ECONOMIC PAPERS EUROPEAN COMMISSION DIRECTORATE-GENERAL FOR ECONOMIC AND FINANCIAL AFFAIRS.

Transcription

1 ECONOMIC PAPERS EUROPEAN COMMISSION DIRECTORATE-GENERAL FOR ECONOMIC AND FINANCIAL AFFAIRS hp://europa.eu.in/comm/economy_finance Number 162 January 2002 EU pension reform- An overview of he debae and an empirical assessmen of he main policy reform opions by Kieran Mc Morrow and Werner Roeger*

2 EU PENSION REFORM- AN OVERVIEW OF THE DEBATE AND AN EMPIRICAL ASSESSMENT OF THE MAIN POLICY REFORM OPTIONS by Kieran Mc Morrow and Werner Roeger* * The auhors are economiss in he Direcorae-General for Economic and Financial Affairs (ECFIN) of he European Commission. Acknowledgemens : The auhors would like o hank A. Dramais, A. Brunila, M. Bui, S. Deroose, H. Oksanen, H. Jansen, M. Thiel, S. Langedijk and especially M. McCarhy, A. Salomaki and D. Cosello for valuable commens and drafing suggesions regarding earlier drafs of his paper. ECFIN/43/02-EN This paper only exiss in English. European Communiies, 2002

3 EU PENSION REFORM -AN OVERVIEW OF THE DEBATE AND AN EMPIRICAL ASSESSMENT OF THE MAIN POLICY REFORM OPTIONS - INTRODUCTORY REMARKS TABLE OF CONTENTS SECTION 1:OVERVIEW OF THE PENSION REFORM DEBATE 1.1 KEY ISSUES AND CONCEPTS 1.2 MAIN DRIVING FORCES BEHIND THE GROWING CALLS FOR PENSION REFORM : HISTORICAL ANALYSIS : WHAT WAS THE ROLE OF DEMOGRAPHIC, LABOUR MARKET AND GENEROSITY FACTORS IN DETERMINING THE PAST GROWTH OF EU AND US PUBLIC PENSION EXPENDITURE? : FUTURE EVOLUTION : WHAT ARE THE IMPLICATIONS OF AGEING FOR GROWTH AND PENSION EXPENDITURE TRENDS IN THE EU AND THE US? 1.3 WHAT ARE THE POSSIBLE POLICY SOLUTIONS? SECTION 2 : DESCRIPTION OF MODEL USED TO ASSESS THE PENSION REFORM OPTIONS 2.1 MODEL'S BASELINE SCENARIO OF THE GROWTH AND PENSION EXPENDITURE IMPLICATIONS IN THE EU OF AGEING POPULATIONS ( ) 2.2 ASSESSING THE EXPLANATORY POWER OF THE AGEING MODEL :ACOMPARISON WITH THE PENSION PROJECTIONS FROM THE EU'S ECONOMIC POLICY COMMITTEE 2.3 CRITERIA FOR EVALUATING THE EFFECTIVENESS OF THE VARIOUS PAYG AND SYSTEMIC PENSION REFORM OPTIONS ANALYSED IN SECTIONS 3 TO 5 SECTION3 : PAYG SYSTEM :ECONOMIC ASSESSMENT OF THE MAINPARAMETRIC /LABOUR MARKET REFORM OPTIONS :WHAT IS NEEDED TO BRING THE SYSTEM BACK INTO EQUILIBRIUM? 3.1 CHANGES TO THE GENEROSITY OF THE PAYG SYSTEM 3.2 AN INCREASE IN THE EFFECTIVE RETIREMENT AGE 3.3 BROAD PACKAGE OF "PAYG" REFORMS :LABOUR MARKET + GENEROSITY + RETIREMENT AGE CHANGES SECTION 4: SYSTEMIC REFORM :2KEY FACTORS TO BE CONSIDERED -INTERNAL RATES OF RETURN +TRANSITION BURDEN 4.1 FORECASTS FOR THE INTERNAL RATE OF RETURN OF THE PAYG + FUNDED PENSION SYSTEMS : (BASELINE SCENARIO +SENSITIVITY ANALYSIS) 4.2 FORECASTS FOR THE TRANSITION BURDEN IN 2000 AND ITS EVOLUTION TO 2050 (BASELINE SCENARIO +SENSITIVITY ANALYSIS) SECTION 5:ECONOMIC ASSESSMENT OF A FULL /PARTIAL SHIFT TO FUNDING % SHIFT TO FUNDING :COMPULSORY SAVINGS OPTION % SHIFT TO FUNDING :VOLUNTARY SAVINGS OPTION 5.3 PARTIAL SHIFT TO FUNDING +STABILISATION OF PAYG SYSTEM SECTION 6:AN "OPTIMAL" EUPENSION REFORM STRATEGY 6.1 PREFERRED SCENARIO :PROGRESSIVE 100% SHIFT TO FUNDING USING A TWO STAGE OPTIMAL TRANSITION PATH 6.2 "OPTIMAL" STRATEGY MAXIMISES THE GROWTH RATE EFFECTS FROM PENSION REFORM (+ FUNDING OFFERS THE POSSIBILITY OF ADDITIONAL GROWTH GAINS) 5

4 SUMMARY OF MAIN POINTS +KEY POLICY CONCLUSIONS REFERENCES ANNEXES ANNEX 1: DETAILED DESCRIPTION OF MODEL ANNEX 2: PREFUNDING SIMULATION :INTERGENERATIONAL EQUITY OBJECTIVE 6

