded private scheme.9

Conversely, no-one earning below the LET would be allowed to contract out

into a funded personal pension. They would stay within the unfunded state

scheme accruing their flat rate entitlement to S2P.10

Those proposals seem to have survived a last minute rewrite of the White

Paper. The rest of the White Paper left an element of choice at least for those

earning above the Lower Earnings Threshold. They will be able to stay in the

unfunded state system. But if they do so they will only accrue the flat rate

component of S2P. By contrast those who opt for a funded private pension

scheme will receive rebates equivalent both to the flat rate and earnings

related elements of the S2P as at present. They will no longer be able to

contract back into the state pay-as-you-go scheme.

Under these proposals there would clearly be a strong financial inducement

for those earning significantly more than the LET to opt for a private pension.

But they would not be required to do so. By allowing employees to stay in the

unfunded state scheme on such disadvantageous terms the government will be

guilty of blatant ?misselling?. This two tier system will also create a very odd

situation for those whose earnings fluctuate above and below the LET.

Moreover, it is rather offensive to treat those on low earnings as second class

citizens. They will not have their own fund into which they could put any

additional savings.

Yet someone earning say half the LET accrues exactly the same entitlement

to S2P as someone earning the LET. The employee earning exactly the LET will

either be permitted or required to have a pension fund and will receive a

rebate sufficient to provide a pension equivalent to the flat rate element of

S2P. The employee earning only half the LET will have to remain in the state

unfunded system. Low earners are apparently deemed unsuitable to hold a


SMF Save Our Pensions Text 16/9/03 12:06 am Page 28

pension fund or unlikely to use it to make additional savings.

I propose that everyone entering the labour market after a certain date

should be treated equally. Over time the whole of the State Second Pension ?

both flat rate and the earnings related elements - should be replaced by an

equivalent Mandatory Funded Second Pension. Everyone earning enough to

entitle them to any S2P should receive a rebate equivalent to that entitlement

payable into their personal or occupational pension fund.

Benefits of a Fully Funded Second Pension

The benefits of such a scheme would be six-fold.

i Meeting the government?s target

It would enable the government to meet its target of reducing the

unfunded element of the nation?s pension liabilities to 40%.

How rapidly it did so would depend on the speed with which it was

phased in.

To do so for all employees not currently contracted out would cost

over twice as much as the rebates of those currently contracted out.

See Tables 1 and 2. These figures reflect the current levels at which S2P

accrues. If this flat rate element was raised as suggested by the ABI to

lift everyone above the level of the Minimum Income Guarantee

indexed to earnings, the rebates for the flat rate element would

increase correspondingly. This would be partly offset by a reduction in

the earnings related element unless the upper earnings limit was also


It might be more feasible to introduce the scheme by age cohorts.

The most gradual way would be (as proposed for Basic Pension Plus) to

require all young people when they get their first job to have a personal

fund into which their rebates would be automatically paid. That would

involve additional rebates mounting initially at a rate of less than ?200

million11 p.a. cumulatively.

In practice it should be possible to include within the scheme from

the start everyone below the age of, say, 30. The initial value of rebates

to fund pensions at the existing level of S2P would then be ?2.2 billion

more than at present, mounting by an additional ?35013 million p.a.

That compares with the current cost of rebates of ?11.5 billion for

everyone who voluntarily contracts out at present


SMF Save Our Pensions Text 16/9/03 12:06 am Page 29

Table 1:

Value of NI Rebates for those Voluntarily Contracted Out of S2P in 2002/03

(GB accruals - ?billion)

Age < 30 Age > 30 All Ages

People earning below LET12 0.1 0.5 0.6

People earning above LET 1.3 9.6 10.9

All earners 1.4 10.1 11.5

Source: DWP/GAD replies to author

Table 2:

Value of NI Rebates for All Earners whether or not currently contracted out

(2002/03, GB accruals, uncapped APP rebates ? ?billion)

Age < 30 Age > 30 All Ages

Those earning below LET

? value of rebates for flat rate S2P 1.0 3.6 4.6

Those earning above LET

? value of rebates for flat rate S2P 2.2 13.5 15.7

? value of rebates for earnings

related accruals 0.7 6.4 7.1

3.9 23.5 27.4

Source: DWP/GAD replies to author

ii Encouraging voluntary saving

As everyone within the scheme would have their own fund they would

be able to make additional contributions into it as they wish. Nearly

half of those who contract out of SERPS do save extra on top of the

National Insurance rebate. The Australian experience suggests that

since people have been required to have a fund it has become much

easier to persuade them to make additional voluntary contributions.

The proportion of those making voluntary contributions on top of the

compulsory minimum has doubled to 40% compared to 20% previously.

On average voluntary pension savings now amount to an additional 3

to 4% of income on top of the 9% compulsory minimum.


SMF Save Our Pensions Text 16/9/03 12:06 am Page 30

Automatic ownership of a personal pension fund would make it

much easier for employees to make additional voluntary savings. But

experience suggests that the proportion who do so is strongly

influenced by advice and encouragement from their employer.

Matching contributions from the employer are a powerful incentive to

save more than the minimum. Research conducted by Watson Wyatt for

Winterthur14 shows that a programme of education in the workplace

can have a powerful multiplier effect. Employers who offered to

contribute an additional 25% of whatever voluntary contributions

employees make to a 401k plan and backed this with a targeted

communications programme achieved almost as high a take up as did

companies who offered a 100% match but carried out no

communications programme. In the USA full tax benefits on 401k plans

are only available to company owners or top management if the plans

are available on the same terms to all employees and a high level of

take up among all employees is achieved. So managers have an

incentive to carry out such communications programmes. It would be

desirable to make similar provision in UK tax law. At present UK

managers are reluctant to offer any advice to employees even to remain

in the company pension fund for fear of falling foul of the Financial

Services Act. This should be revised to give employers the clearest

possible exemption from such liability.

iii Reducing costs

If people automatically have their own fund the costs of running these

funds will be significantly reduced. The costs of persuasion are largely

eliminated and a standardised, mass marketed product has lower

running costs. Moreover, the running costs as a percentage of funds

under management decline as the amount invested rises over time. The

Association of Superannuation Funds of Australia estimates that the

average administration costs would fall from 1.5 % when the average

balance is US$10,000 to 0.5% when the average balance rises to


A reduction in the annual cost of running a scheme from 1% of

capital invested, which is the norm for stakeholder pensions, to 0.5%

could increase the final pension by some 13%. The Vanguard Group in

the US charge just 0.3% on their tracker mutual funds.


