Future of Work and Pension Design: Portuguese Pension System

i Fu u e o Wo k and Pension Design Ped o Miguel da Sil a San os Almeida Po uguese Pension Sys em Disse a ion p esen ed as pa ial equi emen o ob aining he Mas e ’s deg ee in In o ma ion Managemen i NOVA In o ma ion Managemen School Ins i u o Supe io de Es a ís ica e Ges ão de In o mação Uni e sidade No a de Lisboa FUTURE OF WORK AND PENSION DESIGN by Ped o Miguel da Sil a San os Almeida Disse a ion p esen ed as a pa ial equi emen o ob aining he Mas e ’s deg ee in In o ma ion Managemen , wi h a specialisa ion in Knowledge Managemen and Business In elligence Ad iso : P o esso D . Jo ge Miguel Ven u a B a o Co-ad iso : D a. Ca a ina Gal ão No embe 2020 ii ABSTRACT The new o ms o wo k, as such as C owdwo king and F eelance jobs, d i en by he echnologies ad ances and globaliza ion a e in oducing mo e equen ca ee b eaks as well as highe a ia ions on he wo ke 's income which a e incompa ible wi h he cu en old-age p o ec ion sys em in Po ugal. Mos o he s udies pe o med in Po ugal a e a ound he sus ainabili y o he pension scheme and ocus mainly on he pa ame ic changes. Only a ew s udies a e e lec ing a edesign pension scheme like he No ional De ined Con ibu ion (NDC) implemen ed in Sweden, I aly and Poland. The objec i e o his s udy is o p opose a new design o a pension model wi hin he new con ex o he u u e o wo k. The pape analyzes he impac o he u u e o wo k on he cu en Po uguese DB pension model and o he exis en pension models by simula ing pension en i lemen s based on di e en ca ee b eaks and pensionable sala ies oscilla ions. The esul s o he s udy con i med he inadequacy o he ac ual Po uguese DB pension model o non-s anda d wo ke s, and a wo ke wi h mul iple b eaks on he se ice could ha e a signi ican loss on he e i emen pension en i lemen . The s udy also con i med ha o he s andalone design models as he No ional De ined Con ibu ion, Cash Balance, De ined Con ibu ion, and he Sys ems Poin s would p o ide he same inadequacy le els o non-s anda d wo ke s. We concluded ha Be e idge and Bisma ck's plans complemen each o he , and an op imal model wi hin he u u e o wo k would ha e o be based on bo h cha ac e is ics o a Be e idge and Bisma ck models wi h he same le el o impo ance. The s udy shows e idence ha mo ing away om he exis ing DB pension model o a mul i-model, a i s sub-model like a DB plan no linked o he con ibu i e ca ee - a Pu e Be e idge model - wi h he objec i e o “ edis ibu ion” and “po e y elie ”, and a second sub-model like a Bisma ck plan (DC o Cash Balance ype) wi h he objec i e o “consump ion smoo hing” and “insu ance” would imp o e he adequacy and ai ness o he pension e i emen o he non-s anda d wo ke s. KEYWORDS Fu u e o Wo k; Pension Design; Pension Schemes; Social Secu i y Sys ems, Digi al E a. iii ABSTRACTO As no as o mas de abalho, como a p es ação de se iços po aplica i o ou abalho de eelance s, impulsionadas pelos a anços da ecnologia e pela globalização, êm in oduzindo in e upções de ca ei a mais equen es, bem como maio es a iações nos endimen os do abalhado , que são incompa í eis com o a ual sis ema de pensões público em Po ugal. A maio ia dos es udos ealizados em Po ugal, gi a em o no da sus en abilidade do egime de pensões publico e incide p incipalmen e nas al e ações pa amé icas. Apenas alguns es udos analisam o impac o de mudanças es u u ais no a ual modelo de pensões em Po ugal, como o modelo de “No ional De ined Con ibui on” (NDC) implemen ado na Suécia, I ália e Polónia. O p incipal obje i o des e es udo, é p opo um no o desenho de um modelo de pensões em Po ugal den o do no o con ex o do u u o do abalho. O es udo analisa o impac o do u u o do abalho no a ual modelo de po uguês e nou os modelos exis en es, simulando di ei os de pensão de e o ma baseado em di e en es cená ios de in e upções con ibu i as e oscilações de salá ios, du an e a ca ei a con ibu i a. Os esul ados con i mam a inadequação do a ual modelo po uguês de pensões pa a as no as o mas de abalho, e um abalhado com múl iplas in e upções con ibu i as pode e uma pe da signi ica i a na sua pensão de e o ma. O es udo ambém con i mou que ou os modelos como a NDC, “Cash Balance”, Con ibuição De inida e o Sis ema de pon os êm os mesmos ní eis de inadequação de que o modelo de pensões em Po ugal ela i amen e às no as o mas de abalho. Concluímos que os planos do ipo “Be e idge” e “Bisma ck” se complemen am, e um modelo de pensões ó imo e ia necessa iamen e e ambas as ca ac e ís icas de um modelo de Be e idge e Bisma ck com o mesmo g au de impo ância. A p incipal ecomendação do es udo é a al e ação do a ual modelo de pensões de BD, pa a um múl iplo modelo, um p imei o submodelo de plano de BD, mas não inculado à ca eia con ibu i a, como um modelo Pu o de Be e idge em que a p incipal obje i os é ga an i um endimen o mínimo e a edis ibu i idade de inanciamen o do sis ema, complemen ado com um segundo modelo Bisma ck ( ipo DC ou Cash Balance) inculado à ca ei a con ibu i a, e em que p incipal obje i o é sua iza impac o no endimen o pe dido de ido ao e en o de e o ma e p o ege con a o isco de e o ma. O es udo demons a que ado ando ambas ca ac e ís icas “Be e idgean” e “Bisma kian” i ia pe mi i melho es ní eis de adequação das pensões de e o ma pa a as no as o mas de abalho. PALAVRAS CHAVES T abalhos do Fu u o; Desenho Pensões; Plano de Pensões; Sis emas de Segu ança Social, E a digi al. i INDEX 1. In oduc ion .................................................................................................................. 1 2. Li e a u e e iew .......................................................................................................... 5 2.1. Social Secu i y F amewo k .................................................................................... 5 2.2. The p oblem - globally ......................................................................................... 11 2.3. The p oblem in Po ugal ...................................................................................... 12 2.4. The u u e o wo k – adding mo e p essu e ....................................................... 14 2.5. Wha esea ch done in Po ugal o sol e he p oblem? ..................................... 17 3. Me hodology .............................................................................................................. 19 3.1. Resea ch ques ions .............................................................................................. 19 3.2. Resea ch me hod................................................................................................. 19 3.3. Resea ch design ................................................................................................... 19 4. Resul s and discussion ................................................................................................ 26 5. Conclusions ................................................................................................................. 38 6. Limi a ions and ecommenda ions o u u e wo ks ................................................. 43 7. Bibliog aphy ................................................................................................................ 44 8. Annexes ...................................................................................................................... 48 LIST OF ANNEXES Annexe 1 – Pension Models and Unemploymen Insu ance in Po ugal ................................ 