5 INTRODUCTORY REMARKS Over he las decade here has been a noiceable surge in ineres in all of he developed economies in boh he poliics and economics of pension reform. While concerns have always been presen regarding he basic raionale underpinning pension sysems, his recen deepening in ineres has been driven essenially by demographics and he economic and budgeary implicaions of ageing populaions. Ageing raises serious quesions regarding he long erm susainabiliy of he presen PAYG public pension sysem as well as heighening he degree of ineres in he efficacy of he alernaive, non-payg, pension financing mechanisms which are available, mos noably funding. In his regard, here has been a discernible rend in recen years in he pensions debae owards advocaing greaer recourse o capial marke soluions as a way of resolving he pensions crisis alhough a lo of work has also been carried ou in assessing he possibiliies for reform wihin he PAYG sysem iself. This paper will examine he basic case for pension reform, will presen empirical evidence regarding he poency of he various pension reform opions and will address he issue of wheher PAYG or funding is he mos appropriae way of providing reiremen income. As menioned above, a he hear of his whole pension reform debae are demographics, wih a growing awareness amongs policy makers of he many effecs which grey pressure will have, a a global level, progressively over he coming decades. Any objecive assessmen of he likely evoluion of he key underlying deerminans of populaion changes over he nex 50 years suggess ha ageing of he EU's populaion, as wih mos of he developed world, is an inescapable fac, due o he progressive lenghening in life expecancy and he fall in feriliy raes o below he criical hreshold levels required for generaional renewal. The share of elderly people in he overall populaion is presenly of he order of 15 percen in he EC, US and Japan. According o he laes demographic projecions, his share is likely o almos double beween now and 2050 in he case of he EU and Japan, while growing more modesly in he US. While he share of he elderly in he oal populaion also grew over he las number of decades, he overall economic burden on hose in work was aenuaed by he fac ha he populaion of working age was also growing rapidly. This laer luxury of growing numbers enering he labour force, which governmens in he pas could urn o in order o fund he addiional pension and oher age-relaed expendiures is fas disappearing, wih he presen EU raio of four workers per pensioner likely o be cu in half by This change in he old age dependency raio is paricularly problemaic for he PAYG pension sysem since he pensions of he presen generaion of over 65s 1 is paid for hrough he social securiy conribuions of he presen workforce. Consequenly, hese win developmens, i.e. growing shares of he over 65s in he populaion allied o declining numbers in he age groups which radiionally suppored he non-economically acive age cohors, are a he hear of he growing ineres in pension sysems and pension reform. This paper will examine he key hemes which permeae his debae, whils a he same ime aemping o boh quanify he size and naure of he problems o be faced and analysing he respecive meris of he main policy soluions which have been pu forward in he lieraure. While he focus is on 1 65 is used simply for illusraive purposes since he effecive reiremen age in he EU is closer o 60. 7

6 he EU, a comparison of he EU's pension sysem wih ha of he US is also provided and he analysis of he various policy choices is of course applicable o more han jus he Communiy siuaion. In erms of conen, his paper should in fac be seen as a logical follow-up o a paper on "The Economic Consequences of Ageing Populaions" 2 which was published a he end of 1999 and which includes a lo of he basic background research used for he presen analysis. I is no he inenion of he presen paper o repea he ground covered in he earlier work bu references o he laer will be made where appropriae. In addiion, since he primary focus of his paper is on he empirics of pension reform, very lile aemp is made o discuss he more normaive aspecs of he debae. These laer aspecs are discussed in deail in chaper 5 of he European Commission publicaion, "The EU Economy : 2001 Review", which was released by he Direcorae General for Economic and Financial Affairs in November The presen paper is srucured as follows: Secion 1 gives an overview of he pension reform debae, including an analysis of is key underlying driving forces. Much of he laer analysis is based on a comparaive assessmen of he pas and fuure evoluion of he EU and US pension sysems. Secion 2 goes on o describe he model which is used o quanify he effecs of he various pension reform opions, wih an assessmen also being made of is explanaory power by comparing is projecions for fuure pension expendiure wih he equivalen projecions recenly produced by he EU's Economic Policy Commiee (EPC). Secions 3 o 5 provide he resuls of a series of simulaion exercises which sysemaically assess he various parameric and sysemic policy opions available o governmens, wih all he opions being evaluaed relaive o heir growh, budgeary and income disribuion effecs. In addiion o assessing he effecs of a broad combinaion of parameric reforms of he PAYG sysem, secion 3 also looks a he effecs of individual policy iniiaives including changes in he generosiy of he PAYG sysem and increases in he effecive reiremen age. Given he srains which are emerging in he PAYG sysem, secion 4 sars he analysis of he main alernaive o his laer sysem i.e. capial marke funding, and examines he essenial facors o be considered in any sysemic change, such as he inernal rae of reurn differences beween he PAYG and funded sysems and he ransiion coss of moving from he former o he laer. Secion 5 goes on o provide a series of sysemic simulaions ranging from a 100% shif o funding, including compulsory and volunary savings varians, o a parially funded pension sysem wih shares of 75% for he PAYG sysem and 25% for funding. In he final secion of he paper, an aemp is made o disil he key policy messages o be reained from he analysis, wih paricular aenion being devoed o he need for an "opimal" EU pension reform sraegy based in paricular on a progressive 100% shif o funding. Given ha his ransiion process will have o occur over more han one generaion, policy makers mus handle he ransiion pah in as effecive a way as possible. In his regard, during he ransiion phase 2 See Mc Morrow and Roeger (1999) 8

7 governmens mus give prioriy o hose policy measures aimed a offseing he growh loss associaed wih ageing raher han on he budgeary pressures of he laer phenomenon, alhough boh policy goals are, of course, closely inerrelaed. The paper finishes wih a deailed overview of he main poins and key policy conclusions o be reained from he sudy. OUTLINE OF STUDY EU PENSION REFORM GENERAL OVERVIEW OF PENSION REFORM DEBATE (SECTION 1) EMPIRICAL EVALUATION OF PENSION REFORM OPTIONS (SECTIONS 2-5) PREFERRED EU PENSION REFORM STRATEGY (SECTION 6) 1. KEY ISSUES+CONCEPTS 2. COMPARISON OF EU + US PENSION SYSTEMS 3. POLICY SOLUTIONS 1. MODEL DESCRIPTION 2. PAYG REFORMS 3. SYSTEMIC REFORMS % SHIFT TO FUNDING 2. TWO STAGE OPTIMAL TRANSITION PATH 3. PUBLIC + PRIVATE PILLARS 9

