The UK pension system: Key issues - Springer [email protected]. scheme or to a personal or...

elsewhere in Europe and its governmentshave taken measures to prevent apension crisis developing. These measureshave involved making systematic cuts inunfunded state pension provision, andincreasingly transferring the burden ofproviding pensions to the funded privatesector. The UK is not entitled to becomplacent, however, since there remainsome serious and unresolved problemswith the different types of private sectorprovision.

The current system of pensionprovisionA flat-rate first-tier pension is providedby the state and is known as the basicstate pension (BSP). Second-tier or

IntroductionThe UK was one of the first countries inthe world to develop formal privatepension arrangements (beginning in the18th century) and was also one of thefirst to begin the process of reducingsystematically unfunded state provision infavour of funded private provision(beginning in 1980).

This explains why the UK is one ofthe few countries in Europe that is notfacing a serious pensions crisis. Thereasons for this are straightforward: statepensions (both in terms of thereplacement ratio and as a proportion ofaverage earnings) are among the lowestin Europe, the UK has a long-standingfunded private pension sector, itspopulation is ageing less rapidly than

330 Pensions Vol. 8, 4, 330–375 � Henry Stewart Publications 1478-5315 (2003)

The UK pension system:Key issuesReceived (in revised form): 9th April, 2003

David Blakeis Professor of Financial Economics at Birkbeck College in the University of London and Chairman of Square MileConsultants, a training and research consultancy. He is Senior Research Associate, Financial Markets Group, London Schoolof Economics; Senior Consultant, UBS Pensions Research Centre, London School of Economics; Research Associate, Centrefor Risk & Insurance Studies, University of Nottingham Business School; formerly Director of the Securities IndustryProgramme at City University Business School and Research Fellow at both the London Business School and the LondonSchool of Economics. He was a student at the London School of Economics in the 1970s and early 1980s, gaining his PhDon UK pension fund investment behaviour in 1986. In June 1996, he established the Pensions Institute at Birkbeck College(www.pensions-institute.org).

Abstract This paper examines the key issues relating to the UK pension system. Itreviews the current system of pension provision, describes and analyses the reformssince 1980; examines the legal, regulatory, and accounting framework for occupationalpension schemes; assesses the different types of risks and returns from membership ofdefined benefit and defined contribution pension schemes; and investigates themanagement and investment performance of pension fund assets. The paper ends witha discussion of the review of institutional investment in the UK conducted by PaulMyners and published in March 2001.

Keywords: UK pension system; pension reforms; state pensions; private pensions;defined benefit; defined contribution; pension fund management; investmentperformance; Myners review of institutional investment

David BlakePensions Institute,Birkbeck College,University of London,Gresse St,London W1T 1LL,UK.

Tel: �44 (0)20 7631 6410;Fax: �44 (0)20 7631 6403;E-mail: [email protected]

scheme or to a personal or stakeholderpension scheme that has been contractedout of S2P. In such cases both theindividual and the employer contractingout receive a rebate on their NICs (1.6per cent of earnings for the employeeand 3.5 per cent for the employer, unlessit operates a contracted-outmoney-purchase scheme (COMPS), inwhich case the employer rebate is 1.0per cent)6 and the individual foregoes theright to receive a S2P pension. However,there is no obligation on employers tooperate their own pension scheme, nor,since 1988, is there any contractualrequirement for an employee to join theemployer’s scheme if it has one.

There is a wide range of private sectorpension schemes open to individuals.They can join their employer’soccupational pension scheme (if it hasone), which can be any one of thefollowing:

— contracted-in salary-related scheme(CISRS)

— contracted-in money-purchase scheme(CIMPS)

— contracted-out salary-related scheme(COSRS)

— contracted-out money-purchasescheme (COMPS)

— contracted-out mixed-benefit scheme(COMBS)

— contracted-out hybrid scheme(COHS).

A CISRS is a defined benefit (DB)scheme that has not been contracted outof S2P and so provides a salary-relatedpension in addition to the S2P pension.A CIMPS provides a definedcontribution supplement to the S2Ppension. A COSRS must satisfy a‘reference scheme test’ in order tocontract out of S2P, namely provide apension for life from age 65 which isindexed to inflation up to a maximum of

supplementary pensions are provided bythe state, employers and private sectorfinancial institutions, the so-called threepillars of support in old age. The mainchoices are between: a state system thatoffers a pension that is low relative toaverage earnings, but which is fullyindexed to prices after retirement; anoccupational system that offers arelatively high level of pension (partiallyindexed to prices after retirement up to amaximum of 5 per cent p.a.), but, as aresult of poor transfer values betweenschemes on changing jobs, only toworkers who spend most of theirworking lives with the same company;and a personal pension system that offersfully portable (and partially indexed)pensions, but these are based onuncertain investment returns and aresubject to very high set-up andadministration charges, ofteninappropriate sales tactics, and very lowpaid-up values if contributions into theplans lapse prematurely.

Employees in the UK in receipt ofearnings subject to National Insurancecontributions (NICs) will build upentitlement1 both to the BSP2 and, on‘band earnings’ between the lowerearnings limit (LEL) and the upperearnings limit (UEL),3 to the pensionprovided by the State Second PensionScheme (S2P); S2P was introduced inApril 2002, and replaced theState-Earnings-Related Pension Scheme(SERPS), which was introduced in 1978.These pensions are paid by theDepartment of Work and Pensions(DWP) (formerly the Department ofSocial Security (DSS))4 from StatePension Age which is 65 for men and 60for women.5 The self-employed are alsoentitled to a BSP, but not to a S2Ppension. Employees with earnings inexcess of the LEL will automatically bemembers of S2P, unless they belong toan employer’s occupational pension

� Henry Stewart Publications 1478-5315 (2003) Vol. 8, 4, 330-375 Pensions 331

The UK pension system: Key issues

schemes, but with lower unit costsbecause of the savings on up-frontmarketing and administration costs. ASPS is a low-cost PPS with chargescapped at 1 per cent p.a. of the fundvalue and into which contributions of upto £3,600 p.a. can be made irrespectiveof whether the SPS member has madeany net relevant earnings during theyear.7

In 1996, the UK workforce totalled28.5m people, of whom 3.3m wereself-employed.8 The pension arrangementsof these people were as follows:9

— 7.5m employees in SERPS (now S2P)— 1.2m employees in 110,000

contracted-in occupational schemes— 9.3m employees in 40,000

contracted-out occupational schemes(85 per cent of such schemes aresalary-related, although 85 per cent ofnew schemes started in 1998 weremoney purchase or hybrid)

— 5.5m employees in personal pensionschemes

— 1.7m employees without a pensionscheme apart from the BSP

— 1.5m self-employed in personalpension schemes

— 1.8m self-employed without a pensionscheme apart from the BSP.

These figures indicate that 72 per cent ofsupplementary pension scheme membersin 1996 were in SERPS or anoccupational scheme and 28 per centwere in PPSs.10

Table 1 shows the sources ofretirement income in 1997-98. A singleperson had total retirement incomeaveraging 43 per cent of national averageearnings (NAE). Nearly two-thirds ofthis came from state benefits and anotherquarter came from occupational pensions:personal pensions provided only about 5per cent of total retirement income forthe average person.11

5 per cent p.a. where the startingpension is calculated by taking aminimum of 1/80th of the average salaryover the three years prior to retirementfor each year of service in the scheme upto a maximum of 40 years’ service. ACOMPS must have contributions nolower than the contracted-out rebate. ACOMBS can use a mixture of thereference scheme test and the minimumcontributions test to contract out of S2P,while a COHS can provide pensionsusing a combination of salary-related andmoney purchase elements. Individualscan also top up their schemes withadditional voluntary contributions (AVCs)or free-standing additional voluntarycontributions (FSAVCs) up to limitspermitted by the Inland Revenue.

As an alternative, individuals have thefollowing individual pension choices thatare independent of the employer’sscheme:

— personal pension scheme (PPS)— group personal pension scheme

(GPPS)— stakeholder pension scheme (SPS).

A PPS is divided into two components.The first is an appropriate personalpension scheme (APPS) which iscontracted out of S2P and provides‘protected rights’ benefits that stand inplace of S2P benefits: they are alsoknown as minimum contribution orrebate-only schemes since the onlycontributions permitted are the combinedrebate on NICs with the employee’sshare of the rebate grossed up for basicrate tax relief (at 22 per cent). Thesecond is an additional scheme, alsocontracted out, that receives anyadditional contributions up to InlandRevenue limits. A GPPS is a schemethat has been arranged by a smallemployer with only a few employees: itis essentially a collection of individual


Blake

(b) the spouse’s pension was cut from100 per cent of the member’spension to 50 per cent fromOctober 2001 (Social Security Act1986);

(c) the revaluation factor for bandearnings was reduced by about 2per cent p.a. (Pensions Act 1995);the combined effect of all thesechanges was to reduce the value ofSERPS benefits by aroundtwo-thirds.

4 Provided a ‘special bonus’ in the formof an extra 2 per cent NationalInsurance rebate for all PPSscontracting out of SERPS betweenApril 1988 and April 1993 (SocialSecurity Act 1986); provided anincentive between April 1993 andApril 1997 in the form of a 1 per centage-related National Insurance rebateto members of contracted-out PPSsaged 30 or more to discourage themfrom recontracting back into SERPS;age-related National Insurance rebatescontinued in a revised form after April1997 (Social Security Act 1993).

5 Relaxed the restriction on PPSs thatan annuity had to be purchased on theretirement date, by introducing anincome drawdown facility whichenabled an income (of between 35 and100 per cent of a single life annuity)to be drawn from the pension fund

The reforms since 1980

Thatcher-Major reforms to the pensionsystem

The Thatcher Conservative Governmentthat came into power in 1979 becamethe first government in the developedworld to confront head on the potentialcrisis in state pension provision. Thereforms were continued by thesucceeding Major Government. TheseGovernments introduced measures whichdid the following.

1 Linked the growth rate in statepensions to prices rather than NAE,thereby saving about 2 per cent p.a.(Social Security Act 1980).

2 Raised the state pension age from 60to 65 for women over a ten-yearperiod beginning in 2010, therebyreducing the cost of state pensions by£3bn p.a. (Pensions Act 1995).

3 Reduced the benefits accruing underSERPS (which had only been set upin 1978) in a number of ways:

(a) the pension was to be reduced(over a ten-year transitional periodbeginning in April 1999) from 25per cent of average revalued bandearnings over the best 20 years to20 per cent of average revaluedband earnings over the full career(Social Security Act 1986);



Table 1: Sources of retirement income in 1997–1998

Source

Single person Married couples

£per week % of total % of NAE £per week % of total % of NAE

State benefitsa

Occupational pensionsInvestment incomeb

Earningsc

Total

9533147

149

642295

100

271042

43

133904833

304

44301611

100

3826149

87

Notes: aIncludes incapacity benefit, housing benefit, council tax benefit etc; bIncludes income from personalpensions; cWomen in the age range 60-65 and men in the age range 65–70.Source: Department of Social Security (2000), see ref. 9.

Defects in the Thatcher-Major reforms

The main defects of the Thatcher-Majorreforms were as follows.

1 Removing the requirement thatmembership of an occupationalpension scheme could be made acondition of employment. Membershipwas made voluntary and newemployees had to take the activedecision of joining their employer’sscheme: barely more than 50 per centof them did so.

2 No requirement to ensure thattransferring from an occupational to apersonal pension scheme was in thebest interests of the employee, leadingdirectly to the personal pensionsmis-selling scandal that erupted inDecember 1993. Between 1988 and1993, 500,000 members ofoccupational pension schemes hadtransferred their assets to personalpension schemes following highpressure sales tactics by agents of PPSproviders. As many as 90 per cent ofthose who transferred had been giveninappropriate advice. Miners, teachers,nurses and police officers were amongthe main targets of the sales agents.Many of these people remainedworking for the same employer, butthey switched from a goodoccupational pension scheme offeringan index-linked pension into a PPStowards which the employer did notcontribute and which took 25 percent of the transfer value incommissions and administrationcharges. An example reported in thepress concerned a miner whotransferred to a PPS in 1989 andretired in 1994 aged 60. He receiveda lump sum of £2,576 and a pensionof £734 by his new scheme. Had heremained in his occupational scheme,he would have received a lump sumof £5,125 and a pension of £1,791.

(which otherwise remains invested inearning assets) and delaying theobligation to purchase an annuity untilage 75 (Finance Act 1995).

6 Enabled members of occupationalpension schemes to join personalpension schemes (Social Security Act1986).

7 Simplified the arrangements foroccupational schemes to contract outof SERPS by abolishing therequirement for occupational schemesto provide guaranteed minimumpensions (GMPs): since April 1997,COSRSs had to demonstrate only thatthey satisfy the reference scheme test(Pensions Act 1995).

8 Ended its commitment to pay for partof the inflation indexation ofoccupational schemes (Pensions Act1995). Until April 1997, COSRSs hadto index the GMP up to an inflationlevel of 3 per cent p.a. and anyadditional pension above the GMP upto an inflation level of 5 per cent p.a.Since the GMP replaced the SERPSpension which was itself fully indexedto inflation, the Government increasedan individual’s state pension tocompensate for any inflation on theGMP above 3 per cent p.a. But the1995 Act abolished the GMPaltogether and required COSRSs toindex the whole of the pension thatthey pay up to a maximum of 5 percent p.a. (this is known as limitedprice indexation).

9 Improved the security of the assets inprivate sector schemes through thecreation of the Occupational PensionsRegulatory Authority (Opra), acompensation fund operated by thePension Compensation Board (PCB), aminimum funding requirement (MFR)and a Statement of InvestmentPrinciples (SIP) (Pensions Act 1995);Opra, the PCB and the MFR areexamined in more detail below.


Blake

pensions’ (December 1998) turned out tobe much less radical than initiallyanticipated, but nevertheless continuedwith the Thatcher Government’s agendaof attempting to reduce the cost to thestate of public pension provision, and oftransferring the burden of provision to theprivate sector through the introduction ofSPSs.

Nevertheless, there was much greateremphasis on redistributing resources topoorer members of society than was thecase with the Conservatives. Shortly afterthe publication of the Green Paper, theTreasury issued a consultation documenton the type of investment vehicles inwhich stakeholder pension contributionsmight be invested. These proposals willbe examined in turn.

The DSS proposals

The key objectives of the DSS GreenPaper were as follows:

1 Reduce the complexity of the UKpension system, by abolishing SERPS.

2 Introduce a minimum incomeguarantee in retirement linked toincreases in national average earningson the grounds that people who workall their lives should not have to relyon means-tested benefits in retirement;the first-tier BSP will remain indexedto prices, however, and over time willbecome a relatively unimportantcomponent of most people’s pensions.

3 Provide more state help for those whocannot save for retirement, eg thelow-paid (those on less than halfmedian earnings), carers and thedisabled, via the unfunded state system.

4 Encourage those who are able to savewhat they can for retirement, viaaffordable and secure second pillarpensions:

— provided by the state for those onmodest incomes (via a new

As a result of a public outcry, PPSproviders have had to compensatethose who had been giveninappropriate advice to the tune of£13.5bn.

3 No restriction on the charges thatcould be imposed in personal pensionplans, hoping that market forces alonewould ensure that PPSs werecompetitively provided.

4 Giving personal pension schememembers the right to recontract backinto SERPS. This option has turnedout to be extremely expensive for theGovernment because of theback-loading of benefits in DB pensionschemes such as SERPS: benefits accruemore heavily in the later years than theearlier years.12 Despite the financialincentives given to contract out ofSERPS into PPSs, it turned out to beadvantageous for men over 42 andwomen over 34 to contract back intoSERPS once the period of the specialbonus had ended in 1993. Todiscourage this from happening thegovernment has been forced to offeradditional age-related rebates to PPSmembers since 1993. Far from savingthe Government money, the net cost ofPPSs during the first ten years wasestimated by the National Audit Officeto be about £10bn.

The Blair reforms to the pension system

The Blair New Labour Government cameinto power in 1997 with a radical agendafor reforming the welfare state. In theevent, Frank Field, appointed the firstMinister for Welfare Reform at theDepartment of Social Security (DSS) andcharged with the objective of ‘thinkingthe unthinkable’, proved to be too radicalfor the traditional Old Labour wing of theLabour Party and was soon replaced. Theeventual DSS Green Paper proposals ‘Anew contract for welfare: Partnership in



— modernise the system, by abolishingthe weekly means-test, and movingmore into line with the tax systemwhich is based on an annual cycle,thus paving the way for further taxand benefit integration in the future.