48 Annexe 2 - The Wo ld Bank’s Concep ual F amewo k............................................................ 53 Annexe 3 – Po uguese Social Secu i y F amewo k ................................................................. 54 i LIST OF TABLES Table 1 - The i s junc u e: Bisma ck s Be e idge ................................................................... 6 Table 2 - Mul ipilla Pension Taxonomy classi ica ion by he OECD ........................................ 10 Table 3 - Main e o ms o he public pension scheme in Po ugal .......................................... 12 Table 4 - Reduc ion on pension en i lemen s by model – Baseline Scena io .......................... 28 Table 5 - Reduc ion on pension en i lemen s by model – Con e gen Scena io .................... 29 Table 6 – Measu es sco es o he cu en Po uguese DB plan by p o ile ............................. 30 Table 7 – Compa ison measu es sco es by p o ile (DB e sus NDC) ....................................... 30 Table 8 – Compa ison measu es sco es by p o ile (DB e sus DC) .......................................... 31 Table 9 – Compa ison measu es sco es by p o ile (DB e sus Basic Plan+Cash Balance Plan)31 Table 10– Compa ison measu es sco es by p o ile (DB e sus Basic Plan + Poin s Sys ems) . 32 Table 11 – Compa ison measu es sco es by p o ile (Ac ual DB e sus DB wi h new pa ame e s) .......................................................................................................................................... 32 Table 12 – Compa ison measu es sco es by p o ile (DB wi h new pa ame e s e sus Cash Balance) ............................................................................................................................ 33 Table 13 – Replacemen Ra io by sub-Plan (Swiss model= Basic Pension + Cash Balance) .... 33 Table 14 – Measu e sco e o Swiss plan (Basic Pension + Cash Balance)............................... 33 Table 15 – Compa ison measu es sco es by p o ile (Ac ual DB Po uguese plan e sus Newly Ta ge ed Plan + Cash Balance) ......................................................................................... 35 ii LIST OF FIGURES Figu e 1 - Po ugal Da a – Employmen ................................................................................... 16 Figu e 2 - Po ugal Da a - Ea nings ........................................................................................... 16 Figu e 3 - S a is ic Po ugal da a –employmen p o ile in Po ugal ........................................ 20 Figu e 4 - S a is ic Po ugal da a –employmen e olu ion in Po ugal .................................... 20 Figu e 5 – Baseline scena io - mul i b eaks .............................................................................. 21 Figu e 6 – Con e gen scena io - mul i b eaks ........................................................................ 21 Figu e 7 - Impac on he old-age bene i (Baseline scena io) .................................................. 26 Figu e 8 - impac on he old-age bene i (con e gen scena io) ............................................. 27 Figu e 9 – Replacemen Ra io measu e by model and p o ile ................................................. 34 Figu e 10 – Ac ua ial Fai ness measu e by model and p o ile ................................................. 34 Figu e 11 - Impac on he old-age bene i (Baseline scena io) ................................................ 36 Figu e 12 - Impac on he old-age bene i (con e gen scena io) ........................................... 36 1 1. INTRODUCTION The Po uguese pension sys em is based on h ee pilla s o di e ing impo ance: he dominan ea nings- ela ed old-age s a e pension sys em ( i s pilla ), he occupa ional pension p o ision (second pilla ), and he pe sonal pension p o ision ( hi d pilla ). The i s pilla combines an ea nings- ela ed, de ined bene i (DB), manda o y public scheme, comp ising wo sepa a e bu con e gen schemes, wi h mino special sys ems o lawye s and solici o s (CPAS) and olde bank employees (J. M. B a o, 2015) (J.M. B a o, 2016). P i a e-sec o wo ke s a e co e ed by pensions wi hin he gene al social secu i y scheme (Regime Ge al da Segu ança Social (RGSS)). Public se an s en olled be o e Decembe 2005 a e co e ed by pensions wi hin he o me ci il se ice pension scheme (Caixa Ge al de Aposen ações (CGA)), ha has been closed o new en an s since 2006 and new ci il se an s con ibu e o he gene al scheme. Ye , he CGA will con inue o ope a e o mos o he 21s cen u y. Con ibu o y s a e pensions a e inanced on a pay-as-you-go (PAYG) basis by social con ibu ions paid by bo h he employe and employee, complemen ed by a small ac ion o he alue-added ax (social VAT). Addi ionally, he s a e sys em includes non-con ibu o y means- es ed i s - ie pension bene i s, op-up minimum con ibu o y bene i s, and se e al a ge ed assis ance p og ams, ully unded by gene al axes, aimed a p o ec ing hose wi h sho con ibu ion pe iods o low pension en i lemen s mo e gene ally agains old-age po e y. The s a e pension scheme comp ises old age, ea ly e i emen , disabili y, and su i o s' pensions. Toge he wi h supplemen a y bene i s, i was designed o gua an ee he main enance o an app op ia e s anda d o li ing a e e i emen and in he e en o disabili y o dea h (J. M. B a o, 2018) (OECD, 2019). The ac ual Po uguese DB pension model migh ha e been a good i wi hin he dynamics o wo k, economic en i onmen and he demog aphic popula ion a he ime o implemen a ion. The declining o he old-age dependency a io, p ima ily due o he ageing o he Po uguese popula ion esul ing om a decline in e ili y and an inc ease in li e expec ancy, and he economic changes o e he las decades (low economic g ow h, low labou p oduc i i y and sala ies g ow h) ha e been pu ing p essu e on he long e m sus ainabili y o his unding model (B a o J.M. & Coelho E., 2020). As a esul o his con inuous demog aphic and economic changes, he e we e se e al e o ms o he social secu i y sys em in Po ugal in oduced in 1993, 2000, 2002 and 2007. Ne e heless, all hese e o ms we e pa ame ic changes (e.g., adjus ing he quali ying condi ions, indexing e i emen age o li e expec ancy au oma ically, in oducing he sus ainabili y ac o , modi ying he bonus/penal ies o la e/ea ly e i emen ) ins ead o s uc u al changes, designed o p o ide sho -public inance alle ia ion wi hou sol ing he sys em’s long- e m sus ainabili y, adequacy and ai ness p oblems (Ayuso, M., B a o, J. M., and Holzmann, 2020) (J.M. B a o, 2015) (J.M. B a o, 2016) (J. M. B a o, 2018) (Jo ge M. B a o & He ce, 2020). In ea nings- ela ed pension schemes, bene i s a e closely linked o indi idual employmen his o ies, wages and con ibu ions. Unemploymen pe iods and o he olun a y o in olun a y ca ee b eaks (e.g., pa - ime employmen , layo , e mina ion o ixed- e m con ac s, pa en al lea e, childbea ing, child ea ing, illness/disabili y, s udying, mili a y se ice, housekeeping wi hou aising child en o gene al inac i i y pe iods) ha e long- e m sca ing e ec s on u u e labou ma ke possibili ies and ea nings, pe manen ly a ec ing wo ke s’ e i emen income and he pension sys em inancial sus ainabili y (Jo ge M. B a o & He ce, 2020). Empi ical e idence sugges s ha unemploymen b eaks a e he ype o ca ee in e up ion ha is mo e ha m ul o an employee's u u e wage p ospec s and, 8 li ing. Insu ance componen s designed o achie e some a ge s anda d o li ing in e i emen compa ed wi h ha when wo king” (OECD, 2005). All pension schemes design in he OECD coun ies includes a i s ie a ound