8 SECTION 1: OVERVIEW OF THE PENSION REFORM DEBATE 1.1 KEY ISSUES AND CONCEPTS For hose coming o he pension reform debae for he firs ime he erminology used can ofen appear a lile confused, wih Table 1 below aemping o isolae he key issues and conceps which need o be kep in mind by he reader. TABLE 1:KEY ISSUES AND CONCEPTS MAIN QUESTIONS TO BE ADDRESSED ESSENTIAL POINTS TO BE RETAINED 1. WHAT ARE THE MAIN SYSTEMS FOR FINANCING CONTRIBUTORY PENSION INCOME? (NON-CONTRIBUTORY PENSIONS ARE NOT CONSIDERED AS THEY ARE PURELY REDISTRIBUTIVE AND ARE FINANCED FROM GENERAL TAXATION) PAYG V FUNDING : The key issue in erms of conribuory pension sysems is he choice beween he relaive meris of an unfunded sysem such as Pay-As-You-Go (PAYG) and a funded sysem and no as some commenaors would like o sugges beween public and privae provision. Funding can be carried ou in boh he public and privae secors or using a combinaion of boh secors. 2. HOW ARE THESE SYSTEMS FINANCED? 3. WHAT DETERMINES THE LEVEL OF PENSION ONE FINALLY RECEIVES ON RETIRING AND IS THIS DEGREE OF GENEROSITY GUARANTEED OR NOT? CONTRIBUTION RATES : The PAYG and Funded pension sysems are financed simply by financial conribuions from workers : which are channelled via he ax sysem (Social Securiy Conribuions - SSC's) in he case of he PAYG sysem and in he case of funding via regular pension conribuions o various ypes of pension funds. REPLACEMENT RATES +DEFINED BENEFIT (DB) VDEFINED CONTRIBUTION (DC) SCHEMES : The level of ones pension is referred o as he replacemen rae (he % share of ones final salary ha one receives in reiremen). The replacemen rae in he PAYG sysem is of he defined benefi (DB) or "guaraneed" variey and is deermined by a large number of globally esablished facors including he conribuion rae, he number of conribuion years, he indexaion and eligibiliy rules applied ec. Noe ha whils he PAYG's replacemen rae is in heory guaraneed his has no sopped governmens periodically changing key parameers of he PAYG sysem so ha generosiy levels have changed over ime - in a posiive direcion over he las 40 years bu poenially in he opposie direcion over he coming decades as he effecs of ageing sar o "bie". In he case of funding he replacemen rae is increasingly of he defined conribuion (DC) variey (alhough DB schemes are also possible) where he final benefi is no guaraneed and depends on he level of conribuions made and he rae of reurn achieved by he paricular pension fund chosen. 4. IS FUNDING INHERENTLY A CHEAPER FORM OF FINANCING THAN PAYG AND, IF IT IS, WHY IS IT NOT EXCLUSIVELY USED FOR PENSION FINANCING? INTERNALRATES OFRETURN FOR BOTH SYSTEMS : The ulimae cos of any sysem is deermined by he rae of reurn achieved on he conribuions made hroughou ones working life. I is acceped in his paper ha for he average worker ha he inernal rae of reurn of he PAYG sysem is lower (and consequenly he equilibrium conribuion rae is higher) han ha of funding and his is wha one should expec 10

9 in a dynamically efficien economy ha he rae of reurn on capial (i.e. he real ineres rae) is higher han ha of labour (i.e. real wage growh) 3. However, here are a number of oher very imporan facors o be considered in he ulimae choice of sysem such as he fac ha firsly, while he average reurn from funding is higher, he volailiy of hose reurns is also greaer (higher reurn is, in heory a leas, always mached by higher risk); secondly he issue of he ransiion coss of moving from PAYG o funding, and finally a series of more "nebulous" argumens such as he need for balance in he economy as a whole beween human and physical capial invesmen and balance for he individual in erms of a pruden risk diversificaion sraegy (nebulous in he sense ha boh he PAYG and funded sysems can be designed o ensure ha adequae accoun is aken of hese laer facors). 5. HOW IS AGEING LIKELY TO AFFECT BOTH TYPES OF PENSION SYSTEMS AND WHAT SHOULD BE DONE ABOUT IT? 6. WHAT ARE THE BASIC ECONOMICS OF PENSION REFORM? BOTH SYSTEMS ARE EXPOSED TO RISK : Ageing will affec boh sysems of pension provision in a number of differen ways including: he effecs of "greying" populaions on economic growh andconsequenlyonheraesofreurnofbohsysems; he need o offse he effecs of shrinking labour forces and growing dependency raios on SSC's and on pension fund conribuion raes hrough a combinaion of muually reinforcing labour marke and pension sysem reforms; he growing pressure o srenghen he financial basis of he PAYG sysem by linking conribuions and benefis more closely and hereby allaying fears ha he iniial insurance raionale for he PAYG sysem has no given way o an excessively redisribuive agenda, wih he risk from he laer being ha SSC's would increasingly be seen by workers no as pension conribuions bu as simply anoher form of axaion; he need o assess he financial marke risks associaed wih funding and he requiremen for pruden risk diversificaion sraegies and financial marke reforms; policy makers will also need o address he budgeary implicaions of ageing and he need for a fair sysem of inergeneraional burden sharing; finally, ageing will demand a re-assessmen of fundamenal life choices a he individual level regarding he ime o be spen working and he sandard of living one aspires o in reiremen. Pension sysems will have o be adaped o allow for a sufficien degree of flexibiliy for individuals o make hose choices and o see he financial effecs of heir decisions. CASE FOR PENSION REFORM RESTS ON SPILLOVER EFFECTS : A primary economic case for pension reform, on he basis of sandard neoclassical heory, is a difficul one o make since while a funded sysem is in principle superior, here is however no pareo-opimal move 3 Feldsein and Liebman (2001) "...he exisence of a capial sock implies ha individuals could insead (of PAYG) finance heir reiremen by saving and invesing in acual capial goods where hey would earn a real reurn of р. In a dynamically efficien economy, he real rae of reurn р mus exceed he rae of growh of he economy, у (Cass, 1965). Thus each working generaion incurs a loss because i receives a reurn у on is Social Securiy axes ha is less han he reurn р ha i would earn by invesing hose funds in he capial sock" 11