SERPS was replaced by a new S2P inApril 2002: the S2P was initiallyearnings-related but from April 2007becomes a flat-rate benefit, even thoughcontributions are earnings-related, afeature that is intended to provide strongincentives for middle- and high-incomeearners to contract out. The S2P:

— ensures that everyone with acomplete work record receivescombined pensions above the MIG

— gives the low paid earning below£9,500 p.a. twice the SERPS pensionat £9,500 p.a. (implying that theaccrual rate is 40 per cent of £9,500rather than the 20 per cent underSERPS)

— gives a higher benefit than SERPSbetween £9,500 and £21,600 p.a.(average earnings in 1999)

— leaves those earning over £21,600p.a. unaffected (with an accrual rate of20 per cent)

— uprates these thresholds in line withnational average earnings

— provides credits for carers (includingparents with children under 5) andthe disabled.

Stakeholder pensions

New SPSs were introduced in April2001, but are principally intended formiddle-income earners(£9,500–£21,600) with no existingprivate pension provision. They can beused to contract out of S2P.

They are collective arrangements,provided by:

— an employer

unfunded state second pension), and— provided by the private sector for

middle- and high-income earners,with the option of new low-costdefined contribution stakeholderpensions which are likely to replacehigh-cost personal pensions. Butthere will be no extra compulsionto save for retirement at the secondpillar and no additional incentivesover those already existing at thesecond pillar.

The Green Paper proposals formed thebasis of the Welfare Reform and PensionsAct which received the Royal Assent inNovember 1999. The Act deals withfollowing issues.

State pensions

A minimum income guarantee (MIG) of£75 per week was introduced forpensioners in April 1999: it ismeans-tested on a weekly basis (andtapers off if the claimant’s capital exceedsa specified limit) and is indexed toearnings. The MIG significantly increasedthe benefit income of the poorestpensioners, creating a new, higherincome threshold below whichpensioners with no or little savingsshould not fall.

In October 2003, the Governmentintroduced the pension credit (PC),13 theaim of which is not just to end thepenalty on savings, but, for the first time,to reward savings. The PC, which isuntaxed, is designed to make up thedifference between the income apensioner receives from all existingsources (including private pensions andsavings) and the MIG. The PC will:

— reward work and savings inretirement, by abolishing the capitallimits and introducing a cash rewardfor modest savings, earnings orsecond-tier pensions;


Blake

occupational DC plans will attract taxrelief on contributions up to a maximumof 17.5 per cent of earnings (below age36), rising to 40 per cent (above age 61).But contributions up to £3,600 p.a. canbe made into any DC plan regardless ofthe size of net relevant earnings.Contributions in excess of £3,600 p.a.may continue for up to 5 years afterrelevant earnings have ceased. Thereafter,contributions may not exceed £3,600p.a. All contributions into DC plans willbe made net of basic rate tax, withproviders recovering the tax from theInland Revenue and with higher ratetax, if any, being recovered in theself-assessment tax return.

Occupational pensions

Occupational schemes can contract outof the S2P. Employers can again makemembership of an occupational scheme acondition of employment, and employeesare only allowed to opt out if they havesigned a statement of rights being givenup, certified that they have adequatealternative provision, and have takenadvice that confirms that the alternativeis at least as good as the S2P.

The compensation scheme establishedby the 1995 Pensions Act was extendedto cover 100 per cent of the liabilities ofpensioners and those within ten years ofnormal pension age (NPA).

Personal pensions

PPSs can contract out of the S2P. Theyreceive protection in cases of thebankruptcy of the member.

HM Treasury proposals

The Treasury proposals were containedin ‘Helping to Deliver StakeholderPensions: Flexibility in PensionInvestment’ (February 1999). They calledfor the introduction of more flexibleinvestment vehicles for managing pensioncontributions, not only those in the new

— a representative or membership oraffinity organisation, or

— a financial services company.

They are defined contribution (DC)schemes, with the same restrictions as forpersonal pensions, namely that on theretirement date up to 25 per cent of theaccumulated fund may be taken as atax-free lump sum, the remaining fundmay be used to buy an annuity or toprovide a pension income by way of adrawdown facility until age 75 when anannuity must be purchased with theremaining assets.

They have to meet minimumstandards, known as CAT marks (forcharges-access-terms) concerning:

— the charging structure and level ofcharges (a maximum of 1 per cent offund value)

— levels of contractual minimumcontributions (£20)

— contribution flexibility andtransferability (no penalties ifcontributions cease temporarily (up tofive years) or if the fund is transferredto another provider).

The main provisions of the Pensions Act1995 apply to SPSs, covering the annualreport and accounts, the appointment ofprofessional advisers and the Statement ofInvestment Principles.

They are regulated principally byOpra, although the selling of schemesand the supervision of their investmentmanagers is regulated by the FinancialServices Authority (FSA), with thePensions Ombudsman for redress.Employers without an occupationalscheme and with at least five staff mustoffer access to one ‘nominated’ SPS andto provide a payroll deduction facility.

There is a new integrated tax regimefor all defined contribution pension plans.SPSs, personal pension plans and



to earnings. Now the Governmentexplicitly rejected this on the grounds ofboth cost14 and the fact that it wouldbenefit the high-paid as well as thelow-paid, whereas the Government’semphasis was on helping the low-paid.But the problem with keeping the BSPlinked to prices rather than to earnings isthat it will continue to fall relentlessly asa proportion of NAE: it is currently just17 per cent of NAE and will fall to wellbelow 10 per cent by 2025. While theGovernment admits that this will savesubstantial sums of money, it implies theGovernment is effectively abandoning thefirst pillar of support in old age andobliging everyone to rely on the secondand third pillars. The Green Paper talkedabout building on the BSP, but thisimplies building on a sinking ship.

If the Government is genuinelyconcerned about security at theminimum level for all, it should considerfunding the first pillar appropriately byestablishing an explicit fund (like theSocial Security Trust Fund (SSTF) in theUSA) into which it places the NICs ofthose who are in work, while theGovernment itself funds the contributionsof the low-paid, carers and thedisabled.15 The contribution rate couldbe actuarially set to deliver the MIG forall when they retire. It could be ahypothecated part of NICs. In otherwords, the contributions would accrue‘interest’ equal to the growth rate inNAE. The state could explicitly issueNAE-indexed bonds which the SSTFwould buy. This is the only honest wayboth of preserving the value of andhonouring the promises under the firstpillar. The second and third pillars couldthen be formally integrated with the firstpillar, ie the second pillar is used todeliver the tranche of pension betweenthe MIG and the Inland Revenue limits,while the third pillar is used forvoluntary arrangements above the Inland

stakeholder pension schemes, but alsothose in occupational and personalpension schemes. These investmentvehicles were given the name individualpension accounts (IPAs). The main IPAsare authorised unit trusts (AUTs oropen-ended mutual funds), investmenttrust companies (ITCs or closed-endedmutual funds), and open-endedinvestment companies (OEICs).

In comparison with the individualarrangements of existing personal pensionschemes and the poor transferability ofoccupational pension schemes, IPAs offer:

— lower charges: since collectiveinvestment vehicles have much loweroverheads than individual investments;

— greater flexibility: since IPAs are easyto value and transfer betweendifferent stakeholder, personal andoccupational pension schemes,allowing employees to move jobswithout having to change pensionschemes, thereby encouraging greaterlabour market flexibility.

Assessment of the Blair reforms

The Welfare Reform and Pensions Act,while containing some significantimprovements on the existing system,does not fully meet the Green Paper’sown objectives.

Reforms to state pensions

While the abolition of SERPS helped tosimplify the UK’s extremely complexpension system, the proposal to have aMIG (of £75 per week) that differedfrom the BSP (£67.50 per week at thetime) reintroduced substantial complexityat the starting point for state pensionprovision, especially when the differencebetween the two amounts (£7.50 perweek) was initially so small. It wouldhave been far simpler to set the MIGequal to the BSP and to link the latter


Blake

per cent p.a. and rose to as much as 2.2per cent p.a. of fund value for 25-yearpolicies18 are much higher than the 1 percent p.a. CAT-marked limit on SPSs.There may be a range of providers ofSPS to begin with, but the only way fora provider to survive in the long runwill be if it operates at low unit cost ona large scale. This will inevitably lead tomergers among providers and a finalequilibrium with a small number of verylarge providers.

Existing personal pension providersand distribution channels face thesechallenges:

— APPSs face massive competition fromSPSs for future NIC rebates

— SPSs could be better than PPSs formiddle-income groups, leaving PPSsas a choice only for those on highincomes who require and are willingto pay for a bespoke product

— new affinity-based SPSs with gatewayorganisations linking up with pensionproviders (eg AmalgamatedEngineering & Electrical Union with720,000 members and FriendsProvident)

— the Treasury’s proposed PPIs providea low-cost alternative investmentvehicle to the high-cost managedfunds of most PPSs

— Individual savings accounts (ISAs),introduced by the Treasury in April1999 to encourage greater personalsector savings, also provide animportant alternative to PPSs.Contributions into ISAs of up to£5,000 per annum are permitted, andthe investment returns are free fromincome and capital gains tax. Whilenot intended as pension savingsvehicles (they do not attract tax reliefon contributions, for example, unlikestandard pension savings products),ISAs can be used in retirementincome planning, since they enjoy the

Revenue limits. If the first pillar remainsunfunded, there is nothing to preventfuture generations reneging on anagreement which they are expected tokeep but did not voluntarily enter into.

The fact that membership of pensionschemes at the second pillar remainsvoluntary, is highly worrying for reasonsof myopia and moral hazard.Compulsory contributions are seen asone way of dealing with individualmyopia and the problem of moralhazard. Myopia arises because individualsdo not recognise the need to makeadequate provision for retirement whenthey are young, but regret this whenthey are old, by which time it is too lateto do anything about it. Moral hazardarises when individuals deliberately avoidsaving for retirement when they areyoung because they calculate that thestate will feel obliged not to let themlive in dire poverty in retirement.Inevitably, this will lead to substantialmeans testing in retirement.

In short, while the Welfare Reformand Pensions Act has some good points,it fails three tests set by Frank Field for agood state pension system: it is notmandatory, it is not funded and itremains means-tested.16,17

Reforms to private pensions

The Government’s proposal to have amaximum charge of 1 per cent of fundvalue on SPSs will have two dramaticeffects on private sector pensionprovision, especially PPSs.

The first is that it will help to forceeconomies of scale in DC pensionprovision. This is because stakeholderpensions will be a retail product withwholesale charges. To deliver thisproduct effectively, providers will need toexploit massive economies of scale.Charges for personal pension schemeswhich prior to the introduction ofstakeholder pension schemes averaged 1.4



SERPS was an incredibly complexpension system that very few pensionsprofessionals fully understood, let alonemembers of the general public. Whilethere was comment in the media at thetime of these changes to SERPS, verylittle of this seems to have permeated theconsciousness of the mass of thepopulation and the extent of the changeswas little understood. Thirdly, thechanges were introduced with a time lagof 15 to 20 years, so it was easy foreveryone to forget about them.

Even when changes were introducedimmediately, such as the switch in theuprating of the state pension fromearnings to prices, the immediatedifference was relatively small and mostpeople failed to realise how smalldifferences can compound into largeamounts over time.19

A final explanation lies in the fact thatstate pension provision is much lessimportant for most people in the UKthan on the continent, and those forwhom it is important, namely thelow-paid, have little political influence.

The situation on the continent israther different. State pensions provide amuch higher replacement ratio than inthe UK and social solidarity appears tobe a more important objective than it isin the UK. As a consequence, it is muchharder to alter pension arrangements onthe continent, even if the political willto do so is strong, which it clearly isnot.

The legal structure of andregulatory framework foroccupational pension schemes

The trust fund

Most occupational pension schemes inthe UK have been set up as pensiontrust funds. A trust is a legal relationshipbetween individuals and assets, by which

big advantage that they can be cashedin tax free at any time, therebyavoiding the need to purchase apension annuity on the retirementdate.

The second benefit is that it willeffectively force stakeholder pensionfunds to be passively managed, sinceactive management would result in acharge higher than 1 per cent. Asdemonstrated below, active fundmanagers have not demonstrated thatthey can systematically deliver thesuperior investment performance thatjustifies their higher charges. Furtherpassively-managed mutual funds in theUSA, such as Vanguard (which aresimilar investment vehicles to PPIs), havecharges below 0.3 per cent.

The political economy of pension reform

How has it been possible for UKGovernments to reduce the size of statepension provision without significantpolitical protest when similar attempts todo so on the continent have led to streetprotests and strikes (eg in Italy inNovember 1994 and France inNovember 1995)?

Consider the SERPS pension. Whenit was first introduced in 1978, it offereda pension of 25 per cent of the best 20years of band earnings revalued to theretirement date by increases in nationalaverage earnings, with a 100 per centspouse’s pension. Within a quarter of acentury, the value of these benefits hadbeen reduced by two thirds before thescheme was abandoned altogether. Howhas this been achieved so peaceably?There are three main explanations. First,SERPS had only been established a fewyears before changes to it started beingmade, so very few people were drawingthe pension and little loyalty for thescheme had accumulated. Secondly,


Blake

pension scheme must be establishedunder irrevocable trust, with theemployee being a beneficiary under thetrust, and the employer being acontributor. However, the word‘irrevocable’ is not crucial, since the trustdeed can provide for the alteration andwinding-up of the scheme. But the solepurpose of the scheme must be toprovide ‘relevant benefits’ in respect ofservice as an employee, where benefitsare defined as pensions and lump sumspayable on or in anticipation ofretirement or on death. The benefitsmust be made available to the memberor widow/er, children, or dependants.Most trusts have limitations on theirdurations under the 1963 Perpetuitiesand Accumulations Act. However,occupational pension schemes areexempted from these limitations wherethey have received exempt approval fromthe Inland Revenue (under section 163of the 1993 Pensions Schemes Act).

The pension scheme must also appointan administrator to manage the scheme.Under the 1970 Finance Act, theadministrator must be a resident of theUK. Typically the trustees, so long asthey are resident in the UK, areappointed as administrator to the scheme.

The Occupational Pensions RegulatoryAuthority

The 1995 Pensions Act established theOccupational Pensions RegulatoryAuthority (Opra) as the regulatoryauthority for the pensions industry. It isfinanced by an annual levy on pensionschemes. Opra took over most of theresponsibilities of the OccupationalPensions Board (OPB) which had beenset up under the 1973 Social SecurityAct to monitor scheme rules on thepreservation of benefits for early leavers,equal access and contracting-outrequirements. The 1995 Act transferred

assets provided by one individual (thesettlor) are held by another group ofindividuals (trustees) for the benefit of athird group of individuals (thebeneficiaries). The interests of thebeneficiaries are set out in the trust deed.If the trust is a discretionary trust, thetrustees have the freedom of action todispose the income and capital of thetrust as they see fit. The trust servesthree functions: it is the primary sourceof payment of pension entitlements; it isa security for payment; and it is a vehiclefor the collective protection andenforcement of the rights of individualscheme members. The first scheme toadopt this legal vehicle was that ofColmans, the mustard manufacturer, in1900.

There are several reasons why a trustfund came to be preferred to a statutoryfund, its main alternative. A trust fundwas much cheaper to set up than astatutory fund. It was also much moreflexible: the trust deed could be drawnup in virtually any way that suited theemployer, and the employer could ensureeffective control of the fund through hisappointment of the trustees.Nevertheless, a trust is also a usefulvehicle for protecting pension benefits.This is because a trust is a means ofattaching to assets the interests of a wideclass of beneficiaries, including those notyet born. The presence of a trust alsoseparates the assets of the trust fromthose of the employer, a valuable featurein the case of default.

Since trust law had not originally beenestablished to validate pension schemes, itsoon became necessary to put thearrangements on a formal basis. This wasdone in the Superannuation and OtherTrust Funds (Validation) Act of 1927,which permitted the formal validation oftrust funds.