he po e y elie p inciple bu wi h di e en s uc u es. The e a e ou ypes o design he “ i s ie - edis ibu i e schemes” – “social assis ance”, “sepa a e a ge ed e i emen -income p og ams”, “basic pension schemes” and “minimum pensions wi hin ea ning ela ed plans”. All hese p og ams a e manda o y and public pension schemes. The basic pension he bene i is a la bene i paid o e e y pensione independen ly o ecei ing o he incomes om o he sou ces, o i depends only on yea s o wo k (bu no on he pas ea nings). O he sou ces o e i emen income do no a ec he en i lemen o he basic pension. The Ta ge ed plans pay a highe bene i o he poo e pensione s and a educed bene i o be e -o e i ees, i.e., his ype o plans based on a edis ibu ion o bene i s by educing he bene i amoun o weal hy pensione s and inc easing bene i s o needy pensione s. The minimum pension is like a ge ed plans since he e is a le el gua an ee o pension income; howe e , he eligibili y and condi ions a e di e en . In gene al, pensione s o be eligible o ecei e he bene i equi ed a minimum numbe o yea s o con ibu ions. Las ly, he e a e he coun ies which only p o ides a uni e sal social-assis ance in which co e hose who a e no eligible o ecei e an old-age pension by he public pension scheme. The objec i e o he second ie is o ensu e a simila quali y o li e as du ing he ac i e li e pe iod. The e a e di e en models o co e he second ie (J.M. B a o, 2015). De ined Bene i plans (DB plans) Mos o he coun ies ha e pu ely de ined bene i plan (DB), and his is he case o Po ugal. In gene al, his ype o plan, he bene i is de ined based on a cons an acc ual a e linked o he yea s o se ice (con ibu ions) and ea nings o wages. The con ibu ions made by he ac i e membe s, usually in he o m o ax paymen s, a e used o pay he cu en pensione s, and he e o e no sa ed o in es ed. De ined Con ibu ions plans (DC plans) The nex common ype o pension scheme is de ined con ibu ion plan (DC). In his ype o scheme, each wo ke has an indi idual accoun wi h he con ibu ions sa ed and in es ed o capi alisa ion, and he accumula ed capi al a e i emen age con e ed in o an annui y income and, in a e si ua ions, o a Lump sum paymen . The DC schemes o ganised in se e al ways, o example, in Aus alia, employe s equi ed o co e hei employees ia an indus y-wide und o a inancial se ices company. In Hunga y, Mexico and Poland, he employees choose a pension p o ide wi hou employmen pa icipa ion, and in Sweden, employees equi ed o con ibu e o a manda o y indi idual accoun . Mos o he coun ies wi h a manda o y DC plan also ha e ano he ype o scheme as such as poin sys em o he a e o e u n ha indi idual accoun s ea n as complemen a y o gua an ee minimum e i emen p o ec ion. 9 “Poin s Sys ems” Some coun ies do no s uc u e hei pension scheme ei he as a simple DB o DC plan. The i s case is whe e he pension schemes o ganised as “poin s sys ems” – his is he case o he public pension schemes in F ance, Ge man and Slo ak. The wo ke s c edi ed poin s based on hei con ibu ions linked o hei a e age ea nings, and a e i emen , he sum o he poin s mul iplied by a pension- poin alue o con e in o a pension income. The e i emen bene i depends on he alue poin a e i emen age. No ional Con ibu ion plans (NDC) The e a e also he no ional-accoun s schemes – he Public schemes o I aly, No way, Poland, Sweden and La ia. The yea ly con ibu ions made by ac i es as well he in e es on he con ibu ions eco ded on an indi idual accumula ed no ional accoun ha exis s on he books o he managing ins i u ion. A e i emen , each accumula ed accoun con e ed in o a pension income based on a no ional ac ua ial con e sion ac o e lec ing he li e expec ancy a he ime o e i emen . The con ibu ions paid by he ac i e membe s used o pay he cu en pensione s, bu in his case, he con ibu ions linked o he e i emen bene i , and he e is a “no ional” a e e u n on he asse s. In his ype o plan di e s om he simple DB plans, since he demog aphic and economic changes o e ime a e e lec ed in he inal e i emen bene i by adjus ing he no ional e u n o economic luc ua ions o adjus ing he ac ua ial annui y ac o o demog aphic luc ua ions. Usually, coun ies in oduce a gua an eed minimum pension since unlikely he DB plans an adequa e bene i no gua an eed a e i emen age (Holzmann, R., 2006) (Alho, J., B a o, J. M. & Palme , 2013) (Alho, J., B a o, J. M. & Palme , 2013) (Holzmann, 2017) (J. M. & E. M. de F. B a o, 2018) (J. M. B a o, 2019) Cash Balance Plans Finally, he manda o y con ibu ions o Swiss occupa ional plans a e cash balance plans which is like he DC plan, bu he e is a minimum in e es gua an ee on he con ibu ions made o he pension und. In Po ugal, he Social Secu i y F amewo k Law (Law no. 4/2007 o 16 h Janua y) delinea es he ac ual design o he Po uguese Social Secu i y Sys em (annexe 3). The social secu i y sys em based on h ee pilla s – he Ci izenship Social P o ec ion Sys em, he P o iden ial Sys em and he complemen a y sys em; The i s - ie based on he po e y elie and edis ibu ion gua an eed by he Ci izenship Social P o ec ion Sys em included in he subsys em solida i y supplemen ( a ge ed plan), and by he P o iden ial Sys em wi h a minimum pension gua an eed a e 15 yea s o con ibu ion; he second ie based on consump ion smoo hing and insu ance is co e ed by he P o iden ial Sys em which is a de ined bene i plan (DB). (OECD, 2019). Table 2 shows he OECD mul i-pilla pension axonomy o social secu i y schemes inclusi e o Po ugal. 10 Table 2 - Mul ipilla Pension Taxonomy classi ica ion by he OECD Tie : unc ion Fi s - ie : uni e sal co e age, edis ibu i e Second- ie : manda o y, insu ance P o ision Public Public P i a e Type Social assis ance Ta ge ed Basic Minimum Type DB DC Aus alia √ √ Aus ia √ DB Belgium √ √ DB Canada √ √ DB Czech Republic √ √ √ DB Denma k √ √ DB/DC √ Finland √ DB F ance √ √ DB+poin s Ge many √ Poin s G eece √ √ DB Hunga y √ DB √ Iceland √ √ I eland √ √ I aly √ NDC Japan √ DB Ko ea √ DB Luxembou g √ √ DB Mexico √ √ Ne he lands √ √ √ New Zealand √ No way √ √ Poin s Poland √ NDC √ Po ugal √ √ DB Slo ak Republic √ Poin s Spain √ DB Sweden √ NDC √ √ Swi ze land √ √ DB CB Tu key √ √ DB Uni ed Kingdom √ √ √ DB Uni ed S a es √ DB Sou ce: OECD; DB = De ined Bene i , DC = De ined Con ibu ion, CB = Cash Balance, and Poin s = Poin s Sys em. Despi e he a ie y o social secu i y schemes, globally, all schemes we e designed based on he social secu i y p inciples and used ools c ea ed mainly by social secu i y p og ams ini ially es ablished in Ge many and he UK. Fu he mo e, wi h he igh ques ions and he p ope concep ual amewo k empla e, each go e nmen had modelled hei schemes based on p e- de ined objec i es and he ins umen s used o achie e he objec i es- he Bisma ck e sus Be e idge models o a mix be ween he wo models (Ba , 2000). 