SMF Save Our Pensions Text 16/9/03 12:06 am Page 31

iv Spreading wealth

Every working person would eventually own his or her personal pension

fund. This would constitute the largest distribution of wealth since the

spread of home ownership. Working people who earn less than the

Lower Earnings Limit (currently ?11,200 pa) throughout their working

lives would end up owning a substantial pension pot. If they put

nothing apart from their automatic minimum rebate into their personal

funds they would accumulate a fund worth some ?65,000 in today?s

prices by the time they reached age 65.16 Someone earning a salary just

equal to the Lower Earnings Limit who put in additional voluntary

contributions of 3% of salary ? the average amount contributed

voluntarily by their Australian counterparts ? would accumulate a fund

worth over ?115,000. That would finance a second pension indexed for

inflation equal to nearly two and a half times the Basic State Pension.

v Freedom of Choice

Once people have their own fund they have far more freedom to decide

when they should retire. Yet they will also have the incentive to

continue working (possibly on a part-time basis) since this will both

enable them to continue building up their fund and increase the value

of the annuity they can buy for the remaining years of their life. Studies

show that those saving for pensions on a defined contribution basis

work longer than those in defined benefit schemes or solely dependent

on state unfunded pensions.17

vi Changing attitudes

Finally, widespread capital ownership would give people a greater sense

of having a stake in the economy. Everyone would be more conscious of

having a vested interest in policies that encourage profitability and

growth. The sense of ?them and us?, of capital versus labour, and the

populist idea that we can profit by transferring burdens to some

abstraction called "business", would all diminish.

Arguments against funding all State Second Pensions

Of course there also are arguments deployed against funding pensions.

"Funding is pointless" The IPPR18, in particular, argues that there is no

point in encouraging people to save and invest for their future. They say that


SMF Save Our Pensions Text 16/9/03 12:06 am Page 32

people cannot squirrel away a proportion of today?s goods and services like

nuts. All they do is pile up paper claims on future flows of goods and services.

In reality tomorrow?s pensions will have to be paid out of the GDP then being

generated by those then at work. So, they conclude, it makes no difference

whether tomorrow?s pensioners extract their share of tomorrow?s GDP by

exercising paper claims held in the books of pension funds or by the state

imposing taxes reflecting rights accruing in a pay-as-you-go state scheme.

The key element of this argument ? that tomorrow?s pensions will be paid

out of tomorrow?s economic output ? is of course true. But if the rest of the

argument were valid it would follow that the level of saving in an economy is

irrelevant to the future level of income of its inhabitants. There would be no

point in saving at all. In addition to arguing openly against the need to

encourage more saving it is proposed to divert the ?11 billion of NI rebates

currently saved in pension schemes and spend them on raising current


The fallacy of this argument lies in the supposition that because saving is

done via financial instruments it has no effect on future resources available to

pay pensions. Of course it does. The total level of saving in an economy must,

by definition, equal domestic investment plus the acquisition of assets

overseas. Prices and interest rates adjust to bring that about. So extra saving

must result in more domestic investment or more overseas assets or both. Of

course, this assumes that extra private saving is not offset, or financed by, an

increase in public sector borrowing.

Even if the level of domestic investment were entirely unresponsive to

more saving, extra savings would be invested abroad. That would mean that

tomorrow?s pensions could be financed by claims on foreign economies rather

than taxing our own.

Finally, levying heavy taxes on the future British economy to pay for

unfunded pensions may depress activity and drive it abroad whereas

drawing dividends on investments across the world will not have such an


"Funded investments are too risky". Some argue that recent stock market

falls show that we should not encourage people to invest in the capital

markets for their retirement. Saving for retirement is a long-term process. Even

after the recent bear market pension funds show a good return over the long

term. In any case, it is always a mistake to base policy on short-term

movements in markets ? still less on extrapolating recent trends. In fact the


SMF Save Our Pensions Text 16/9/03 12:06 am Page 33

best time to start broadening reliance on investment in pension funds is after

a stock market fall rather than at the top of a bull market. The Australians

believe their compulsory saving scheme got off to a good start precisely

because it started after the sharp 1987 stock market fall.

Nonetheless, the greater uncertainty pertaining to funded pensions may

affect people at different income levels differently. Low earners might be

expected to be more risk averse than the better off. That is why I believe the

state should guarantee the flat rate element of S2P. If for any reason the

portion of a person?s fund built up from their flat rate rebates does not reach

the level necessary to finance a pension equal to the S2P, the state will top up

that person?s second pension to the level of S2P they would have accrued. (The

extra earnings related rebate which people get if they earn above the Lower

Earnings Threshold would not carry a guarantee. So fund managers would

notionally divide each person?s fund into a flat rate and earnings related


Two objections to such a guarantee have been put forward. The first is that

the state could not afford the cost if markets did very badly. In practice, the

guarantee would only be called in respect of part of the pension for some

pensioners in a specific age cohort. By definition the cost to the taxpayer of

supplementing part of the pension of a section of pensioners cannot be greater

than the cost of paying the entire pension on a pay-as-you-go basis.

The second objection is more material. It is that, on its own, a guarantee of

that kind could lead to investment in excessively risky securities. Fund

managers would know that if the investments failed, their clients? pensions

would be topped up by the state. Yet if the high risks produced high rewards

they would get more than the standard pension.

I propose that fund managers would be required to invest on a Prudent

Person basis, as if that guarantee did not exist. Such rules already exist. They

can be enshrined in statute as in Australia which has largely prevented such

imprudent investments. The guarantee would only apply to the part of each

person?s fund built up from their flat rate rebates. Fund managers would also

be investing the earnings related rebates which would not be guaranteed.

Managers could be required to adopt the same risk profile for both parts of the


Means tested benefits already in theory provide a similar sort of guarantee

and incentive to invest in high risk investments for people approaching

retirement with a pension fund insufficient to lift them clear of the means


SMF Save Our Pensions Text 16/9/03 12:06 am Page 34

testing trap. Yet there is no evidence that financial advisers or fund managers

do put such clients? investments in high risk assets in practice. So this problem

may be more theoretical than real.

"Pay-as-you-go reflects the contract between the generations".

Continental politicians invoke this argument to justify their reliance on future

taxpayers to pay for the pensions they have promised. They say, in effect, that

parents bear the cost of raising their children and in return have always looked

to their children to support them in old age. In mature societies each

generation likewise looks to the next to support it in retirement. That

argument no longer works. If such a contract ever existed this generation has

broken it. They are having far fewer children and are living far longer than

their parents. They have little moral claim on a diminished future work force to

support them in retirement for far longer than they supported their parents?


"One generation will pay twice". Some argue that, although it is desirable

that everyone should save for their retirement, we cannot reach that goal

without making the transitional generation pay twice. The current generation

will have to pay for their parents? pay-as-you-go state pensions while also

building up savings for their own funded pension. That was an issue I had to

deal with in my proposal to move to a fully funded basic state pension (see

Annex B). But it does not apply in the case of the State Second Pension. If

anything, the reverse is the case. The S2P is a new benefit and is partly funded

through voluntary contracting out (as was SERPS, which was still building up).

