Pension Prefunding, Ageing, and Demographic Uncertainty

Transcription

1 Penson Prefundng, Ageng, and Demographc Uncerany by Jukka Lassla and Tarmo Valkonen The Research Insue of he Fnnsh Economy (ETLA) February 28, 2 Forhcomng n Inernaonal Tax and Publc Fnance Absrac: Penson prefundng can be used o smooh conrbuon raes n economes where ageng wll ncrease penson expendure. Bu how exensve should prefundng be n a defned benef penson sysem when here s consderable uncerany concernng fuure moraly, ferly, and mgraon? We sudy he prefundng rules n he Fnnsh earnngs-relaed penson sysem wh an OLG smulaon model. The resuls show ha ncreasng he degree of prefundng could yeld a more even nergeneraonal oucome and make fuure generaons poson beer, bu s que possble o overshoo and harm curren generaons oo much. Makng he degree of prefundng ferly-dependen appears o be a useful alernave. Wh declnng ferly, curren large cohors would pay modesly ncreased conrbuons. The accumulaed funds, however, wll be huge n relaon o he wage blls of smaller fuure cohors. Ths research was carred ou wh fnancal suppor from ageng programme of he Academy of Fnland and he Cenral Penson Secury Insue n Fnland. We have grealy benefed from dscussons wh he Insue's economss Mkael Forss, Lasse Kosknen, Bo Lundqvs, Ismo Rsku, Janne Salonen and Rejo Vanne. We also hank Juha Alho for provdng us wh hs smulaons, Lans Bovenberg and an anonymous referee for commens, and Eja Kaupp for he model programmng.

2 . INTRODUCTION The debae concernng penson fundng has wo man branches. One, perhaps somewha passng, suppors a fundng mehod beween PAYG and fully funded sysems. Ths s also found under he headng of penson prvasaon. The oher emergng branch concerns prefundng as a ool for preparng for he effecs of populaon ageng on penson expendures. The former branch relaes fundng more o defned-conrbuon sysems, whereas he laer ncludes ssues relaed specfcally o prefundng n defned-benef penson sysems. Ths paper belongs o he laer caegory. Our paper makes he pon ha once he decson o prefund has been made, he rules for fundng have o be decded, and hese rules are crucal for he success of he prefundng polcy. The rules defne he amouns of new fundng and whdrawals from he penson funds. We concenrae on rules ha relae prefundng o penson rghs and demographc developmens, and dscuss only shorly oher ypes of fundng. Our sarng pon s he curren Fnnsh prvae-secor earnngs-relaed penson sysem. I s based on a defned benef prncple, bu s a he same me parally funded wh forward-lookng fundng rules. We analyse he mpacs of changng he prefundng rules, keepng he deermnaon of he benefs nac and leng he conrbuon rae adjus. To do so, we smulae he polces usng a numercal overlappng-generaons model. The paper also characerses he mpacs of demographc uncerany on he penson sysem and he aggregae economy. An analyss of sochasc populaon forecass shows ha he uncerany s much larger han generally s recognsed consderng he me horzons used by penson sysems. Ths uncerany boh ncreases he need for preparng for he fuure and makes he creaon of good fundng rules a demandng ask. The paper s organsed as follows. Secon 2 surveys recen relevan leraure and descrbes some exsng prefundng pracces, especally hose n Fnland. In Secon 3 we descrbe boh he sochasc populaon smulaons and he overlappnggeneraons general equlbrum model ha uses hese smulaons as npus. Secon 4 analyses he effecs of ncreasng he degree of prefundng n he Fnnsh penson sysem, and Secon 5 dscusses he effecs of complemenng he curren rules wh a ferly-dependen componen. Secon 6 concludes. 2 PREFUNDING IN DEFINED BENEFIT SCHEMES 2. Background The penson reform leraure s passng he sage of vgorous debae amongs hose n favour of and agans prvasaon of he penson sysem. The dscusson s now much more focused, e.g., separang acual prvasaon from porfolo dversfcaon and prefundng (see Geanakoplos e. al 999). An mporan conrbuon o he dscusson was he observaon ha, n presen value erms, here s nohng o be 2