11 2.2. T HE PROBLEM - GLOBALLY Nicolas Ba (2000) di e en ia es isk and unce ain y as o “ he isk, he p obabili y o po en ial ou comes is known o es imable, wi h unce ain y i is no .” This dis inc ion is c ucial since we can ac ua ially measu e he isk bu no he unce ain y. Public and p i a e pension schemes ace bo h isks and unce ain ies. Ba hen iden i ies wo so s o unce ain y: “Mac oeconomic shocks” and “Demog aphic shocks” (Ba , 2000). Indeed, hese wo ac o s we e he ini ia o s o challenging he sus ainabili y o pension schemes in Wes e n Eu ope, s a ing in he 1970s and onwa ds. The economic c isis in he 1970s, oge he wi h ongoing demog aphic al e a ions, had damaged se e al go e nmen budge s. The inancing o public pension schemes, in pa icula , he pay-as-you-go sys ems p e alen in he Bisma ck and Be e idge sys ems wen unde p essu e s a ed i s in he UK, and la e in I aly and Sweden wi h he public de ici c isis in ea lie 1990s (Ebbinghaus, 2009). Many coun ies in oduced e o ms on hei pension schemes in o de o main ain sus ainable hei pension schemes and many coun ies ha e unde aken signi ican e o ms, and o he s a e now expe iencing a simila p ocess (Ignacio Conde-Ruiz & Gonzalez, 2014). (Felds ein, 2001) Felds ein p edic ed he axes o pay he u u e bene i s would need o ise sha ply due o he ageing o he Eu opean popula ion. Felds ein s a ed ha his is no a empo a y issue bu a pe manen issue. The ax a e will ha e o ise by abou 50 pe cen in mos coun ies wi h e en mo e signi ican inc ease in some coun ies i no changes made o he s uc u e o he exis ing PAYG schemes (Felds ein, 2001). Based on Ignacio and Gonzalez, he e is no doub ha he ageing popula ion is damaging ha d he Pays-As-You-Go (Ignacio Conde-Ruiz & Gonzalez, 2014). Based on Schuldi, he economic p essu e will be mo e se e e on he public pension schemes since he demog aphic changes will lead o massi e ageing o he popula ion om 2020 onwa ds. He men ioned wo signi ican demog aphics changes – he inc easing li e expec ancy and declining e ili y a es wi h mo e subs an ial incidence wi hin wes e n Eu ope. Bo h Be e idge and Bisma ck models a e highly ulne able o hese demog aphic changes. Bo h models ha e he pension paid o pensione s based on he con ibu ions (in he o m o ax) paid by he ac i e membe s, and he e o e when he a io be ween ac i es and pensione s declines, i.e., he e a e ewe ac i es o pay he pensions in paymen o he pensione s, he sys em will come unde p essu e. (Schludi, 2005). Wha ha e done Wes e n Eu ope coun ies o sol e he p oblem? Despi e all he changes obse ed in he public pension sys ems s a ed in he 1970s and onwa ds, in Wes e n Eu opean coun ies, only a ew coun ies such as Sweden and I aly pe o med s uc u al changes on he public sys em by mo ing o a no ional de ined bene i (NDC) and eo ganised hei sys em ho oughly o make hem mo e sus ainable o he u u e. Sweden was he unique coun y ha mo ed om a basic pension plus ea nings- ela ed second pension (Be e idge ype) o a con ibu o y ea nings- ela ed (Bisma ck Type) bu keeping uni e sal. Fu he mo e, Sweden and Denma k in oduced unded pensions in he i s ie as a manda o y public pension. I aly did no in oduce he unded pa as done by Sweden bu ins ead ans e ed he manda o y end-o -se ice pay (“TFR” plans) in o a p i a e pension. In Ge many, he bi h o he i s Bisma ck social secu i y scheme in oduced a new laye in he public sys em, he olun a y pe sonal pension (“Ries e ”) and p omo ed collec i e occupa ional pension. The a ionali y is ha due o demog aphic changes, he p inciple o consump ion smoo hing he inspi a ion o he Bisma ck design is no mee any longe , and he e o e, he need o in oduce a second laye on he public pension scheme o ill he gap. In 12 coun ies wi h he occupa ional second ie s, Swi ze land and he Ne he lands, he pension e o m has been less p onounced, bu i is in he poli ical agenda. (Ebbinghaus, 2009). In conclusion, in Wes e n Eu opean coun ies, he e has been a ans o ma ion on he public schemes, and as men ioned by Ebbinghaus, despi e all he changes, “Ye , in he long- un hese ins i u ional changes may be he i s s ep owa ds a mo e subs an ial change in he public-p i a e mix o come in he u u e.” (Ebbinghaus, 2009). 2.3. T HE PROBLEM IN P ORTUGAL Po ugal is also acing he same issues as o he coun ies in Wes e n Eu ope since he Bisma ck and he Be e idge models inspi e he Po uguese social secu i y model. In he pas , he e we e se e al amendmen s o he social secu i y amewo k based on Law No. 28/84 o 14 h Augus . Table 4 illus a es he main e o ms pe o med o he public pension scheme in Po ugal since 1990. Table 3 - Main e o ms o he public pension scheme in Po ugal Sou ce: OECD epo (2019) Howe e , all hese modi ica ions we e a a pa ame ic le el, a he s uc u al le el. (Manuel e al., 2006). The ac ual s uc u e o he social secu i y sys em in Po ugal main ains he PAYG s uc u e whe e he pilla s o his model a e i mly based on he equilib ium be ween he numbe s o axpaye s and he numbe o pensione s ( o be mo e p ecise, be ween he numbe o con ibu o s and pensions) which hea ily depend on he dynamics o he popula ion size and i s age s uc u e (old-age dependency a io). A epo issued by OECD in he yea 2019 ela ed o he old-age pension sys em in Po ugal, among all OECD coun ies, Po ugal is he coun y wi h he mos signi ican decline (in pe cen age) in he size o ac i e popula ion be ween 2015 and 2045, pu ing e en mo e p essu e o he ac ual, and al eady agile, public pension scheme in Po ugal. A he na ional le el, he s udies emo e o he 1990s jus a ew yea s om he implemen a ion o social secu i y based on he Be e idge model. Consequen ly, simila o o he coun ies, he sus ainabili y o he PAYG schemes has been s udied and discussed o se e al yea s and exis s se e al 13 s udies on his heme in Po ugal appoin ing o he u gency o edesigned he public pension scheme in Po ugal (OECD, 2019). In 1995, Mendes ques ioned he implemen a ion o he uni e sal social secu i y inspi ed by he Be e idge model in 1984 due o demog aphic and employmen ends in Po ugal. He an icipa ed he ma u a ion o he Be e idge model by jus obse ed o he coun ies ha adop ed he Be e idge model ea lie han Po ugal and we e acing issues wi h he public pension sus ainabili y. Mendes p edic ed a signi ican inc ease in he ax con ibu ions paid by he ac i e membe s o he cu en pensione s, which would no pe mi indi iduals and employe s o con ibu e o he second and he hi d pilla o he pension sys em (Mendes, 1995). In 1997, due o he c isis and bank up cy o he Eu opean Social Secu i y schemes, Sil a pe o med a esea ch s udy which con i med he p e ious s udy ela ed o he ma u a ion o he PAYG schemes. Acco ding o Sil a, he Po uguese social secu i y sys em is ma u e, and i would no be sus ainable in he long e m due mainly o wo ac o s – due o demog aphic p essu e – dec ease on he e ili y a es and inc ease on he li e expec ancy, and due o he e o m on educa ion (ex ension in yea s o he manda o y schola ship) which delay he s a o he ac i e li e o Po uguese ci izens (Sil a, 1997). In 2004, Pe ei a Da Sil a o esaw he p obabili y o he popula ion aged o e 65 would be inc eased by 70 pe cen in he yea s be ween 2000 and 2050. A he same ime, he p edic s ha he p opo ion o age unde 15 will emain cons an , dec easing om 20 pe cen in 2000 o 19 pe cen in 2019. The da a p ojec ions e ealed a signi ican di e gence be ween he e olu ion o pensione s and ac i e membe s h ea ening he equilib ium o he PAYG scheme in Po ugal. He concluded he s udy ha in o de o ensu e he sus ainabili y o he public pension scheme a e he yea 2050, signi ican changes need o be made o he ac ual public pension sys em in Po ugal (Pe ei a Da Sil a, Tomé Calado, & Medei os Ga cia, 2004). In 2006, Rod igues concluded ha popula ion ageing is causing se e e isk o he sus ainabili y o public pension schemes and inc eases axes o inance he public pension sys em would be ine i able i no changes in he cu en social secu i y sys em would be pe o med a he s uc u al le el. In he las couple o yea s, se e al s udies appea s essing he u gency o he need o immedia e s uc u al changes in he public pension scheme in Po ugal (Rod igues & Pe ei a, 2006). In ecen yea s, Po ugal has implemen ed nume ous ( empo a y and pe