If previous generations wanted to have a pension in addition to their Basic

State Pension they had to save for it. They did not pass the obligation on to the

next generation. That is what we are doing to the extent that we give people

rights to an Additional State Pension without requiring them to make genuine

savings and investments to pay for it.

"Means testing makes it pointless for the poor to save". The Institute for

Fiscal Studies has highlighted this problem. But their point is that the current

level of means tested benefits for retired people makes it uneconomic to save

voluntarily for a small pension on top of the compulsory state pensions. I

propose a change in the balance between compulsory minimum provision and

means tested benefits that will eliminate that problem. Under the present

arrangements people are already compelled to set aside a certain amount for a

second pension. As far as this compulsory provision is concerned the

disincentives are irrelevant. Nor do the disincentives affect whether the


SMF Save Our Pensions Text 16/9/03 12:06 am Page 35

compulsory provision for a second pension should be via a state pay-as-you-go

scheme or a private funded one.

9 "the proposed system?compulsory funded pensions for those earning over [the LET]" page 105:

A new contract for welfare: Partnership in Pensions Cmd 4179 December 1998.

10 ibid "..we expect the new State Second Pension to become a flat rate scheme for those on lower

earnings, with those on moderate and higher earnings joining a funded pension (with

contracted out rebates continuing to be earnings related)."

11 Estimated from figures provided by DWP/GAD for total costs of rebates for year cohorts of

different ages whether or not currently contracted out.

12 For people earning below the LET and for people contracted out through occupational schemes

who earn below the HET (?24,000pa in 2002/03), the rebate does not reflect the full value of

S2P. S2P top ups are paid in retirement in order to compensate for this. Expenditure on such top

ups is excluded from the figures shown.

13 See footnote 12.

14 Response by Winterthur Life UK Ltd to DWP Green Paper "Simplicity, Security and Choice".

15 Schieber SJ & Shoven JB: "Administering a Cost Effective National Program of Personal Security

Accounts" NBER, Cambridge MA, December 4, 1998.

16 Assumes earnings and thresholds rise 1.5% p.a. faster than prices.

17 Prof Blake, The Pensions Institute.

18 A New Contract for Retirement, R.Brooks, S. Regan & P. Robinson, IPPR.


SMF Save Our Pensions Text 16/9/03 12:06 am Page 36



If the government chooses not to replace the pay-as-you-go State Second

Pension by a Mandatory Funded Second Pension, the only way they can hope

to meet their 60% funding target is by raising the State Pension Age.

They have already begun to debate this issue by pointing out the truism

that the only solution to the pensions crisis is "either to save more, to work

longer or a mix of both"19 The government is also trying ?to lodge the idea in

the public consciousness that retirement at 70 will mean a higher pension?, as

Alan Pickering has pointed out.

If the government goes down this route, the sensible approach would be to

index the State Pension Age to life expectancy. On average, life expectancy of

those aged 65 has increased by about 1 month a year over the last twenty

years. If the government raises the State Pension Age at that rate in future,

someone now aged 21 would have to work until they were aged 68 years and

8 months before drawing their state pension.

Ideally, however, the government should not be telling people how long

they must work ? still less that they must work ever longer. The great

advantage of enabling everyone to build up their own pension fund is not just

that it helps meet the government?s funding target without raising the SPA. It

also leaves people free to decide when to retire once their fund is adequate to

provide a decent pension (i.e. greater than the Minimum Income Guarantee).20

Moreover, it gives everyone a double incentive to work and save longer. Every

extra year they work gives them a year longer to save and a year less

retirement to finance, enabling them to buy a larger pension for a given sum.

Those advantages of working longer are substantial. Take a fairly typical

man whose pay rises at 3% p.a. in real terms until he is 50 and then remains

flat. If he wants to retire at 65 on an indexed second pension worth half his

final salary he would need to save 12.1% of his salary each year throughout his

working life. However, if he intends to keep working until age 70, he will need

to save only 7.5% of his salary. Put another way, the cost of an indexed single

life annuity worth ?10,000 p.a. at age 65 is ?195,000. Buying the same annuity

at age 70 would cost ?156,000.21

One way the government could reinforce that incentive to delay


SMF Save Our Pensions Text 16/9/03 12:06 am Page 37

retirement, without altering the State Pension Age, would be to use the tax

system. The major tax advantage of saving for a pension is what Nigel Lawson

called "the anomalous but much loved lump sum". At present up to 25% of a

personal pension fund (excluding that portion derived from NIC rebates) can

be withdrawn tax free as a lump sum. The lump sum can be spent as the saver

wishes. It does not have to be used to buy a retirement income. The age at

which that valuable privilege may be exercised could be raised in line with

average longevity. People would still be free to withdraw their funds, including

the tax free element, before that age as long as they used all the money to

provide an annuity or draw down income sufficient to avoid becoming

dependent on means tested retirement benefits.

A number of criticisms could be made of simply leaving people free to

decide whether to cease working once their fund is adequate to escape means

tested benefits:-

i Most people will opt to retire and take their pension at the earliest

permissible date even if their pension is meagre. Only 8% of men work

beyond their State Pension Age of 65 though nearly a third of women

retire after 60.22 It is certainly true that very few people avail

themselves of the current option of deferring the state pension even

though they can thereby increase their future state pension. The

increase is one seventh of 1% for every week deferred, which is about

7.4% for each year of deferral, up to a maximum of 5 years. The lack of

interest is partly because this option is little known. Also the

enhancement is not particularly generous. The current terms are

roughly neutral on an actuarial basis. They will therefore be less than

neutral for those with below average life expectancy. That is why I

decided to improve the terms to one fifth of 1% per week (10.4% for

each year?s deferral) and to remove the limit on the number of years of

deferral. The government?s decision to bring that forward is welcome.

The experience of the self-employed suggests that when people

own their personal pension fund the double incentive to work longer is

effective. Most self-employed people own a money purchase pension

fund. In recent years their average age of retirement has increased

relative to that of members of Defined Benefit Schemes, including

those relying on SERPS.23


SMF Save Our Pensions Text 16/9/03 12:06 am Page 38

Fig 1. Employment, unemployment and inactivity rates for men aged 50-64;

UK; spring 1984 to spring 2001

ii People in manual/stressful/low paid jobs do not want to, and often

cannot, work past the State Pension Age. To those engaged in

stimulating jobs the idea of working beyond 65 might be positively

appealing even if it had no impact on their eventual pension. But is it

reasonable to expect the man operating a pneumatic drill to go on

working into his late sixties? Nearly half of all men have ceased to work

by age 62.24 Early withdrawal from the labour market is particularly

prevalent among the less skilled. 41% of men aged 50 to 64 with no

qualifications were inactive compared with 22% of those with GCSElevel

or higher. Three quarters of men aged between 50 and 64 who are

economically inactive say they do not want a job. 62% of plant and

machine operatives gave their reason for inactivity as being long-term

sick. By contrast less than a quarter of managers and professionals were

long-term sick and over 55% said they had retired ? often with an early

occupational pension.25











2000 1998 1996 1994 1992 1990 1988 1986 1984

In employment

Economically inactive

ILO unemployed

SMF Save Our Pensions Text 16/9/03 12:06 am Page 39

Even if the manual groups had possessed a personal pension fund,

the prospect of adding to it would have been unlikely to persuade many

of them to work longer to build up a better pension. However, the

reluctance of lower earners to work up to, let alone beyond, the State

Pension Age is already a problem under the present system. The changes

I have proposed so far would not solve that problem but nor would they

make it worse. The solution is to improve the working of the labour

market so that people in their 50s can find jobs which are neither

physically stressful nor demanding of lengthy experience.