3 ganed from a ranson from PAYG o a funded sysem, even hough he laer would offer a permanenly hgher rae of reurn, see Snn (2). Ths does no mean ha fundng canno ncrease socal welfare under several ypes of welfare funcons. Fundng s now mosly seen as a mehod of redsrbung he nergeneraonal ncdence of coss due o ageng of he populaon. Snn (2) noes ha prefundng should be jus an alernave o rasng more chldren. Funds would replace mssng human capal wh real capal. Those generaons, whch do no bear enough chld rearng coss, mus save correspondngly more n order o manan he nergeneraonal farness of he penson sysem. Anoher relaed reason for prefundng would be ncreasng longevy. Bohn (999) clams ha a defned benef sysem s ex-ane more effcen han a defned conrbuon sysem n a world wh demographc uncerany. In he case of a negave ferly shock, he small cohors benef a lo from he favourable wage and neres rae movemens. The hgher axes only neuralse par of hs gan. The concluson s based on a smulaon of a closed economy wo-perod OLG model. The analyss assumes ha ncreasng conrbuon raes are no problemac as such. Changng crcumsances may warran a change n he degree of fundng whn he penson sysem. Orszag and Orszag (2) noe ha commng o he up-fron coss assocaed wh movng from a PAYG sysem o a prefunded sysem whou any assurance ha he coss could be recapured f he reform urns ou o be less benefcal han had been ancpaed may be unwse. Therefore fundng rules should be flexble. If he am s o sablze he conrbuon rae n a defned benef sysem, many knds of flexbly n prefundng mgh be warraned. Shor-erm varaons n he ax base due o busness cycles can be addressed by buffer funds. Buffers are also used o allevae he nergeneraonal burden due o he reremen of he baby boom generaons. We consder he mos mporan long-erm rsk o be devaons n demographc varables, perhaps gong beyond he known varaon n he sze of curren age cohors. Ferly and he penson sysem have been conneced n many sudes. Snn (998a) emphasses he role of a PAYG sysem as an nsurance for ndvduals agans no beng able o have chldren, or havng ungraeful chldren who are no wllng o ake care of her parens. He noes ha he sysem unforunaely also faclaes free-rdng volunarly no havng chldren. The queson wheher penson benef should be relaed o he number of chldren he reree has had has also been addressed, see e.g. Snn (998b) and he references heren. Many curren publc penson schemes already have feaures whch promoe ferly. The ofen used measure s o allow penson accrual o ake place also durng maerny leave. The resul s vald only when he households dscoun rae s he same as he rae of reurn on capal. Ths rules ou dsoronal capal ncome axaon. There are also some oher qualfcaons, no drecly lnked o he prefundng ssue, bu relevan o acual penson sysems. For example, n PAYG sysems he benefs are ofen defned so ha he lnk beween pad conrbuon and accrued penson rgh s no gh. A shf o a defned conrbuon sysem would elmnae hs ncenve problem. 3

4 Penson prefundng s a fundamenal par of mos volunary penson schemes. Palacos and Pallares-Mrallès (2) classfy he mandaory schemes accordng o he fnancng prncples and fnd ha abou half of mandaory schemes are purely payas-you-go fnanced. Abou a hrd have some reserves, even hough hey are based on a defned benef (DB) prncple. In mos cases hese funds are, however, ransory and dsspae afer he schemes maure. The reasons gven for prefundng vary. In Sweden, he nal dea was o neuralze he negave effec of a PAYG scheme on prvae savng and o use he reserves as a buffer agans cyclcal varaons (Sundén 2). In Fnland, he offcal explanaon for paral fundng was he need o smooh he expendure pressures due o populaon ageng, bu also oher nerpreaons have been gven. The Employees Pensons Ac (TEL) was creaed n negoaons beween rade unons and employers organsaons. Shor-erm barganng consderaons and he need for polcal suppor may well have been mporan facors (see Lassla and Valkonen 2a) Penson prefundng n Fnland: old-age pensons n he prvae secor The man prvae-secor earnngs-relaed penson sysem n Fnland, he so-called TEL scheme, s parally funded, bu fundng does no affec penson benefs. I only affecs he level and me pah of conrbuons. Fundng s collecve bu based on ndvdual penson rghs. Currenly he man prefundng rules are as follows. A par of old-age penson benefs, payable afer age 65, s funded for each employee. Fundng akes place beween ages 23 54, so only benefs accrued durng hose years are (parally) funded. The degree of nal fundng s below one-hrd 2 : of he.5 (of wage ncome) penson rgh accrung every year,.5 s funded. The presen value of accrued rghs s calculaed usng a 3 dscoun rae and a moraly able. No fundng s done for benef ncreases due o ndexaon. Fundng of dsably and unemploymen penson akes place when he case occurs. The nal funded share s 8 and, agan, no fundng s done for benef ncreases due o ndexaon. The pensons are pad, and funded, only unl age 65. Afer ha he pensoner receves old-age penson. In he followng we concenrae on old-age pensons. We frs descrbe he prefundng of he average employee's fuure benefs, and how he fund s run down durng he reremen years. Then we aggregae o he oal populaon level, o be able o presen he dynamcs of oal funds, he mechansm of conrbuon deermnaon, and he use of he yeld of funds. 2 There s no specfed arge for he share ha s funded. Before 997 fundng beween ages was full, n he sense ha f here would have been no wage nflaon, he 5 nomnal yeld requremen of he funds would have resuled n funds suffcen o pay ou, whou any PAYG fnancng, exacly he benefs whose rghs had accrued beween ages Needless o say, here was nflaon boh n prces and n wages, and fundng was far from full. The changes made n 997, descrbed n Lassla and Valkonen (2a), were calbraed so ha he requred fundng would say a he prevalng level, and.5 s a resul of ha calbraon. 4