manen ) pa ame ic pension e o ms aimed a educing public pension expendi u es, wi h li le ma gin o add essing income adequacy conce ns in an al eady highly aged socie y. The main policy ini ia i es we e d i en by he sho - e m need o iscal consolida ion a he han by a long- e m p ospec o he design o pension sys ems. Pa ame ic e o ms include: (i) he in oduc ion o a sus ainabili y ac o in he pension o mula; (ii) an inc ease in he s a u o y e i emen age; (iii) mo ing om bes yea s o li e ime a e age ea nings; (i ) a new indexa ion mechanism linked o p ices and eal GDP g ow h; ( ) he in oduc ion o bonuses o la e e i emen and penal ies o ea ly e i emen ; ( i) suspension o ea ly- e i emen egimes; ( ii) nominal pension cu s; and ( iii) means- es ing o non-con ibu o y bene i s; (ix) con e gence o ules be ween RGSS and CGA egimes. Long- e m p ojec ions show ha sys ems a e no sus ainable om he economic, inancial and demog aphic poin o e iew, and despi e he se e al wa es o pension e o ms made o e he las 20 yea s, he popula ion ageing and economic changes d i en by social and echnological changes in he labou ma ke will con inue o impac he public pension schemes and change he li ing condi ions o pensione s (J.M. B a o, 2012) (J. M. B a o, 2012) (Jo ge Miguel B a o, A onso, & Gue ei o, 2013) (B a o, J. M., Gue ei o, G., A onso, 2014) (J. M. B a o, 2017). 14 2.4. T HE FUTURE OF WORK – ADDING MORE PRESSURE S anda d and non-s anda d o m o wo k Adding e en mo e p essu e o he public pension models is he apid ans o ma ion o he wo king amewo k d i ing by he echnological ad ances and globalisa ion. The s anda d wo king model, used as a base o a pension design, is mo ing om ull- ime wo k unde an open-ended employmen con ac o a ypical o ms o wo k wi h ea u es incompa ible wi h ac ual pension designs (Schoukens & Ba io, 2017). Th ee main cha ac e is ics de ine a s anda d o m o wo k: i) di ec employmen ela ionship whe e he e is a clea dis inc ion o bila e al ela ion be ween a single employe and employee; ii) Labou s abili y whe e he wo k a angemen s a e o an inde ini e pe iod and only e mina es unde pa icula si ua ion egula ed by labou law, and iii) Income Secu i y whe e he employee ecei es a ixed income paid egula ly by he employe in exchange o hei se ices. The ac ual design o he public pension scheme is s ongly co ela ed wi h hese h ee concep s. i.e., he majo i y o he public pension schemes a e s uc u ed as an income eplacemen unde he pe cep ion ha he s anda d o wo k is he p ima y income sou ce. The u u e o wo k is challenging somehow hese h ee concep s whe e he e is no clea dis inc ion be ween a di ec single ela ionship employe and employee, he income made by he employee along he wo k ca ee is unp edic able as well he du a ion o he wo k assignmen s (Schoukens & Ba io, 2017). The non-s anda d o ms o wo k ha e he ollowing issues ela ed o old-age p o ec ion co e age due o he lack o one o all o he h ee main ea u es ha de ine he s anda d o m o wo k (Runde & Milne , 2018) (ILO, 2019): - Labou s abili y: Uns able wo k ca ee s, discon inuous ca ee s and sho wo king hou s esul ing in inadequa e co e age o limi ed bene i s. Pension schemes o en equi e minimum yea s o con ibu ions o eligibili y o ecei e a e i emen bene i , and he e o e indi idual co e ed by he a ypical wo k o ms a e usually pushed ou o social secu i y sys em, e en hough hey may accumula e o e ime a mul i ude o ixed/pa - ime wo k assignmen s bu no enough o be eligible o social insu ance pu poses; - Income Secu i y: Usually, lowe ea nings and wi h high a iance on he ea nings. O en, insu ance-based and non-con ibu o y bene i s co-exis , and indi iduals can access only he la e (e.g. only basic pension and no ea nings- ela ed). - Di ec single employe ela ionship: Di icul y o ha e a clea dis inc ion be ween employe and he employee, and his is a cen al elemen o any social secu i y sys ems o paying dues ( inancing); Who should be liable o he employe con ibu ions i an employe is no well iden i ied? Sel -employed can choose o decla e lowe -con ibu ion base, bu also ecei ed lowe bene i en i lemen ; 15 - Po abili y: lack o egula ion o ensu e labou ma ke mobili y. Acqui ed en i lemen s los du ing a change in employmen s a us o job (e.g. ollowing a ansi ion om dependen o sel -employmen when igh s di e ac oss employmen s a uses, o be ween jobs i hey ied o a speci ic employmen ela ionship). Addi ionally, some ac i i ies do no ise o pension en i lemen s (e.g. sel -employmen in many ju isdic ions and casual o i egula wo k in some); - Taxa ion o pensions: The cu en “Exemp -Exemp -Taxed” (“EET”) egime adop ed in mos coun ies, whe e bo h con ibu ions and e u ns on in es men a e exemp ed om axa ion while bene i s a e ea ed as axable income upon wi hd awal, and he inc easing mobili y o e i ees challenges he adi ional de e ed axa ion a e mo i a ing signi ican ax a bi age ac oss coun ies (J. M. B a o, 2016) (J. M. B a o, 2018). Bisma ck e sus Be e idge models Analysed he e ec o he non-s anda d o ms on he Bisma ck e sus Be e idge models, he e is a gene al pe cep ion ha he e is limi ed and lowe access o he “non-s anda d wo ke s” in he Bisma ck sys ems since he bene i is linked di ec ly o he con ibu ions paid based on he sala ies; he e o e, wi h mo e equen ca ee b eaks and lowe wages. On he o he side, he Be e idge sys ems be e designed o deal wi h he impac o he e ec s o he non-s anda d o ms o wo k since he bene i is usually a la - a e bene i . Al hough his may be ue, howe e , mos o he coun ies do no ha e a pu e “Be e idge model”, and usually he e is a second- ie wi h p i a e- scheme wi h a high le el o bene i s, whe e in mos si ua ions, he non-s anda d o wo ke s no p o ec ed by his second- ie . Addi ionally, i is ques ionable whe e Be e idge sys ems, wi h an inc ease on mo e equen b eaks o wo k and lowe sala ies, migh be no inanced only by axes wi hou he suppo o wo k con ibu ions. Fu he mo e, his b ings o ano he issue abou he ai ness o he adi ional o ms o wo k inancing he i s - ie uni e sal schemes almos en i ely(Schoukens & Ba io, 2017). Non-s anda d o ms o wo k in Po ugal A ypical o ms o wo k as such as empo a y wo k, pa - ime wo k and sel -employmen ep esen al eady a hi d o all ype o employmen ela ionship in he OECD coun ies (Schoukens & Ba io, 2017). Howe e , many o he o ms o non-s anda d o wo k a e eme ging as such as pla o ms wo ke s, lexible o on-call con ac s ha also challenge he pension s uc u es wo ldwide. In his pape , he ocus is on wo ypes o non-s anda d o he wok - “Tempo a y employmen ” and Sel - employmen ” since hese a e he p edominan ype o non-s anda d wo ke s in Po ugal. 16 Figu e 1 - Po ugal Da a – Employmen Sou ce: OECD In he dashboa d abo e based on da a p oduced by he OECD o Po ugal, he e is a s eep dec ease in he non-s anda d o ms o wo k (sel -employmen , pa - ime and empo a y wo ke s), om he yea 2012 and onwa ds (g aph II), bu his dec ease linked o he unemploymen a e ha a ec ed mo e he non-s anda d wo ke s as shown in g aph IV. This explici shows e idence ha he in e up ions o wo k a ec mo e he non-s anda d ypes o wo k. Based on da a, he sel - employmen wo ke s a e he mos a ec ed by he dec ease in he employmen a e (g aph I). The non-s anda d wo ke s ep esen almos 45% o he employmen in he yea 2018. Figu e 2 - Po ugal Da a - Ea nings Sou ce: OECD The wo g aphs abo e ep esen he incidence o low pay wo ke s (ea ning less han wo- hi ds o median ea nings) and high paymen wo ke s (ea nings mo e han one-and-a-hal ime median ea nings) o