There are signs that this is happening both in the USA and the UK.

In the USA, the decline in participation rates among men aged

between 60 and 64 halted in the early 1990s and has begun to reverse

since then particularly among ?high school drop outs? ? the group with

the lowest participation rate26.

Likewise in the UK the employment rate among men aged 50 to 64

has risen by 5 percentage points since its nadir in 1993.27

Given appropriate job opportunities, the possession of a personal

pension fund may well encourage people to defer retirement or to take

on part-time jobs to supplement an early retirement pension.

iii Lower income/manual workers have lower than average life expectancy

so they subsidise the pensions of longer-lived, higher income groups.

This is true of the pay-as-you-go state scheme as well as private funded

pensions. At age 65 the life expectancy of an unskilled male manual

worker is over four years less than that of a male professional.28 If both

types of employee pay premia into the same fund over their working

lives, the typical professional will draw out four years more pension

than the manual worker.

That is not merely inequitable, it is positively perverse. The less well

off are subsidising the rich.

The situation is accentuated at present by the fact that the

annuities market is dominated by the better off who happen to be


Annuity rates reflect this since they are calculated on the basis of

the mortality of actual annuitants rather than the average population.

As a result an average member of the population would have to pay 7

or 8% more for an annuity than their average life expectancy warrants.


SMF Save Our Pensions Text 16/9/03 12:06 am Page 40

Fig 2. Life expectancy for manual and non-manual groups

Making personal pensions universal and mandatory would

automatically help remedy this. It would do so by bringing the

characteristics of those buying annuities into line with the total

population. However, the fundamental problem of the less well off

subsidising the longer-lived better off would persist.

Funded pensions would also make it possible to do something about

that. In the first place annuity providers could be encouraged to relate

annuity rates to the life expectancy of different income or social

groups.29 At present "insurers in the UK rarely use socio-economic

factors in pricing" annuities.30

Those who have lowest life expectancy ? generally those on lowest

incomes ? should get the best annuity rate when converting their fund.

Annuity providers do offer advantageous annuity rates to smokers and

higher ?impaired life? annuities to individuals with specific life










1972-76 1997-99 1972-76 1997-99







at age 65



17.4 17.4


SMF Save Our Pensions Text 16/9/03 12:06 am Page 41

threatening conditions. Some providers also vary their rates according

to the size of fund converted into an annuity ? because wealthy people

with larger funds tend to live longer. Anecdotal evidence suggests that

the cost per ?10,000 in the fund of buying a large annuity (say ?50,000

p.a.) can be up to 19% more than for a small annuity (?2,000 p.a.).

Some 10% of the total annuity market (20% of the open market

business) is already written on special bases like these. This element is

rising rapidly.

However, annuity providers do not yet offer differential annuity

rates related to past earnings or socio-economic group. It is not clear

why they do not. It may partly be that the providers do not have

information on individuals? life-time earnings. To obtain and verify it

would involve disproportionate cost. The government might be able to

overcome this by making available information on historic earnings

from the National Insurance Recording System. This could possibly be in

a form equivalent to a tax code.

Another option would be to encourage the development of pension

funds limited to members of particular occupations. Some funds

limited to people employed in a specific industry already exist -

notably the Buildings and Civil Engineering Benefit Schemes (B&CE),

and also the Heating and Ventilation Engineers. The B&CE is the largest

supplier of stakeholder pensions. It guarantees zero charges on these

policies up to 2006. The Australian system is largely based on industrywide

funds. These pool the contributions of all employment levels

within each industry. Pooling of investment funds on an occupational

basis will only be advantageous for members if the average life

expectancy is below that of other insured lives and if the scheme

provides annuities which reflect that. B&CE for example can offer its

members annuities some 18% higher for a given sum than is the norm

in the open market.

Schemes limited to manual workers could result in them receiving

an annuity perhaps 30% higher per ?1,000 of savings than a manager.

It is not at all clear why such annuities are not on offer. This may be

because the providers have felt that such provision would be politically

unacceptable. If so they need to be persuaded otherwise.

A third option is for the State to relate employees? rebates to the life

expectancy of their income group. That would mean paying larger


SMF Save Our Pensions Text 16/9/03 12:06 am Page 42

rebates to those on low incomes. Under the State Second Pension

accrual rates are biased in favour of those with earnings below the

lower earnings threshold (currently ?11,200). However, this is not fully

reflected in the rebates available for the few who contract out of the

state scheme in that income bracket.

19 Hansard col 403 13 Jan 2003 Andrew Smith Secretary of State for Work and Pensions.

20 Anyone wishing to use their fund to retire before the SPA would be required to buy a fixed term

annuity for the period up to the SPA equal to at least the MIG followed by an annuity equal to

the difference between the MIG and the BSP. Legal & General have advocated permitting

purchase of such fixed term annuities: L & G?s response to "Modernising Annuities" 9th April


21 The calculation assumes net real return on savings of 4% per annum; earnings rise by 3% in real

terms between age 21 and 50 then remain stable in real terms until retirement at 65, or 70 if the

individual continues working. Typical annuity rates are taken from the FSA comparative tables

for a male non-smoker single life: ?100,000 buys ?5,124 p.a. at 65 and ?6,429 at 70.

22 "The Dynamics of Retirement: Analyses of the Retirement Surveys" by Richard Disney, Emily

Grundy and Paul Johnson. DWP Research Report No. 72.

23 Also see "The Impact of Wealth on Consumption and Retirement Behaviour in the UK", David

Blake, The Pensions Institute.

24 Hansard 13th May 203 DWP written answer to Peter Lilley MP.

25 "Patterns of Economic Inactivity Among Older Men" Catherine Barham, ONS Labour Market

Trends June 2002.

26 ?The Puzzle of Later Male Retirement? Richard Johnson Economic Review, Q3 2002, Federal

Reserve Bank of Kansas City.

27 "Patterns of Economic Activity among Older Men" Catherine Barham, Office of National

Statistics in Labour Market Trends June 2002.

28 The gap between Social Classes I and V widened from 2.6 years to 4.1 years between 1972-6 and

1997-9. "Trends in Life Expectancy by Social Class" ONS 18th February 2002.

29 Annuities purchased with the compulsory element of pension savings must be priced on a

unisex basis. I do not propose any alteration in this.