5 Prefundng on he ndvdual level Every year, a new penson rgh accrues for each worker, and a par of he presen value of he rgh s prefunded. For someone already rered, he money prefunded n her workng years s used o pay a par of her penson. Equaons () and (2) below descrbe hese fundng rules for he average worker and pensoner n each age group n perod. The gross labour ncome of he average worker n age group s denoed by g. IN () h = a kg M h j,, S, j + r ) j= /( = 23,,54 ndvdual accumulaon rule OUT IN TEL j (2) h, = h + j, j ( + r ) = 65,,M ndvdual decumulaon rule j= 23 Accordng o equaon (), a share a of he presen value of he penson rgh accrung n perod o workers aged beween s pu n he funds. The presen value ncludes all old-age penson years, from 65 o a maxmum age denoed by M. The labour ncome g creaes a penson rgh for each year from age 65 onwards. For prefundng purposes, he magnude of hs rgh s evaluaed gnorng all fuure changes due o wage or prce developmens. Thus he value of he rgh s smply kg for each reremen year. Currenly, k s.5. The dscoun facor ncludes boh an neres rae and survval probables. The ex h ane fund rae of neres r, used here only n he prefundng rule (), s admnsravely se. The erm S n () descrbes he expeced effecs of moraly. Only a share S of hose n age group n perod s expeced o be alve a age j. Equaon (2) saes ha, for a rered person, he amoun prefunded earler (when he curren pensoner was beween ages 23-54) for perod 's penson, and he neres accrued o hose funds, s used o pay a par of he person's penson (he res comes from he PAYG par). The neres accrued s calculaed ex pos usng anoher admnsravely se neres rae, he so-called TEL-calculaed neres rae, assumed here o be consan for a smpler exposon. In pracse follows approxmaely he average marke yeld plus a margn, and mus no be lower han he fund rae n equaon (). Equaons () and (2) are n pracse neresng only o penson companes, as hey are used n calculang her penson lables. Each company s responsble for he prefunded pars of he pensons of hose nsured n he company. The companes are jonly responsble for he res of he pensons. Bu he equaons are mporan, of course, for he aggregae dynamcs of he penson sysem, especally for he level and me pah of he conrbuon rae. Aggregae penson funds and he conrbuon rae The oal amoun of new fundng n perod s obaned by mulplyng he average ndvdual fundng n age group, descrbed n equaon (), by he number of workers n n he age group, and summng over all age groups. Ths s done n equaon (3). The 5

6 oal amoun whdrawn from he funds s obaned analogously (equaon 4). Three oher aggregaes are defned n equaons (5) o (7): he oal wage bll, from whch he penson conrbuons are colleced, he oal amoun of old-age earnngs-relaed penson expendure, where he average ndvdual pensons are denoed by z, and he oal amoun of oher ransfers from he penson secor. 54 (3) A = n = 23, h IN, new fundng, oal M OUT (4) W = n, h, whdrawals from he funds, oal = 65 (5) G = 64 = 8 n, g, oal conrbuon base (wage bll) (6) Z = M = 65 Z (7) S = M = 8 n n z, oal old-age penson expendure, s Z,, oher pensons from he TEL sysem The dynamcs of he oal amoun of funds, H, follow from equaon (8) and he conrbuon rae s deermned as a resdual from equaon (9). TEL (8) H = H ( + r ) + A W l Z TEL (9) τ G Z + S + A W H r r ) ( = Conrbuons mus be suffcen o cover ha par of penson expendure whch does no come from whdrawals from he funds, plus new fundng and ransfers, mnus he dfference beween he real yeld r on he funds and he TEL-calculaed neres rae 3. To beer see he role of demographcs n he deermnaon of conrbuons, we nser equaons (3) - (7) no equaon (9) and reorder. l () τ M M 54 M OUT IN Z n, z, n, h, n, h, n, s, = 65 = 65 = 23 = 8 H ( r = = 8 n, g, = 8 n, g, = 8 n, g, = 8 n, g, = 8 n, r g TEL, ) The frs erm on he RHS s he rao of penson expendure o he wage bll. The man demographc varable s he number of pensoners relave o wage-earners. The second erm s lke he frs one: he number of pensoners relave o wage-earners s he crucal demographc facor. The hrd erm, new fundng relave o he wage bll, s demographcally raher nvaran: he numeraor depends on he number of people 3 The yeld dfferenal vares beween penson companes, and s he man facor n her compeon for cusomers, whch are employers provdng penson nsurance for her workers. 6

7 beween ages 23 and 54, and he denomnaor on hose beween ages 8 and 65. The overlap s 32 years, and n addon he labour force parcpaon raes for persons below he age of 23 and above he age of 54 are low. The fourh erm ncludes he oal adul populaon n he numeraor and he workng-age populaon n he denomnaor. I vares wh demography, bu should be kep n mnd ha oher ransfers from he penson sysem are raher small compared o old-age pensons. The ffh erm, he neres accrual over he TEL rae, shows ha he amoun of funds relave o he wage bll s mporan for he conrbuon rae f he acual yeld exceeds he admnsravely se rae. In he followng we assume ha he wo raes are equal. If he goal s o smooh he me pah of conrbuons, equaon () shows hen he approxmae effec of Fnnsh-ype prefundng. The frs par s he normal pure PAYG effec, whch s reduced by he second erm, n a way smlar o reducng he penson benef level. The cos of hs reducon s he hrd erm, whch s roughly consan over me. So he varably n he PAYG par s reduced by ransformng a par of o a consan. The queson we ask s: should we ncrease he role of hs smoohenng? The reasonng s smple: currenly he amoun whdrawn from he funds reflecs only he number of curren pensoners. I perhaps should reflec also he number of curren workers. The hrd erm wll hen also reflec he rao of curren workers o fuure workers. I becomes more responsve o demographc varaons. Besdes fundng, here are oher ways a penson sysem can prepare for dfferen elemens of demographc uncerany. The nergeneraonal redsrbuon effecs of unceran longevy could be allevaed ex ane by lnkng pensons o lfe expecancy (see Sherman, 999). Only surprses n longevy afer reremen age would hen affec he conrbuon rae. Ne emgraon can be a very dffcul problem, because may leave very lle me o adjus. One possbly s ndexng he benefs o he developmens of he wage bll (see Dsney, 999). Paral raher han full ndexng may be called for, as he laer may make he benefs very unpredcable, whch should be avoded as rered persons usually have very lmed possbles o reac o benef cus. Ferly changes, on he oher hand, affec he penson sysem wh a longsh delay. Lassla and Valkonen (2b) compares longevy-adjusmen and ndexaon of he benefs o oal wages o he ferly adjusmen n prefundng as devces of nergeneraonal redsrbuon of demographc rsks. 3. THE MODEL Our man ool s an overlappng-generaons general equlbrum smulaon model. I uses as npus dfferen demographc scenaros, produced by sochasc populaon smulaons. Techncally hs s sraghforward. Inellecually here s a gap: here s no uncerany n our OLG model. The households and frms n he model do no ake no accoun demographc uncerany n her decsons. They have perfec foresgh of he fuure populaon developmen n each alernave. Alhough would be more sasfacory o have uncerany ncluded, no essenal changes n resuls could be expeced. Prefundng changes he mng of conrbuons, no he level of benefs or he nsurance he penson sysem provdes agans oulvng one s savngs. I s 7