ull- ime employees. The o e all dec ease on he a e age ea nings (g aph I) ma ches wi h he dec ease o he a e age ea nings o he lowe -ea nings p o ile. The highe -ea nings p o ile main ains a cons an o a sligh inc ease in he a e age ea nings o e he yea s. The dec ease o lowe pay wo ke s likely co ela ed o he sel -employed wo ke s and empo a y wo ke s, usually wi h lowe ea nings o wi h mo e luc ua ions o e he yea s han he s anda d o ms o wo k. 17 Insigh s om OECD In 2019, he OECD sugges ed wo basic concep s o ackle some o he challenges acing he old-age social p o ec ion o he non-s anda d o ms o wo k. The OECD sugges ed ying en i lemen s o indi idual wo ke s a he han speci ic employmen ela ionships o o do he opposi e and un ie bene i s om con ibu ions. The i s concep , OECD sugges ed, “Indi idualisa ion o social p o ec ion” simila o he NDC plans as implemen ed by se e al coun ies like Sweden, I aly and Poland. On his ype o plan, social p o ec ion con ibu ions made by wo ke s, employe s o he s a e on he wo ke behal linked o an indi idual no ional accoun – his measu e would ackle he high a ia ions o he sala ies as well combining incomes om di e en sou ces. Addi ionally, his concep p ese es igh s du ing job changes and ca ee b eaks and would sol e he “double con ibu ion p oblem”. On he o he hand, since his is an indi idual use, his solu ion would no inco po a e he isk-sha ing undamen al o any insu ance, he e o e, would no p o ec agains ca as ophic isks as such disabili y. Addi ionally, i is no he igh solu ion o many low-income wo ke s and pa - ime wo ke s unless he go e nmen would sponso lowe -income wo ke s. The second concep o un ying social p o ec ion om he employmen ela ionship would be by making Social Secu i y mo e Uni e sal – un ie e i emen bene i s o con ibu ions, i.e. adop a pu e Be e idge model o all ci izens. I is he ype o solu ion he e is he isk o c owding ou employe con ibu ions (OECD, 2019). 2.5. W HAT RESEARCH DONE IN P ORTUGAL TO SOLVE THE PROBLEM ? The e we e ew esea ch pape s published in he las 15 yea s o assess he long- e m sus ainabili y o he Po uguese Pension Sys em empi ically, and mos o hem ocused on pa ame ic changes, despi e he ac , as concluded om ecen esea ch pape s, ha he Po uguese Pension Sys em will no be sus ainable wi h he cu en PAYG design. Rod igues measu ed he long- e m inancial sus ainabili y o he public pension sys em in Po ugal and e alua ed a ious e o m op ions in a dynamic gene al equilib ium amewo k. The s udy ound ha he e a e limi s o wha can be achie ed h ough pa ame ic e o ms o he PAYG sys em. Indeed, o ensu e long- e m inancial sus ainabili y, as well as he adequacy o pension income, a public second pilla needs o be e ec ed once he i s pilla is signi ican ly scaled back. Re o ming he PAYG sys em canno go much u he han sol ing an es ima ed 47.6% o he p oblem o inancial sus ainabili y. E en i he PAYG pilla we e inancially sus ainable in he long e m, se ing up a second pilla would be ad ised (Rod igues & Pe ei a, 2006). Pina Manso analysed he impac o a simula ed ansi ion o an NDC scheme on he Pension Sys em in Po ugal. The esul s sugges ha e en a e wo decades o pa ame ic e o ms and pension adjus men s, he e is s ill a deeply oo ed sus ainabili y p oblem in he Po uguese Pension Sys em. The hesis used a Dynamic Mic osimula ion Model (DYNAPOR) o analyse he impac o a ansi ion om a adi ional De ined Bene i Pay-As-You-Go pension scheme o a No ional De ined Con ibu ion sys em on bo h he inancial and he social sus ainabili y o he pension sys em in Po ugal. The esul s showed ha while he NDC scena ios ou pe o m he DB-PAYG sys em in e ms o inancial sus ainabili y, i does so a he cos o he social componen . The a ious ea u es o he NDC pension 24 The pension poin cos and he alue o he poin s a e e alua ed by indexa ion a es de ined by he plan. De ining pa ame e , u, as he indexa ion a e o pension cos , and pa ame e , n, o indexa ion a e o pension alues, such as   =   (1+  ), and   =   (1+  ), we can ew i e he pension bene i o a poin sys em as ollows: =  (1+  )        And he eplacemen a io de ined as ollows:   =(1+  1+          The No ional accoun s wo k like poin s, bu he uni alue is exp ess in cu ency a he han in he numbe o poin s. In lows a e equal o con ibu ions (in cu ency), i.e., wages mul iplied by he con ibu ion a e. The no ional capi al is inc eased each yea by he no ional in e es a e, . A e i emen , he accumula ed no ional capi al is di ided by a no ional ac ua ial annui y ac o ,   as de ined by he plan ules. The pension bene i o NDC de ined as ollows:  =  (1+  )      and he eplacemen a io is:   =(1+  1+         The DC wo ks simila ly wi h he NDC plan. Howe e , he con ibu ions (in cu ency) made du ing he yea a e capi alized by he ac ual in e es a es depending on he in es men s a egy, e u n asse s, ins ead o es ablished no ional a e , and he annui y ac o is based on he ma ke condi ions a e i emen age, ins ead o a no ional annui y ac o . The accumula ed con ibu ions made o he social secu i y sys ems a e i emen age a e as ollows:   =   (1+  )    Whe e,   is he capi aliza ion o he con ibu ions made by he pensione du ing he ac i e li e. In he cases whe e he con ibu ions a e made in he o m o axes, and no capi alized, he   is null on he o mula abo e. 25 The es ima ed “ ai ” bene i a e i emen ob ains by di iding he accumula ed con ibu ions (in he o m o ax paymen s) h ough all ac i e li e ime by he expec ed u u e li e ime a e i emen age,   based on he assumed mo ali y dec emen s. Fai =    The “ac ua ial ai ness” measu ed used on his pape is he e i emen pension bene i de ined by he model di ided by he es ima ed pension ai bene i based on he accumula ed con ibu ions o ax paid, i.e. Ac ua ial Fai ness Measu e=  (  ℎ  )  Fai   (Boulhol, 2019) 26 4. RESULTS AND DISCUSSION Why doesn’ he cu en Po uguese pension scheme i well wi hin he con ex o he u u e o wo k? The g aph in igu e 7 shows he pe cen age educ ion in he old-age bene i compa ing wi h he a e age wo ke ha did no su e wo k in e up ions du ing he ac i e li e. The modelling assumes ha pos -in e up ion ea nings esume o ha o he wo ke who has no su e ed a ca ee b eake . Figu e 7 - Impac on he old-age bene i (Baseline scena io) Sou ce: Sou ce: Au ho ’s p epa a ion; No es: AW-=A e age Wage, NMW= Na ional Minimum Wage. The longe he pe iod o a ca ee b eak, he highe is he educ ion on he old-age pension en i lemen . Fo he nons anda d wo ke s ha ha e an in e up ion o 15 yea s will lose 37% o he old-age pension agains 25% o he a e age wo ke co e ed by he unemploymen p og am. As expec ed, he longe he b eak o se ice, he highe is he educ ion impac on he old-age pension en i lemen . The a e age wo ke co e ed by he unemploymen p og am, he old-age pension can be educed by 1% o 25% as he in e up ion inc ease om 1 o 15 yea s, and o wo ke s no co e ed by unemploymen o all ea nings p o iles, he old-pension en i lemen educed be ween 2% o 35% as he in e up ion inc ease om 1 o 15 yea s. The gap be ween he wo ke p o ec ed by he unemploymen p og am and he wo ke no p o ec ed by he unemploymen p og am inc eases signi ican ly a e i e yea s o b eak. The di e ence in he wage sala ies has li le impac since we assumed he pos -in e up ion ea nings esume o ha o he wo ke who has no su e ed a ca ee b eake . The second g aph below assumes a loss o 10% o each yea o in e up ion on he pos -in e up ion wage sala y, and hen ca ch-up wi h he baseline scena io in i e yea s. -50.00% -45.00% -40.00% -35.00% -30.00% -25.00% -20.00% -15.00% -10.00% -5.00% 