30 ibid page 5.


SMF Save Our Pensions Text 16/9/03 12:06 am Page 43


The recent gloom about the cost of providing pensions has focussed on the fall

in the stock market and the increase in longevity. There is little the government

can do about the former or would want to do about the latter. Remarkably

little thought has been given to reducing the non investment costs of

providing pensions. Yet this is the one area where improvements could be

made which could ameliorate the position.

Pensions involve two stages: the accumulation of a pension fund and the

subsequent payment of an annuity or pension. The first stage has by far the

highest costs. But there are also opportunities to improve value in annuity

provision to which very little attention has been paid.

Costs of accumulating pension funds

The costs of accumulating pensions can be surprisingly high ? especially on

personal pensions. A study in 199931 put the total cost at some 36% of the

accumulated balance over a working lifetime. A significant proportion of this

cost came from the cost of transferring accounts, opening duplicate accounts

and ceasing to contribute to paid up accounts. Even the 1% per annum cap on

costs which providers of stakeholder pensions are permitted to charge can still

reduce the accumulated balance over a working life by between 20% and 25%.

Put another way, a charge of 1% per annum on funds in a personal pension

would absorb nearly a quarter of the real return of 4% per annum projected by

the Government Actuary.

The factors which have tended to swell the costs of personal pensions are the

costs of persuasion, advice, compliance, risk of lapse and/or switching to another

provider all on top of the actual cost of managing the fund and administering

each account. An analysis32 of Life Office expenses showed that two thirds of

their costs are accounted for by the cost of acquiring customers ? especially

paying commissions. Even mutual funds with commission free sales spend half

their expenses on acquiring customers. One reason occupational funds have

lower costs33 is that they do not need to attract members. Australian figures

suggest that the costs of running occupational schemes are typically half those

of personal pensions. Preliminary results from a study34 of British occupational

schemes suggested that their costs, though lower than those of personal

pensions, are not far below the 1% annual limit set for stockholder pensions.


SMF Save Our Pensions Text 16/9/03 12:06 am Page 44

The government has recently indicated that, rather than reducing the 1%

cap on stakeholder fees, it is prepared to consult about raising it. Is there any

scope for reducing those costs? The last Conservative government initiated a

number of steps to bring about greater transparency, simplicity and

competition to bear down on costs. These have been carried forward with the

Pickering Report, the Sandler Report and the Inland Revenue report on Tax

Simplification. These make many welcome recommendations but at best they

will bring about marginal reductions in costs and fees.

The simple truth is that there is only one way to cut costs ?at a stroke?. That

is to make it compulsory for everyone to have their own pension fund

throughout their working lives. Compulsion reduces the need for persuasion ?

the most costly element in the process; dramatically reduces the need for

advice; means that the product can be standardised and increases the volume

of investment over which costs can be spread.

The Australian experience35 is revealing in this respect. Even though there is

no cap on costs they are coming down steadily as a proportion of funds under

management. And they are expected to fall significantly further as the average

balance rises. The Australian experience is one to learn from - not to copy

slavishly. They have paid remarkably little attention until recently to the level

or structure of costs and charges. Their costs could well have come down

further and faster if they had done so earlier ? for example by automatically

merging the multiple accounts which arise when people move jobs or by

imposing a cap on charges.

Over time it might be possible to reduce the charges on the compulsory

element of saving in UK Stakeholder type funds from 1% to as low as 0.5% per

annum. That sounds a small saving. In fact it could boost the value of the

pension accumulated over a working life by 13%.

Costs were coming down substantially in the UK even before the cap on

stakeholder costs was introduced. Table 3 shows that they fell by nearly half

between 1989 and 1997. It also indicates that very significant scope remains

for further reductions given that the quarter of companies with lowest costs

have an expense ratio less than half that of the quartile with the highest costs.

At the same time, the cap could be retained at a higher level (or

conceivably removed entirely) on any extra savings above the mandatory level

that people were encouraged to put into their personal funds. This would

restore the ?reward for persuasion? which has largely disappeared at the present

level of stakeholder provision.


SMF Save Our Pensions Text 16/9/03 12:06 am Page 45

Table 3

Life office expenses

1997 1994 1989

Average expenses ratio* (basis points) 130 192 236

Expense ratio ? first quartile* (basis points) 62 123 174

Expense ratio ? third quartile* (basis points) 156 171 270

Average share of business acquisition in total costs* (percent) 65 69 73

Average share of commissions in total costs* (percent) 31 32 33

Sample size 146 157 173

*weighted average ratio of expenses to funds invested.

Source: Murthi, Orszag & Orszag March 1999 based on Synthesis Life analysis of statutory returns to DTI.

Reducing the Cost of Annuities

There has been great resentment among those reaching retirement in recent

years about the decline in annuity rates. In 1993 ?10,000 would buy a 65 year

old male a lifetime income of ?1245 p.a. In 2003 it buys just ?741 p.a.36

This is partly due to falling interest rates reflecting declining inflation. So it

means retired people who choose a cash annuity, as most do, will not see the

purchasing power of their incomes eroded so rapidly.

The other reason is the sharp increase in assumptions of life expectancy.

Insurance companies are very good at pooling risks across a population with

stable characteristics. They are less good at predicting trends in the

characteristics of the whole population ? still less at coping with possible but

unpredictable events like a magic bullet cure for cancer.

All they can do is make provision against trends and events which are

possible though unlikely. Those provisions have to be factored into the cost of

annuities. So it is more likely than not that annuitants will have paid for

greater longevity than they will collectively enjoy.

The uncertainty about future mortality rates is greatest in relation to the

more distant future and the later stages of life.

In pricing an annuity for those currently retiring, the providers can assume

that the proportion of those aged 65 who will die in the next five years will

reflect recent mortality rates and trends among 65 to 70 year olds. Any

significant medical advances affecting the period are likely to be in the

pipeline and therefore identifiable. But it becomes progressively harder to be

certain about future mortality rates for older ages. The ages showing greatest

improvements in mortality seem to be steadily rising.


SMF Save Our Pensions Text 16/9/03 12:06 am Page 46

The Insurance Regulations 1994 require that the price of annuity liabilities

"be determined on the basis of prudent rates of mortality". So annuity

providers should err on the side of caution ? which increases the cost of

annuities. Some argue that they may nonetheless be significantly

underestimating the risk of mortality rates improving.37 If so they risk being

unable to pay the annuities they have promised, should the proportion of

people living to a great age increases beyond what they have provided for.

Either way annuitants will lose out.

There is therefore a plausible case for government bearing the ?risk? of

financing pensions for the later years of retirement.

The situation is analogous to that of financing residential care. When the

last Conservative government was considering how to encourage greater

private provision we discovered that the average period spent in residential

and nursing care was fairly short. But that average included a minority of

elderly people who spent many years in residential care. The insurers faced an

unquantifiable risk that this minority and their duration of stay would

increase. We found that if government accepted the responsibility for funding

residential and nursing care for those who stayed well beyond the average

term, insurers could reduce their premia disproportionately and more providers

would enter the market.