8 possble ha changng he prefundng rules would reduce precauonary savng, f credbly of he penson sysem ncreases, bu wh perfec capal mobly hs would have only small effecs. The demographc scenaros are ndependen of he OLG model. I would have been very demandng o endogenze ferly, longevy and mgraon whn he model, and he changes n economc oucomes, caused by dfferen prefundng polces, are no lkely o generae sgnfcan demographc reacons. 3. Demographc uncerany Alho (998) has made sochasc populaon forecass for Fnland. Based on prevous forecas errors n ferly, longevy, and mgraon, he has modelled he uncerany n demographcs, and has made 5 smulaons wh he model. He provded wo sample pahs for hs sudy. They are also repored n Lassla and Valkonen (999). The 5 smulaons were ordered wh respec o her age rao sasc n 23. The age rao s he number of people over 6 years dvded by he number of people beween ages From hs orderng, he wo samples closes o he lms of he 8 confdence nerval n 23 were chosen. The hgh age rao s a resul of declnng ferly and ne emgraon, and he low age rao s manly he resul of declnng lfe expecancy. Fgure. Rao of populaon n ages over 6 o he populaon n ages KELA-Eurosa Low age rao Hgh age rao ETLA Age rao: rao of populaon n ages over 6 o hose beween Hgh (low) age rao: neares realsaon o he upper (lower) lm of he 8 confdence nerval of he age rao n 23. Source: calculaons by J. Alho Fgure shows hese realsaons, as well as he baselne scenaro. The 8 confdence nerval for he age rao n 23 s wde: he lms are.63 and

9 Ths can be roughly ranslaed no penson language as follows. There s a probably ha Fnland wll have 8 pensoners or more for every workng-age persons, and a probably ha here wll be 6 pensoners or less for every workng-age persons n 23, provded ha he acual reremen age does no ncrease subsanally. In any case, he ncrease wll be large, as he curren rao s 3.5 pensoners for workng-age persons. 3.2 OLG Model FOG s an Auerbach-Kolkoff-ype, perfec foresgh numercal overlappng generaons model. There are fve secors: households, enerprses, a governmen, a penson fund and a foregn secor. The labour, goods and capal markes are compeve and prces balance supply and demand perod-by-perod. There s no money or nflaon n he model. Households and frms are forward-lookng decson makers. Household behavour Households maxmse he uly from consumpon and lesure n dfferen perods and he beques ha hey gve. The lfe-cycle plan for he household sarng s work lfe a me = s he soluon o he followng maxmsaon problem subjec o he perodc uly funcon (2), lfeme budge consran (3) as well as he deermnaon of gross labour ncomes (4), he reference penson (5), old-age pensons (6) and he dscoun facor (7): Max, () c l B T U γ B γ + µ T = ( + δ ) ( + δ ) γ, T, (2) (3) U = ρ ρ c + αl ( ) ρ T W g = T w T w T T C C ( τ ) Z R + z ( τ ) R + R B + s + s = c p ( + τ ) R R T B T = T w+ (4) g = ( l ) e w = = = TW w ref T (5) z = θ g = w W λ p p C TW C λ (6) λ2 λ 2 C ref w p = z C w T p W TW z. 9

10 (7) R = S +, ( r) Households consder he possbly of early deah by dscounng fuure consumpon and ncomes by a facor whch ncludes boh he neres rae and he age-specfc survval probably. The varable c descrbes consumpon, p C s prce, l s lesure, and of he consan parameers γ s he elascy of neremporal subsuon, δ s he rae of me preference and ρ s he elascy of subsuon beween consumpon and lesure. Households receve a beques B a he age of and gve a beques B before dyng. The parameer µ deermnes he srengh of he T joy-of-gvng beques move. The aggregae amoun of he generaon specfc ransfers S s deermned o balance he revenues and expendures of he cenral governmen. A lfe-cycle plan s made a he age of 2, and people plan o rere a he age of T w +. The budge consran (3) says ha dscouned lfeme wage and w penson ncome equals dscouned consumpon expendure. The erms τ andτ C are ncome ax and value added ax parameers. The θ parameers descrbe how he penson rghs are relaed o career earnngs. More wegh s gven o he las workng years, because n pracse he pensonable wage s aggregaed over he las years of each employmen conrac. As benefs depend on and conrbuons are pad from wages, here s an ndrec connecon beween conrbuons and benefs a he ndvdual level. Ths connecon s no one-o-one, however, as all career earnngs are no weghed equally. The ndexng of accrued penson rghs s dfferen n workng years and n reremen: λ s currenly.5 and λ s.2. The acual equaons of he smulaon model are he frs-order condons 2 derved from he opmsaon problem. Decson problem of he frms Frms choose he opmal amoun of nvesmen and use of labour o maxmse he prce of her shares. The marke value of he frm s deermned as a dscouned sum of fuure dvdends. The problem can be presened as maxmsng n he begnnng of perod he dvdends D dsrbued durng he perod plus he value of he frm V a he end of he perod, subjec o he amoun of nal capal sock, he cash-flow equaon of he frm (9), he CES producon funcon F (2), he accumulaon condon of he capal sock and he nvesmen adjusmen cos funcon (8) Max L I, K, F (9) D = ( ) K (2), he deermnaon of he frm's deb D + V. subjec o: G (23). l F d F F F K [ p F G ( + τ ) w L r B ] + B B p I F B (22)

11 (2) ( ) ( )( ) ( ) β F β F β β F = A εk + ε ν L, (2) K = d ) K + I F (, K (22) B = bp K and (23) 2 I = G ξ. K Equaons (8) and (9) have been smplfed by leavng ou he capal ncome ax F K erms. The prce varables p, p descrbe he prces of value added and he capal d un. r s he domesc neres rae, whch generaes neres flows o be dsrbued F durng perod. The ypcal CES producon funcon parameers are as follows: A s he scale parameer, ε s he share parameer and β s he subsuon parameer. ν descrbes he rae of producvy growh of labour. The accumulaon of capal K s explaned by usng he deprecaon rae λ and he amoun of new nvesmens I. The parameer b descrbes he collaeral value of he capal sock. In he las equaon, he parameer ξ deermnes he scalng of he nvesmen adjusmen coss. Three of he four frs-order condons of he consraned opmsaon are used as model equaons, he fourh beng he ransversaly condon. Markes The model ncludes four markes, whch clear every perod. In he labour marke, frms demand labour accordng o he margnal producvy of labour rule. Households' aggregae labour supply s dvded beween publc and prvae employmen. The wage rae s deermned by equang supply and demand n he labour marke. Frms are sole supplers n he marke for he domesc good. The produc s used by oher frms as par of he compose nermedae and nvesmen goods, by households as par of he compose consumpon good and by foregn agens. The demand of domesc agens and he prces of he compose goods are deermned by a cos mnmsng procedure. Domesc demand for he fxed-prce mpored good s also deermned by mnmsng he coss of he compose goods. The perfecly elasc supply adjuss o demand n hs marke. The fourh marke s he capal marke, n whch savng and nvesmen are balanced d by he domesc neres rae r. Toal savng s he sum of domesc savng and foregn porfolo nvesmens. The presenaon above descrbes only he relevan pars of he model. The acual model ncludes, e.g., governmen wh an neremporal budge consran, and rade

12 and capal flows wh he res of he world 4. The modelled penson sysem also ncludes he employee s conrbuon rae and prefunded dsably and oher early pensons, whch are no mporan for our resuls. The model s un perod s fve years, and he annual fundng rules have been modfed accordngly. The model s descrbed n more deal n Valkonen (999) and Lassla (2). Acuary rao As an nergeneraonal measure of he connecon beween benefs and conrbuons we defne he followng. The acuary rao s he rao of a cohor s dscouned benefs from he penson sysem o s dscouned sum of paymens o he penson Z sysem. The benefs nclude old-age pensons, denoed by z, and, combned no s, dsably and unemploymen pensons and all oher pensons from he earnngsrelaed penson sysem. T T T W (24) Α R = Z + l s R zr / τ gr = = TW + = If he acuary rao equals, he cohor ges he same rae of reurn on penson conrbuons han would have go from fnancal markes. A rao less han descrbes a ypcal suaon n PAYG penson schemes, afer he nal generaons who have a rao n excess of uny. The acuary rao s closely relaed o he money s worh calculaons famlar from he penson reform dscusson n he Uned Saes. 4. BASELINE SCENARIO Our baselne smulaon ncludes he neracon beween he populaon, he penson sysem and he res of he economy. The ageng of he populaon s due o he connuous ncrease n lfe expecancy and he low brh rae. Accordng o he offcal forecas, he Fnnsh populaon reaches a maxmum n he begnnng of he 22'es. The man pon s, however, ha he workng-age populaon sars o dmnsh already afer a few years and he amoun of pensoners ncreases srongly, snce he babyboom generaons rere. The low reremen age and he maurng of he penson sysem magnfy he economc burden. The forward-lookng frms reac o he expeced reducon n he labour force by lowerng he growh rae of nvesmens and by developng labour-savng producon mehods. The excess demand for labour acceleraes he growh of wages, even hough he employer's penson conrbuon rae rses. The wage growh s also suppored by he dampenng growh of expors, whch mproves he erms of rade of he economy. 4 The amoun of labour marke effcency losses due o he hgher penson conrbuons depends on boh he deals of he penson sysem and on he ax srucure of he economy. In he Fnnsh case, he overall axaon of earned ncomes s hgh even before ageng sars o have an effec. The dual ncome ax sysem allows he capal ncomes o be axed by much less. These feaures emphasse he benefs of labour ncome ax smoohng, whch could be acheved va prefundng of pensons. 2

13 From he pon of vew of he workers, he oucome s, however, negave, snce he hgher consumer prces and he hgher employee's conrbuon rae lowers he rae a whch he purchasng power of wages ncreases. The penson savng dmnshes gradually and also aggregae household savng declnes due o ageng. The nsuffcen savng fnally urns he curren accoun surplus no a permanen defc. Wh he lower margnal produc of capal he marke value of he frms drops from he nal growh pah. Invesors are compensaed by hgher dvdends, whch can be dsrbued due o he reduced need for nvesmen fnance. Accordng o he baselne scenaro, prvae secor penson expendures rse abou 2 percenage pons and he average penson conrbuon rae rses abou percenage pons of he correspondng aggregae wages durng he frs half of he cenury. The gap s explaned by he fac ha he penson fundng ncreases. The larger nvesmen reurns of he fund mgae he needed long-erm hke n he penson conrbuon rae. Whle he benef deermnaon rules reman nac, he hgher conrbuon rae neverheless weakens sgnfcanly he acuaral farness of he penson sysem. 5. EFFECTS OF CHANGING THE PREFUNDING RULES 5. Increasng he degree of prefundng We sudy he effecs of rasng he degree of prefundng (parameer a n equaon ) of he prvae secor old-age pensons from one-hrd o wo-hrds ( p.a.) and o one (.5 p.a.) from 25 onwards. Fundng sll akes place only beween ages and he dsably penson fundng remans unaffeced. The developmen of he oal conrbuon rae depends on he demographc rends. In he baselne alernave (KELA-Eurosa), ncreasng fundng o wo-hrds resuls n a roughly horsonal pah for he conrbuons. A hgher degree of fundng would be over-reacng: he conrbuon rae could be lowered afer 225. Ths overshoong would be even more pronounced n he low age rao scenaro. Bu n he low ferly (hgh age rao) case a 2/3 degree of fundng would no suffce o yeld sable conrbuon raes, see Fgure 2. Smlar developmens can be seen n acuary raes (Fgure 3). Increasng he degree of prefundng could yeld a more even oucome and make fuure generaons poson beer, bu s que possble o overshoo and unnenonally harm curren generaons oo much. The degree of prefundng s no an easy polcy nsrumen o use. 5.2 Addng ferly o he deermnans of prefundng 5 Lenghenng he ndvdual fundng perod o age 65 would have only small effecs. Many workers have already rered, as he acual reremen age s below 6 years on average. Ineres ncome would accrue for a relavely shor me. Conrbuons would nally rse by abou percenage pons, and hen fall permanenly below he baselne by abou.2.3 percenage pons n all demographc alernaves. Inergeneraonal balances would reman praccally unchanged. 3

14 Curren prefundng rules provde some auomac smoohng of penson conrbuon raes n he case of changng demographcs. As fundng akes place for each worker, he more here are rered workers he more here are also funds. And hese funds are used o pay a par of he penson benefs of hese rered workers. Thus he conrbuon raes need no rse n full proporon o he number of rerees. If we hnk of he rao of pensoners o workers, curren prefundng rules reac o he numeraor. To furher sablze he conrbuon rae, he denomnaor should also be aken no accoun. Ths s he subjec of hs secon. Table. Toal conrbuon rae and asses of he penson funds n Ferly dep. prefundng Kela- Eurosa Low age rao Hgh age rao Curren rules Toal conrbuon rae, Prefundng rae 2/3 Prefundng rae Curren rules Asses, per cen of oal wages Ferly dep. prefundng Prefundng rae 2/3 Prefundng rae We amend curren old-age penson fundng rules so ha, for each fundng cohor ( a cohor beween ages 23 55), he amoun funded depends also on he sze of he fundng cohor relave o he sze of recenly born cohors. The dea s ha we can esmae from he sze of recenly born cohors he sze of he work force n hose fuure perods when he fundng cohor wll be rered. Ths ferly effec vares beween cohors, and for each cohor vares over me. 4

15 The ferly correcon changes equaon () no ( ): M h j,, S, j /( + r ) j= 65 IN ( ) h = ab, kg where b n f ( n,..., n ), =, /, +, The share of he presen value of benefs ha s prefunded s no longer a, bu nsead a mulpled by he erm b, whch depends on he sze of he fundng cohor (a s brh year), compared o he brh szes of he younger cohors. If he fundng cohor s bgger han he younger cohors, b exceeds uny and hus fundng s ncreased. If ferly ncreases and younger cohors are bgger, fundng declnes compared o curren rules. If he fundng cohor and younger cohors are equal n sze, b equals. The funcons f are formulaed so ha, n effec, he fuure generaons are represened wh weghs correspondng o her presence n he labour force durng he years when he now fundng cohor s rered. All hose fuure generaons canno be ncluded n he ferly correcon: fundng ends a age 54, and he younges cohor ncluded s hus 53 years younger han he fundng cohor. Bu when he cohor reaches, e.g., he age of 9, here are persons 72 years younger payng penson conrbuons. Sll, he ferly correcon capures a szeable fracon of he fuure labour force. Deals of he f funcons are presened n Lassla and Valkonen, 999. Table 2. Toal conrbuon rae and asses of he penson funds n 27, baselne and wh hgher neres rae Kela-Eurosa - curren rules - prefund. rae 2/3 - ferly depend. Toal conrbuon rae, Baselne scenaro Hgher neres rae Asses, per cen of oal wages Baselne scenaro Hgher neres rae Low age rao - curren rules prefund. rae 2/ ferly depend Hgh age rao - curren rules prefund. rae 2/ ferly depend Fgure 4 shows ha ferly-dependen prefundng would no make much dfference n he basc populaon scenaro and n he low age rao case. Bu would make a huge dfference f ferly were o declne, as n he hgh age rao case. The conrbuon rae would hen be somewha hgher han wh curren rules for abou 3 5

16 years, bu afer ha he rae would sablse below 4 per cen nsead of rsng o 5 per cen. From Fgure 4 seems ha a raher modes ncrease n conrbuon raes, alhough for a raher long perod, faclaes a very szeable declne laer. Ths s precsely wha happens. The smple explanaon s ha, wh declnng ferly, here are many more people payng he modesly ncreased conrbuons han here wll be n he fuure payng as conrbuons he PAYG par of curren large cohors. Wh many payers he funds accumulae and wll be szeable n relaon o he wage blls of smaller fuure cohors. Ferly-dependen prefundng also makes acuary raes more even, especally n he case of a hgh age rao, see Fgure 5. If a polcal decson on ncludng ferly as an addonal deermnan of prefundng were o be made, s echncal mplemenaon would be sraghforward. The admnsraon of he Fnnsh sauory earnngs-relaed penson scheme s decenralsed and handled by prvae penson nsuons. Each nsuon s responsble for he funded par of he pensons of hose nsured n he nsuon, and all nsuons are jonly responsble for oher pars. Every year each nsuon calculaes he requred amoun of new fundng, based on ndvdual daa. Complemenng hese daa wh cohor-wse ferly nformaon would pose no problems. Of course he exac formulas would have o be agreed upon. Macroeconomc and welfare effecs In all he polces descrbed above he pre-fundng rae s nally acceleraed. Therefore, he macroeconomc oucomes are smlar and s suffcen o descrbe he overall rends. Shfng he mng and scale of pre-fundng necessarly nvolves adjusmens n he conrbuon rae. From he pon of vew of he aggregae economy, does no maer wheher he employer's or employee's rae s shfed, he mpac on ne wages s dencal. In a defned benef sysem, a hke n he conrbuon rae due o an ncrease n pre-fundng represens a pure ax. Consequenly, lowers he ncenve o supply of labour durng he acve fundng phase. The overall effcency oucome s, however, lkely o be margnally posve, snce he larger nvesmen ncomes cu he fuure op rae of conrbuons. The nal reducon n labour supply due o he jump n he conrbuon rae reduces he margnal producvy of capal and spurs frms o lower somewha her nvesmen rae. Laer hs rend reverses. Household savng reacs n a smlar way o he varaon n he penson conrbuon rae. The ncrease n penson savng domnaes n he capal markes, cung he curren accoun defc permanenly. The overall mpacs are, as expeced, n he case of hgher fundng degree polces more promnen han n he case of brh rae adjusmen polcy, when he demographc rends follow he offcal forecass. The benefcal macroeconomc oucomes of he laer polcy are evden only n he hgh age rao scenaro. In hs 6

17 case he gans of acve pre-fundng really srke ou boh because of he srongly reduced dsorons n labour supply and he markedly lower foregn ne deb, whch would oherwse burden heavly he small fuure household generaons. Any ncrease n pre-fundng reduces he welfare of he curren workers due o he lower ne wage. Correspondngly, he fuure households gan. For he fnal evaluaon, one should herefore be able o add he gans and losses of varous household generaons. Ths procedure s, however, subjec o many objecons. The man feaures can anyhow be observed from he fgures descrbng he relave compensaed varaons by generaons (see Fgure 6). They show ha a gan from acve prefundng s possble also n he baselne case, bu n he low ferly rae scenaro he mprovemen n welfare s large. Ths apples especally o he polcy of lnkng he fundng rae o ferly. 6. CONCLUSIONS We have consdered he performance of penson prefundng rules n a defned-benef sysem and n an envronmen where here s consderable uncerany concernng fuure demographc developmens. In he rules, new fundng and whdrawals from exsng funds are relaed o accrued ndvdual penson rghs, as s currenly he suaon n he Fnnsh penson sysem, and o ferly. Penson prefundng can smooh conrbuon raes n economes where ageng wll ncrease penson expendure. Bu funds ha are suffcen n one demographc oucome are oo small n anoher and may be oo hgh n a hrd one. Increasng he degree of prefundng, n relaon o accrued penson rghs, n he Fnnsh earnngs-relaed penson sysem could yeld a more even nergeneraonal oucome and make fuure generaons poson beer. The lkelhood of demographc scenaros n whch hs would happen s large wh small ncreases n fundng bu ges smaller wh larger ncreases n fundng. I s que possble o overshoo and harm curren generaons oo much. Thus, a fxed degree of prefundng s hardly he bes choce. Makng he degree of prefundng ferly-dependen would no have much mpac n he basc populaon scenaro bu f ferly were o declne, a raher modes ncrease n conrbuon raes would faclae a very szeable declne laer. Wh declnng ferly, here are more people payng he modesly ncreased conrbuons han here wll be people n he fuure payng as conrbuons he PAYG par of curren large cohors. Wh many payers he funds do accumulae and wll be huge n relaon o he wage blls of smaller fuure cohors. Fundng rules ha combne accrued penson rghs and ferly seem o be worhy of furher nvesgaon. Desgnng prefundng rules ha work well n an unceran demographc envronmen s no a rval maer. Tha should no be used as an argumen for no preparng for he fuure, bu nsead as a sarng pon o sudy alernaves and compare her performance by, e.g., smulaon sudes. Alhough he ools for such sudes are no deal ye, hey provde valuable nsghs no he properes of varous prefundng rules under dfferen demographc oucomes. 7

18 REFERENCES Alho, J.M. (998): A sochasc forecas of he populaon of Fnland. Sascs Fnland, Revews 998/4. Helsnk. Auerbach, A.J. and J.L. Kolkoff (987): Dynamc Fscal Polcy. Cambrdge Unversy Press. Cambrdge Bohn, H. (999): Socal Secury and Demographc Uncerany: The Rsk Sharng Properes of Alernave Polces. NBER Workng Paper No. W73. Dsney, R. (999): Noonal Accouns as a Penson Reform Sraegy: An Evaluaon Socal Proecon Dscusson Paper No. 9928, World Bank, Washngon DC. Geanakoplos, J., O.S. Mchell and S. P. Zeldes (999): Socal Secury Money's Worh. In Prospecs for Socal Secury Reform. Eds. O.S. Mchell, R. Myers, and H. Young. Penson Research Councl. Phladelpha, PA: Unversy of Pennsylvana Press, 999: Lassla, J. (2): Wage Formaon by Majory Vong and he Incenve Effecs of Pensons and Taxaon. Fnnsh Economc Papers, Vol.3, pp Lassla, J. and T. Valkonen (999): Penson prefundng and ageng n Fnland (n Fnnsh), Helsnk: ETLA B 58. Lassla, J. and T. Valkonen (2a): Prefundng n a Defned Benef Sysem - he Fnnsh Case. In M. Feldsen and H. Seber (edors): Socal Secury Penson Reform n Europe. NBER conference volume, forhcomng from The Unversy of Chcago Press. Lassla, J. and T. Valkonen (2b): Penson Polces and Demographc Rsks. Paper presened a he Year 2 Inernaonal Research Conference on Socal Secury, organzed by he Inernaonal Socal Secury Assocaon (ISSA) on Sepember 2 n Helsnk. Orszag, J. M. and P. R. Orszag (2): The Benefs of Flexble Fundng: Implcaons for Penson Reform n an Unceran World. Paper presened a he World Bank ABCDE 2 Conference. Palacos, R. and M. Pallares-Mrallès (2): Inernaonal paerns of penson provson. Socal Proecon Dscusson Paper No. 9. World Bank, Washngon DC. Scherman, K.G. (999): The Swedsh Penson Reform. ILO, Issues n socal proecon, dscusson paper No. 7. Geneve. Snn, H-W. (998a): The Pay-As-You-Go Penson Sysem as a Ferly Insurance and Enforcemen Devce. CES Dscusson Paper No. 54. Unversä München. 8

19 Snn, H-W. (998b): Commen on Persson, n Seber, H. (ed.) Redesgnng Socal Secury. Insu für Welwrschaf an der Unversä Kel. Snn, H-W. (2): Why a Funded Penson Sysem Is Useful and Why I Is No Useful. NBER Workng Paper No 7592, March 2. Sundén, A. (2): How wll Sweden s new penson sysem work? An Issue n Bref, Cener for Reremen Research a Boson College, March 2, No. 3. Valkonen, T. (999): The Fnnsh Corporae and Capal Income Tax Reform: A General Equlbrum Approach. ETLA A 29. Helsnk. Avalable onlne a 9

20 Fgure 2. Effecs of ncreasng old-age penson prefundng on penson conrbuons and funds Toal conrbuons Funds, of wages 5 KELA-Eurosa KELA-Eurosa Low age rao Low age rao Hgh age rao Hgh age rao ETLA Base case Fundng degree 2/3 Fundng degree 2

21 Fgure 3. Effecs of ncreasng old-age penson prefundng on acuary rae KELA-Eurosa Low age rao Hgh age rao Base case Fundng degree 2/3 Fundng degree Cohors denoed by he year of enerng labour force ETLA 2

22 Fgure 4. Effecs of ferly-dependen prefundng on penson conrbuons and funds Toal conrbuons Funds, of wages 5 KELA-Eurosa KELA-Eurosa Low age rao Low age rao Hgh age rao Hgh age rao Base case Ferly-dependen prefundng ETLA 22

23 Fgure 5. Effecs of ferly-dependen prefundng on acuary raes KELA-Eurosa Low age rao Hgh age rao Base case Cohors denoed by he year of enerng labour force. Ferly-dependen prefundng ETLA 23

24 Fgure 6. Welfare effecs Increased fundng degree Ferly-dependen prefundng Low age rao KELA-Eurosa Hgh age rao Cohors denoed by he year of enerng labour force ETLA 24