0.00% 0 3 6 9 12 15 % dec ease Yea s o b eak AW MNW 0.5 AW 0.75 AW 1.5 AW 2 AW 3 AW 27 Figu e 8 - impac on he old-age bene i (con e gen scena io) Sou ce: Sou ce: Au ho ’s p epa a ion; No es: AW-=A e age Wage, NMW= Na ional Minimum Wage. Iden ical o he i s scena io, smalle b eaks ha e li le impac on he old-age pension en i lemen s. Howe e , a e h ee yea s o in e up ion, he pension en i lemen can be educed by 5% wi h h ee yea s b eak o 37% wi h 15 yea s b eak o wo ke s co e ed by unemploymen p og am and om 8% wi h h ee yea s b eaks o 49% wi h 15 yea s b eak o wo ke s no co e ed by unemploymen p og am o all p o ile’s ea nings. The e is an e idence o he inadequacy o he ac ual Po uguese pension model o nons anda d wo ke s wi h mo e equen in e up ions on hei ca ee compa ed wi h he s anda d wo ke s, and in his case, hese wo ke s could lose be ween 2% o 50% o he pension en i lemen depending on he leng h o in e up ions. The e is also e idence o nons anda d wo ke s no co e ed by he unemploymen p og am ha can lose mo e han 5% o 12% compa ed wi h a e age wo ke s wi h he same leng h o pe iods b eaks du ing he ac i e li e. In summa y, he s udy eaches he same conclusion has B a o (He ce & B a o, 2017). The exis ence o con ibu ion c edi s a angemen s somehow p o ec s wo ke s om losing pension en i lemen s, and he mul iple b eaks ha e a highe impac han he single b eaks, mo e equen in e up ions ha e a mo e signi ican impac on he pension en i lemen s, and he ac ual DB sys em is inadequa e o nons anda d wo ke s. A e o he models be e o han he DB model o applied by he Po uguese old-age pension scheme? The ables below show he pe cen age o educ ion on he old-age pension en i lemen s based on he Baseline scena io and he “con e gen ea nings” scena io. -50.00% -45.00% -40.00% -35.00% -30.00% -25.00% -20.00% -15.00% -10.00% -5.00% 0.00% 0 3 6 9 12 15 % dec ease Yea s o b eak AW MNW 0.5 AW 0.75 AW 1.5 AW 2 AW 3 AW 28 Table 4 - Reduc ion on pension en i lemen s by model – Baseline Scena io Sou ce: Au ho ’s p epa a ion; No es: DB=De ined Bene i ; NDC=No ional De ined Con ibu ion; DC (o FDC)=Funded De ined Con ibu ion; AW-=A e age Wage, NMW= Na ional Minimum Wage. The loss o pension en i lemen is lowe o he a e age wo ke since he s udy assumed ha he Po uguese unemploymen p og am does no co e he nons anda d wo ke . All he models ha e simila losses on he pension en i lemen s— he same si ua ion o he “con e gen ” scena io, as shown below. Howe e , o he “Poin s Sys ems” he e is a maximum loss educ ion o 37% o highe ea nings p o ile because o he ceiling amoun . In summa y, he o he pension models like he cu en DB Po uguese pension model shows e idence o signi ican pension en i lemen losses o he nons anda d wo ke s and biased di e ences be ween he a e age wo ke s and nons anda d wo ke s. 1 y mul iple b eaks o wo k impac Baseline AW MNW 0.5 AW 0.75 AW 1.5 AW 2 AW 3 AW Pu e DB plan 0% -2% -7% -7% -7% -7% -7% -7% NDC Plan 0% -2% -7% -7% -7% -7% -7% -7% Basic Pension + Cash Balance 0% -4% -7% -7% -7% -7% -7% -7% Floo O se Plan (DC) 0% -3% -7% -7% -7% -7% -7% -7% Poin s Sys ems 0% -2% -7% -7% -7% -7% -7% -7% 2 y s mul iple b eaks o wo k impac Baseline AW MNW 0.5 AW 0.75 AW 1.5 AW 2 AW 3 AW Pu e DB plan 0% -5% -15% -14% -14% -14% -14% -14% NDC Plan 0% -5% -15% -14% -14% -14% -14% -14% Basic Pension + Cash Balance 0% -9% -14% -14% -14% -14% -14% -14% Floo O se Plan 0% -6% -15% -14% -14% -14% -14% -14% Poin s Sys ems 0% -5% -14% -14% -14% -15% -15% -15% 3 y s mul iple b eaks o wo k impac Baseline AW MNW 0.5 AW 0.75 AW 1.5 AW 2 AW 3 AW Pu e DB plan 0% -10% -22% -22% -22% -22% -22% -22% NDC Plan 0% -9% -22% -22% -22% -22% -22% -22% Basic Pension + Cash Balance 0% -14% -21% -20% -21% -22% -21% -21% Floo O se Plan (DC) 0% -10% -22% -22% -22% -22% -22% -22% Poin s Sys ems 0% -10% -21% -21% -21% -22% -22% -22% 4 y s mul iple b eaks o wo k impac Baseline AW MNW 0.5 AW 0.75 AW 1.5 AW 2 AW 3 AW Pu e DB plan 0% -18% -30% -29% -29% -29% -29% -29% NDC Plan 0% -16% -30% -29% -29% -29% -29% -29% Basic Pension + Cash Balance 0% -21% -28% -27% -28% -29% -28% -28% Floo O se Plan (DC) 0% -18% -30% -29% -29% -29% -29% -29% Poin s Sys ems 0% -17% -29% -29% -29% -30% -30% -30% 5 y s mul iple b eaks o wo k impac Baseline AW MNW 0.5 AW 0.75 AW 1.5 AW 2 AW 3 AW Pu e DB plan 0% -25% -38% -37% -37% -37% -37% -37% NDC Plan 0% -22% -38% -37% -37% -37% -37% -37% Basic Pension + Cash Balance 0% -28% -36% -34% -35% -36% -35% -35% Floo O se Plan (DC) 0% -26% -38% -37% -37% -37% -37% -37% Poin s Sys ems 0% -25% -37% -37% -37% -37% -37% -37% 29 Table 5 - Reduc ion on pension en i lemen s by model – Con e gen Scena io Sou ce: Au ho ’s p epa a ion; No es: DB=De ined Bene i , NDC=No ional De ined Con ibu ion, DC (o FDC)=Funded De ined Con ibu ion; AW-=A e age Wage, NMW= Na ional Minimum Wage. How should be he Po uguese pension scheme design o i wi hin he con ex o he u u e o wo k? The adequacy o a pension model measu ed by he le el o eplacemen a io, i.e., lowe eplacemen a io measu es imply an inadequa e pension model, and highe eplacemen a ios be e he adequacy o he pension model. On he o he hand, he ac ua ial ai ness, as desc ibed in sec ion h ee, measu es he ai ness be ween he con ibu ions and he bene i o he pension model. Lowe he ac ua ial ai ness measu e be e is he balance be ween he in lows and ou lows o he pension model, and highe is he ac ua ial ai ness measu e mo e unbalanced is he link be ween ou lows and in lows o he pension model. The e o e, he bes model should ha e he maximum eplacemen a io and he minimum ac ua ial ai ness measu es, and i is he balance be ween hese wo measu es ha we may ind he op imal pension model. Addi ionally, highe ac ua ial ai ness measu es o lowe ea nings p o iles and lowe ac ua ial ai ness measu es o highe p o ile ea nings implies a dis ibu i eness ea u e o a pension model wi hin gene a ions. (Con e gen scena io) 1 y mul iple b eaks o wo k impac Baseline AW MNW 0.5 AW 0.75 AW 1.5 AW 2 AW 3 AW Pu e DB plan 0% -3% -8% -8% -8% -8% -8% -8% NDC Plan 0% -3% -8% -8% -8% -8% -8% -8% Basic Pension + Cash Balance 0% -5% -8% -7% -8% -8% -7% -7% Floo O se Plan (DC) 0% -3% -8% -8% -8% -8% -8% -8% Poin s Sys ems 0% -3% -8% -8% -8% -7% -7% -7% 2 y s mul iple b eaks o wo k impac Baseline AW MNW 0.5 AW 0.75 AW 1.5 AW 2 AW 3 AW Pu e DB plan 0% -8% -17% -17% -17% -17% -17% -17% NDC Plan 0% -7% -17% -17% -17% -17% -17% -17% Basic Pension + Cash Balance 0% -11% -16% -16% -16% -16% -16% -15% Floo O se Plan 0% -8% -17% -17% -17% -17% -17% -17% Poin s Sys ems 0% -8% -16% -16% -16% -15% -15% -15% 3 y s mul iple b eaks o wo k impac Baseline AW MNW 0.5 AW 0.75 AW 1.5 AW 2 AW 3 AW Pu e DB plan 0% -15% -27% -26% -26% -26% -26% -26% NDC Plan 0% -13% -27% -26% -26% -26% -26% -26% Basic Pension + Cash Balance 0% -19% -26% -24% -26% -26% -25% -23% Floo O se Plan (DC) 0% -15% -27% -26% -26% -26% -26% -26% Poin s Sys ems 0% -14% -26% -26% -26% -22% -22% -22% 4 y s mul iple b eaks o wo k impac Baseline AW MNW 0.5 AW 0.75 AW 1.5 AW 2 AW 3 AW Pu e DB plan 0% -25% -35% -37% -34% -37% -37% -37% NDC Plan 0% -22% -38% -37% -37% -37% -37% -37% Basic Pension + Cash Balance 0% -30% -37% -34% -36% -36% -34% -32% Floo O se Plan (DC) 0% -26% -38% -37% -37% -37% -37% -37% Poin s Sys ems 0% -26% -37% -37% -37% -30% -30% -30% 5 y s mul iple b eaks o wo k impac Baseline AW MNW 0.5 AW 0.75 AW 1.5 AW 2 AW 3 AW Pu e DB plan 0% -37% -47% -48% -46% -49% -49% -49% NDC Plan 0% -33% -50% -49% -49% -49% -49% -49% Basic Pension + Cash Balance 