By analogy, the government could assume responsibility for paying S2P for

the most elderly ? say, those living beyond 85.38 People would then only need

to accumulate a fund to buy a 20 year fixed term annuity at their S2P level to

cover their retirement between 65 and 85. If they wanted a retirement income

higher than the S2P level they would have to buy an open ended annuity for

the whole period of their retirement.

In Chapter 3 I postulated that the government may decide to raise the

State Pension Age (SPA) by about a month each year in line with the rise in

average life expectancy. In that case the state should be committed to start

paying S2P to each cohort 20 years after their SPA, rather than specifically at

age 85. Everybody would then be required and enabled during their working

life to build up a Mandatory Second Pension fund to provide for a fixed period

annuity for the first 20 years of their retirement.

It is hard to calculate the cost of such fixed period annuities relative to full

life annuities. The mortality tables suggest that 87% of the cost of an ordinary

annuity for a 65 year old covers the period to age 85. But the funding need

should be reduced by more than 13% since providers will no longer need to


SMF Save Our Pensions Text 16/9/03 12:06 am Page 47

build up a ?provision? for the major uncertainty relating to the later years.

It is difficult to tell the actual level of annuity companies? costs and

provisions against risk. A number of studies have suggested that annuities are

quite keenly priced. These studies are based on comparing the cost of

purchasing an actual annuity with the cost of government bonds necessary to

provide that income stream for a population using published mortality


This methodology suggests that costs (including provisions) amount to

about 5 or 6% of the price of an annuity. However, the annuity providers may

be investing partly in corporate bonds and equities and using the higher yields

to offset a higher level of costs. If so, the cost margin may be substantially

higher than 6%.

So there may be significant scope for cost reductions if government takes

on the cost of S2P for the over 85s.

31 ?The Charge Ratio on Individual Accounts: Lessons from the UK Experience? by Murthi, Orszag &

Orszag (March 1999).

32 Ibid page 44.

33 This is well-documented in Australia in ?Superannuation Fees and Competition? by Michael Rice

and Ian McEwin of Phillips Fox (9 April 2002).

34 Appendix II of Murthi, Orszag & Orszag (March 1999).

35 ibid.

36 NAPF.

37 "?the new Continuous Mortality Investigation projection basis significantly underestimates

likely future mortality improvements ? Life Offices writing annuity business on competitive

terms may be making significant losses ? some pension scheme liabilities may be

underestimated by as much as 30%." ?Mortality in the Next Millennium? by Richard Willets FFA.

38 An alternative way for government to bear this risk would be to issue bonds whose annual

coupons reflect the proportion of the population of retirement age on the issue date who

remain alive in each subsequent year ? as proposed by David Blake and William Burrows in

"Survivor Bonds: Helping to Hedge Mortality Risk" The Journal of Risk and Insurance 2001 vol 68

no 2.


SMF Save Our Pensions Text 16/9/03 12:06 am Page 48



Both employees and the self-employed should be required to contribute to a

Mandatory Funded Second Pension at least sufficient, in conjunction with

their Basic State Pension, to lift their retirement income above the means

tested benefit level ? the MIG.

The Mandatory Funded Second Pension should provide a flat rate element

at least equal to the difference between the Minimum Income Guarantee

(MIG) and Basic State Pension (BSP). It could also provide an earnings related

element of pension mirroring that currently included in S2P.

It would be virtually impossible for the BSP and Second Pension to keep

pace with the MIG if government continues to raise the MIG in line with

average earnings and the BSP only in line with prices. Instead both should be

statutorily up rated in line with prices and any increases above that should be

the same additional cash amount for the BSP as for the MIG as and when the

public finances permit.


Ultimately everyone in work should have a Mandatory Funded Second Pension

? either a personal or occupational pension - into which would be paid NIC

rebates sufficient to fund a pension/annuity equivalent to the S2P during the

first twenty years of their retirement. Thereafter the S2P will be paid on a payas-

you-go basis by the State.

However, this move to compulsory funding could be phased in. Those born

less than 30 years before the start of the scheme would be required to have

such a fund from that point or when they entered employment. Those aged

over 30 when the scheme starts would retain the current option to remain in

the state pay-as-you-go second pension or contract out into an equivalent

funded scheme.

Cutting Costs

Requiring everyone to have a pension fund is the most effective way to cut the

costs of running those funds. It eliminates the cost of persuasion, dramatically

reduces the need for advice, simplifies and standardises and spreads costs far

more widely. It should eventually be possible to reduce the cap by up to a half.


SMF Save Our Pensions Text 16/9/03 12:06 am Page 49

That would increase the value of final pensions by as much as 13%.

The cost of annuities can also be cut disproportionately if government

assumes responsibility for paying the Second Pension from age 85. That would

transfer 13% of the cost of expected longevity to the state. But the cost of

annuities for a maximum 20-year period are likely to be reduced by

significantly more than that amount because it would no longer be necessary

to make provision for unlikely but possible increases in longevity beyond that


Pension Age

People would be free to draw an income from their compulsory second pension

fund earlier than the State Pension Age (SPA) if it is sufficient to provide an

income above the MIG. They would, however, have a double incentive to go on

working and saving beyond the SPA.

Should the government fail to adopt these proposals it will be forced to

raise the SPA. If it does so the logical approach would be to raise it roughly in

line with life expectancy, i.e. by one month each year. It could, instead, raise in

this way the age at which people can draw their tax free lump sum unless they

are using it to buy an annuity.

These proposals would make it unnecessary to change the SPA though the

government might decide to do so anyway. In which case it would be logical to

increase the age at which the state takes over funding the second pension on

a similar basis.

To remove the unfairness whereby those on lower incomes who have

shorter life expectancy subsidise the better off, longer-lived groups the terms

of annuities should reflect the relationship between life expectancy and


To make this possible, the Government should make available to annuity

providers information from National Insurance records summarising people?s

lifetime earnings in a form similar to their tax code. Alternatively National

Insurance rebates could be adjusted to reflect the relationship between

earnings and life expectancy.

Funds or annuities which more fairly reflect that relationship will make it

easier for lower income groups to exercise genuine choice over the age at

which they retire.


SMF Save Our Pensions Text 16/9/03 12:06 am Page 50



This summary of how the present system works may at least convey to the

reader how complex it has become. Much of that complexity arises from recent

attempts to mitigate the disincentive effects of means testing. Unfortunately

that very complexity itself undermines the savings culture and causes many

otherwise prudent people to shy away from making any pension provision at



The state system for retirement income has three components: means tested

retirement benefits; the Basic State Pension; and the Additional State Pension

or rebates to obtain an equivalent private pension.