10 from a PAYG o a funded sysem because of he ransiion coss, wih one generaion a leas having o suffer a loss in he ransiion process (i.e. lower consumpion / higher savings by eiher individuals or governmens). Consequenly, he case for welfare enhancing pension reform is normally made on he basis of posiive secondary or spillover effecs from reforming paricular aspecs of a counry's pension sysem, on he efficiency of he economies labour and financial markes; on he budgeary gains or finally on he savings and overall growh performance of he economy. The scope for welfare enhancing reform is large since he PAYG and funded pension sysems are each quie disinc, wih significan differences linked o he basic design of, and incenives buil ino, he respecive sysems. The key economic quesion herefore is which pension sysem or combinaion of sysems is likely o have he greaes effec on promoing savings, providing work incenives and generally enhancing overall growh prospecs in a counry which is faced wih a dramaic ageing of is populaion. The acual pension sysem a counry ends up wih reflecs no only he above economic consideraions bu also a hos of counry-specific disribuional and poliical prioriies. 7. WHAT IS THE PRIMARY FOCUS OF THE PRESENT PAPER? ECONOMIC, BUDGETARY AND POLITICAL SUSTAINABILITY OF PENSION REFORM : This paper's primary focus is on he economics of pension reform and especially he growh and budgeary consequences of he various pension reform opions, in he face of unprecedened demographic change. Poliical susainabiliy, in he form of he income disribuion effecs of he reforms se in place, is also considered. In erms of pension sysems, he paper assesses he key issues o be considered in any shif from PAYG o funding, namely he inernal raes of reurn of boh sysems and he ransiion burden. The insighs from his analysis are hen used o pu forward an "opimal" EU pension reform sraegy. 1.2 MAIN DRIVING FORCES UNDERLYING THE GROWING CALLS FOR PENSION REFORM The pension reform debae is clearly being driven by a large number of inerrelaed economic facors bu one of he mos significan and visible is he growh in he cos of reiremen income provision over he las 40 years and he prospec of a massive increase over fuure decades as he effecs of ageing really sar o emerge. In his regard public pension expendiure as a percen of GDP has risen in he EU from abou 6% in 1960 o over 12% 4 in 2000 and on he assumpion ha no acion will be aken o address his siuaion, pension spending could reach unsusainable levels close o 20% of GDP 5 in he coming decades. Linked o his alarming expendiure oulook are 4 Pension expendiure figures are aken from OECD sources due o he fac ha a longer ime series is available. The figure of 12% for 2000 for he EU is however in keeping wih he roughly equivalen figures from Eurosa's ESSPROS daabase for he lae 1990's. I is imporan o sress ha due o differen definiions and conceps used here are someimes large discrepancies beween differen published sources, wih for example he EPC's Ageing Working Group esimaing ha pension expendiure in he EU in 2000 was of he order of 10 ½% of GDP. 5 This figure of 20% assumes ha he saring poin for pension expendiure as a % of GDP is he OECD figure of 12% in 2000 and no he 10 ½% of GDP assumed by he EPC. 12

11 of course economic growh consideraions wih obvious concerns being expressed regarding he negaive effecs of hese developmens for fuure changes in he living sandards of ciizens. This secion will look a he main facors which have been driving changes in pension spending over he pas number of decades and which will undoubedly deermine changes in he fuure. These facors are as follows: PURE DEMOGRAPHICS : by his is mean changes in feriliy raes and life expecancy which impac on he size of he working age and pensioner populaions in an economy. LABOUR MARKET INFLUENCES : This broad heading looks a all he facors which influence he share of he working age populaion which end up as effecive conribuors o he PAYG pension sysem - namely he numbers in employmen who pay he social securiy conribuions needed o pay he pensions of he presen reired populaion. GENEROSITY OF PENSION SYSTEM : This is he final facor which deermines he overall level of pension expendiure in an economy and is ofen proxied by he gross replacemen raio. There are wihou doub a lo of oher facors in he pension sysem which should be aken ino accoun in any deailed assessmen of expendiure rends, such as eligibiliy and indexaion rules, bu hese are no direcly allowed for in he calculaions which are given in his secion. For he presen exercise, i is sufficien o have a rough breakdown of pension expendiure over he las 40 years and his can be carried ou using simple relaionships which isolae he main driving forces in erms of expendiure changes. Pensions expendiure over ime is essenially driven by changes in he economic dependency raio (numbers in employmen / numbers of pensioners) and adjusmens o he generosiy of he sysem. In effec he economic dependency raio is deermined by changes in he old age dependency raio and changes in employmen, wih generosiy developmens being measured by modificaions o he gross replacemen raes of he respecive EU and US pension sysems : HISTORICAL ANALYSIS WHAT WAS THE ROLE OF DEMOGRAPHIC, LABOUR MARKET AND GENEROSITY FACTORS IN DETERMINING THE PAST GROWTH OF EU AND US PUBLIC PENSION EXPENDITURE? I is imporan o sress again a he ouse ha his analysis of he facors which have been responsible for he increase in pension spending over he period is purely illusraive, wih definiive breakdowns ino he hree componens of pure demographics, labour marke influences and generosiy no being feasible given he large array of influences which are involved and because of obvious daa consrains. Despie his reservaion he breakdown provided is, in he opinion of he auhors, a reasonable picure of he main influences a work, wih an idenical mehodology being applied o boh he EU and US. 1. DEMOGRAPHIC INFLUENCES : The main poins o be reained from graph 1 are as follows : 13

12 Wih regard o feriliy raes i is clear ha a significan break occurred in boh he EU and he US around , wih he numbers in he 0-14 year old age group declining significanly in boh areas following srong growh in he 1960's. However, while his negaive paern has persised in he EU in he period up o 2000 here has been quie a significan recovery in he youh populaion in he US in he 1980's and early nineies. In erms of life expecancy a birh, boh areas have experienced a seady upward movemen in longeviy since he 1960's wih life expecancy of around 70 in 1960, growing by 7 o 8 years in boh he US and EU respecively up unil he year Regarding he populaion of working age he US has always had a higher growh rae, wih increases in excess of 1 ½ % on an annual average basis in he 1960's and 1970's and close o 1% annually in he 1980's and 1990's. The EU on he oher hand has experienced a much slower pace of increase in he working age populaion wih he average over he las 4 decades being of he order of ½ a % annually. In erms of he number of pensioners here has been a seady upward shif in he share of pensioners in he overall populaion, wih he share of he over 65s in he oal populaion growing from 9% o 13% of he oal in he US over he period and from 11% o 16% in he EU over he same period. Finally, wih regard o he old age demographic suppor raio, which is simply he inverse of he dependency raio, here has been a significan change over he las number of decades in he number of people in he populaion of working age relaive o he number of people over 65. In 1960 he demographic suppor raio was 6.5 and 6.1 in he US and EU respecively, by 2000 his raio of people aged relaive o people aged 65 and over had fallen o 5.3 and 4.1 in he US and he EU respecively. Clearly herefore in erms of pure demographics he EU has been more severely affeced over he las 40 years by falling birh raes and working age populaion changes han has he US. GRAPH 1:PURE DEMOGRAPHICS 2 Annual % Change Growh Rae of 0-14 Year Olds (EU15 V US) US 80 Years Life Expecancy (EU15 V US) EU US EU While a large array of facors have been pu forward as possible explanaions for his break in rend, he growing use of he conracepive pill was undoubedly one of he conribuory elemens. 14