In order to receive exempt approvedstatus from the Inland Revenue, a



— apply for a court injunction toprevent the misuse ormisappropriation of scheme assets;

— apply for a court order requiring therestitution of scheme assets where it issatisfied that they have beenmisappropriated, eg where schemeassets have been loaned to theemployer in contravention of legalrequirements;

— direct trustees to pay members’benefits, eg where an employerdeducts pension contributions fromearnings but does not pass them on tothe scheme;

— require the production of anydocument relating to a particularpension scheme from a trustee,manager, professional adviser oremployer;

— enter premises where schememembers are employed, or wheredocuments relating to schememembers are kept, or where theadministration of a pension scheme iscarried out, and to question anyperson on those premises who may beable to provide relevant information.

The Pensions Act requires every pensionscheme to appoint an auditor and anactuary. The Act imposes a specificobligation on the auditor and actuary toreport to Opra if they have ‘reasonablecause’ to believe that there has been abreach of duty relevant to the scheme’sadministration by the employer, trustees,administrator or a professional adviser.The auditor and actuary are protectedfrom any claim of breach of privilege ifthey ‘blow the whistle’ but face civilpenalties or even disqualification if theyfail to meet these requirements. The Actalso requires the appointment of aprofessional fund manager where ascheme has investments regulated by the1986 Financial Services Act. The auditor,actuary and fund manager are classified as

the contracting-out arrangements to theNational Insurance Contributions Office(NICO) of the Inland Revenue anddisbanded the OPB.

Opra has extensive powers, includingthe power to:

— remove or suspend a trustee wherethere has been a ‘serious or persistentbreach’ of his or her duties, whereproceedings have been commencedagainst him or her for an offenceinvolving dishonesty or deception,where a bankruptcy petition has beenpresented against him or her, orwhere an application has been madeto disqualify him/her as a companydirector;

— appoint a new trustee if an existingtrustee has been removed ordisqualified under the Pensions Act,or in order to secure ‘the properadministration of the scheme’ or ‘theproper use or application of the assetsof the scheme’;

— wind-up schemes if it is satisfied thatthe scheme ought to be replaced by adifferent scheme, that the scheme isno longer required, or that awinding-up is necessary to protect theinterests of the generality of thescheme members;

— modify schemes to enable a schemeto reduce or eliminate a statutorysurplus, to enable surplus assets to bedistributed to the employer in thecase where a scheme is beingwound-up, or to enable a scheme tobe contracted out during a prescribedperiod;

— impose civil penalties for misconduct,eg making a payment to theemployer from the scheme assetscontrary to s. 37, or failure to obtainan actuarial valuation and certificate inaccordance with s. 57, or failure tomaintain a payment schedule or makea statement of investment principles;


Blake

beneficiaries and to act impartiallybetween the interests of different classesof beneficiaries. They have to act inaccordance with the trust deed and rulesof the scheme, within the framework oftrust law and the statutory regulations ofOpra. They also have to act prudently,conscientiously, honestly, and with theutmost good faith. But there are nospecific rules on the number of trusteesor for the scheme to have anindependent trustee. There is also norequirement for a trustee to have anyspecial training or to meet anyprofessional standard. It is possible for thetrustee to be a limited company: indeedmore than half of existing pensionschemes have corporate trustees (althoughmost of these are not independent of theemployer).

Trustees have a fiduciary duty underthe 1961 Trustee Investments Act topreserve the trust capital and to apply thecapital and its income according to thetrust deed. This means that trustees areultimately responsible for the safe custodyof scheme assets and for ensuring thatthe benefits provided under scheme rulesare duly delivered to scheme members.Trustees generally have wide investmentpowers, including powers to borrow.Indeed, the failure to invest, or at leastplace funds on deposit, might maketrustees liable to make up the lostincome. Scheme members can sue forcompensation if they suffer loss as aresult of negligence by trustees under the1925 Trustee Act.

Trustees (and their investment advisers)also have to abide by the FinancialServices Act 1986. It is a criminaloffence to carry on investment businessin the UK unless either authorised to doso or exempted from the provisions ofthe Act. Pension fund managers areregulated by the Financial ServicesAuthority (FSA) (formerly IMRO — theInvestment Managers Regulatory

‘professional advisers’ under the PensionsAct and will have to be members of arecognised professional body. There is norequirement under the Act to appoint alegal adviser. However, legal advisers to apension scheme are exempt from therequirement to blow the whistle on thetraditional grounds of legal professionalprivilege, except where there is reason tobelieve that the pension scheme is beingused for money laundering purposes.

The Act also established the PensionsCompensation Board (PCB) to administera compensation scheme. Certainconditions will need to be met before thecompensation provisions apply:

— the scheme must be established undertrust

— the employer must be insolvent— the value of the scheme assets must

have been reduced as a result of anillegal act and, in the case of asalary-related scheme, to less than 90per cent of the value of the liabilities.

The amount of any compensation isdetermined by regulations, but will notexceed 90 per cent of the loss at theapplication date and, in the case of asalary-related scheme, will be limited towhatever is necessary to restore thescheme to a 90 per cent solvency level.The PCB has power to make drip-feedpayments to a defrauded scheme in anemergency situation where it accepts thatthere are grounds for compensation andwhere the trustees would not otherwisebe able to make pension payments. Thescheme will be financed by a special levyon occupational pension schemesimposed after the compensatable eventhas taken place.

Trustees

The role of the trustees is to operate thepension scheme in the best interests of its



company director. The assets of thescheme cannot be used to pay a fineimposed on a trustee; previously schemeassets could be used to indemnify atrustee who inadvertently committed abreach of trust.

Thirdly, the Act allows for theappointment of member-nominatedtrustees (MNTs) unless schememembers have specifically voted againstthis. Once appointed, MNTs can onlybe removed with the agreement of allother trustees; previously, the employerhad the exclusive power to appointand remove trustees. Where theappointment of MNTs has beenapproved, the scheme’s membership isentitled to elect one-third of the totalnumber of trustees, with a minimumof two MNTs for large schemes andone in schemes with fewer than 100members. The MNT does not have tobe a scheme member, although theemployer has the power to block theappointment of a non-scheme memberof whom he does not approve;however, the employer cannot blockthe appointment of a scheme memberwho has been elected. A schememember includes any ‘active, deferred,or pensioner member’ of the scheme.MNTs have the same fiduciaryresponsibilities as other trustees to actin the best interests of all members.

Fourthly, the Act imposes on trustees aduty of care to invest the assets of thefund in an orderly and correct manner.Section 33 prohibits the trust deed orrules from restricting a trustee’s liabilitywhere that duty of care is not properlyobserved. Section 34 gives pensiontrustees a general power to makeinvestments ‘of any kind as if they wereabsolutely entitled to the assets of thescheme’. Trustees are, however,permitted to delegate authority to a fundmanager. In the case of a discretionaryfund manager, with powers to make

Organisation). Under s. 19 of the Act,trustees who are not involved in dailyinvestment decision-taking for theirschemes do not have to be regulatedunder the Act. However, the FSA makesregular inspection visits to occupationalpension schemes. In general, it finds thatmost schemes are well-run, althoughsome schemes have been criticised forinadequate record-keeping and failing toensure that administrative staff areproperly trained.

Trustees have substantial discretionover who benefits in the event of amember dying, especially if the memberwas unmarried or had no one who wasfinancially dependent on him or her. If,for example, a man was married (even ifthe wife was financially independent) orhad parents, or children up to the age of18 who were financially dependent onhim, the case would be clear-cut: theywould receive a widow’s or dependant’spension. If the man had a financiallyindependent common-law wife, the caseis also clear-cut: the common-law wifewould not receive a widow’s pension. If,however, the common-law wife wasfinancially dependent, she might receivea pension at the trustees’ discretion.

There is no restriction on whoreceives the tax-free lump sum in theevent of death in service. It can go towhoever is nominated by the member. Ifno one is nominated and there are nodependent relatives, it will go into themember’s estate and be taxed.

The 1995 Pensions Act had a majorimpact on trustees. First, it placed alltrustees under the supervision of Opra.Secondly, it specified who could notserve as a trustee. For example, a schemeauditor or actuary cannot serve as atrustee. Neither can ‘unsuitable persons’such as anyone who has been convictedof an offence involving dishonesty ordeception, an undischarged bankrupt, orany person disqualified from acting as a


Blake

one-size-fits-all structure and in March2001, the government announced that itwould replace the MFR with a fundingstandard that was scheme-specific. Eachof these funding standards will beconsidered in turn.

The MFR

The MFR came into effect in April1997. It applied to all occupationalschemes except occupational moneypurchase schemes, public service schemesestablished by statute, local governmentschemes, schemes with a governmentguarantee and unapproved schemes. Itobliged schemes subject to MFR toensure that ‘the value of the assets of thescheme are not less than the amount ofthe liabilities of the scheme’. Theprocedure for doing this were set out inthe Occupational Pension Schemes(Minimum Funding Requirement andActuarial Valuations Regulations) 1996(SI 1996/1536) and Guidance Note 27from the Institute and Faculty ofActuaries.

Trustees subject to MFR were obligedto prepare and maintain a schedule ofcontributions indicating the rate of bothemployer and member contributions andthe dates on which the contributions hadto be paid into the fund. The actuaryhad to give an opinion on whether thecontributions were ‘adequate for thepurpose of securing that the minimumfunding requirement will continue to bemet throughout the prescribed period[the next five years] or, if it appears tohim that it is not met, will be met bythe end of that period’.

A ‘serious underprovision’ arose in thecase where the scheme’s assets were lessthan 90 per cent of its liabilities. Theemployer was required to make up thedifference to 90 per cent through a cashinjection or other means such as a bankletter of credit for the amount of shortfalland for the life of the schedule of

day-to-day investment decisions, thetrustees will not normally be heldresponsible for any act or default of thefund manager, as long as they have takenall reasonable steps to satisfy themselvesthat the manager ‘has the appropriateknowledge and experience for managingthe investments of the scheme’. In thecase of other types of fund manager, thetrustees will normally be held responsiblefor any act or default by them.

Fifthly, the Act requires trustees toprepare and maintain a Statement ofInvestment Principles (SIP). The SIPspecifies the strategic objectives of thepension fund and must cover thefollowing issues:

— the kinds of investments held— the balance between different kinds of

investments— risk and the need for the

diversification of investments— expected return on investments— the realisation of investments.

In preparing the SIP, the trustees arerequired to take written advice from ‘aperson who is reasonably believed by thetrustees to be qualified by his ability inand practical experience of financialmatters and to have the appropriateknowledge and experience of themanagement of the investment of suchschemes’.

Finally, trustees are required tointroduce arrangements for resolvinginternal disputes between schemeadministrators and scheme members.

The minimum funding requirement andscheme-specific funding standard

The minimum funding requirement(MFR) was introduced by the 1995Pensions Act in response to the Maxwellscandal.20 However, there was almostimmediate criticism of its inflexible,



proposals were recommended by the2001 Myners review of institutionalinvestment.20 The new standard isexpected to come into force in 2005.

Trustees and their advisers will berequired to take a view on the appropriatefunding and investment of the scheme inthe light of the scheme’s specificcircumstances and those of the sponsoringemployer. Associated with this will be astrong regime of transparency anddisclosure. The trustees and advisers ofeach scheme will be required to publish ascheme-specific funding statement which‘sets out in a clear and straightforwardway how it sees its liabilities growing overtime and how, through contributions tothe fund and growth in the value of theassets through investment returns, itproposes to meet its liabilities’. Thefunding statement will specify:

— the funding objectives for the scheme— the fund’s investment policy and

projected returns on its assets— assumptions for projecting its

liabilities, including the range ofeconomic scenarios considered

— a contribution schedule agreed by thetrustees and the employer.

The statement will be drawn upassuming that the employer will continuein existence, that is, on a long-term orongoing basis, and the trustees will berequired to assess and report on thestrength of the employer’s covenant. Theemployer is therefore expected to befully involved in discussions aboutfunding and investment plans, and inagreeing the required contribution rates.Interested parties, such as schememembers, their representatives (such astrade unions and pensioner supportgroups), and the company’s shareholders,can scrutinise the scheme’s funding andinvestment plans and assess whether theyare realistic and appropriate.

contributions. The employer had torestore the scheme’s solvency level to 90per cent within three years of the seriousunderprovision being discovered. Wherethe solvency level was between 90 and100 per cent, it had to be restored to100 per cent and so satisfy the MFR bythe end of the current schedule ofpayments, namely five years.

A number of problems emerged withthe MFR:

— it did not guarantee that the pensionwould be paid in full: a pension fundthat fully meets the MFR might onlyhave funds sufficient to purchasearound 70 per cent of the pensionsdue to active members if the sponsorbecomes insolvent, mainly because theclaims of retired members are metfirst

— it was highly sensitive to the way inwhich pension liabilities were valued

— it restricted pension funds frominvesting in an optimal mix of assets,by encouraging pension fundmanagers to reduce their weighting in‘volatile’ asset categories such asequities.

The Faculty and Institute of Actuaries, inits 2000 publication ‘Review of theMinimum Funding Requirement’,concluded that the MFR ‘cannot bemade to work as a statutory standard’. Itaccepted that there was an ‘inherentconflict between the MFR whichimposes a risk of short-term fluctuationsin funding requirements and thelong-term asset allocation to produce thebest financial results for pension fundmembers’.

A scheme-specific funding standard

In March 2001, the Governmentannounced that it would replace theone-size-fits-all MFR with a long-term,scheme-specific funding standard.21 These


Blake

The Government argues that: ‘Theseproposals will provide protection formembers of all defined benefit schemesand will encourage an intelligent andthought-through approach to planninginvestment and contributions policy.They do not distort investment as theMFR does, because they do not involvethe valuation of liabilities using statutoryreference assets which create artificialincentives for schemes to invest in thoseassets. Employers that wish to go onoffering defined benefit schemes will findit easier to do so under these proposals.At the same time, the proposals willmake it more difficult for those that wishto walk away from the pension promisesthat they have made.’

The accounting framework foroccupational pension schemes

Financial Reporting Statement 17

In November 2000, the AccountingStandards Board (ASB) issued a newFinancial Reporting Standard (‘FinancialReporting Standard 17 — RetirementBenefits’) with the objective of replacingSSAP24, the existing accounting standardfor reporting pension costs in DBpension schemes. The principal changesare that:

— actuarial gains and losses will berecognised fully and immediately(rather than amortised over a periodof up to 15 years)

— scheme assets and liabilities will bevalued by reference to current marketconditions.

The consequence of this could be greatervolatility of pension costs year on yearand greater volatility in the balancesheet.

Prior to the introduction of SSAP24(Accounting for Pension Costs) in 1988,

Each scheme will have to compareitself against the funding statement ona regular basis. If the scheme finds thatit is not adequately funded then it willhave to produce a recovery plan forreturning the fund to full fundingwithin three years. The key objectiveis to ensure that the scheme is fundedto meet the benefits in full in the longterm. The scheme will be required tofile the recovery plan with Opra andreport annually on progress against it.Opra will have some discretion toallow extensions to the deadline forreturning funding to an adequate level,in the light of the specificcircumstances of the scheme.

The trustees, actuaries and auditorswill have whistleblowing duties to reportto Opra if contributions are not paid inaccordance with the recovery plan. Inparticular, the scheme actuary will have astatutory duty of care towards schememembers. This will be particularlyimportant for smaller funds where theremay not be people or organisations withthe required skill or interest to exerciseeffective scrutiny of the scheme’s fundingstatement. The actuary will have anexplicit duty to consider the implicationsof funding plans for the scheme membersand beneficiaries. The actuary will have aduty to report to Opra if contributionsare not being paid according to thefunding statement; if there are any delaysin drawing up a recovery plan in ascheme that is underfunded; and ifcontributions to an underfunded schemeare not being paid in line with therecovery plan.

There will also be an extension of thefraud compensation scheme. The level ofcompensation for fraud will be increasedto cover not simply the MFR liabilitiesas at present, but the full cost of securingmembers’ accrued benefits (or theamount of the loss from fraud,whichever is the lesser).



of gains and losses in the statement ofrecognised gains and losses, not in theP&L

— the financial statements containadequate disclosures.

FRS17 will have the following effectswhen it is fully in force for year-endsafter 2005.

Scheme assets

Scheme assets will be included at theirfair value on the company’s balance sheetdate. This, in turn, will require anannual update of the scheme’s actuarialvaluation. The expected return onscheme assets will be calculated as theproduct of the expected long term rateof return and the market value (at thestart of the period).

Actuarial liability

The actuarial liability will be calculatedusing the projected unit method and anAA corporate bond discount rate,although the actual discount rate usedcan be based on gilt yields with aconstant risk premium of, say, 1 per cent.This rate will generally be lower thanthat used under SSAP24 which is basedon the assumed returns on the pensionfund assets and so includes an equitycomponent. The discount rate should beof equivalent currency and term as thescheme liability; however, the ASBargues that ‘in theory, different discountrates should be applied to cash flowsarising in different periods, reflecting theterm structure of interest rates. Inpractice, acceptable results may beachieved by discounting all the cashflows at a single weighted averagediscount rate’.23

The AA corporate bond yield waschosen because this was the yield used inthe equivalent US accounting standard,FAS87. FAS87 adopted this particularyield because it matched the asset class

employers accounted for pension schemeson a cash basis. Under SSAP24, theprofit and loss account is charged with‘regular pension cost’ which is designedto be a stable proportion of pensionablepay. Any variations from regular cost arespread forward and charged to profit andloss (P&L) gradually over the averageremaining service lives of the employees.Assets and liabilities are reported atactuarial value rather than fair value.

A number of problems emerged withSSAP24:

— too much flexibility in choosing thevaluation method and in accountingfor the resulting gains and losses

— inadequate disclosure requirementsand lack of transparency

— inconsistency between the pensionassets and liabilities in the company’sbalance sheet and the actual surplus ordeficit in the scheme

— inconsistent with internationalaccounting standards (eg FAS87(Employers’ Accounting for Pensions)and IAS19 (Accounting forRetirement Benefits in the FinancialStatements of Employers)) which hadmoved towards a market basis forvaluing scheme assets.

The objectives of FRS17 are to ensurethat:

— the employer’s financial statementsreflect the assets and liabilities arisingfrom retirement benefit obligationsand any related funding, measured atfair value

— the operating costs of providingretirement benefits are recognised inthe periods the benefits are earned byemployees

— financing costs and any other changesin the value of the assets and liabilitiesare recognised in the periods they arise

— there will be immediate recognition


Blake

most cases, the vesting of suchimprovements is immediate, so the costis charged immediately to the P&Laccount without offset against the surpluseven if it is funded from a surplus.

Profit and loss account

The P&L charge will be split between:

— operating costs: which includescurrent service costs and past servicecosts;

— financing costs: which includesinterest costs (the pension liabilitydiscount) and the expected return onassets.

Any overpaid/unpaid contributions arerepresented as debtor/creditor due withinone year.

Actuarial gains and losses

SSAP24 and IAS19 allow differencesbetween actual and expected outcomesto be spread in the P&L over a numberof years and to defer a hard core (the 10per cent corridor) indefinitely.

FRS17, in a radical departure fromconventional practice, requires immediaterecognition of actuarial gains and lossesthrough a new account, the statement ofrecognised gains and losses (STRGL).The asset returns in the pension fund aredivided into two parts which arerecognised separately in the P&L andSTRGL. The financing item in the P&Lwill show an expected asset return,which is designed to be reasonably stableover time. The differences betweenrealised and expected asset returns areshown in the STRGL, as are changes inactuarial assumptions and differencesbetween these assumptions and actualexperience in respect of the liabilities. Afive-year history of these differences isrequired to enable users of the accountsto assess the accuracy of the forecastreturns.

that a US insurance company, taking onthe liabilities of an insolvent pensionplan, would use to invest the scheme’sremaining assets. The same yield wassubsequently adopted by the InternationalAccounting Standards Committee inIAS19.

At the end of each accounting year, apension scheme member will have earnedan additional year of service: this currentservice cost is classified as an operatingcost in FRS17. Also by the end of theyear, the member’s pension liability willhave risen because it is one year closer tobeing delivered (this is denoted theinterest cost or pension liability discount),but this will be offset by the expectedreturn generated on the assets backing theliability: the difference is denoted the netfinancing cost in FRS17.

The current service cost will be higherthan the regular cost under SSAP24. Onthe other hand, under FRS17 thediscount rate (and hence the interest costrelating to the liability) is likely to belower than the expected return onscheme assets, so that the net financingcost for the pension scheme is likely tobe a credit.

Surplus or deficit

The net defined benefit pension asset orliability, after attributable deferred tax,will be shown after other net assets inthe balance sheet. FRS17 limits thesurplus recognised by the employer tothe amount that the employer couldrecover through reduced contributionsand agreed refunds.

Past service costs

Past service costs arise whenever animprovement in benefits is backdated (egthe award of a spouse’s pension). UnderSSAP24, they may be set against anysurplus, with any excess cost charged tothe P&L. With FRS17, they are chargedto P&L over the period of vesting. In



— the balance sheet shows the deficit orrecoverable surplus in the scheme;

— the total profit and loss charge ismore stable than it would be if themarket value fluctuations were spreadforward.

The Association of Chartered CertifiedAccountants (ACCA) argued that thespreading forward of gains/losses overaverage service lives is better thanimmediate recognition because of thelong-term nature of pension costs, theuncertainty over the estimates of keyyields, and the conformity with currentinternational standards (eg IAS19).Although the various components mightbe separately disclosed, the ACCApreferred the pension cost to be chargedas a single item in operating cost.

The FIA argued that, while FRS17will make ‘the respective risks andrewards borne by companies andshareholders more transparent to theshareholders’, there would be ‘adverseimpacts on pension scheme members,because it will introduce new volatilityinto the assessment of pension costs andliabilities’. As a consequence, sponsors ofDB schemes could become morereluctant to improve benefits since thesewould be immediately reflected incompany P&L, even if funded fromsurplus assets.

The long-term effect of FRS17 onasset allocation is not clear. On the onehand, as in the case of the MFR, the useof a specific discount rate for liabilities(such an AA corporate bond yield) mightinduce funds to adopt a morebond-based investment strategy. On theother hand, by excluding the impact ofequity risk on the P&L, FRS17 providescompanies with an incentive to raise theequity component of their pension fundin order to generate higher expectedasset return and profit figures. However,anecdotal evidence suggests that pension

Assessing FRS17

FRS17 will have three major impacts.

— It will reduce the volatility of theP&L but cannot eliminate it, sincechanges in realised market rateseventually flow through to the P&Lvia consequential changes in thelong-term expected returns on bothassets and AA corporate bonds.

— It will increase the volatility of thebalance sheet due to the inclusion ofthe net pension asset or liability andthis may trigger loan covenants orborrowing limits.

— There will be increased complexity ofthe financial statements arising fromnon-cash pension items, eg currentservice cost and amortisation of pastservice costs within operating cost,and the unwinding of the pensionliability discount and the expectedreturn on assets within financing costs.

International accounting standards dealwith this volatility by averaging themarket values over a number of yearsand/or spreading the gains and lossesforward in the accounts over theremaining service lives of the employees.But the consequences are that thebalance sheet does not represent thecurrent surplus or deficit in the schemeand that charges to P&L are infected bygains and losses that arose many yearspreviously.

With FRS17, the P&L shows therelatively stable ongoing service cost,interest cost and expected returns onassets measured on a basis consistent withinternational standards. The effects of thefluctuations in market values, on theother hand, are not part of the operatingresults of the business and are treated inthe same way as revaluations of fixedassets, ie are recognised immediately inthe STRGL. This has two advantagesover the international approach:


Blake

pension benefit that is defined. In theUK, for example, most DB schemes arearranged by companies and are known asoccupational final salary schemes, sincethe pension is some proportion of finalsalary, where the proportion depends onyears of service in the scheme. A typicalscheme in the UK has a benefit formulaof one-sixtieth of final salary for eachyear of service up to a maximum of 40years’ service, implying a maximumpension in retirement of two-thirds offinal salary, and with the pension indexedto inflation up to a maximum of 5 percent per annum (ie limited priceindexation). In contrast with a DCscheme, what is defined is thecontribution rate into the fund, eg 10per cent of earnings. The resultingpension depends solely on the size of thefund accumulated at retirement. Suchschemes are also known as moneypurchase schemes. The accumulated fundmust be used to buy a life annuity froman insurance company (although in theUK, up to 25 per cent of the fund canbe taken as a tax-free lump sum on theretirement date).

DB schemes

DB and DC schemes have different costsand benefits. DB schemes offer anassured (and in many cases a relativelyhigh) income replacement ratio inretirement. People in retirement canexpect to enjoy a standard of living thatis related to their standard of living justprior to retirement. But this is the caseonly for workers who remain with thesame employer for their whole career.Fewer than 5 per cent of workers in theUK do this: the average worker changesjobs about six times in a lifetime.25

Every time workers switch jobs theyexperience a ‘portability loss’ in respectof their pension entitlement. This isbecause DB schemes are generally

funds are increasing rather than reducingtheir weighting in bonds in preparationfor the introduction of FRS17.

Other objections have been putforward:

— the P&L depends on an assessed orexpected figure for asset returns;

— there are potentially two differentvaluation results, the trustees’ fundingvaluation and the company’saccounting valuation; companiesprefer to align the two types ofvaluation, if possible using the weakerfunding basis, thereby reducing thesecurity of benefits;

— despite the greater transparency fromusing market values, there can besubstantially different investmentconditions if companies use differentmeasurement dates, even if these datesare only a short time apart;

— a pension scheme deficit has to bededucted from distributable reserves,thereby lowering dividend cover andpossibly forcing a company to pass adividend payment. Somecommentators have suggested that thisis what should happen if companiesmake a pension promise and do nothave the resources to cover it;

— the use of the projected unit methodto determine pension liabilities isinconsistent with the MFR, eventhough it gives a more realisticmeasure of the true eventual liability;

— unlike the USA, AA bonds are not asignificant investment category in theUK: their weighting was just 7 percent of the total UK bond market inDecember 2000.

The risks and returns in fundedschemesThere are two main types of fundedscheme: the DB scheme and the DCscheme.24 With a DB scheme, it is the



leaver valued in terms of their projectedsalary at retirement which is likely to behigher. Long stayers are thereforesubsidised at the expense of early leavers.In the UK, the portability loss is morecommonly known as a ‘cash equivalentloss’.

For a typical worker in the UKchanging jobs six times during theircareer, Table 2 shows that the portabilityloss lies between 25 and 30 per cent ofthe full service pension (ie the pension ofsomeone with the same salary experiencebut who remains in the same scheme alltheir working life). Even someonechanging jobs once in mid-career canlose up to 16 per cent of the full servicepension. It is possible to reduceportability losses by, for example,indexing leaving salaries between the

provided by specific employers and whena worker changes jobs they have tomove to a new employer’s scheme.When they do so, they will either take atransfer value equal to the cashequivalent of their accrued pensionbenefits with them or leave a deferredpension in the scheme that they areleaving. Accrued benefits are valued lessfavourably if someone leaves a schemethan if they remain an active member ofthe scheme. This is because schemeleavers (whether they choose a transfervalue or a deferred pension) have theiryears of service valued in terms of theirleaving salary (although this is upratedannually to the retirement date by thelower of the inflation rate or 5 per cent),whereas continuing members will havethe same years of service as the early


Blake

Table 2: Portability losses from defined benefit schemes (percentage of full service pension received atretirement)

Workertype

Job separationassumptions1

Transfervalue2

Deferredpension3

DCpension(employer-run)4

Personalpension(employercontributions)5

Personalpension(no employercontributions)6

Average UK worker(MFR assumptionsrealised)7

Average UKmanual worker

Average UKnon-manualworker

ABC

ABC

ABC

757184

757184

757184

757184

888696

868394

71

78

79

61

66

68

37

45

44

Notes:1 This table presents estimates of the size of the portability losses experienced by three different types of UK

workers (based on typical lifetime earnings profiles) under three different sets of job separationassumptions: A — separates at ages 28, 29, 30, 40 and 57; B — separates at 26, 27, 30, 31, 38, 44 and55; C — separates at 45. The loss is expressed in the form of a reduced pension compared with whateach of the three workers would have received had they remained in a single scheme for their whole career.

2 Leaving worker takes transfer value to new scheme.3 Leaving worker leaves deferred pension in leaving scheme.4 Leaving worker transfers into employer-run DC scheme.5 Leaving worker transfers into personal pension scheme where the employer also contributes.6 Leaving worker transfers into personal pension scheme where the employer does not contribute.7 The MFR assumptions are the assumptions specified in the 1995 Pensions Act concerning future inflation,

earnings growth and investment returns that must be used by UK pension funds from April 1997 todetermine the minimum contribution level needed to meet projected pension liabilities.

Source: Blake, D. and Orszag, J. M. (1997) ‘Portability and Preservation of Pension Rights in the UK, Reportof the Director-General’s Inquiry into Pensions’, Vol. 3, Office of Fair Trading, London, July, Appendix E, Table5.8, p. 74.

simpler structure). Individual DCschemes in the UK take around 2.5 percent of contributions in administrationcharges and up to 1.5 per cent of thevalue of the accumulated assets in fundmanagement charges. The Institute ofActuaries has estimated that all thesecosts are equivalent to a reduction incontributions of between 10 and 20 percent; in contrast, the equivalent costs ofrunning an occupational scheme workout to between 5 and 7 per cent ofannual contributions. On top of this,most of the costs associated with anindividual DC scheme relate to theinitial marketing and set-up. To reflectthis, charges are also frontloaded, ie theyare extracted at the start-up of a schemerather than spread evenly over the life ofthe scheme. In many schemes, much ofthe first two years of contributions areused to pay sales commissions. This has adramatic effect in reducing the surrendervalue of a scheme if contributions ceaseearly on and it is transformed from anongoing to a paid-up basis. Thecumulative effect of these charges inrespect of DC schemes is shown in Table3. Over a 25-year investment horizon,the average scheme with a fullcontribution record takes around 19 percent of the fund value in charges, whilethe worst scheme provider takes around28 per cent.26 There is also evidence of asubstantial absence of persistency inregular premium personal pensionpolicies. Table 4 shows that the estimatedaverage lapse rate is 27 per cent aftertwo years and 53 per cent after fouryears: it is 84 per cent after 25 years(assuming a 6.5 per cent annual averagelapse rate after 4 years). The lapserate-adjusted reduction in contributionsfor a 25-year policy is 62 per cent: theeffective contributions into this schemefor a typical policy holder are just 38p inthe £.27

Further, although individual DC

leaving and retirement dates to thegrowth in real earnings or by providingfull service credits on transfers betweenjobs, but this is not common in the UK(except on transfers between differentpublic sector occupational pensionschemes).

There are also risks to the sponsors ofDB schemes that have becomeincreasingly apparent in the last fewyears, in particular, the emergence oflarge actuarial deficits as a result of fallingequity markets and the increasinglongevity of pensioners. The first factorhas led to a substantial decline on theasset side of the balance sheet of manypension schemes, while the second factorhas significantly increased their liabilities.As a consequence of this most DBpension schemes in the UK have closedtheir doors to new members in anattempt to cap their liabilities (DBschemes are funded on a balance of costbasis, so the employer is obliged to coverany deficit). In their place employershave established DC schemes for newemployees since this allows costs to bestrictly controlled. But as a result, anumber of risks have been transferred topension scheme members as outlinedbelow.

DC schemes

With DC schemes, it is important todistinguish between the accumulationand decumulation stages.

The accumulation stage

DC schemes have the advantage ofcomplete portability when changing jobs.However, individual DC schemes (suchas personal pension schemes) tend tohave much higher operating costs thanoccupational DB schemes (althoughoccupational DC schemes may havelower operating costs than occupationalDB schemes on account of their much



Moving to a DC scheme involves a‘backloading loss’ in addition to the cashequivalent loss incurred when leaving aDB scheme. The backloading loss arisesbecause benefits are backloaded in finalsalary schemes but not in DC schemes;this follows because salary and thereforeaccrued benefits generally increase withyears of service. Individuals transferringto a DC scheme (with age-independentcontributions) forego these backloadedbenefits: the marginal benefit from anadditional year’s membership of a DCscheme is simply that year’s contributions(plus the investment returns on these)which are usually a constant proportion ofearnings. If the DC scheme happens tobe a personal pension scheme then thereare also initial and annual charges to pay,plus the possible loss of the employer’scontribution. The impact of these factorscan be extensive as the above portabilitylosses indicate.28

Another problem with DC schemes,

schemes are portable between jobs, theyare not fully portable between schemeproviders or even between differentinvestment funds operated by the sameprovider. Transfers between personalpension scheme providers, for example,can incur charges of between 25 and 33per cent of the value of the assetstransferred. Transfers from DB schemesinto DC schemes can cost even morethan this. Table 2 shows that, even if aworker changes jobs only once inmid-career and moves out of a DBscheme, he would receive a reducedpension of 71–79 per cent of the fullservice pension if he moved to anemployer-run DC pension (with thesame total contribution rate as the DBscheme and no extra charges); 61–68 percent if he moved to a personal pensionscheme (where the employer alsocontributes); and only 37–44 per cent ifhe moved to a personal pension scheme(without employer contributions).


Blake

Table 3: Percentage of DC fund value represented by charges

5 years 10 years 15 years 20 years 25 years

Regular premium scheme(£200/month)Best commission-free fundBest commission-loaded fundIndustry averageWorst fund

Single premium scheme(£10,000)Best commission-free fundBest commission-loaded fundIndustry averageWorst fund

3.14.0

11.619.2

3.83.89.6

17.4

4.14.1

13.022.0

7.17.1

13.320.5

7.27.4

14.824.6

9.29.2

16.327.0

8.58.9

17.728.2

10.610.619.132.9

9.810.619.027.8

10.410.421.938.2

Source: Money Management (October 1998).

Table 4: Persistency rates for regular premium personal pension plans (percentages)

Company representatives: after Independent financial advisers: after

1 year 2 years 3 years 4 years 1 year 2 years 3 years 4 years

1993199419951996

84.183.785.586.6

72.372.875.0

63.664.4

56.7 91.591.390.890.2

83.382.181.6

76.674.5

70.5

Source: Personal Investment Authority (1998), see ref. 27, Table 1.

around 9 per cent of employee earningscompared with 15–18 per cent foroccupational DB schemes. Nevertheless,administration costs are much lower withoccupational DC schemes than withpersonal pension schemes, so even ifemployers made the same contributioninto an employee’s personal pensionscheme as into their own DC scheme,the final pension would still be lower inthe personal pension scheme.

Asset risk is not the only risk borneby DC scheme members and theirdependants. They also bear some of theother types of risk, namely, ill-health,disability and death-in-service. In DBschemes, these risks exist, but aretypically carried by the scheme sponsor.In DC schemes, protection against theserisks has to be purchased directly by themember as additional insurance policies.

Nevertheless, Table 5 shows that, aslong as individuals join a DC scheme ata sufficiently early age and maintain theircontribution record over a sufficientlylong investment horizon (and so get thebenefits of compounded returns), adecent pension in retirement can beachieved for a modest contribution rate.The table indicates that a 25-year oldmale can expect a pension of two-thirdsof final salary (the maximum availablefrom a DB scheme in the UK) with atotal net contribution rate of just under

in practice, is that total contributions intothem tend to be much lower than withDB schemes. In a typical DB scheme inthe UK, the employee’s contribution isabout 5–6 per cent of employeeearnings, while the employer’scontribution is double this at about10–12 per cent.29 The size of theemployer’s contribution is not widelyknown among employees; and, to anextent, the size of the employer’scontribution is irrelevant from theemployee’s viewpoint, since the pensiondepends on final salary, not on the levelof contributions. This is not the casewith DC schemes where the size of thepension depends critically on the size ofcontributions. When personal pensionschemes first started in the UK in 1988,most employers refused to contributeanything towards these schemes andmany workers were not fully aware ofthe penalty in terms of the reducedpension they were incurring as a result offoregoing the employer’s contribution.

All new occupational schemes beingestablished in the UK are DC schemes.The average employee contribution intosuch schemes is 3 per cent, while theaverage employer contribution is againdouble at 6 per cent (although someemployers only match the employee’scontribution).30 Total contributions intooccupational DC schemes are therefore



Table 5: Contributions needed to achieve a pension of two-thirds final salary

Age commencement maleRequired contributions(% of salary)

Maximum contributions(% of salary)

25 10.90 17.530 13.41 17.535 16.81 17.540 21.66 20.045 28.92 20.050 40.81 25.055 64.15 30.060 129.83 35.0

Assumptions: Male retiring at age 65; no previous contributions into any other pension scheme; salaryincreases by 3% p.a.; investment return 6% p.a.Source: Blake (1997), see ref. 10, Table 10.2.

to provide an income stream that can beused to pay the annuity.35

Even worse, the market for deferredannuities is extremely thin, particularly atdistant starting dates (where the market isvirtually non-existent). Where deferredannuities are available, they are very poorvalue for money. Deferred annuities areparticularly important in the case where aDB scheme is wound up, say, as a result ofthe insolvency of the sponsoring company.The assets of the scheme, which is oftenin deficit at the time (since the company,recognising its serious financial position,usually ceases making contributions intothe scheme some time before theinsolvency is formally declared), areinsufficient to pay the current and futurepension liabilities in full. In the past, theresidual assets in the scheme were used tobuy non-profit policies for currentpensioners and deferred annuities fordeferred pensioners. But fewer and fewerinsurance companies are willing to selldeferred annuities because of theuncertainties attached to forecastingmortality improvements.

Insurance companies use thegovernment bond market to protectthemselves against both interest rate andinflation risk. When an insurancecompany sells a level annuity it uses theproceeds to buy a fixed-incomegovernment bond of the same expectedterm as the annuity (typically 15 years)and then makes the annuity paymentsfrom the coupon payments received onthe bond. Similarly, when an insurancecompany sells an indexed annuity, it buysan index-linked bond of the sameexpected term as the annuity; few, if any,insurance companies sell indexedannuities with expected maturitiesbeyond that of the most distant tradingindexed-linked gilt. As a consequence,interest and inflation risk are transferredto the annuitant. If a DC schememember retires during an interest rate

11 per cent of earnings. The requiredcontribution rate rises sharply with age,however. Someone joining at 35 wouldneed a contribution rate of around 17per cent, and by the age of 40, therequired contribution rate is above themaximum permissible under theregulations establishing such schemes.

The distribution stage: Annuities

The weak tail of DC pension provisionlies in the distribution stage and relatesto the annuities market. The market forimmediate annuities is highlyconcentrated: of around 200 authorisedlife companies in the UK, only aboutten are serious providers of life annuitiesat any given time.31 There are a numberof problems facing both annuitants andannuity providers.32 First, there is anadverse selection bias associated withmortality risk. This is the risk that onlyindividuals who believe that they arelikely to live longer than the average forthe population of the same age will wishto purchase annuities. Secondly, mortalitytends to improve over time and therecan be severe consequences if insurancecompanies underestimate mortalityimprovements. Insurance companies addsubstantial cost loadings to cover theserisks, something of the order of 10–14per cent of the purchase price.33 Thirdly,there is inflation risk, the risk that withlevel annuities, unanticipated highinflation rapidly reduces the real value ofthe pension. Fourthly, there is interestrate risk. Annuity rates vary substantiallyover the interest rate cycle. They arerelated to the yields on governmentbonds of the same expected term; andsince these yields vary by up to 150 percent over the cycle,34 annuity rates willvary by the same order of magnitude.Finally, there is reinvestment or mismatchrisk arising from an inadequate supply oflong-maturing assets, such as governmentfixed-interest and indexed-linked bonds,


Blake

thereby gaining from any rise in the stockmarket over the life of the options.However, there are very few providers ofthese products in the UK.

A possible solution for thepost-retirement period is provided byvariable annuities. These were first issuedin 1952 in the USA by theTIAA-CREF.37 In the UK, they are betterknown as unit-linked or with-profitannuities, but only a few insurancecompanies offer them. A lump sum isused to buy units in a diversified fund ofassets (mainly equities) and the size of theannuity depends on the income andgrowth rate of assets in the fund. Theannuity can fall if the value of the assetsfalls substantially, so there is somevolatility to the annuity in contrast with alevel annuity. But since the pension froma level annuity is based on the yield ongilts, it is likely that the pension from avariable annuity, based on the return onequities, will generate a higher overallincome (assuming that the duration of theannuity is sufficiently great).

The Government could also do moreto ameliorate these market failures in theprivate provision of annuities which arise,in part, from aggregate risks that arebeyond the abilities and resources ofprivate insurance companies to hedge. Anumber of proposals have been suggestedrecently. For example, in order to helpthe private sector hedge against inflationrisk more effectively, the Goode Report(s. 4.4.44)38 suggested that thegovernment introduce a new type ofbond, with income and capital linked tothe retail price index, but with paymentof income deferred for a period. Suchbonds were given the name ‘deferredincome government securities’ (DIGS):they could be introduced with differentstarting and termination dates and wouldallow all deferred pensions to be indexedto prices. DIGS were never officiallyintroduced, but the introduction of the

trough (as happened in the mid to late1990s), he can end up with a very lowpension. Similarly, if a 65-year oldannuitant chooses an indexed annuity, hewill receive an initial cash sum that isabout 30 per cent lower than a levelannuity, and, with inflation at 3 per centp.a., it would take 11 years for theindexed annuity to exceed the levelannuity.36 Since retired people tend tounderestimate how long they willcontinue to live, most prefer to buy alevel annuity and thereby retain theinflation risk. In 1995, as a result offalling interest rates, the UK Governmentwas pressed into allowing incomedrawdown: it became possible to delaythe purchase of an annuity until annuityrates improved (or until age 75,whichever was sooner) and in theinterim take an income from the fundwhich remained fully invested.

However, until very recently, theinsurance industry (especially in Europe)has been reluctant to offer products thathelp annuitants hedge the risks, especiallyinterest rate risk, that they have beenforced to assume. Yet a whole range offinancial instruments and strategies isavailable to enable them do this. Thesimplest strategy, based on the principle ofpound cost averaging, involves a plannedprogramme of phased deferred annuitypurchases in the period prior toretirement which must be of sufficientlength to cover an interest rate cycle (say,5–7 years). A more sophisticated solutionfor the pre-retirement period is protectedannuity funds which employ derivativeinstruments. One example places afraction (eg 95 per cent) of the funds ondeposit and the rest in call options onbond futures contracts: if interest rates fallduring the life of the option, the profit onthe options will compensate for the lowerinterest rate. Another example places afraction of the funds in bonds and the restin call options on an equity index,



critical task of determining the attitudeto risk of the scheme member has beenundertaken. This usually involvesassessing the degree of risk tolerance ofthe scheme member. The greater thedegree of risk tolerance, the greater therisk that can be borne by the scheme’sassets and hence the greater the expectedvalue of the pension fund at theretirement date. This can be explained interms of the risk-adjusted expected valueof the asset portfolio which is defined asthe expected value of the pension assetsnet of a risk penalty, where the latterequals the ratio of the volatility of thefund’s assets to the member’s degree ofrisk tolerance. The higher the asset riskand the lower the risk tolerance, thegreater the risk penalty. The fundmanager’s task is to maximise therisk-adjusted expected value. It is possibleto increase the expected value of thepension assets by taking on more risk,but if too much additional risk is takenon, the risk-adjusted expected value willfall, especially if risk tolerance is low.The risk penalty shows the cost of takingon more risk.

Personal pension DC schemes in theUK are provided by financial institutionssuch as insurance companies, banks,building societies, unit trusts (ieopen-ended mutual funds), investmenttrusts (ie close-ended mutual funds), andopen-ended investment companies. Thescheme provider will offer the schememember a choice of investment vehiclesin which the pension assets willaccumulate, ranging from low risk (eg adeposit administration scheme), throughmedium risk (eg an endowment schemefrom an insurance company) to high risk(eg a unit-linked scheme). The depositadministration scheme is targeted at ascheme member with a very low degreeof risk tolerance, while the unit-linkedscheme is targeted at a scheme memberwith a high degree of risk tolerance.

gilt strips market in 1997 could helpinsurance companies construct themsynthetically. Similarly, the introductionof LPI bonds would allowpost-retirement inflation risk to behedged more effectively.

But the main causes of market failureare the risks associated with adverseselection and mortality. Making secondpensions mandatory rather than voluntarywould do much to remove the adverseselection bias in the demand forannuities.39 The Government could alsohelp insurance companies hedge the riskassociated with underestimating mortalityimprovements by issuing ‘survivorbonds’, a suggestion made in Blake andBurrows.40 These are bonds whose futurecoupon payments depend on thepercentage of the population ofretirement age on the issue date of eachbond who are still alive on the date ofeach future coupon payment. For a bondissued in 2000, for instance, the couponin 2010 will be directly proportional tothe amount, on average, that aninsurance company has to pay out as anannuity at that time. The insurancecompany which buys such a securitybears no aggregate mortality risk and, asa consequence, cost loadings fall. Thereis therefore much that could be done byboth government and the insuranceindustry to improve the market forannuities which remain the weak tail inDC pension provision.

The management of the pensionfund assetsDB and DC schemes are managed invery different ways.

DC schemes: Maximising risk-adjustedexpected value

The optimal management of a DCscheme is fairly straightforward, once the


Blake

employer’s contribution rate) into thefund. In practice, there are usually sometolerance limits. In the UK, for example,it is permissible for the value of assets tovary between 90 per cent and 105 percent of the value of liabilities (althoughthe surplus and deficit valuation basesdiffer). If the value of assets exceeds the105 per cent limit (on the statutoryvaluation basis), the scheme has up to 5years to reduce the value to 100 per centof liabilities (Finance Act 1986). Themost common means of doing this is theemployer’s contribution holiday, althoughother means are available: an employee’scontribution holiday, improved pensionbenefits or selling off financial assets, theproceeds from which are returned to thesponsor subject to a 35 per cent tax. Ifthe value of assets falls below 90 per centof the value of liabilities (on the MFRvaluation basis), the scheme has threeyears to raise the value of assets to 90per cent of liabilities and up to a further7 years to raise it back to 100 per cent(Pensions Act 1995). The most commonmeans of doing this is additionalemployer contributions (ie deficiencypayments).

Secondly, if the assets in the pensionfund are selected in such a way that theiraggregate volatility matches that of theliabilities, then the surplus risk can bereduced to zero. This requires the assetsin the pension fund to have both thesame variance as the pension liabilitiesand to be perfectly correlated with them(although it is unlikely in practice thatfinancial assets with these preciseproperties exist, unless governments inthe near future begin to issuezero-coupon wage-indexed bonds). This,in turn, requires the assets to constitute a‘liability immunising portfolio’, that is, aportfolio that immunises (or hedges) theinterest rate, real earnings growth rateand inflation rate risks embodied in thepension liabilities.43

However, it is arguable whetherlow-yielding deposits are a suitableinvestment vehicle for long-horizoninvestment programmes such as pensionschemes. Other asset categories, such asequities, have, in the past, offered muchhigher long-run returns. Furthermore,equities may have high short-termvolatility, but long-run returns have beenmuch more stable. Investing in depositadministration schemes or bonds hasbeen described as a strategy of ‘recklessconservatism’: these assets, while havingstable capital values in nominal termsover short horizons, do not tend to havelong-term returns that match the realgrowth rate in earnings. Despite this,surveys of personal pension schememembers in the UK and elsewhere tendto show that fear of short-term capitalloss drives many individuals towardsinvestment strategies that are recklesslyconservative in the long run.Nevertheless, once a scheme member hasselected a particular type of scheme, thefund manager’s task is to choose the assetmix (between equities, bonds, propertyetc) that maximises the risk-adjustedexpected value of the assets.41

DB schemes: Asset-liabilitymanagement

The appropriate investment managementstrategy for pension funds running DBschemes is asset-liability management(also called surplus risk management).42

This involves constructing a portfolio offinancial assets that (together withpromised future pension contributions)matches the pension liabilities in two keyrespects: size and volatility.

First, if pension schemes are alwaysfully funded, so that assets are alwayssufficient to meet liabilities in full, thenthe surplus in the fund will always bezero. This can be achieved by adjustingthe contribution rate (especially the



possible outcomes, and, in the lattercase, spells out the extent of the risksthat the pension fund sponsor faces.

With stochastic modelling on theother hand, tools such as monte carlosimulation are used to prepare adistribution of possible outcomes forboth assets and liabilities at the end ofthe relevant time horizon and thesponsor is presented with a range ofcontribution rates needed to achievefull funding over the period. The mostsensitive factor in any stochastic ALMmodel is the size chosen for the equityrisk premium, the projected excessreturn on equities over bonds. Smallincreases in this premium will tend tosingle large switches in the SAA infavour of equities and away frombonds.

There are two main uses of ALM.The first is to indicate theconsequences of adopting any particularinvestment strategy. The second is todiscover alternative strategies thatincrease the likelihood of meeting thefund’s objectives. Proponents ofasset-liability modelling argue that thestrategy allows pension funds togenerate higher returns without anyconsequential increase in risk. Themodelling exercise might indicate, forexample, that if current investmentreturns are sustained, there would beno need to change the employercontribution rate into the pension fundover the next five years. However, theworst-case scenario might indicate theemployer contribution rate might haveto rise by 10 per cent over the nextfive years. The exercise therefore allowsthe scheme sponsor to plan for thispossibility. As another illustration, themodelling exercise might indicate thatbecause a pension fund is maturing, itshould switch systematically out ofequities into fixed-income bonds (inthe five or so years prior to

Structuring the liability immunisingportfolio is the most important part ofdetermining the fund’s strategic assetallocation (SAA). The SAA is usuallydetermined by the funds’ actuary orinvestment consultant. Given the natureof the fund’s liabilities (which aretypically indexed to real wage growth),the liability immunising portfolio duringthe early life (ie immature stage) of apension scheme will contain a highproportion of equities and other ‘real’assets such as property, on the groundsthat, the shares of factors of productionin national income tend to berelatively stable, so that the returns tocapital (equity) and land (property) willover the long run match that onlabour (real wages). The actuary’sadvice will be based on anasset-liability modelling (ALM) exercise.ALM is a quantitative technique usedto help structure asset portfolios inrelation to the maturity structure ofliabilities. There are two commonversions of ALM, one based onscenario analysis, the other besed onstochastic modelling. Both versionsinvolve forecasts about how a pensionfund’s liabilities are going to accrueover a particular time horizon, thatmight be five, ten or 15 years ahead.To do this, assumptions concerningsalary growth rates, staff turnover, andthe age distribution and sexcomposition of the workforce have tobe made. Then forecasts concerning thefunding position of the pension schemehave to be generated. This involvesmaking projections of futurecontribution rates and also assessing thevalue of assets in relation to accruedliabilities. These forecasts andprojections can be made under differentscenarios concerning likely outcomes.Typically three scenarios are adopted:most likely, best-case and worst-case.This provides a realistic range of


Blake

The investment performance ofpension fund assetsGood or bad investment performance byDB and DC pension schemes have verydifferent consequences for schememembers. With DB schemes, theinvestment performance of the fund’sassets are of no direct relevance to thescheme member, since the pensiondepends on the final salary and years ofservice only and not on investmentperformance. The scheme member canrely on the sponsoring company to bailout the fund with a deficiency paymentif assets perform very badly. In extremecircumstances, however, it is possible fora firm and possibly the scheme tobecome insolvent. Of course, if the assetsperform well, the surplus is retained bythe sponsor.

However, investment performance iscritical to the size of the pension in thecase of a DC scheme: scheme membersbear all the investment risk in suchschemes. Scheme members, especiallypersonal pension scheme members, canfind themselves locked into a poorlyperforming fund, facing very high costsof transferring to a better performingfund. In addition, the type of funds inwhich personal pension scheme membersinvest can and do close down and thenthe assets do have to be transferred to adifferent fund. In this section, weexamine the investment performance ofpension scheme assets, beginning withthose of DC schemes.

Investment performance of DC schemes

The anticipated return in a high-riskinvestment vehicle must be greater thanthat in a low-risk investment vehicle, butthere can be wide differences in realisedreturns, even for schemes in the same riskclass. Blake and Timmermann44

conducted a study of the investmentperformance of unit trusts in the UK, one

retirement), which are more likely tomeet pension liabilities with lower riskof employer deficiency payments; this isknown as ‘lifestyle’ fund management(or ‘age phasing’).

Some fund managers are concernedthat ALM gives an unwarranted role tooutsiders, such as actuaries, in designingthe strategic asset allocation. Actuarieshave always had a role in determiningthe value of a pension scheme’s liabilities.But with the advent of ALM, actuarieshave begun to have a role in setting thelong-term or strategic asset allocationover, say, a ten-year horizon. Some fundmanagers claim they are being reducedto the subsidiary role of determiningtactical asset allocation (or market timing)and stock selection relative to this newlong-term strategic asset allocationbenchmark. However, not all fundmanagers are critical of the redefinitionof their respective roles. Many fundmanagers have positively welcomed theformal separation of long-term policydecisions from short-term tacticaldecisions that ALM allows.

Another potential problem concernsthe interpretation of measures ofinvestment performance in the light ofthe technique. ALM justifies differentpension funds pursuing differentinvestment policies. For example, small,fast-growing funds might pursue veryaggressive investment policies, while largemature funds might adopt more passiveinvestment policies. This makes it verydifficult to interpret a single performanceleague table drawn up on the assumptionthat all funds are pursuing the sameobjective of maximising expected returns.

Performance measurement serviceshave begun to take this into account byconstructing peer-group performanceleague tables, drawn up for sub-groups offunds following similar objectives.Performance measurement is nowdiscussed in more detail.



this to occur: the modal duration oftrusts was 4.25 years (51 months), butthe average duration was about 16 years.Across the unit trust industry, the averagereturn on funds that survived the wholeperiod was 13.7 per cent p.a., while theaverage return on funds that werewound up or merged during the periodwas 11.3 per cent p.a. This implies that atypical personal pension scheme membermight find him or herself locked into anunder-performing trust that is eventuallywound up or merged into a moresuccessful fund, experiencing anunder-performance of 2.4 per cent p.a.,over a 16-year period. This translatesinto a fund value that is 19 per centlower after 16 years than a fund that isnot wound up or merged. So it seemsthat in practice personal pension schememembers cannot rely on the markets toprovide them with a painless way ofextricating them from anunder-performing fund. They have to doit themselves, paying up to one-third ofthe value of their accumulated fund intransfer charges.

The investment performance of DBschemes

There are about 150,000 small definedbenefit pension schemes in the UK, most

of the key investment vehicles for DCschemes. Table 6 shows the distribution ofreturns generated by unit trusts operatingin the four largest sectors. These figuresindicate enormous differences inperformance, especially over the long lifeof a pension scheme. For example, the 4.1percentage point per annum differencebetween the best and worst performingunit trusts in the UK Equity Growthsector leads, over a 40-year investmenthorizon, to the accumulated fund in thetop quartile being a factor of 3.2 timeslarger than the accumulated fund in thebottom quartile for the same pattern ofcontributions. The 5.9 percentage pointper annum difference between the bestand worst performing unit trusts in theUK Smaller Companies sector leads to aneven larger fund size ratio after 40 yearsof 5.3.

So personal pension scheme memberscan find themselves locked into poorlyperforming funds. But should it not bethe case in an efficient capital marketthat systematically under-performingfunds fail to survive and are taken overby more efficient fund managers? Lunde,Timmermann and Blake45 investigatedthis possibility. They found thatunder-performing trusts are eventuallymerged with more successful trusts, butthat on average it takes some time for


Blake

Table 6: Distribution of returns generated by UK Unit Trusts, 1972–1995

Sector Top quartile MedianBottomquartile

Ratio offund sizes

UK Equity GrowthUK Equity GeneralUK Equity IncomeUK Smaller Companies

16.014.315.418.7

13.613.414.015.5

11.913.112.412.8

3.21.42.35.3

Note: The first three columns are averages measured in percentages per annum for the sample period1972–1995; the last column gives the ratio of fund sizes after 40 years based on the top and bottom quartilereturns. The formula is (assuming the same contribution stream):

(1 � rT )T � 1rT

�(1 � rB )T � 1

rB

where rT � 0.160, rB � 0.119 and T � 40, etc.Source: Blake and Timmermann (1998), see ref. 44, and Lunde, Timmermann and Blake (1999), see ref. 45.

The investment performance of UKdefined benefit pension fund managersbetween 1986 and 1994 has beeninvestigated in Blake, Lehmann andTimmermann.47–49 The data set usedcovers the externally-appointed activefund managers of more than 300medium-to-large pension funds with amandate agreement to ‘beat the market’.The UK pension fund industry is stillhighly concentrated and most of theseactive fund managers come from just fivegroups of professional fund managers(Deutsche Asset Management, MerrillLynch Investment Management, UBSAsset Management, Schroder InvestmentManagement and Gartmore PensionFund Managers).

While the average or medianperformance has been very good overthe sample period, important implicationsconcerning the behaviour of fundmanagers can be derived from anexamination of the distribution of thisperformance about the median. Table 7shows the cross-sectional distribution ofreturns realised by the pension funds inthe sample over the period 1986–94 inthe most important individual asset classesas well as for the total portfolio. Thesemi-interquartile range is quite tight,below 2 percentage points for most assetclasses and only just over 1 percentage

with fewer than 100 members in each.Virtually all these schemes are managedon a pooled basis by insurancecompanies. There are about 2,000 largeschemes, including 70 or so with assetsin excess of £1bn each.46 As indicatedabove, the investment performance ofthese funds is much more important forthe scheme sponsor than for the schememember. The recent history of the UKpension fund industry embraces a periodof substantial deficiency payments in the1970s (arising from the UK stock marketcrash in 1974–1975), and the build up ofhuge surpluses during the bull markets ofthe 1980s and 1990s and the decentagain into deficits as a result of the 30per cent fall in the UK equity marketbetween 2000 and 2002. The surplusesenabled sponsors to reduce theircontributions into their schemes (ie totake employer’s contribution holidays). Inother words, during the 1980s and1990s, UK pension scheme sponsorsbenefited enormously from theinvestment successes of their fundmanagers. However, poor recentinvestment performance has encouraged alarge number of employers to closedown their final salary schemes andreplace them with defined contributionschemes in which the scheme memberassumes all the investment risk.



Table 7: Fractiles of total returns by asset class for UK Managed Funds, 1986–1994 (average annualised percentages)

UKequities

Internationalequities UK bonds

Internationalbonds

UKindexbonds

Cash/otherinvestments

UKproperty Total

Minimum5%

10%25%50%75%90%95%MaxMax–Min

8.5911.4311.8512.4413.1313.9314.8115.4617.398.80

4.428.599.039.64

10.6511.7612.5213.1414.6810.26

6.599.449.95

10.4310.7911.2211.7012.0517.2310.64

–0.642.187.568.30

11.3713.3714.5518.1526.3426.98

5.597.207.817.918.228.458.808.69

10.074.48

2.675.467.608.97

10.2511.2514.2016.1319.7317.06

3.055.076.588.038,759.99

10.8411.3613.5310.48

7.2210.6010.9611.4712.0612.5913.1313.3915.037.81

Note: The table shows the fractiles of the cross-sectional distribution of returns on individual asset classes as well ason the total portfolio.Source: Blake, Lehmann and Timmermann (2002), see ref. 48, Table 1.

fund manager’s fee is taken into account.Further, only 42.8 per cent of fundsout-perform the market average. Themain explanation for this is the relativeunder-performance in UK equities, thelargest single category with an averageportfolio weighting of 54 per cent overthe sample period; the averageunder-performance was –0.33 per centp.a. and only 44.8 per cent of UKpension funds beat the average return onUK equities. To be sure, relativeperformance is better in other assetcategories, especially UK andinternational bonds, but the portfolioweights in these asset categories are notlarge enough to counteract the relativeunder-performance in UK equities.

Tables 7 and 8 together indicate howclose the majority of the pension fundsare to generating the average marketreturn. The median fund generated anaverage total return of 12.06 per centp.a., just 12 basis points short of theaverage market return, and 80 per centof the funds are within one percentagepoint of the average market return. Thissuggests that, despite their claims to beactive fund managers, the vast majorityof UK pension fund managers are notonly herding together, they are alsocloset index matchers.

There are some other features of UKpension fund performance worthy of

point for the total portfolio return. Thissuggests evidence of a possible herdingeffect in the behaviour of pension fundmanagers: fund managers, although theirfee is determined by their absoluteinvestment performance, are appointedand evaluated on the basis of their relativeperformance against each other andtherefore have a very strong incentivenot to under-perform the peer-group.50

The fund managers in the sample areactive managers who have won mandateson the basis of promises to beat themarket: they are not passive managersattempting to match the market. If theywere genuinely pursuing active strategies,there would be a wide dispersion inperformance as is observed in the USA.There is a tight dispersion ofperformance about a median. From thisit can be concluded that the active fundmanagers are herding to avoid deliveringpoor relative performance (which putstheir mandate at risk). Despite this, thedifference between the best and worstperforming funds is very large, as the lastrow of Table 7 indicates.

Table 8 shows how well UK pensionfunds have performed in comparisonwith other participants in the market.The fourth column shows that theaverage UK pension fundunder-performed the market average by0.45 per cent p.a.; and this is before the


Blake

Table 8: Performance of UK pension funds in comparison with the market, 1986–1994 (percentages)

Averageportfolioweight(%)

Averagemarketreturn(%)

Averagepensionfund return(%)

Averageout-performance(%)

Percentageout-performers

UK equitiesInternational equitiesUK bondsInternational bondsUK index bondsCash/other investmentsUK propertyTotal

53.719.57.62.22.74.58.9

13.311.1110.358.648.229.909.00

12.18

12.9711.2310.7610.038.129.019.52

11.72

–0.330.120.411.39

–0.100.890.52

–0.45

44.839.877.368.851.759.539.142.8

Note: International property is excluded since no market index was available.Source: Blake, Lehmann and Timmermann (1999, 2002), see refs 47–48.

We found that 32 per cent of thequartile containing the largest fundswere also in the quartile containing theworst performing funds, whereas only15 per cent of the quartile containingthe smallest funds were also in thequartile of worst performing funds.These results confirm the often-quotedview that ‘size is the anchor ofperformance’: because large pensionfunds are dominant players in themarkets, this severely restricts theirabilities to out-perform the market.

The final result concerns the abilitiesof UK pension fund managers in activefund management, that is, in theirattempts to beat the market incomparison with a passive buy and holdstrategy. The most important task ofpension fund managers is, as we haveseen above, to establish and maintain thestrategic asset allocation. This isessentially a passive management strategy.However, fund managers claim that theycan ‘add value’ through the activemanagement of their fund’s assets. Thereare two aspects to active management:security selection and market timing (alsoknown as tactical asset allocation).Security selection involves the search forundervalued securities (ie involves thereallocation of funds within sectors) andmarket timing involves the search forundervalued sectors (ie involves thereallocation of funds between sectors).The total return generated by fundmanagers can be decomposed into thefollowing components:

%Strategic asset allocation 99.47Security selection 2.68Market timing –1.64Other –0.51

Total 100

99.47 per cent of the total return

note. First, there is some evidence ofshort-term persistence in performanceover time, especially by the best andworst performing fund managers. Forexample, we found that UK equity fundmanagers in the top quartile ofperformance in one year had a 37 percent chance of being in the top quartilethe following year, rather than the 25per cent that would have been expectedif relative performance arose purely bychance. Similarly, there was a 32 percent chance of the fund managers in thebottom quartile for UK equities for oneyear being in the bottom quartile thefollowing year. There was also evidenceof persistence in performance in the topand bottom quartiles for cash/otherinvestments, with probabilities ofremaining in these quartiles the followingyear of 35 per cent in each case.However, there was no evidence ofpersistence in performance for any otherasset category or for the portfolio as awhole. Nor was there any evidence ofpersistence in performance over longerhorizons than one year in any assetcategory or for the whole portfolio. Thissuggests that ‘hot hands’ in performanceis a very short-term phenomenon.

Secondly, there was some evidence ofspillover effects in performance, but onlybetween UK and international equities.In other words, the funds that performedwell or badly in UK equities alsoperformed well or badly in internationalequities. This suggests that some fundmanagers were good at identifyingundervalued stocks in different markets.This result is somewhat surprising sincethe world’s equity markets are much lesshighly integrated than the world’s bondmarkets, yet there was no evidence ofspillover effects in performance acrossbond markets.

Thirdly, there was evidence of a sizeeffect in performance. Large fundstended to under-perform smaller funds.



about. Nevertheless, the industry and itsdecision-taking structures face forbiddingchallenges: an ageing population,unrecognisably different labour markets,shifting employer attitudes. In the world wenow face, an ever-higher premium is likelyto be placed on efficiency and flexibility.The review finds that savers’ money is toooften being invested in ways that do notmaximise their interests. It is likely to followtoo that capital is being inefficiently allocatedin the economy. The review sets out ablueprint for change, to drive clearerincentives and tougher customer pressuresthroughout the savings and investmentindustry.’

The review identified the following maindistortions:

— pension fund trustees, who are the verycentre of the system, are being askedto take crucial investment decisions,yet many lack resources and expertise.They are often unsupported byin-house staff, and are rarely paid;

— as a result, they rely heavily on anarrow group of investment consulting(mainly actuarial) firms for advice.Such firms are small in number, havea narrow range of expertise and littleroom for specialisation. Furthermore,their performance is not usuallyassessed or measured;

— a particular consequence of thepresent structure is that asset allocation(the selection of which markets, asopposed to which individual stocks,to invest in) is an under-resourcedactivity. This is especially unfortunategiven the weight of academicevidence suggesting that thesedecisions can be critical determinantsof investment performance;

— a lack of clarity about objectives at anumber of levels. Fund managers arebeing set objectives which, takentogether, appear to bear little coherentrelationship to the ultimate objective

generated by UK fund managers can beexplained by the strategic asset allocation,that is, the long-run asset allocationspecified by pension scheme sponsors onthe advice of their actuaries following anALM exercise. This is the passivecomponent of pension fund performance.The active components are securityselection and market timing (or TAA).The average pension fund wasunsuccessful at market timing, generatinga negative contribution to the totalreturn of –1.64 per cent. The averagepension fund was, however, moresuccessful in security selection, making apositive contribution to the total returnof 2.68 per cent. But the overallcontribution of active fund managementwas just over 1 per cent of the totalreturn (or about 13 basis points), whichis less than the annual fee that active fundmanagers charge (which ranges between 20basis points for a £500m fund to 75basis points for a £10m fund).51

The Myners review of institutionalinvestmentIn March 2001 the HM Treasury-sponsored review of institutionalinvestment chaired by Paul Myners,Chief Executive of Gartmore, waspublished.52 Its recommendations wereimmediately accepted in full by theGovernment.53 The report called for anew approach to institutional investment,identified a series of current distortions toeffective decision making, and suggestedways of tackling them.

In introducing his report, Paul Mynerssaid:

‘Our funded pensions system, ourhighly-developed equity culture and theprofessionalisation of investment in the UKare an enviable success story. I pay tribute tothe commitment and dedication ofinstitutions and their advisers in bringing this


Blake

approach taken on corporate governanceby the Cadbury (and later) codes. Theseprinciples would apply first to pensionfunds and subsequently to otherinstitutional investors. As with theCadbury code, they would not bemandatory. However, where a pensionfund chose not to comply with them, itwould have to explain to its memberswhy not.

The proposed set of principles for DBpension schemes is as follows.

— Effective decision making. Decisionsshould be taken only by persons ororganisations with the skills,information and resources necessary totake them effectively. Where trusteeselect to take investment decisions,they must have sufficient expertise tobe able to evaluate critically anyadvice they take. Trustees shouldensure that they have sufficientin-house staff to support them in theirinvestment responsibilities. Trusteesshould also be paid, unless there arespecific reasons to the contrary. It isgood practice for trustee boards tohave an investment subcommittee toprovide appropriate focus. Trusteesshould assess whether they have theright set of skills, both individuallyand collectively, and the rightstructures and processes to carry outtheir role effectively. They shoulddraw up a forward-looking businessplan.

— Clear objectives. Trustees should set outan overall investment objective forthe fund that: (i) represents their bestjudgment of what is necessary tomeet the fund’s liabilities, given theirunderstanding of the contributionslikely to be received from employer(s)and employees; and (ii) takes accountof their attitude to risk, specificallytheir willingness to acceptunder-performance due to market

of the pension fund, namely to meetits pension obligations;

— fund managers are often set objectiveswhich give them unnecessary andartificial incentives to herd. So-calledpeer-group benchmarks, directlyincentivising funds to copy otherfunds, remain common. Risk controlsfor active managers are increasinglyset in ways which give them littlechoice but to cling closely to stockmarket indices, making meaningfulactive management near-impossible;

— there is also extreme vagueness aboutthe timescales over which fundmanagers’ performance is to bejudged. This is a real (but whollyunnecessary) cause of short-termism infund managers’ approach toinvestment;

— fund managers remain unnecessarilyreluctant to take an activist stance inrelation to corporate under-performance, in companies wherethey own substantial shareholdings,even where this would be in theirclients’ financial interests;

— finally, an important cost to pensionfunds, namely broking commission, issubject to insufficient scrutiny. Clearerand more rigorous disciplines couldbe applied to these costs, which aresubstantial;

— in the life insurance industry,competition, though intense, tendsnot to focus directly on investmentperformance, and this issue needs tobe tackled if stronger incentives toefficient investment decision-makingin the industry are to be created.

The review makes a number of proposalsto deal with these distortions. The keyproposal is the introduction of astatement of the principles of institutionalinvestment, incorporating a short set ofclear principles of investment decisionmaking. The idea is modelled on the



of any set of financial instruments,without clear justification in the lightof the specific circumstances of thefund. The mandate should incorporatea management fee inclusive of anyexternal research, information ortransaction services acquired or usedby the fund manager, rather thanthese being charged to clients.

— Activism. Making intervention incompanies, where it is inshareholders’ interests, a duty for fundmanagers. The mandate shouldincorporate the principle of the USDepartment of Labor InterpretativeBulletin on activism. Managers shouldhave an explicit strategy, elucidatingthe circumstances in which they willintervene in a company; the approachthey will use in doing so; and howthey measure the effectiveness of thisstrategy. The US Department ofLabor Interpretative Bulletin 26 onactivism is as follows.• The fiduciary act of managing plan

assets that are shares of corporatestock includes the voting of proxiesappurtenant to those shares of stock.

• The fiduciary obligations ofprudence and loyalty to planparticipants and beneficiaries requirethe responsible fiduciary to voteproxies on issues that may affect thevalue of the plan’s investment.

• An investment policy thatcontemplates activities intended tomonitor or influence themanagement of corporations inwhich the plan owns stock isconsistent with a fiduciary’sobligations under the 1974Employee Retirement IncomeSecurity Act when the responsiblefiduciary concludes that there is areasonable expectation that activitiesby the plan alone, or together withother shareholders, are likely toenhance the value of the plan’s

conditions. Objectives for the overallfund should not be expressed in termswhich have no relationship to thefund’s liabilities, such as performancerelative to other pension funds, or toa market index.

— Focus on asset allocation. Strategic assetallocation decisions should receive alevel of attention (and, whererelevant, advisory or managementfees) that fully reflect the contributionthey can make towards achieving thefund’s investment objective. Decisionmakers should consider a full range ofinvestment opportunities, notexcluding from consideration anymajor asset class, including privateequity. Asset allocation should reflectthe fund’s own characteristics, not theaverage allocation of other funds.

— Expert advice. Contracts for actuarialservices and investment advice shouldbe opened to separate competition.The fund should be prepared to paysufficient fees for each service toattract a broad range of kinds ofpotential providers.

— Explicit mandates. Trustees shouldagree with both internal and externalinvestment managers an explicitwritten mandate covering agreementbetween trustees and managers on: (i)an objective, benchmark(s) and riskparameters that together with all theother mandates are coherent with thefund’s aggregate objective and risktolerances; (ii) the manager’s approachin attempting to achieve theobjective; and (iii) clear timescale(s) ofmeasurement and evaluation, suchthat the mandate will not beterminated before the expiry of theevaluation timescale other than forclear breach of the conditions of themandate or because of significantchange in the ownership or personnelof the investment manager. Themandate should not exclude the use


Blake

approximations involved in indexconstruction and selection; considerexplicitly for each asset class invested,whether active or passive managementwould be more appropriate given theefficiency, liquidity and level oftransaction costs in the marketconcerned; and where they believeactive management has the potentialto achieve higher returns, set bothtargets and risk controls that reflectthis, giving managers the freedom topursue genuinely active strategies.

— Performance measurement. Trusteesshould arrange for measurement ofthe performance of the fund andmake formal assessment of their ownprocedures and decisions as trustees.They should also arrange for a formalassessment of performance anddecision making delegated to advisersand managers.

— Transparency. A strengthened Statementof Investment Principles should setout: (i) who is taking which decisionsand why this structure has beenselected; (ii) the fund’s investmentobjective; (iii) the fund’s planned assetallocation strategy, including projectedinvestment returns on each asset class,and how the strategy has been arrivedat; (iv) the mandates given to alladvisers and managers; and (v) thenature of the fee structures in placefor all advisers and managers, andwhy this set of structures has beenselected.

— Regular reporting. Trustees shouldpublish their Statement of InvestmentPrinciples and the results of theirmonitoring of advisers and managersand send them annually to membersof the fund. The Statement shouldexplain why a fund has decided todepart from any of these principles.

The following principles are proposed forDC schemes:

investment, after taking intoaccount the costs involved. Such areasonable expectation may exist invarious circumstances, for example,where plan investments in corporatestock are held as long-terminvestments or where a plan maynot be able to easily dispose suchan investment.

• Active monitoring andcommunication activities wouldgenerally concern such issues as theindependence and expertise ofcandidates for the corporation’sboard of directors and assuring thatthe board has sufficient informationto carry out its responsibility tomonitor management. Other issuesmay include such matters asconsideration of the appropriatenessof executive compensation, thecorporation’s policy regardingmergers and acquisitions, the extentof debt financing and capitalisation,the nature of long-term businessplans, the corporation’s investmentin training to develop its workforce,other workplace practices andfinancial and non-financial measuresof corporate performance. Activemonitoring and communicationmay be carried out through avariety of methods including bymeans of correspondence andmeetings with corporatemanagement as well as byexercising the legal rights of ashareholder.

— Appropriate benchmarks. Trustees should:explicitly consider, in consultationwith their investment manager(s),whether the index benchmarks theyhave selected are appropriate; inparticular, whether the construction ofthe index creates incentives to followsub-optimal investment strategies; ifsetting limits on divergence from anindex, ensure that they reflect the



every defined benefit pension fundwould be required to publish:• the current value of its assets and in

what asset classes they were invested• the assumptions used to determine

its liabilities• planned future contributions• its planned asset allocation for the

following year or years• the assumed returns and assumed

volatilities of those returns for eachasset class sufficient to meet theliabilities

• a justification by the trustees of thereasonableness of both their assetallocation and the investmentreturns assumed in the light of thecircumstances of the fund and ofthe sponsor

• an explanation of the implicationsof the volatility of the investmentvalues for possible underfunding,and a justification by trustees ofwhy this level of volatility is judgedto be acceptable;

— pension fund surpluses. The LawCommission to be asked whether itcan suggest greater legal clarity aroundthe ownership of surplus pension fundassets, and reduction of the rate of taxon distributed pension fund surpluses;

— private equity. Investment in privateequity should benefit from theframework set out by the principlesand from the replacement of theMFR. The review also made anumber of proposals which takeaccount of the special nature of privateequity as an asset class for institutionalinvestors, including changes to themaximum number of partners in alimited partnership and changes to thetaxation of investments in limitedpartnerships. It also calls for the BritishVenture Capital Association to takeaction to improve transparency anddisclosure about issues such asinvestment returns and compensation;

— when selecting funds to offer asoptions to scheme members, trusteesshould consider the investmentobjectives, expected returns, risks andother relevant characteristics of eachsuch fund;

— where a fund is offering a defaultoption to members through acustomised combination of funds,trustees should ensure that anobjective is set for the option,including expected risks and returns;

— schemes should, as a matter of bestpractice, consider a full range ofinvestment opportunities, includingless liquid and more volatile assets. Inparticular, investment trusts should beconsidered as a means of investing inprivate equity;

— the Government should keep underclose review the levels of employerand employee contributions to DCpensions, and the implications forretirement incomes.

In commenting on the proposed set ofprinciples, Paul Myners said: ‘Theprinciples may seem little more thancommon sense. In a way they are — yetthey certainly do not describe the statusquo. Following them would requiresubstantial change in decision-makingbehaviour and structures.’ The reportcalled for the industry to adopt theprinciples voluntarily within two years,but if necessary the government shouldlegislate to require disclosure againstthem.

The review made a series of otherproposals. The main ones relate to:

— minimum funding requirement. Thereplacement of the MFR with aregime based on transparency anddisclosure, under which pension fundswould report publicly on the currentfinancial state of the fund and onfuture investment plans. Each year,


Blake

individual behaviour and characteristics,for example, how often someone changesjobs and their attitude to risk. The morefrequently someone changes jobs and themore risk tolerant they are, the moreappropriate it will be for them to choosea DC scheme.

However, even if someone has chosenthe appropriate pension scheme inprinciple, weaknesses in the design oftheir scheme can lead to lower pensionsthan otherwise need be the case. Oneillustration of this concerns investmentperformance: it affects the net cost to thesponsor of a DB scheme and the netpension benefit to the member of a DCscheme. It has been shown in this paperthat, on average, UK pension funds haveunder-performed the market, and whilethere has been a wide dispersion ofperformance by individual fundmanagers, most of them appear to herdaround the median fund manager.Furthermore, fund managers have notbeen especially successful at active fundmanagement: virtually the same or betterreturns could have been generated ifpension funds had invested passively inindex funds. In addition, fundmanagement costs would have beenlower and the dispersion in returns acrossfund managers would have been reduced.Another example concerns charges. It ismost unlikely that good investmentperformance can compensate for highcharges, and we have seen that it isequally unlikely that above-averageinvestment performance can be sustainedfor a significant period of time.Well-designed pension schemes wouldtake these factors into account.

Some important policy conclusionsemerge from this analysis. First, ifgovernments want to see well-designedpensions in the private sector, they mustprovide an infrastructure that helps theprivate sector deliver these. Theregulatory framework should be kept as

— compensation. The level ofcompensation provided by thePensions Compensation Board fornon-pensioner members should beincreased to cover not simply the 90per cent of MFR liabilities as atpresent, but something closer to thecost of securing members’ accruedrights (or the amount of the loss,whichever is the lesser);

— independent custody. There should be astatutory requirement for funds tohave independent custody of assets.

ConclusionOver the last 25 years, governments havehad two major impacts on pensionprovision in the UK. First, they havereduced the cost of providing statepensions by reducing the level of benefitsfrom the state schemes. Secondly, theyhave encouraged greater and moreeffective private sector provision,although the Conservative and LabourGovernments have done this in quitedifferent ways. The Thatcher-MajorGovernments made privatesupplementary pension arrangementsvoluntary and used tax incentives toencourage consumers to join personalpension schemes, but they left it to themarket to determine the structure andefficiency of these schemes. The resultwas schemes that exhibited very highfront-loaded charges, because retailcustomers tend not to be skilled atassessing the cost-effectiveness of retailfinancial products.54 In contrast, the BlairGovernment, recognising the marketfailure arising from poorly informedconsumers, imposed restrictions on thestructure of stakeholder pension schemesthat helped to force economies of scaleand hence lower charges.

The suitability of the two key types ofprivate funded scheme, DB or DC, toparticular workers depends on both



some of these issues. For example, thenew stakeholder pension schemes havean upper limit placed on the charges thatcan be imposed and this will effectivelyrule out the active management of theassets in such schemes; and, in the USA,the Government is considering a rangeof options for dealing with the growingburden of social security, including theestablishment of individual privatisedaccounts and the investment of part ofthe Social Security Trust Fund inequities.