Means Tested Retirement Provision (MIG & PC)

Minimum Income Guarantee (MIG)

There has been a means tested safety benefit since the welfare state was

established. Originally it was called National Assistance, then Income Support.

In 1999 Income Support for those above the State Pension Age (SPA) was renamed

the Minimum Income Guarantee (MIG) and set at a level higher than

the Income Support level. In 2003/4 its basic value was ?102.10 per week for a

single person and ?155.80 for a couple.

Anyone over the SPA with income below the MIG is entitled to have their

income topped up to the MIG level (subject to an assets test). Every extra ?1 of

pension, up to this level, therefore results in ?1 less of means tested benefit.

Pension Credit (PC)

To mitigate this disincentive, Pension Credit is being introduced from October

2003. It will have two components ? the Guarantee Credit which tops people?s

income up to the MIG level and the Savings Credit to partly compensate those

whose retirement income (additional to the BSP) results in a loss of

entitlement to Guarantee Credit. Retired people will be entitled to Savings

Credit equal to 60p for every ?1 of extra pension in the range between the

value of the full Basic State Pension and the MIG. For those whose extra


SMF Save Our Pensions Text 16/9/03 12:06 am Page 51

pension income takes them somewhat above the MIG level their entitlement to

Savings Credit is reduced by 40p for every ?1 by which their pension exceeds

the difference between the BSP and the MIG. So instead of losing ?1 for every

?1 of extra pension in the range between the BSP and the MIG people lose 40p

for every ?1 in a range two and a half times as great.

Basic State Pension (BSP)

Everyone in work pays National Insurance Contributions (NICs) towards their

Basic State Pension (BSP) if their earnings are above a minimum threshold

called the Lower Earnings Limit (LEL).

The LEL is by convention set at about the same level as the BSP (since it

would be odd to force people to save to have a higher income in retirement

than they have in work). The LEL is ?4,004 in 2003/4.

Entitlement to BSP builds up on the basis of the number of periods in

which people make NICs or receive NIC credits (regardless of the amount of

NICs they pay in any period). It takes 44 years of contributions and credits for

a man, and 39 years for a woman, to earn full BSP. (The number of years

required for full pension can be reduced for periods caring for young children

or disabled relatives.)

The BSP in 2003/4 is ?77.45 per week for a single pensioner and ?123.80

per week for a couple. It is up rated each year by at least the rate of inflation.

Additional State Pension (ASP)

The first Additional State Pension (ASP) ? additional to the BSP ? was the very

modest Graduated Retirement Pension introduced in 1961. It was replaced by

the State Earnings Related Pension Scheme (SERPS) in 1978 which was

replaced by the State Second Pension (S2P) in 2002.

State Earnings Related Pension (SERPS)

The State Earnings Relation Pension (SERPS) covered only employees (not the

self-employed) who earned above the LEL and were not contracted out.

Each year employees? earnings between the Lower and Upper Earnings

Limits were taken into account. That element of each person?s earnings was up

rated in line with rises in an index of average earnings until they reached the

State Pension Age.

The rules for calculating the SERPS pension entitlement have changed over

time. Originally the SERPS pension was to equal 25% of the average of the best


SMF Save Our Pensions Text 16/9/03 12:06 am Page 52

20 years of these re-valued earnings. Subsequently the pension was to equal

20% of the re-valued relevant earnings averaged over a full working life (44

years for men, 39 for women).

Once in payment a SERPS pension is increased in line with prices.

Contracting Out of SERPS

Employers running an occupational pension scheme offering benefits

equivalent or superior to SERPS can contract out of SERPS all their employees

who are members of the scheme. Likewise any individual employee who takes

out an Approved Personal Pension (i.e. one offering roughly equivalent or

greater benefits) can contract out of SERPS. In both cases rebates from

employer and employee NICs will be paid directly into their pension funds. The

rebates are calculated by the Government Actuary to be sufficient, when

invested over a working life, to pay for a pension or annuity equivalent to the

SERPS rights foregone.

Rebates payable into Approved Personal Pension Funds were related to the

employee?s age.

On average rebates are equivalent to nearly 5% of relevant earnings.

State Second Pension (S2P)

Like SERPS, S2P covers only employees (not the self-employed) earning above

the LEL who are not contracted out. Pension entitlement is also based on each

year?s earnings between the Lower and Upper Earnings Limit, re-valued in line

with average earnings up to State Pension Age.

However, the pension entitlement is not simply proportionate to this

element of earnings. S2P is very redistributive towards those earning less than

the Lower Earnings Threshold (LET). This is set at ?11,200 in 2003/4. It is

intended to be roughly half national average earnings and to rise in line with


Everyone earning between the LEL and LET is treated as if they earned the

LET. So they accrue the same amount of S2P. The accrual rate for that element

is 40% (i.e. twice the previous SERPS accrual rate). The accrual rate for

earnings above the LET up to the Higher Earnings Threshold (HET) is only 10%.

The HET is set each year so that the average accrual rate at that point is 20%.

That is ?25,600 in 2003/4. The accrual rate is 20% for earnings above the

Higher Earnings Threshold (HET) and up to the Upper Earnings Level (LEL). The

UEL is the limit beyond which higher earnings do not attract higher


SMF Save Our Pensions Text 16/9/03 12:06 am Page 53

contributory benefits or charges. It is set at ?30,940 in 2003/4.

Entitlement to S2P equals the total of the relevant elements of each year?s

earnings, re-valued and weighted by accrual rates and averaged over a full

working life.

S2P in payment will also be up rated annually in line with prices. It is


Contracting Out of S2P

The arrangements are similar to SERPS. However, those earning less than the

LET can only contract out in respect of their actual earnings. If they do so, they

still accrue an element of S2P in respect of the notional earnings attributed to

them ? that is the difference between their actual earnings and the LET.

Those earning more than the LET can contract out and receive rebates

equivalent to their total S2P entitlement.


SMF Save Our Pensions Text 16/9/03 12:06 am Page 54



On 5th March 1997 the Prime Minister, John Major, and I announced a

revolutionary plan to reform the state pension system called Basic Pension


The Plan involved three key elements:

A Personal Fund - all young people entering the labour market would

be given their own fund.

Rebates - from their National Insurance Contributions would be

invested in their fund sufficient to finance their basic state pension (a

flat rebate of ?9 per week rising with inflation) and a compulsory

second pension (5% of earnings).

Guarantee - the state would guarantee that fund holders would receive

at least their Basic State Pension. If the fund?s performance were

inadequate for any reason, the state would top up their pension from

that fund to equal the BSP level.

We said the plan would bring about the largest extension of personal

ownership of wealth since the spread of home ownership ? and in so doing

resolve one of the major issues facing modern governments ? providing

decent, secure pensions for increasing numbers of elderly people.

Its aims were to guarantee the Basic State Pension; to enable future

pensioners to share in economic growth; to give a massive boost to investment

and ultimately to relieve taxpayers of their biggest burden.