13 2 Annual Average % Change US C : Populaion of Working Age (EU15 V US) 7 Raio D : Demographic Suppor Raio (Populaion of Working Age relaive o Pensioners) (EU15 V US) US EU15 5 EU 's 1970's 1980's 1990's LABOURMARKET INFLUENCES: The demographic suppor raio only provides par of he sory in erms of pension expendiure pressures in he EU and he US. The demographic raio doesn accuraely reflec he economic burden on he acive proporion of he populaion of working age (i.e. he labour force) and in paricular on hose acually in employmen since i is only hose wih jobs which are financing, hrough heir social securiy conribuions, he governmen's pension ransfers o he non-acive pensioner populaion. I is clear ha his laer economic dependency burden on curren labour income is much heavier for regions or counries wih low employmen raes and graph 2 shows clearly ha his is paricularly he case wih he Communiy relaive o he US. The main poins o be reained from he graph are as follows : In erms of paricipaion raes, he las 40 years has seen very lile change in he EU in he proporion of he working age populaion which is acively paricipaing in he labour force. By conras, he US has experienced a 13 percenage poins increase in paricipaion raes rising from 71% in he 1960's o 84% in he 1990's. This has resuled in annual average growh raes of he US labour force of 1 ½ - 2% in each of he las 4 decades compared wih raes of ½ a % a year in he EU over he same period. Coupled wih large increases in labour force growh, he US also experienced roughly equivalen raes of growh in employmen since he rae of srucural unemploymen in he US over he las 40 years has been reasonably sable a 5-6% of he labour force. As agains his he EU performance on all couns looks raher dismal, wih poor growh in he populaion of working age being compounded by virually no change in labour force paricipaion raes and an increase in srucural unemploymen from around 2% of he labour force in he 1960's o close o 10% a he end of he 1990's. In addiion, European workers appeared o make greaer recourse o reiring early over he period wih he effecive reiremen age falling in he EU by 5 ¼ years o less han 60 over he period compared wih a fall of 3 ½ years in he US o 63 over he same period. Given he preceding commens i is hardly surprising o find ha he employmen performance in he EU compared wih he US over he las 40 years has ranged from less han a ¼ of a % growh in he 1960s, on an annual average basis, o ½ a % in he 1980's and 1990's compared wih annual raes of growh which ranged from 1½-2% a year in he US over he same period. 15

14 Wih such a weak labour marke performance in he EU compared wih he US, changes in he economic suppor raio in boh areas have evolved in very differen ways. As graph 2 shows, he economic suppor raio did no change very much in he US over he las 40 years, saying a roughly 4 workers per pensioner - in fac i improved marginally in he 1990's. Over he same period he EU raio deerioraed from having a broadly similar raio o he US in 1960 of 4 workers o 1 pensioner o a raio of 2 ½ workers supporing 1 pensioner in Since pension expendiure as a proporion of GDP is essenially made up of he economic suppor raio plus changes in generosiy one can quickly see ha he weak EU labour marke performance over he las number of decades has no aced o offse he pure demographic changes coming from reduced birh raes and increased life expecancy and has consequenly led o sharp increases in he pension burden. In conras o his he US labour marke has fully offse he pure demographic effecs so ha no upward pressure has occurred in erms of US pension spending from he economic dependency raio. GRAPH 2:LABOUR MARKET INFLUENCES A : Labour Force Paricipaion Raes B : Effecive Reiremen Age 90 % of Populaion of Working Age 69 Years US US EU15 65 EU 's 1970's 1980's 1990's Annual Average % Change C : Employmen 5 Raio D : Economic Suppor Raio (Number of Workers per pensioner) US EU US EU 's 1970's 1980's 1990's GENEROSITY OF THE SYSTEM : The las piece of he pension expendiure sory comes from changes in he generosiy of he respecive pension sysems in he EU and he US. This is undoubedly he mos difficul calculaion o be made and he one 16

15 around which mos conroversy revolves. Graph 3 below shows he OECD's synheic indicaor for he gross replacemen rae applying in he EU and he US over he period I is an admiedly crude aemp o rack he changes in he generosiy of he respecive sysems bu neverheless can be broadly corroboraed using evidence from oher daa sources a he naional level. Wha he graph is simply saying is ha he generosiy of he EU and US public pension sysems have boh improved over he las number of decades, wih US pensioners gaining relaively more as shown by a narrowing of he generosiy gap wih he EU public pension sysem from nearly 14% poins in 1960 o 6% poins in he mid 1990's. GRAPH 3:GENEROSITY OF PENSION SYSTEM % Gross Replacemen Rae (OECD Synheic Indicaor) EU15 US COMBINED IMPACT OF DEMOGRAPHIC + LABOUR MARKET + GENEROSITY CHANGES ON OVERALL PENSION EXPENDITURE : The final graph in his secion brings ogeher he informaion gahered regarding all he demographic, labour marke and generosiy changes which have occurred since 1960 and presens he oal change in pension expendiure over he las 40 years 7 and he proporion of he change which can be roughly aribuable o he facors menioned. In erms of overall pension expendiure changes, using he broades definiion of pension spending which includes old age, survivors and disabiliy pensions, Graph 4 shows ha he increase in he oal public expendiure on pensions was 3 ¼% poins of GDP in he case of he US and 6 ¼% poins in he EU over he period In erms of he driving facors behind hese overall changes i is clear ha if he EU had experienced he same labour marke performance as he US i would have resriced he change in oal pension expendiure o ha which occurred in he US. In he US since he pension spending pressures emanaing from pure demographic facors were neuralised by favourable labour marke changes he only facor forcing up US pension expendiure over he period was changes in he generosiy of he sysem. In he case of he EU, demographic, labour marke and generosiy changes all conribued o he large increase in pensions spending. The Communiy s old age economic dependency raio in he year 2000 of 38% (i.e. equal o a suppor raio of 2 ½ workers per pensioner) compares wih a raio of less han 26% in he US (4 workers per pensioner). This 12 % poins difference 7 The 1995 generosiy levels are simply exrapolaed forward o 2000 o allow a comparison for all hree ses of variables (demographic, labour and generosiy) over he full 40 year period

16 in he ransfer burden facing EU workforces, relaive o hose in he US, mus of course a some poin reflec hemselves eiher in higher axes / social securiy conribuions in order o finance he addiional age relaed ransfers or in a lowering of benefi paymens o recipiens. The laer choice beween increased axes or reduced benefis is a difficul one and from he scenarios presened laer on in his paper is one which will increasingly have o be made over he nex number of decades. Such choices are, however, more easily made, and he solidariy beween he working and dependen populaions is poenially less srained, when he financing of he ageing burden can be spread over a greaer number of workers. This is he case in he US because of heir higher labour force paricipaion raes and lower srucural unemploymen raes compared wih here in Europe. These differences in he respecive performances of he US and he EU on he pension spending side were of course refleced in erms of changes on he revenue side of public pension sysems. The social securiy conribuions imposed on US workers rose from 7% of wages in he 1960's o 12 ½% in he 1990's (wih a 50/50 spli beween employers and employees). In he case of he EU, social securiy conribuions rose from 11 ¼% of wages in he 1960's o around 19% 8 in he 1990's. Finally, i is imporan o reierae ha hese pension expendiure figures only refer o he public PAYG sysem in he EU and he equivalen social securiy sysem in he US. In he case of he EU nearly 90% of reiremen income provision comes via he PAYG sysem 9 wih he remainder coming from various funding pillars. In he case of he US, he breakdown beween social securiy and funding is roughly 2/3 o 1/3 respecively. In erms of he oal income of pensioners, he PAYG sysem provides roughly 55-60% of he oal income of pensioners in he EU, wih he US social securiy sysem conribuing roughly 40% %ofgdp GRAPH 4:PENSION EXPENDITURE IN EU AND US : A:EU15VUS Pension Expendiure EU %ofgdp B : Conribuions o he change in EU15 + US Pension Expendiure EU EU15 US US 6 4 US 2 0 US EU15 US EU Demographics Labour Marke Generosiy Toal Change 8 Noe: This figure for 2000 of 19% is aken from OECD sources and is higher han he figure of 16% used in he simulaions in Secion 2. The reason for he difference is explained by he fac ha he model used for he simulaions is calibraed on he EPC pension expendiure figure of 10 ½% of GDP in 2000 whereas his figure of 19% is based on he higher OECD esimae for pension expendiure of 12%. 9 This breakdown beween PAYG and funding only refers o he EU average wih big variaions in he respecive shares of he individual EU counries. 18

17 1.2.2 : FUTURE EVOLUTION WHAT ARE THE IMPLICATIONS OF AGEING FOR GROWTH AND PENSION EXPENDITURE TRENDS IN THE EU AND THE US? In addiion o he insighs provided by an hisorical evaluaion of he main driving facors behind he growh in pension expendiure over he las number of decades, an undersanding of fuure pressures in his area is also imporan in erms of he pension debae. A recen join EPC / OECD exercise was carried ou o assess he budgeary consequences of ageing populaions over he period o 2050 and he resuls of his exercise are given in Graph 5. This graph shows clearly ha ageing populaions will exer srong upward pressure on public expendiure over he nex number of decades in boh he US and he EU bu he effecs will be subsanially greaer in he EU. If one looks a he purely demographic facors one sees ha his facor alone will boos EU pension expendiure by 6½% poins compared wih 2½% poins in he US. While his large demographic pressure in he EU is forecas o be offse by declines in generosiy and some labour marke improvemens, i is clear ha he EU is sill facing much greaer pension expendiure pressures over he nex 50 years compared wih he US. GRAPH 5:CHANGE IN PENSION EXPENDITURE IN THE EU + US ( )* 8 6 %ofgdp EU Conribuions o he change in Pension Expendiure EU + US ( ) 4 2 US EU EU US 0-2 EU US US US -4 Dependency Raio Employmen Raio EU Benefi Raio Eligibiliy Raio Toal Change Source : EPC, OECD * Relaive o Graph 4B he breakdown of pension expendiure in he presen graph can be roughly compared as follows : dependency raio = demographics; benefi raio = generosiy; employmen + eligibiliy raios = labour marke. Finally, i may be useful o recall he conclusions of Mc Morrow and Roeger (1999) wih regard o he overall economic impac of ageing populaions on he economic growh performance of he EU and US over he coming decades. The simulaions underpinning he laer analysis were carried ou using he Commission services QUEST II model and suggesed ha per capia living sandards in he EU and he US were likely o fall significanly over he nex 50 years due o he direc influence of he ageing process. The effec of ageing populaions in erms of slowing he rae of growh of poenial oupu would also make he budgeary implicaions of ageing more difficul for he individual economies o bear. If he scenarios prove accurae, economic agens in Europe could be looking a roughly an annual half poin reducion 19

18 in poenial growh raes from he presen 2 ¼ percen rae of growh o an average rae of 1 ¾ percen over he period , represening a cumulaive GDP loss of nearly 20 percen. Cumulaive reducions of 10% are prediced for he US, which would be equivalen in growh rae erms o an annual average loss of around a ¼ of a % poin over he nex 50 years. While i can be argued ha he laer no policy change simulaion may be unrealisic in ha governmens are unlikely o sand idly by, i neverheless gives an idea of he scale of he ask faced by policymakers in devising policy measures aimed a avoiding, or a leas cushioning, he poenial shock o people's living sandards over he coming decades. GRAPH 6:GROWTH EFFECTS OF AGEING POPULATIONS ( ) 5 % IMPACT OF AGEING ON THE LEVEL OF GDP IN THE EU AND THE US ( % Change Relaive o Baseline ) US EU Source : Mc Morrow and Roeger (1999) In conclusion, herefore, he EU will face a more difficul ime over he coming decades compared wih he US in coming o erms wih he implicaions of a significan ageing of is populaion wih regard o boh he public pension expendiure implicaions and in erms of is impac on overall sandards of living. When one looks a he acual numbers involved one can more easily grasp he magniude of he challenge o be faced by he EU. Over he pas 40 years he number of people aged 65 and over in he EU increased by roughly 27 ½ million. This, however, did no pose any major economic problems since he working age populaion rose by subsanially more (47 million) and easily suppored he addiional economic burden. The nex 50 years will see a dramaic urnaround in hese rends, wih he number of over 65s growing by an addiional 42 million bu wih he working age populaion acually declining by 41 million. By comparison, while he US will winess similar absolue increases in is over 65s populaion o ha of he EU (i.e. 41 million), unlike he EU i will no experience a decline in he numbers in he working age populaion - in fac i is projeced o winess a growh of 30 million in his age caegory. On he basis of hese laer absolue numbers one can more readily comprehend he dauning naure of he challenge which his ageing burden places on he EU's economic sysem relaive o ha of he US. Given ha he economic and budgeary scale of he ageing problem looks likely o be so much greaer in he EU han in he US, he res of his paper will focus exclusively on he possible policy soluions available o EU governmens o ackle he pension expendiure and growh implicaions of his ageing phenomenon, alhough a large number of he possible soluions will be equally applicable o he US siuaion. 20

19 1.3 WHAT ARE THE POSSIBLE POLICY SOLUTIONS? ESSENTIAL POLICY PROBLEM : Wihou reform, ageing will dramaically increase he coss of EU public PAYG pension schemes, wih he financing of he addiional spending pushing he burden of social securiy conribuions on workers o levels approaching 30% of heir wages. Furhermore, ageing poses no only specific susainabiliy concerns for he public pension sysem bu also, and more imporanly, serious worries regarding he overall economic growh impac on economies. BROAD OVERVIEW OF MACRO POLICY SOLUTIONS : Policy acion o address he growh and budgeary implicaions of ageing populaions needs o be direced a wo key objecives, in order of prioriy : 1. ADDRESSING THE NEGATIVE ECONOMIC GROWTH IMPLICATIONS OF AGEING THROUGH THE FOLLOWING MEASURES : A: BOOSTING SAVINGS AND CAPITAL ACCUMULATION : Savings and invesmen have a poenially imporan role o play in helping o offse he effecs of rising dependency burdens, wih higher savings leading o higher producive invesmen and higher long-run growh. B: ACTION TO EXPAND THE EFFECTIVE LABOUR SUPPLY : Policies o increase he workforce and o reduce non-cyclical unemploymen would be an appropriae response o he adverse demographic developmens and would bring a riple gain : an increase in poenial oupu; a reducion in public expendiure on he elderly, due o he slower rae of increase in he effecive dependency raio; and higher ax revenues. Possible acions in his area 10 are: firsly, exending he working lifeime beyond he presen effecive reiremen age of less han 60, a leas o he sauorily imposed limi of 65 or even o go furher by linking he sauory age o changes in life expecancy; and secondly, labour marke reforms aimed a raising paricipaion raes and ackling srucural unemploymen. C: STRUCTURAL REFORM INITIATIVES AIMED AT OFFSETTING THE EFFECTS OF AGEING VIA PRODUCTIVITY IMPROVEMENTS : In he case of he EU, he policy opions available o permanenly raise he rae of growh of TFP are geing fewer and fewer. The EU has made enormous progress over he las number of decades in puing in place he policy sraegies which are mos ofen linked in he lieraure wih he reaping of subsanial dynamic gains such as a sable macroeconomic environmen (i.e. EMU) and an open rading environmen, boh inernally (i.e. he single marke programme) and exernally hrough he World Trade Organisaion (WTO). While hese openness and low inflaion effecs on growh raes have already been largely reaped, he EU can sill coninue o derive saic gains from regulaory reform o improve he funcioning of produc, labour and capial markes. In he case of pension reform, financial marke resrucuring is a paricularly imporan area for acion. 10 While increased immigraion is anoher possibiliy for expanding he labour force, i is no regarded as a poen policy arge because of he pracical limis o such a policy due o he poliical sensiiviies involved, no only in he hos counries bu also in he migrans home counry because of he likely "brain drain" impac of immigraion programmes geared o aracing highly skilled poenial migrans. Furhermore, given he magniude of he ageing crisis, normal immigraion paerns could a bes only provide a small proporion of any overall soluion. 21

20 2. EFFECTIVELY DEALING WITH THE PUBLIC EXPENDITURE PRESSURES OF AGEING : IN THIS REGARD POLICIES MUST BE EVALUATED AGAINST 3 OBJECTIVES A: Are policy reform acions posiive from an economic growh perspecive? The higher he growh rae he greaer he degree of budgeary relief achieved. B: How effecive are he individual policy measures in puing he public finances on a susainable fooing? C: How susainable are he policies which have been se in place in a long run conex? Are here big changes implied in he relaive income posiion of pensioners and of he working age populaion? Is such a reform pah poliically susainable given he ineviable growh in he poliical power of he reired populaion over he coming decades? SPECIFIC PENSION POLICY SOLUTIONS : As can be easily garnered from he above discussion regarding macro policy soluions o he ageing problem, pension policy mus conribue o he overall economy-wide measures being adoped o ackle his problem. The mos effecive way in which i can do his is hrough assessing he various pension policy suggesions agains he above menioned 3-pronged policy sraegy. Table 2 presens he main policy suggesions in he pension area, wih he res of his paper concerned wih evaluaing hese various policy sraegies agains he hree objecives of boosing growh, conaining budgeary pressures and avoiding negaive income disribuion consequences. Whils discussing hese sraegies in he coming secions, readers will hopefully find answers o he following key quesions which mus be asked abou he operaion of any pension sysem : Wha are he advanages and disadvanages of boh he PAYG and funded approaches o reiremen income provision? Which is he cheapes pension financing opion for individuals (i.e. wha is he equilibrium conribuion rae -ECR- for boh he PAYG and funded sysems )? Wha ype of reforms are needed o keep he ECR of he PAYG sysem consan a is 2000 level (i.e. wha will i ake o offse he effec of ageing on he PAYG sysem over he nex 50 years)? Given he risks which boh he PAYG and funded sysems are exposed o, wha is he bes long run risk diversificaion sraegy? Wha are he coss associaed wih a move from PAYG o funding? Wha measures need o be aken o proec he naural benefis of funding? Wha is a sensible pension policy for he EU for he coming decades? TABLE 2:POLICY REFORM OPTIONS :MAIN ISSUES TO BE CONSIDERED BASIC CHOICE FOR PENSION REFORM :PARAMETRIC /LABOUR MARKET REFORMS OF THE PAYG SYSTEM VSYSTEMIC REFORMS IN THE FORM OF A SHIFT TO FUNDING OR SOME COMBINATION OF BOTH SYSTEMS* 22

21 PARAMETRIC /LABOUR MARKET REFORMS OF THE PAYG SYSTEM REFORM OPTIONS : While "PAYG" reforms should essenially only refer o he mechanics or parameers of he PAYG sysem, namely he replacemen rae, he conribuion rae, he eligibiliy crieria ec, in secions 2 and 3 of his paper hese reforms refer o any acions of governmens o place he PAYG sysem on a more susainable long erm fooing and consequenly he following reforms are discussed: Labour marke reforms such as increases in paricipaion raes and falls in srucural unemploymen. Reducions in he generosiy of he PAYG sysem via direc replacemen raio changes or indirecly via changes in indexaion rules (price V wages) or modificaions in erms of accrual raes / reference earnings ec. Increase in he effecive reiremen age. MAIN ISSUES TO BE CONSIDERED Gainsfromreforminermsoffinancingcoss and incenive effecs. Inergeneraional equiy consideraions. Assessing he exen o which he PAYG sysem will be affeced by he change in demography - is i really more exposed o demographic risk han a funded sysem or is i simply a quesion of adjusing he passiviy raio which will bring he sysem back ino equilibrium by adjusing he raio of years worked o years in reiremen? Imporance o be aached o reaining he naural advanages of PAYG including universal coverage, defined benefis, absence of financial marke risk and inergeneraional solidariy. Balance o be sruck beween he insurance and redisribuive goals of he sysem. The growing demands for greaer ransparency in PAYG sysems and for a closer acuarial link beween individual conribuions and individual benefis for labour supply and economic growh reasons. A he momen his link beween conribuions and benefis does no exis since he conribuion rae from curren workers o he PAYG sysem is simply dicaed by he paymens o be made o curren pensioners, wih he sysem, in heory a leas bu no in pracice, always in balance. SYSTEMIC REFORMS :FULL OR PARTIAL SHIFT TO FUNDING REFORM OPTIONS : The essenial issue o be assessed in secions 4 and 5 of he paper is he exen o which governmens should abandon he "unfunded" PAYG sysem in favour of a sysem of funding which is based on he accumulaion of a sock of capial asses amassed hroughou one's working lifeime. Any policy acion aimed a increasing he share of funding needs o ake he following facors ino consideraion : wha are he ransiion coss involved; wha is he jusificaion in erms of he financial cos of he sysem; wha is he economic jusificaion in erms of he benefis for savings and growh, for labour supply and for financial marke developmen. MAIN ISSUES TO BE CONSIDERED Should funding be conrolled by he public or privae secors or some combinaion of boh? Should funded schemes involve compulsory or volunary savings? Financial marke risk and he volailiy of reurns o defined conribuion schemes - role and cos of insurance mechanisms o limi he losses. Cos of ax incenives for encouraging he ake-up of privae pension plans. Design feaures of he funded sysem which need o be addressed, paricularly he relaively high adminisraion coss associaed wih funding. Does funding really lead o higher savings, lower labour marke disorions and more efficien financial markes or is he only real difference is higher rae of reurn? Acuarial and inergeneraional fairness coupled wih higher raes of reurn are imporan advanages for funding bu are hese more han offse by he ransiion coss o such a sysem coupled wih any addiional adminisraion, ax incenive and volailiy coss when he funded sysem is operaional? How significan are he financial marke gains likely o be, given he raher mixed evidence from he lieraure and he fac ha he EU already has a sophisicaed financial secor. Arguably, regulaory issues and effecive levels of compeiion are more imporan concerns for policy makers in erms of efficiency gains in his secor raher han a shif o funding. 23

22 ISSUES COMMON TO THE PAYG AND FUNDED SYSTEMS Are boh sysems equally affeced by he lower economic growh oulook associaed wih ageing? How well do boh sysems handle demographic, macroeconomic and financial marke risks? Are boh sysems affeced by poliical risk in erms of ime inconsisency? * This breakdown ino "PAYG" and sysemic reforms is purely for illusraive purposes since reform effors in pracice do no fall nealy ino he laer wo caegories. However, his simplisic breakdown is concepually useful in isolaing he imporan lines of argumen in he pension reform debae. This simpliciy is paricularly eviden for he "PAYG" reforms, wih measures included, such as increases in he reiremen age and labour marke reforms, which are equally applicable o he funded as well as he PAYG sysem, wih hese reforms included in he "PAYG" secion simply o show he exen of he acion which will be needed o sabilise he PAYG sysem. Furhermore, his breakdown ino PAYG and sysemic does no accuraely reflec he inherenly more "hybrid" naure of naional pension sysems. For example, many naional PAYG sysems have elemens of funding eiher of he prefunding ype (which is discussed in Annex 2) or of he noional defined conribuion ype which is now operaing in boh Ialy and Sweden. Many of hese "hybrid" sysems are an aemp o ap ino he "naural" advanages of boh he PAYG and funding opions, wih he objecive of keeping he financing coss of he overall sysem as low as possible. The noional defined conribuion approach has addiional objecives such as insulaing he PAYG sysem from he effecs of ageing whils simulaneously addressing design fauls wih he original sysem, in paricular he lack of a clear acuarial link beween conribuions and benefis which could lead o labour supply disorions. While he design of pension sysems may differ, he basic economic raionale is more homogeneous, wih he crucial economic consideraions being he following : o avoid disorions wherever possible, in paricular by creaing close links beween conribuions and benefis and by having a high degree of ransparency in he sysem; o boos savings and o keep economic growh consideraions a he cenre of pension policy deliberaions. The above lis of reform opions and he issues which need o be considered shows he difficuly of coming o any firm conclusions in relaion o pension reform simply by examining he lieraure. The case for pension reform is essenially an empirical quesion which mus be assessed using a se of parameers which are specific o a paricular counry or group of counries such as he EU. The objecive of secions 2-5 is o presen he essenial deails of he EU's pension sysem and o empirically evaluae he above pension reform sraegies agains he specific background circumsances peraining in he EU in erms of demographics and labour marke performance. Secion 6 hen ries o bring all he main srands of he pension reform debae ogeher o creae an "opimal" EU pension reform sraegy, wih he final choice for policy makers being clearly dicaed by more han he economic consideraions enunciaed in he presen paper. OVERVIEW OF PENSION REFORM DEBATE KEY ISSUES + CONCEPTS (BASIC ECONOMICS OF PENSION REFORM) POSSIBLE POLICY SOLUTIONS (REFORM PAYG V SHIFT TO FUNDING) COM PARISON OF EU + US PENSION SYSTEMS ( ) HISTORICAL CHANGE IN PUBLIC EXPENDITURE ON PENSIONS (% POINTS OF GDP) FUTURE EFFECT OF AGEING ON PUBLIC EXPENDITURE ON PENSIONS+ON GROWTH EU + 6 ¼ US + 3 ¼ PENSIONS (EPC / OECD) (% POINTS OF GDP) EU +3 US +1 ¾ GROW TH RATE (ANNUAL CHANGE) EU - ½ US - ¼ 24