However, the greatest impediment tohaving a decent pension in retirement isinadequate pension savings made duringthe working lifetime. There is a strongcase for arguing that only with sufficientmandatory minimum contributions into afunded pension scheme (with creditsgiven to those on very low earnings) cana decent pension be achieved, but fewgovernments seem willing to confrontthis issue; the UK mandatory minimumfor the second pension (equal to thecontracted-out rebate on NationalInsurance contributions of 4.6 per centof earnings) is not sufficient to build toan adequate pension (as Table 5 showed)and the Welfare Reform and PensionsAct explicitly ruled out additionalcompulsory contributions.

References and notes1 NICs also build up entitlement to health service,

sickness, disability and incapacity benefits and thejob seeker’s allowance.

2 Worth £75.50 per week for a single person in2002–2003, while national average earnings were450 per week, suggesting a replacement ratio ofabout 17 per cent.

3 The LEL was £75 per week and the UEL was 585per week in 2002–2003.

4 The Department of Social Security (DSS) wasrenamed the Department of Work and Pensions inJune 2001.

5 The State Pension Age for women is beingprogressively raised to 65 over the period 2010–20.

6 The non-contracted out National Insurancecontribution rate in 2002–2003 for employees was10 per cent of earnings between £89 per week and

simplified as possible in order tominimise compliance costs, and chargingstructures should be made simple andtransparent to enable consumers identifythe most competitive providers moreeasily. Governments could also help keepcosts down or improve benefits in otherways: for instance, by enablingeconomies of scale to be exploited morefully (eg establishing a central clearinghouse to channel contributions in thecase of DC schemes) or by introducing acommon set of actuarial assumptions, asin Holland, which would enable fullservice credits to be transferred betweenschemes when workers change jobs,thereby improving the portability rightsof members of DB schemes.Governments could help the privatesector cope with the market failures thatprevent or at least make it difficult forindividuals to hedge certain risks, egsurplus risk could be hedged moreeffectively through the introduction ofzero-coupon wage-indexed bonds andmortality risk could be hedged throughthe introduction of survivor bonds.

Secondly, if governments wish topromote the efficient investmentmanagement of pension assets, theyshould not put in place regulations thatdistort pension fund asset allocations thatcould happen with FRS17 or theproposed revision to the MFR; this hasbeen explicitly recognised in the Mynersreport on institutional investment. Theyshould also encourage the introduction ofappropriate incentives, such as greatertransparency in published performancedata and the adoption ofperformance-related fund managementfees.55 This would encourage the lesstalented fund managers to invest in indexfunds, with consequential benefits interms of lower fund management chargesand a lower dispersion of performance.56

There is evidence that governmentsare becoming more aware of at least


Blake

16 Field, F. (1996) ‘Stakeholder Welfare’, Choice inWelfare Series No. 32, Institute of Economic Affairs,London.

17 Field, F. (1996) ‘How to Pay for the Future:Building a Stakeholders’ Welfare’, Institute ofCommunity Studies, London.

18 Money Management, October 1998.19 Had the indexation of the BSP been preserved to

the growth rate in national average earnings since1980, the BSP would have been £95 per week in1999 rather than £66.75 (Daily Telegraph, 31st July,1999).

20 In 1991, Robert Maxwell stole £400m from thepension funds of his companies, Mirror GroupNewspapers and Maxwell CommunicationCorporation. He had been a trustee of both thesepension funds.

21 This section draws heavily on Department of SocialSecurity and HM Treasury (2001) ‘Security forOccupational Pensions: The Government’s proposal’,London, March.

22 Myners, P. (2001) ‘Institutional Investment in theUnited Kingdom: A Review’, HM Treasury,London, March.

23 Accounting Standards Board (1997) ‘Discounting inFinancial Reporting’, ASB Publications, MiltonKeynes.

24 However, there is an increasing number of hybridschemes being introduced which combine features ofboth DB and DC schemes. It is also possible tohave unfunded DB and DC schemes.

25 Burgess, S. and Rees, H. (1994) ‘Lifetime Jobs andTransient Jobs: Job Tenure in Britain 1975–91’,Centre for Economic Policy Research, London;Gregg, P. and Wadsworth, J. (1995) ‘A ShortHistory of Labour Turnover, Job Tenure, and JobSecurity, 1973–93’, Oxford Review of Economic Policy,Vol. 11, pp. 73–90.

26 It is the high costs associated with individualpersonal pension schemes in the UK that has ledmany small companies without the resources to runeither occupational DB or occupational DC schemesto establish GPPs which have lower unit costs thanpersonal pension schemes. The high costs of personalpension schemes were also responsible for thegovernment setting up low-cost stakeholder pensionschemes.

27 The lapse rates come from Personal InvestmentAuthority (1998) ‘Survey of Persistency of Life andPension Policies’, London, October, while lapserate-adjusted reductions in contributions areestimated in Blake, D. and Board, J. (2000)‘Measuring Value Added in the Pensions Industry’,Geneva Papers on Risk and Insurance, Vol. 25, pp.539–567.

28 There are other costs which are more difficult toquantify, the most important of which are searchand information costs. Office of Fair Trading (1997)‘Report of the Director General’s Inquiry intoPensions’, London, July found (on the basis of asurvey it conducted) that most people in the UKdid not regard themselves as being financially literate

the UEL, while for employers it was 11.8 per centon all earnings above 89 per week.

7 UK private pension schemes benefit from an EETsystem of tax breaks: the contributions into schemesare exempt from tax, the investment returns (withthe exception, since 1997, of dividend income onUK equities) are exempt from tax, and the pensionis taxed (with the exception of a tax-free lump sumequal to 1.5 times the final salary in the case of adefined benefit scheme and 25 per cent of theaccumulated pension fund in the case of a definedcontribution scheme).

8 Economic Trends Annual Supplement 1999(Table 1).

9 Department of Social Security (1998) ‘A NewContract for Welfare: Partnership in Pensions’,Department of Social Security, London (Cm 4179,December), Table 1.0; National Association ofPension Funds (1997) ‘Annual survey’, NAPF,London; and estimates by the Government Actuary’sDepartment.

10 For more details of the UK pension system, seeBlake, D. (1997) ‘Pension Choices and PensionsPolicy in the United Kingdom’, in S. Valdes-Prieto(ed.), ‘The Economics of Pensions: Principles,Policies and International Experience’, CambridgeUniversity Press, New York, pp. 277–317; Blake, D.(2003) ‘Pension Schemes and Pension Funds in theUnited Kingdom’, Oxford University Press, Oxford(second edition); Fenton, J., Ham, R. and Sabel, J.(1995) ‘Pensions Handbook’, Tolley Publishing,Croydon; Reardon, A. M. (2002) ‘PensionsHandbook’, Zurich Publishing, London; PensionsProvision Group (1998) ‘We All Need Pensions:Prospects for Pension Provision’, The StationeryOffice, London.

11 This is partly because personal pension schemes haveonly been around since 1988.

12 Although the backloading effect is lower in averagesalary schemes (such as SERPS) than in final salaryschemes (such as a typical occupational scheme).

13 Department of Social Security (2000) ‘The PensionCredit: A Consultation Paper’, Department of SocialSecurity, London (Cm4900, November).

14 An additional £3bn per year (Daily Telegraph, 31stJuly 1999).

15 In fact, the Conservative Government in the UKannounced in March 1997 plans to privatise theentire state pension system from the turn of thecentury and to end its unfunded nature. Allindividuals in work would receive rebates on theirNICs which would be invested in a personalisedpension account. The initial costs in terms ofadditional taxation were estimated to be £160m inthe first year, rising to a peak of £7bn a year in2040. However, the long term savings to thetaxpayer from the end of state pension provisionwere estimated to be £40bn per year (all in 1997prices). The proposals were put on hold as a resultof the Conservative Government’s defeat in the May1997 General Election (see ‘Basic Pension Plus’,Conservative Central Office, 5th March, 1997).



account of recent government budget surpluses andthe consequential absence of new gilt issues, seeBishop, G. (1999) ‘Why are Long Gilts the RichestBonds in the World, and Getting Richer?’,SalomonSmithBarney, London, October.

36 Khorasanee, M. Z. (1996) ‘Annuity Choices forPensioners’, Journal of Actuarial Practice, Vol. 4,pp. 229–255.

37 Teachers Insurance and Annuity Association ofAmerica, College Retirement Equity Fund,www.tiaa-cref.org

38 Goode, R. (1993) ‘Pension Law Reform: Report ofthe Pension Law Review Committee’, CM 2342-I,HMSO, London.

39 There is a growing body of support formandatory contributions into second pensions,including Field, F. and Owen, M. (1993) ‘PrivatePensions for All: Squaring the Circle’, FabianSociety Discussion Paper No. 16, London; Borrie,G. (Chairman) (1994) ‘Social Justice — Strategiesfor National Renewal’, Report of the Commissionfor Social Justice, Vintage, London; World Bank(1994) ‘Averting the Old-Age Crisis’, OxfordUniversity Press, Oxford; Dahrendorf, R.(Chairman) (1995) ‘Wealth Creation and SocialCohesion in a Free Society’, Report of theCommission on Wealth Creation and SocialCohesion, Xenogamy, London; Anson, J.(Chairman) (1996) ‘Pensions 2000 and Beyond’,Report of the Retirement Income Enquiry,London; as well as surveys of customersconducted by NatWest Bank and Coopers &Lybrand (reported in Field (1996, see ref. 17above, pp. 52–53)). Compulsory contributions areseen as one way of dealing with individualmyopia and the problem of moral hazard. Thefirst issue arises because individuals do notrecognise the need to make adequate provision forretirement when they are young. The latterproblem arises when individuals deliberately avoidsaving for retirement when they are youngbecause they know the state will feel obliged notto let them live in dire poverty in retirement.

40 Blake, D. and Burrows, W. (2001) ‘Survivor Bonds:Helping to Hedge Mortality Risk’, Journal of Riskand Insurance, Vol. 68, pp. 339–348.

41 See Blake, D. (2003) ‘Pension Schemes and PensionFunds in the United Kingdom’, Oxford UniversityPress, Oxford, Second edition, ch. 13 or Blake, D.(2000) ‘Financial Market Analysis’, Wiley,Chichester, ch. 14.

42 See, eg Fabozzi, F. and Konishi, A. (eds) (1991)‘Asset-Liability Management’, Probus, Chicago.Formally the fund manager’s objective with a DBscheme is to minimise surplus risk each periodsubject to the condition that the surplus is alwayszero. The control variables in this dynamicprogramming exercise are the contribution rate intothe fund and the composition of the assets in thefunds (ie the portfolio weights or the assetallocation). See Blake, D. (1992) ‘Modelling PensionFund Investment Behaviour’, Routledge, London,

and also they did not tend to shop around (80 percent of the survey’s respondents had little or nointerest in financial matters and 85 per cent ofrespondents who had sought advice on pensions hadused only one source). Traditional providers ofpensions (such as insurance companies) wereregarded as offering complex products that weredifficult to understand and therefore requiredadditional training by sales staff. Newer providers(such as direct-selling pension providers) wereregarded as offering pension products that were easyto understand and therefore to sell. The tax ruleswere also regarded as a major source of confusion.

29 National Association of Pension Funds (1997)‘Annual Survey’, NAPF, London; GovernmentActuary’s Department (2000) ‘Occupational PensionSchemes 1995 — Tenth Survey by the GovernmentActuary’s Department’, The Stationery Office,London.

30 National Association of Pension Funds (1997)‘Annual Survey’, NAPF, London.

31 Department of Trade and Industry Returns 1997(now known as Financial Services AuthorityReturns). The top five providers account for about60 per cent of sales.

32 Blake, D. (1999) ‘Annuity Markets: Problems andSolutions’, Geneva Papers on Risk and Insurance, Vol.24, pp. 358–375.

33 MacDonald, A. (1996) ‘United Kingdom’, inMacDonald, A. (ed.) ‘The Second Actuarial Studyof Mortality in Europe’, Groupe Consultatif desAssociations D’Actuaires des Pays des CommunautesEuropeennes, Brussels found that mortality forecasterrors of 15–20 per cent over intervals of 10 yearsare not uncommon. US studies (eg Mitchell, O. S.,Poterba, J. M., Warshawsky, M. J. and Brown, J. R.(1999) ‘New Evidence on the Money’s Worth ofIndividual Annuities’, American Economic Review, Vol.89, pp. 1299–1318; and Poterba, J. M. andWarshawsky, M. J. (1998) ‘The Cost of AnnuitisingRetirement Payouts, Working Paper’, EconomicsDepartment, Massachusetts Institute of Technology,Cambridge, MA) found that the deduction from theactuarially fair value of an annuity for a 65-year oldUS male was 15 per cent if the male was a typicalmember of the population as a whole (calculatedusing the mortality tables for the whole US malepopulation) and 3 per cent if the male was typicalof the population buying annuities voluntarily(calculated using the select mortality tables for maleannuity purchasers), implying a 12 per centdeduction for the greater mortality risk. SeeFinkelstein, A. and Poterba, J. (1999) ‘SelectionEffects in the Market for Individual Annuities: NewEvidence from the United Kingdom’, DiscussionPaper, Economics Department, MassachusettsInstitute of Technology, Cambridge, MA; which,using UK data, estimates cost loadings for 65-yearold males of 10 per cent.

34 Credit Suisse First Boston (1999) ‘Equity-GiltStudy’, CSFB, London.

35 This is currently a serious issue in the UK on


Blake

Settlements, Basel, which reports a survey of UKand US fund managers in which they acknowledgethe existence of a herding effect.

51 Pensions Management, September 1998.52 Myners, P. (2001) ‘Institutional Investment in the

United Kingdom: A Review’, HM Treasury,London, March.

53 Department of Social Security and HM Treasury(2001) ‘Security for Occupational Pensions: TheGovernment’s Proposal’, London, March.

54 Office of Fair Trading (1997) ‘Consumer Detrimentunder Conditions of Imperfect Information’,Research Paper 11, London; Office of Fair Trading(1999) ‘Vulnerable Consumers and FinancialServices’, Report 255, London.

55 The fund manager benefits by sharing someproportion of the outperfromance of the benchmarkindex; there is also a penalty for underperformance,although it comes in the form of a credit against thefuture fee rather than as a cash refund in the quarterin which the underperformance occurs.

56 Even though no evidence was found that fundmanagers could systematically outperform themarket, it would be difficult for the Government torequire pension fund managers to use indexmatching. There would be no clear consensus onwhich index to match (the FTSE100 index, theFTA All Share index, a European index or a globalindex). Also there is a risk that market inefficienciescould emerge if large institutional investors such aspension funds were prevented from searching forunder- and over-valued stocks; the author foundthat the only source of value-added in active fundmanagement was security selection.

for an analysis of UK pension fund investmentbehaviour over a period when DB schemes werebroadly unconstrained by their liabilities and hencepursued investment strategies more akin tomaximising risk-adjusted expected value.

43 For more details see Blake, D. (1998) ‘PensionSchemes as Options on Pension Fund Assets:Implications for Pension Fund Management’,Insurance: Mathematics and Economics, Vol. 23, pp.265–288.

44 Blake, D. and Timmermann, A. (1998) ‘MutualFund Performance: Evidence from the UK’,European Finance Review, Vol. 2, pp. 57–77.

45 Lunde, A., Timmermann, A. and Blake, D. (1999)‘The Hazards of Mutual Fund Underperformance’,Journal of Empirical Finance, Vol. 6, pp. 121–155.

46 Pension Schemes Registry and GovernmentActuary’s Department (2000).

47 Blake, D., Lehmann, B. and Timmermann, A.(1999) ‘Asset Allocation Dynamics and Pension FundPerformance’, Journal of Business, Vol. 72,pp. 429–447.

48 Blake, D., Lehmann, B. and Timmermann, A.(2002) ‘Performance Clustering and Incentives in theUK Pension Fund Industry’, Journal of AssetManagement, Vol. 3, pp. 173–194.

49 See Lakonishok, J., Shleifer, A. and Vishny, R.(1992) ‘The Structure and Performance of theMoney Management Industry’, Brookings Papers:Microeconomics, pp. 339–391, which discusses similarresults in the USA.

50 See Davis, E. P. (1988) ‘Financial Market Activity ofLife Insurance Companies and Pension Funds’,Economic Paper No.21, Bank for International



The UK pension system: Key issues - Springer [email protected]. scheme or to a personal or...

Documents

Transcript of The UK pension system: Key issues - Springer [email protected]. scheme or to a personal or...