If the extra investment boosted the average growth rate by just one

twentieth of one per cent (e.g. from 2.25% to 2.30% pa) it would generate

sufficient extra tax revenues to be self-financing. In any case the ?double

funding? cost was to be mitigated by changing the timing of tax relief on

saving for the generation covered by the scheme from an up front to a PEPs

basis. Pensions for the generations covered by BPP would therefore be free of



SMF Save Our Pensions Text 16/9/03 12:06 am Page 55




Much has been written about funded pension systems in Chile and other Latin

American countries, Singapore?s centrally operated funded system and new

systems being developed in former communist countries.

But these countries have totally different political, institutional and

economic frameworks from the UK. So any lessons are unlikely to be directly

applicable to us.

A far more relevant experience is that of Australia ? a developed country

that shares our Anglo-Saxon institutions. Moreover, it started from a heavily

means tested universal system of state pensions rather similar to that which is

being created in the UK by the present Labour government.

Australia recognised nearly two decades ago that the disincentives and

resentment inherent in a means tested state system lead logically to

compulsory savings.

Before 1983 Australia had a universal Age Pension funded from taxation and

set at a low level by OECD standards. It was means tested against both

income and assets. So it resembled Income Support or the MIG/Pension Credit

system being developed in the UK.

It acted as a strong disincentive to save for retirement for those on modest

earnings. There were generous tax incentives for occupational and personal

pensions but they were generally sufficient to outweigh the disincentive effect

of means testing the Age Pension only for those on or above average earnings.

Even those who did save for a pension were often tempted to draw their

savings as a lump sum and spend them in order to qualify for the means tested

state Age Pension and related benefits such as health care cards and

discounted transport.

Australia?s current system was phased in over more than a decade under

three successive Labour governments.

The first pillar remains the means tested Age Pension funded out of taxation.

This is set at 25% of Male Total AverageWeekly Earnings. Its value in 2002/03 was


SMF Save Our Pensions Text 16/9/03 12:06 am Page 56

A$11,164 per annum for a single person. (This was equivalent to ?78 per week.)

A single person can have private income from other sources up to A$3,016

and still receive the full rate of Age Pension. Thereafter the Age Pension is

reduced by 40 cents for every extra dollar of private income until it phases out

at A$31,304.

The second pillar is a system of compulsory contributions paid by employers

for all employees. Since July 2002 this contribution has been set at 9% of

earnings. The self-employed are excluded as are most of those earning less

than a Lower Earnings Limit of A$450 per month. There is also an Upper

Earnings Limit of about A$90,000 per annum beyond which the contribution is

not compulsory.

Some 88% of the workforce are within this system as against 40% with

private pensions before it began.

The contributions are paid into each employee?s personal account in a

privately managed retirement fund. Not-for-profit trustee superannuation

funds were established on an industry basis.

The intention is to build up a fund sufficient to buy an annuity worth,

together with the means tested Age Pension, about two thirds of preretirement

income. Table 4 shows projections for different contribution periods

and income levels.

Table 4 Replacement income generated by Australian compulsory superannuation scheme

plus means tested Age Pension

30 years of 40 years of

contributions contributions

Total average salary % of gross % of net % of gross % of net

pre- retirement pre-retirement pre-retirement pre-retirement

A$20,000 70 79 82 90

A$40,000 45 55 58 70

A$60,000 37 48 50 62

Source: Superannuation over two decades: The politics of pension reform in Australia. Senator The Hon Nick Sherry Jan 2003.

The third pillar is voluntary savings. It had been anticipated that voluntary

savings would decline as people assumed that the compulsory level was the

?correct? amount. In fact, the level of additional voluntary contributions into

the superannuation schemes has doubled from 20% to 40%. On the other

hand employers who had contributed more than the compulsory 9% have

tended to reduce to that level.


SMF Save Our Pensions Text 16/9/03 12:06 am Page 57


The cost of administering superannuation schemes in Australia averages 1.2%

of the total amount invested and is declining as this increases.

This is comparatively low given that the system is still young, there is no

cap on costs and there is no central clearing house like the UK National

Insurance system that collects and distributes the NI rebates in Britain.

Moreover, the Australian system has unexpectedly generated a multiplicity of

small accounts as individuals change jobs and industries. On average there are

2.5 accounts for every member. It is proposed that accounts should

automatically be consolidated. This should significantly reduce costs further.

There is no requirement that charges be proportionate to the amount

invested so those on low incomes are not protected from bearing a

disproportionate level of costs in some schemes.


Industry-wide funds are run on a trustee basis with half the trustees appointed

by the employer and half by the unions/employees. To make a decision requires

a two thirds majority.

The ?Prudent Person? rules have been translated from Common Law to

Statute. They require investment managers to invest at arms length, diversify

and match risks with liabilities. Company schemes must reduce the amount

invested in the firm to a maximum of 5% of the total.

There is a compensation fund providing 100% cover for theft or fraud. So

far only one fund has been reported involving A$30m out of a total of A$530


Tax Issues

As compulsion removed the necessity for tax breaks to encourage saving, the

tax regime for pension funds has become less generous. Indeed tax is now

vying with declining administrative costs to become the major cost for pension


A key and, to UK observers, unusual feature of the Australian system has

been that savers have no obligation to convert their pension savings into an

annuity on retirement. They can take 100% as a lump sum and put it to any use

they choose. There is now a tax charge on exit, but savers can still take the

whole lump sum and spend it on anything. The amounts accumulated under

the new compulsory scheme by those reaching retirement are small so far.


SMF Save Our Pensions Text 16/9/03 12:06 am Page 58

About two thirds are taken as a lump sum ? albeit often to repay mortgages or

other debts. Australia may need to change these rules. The clear lesson for the

UK, however, is that the primary call on retirement savings should be to ensure

retired people are not dependent on means tested benefits.

Savings Ratio

After the introduction of compulsory savings in Australia there were at first

signs that the long-term declining trend in household savings was beginning

to come to an end or even reverse. But the downtrend resumed towards the

end of the 1990s. Other Anglo-Saxon countries experienced a similar longterm

decline, worsening in recent years. This may reflect a growing willingness

to borrow as buoyant housing and stock markets increased the value of

people?s assets. In Australia the knowledge that funds accumulated in

compulsory superannuation accounts can be used to repay debts may even

encourage people to borrow during their working lives.

Public Sector Employee Pensions

Apparently public sector pension schemes moved from being defined benefit

to defined contribution, similar to those in the private sector, with remarkably

little fuss.

39 Fuller details are available on my website ? ? under Speeches and Articles

Feb 1999.

40 The average exchange rate during 2002/3 was A$2.754/?.


SMF Save Our Pensions Text 16/9/03 12:06 am Page 59


SMF Save Our Pensions Text 16/9/03 12:06 am Page 60

Go Back



Current Location:


Home / Publications / 2003

Save our Pensions


Search this Site



Contact Peter Lilley MP: