EUROPEAN. Setingtargetsforgovernmentbudgets inthepursuitofintergenerationalequity. EconomicPapers358 April2009. HeikkiOksanen EUROPEANCOMMISSION

Transcription

1 EUROPEAN ECONOMY EconomicPapers358 April2009 Seingargesforgovernmenbudges inhepursuiofinergeneraionalequiy HeikkiOksanen EUROPEANCOMMISSION

2 Economic Papers are wrien by he Saff of he Direcorae-General for Economic and Financial Affairs, or by expers working in associaion wih hem. The Papers are inended o increase awareness of he echnical work being done by saff and o seek commens and suggesions for furher analysis. The views expressed are he auhor s alone and do no necessarily correspond o hose of he European Commission. Commens and enquiries should be addressed o: European Commission Direcorae-General for Economic and Financial Affairs Publicaions B-1049 Brussels Belgium Ecfin-Info@ec.europa.eu This paper exiss in English only and can be downloaded from he websie hp://ec.europa.eu/economy_finance/publicaions A grea deal of addiional informaion is available on he Inerne. I can be accessed hrough he Europa server (hp://europa.eu ) KC-AI EN-N ISBN ISSN DOI /15403 European Communiies, 2009

3 Seing Targes for Governmen Budges in he Pursui of Inergeneraional Equiy a by Heikki Oksanen b Direcorae-General for Economic and Financial Affairs, European Commission April 2009 Absrac The Sabiliy and Growh Pac (SGP) adoped in 1997 originally se budge balance as he medium-erm objecive (MTO) for all EU Member Saes, o creae a safey margin under he 3% of GDP defici ceiling. In a reform in 2005, MTOs were made counry-specific and dependen on iniial deb and he poenial growh rae. They were agreed for a ransiion period, unil he crieria for aking ino accoun implici pension liabiliies under ageing populaions were esablished. Preparaions for his are currenly (early 2009) ongoing. Agains his backdrop his paper explores possible benchmarks for seing MTOs under alernaive rules for he public pension sysem, based on he principle of acuarial neuraliy across generaions. The effecs of privaisaion of he mandaory pension sysem are spelled ou. The ongoing revision of he naional accouns will esablish new accouns for he (implici) pension liabiliies of governmens. This will allow arges o be se for explici and implici governmen deb in he ligh of inergeneraional equiy. One of he conclusions o emerge from curren rends and policies is ha reassessing he generosiy of public pensions, especially he generously low reiremen age in Europe, will remain high on he poliical agenda, as oherwise he burden o be shouldered by fuure generaions will coninue o increase (JEL codes: H11, H55, H6). Keywords: Medium-erm objecives for governmen budges, Public pensions, Populaion ageing, European Union Sabiliy and Growh Pac. a A previous version of his paper was prepared for he 10h Anniversary of he Euro workshop, November 2008, Middlebury College, USA. Wihou implicaing hem, I hank he paricipans for heir commens. I also hank Middlebury College for publishing i in heir working paper series (Oksanen, 2009b). Help from Paul van den Noord in finalising he curren version for publicaion is grealy acknowledged. b Mailing address: Direcorae-General for Economic and Financial Affairs, European Commission, BU-1, 5/179, B-1049 Brussels, Belgium, phone: , Heikki.Oksanen@ec.europa.eu. This paper was wrien by Heikki Oksanen, Direcorae-General for Economic and Financial Affairs. European Communiies, The views expressed in his paper are hose of he auhor and do no necessarily reflec he views of he European Commission.

4 Seing Targes for Governmen Budges in he Pursui of Inergeneraional Equiy by Heikki Oksanen Conens 1. Inroducion and ouline 2 2. The Excessive Defici Procedure and he SGP: origins and aims 3 3. The revision of he SGP in Ageing-relaed public expendiure projecions 9 5. Using he expendiure projecions for he susainabiliy gap indicaor The inergeneraional disribuion of coss under alernaive fiscal rules 13 Descripion of he pension model 14 Pure PAYG and consan deb raio 16 A new policy rule: acuarial neuraliy across generaions 16 The calibraion 18 The numerical resuls 19 Table 1: Public finances and pensions under acuarial neuraliy 21 Table 2: Public finances and pensions under acuarial neuraliy: one-hird privaisaion 21 Oher ageing-relaed expendiures 22 Table 3: Overall financial implicaions of ageing Possible implicaions for seing fiscal arges 24 Consequences of pension reforms for public deb and defici 24 Measuremen and use of IPD esimaes 26 Proposed use of he susainabiliy gap for seing MTOs Wider implicaions Concluding remarks 33 Posscrip 35 References 36 1

5 1. Inroducion and ouline This paper explores he issue of seing arges for governmen budge deficis in he pursui of inergeneraional equiy in he conex of he European Union's (EU) Sabiliy and Growh Pac (SGP), as revised in The origins and aims of he SGP are firs recalled (secion 2). The revision in 2005 is hen explained, emphasising ha he medium-erm objecives (MTOs) for governmen deficis became counry-specific and dependen on iniial deb and he poenial growh rae (secion 3). Since 2001 an exercise joinly underaken by he European Commission and he Member Saes has produced commonly agreed projecions on ageing-relaed public expendiures over he long erm for each Member Sae, wih he second round published in 2006, and he hird round in he pipeline for 2009 (secion 4). These projecions have been used since 2001 o analyse he requiremens of he financial susainabiliy of public finances and made available o he ECOFIN Council Minisers of he EU (secion 5). Nex, we specifically address he quesion of seing budgeary arges under alernaive rules for he public pension sysem, based on he crierion of acuarial neuraliy across generaions as a benchmark for inergeneraional equiy, as oulined by Beesma and Oksanen (2008). The effecs of swiching o parial funding and privaisaion of he mandaory pension sysem, under a simple model, are spelled ou (secion 6). The implicaions of his simple model for seing fiscal arges are hen discussed, focusing in paricular on he implici pension deb of governmens. This is ineresing in ha, in 2005, he ECOFIN Council agreed ha MTOs should become counry-specific and dependen on iniial deb and he poenial growh rae, while also agreeing ha hey were only se for a ransiion period unil he crieria for aking implici pension liabiliies ino accoun under ageing populaions are esablished. The analysis here is designed o explore he possible quaniaive implicaions of such crieria. Special aenion is given o he work of he saisical communiy on he nex revision of he naional accouns, where new accouns will be esablished for he (implici) pension liabiliies of governmens (secion 7). Given he baseline projecions for ageing-relaed expendiures, he oucome from he acuarial neuraliy rule oulined in his paper is a significan reducion in explici public deb, wih crucial implicaions for financial markes. This and wider implicaions for 2

6 global financial flows are briefly discussed (secion 8), and he paper concludes wih a summary of resuls and he main implicaions for policy design (secion 9). 2. The Excessive Defici Procedure and he SGP: origins and aims When he EU heads of sae and governmen agreed o esablish an economic and moneary union wih a single currency in Maasrich in December 1991, heir primary concern had o be he sabiliy and he credibiliy of he new currency. Wihou his, he single currency would no have been able o deliver he advanages of a more efficienly funcioning single marke o any of he paricipans. In paricular, sabiliy and credibiliy were indispensable for Germany, which was supposed o give up is srong Deuschmark. To achieve his aim, he new EU Treay delegaed moneary policy o he European Cenral Bank, wih independen saus. To eliminae he possibiliy of irresponsible fiscal policies disurbing moneary policy or he economy in any oher ways, sound public finances were made one of he guiding principles of economic policy (Aricle 4 TEC 1 ). This was ensured by prohibiing (1) cenral bank financing of governmens, (2) heir privileged access o oher financial insiuions and (3) he bail-ou of debs of any public eniy wih he help of he European Communiy or he Member Saes. In parallel, (4) seing price sabiliy as he primary objecive of he ECB was designed o preven bailou of debs of a Member Sae hrough inflaion (i.e. eroding he real deb burden by inflaion was eliminaed). This arsenal of measures geared o prevening divergences from sound public finances was furher complemened in he Treay wih an Excessive Defici Procedure (EDP), wih reference values for deficis (3% of GDP) and deb (60% of GDP). The EDP can ulimaely lead o fines being imposed on any Member Sae ha does no correc is Excessive Defici before he deadline se by he EU Council (Aricle 104 TEC). The wo reference values may well be quesioned for heir simpliciy, bu one should noe ha, according o he Treay (Aricle 104), he EDP was supposed o be implemened wih a view o indenifying gross errors, and ha discreion was lef o he Council of Minisers (ECFIN Council) in using he reference values, as i made i no only possible bu even obligaory o ake ino accoun all oher relevan facors. Thus, 1 References are made o he Treay Esablishing he European Communiy (TEC) as in force in 2008; all Aricles referred o here originae from he Maasrich Treay of 1991, and hey are essenially mainained in he Reform Treay under raificaion in

7 he Maasrich Treay aimed o srike a balance beween he simpliciy of he rules required for he poliical process and he complexiy of he underlying economic raionale. Beween he signing of he Treay and is enry ino force on 1 November 1993, he European Union experienced wo major currency crises (laer half of 1992 and Augus 1993). The credibiliy of he single currency plan was hi hard, and i soon became clear ha he decision o esablish moneary union was no going o be aken by he end of 1996, he firs possible saring dae if he majoriy of he Member Saes fulfilled enry crieria. However, i hen became more and more likely ha moneary union would sar on 1 January 1999, as he Treay sipulaed ha i would sar wih hose counries ha were eligible, regardless of how many hey were. Preparing for his evenualiy made i necessary o provide furher assurances o he German public and poliicians ha all poenially paricipaing counries would be commied o pursuing sound public finances. The record from he previous 20 years was no encouraging: in he counries ha ulimaely adoped he single currency public deb had grown from 30% of GDP before he firs oil crises in he early 1970s o 60% a he ime of signing he Maasrich Treay, and o above 70% by he mid-1990s. In 1991, i was 100% in Ialy and nearly 130% in Belgium wo founding members of he EU and i coninued o rise significanly unil In Germany iself and in France i had risen from around or below 40% in 1991 o nearly 60% in Thus, he prospecs of esablishing moneary union on solid ground were no very promising. There was no only he quesion of wheher counries acing (wisely) in heir own ineres migh possibly harm one anoher (which is one of he issues sill debaed by expers) bu, more fundamenally, some people saw a risk ha shor-sighed poliicians would harm boh heir own and all oher counries by pursuing irresponsible policies. These fears (alhough no necessarily pu in such sraighforward erms a he ime) led he German Finance Miniser o come up wih a proposal for a so-called Sabiliy Pac, which would lead o auomaic sancions for counries wih excessive deficis. However, auomaism was no legally possible as, according o he Treay, i is for he Council o ake decisions. This evenually led o agreemen regarding he regulaion on speeding up and clarifying implemenaion of he excessive defici procedure. I se a sric ime schedule for he consecuive sages under he EDP, deailed various crieria o be applied in he procedure and esablished a scale (as a percenage of GDP) of evenual fines. As a 4

8 rule appeared in a couple of sraegic places in he ex o make i comply wih he discreion reserved for he Council. This regulaion became he correcive arm of wha was labelled he Sabiliy and Growh Pac o reflec he view of some of he Member Saes (in paricular, France) ha he ask of he EU fiscal framework was also o suppor economic aciviy. 2 This correcive arm was mean o eliminae irresponsible policies in he shor and medium erm by ensuring ha he sancions became a credible hrea. The oher par of he SGP, he regulaion on he srenghening of surveillance of budgeary posiions and he surveillance and coordinaion of economic policies aimed o preven excessive deficis by requiring counries o srive for a budge ha is close o balance or in surplus in he medium run (he prevenive arm ). The safey margin below he 3% of GDP defici was designed o allow he auomaic sabilisers o do heir work, unless he economy falls ino a very severe recession. As par of his regulaion, euro-area members mus submi a Sabiliy Programme every year, in which hey se ou heir budgeary pah and he underpinning measures for he coming years. The oher EU counries submi similar Convergence Programmes, wih he subex ha hey are sriving o comply wih he convergence crieria se for adoping he euro. Issues relaing o he long-erm susainabiliy of public finances in connecion wih ageing populaions were no menioned in hese original regulaions. Of course, sicking consisenly o balanced budges (on average over ime) implies ha in a growing economy (nominal growh is decisive here) deb o GDP raio converges (slowly) o zero. Thus, such a policy rule, if consisenly pursued, has persisen consequences for public deb. However, (probably) he firs ime populaion ageing appears in a ex submied o he ECOFIN Council was on 1 May 1998 when he decision o adop he single currency was aken. The declaraion under srucural reforms alks abou addressing all aspecs of social securiy sysems in view of ageing populaions (ECOFIN Council, 1998). This did no come ou of he blue, as groundwork had already been sared and published by he Direcorae-General for Economic and Financial Affairs (DG ECFIN) of he European Commission (for example in 1996), and by he OECD. Menioning ageing 2 For more deails, see Eichengreen and Wyplosz (1998); Fischer e al. (2006) classify he various proposals o amend he Pac; prominen examples are Wyplosz (2005), von Hagen (2002), Faas e al. (2003), Fioussi (2002), Blanchard and Giavazzi (2004). The proposals ha focus on long-erm susainabiliy are especially relevan o his paper: Pisani-Ferry (2002), Calmfors and Corsei (2004) and Buier and Grafe (2004). 5

9 populaions in he conex of moving o full moneary union no doub gave new impeus o ackling his challenge. In 1999, new repors and sudies were published by DG ECFIN, and in lae 1999 a working group for ageing populaions was appoined under he Economic Policy Commiee (AWG/EPC). The firs round of demographic and ageing-relaed public expendiure projecions was published in 2001, and he issue of he long-erm susainabiliy of public finances became a permanen iem on he policy agenda and in various repors. 3. The revision of he SGP in 2005 The revision of he SGP agreed upon a he ECOFIN Council and endorsed by he European Council in March 2005 was riggered by somehing quie differen from he opics in his paper. Bu as he opporuniy was here, he revision also addressed he issue of meeing he challenges of populaion ageing. The SGP had been pu on hold a he end of 2003 as a resul of he failure o apply he procedure in he case of he Excessive Deficis of Germany and France. Apar from he fac ha in many Member Saes deficis exceeded 3%, he paricular issue ha riggered he crisis was an issue of a Member Sae in Excessive Defici (Germany) ha had implemened he recommendaions i had received, bu exogenous facors had urned ou o be more unfavourable han expeced and he defici herefore did no decline. The Commission considered ha i was legally obliged o recommend moving o he nex sage of he procedure (i.e. one sep closer o poenial sancions), while Germany, suppored by France, waned o reurn o he previous recommendaions and revise hem (Korkman, 2005, p. 117). This dispue caused a deadlock in he Council, as he required qualified majoriy was no found under he correc legal procedure for a decision. In he subsequen ruling, he Cour of Jusice of he European Communiies (2004) considered (among oher hings) ha he recommendaions can indeed be amended a a laer sage by he Council, saing also ha his can legally only happen on a fresh recommendaion from he Commission (paragraph 92). Thus, under he original SGP he Commission was in a posiion o block a revision of a previous recommendaion by no puing any recommendaions on he able, bu if i did so he Council was, by a qualified majoriy, free o decide wha i waned. This was a clarificaion of he legal procedure under he original SGP, which wen agains he legal inerpreaion of he Commission in

10 The credibiliy of he SGP had suffered in he legal dispue. Seen in a posiive ligh, he siuaion provided an opporuniy for a review of he SGP. One of he main oucomes was ha he possibiliy of repeaing he seps was made explici in he revised SGP: i allows recommendaions o be revised and deadlines for correcing Excessive Deficis o be exended if unexpeced adverse economic evens occur. Similarly, he revision comprises several improvemens and clarificaions on shor- and medium-erm budge managemen designed o preven pro-cyclical fiscal behaviour. Specifically, budgeary adjusmen should be judged in erms of he cyclically adjused balance, ne of one-off iems and emporary measures. While implemenaion of he SGP had so far emphasised public deficis, in he revision he Council gave due consideraion o safeguard he susainabiliy of public finances in he long run, o promoe growh and o avoid imposing excessive burdens on fuure generaions (European Council, 2005). Thus, he revised SGP includes special provisions for reaing srucural reforms, including pension reforms. I had become obvious ha he similarly expressed medium-erm objecive (MTO) for budge balance for all counries under he original SGP was a simplificaion wihou sound economic raionale: i was sufficien o avoid excessive deficis bu i was no an economically jusified budge balance arge for counries in very differen posiions wih regard o iniial deb and ageing-relaed expendiure projecions. The join work on expendiure projecions had highlighed ha he Member Saes deviaed considerably in his regard, and i became clear ha MTOs should be seen as he link from he challenges of populaion ageing o budge plans for he medium erm (unil he end of he Sabiliy and Converge Programme period). Therefore, he revised SGP made MTOs counryspecific, subjec o possible revision in he case of major srucural reforms and in any case every four years. However, i was no ye possible o find a more lasing soluion for seing MTOs, and hus hey were se provisionally by aking only wo facors ino accoun: iniial deb posiion and poenial growh. Alhough references are made o srucural reforms in general and ageing-relaed issues in paricular, heir implicaions for he provisional MTOs remained limied, mosly o he adjusmen pah. Essenially, for counries ha already have adoped he euro or paricipae in ERM-II, he MTO se in 2005 ranges from a minimum of -1% of GDP for low-deb and high-poenial-growh counries o balance or surplus for high-deb or low-poenial-growh counries (ECOFIN Council, 7

11 2005). The baseline adjusmen pah was deermined so ha a counry ha has no ye reached is MTO should reduce is cyclically adjused defici annually by a leas 0.5% of GDP. The revised SGP also addressed he impac of sysemic pension reforms on MTOs, as a paricular issue arises when par of he public pension sysem is replaced by a privaesecor-managed fully funded ier. This issue was riggered by he enlargemen of he EU in 2004: Poland s mandaory pension sysem conained an imporan second pillar and i became necessary o clarify he rules. In essence, parial privaisaion of public pensions riggers an immediae reducion in pension conribuions paid o he firs pillar while pensions semming from he previously accrued righs need o be honoured, and herefore expendiures only go down much laer, over he ensuing decades. A sock of asses will be buil up in he newly esablished funded pillar, bu his will no be par of he governmen accouns as, according o he decision by Eurosa (2004), funded defined-conribuion schemes should be recorded as par of he privae secor. 3 To cope wih he reform, he governmen may issue deb, making some of he implici pension deb explici. Should his be an accepable consequence of he reform? The problem for he SGP is ha he public defici and deb increase, while he fall in implici liabiliies due o he reducion of fuure pension paymens from he PAYG pillar is no recognised in he naional accouns used for assessing compliance wih he SGP. 4 The resul was ha he correcive arm of he revised SGP gives some leeway, as follows: consideraion o he ne cos of he reform will be given for he iniial five years afer a Member Sae has inroduced a mandaory fully-funded sysem, so ha during he five years 100, 80, 60, 40 and 20 percen of he ne cos of he reform o he publicly managed pillar will be aken ino accoun (European Council, 2005; for a deailed presenaion, see European Commission, 2007). This leeway is indeed very limied as he reference defici level as a share of he cos falls over ime and is resriced o five years only, while he ransiion under pension reforms ypically lass for decades. 3 This decision by Eurosa concerned defined-conribuion, funded pension sysems ha may be managed by he governmen. I considered ha he fund s asses are ulimaely owned by he paricipans, who bear he risk associaed wih he reurn on he asses. These sysems should herefore be classified in he privae secor. Wih regard o defined-benefi schemes, an imporan crierion is he degree of funding. The Duch occupaional defined-benefi sysem is classified in he privae secor (as i is fully funded), while he Finnish parially funded mandaory defined-benefi sysem falls wihin he general governmen (as he degree of funding is only abou a quarer). 8

12 This was one of he issues a very concree one relaed o populaion ageing and pension reforms faced by he SGP revision, bu i was undersood ha even more fundamenal issues were no saisfacorily resolved. This was openly saed in 2005 by he ECOFIN Council, which said ha he new MTOs were se for a ransiion period unil he crieria and modaliies for aking ino accoun implici liabiliies [relaed o increasing expendiures in he ligh of ageing populaions] are appropriaely esablished and agreed by he Council (ECOFIN Council, 2005). This ransiion period was supposed o be a mos four years. Seing up he cieria is currenly (early 2009) on he able of he commiees ha prepare ECOFIN Council decisions. 4. Ageing-relaed public expendiure projecions Alhough he ECOFIN Council was no ye able, in 2005, o se well-based rules for aking he implici pension deb ino accoun, serious groundwork had already been done in he EU since 1999 in he Ageing Working Group of he Economic Policy Commiee (AWG/EPC). As said above, he firs round of demographic and ageing-relaed public expendiure projecions had been published in 2001, and hey laid he basis for analysing he long-erm susainabiliy of public finances in various EU repors. The Sockholm European Council in 2001 had agreed on a hree-pronged sraegy o ensure he susainabiliy of public finances and public policies vis-à-vis ageing populaions by (i) reducing deb a a fas pace; (ii) raising employmen raes and produciviy; and, (iii) reviewing and, where necessary, reforming pension, healhcare and long-erm care sysems. These had also become regular feaures in he Sabiliy and Convergence Programmes of he Member Saes (updaed annually), and he second round of joinly produced projecions was already in he pipeline in 2005 (Economic Policy Commiee and European Commission, 2001 and 2006a-b). To highligh some of he complex issues semming from, for example, he diversiy of social securiy sysems in he EU, we give here a summary of he challenge of ageing based on he projecion round of 2006 (he hird will be available in 2009). I provides projecions for populaion ageing-relaed public expendiures in he EU Member Saes (EU-25 a he ime) from 2004 o 2050 (henceforh referred o as EPC projecions ). 4 Tabellini (2003) makes his poin. Oksanen (2004) presens a numerical illusraion of he effec of parial privaisaion on public defici and deb. 9

13 They are based on demographic projecions, a commonly agreed se of macroeconomic assumpions regarding he labour force, produciviy growh and real ineres raes, and he policies or policy rules prevailing in Pensions are largely unfunded in mos EU counries. Before EU enlargemen in 2004, only he Neherlands, Denmark and he Unied Kingdom feaured a subsanial funded pension pillar. 5 If sricly applied, a pure PAYG pension sysem implies ha conribuions ino he sysem exacly mach pension paymens. Such a sysem hus affecs neiher he governmen s defici nor is deb. However, he consequence of increasing expendiure under a pure PAYG sysem is ha over he coming decades PAYG pension conribuions (or oher axes) have o increase subsanially. Cumulaing all ageing-relaed expendiures, we see ha he projeced increase by 2050 is 5%-poins of GDP (or more) in half of he EU Member Saes (Economic Policy Commiee and European Commission, 2006a-b). If his were o be financed from curren ax revenues, hen he sauory ax raes would have o increase by a leas 8-10%-poins (depending on he ax base). This raises serious quesions abou he negaive consequences of such ax hikes, including lower labour supply and ax evasion. Thus, one of he firs lessons is ha susainabiliy of public finances is no somehing ha in essence can be defined solely wih reference o defici and deb. Broader issues of accepabiliy and feasibiliy of he ax rae echnically required o cover he projeced expendiures also need o be covered by he analysis. In 2004, EU-25 pension expendiures were on average 10.6% of GDP, ranging from 4.7% of GDP in Ireland o 14.2% in Ialy. The average increase in he EU-25 by 2050 is a relaively modes 2.2%-poins of GDP. This is much less han he increase in he oldage dependency raio, which (oher hings being equal) would imply an 8.1%-poin increase in expendiure. The counervailing facors are a projeced increase in he employmen rae of prime-age workers and in he reiremen age, and, noably, a decrease in he average pension relaive o he average wage. The projeced average expendiure 5 Economic Policy Commiee and European Commission (2006a), pp , provides an overview of European pension sysems; Economic Policy Commiee and European Commission (2006b), pp , conains a scheme of he pension reforms in he EU. See also Whieford and Whiehouse (2006). In Finland, he sauory pension sysem is 20-25% funded, and consiss of invesmens mainly in asses oher han Finnish governmen bonds. I is classified wihin general governmen accouns; consequenly (e.g. in 2006), while gross public deb was 39%, ne deb was negaive, a -24% of GDP (Minisry of Finance of Finland, 2007). 10

14 increase of 2.2%-poins of GDP also conceals large differences beween Member Saes, from a 5.9%-poin decrease in Poland o a 12.9%-poin increase in Cyprus. Healhcare and long-erm-care expendiures are projeced o rise, on average, a roughly he same relaive speed as pensions. The dispersion across counries is smaller han for pensions, presumably parly because a common mehodology was applied o projecing hese coss. Based on ageing-relaed expendiure projecions, and aking ino accoun curren defici and deb raios, he European Commission (2006, pp ) classifies EU Member Saes ino hree groups as regards risks o he susainabiliy of public finances: (1) highrisk : he Czech Republic, Greece, Cyprus, Hungary, Porugal and Slovenia, where he projeced increase in pension expendiure ranges from 5.6 o 12.9%-poins of GDP; (2) medium-risk : Belgium, Germany, Spain, France, Ireland, Ialy, Luxembourg, Mala, Slovakia and he UK, where he projeced increase ranges from zero o 7.4%-poins of GDP; and (3) low risk : Denmark, Esonia, Lavia, Lihuania, he Neherlands, Ausria, Poland, Finland and Sweden, where he projeced change ranges from a decrease of 5.9% o an increase of 3.5%-poins of GDP. The EPC projecions should be regarded wih considerable cauion. For example, he significan decreases in projeced expendiures, noably in Poland, Esonia and Mala, and he very small increase for Ialy, sem from he policy rules in force in There is serious doub, however, abou he poliical susainabiliy of hose rules (and possibly also heir inerpreaion). This doub mainly concerns he indexaion of pensions: he rule ha is formally in place (e.g. price indexaion only), and herefore assumed for he projecion, may no be poliically susainable. 6 The EPC projecions do show, however, ha ageing-relaed public expendiures pose a severe challenge o almos all EU Member Saes. In his paper we refer o EU-average figures for illusraive purposes, while a closer analysis would require a careful look ino he EPC projecion for each Member Sae separaely. 6 Poland is an example: in 2005 he indexaion rule was changed o inflaion only, abolishing he parial indexaion o real wages. In line wih he common mehodology, he projecions were based on his new rule, implying a significan reducion in projeced replacemen raes. However, he original indexaion rule was resored in 2007 (Republic of Poland, 2008). As even he resored rule can be regarded as conflicing wih he basic principles of he Polish Noional Defined Conribuion sysem, 11

15 5. Using expendiure projecions for he susainabiliy gap indicaor Soon afer compleing he firs round of ageing-relaed expendiure projecions in 2001, hese projecions were used o consruc a quaniaive indicaor for he long-erm sabiliy of public finances along he lines of Buier (1985). An indicaor called susainabiliy gap is based on he consan ax rae ha would fulfil he ineremporal budge consrain (ITBC) of a governmen, given he projeced expendiures and servicing he curren deb. This ax rae is compared o he curren ax rae, and he difference is called he gap (see European Commission, 2006; also Buier and Grafe, 2004). The gap derived from an infinie-horizon calculaion is now more prominen, while a gap derived from a calculaion up o 2050 was firs used. The Board of Trusees of he US Social Securiy Sysem (Board of Trusees, 2007) uses he same mehod. However, i is by no means self-eviden ha his indicaor is an exhausive measure of he disance from susainabiliy. To sar wih, for a given projeced expendiure increase an infinie number of ax rae pahs exis ha fulfil he ITBC of he governmen, while he scenario wih a consan ax rae is only one of hem. We have already noed above ha he sandard pure PAYG rule is one of hem: axes are increased along wih he increase in expendiure. This, combined wih a rule o keep he public deb raio consan, is cerainly one scenario ha echnically fulfils ITBC of he governmen. Thus, here are several argumens for looking a a broader range of ax rae scenarios. Firs, he required ax increases may no be economically feasible (due o disoring labour supply, for example, or even moving he sysem o he righ of he peak of he Laffer curve). Second, subsequen generaions generally differ from each oher wih regard o heir feriliy, longeviy, reiremen age and pension benefis. Hence, o achieve inergeneraional equiy, one migh envisage ha hey also pay differen pension conribuions and oher axes (Sinn, 2000 and 2004). Third, i is possible ha MTOs should no be se solely on he expendiure projecion based on curren policies. I migh well be ha for many EU Member Saes he emphasis should firs be pu on designing reforms ha reduce hese expendiures (e.g. European Council, 2005), and MTOs should only be based on he revised projecions. When designing a comprehensive policy one may reasonably expec furher changes o he indexaion rule or occasional increases in pension payous. 12

16 package, alernaive calculaions for MTOs could be made condiional on several reform opions. I is clear, as is also acknowledged by he European Commission (2006, p ), ha he SGP incorporaes inergeneraional equiy neiher explicily nor sysemaically. One reason for his is ha disribuional issues are a maer of poliical preferences expressed and implemened a naional level aking acions affecing he fundamenal principles of naional social securiy is even explicily excluded from he compeence of he European Union (Aricle 137 TEC; his is mainained in he Reform Treay of 2007, which is pending raificaion). A furher reason is he lack of daa: since he EPC projecions do no provide daa by age cohor, hey do no allow he link o be examined beween he coss and benefis of each subsequen generaion. However, pensions (and oher ageing-relaed expendiures) are a large iem in governmen budges; hey deeply concern inergeneraional burden sharing and are hus anamoun o rolling over public deb from one generaion o he nex. There is herefore a compelling case for analysing alernaive policy opions from he poin of view of inergeneraional equiy. The aim of he following secion is o provide a simple overall framework o ake accoun of he inergeneraional equiy embodied in he public pension sysem when esablishing fiscal arges. 6. The inergeneraional disribuion of coss under alernaive fiscal rules This secion is based on a model presened by Beesma and Oksanen (2007, 2008) o address he consequences, under populaion ageing, of differen fiscal and pension arrangemens for public deficis, public deb, implici pension liabiliies and he balance of conribuions and benefis per generaion. The formal presenaion is kep o he minimum here, and also only he main implicaions are described. We deviae from sandard analyses of public pensions sysem ha assume ha hey are operaed on a pure PAYG basis (i.e. ha he conribuion rae is deermined by expendiure). Insead, an imporan saring poin in his insance is ha, under populaion ageing and changes o he rules for public pensions (parly brough abou by he former), mainaining a pure PAYG sysem does no necessarily have a solid foundaion. The argumen in Beesma and Oksanen (2008), as earlier pu forward by Sinn (2000) and Oksanen (2004), goes as follows: under a pure PAYG wih fixed parameers, if feriliy 13

17 declines and/or longeviy increases o a permanenly new level, he firs generaion wih he new demographic characerisics will gain a he expense of all fuure generaions, as i will have paid he conribuions deermined by he longeviy of he previous generaion bu will enjoy he benefis over a longer reiremen period (his is a generalisaion of he inference concerning he windfall o he firs generaion(s) when a pure PAYG sysem is esablished). To correc his unequal reamen, some of he policy parameer values should be changed: if he benefi raes are mainained, he conribuion rae should be increased immediaely i becomes eviden ha feriliy has declined and longeviy is increasing; or he fuure benefis should be reduced by decreasing he pension accrual raes and/or by increasing he reiremen age. The principle ha similar generaions should pay he same price for similar pension benefis opens up a wide range of reform opions, wih he salien feaure ha he sysem may no longer coninue as a pure PAYG, bu may move o parial funding. More specifically, i may accumulae eiher asses or deb, even permanenly (deb should be accumulaed if he accrued righs in he public pension sysem are significanly reduced, for example, in he conex of privaisaion while respecing he righs accrued before he reform; his means ha he implici deb becomes explici). 7 Descripion of he pension model In he model, workers are idenical, hey work for a cerain number of years and hey receive a pension when hey reire. We sudy he effecs of feriliy decline and longeviy increase. Reiremen age and pension accrual (and hence replacemen rae) may change. Feriliy per person is f children. Pensions can be parly (or fully) financed ou of curren workers conribuions or hey can be parly (or fully) financed ou of he asses accumulaed from conribuions in he pas. The discouned pension benefis o be 7 The argumen above is differen from he argumens pu forward, for example, by Kolikoff (1997) and Feldsein and Samwick (2002) ha privaisaion and funding provide efficiency gains o be shared beween he curren and fuure generaions. If valid, heir argumens also apply wihou populaion ageing, while he argumen above should prove ha here migh be a case for moving o parial funding under populaion ageing jus for he sake of neural burden sharing, even wihou he possible efficiency gains. Orszag and Sigliz (2001) righly poin ou ha moving o parial funding of public pensions means pre-funding in he broad sense so ha he sysem accumulaes asses, while esablishing individual accouns is a separae maer. The neuraliy principle referred o here is more generally applicable han Musgrave s (1986) fixed relaive posiion rule for a pure PAYG sysem, as promoed, for example, by Esping-Andersen e al. (2002). I is worh emphasising here ha a mandaory public pension sysem (for paernalisic eliminaion of people s shor-sighedness, and o avoid free riding under social insurance based on Rawlsian principles) need no be a pure PAYG, bu ha i can be funded o any degree, and his degree can also change over ime. 14

18 received by curren workers are ermed he implici pension deb (IPD, also called accrued-o-reference-dae liabiliy ) of he public secor. The key variables and parameers are bes defined wih he help of he consolidaed public secor (governmen including he pension sysem) insananeous simplified budge consrain: i owns (ne) financial asses A in period (public deb amouns o negaive asses), and all axes are levied on he wage bill. 8 The budge consrain is (1) cwl + ( ρ 1) A 1 = π 1wR + A A, 1 where c is he ax rae, w is he (gross) wage rae, 9 L is he effecive labour supply of workers in period, ρ is he financial marke ineres rae facor (he ineres rae is ρ -1), π -1 is he pension accrual rae and R is he effecive number of he elderly. Here, L, where is he number of new enrans o he labour force in period, and l is = ll % L % he number of years spen in work in period, divided by he number of years spen in work in he reference period 0. Furher, since periods refer o generaions here, and as he number of years during which a pension benefi is received generally differs from he number of years ha individuals pay conribuions o he pension sysem, we define σ as he number of years spen in reiremen in period divided by he number of years spen working in period -1, i.e. R = σ L -1. Finally, w 1+ g = w 1 1 ρ = (1 + g )(1 + μ ), where L is he wage bill growh facor, and μ >0 is an exogenous mark-up of he L financial marke ineres rae facor on he wage bill growh facor. Hence, μ is no influenced by he demographic shocks considered below. 10 The accrual rae as a share of π n 1 π n 1 he wage ne of conribuions,, is se by policy. Using, one hen derives he appropriae accrual rae π -1 as a share of he gross wage w For convenience, we label all primary revenues of he public secor as axes, even hough in our model hey consis mosly of pension conribuions. The reason is ha primary revenues also include axes colleced (from wages) o service he iniial explici deb of he governmen. From now on, wage sands for gross wage ha is, he wage before pension conribuions are paid. Noe ha for seing up he accouning framework, he expression for he ineres rae is merely an ideniy; for he main resuls below, however, we need o assume ha μ is exogenous. 15

19 Pure PAYG and consan deb raio We firs spell ou he policy rule ha keeps he financial posiion of he public secor unchanged (i.e. asses as a share of he oal wage bill are kep consan a a level a = a0 ). The oal ax rae hen consiss of a componen ha covers he curren pension oulays and a second componen ha capures he cos of deb servicing so ha a is kep consan. This implies ha (under he assumpion of posiive wage bill growh) he budge balance is in defici (surplus) if governmen ne asses are negaive (posiive). A new policy rule: acuarial neuraliy across generaions Now, we inroduce he following policy rule: saring from he curren period, for he (known or projeced) demographic characerisics (feriliy, reiremen age and longeviy) and pension accrual peraining o curren workers and he ineres mark-up projeced for he nex period, he ax rae is se a a level ha will be indefiniely susainable under he hypoheical siuaion ha no changes o hese facors will occur in he fuure. This rule essenially sipulaes ha, from now on, (hypoheical) idenical generaions will be reaed idenically by he governmen. Noe ha we do no assume ha hese facors will remain unchanged indefiniely. Raher, he rule amouns o how he ax rae is deermined from now on as a funcion of he demographic variables, pension policy parameers and he ineres mark-up. If and when one or more of hese facors change for a new generaion of workers, he same principle as jus defined applies again. For example, when he life expecancy of a new working generaion is projeced o increase, he ax rae on his generaion s income is raised. We now derive he formula for seing he ax rae implied by his rule. Where applicable, a superscrip indicaes ha he variable is based on he informaion available in period. For example, μ denoes he ineres mark-up in period known in period, while μ + 1 denoes he ineres mark-up projeced for period +1 under he informaion available in period For clariy, we also aach a superscrip o he lengh of he working life variable,, o indicae ha i is he value known in period when he ax rae for workers in ha period is deermined. Variable θ below also carries a superscrip as for each period i has boh an ex ane and an ex pos value; for a deailed explanaion of our rule, see he Appendix. 16 l

20 We define θ as he IPD as a share of he oal wage bill. For he burden from he pas (period -1) and given he (known) parameer values in period, i is (2) θ π 1 1( l 1 / l ) ( 1+ μ ) σ IPD 1 1 = = 1 w 1L 1 f 1, while for period, given he known parameer values and hose projeced for +1, i is (3) θ IPD πσ = = wl f + ( 1 1+ μ + 1 ). Higher feriliy f and longer working life l raise he oal wage bill and hence reduce he IPD as a share of he wage bill. An increase in he pension accrual rae π or a (projeced) increase in he number of years in reiremen relaive o he lengh of working life σ naurally raises i. An increase in he ineres rae mark-up implies heavier discouning of fuure liabiliies and hus has a depressing effec on θ. Using he above expressions, he ax rae implied by he new policy rule is found o be (see he Appendix o Beesma and Oksanen, 2008): 1+ μ μ θ θ μ+ 1 a (4) c = + 1 ( 1 a 1) Hence, each generaion of workers firs conribues o he public accouns o share he burden semming from pas decisions on pensions and oher expendiures and revenues by paying he ineres mark-up on he sum of implici pension liabiliies and explici public deb (possibly correced for a change in he mark-up see he firs erm on he righ-hand side), and hen pays he full presen value of is own fuure pensions (he second erm). An increase in he lengh of he working life implies ha he IPD accumulaed in he pas is shouldered by more labour inpu, which means a reducion in he conribuion rae. Now ha we have shown hese implicaions of our policy rule we label i as acuarial neuraliy across generaions: he burden of iniial explici and implici public deb is shared equally beween curren and fuure generaions and each generaion covers he acuarial value of is own fuure pensions. Like any acuarial calculaion implemenaion l > 1 l 1 17

21 of he rule is parly based on fixed policy parameers (e.g. he pension accrual rae and he reiremen age) and parly on projeced values (longeviy and ineres rae). The expression for he ax rae in equaion (4) is applicable for any changes, permanen or emporary, in he demography, reiremen age, generosiy of pensions and ineres rae margin in subsequen periods. Under his rule, while respecing he righs of curren reirees regardless of wha hey paid for hem, he balance of pension conribuions and benefis of each generaion, saring from he ime a which his rule is firs implemened, is enirely separae from he characerisics and pension policy choices of oher generaions. The rule and is implicaions can be viewed as a naural benchmark for neural reamen of differen generaions by he governmen. This conrass wih pure PAYG pension financing, which generally produces sysemaic redisribuion across generaions under populaion ageing. An imporan implicaion of our rule is ha he sum of explici public deb (-a) and implici liabiliies as a share of he wage bill (labelled he oal deb raio ) evolves as follows: 1+ μ 1 1 = 1 θ 1 1 μ a (5) θ a ( l l ) /. This equaion hus implies ha he oal deb raio remains consan if μ = μ + 1 and 1 l 1 = l. An increase in he lengh of working life or an increase in he ineres mark-up implies a reducion in oal deb. Oher hings being equal, higher implici liabiliies (a higher θ ) require higher pension conribuions, which implies an offseing reducion in he public deb. Noe ha (5) is applicable under any iniial degree of funding, which hen evolves as a funcion of facors in he formula for. In paricular, if he sysem were iniially fully funded, i would remain so under he acuarial neuraliy rule. a c The calibraion Our numerical examples are largely based on he following sylised calibraion aken from Oksanen (2005). The uni period corresponds o 30 years, which is roughly he average childbearing age of women in Europe. I is also roughly he average age difference beween a reired person (70) and a worker (40). Throughou, we assume ha he annual growh rae of he nominal wage per worker is 3.28%, which sems from a 18

22 uni real wage growh rae of 1.75% and an inflaion rae of 1.5% per annum. The ineres rae mark-up over he growh of he oal wage bill is 1.5%-poins per annum. The economy sars in period 0 in a seady sae in which people work for 40 years and spend 18 years in reiremen (hence, σ = 0.45). These figures are hus used o scale pension conribuions and pensions o correspond o realisic figures, alhough he formal model works wih a 30-year period. Feriliy, moreover, iniially preserves a consan populaion. Furher, he uni pension is iniially se a 55% of he wage afer pension conribuions (i.e., n π 1 = 0.55, see Appendix), so as o make he iniial figures comparable wih hose for 2004 in he EPC projecions. Iniial public deb is 60% of annual GDP. Period 1 includes boh a 20% fall in feriliy (roughly corresponding o a fall from 2.1 children per woman, which is needed for full replacemen, o 1.7, which is close o he curren average in Europe) and an increase in longeviy by hree years (his corresponds o one year for each en-year period). Period 2 includes a furher increase in longeviy by hree years. This allows us o mach quie closely he assumpions regarding he increase in longeviy in he EPC projecions for he EU average (an increase in life expecancy (a birh) for males (6.3 years) and for females (5.1 years) from 2004 o 2050). The numerical resuls We skip here he resuls for he pure PAYG and concenrae on he ime pahs of he mos relevan variables under various policy opions ha observe he acuarial neuraliy rule. We consider wo possible policy measures o conain he rising ageing burden: an increase in he reiremen age and a reducion in he replacemen rae; he ax rae is hen residually deermined. We also consider (parial) privaisaion of he public pension pillar, where he laer is parly replaced by a mandaory, funded privae pillar. Mos figures are expressed in percenages of GDP, assuming for simpliciy ha he oal wage bill (including pension conribuions) is a consan 60% of GDP. Table 1 illusraes he implicaions of he acuarial neuraliy rule. The baseline assumes no change in he generosiy of he benefis or in he reiremen age. The fall in feriliy (via a reducion in he discoun rae) and he rise in life expecancy boh raise he IPD as a share of GDP. In anicipaion of his increase, axes already rise in period 1 (he sysem moves o parial funding). Hence, he rise in he implici deb raio is followed by an 19

23 equivalen reducion in he (explici) public deb raio (or increase in he public asse raio), and he defici urns ino a surplus from period 1 onwards in all panels. The surplus is larges in period 1, when he populaion is affeced by wo shocks simulaneously (he rise in workers life expecancy and he fall in he feriliy rae). A reducion in he accrual rae (panel 2.1) produces a smaller implici deb/gdp raio, hereby requiring smaller surpluses han under he baseline. This effec is furher amplified when he reducion in he accrual rae is combined wih an increase in he reiremen age. The expendiure increase of 2.7%-poins of GDP in panel 2.2 in Table 1 roughly corresponds o he EPC projecion for he EU. All acuarially neural scenarios clearly show ha, as long as he demographic change is permanen, he financial posiion of he governmen should also change permanenly. Depleing public asses once he ageing process has ended is herefore excluded, unless pension accrual is drasically reduced. Table 2 considers he laer opion induced by parial privaisaion (possibly in combinaion wih oher measures) in which implici pension deb is swapped for (explici) public deb. This parial privaisaion is implemened wih a one-hird reducion in he accrual rae expressed as a share of he gross wage rae. Under acuarial neuraliy axes paid o he governmen fall immediaely (he conribuions o he newly esablished privae second pillar come on op of hese axes). Obviously, as accrued pension righs are respeced, public pension expendiure falls only gradually. Comparing he new seady sae wih period 0, he public deb/gdp raio falls only slighly (panel 1). In period 1 i exceeds he 60% limi of he SGP, while he defici raio in ha period remains jus marginally below he 3% limi. Adding o his a reducion in he accrual rae leads o a furher swap of deb in panels , where he defici raio is in breach of he SGP in period 1, as is he deb raio from period 1 onwards. 20

24 Table 1: Public finances and pensions under acuarial neuraliy Period change 1. Acuarial neuraliy baseline: ne accrual rae 55%; fixed reiremen age pension exp/gdp, % ax rae public deb/gdp, % IPD/GDP, % oal deb/gdp, % budge surplus/gdp, % Acuarial neuraliy accrual rae reduced o 48%; fixed reiremen age pension exp/gdp, % ax rae public deb/gdp, % IPD/GDP, % oal deb/gdp, % budge surplus/gdp, % Acuarial neuraliy accrual rae reduced o 48%; increase in working life ( ) pension exp/gdp, % ax rae public deb/gdp, % IPD/GDP, % oal deb/gdp, % budge surplus/gdp, % Noes: (1) The ax rae consiss mainly of pension conribuions and is expressed as a percenage of he oal wage cos. (2) The final column change gives he %-poin change from period 0 o he new seady sae, excep for he budge surplus/gdp raio, where i gives he %-poin change from period 0 o he lowes or highes level. Table 2: Public finances and pensions under acuarial neuraliy: one-hird privaisaion Period change 1. Acuarial neuraliy 55% accrual rae; fixed reiremen age pension exp/gdp, % ax rae public deb/gdp, % IPD/GDP, % oal deb/gdp, % budge surplus/gdp, % Acuarial neuraliy accrual rae reduced o 48%; fixed reiremen age pension exp/gdp, % ax rae public deb/gdp, % IPD/GDP, % oal deb/gdp, % budge surplus/gdp, % Acuarial neuraliy accrual rae reduced o 48%; increase in working life ( ) pension exp/gdp, % ax rae public deb/gdp, % IPD/GDP, % oal deb/gdp, % budge surplus/gdp, % Noes: see Table 1. 21