0% -43% -49% -44% -48% -47% -45% -40% Floo O se Plan (DC) 0% -38% -50% -49% -49% -49% -49% -49% Poin s Sys ems 0% -38% -50% -50% -50% -37% -37% -37% 30 The cu en eplacemen a ios and ac ua ial ai ness measu es o he DB Po uguese model o he di e en p o iles assuming no in e up ions in wo k a e as ollows: Table 6 – Measu es sco es o he cu en Po uguese DB plan by p o ile Sou ce: Au ho ’s p epa a ion; No es: DB=De ined Bene i ; Measu e Sco e de ined on sec ion 3.2.; AW-=A e age Wage, NMW= Na ional Minimum Wage. The ac ua ial ai ness measu e o he a e age wo ke is lowe compa ed wi h he o he p o iles due o he sligh ly highe con ibu ions a es du ing ac i e li e pe iod. The e is li le edis ibu ion wi hin gene a ions in he DB Po uguese model since he a iance o he ac ua ial ai ness measu es be ween highe ea nings p o iles, and lowe ea nings p o ile is minimal. Howe e , he ac ua ial ai ness measu es a e conside ably highe o all ea nings p o ile pu ing in cause he sus ainabili y o he ac ual pension model. In summa y, he majo issue o he DB pension model in Po ugal is he conside able highe ac ua ial ai ness measu es be ween all p o iles and low le els o edis ibu ion wi hin he same gene a ion. The e o e, he i s s ep o design a pension model adequa e o nons anda d and s anda d wo ke s is o ha e lowe ac ua ial ai ness measu es o highe ea nings p o iles and inc ease he le els o edis ibu ion wi hin he same gene a ion. Wha would be he ac ua ial ai ness o o he pension models wi h he same eplacemen a io measu es as he DB plan o Po ugal o he a e age wo ke ? The models a e easily adjus ed o ob ain he same eplacemen a io measu e o he cu en DB Po uguese DB plan by eadjus ing he pa ame e s o each pension model o he a e age wo ke . Fo he NDC pension model, a eplacemen a io o 46% a ained o an a e age wo ke by se ing up he no ional asse e u n o 3% and discoun a e used on he no ional con e sion ac o o 3.25% ins ead o he assumed hypo he ical discoun a e o 3.00%. Table 7 – Compa ison measu es sco es by p o ile (DB e sus NDC) Sou ce: Au ho ’s p epa a ion; No es: Pu e DB= he cu en Po uguese De ined Bene i plan; NDC plan = No ional De ined Plan; Measu e sco es de ined on sec ion 3.2.; AW-=A e age Wage, NMW= Na ional Minimum Wage. The ac ua ial ai ness sligh ly imp o ed o he a e age wo ke bu no su icien ly enough o be conside ed a signi ican ad an age. Howe e , o he nons anda d wo ke s, he ac ua ial ai ness measu es imp o ed conside able, bu he eplacemen a io measu e dec ease be ween 5 o 10 basis poin s. The edis ibu i e le el wi hin gene a ions by analysed he di e ence in a iances o he ac ua ial ai ness measu es be ween p o iles is inexis en on his ype o model. Fo he DC pension model, a eplacemen a io o 46% a ained o an a e age wo ke by assuming a e u n on asse s by 2.90% p.a. P o ile AW MNW 0.5 AW 0.75 AW 1.5 AW 2 AW 3 AW Pu e DB plan Replacemen Ra io 46% 45% 50% 46% 45% 45% 44% Ac ua ial Fai ness 96% 131% 149% 132% 127% 124% 120% DB e sus NDC B eaks o wo k impac Baseline AW MNW 0.5 AW 0.75 AW 1.5 AW 2 AW 3 AW Pu e DB plan Replacemen Ra io 46% 46% 45% 50% 46% 45% 45% 44% Ac ua ial Fai ness 96% 96% 131% 149% 132% 127% 124% 120% NDC Plan Replacemen Ra io 46% 46% 38% 39% 39% 39% 39% 39% Ac ua ial Fai ness 88% 88% 84% 86% 86% 86% 86% 86% 31 Table 8 – Compa ison measu es sco es by p o ile (DB e sus DC) Sou ce: Au ho ’s p epa a ion; No es: Pu e DB= he cu en Po uguese De ined Bene i plan; DC (o FDC) plan = Funded Con ibu ion De ined Plan; Measu e sco es de ined on sec ion 3.2.; AW-=A e age Wage, NMW= Na ional Minimum Wage. The e is e idence ha he DC model is a pension design mo e a ac i e han he DB model in e ms o sus ainabili y because i has shallow ac ua ial ai ness measu es o all p o iles. The eason is ha he isk is wi hin he wo ke s. The 20% ac ua ial ai ness is due mos ly due o he di e ence be ween he ac ua ial ac o used o con e he accumula ed con ibu ions in o an annui y and he expec ed a e age li e ime o an indi idual a age 66.5 yea s. Like he NDC model, he eplacemen a io is lowe o nons anda d wo ks compa ed wi h he Po uguese DB plan, and he edis ibu ion wi hin a gene a ion is inexis en . Howe e , he mos signi ican ad an age o he DC plan is he con ibu ions made by he axpaye s a e in es ed and consequen ly capi alised, and he con ibu ions unde he NDC used di ec ly o pay he pension o he cu en pensione s. This has an impac o a ound 70% inc ease on he ac ua ial ai ness measu e. Fo he Swiss pension model, a eplacemen a io o 46% a ained o an a e age wo ke by changing he minimum pension ac o o 1.65 (pa ame e o he Basic Plan). Table 9 – Compa ison measu es sco es by p o ile (DB e sus Basic Plan+Cash Balance Plan) Sou ce: Au ho ’s p epa a ion; No es: Pu e DB= he cu en Po uguese De ined Bene i plan; Measu e sco es de ined on sec ion 3.2.; AW-=A e age Wage, NMW= Na ional Minimum Wage. The e is e idence ha he Swiss model is be e compa ed wi h he DB Po uguese model by sco ing high on all measu es. The eplacemen a io measu es a e highe o lowe ea nings han o highe ea nings. The ac ua ial ai ness measu es a e signi ican ly be e compa ing wi h he DB Po uguese model, and he e is a be e equilib ium be ween eplacemen a io and ac ua ial ai ness measu es. Fo highe ea nings, he eplacemen a io is lowe compa ed wi h he Po uguese DB plan because o he maximum pension limi on he basic pension. B eaks o wo k impac Baseline AW MNW 0.5 AW 0.75 AW 1.5 AW 2 AW 3 AW Pu e DB plan Replacemen Ra io 46% 46% 45% 50% 46% 45% 45% 44% Ac ua ial Fai ness 96% 96% 131% 149% 132% 127% 124% 120% Floo O se Plan (DC) Replacemen Ra io 46% 46% 38% 39% 39% 39% 39% 39% Ac ua ial Fai ness 20% 20% 20% 20% 20% 20% 20% 20% Baseline AW MNW 0.5 AW 0.75 AW 1.5 AW 2 AW 3 AW Ac ual DB Plan Replacemen Ra io 46% 46% 47% 53% 48% 43% 40% 35% Ac ua ial Fai ness 97% 97% 139% 164% 140% 117% 101% 77% Swiss Plan - Basic Pension+Cash Balance Replacemen Ra io 46% 46% 50% 60% 51% 40% 37% 32% Ac ua ial Fai ness 66% 66% 109% 147% 107% 65% 51% 30% 32 Fo he Poin s Sys em, a eplacemen a io o 46% a ained o he a e age wo ke by changing he pension alue o 16.75 eu os. Table 10– Compa ison measu es sco es by p o ile (DB e sus Basic Plan + Poin s Sys ems) Sou ce: Au ho ’s p epa a ion; No es: Pu e DB= he cu en Po uguese De ined Bene i plan; Measu e sco es de ined on sec ion 3.2.; AW-=A e age Wage, NMW= Na ional Minimum Wage. The esul s o he poin s sys em pension models a e iden ical o he DB Po uguese plan o all p o iles excep o he highe ea nings because o he ceiling, bu his also gi es he possibili y o highe -ea ning p o iles o sa e he excess in a p i a e pension scheme. In summa y, somehow all models we e be e o han he ac ual DB Po uguese model in e ms o ac ua ial ai ness measu es, bu in e ms o adequacy, he DB Po uguese model is sligh e be e han he o he models excep o he Swiss model o lowe ea nings p o ile and he a e age wo ke . Could we change he pa ame e s o he DB Po uguese model o imp o e he ac ua ial ai ness and keep he same eplacemen a io o a e age wo ke s? The answe is yes. By se ing up highe acc ual a es o he 1 s ie om 2.30% p.a. o 2.50% p.a. and lowe acc ual a es o he 3 d ie , 4 h ie and 5 h ie om 2.20%, 2.10%, 2.00% o 2.00%, 1.50% and 0.75%, espec i ely. Table 11 – Compa ison measu es sco es by p o ile (Ac ual DB e sus DB wi h new pa ame e s) Sou ce: Au ho ’s p epa a ion; No es: DB= De ined Bene i plan; Measu e sco es de ined on sec ion 3.2.; AW-=A e age Wage, NMW= Na ional Minimum Wage. Changing he DB plan o highe a es o lowe ie s and lowe a es o highe ie s would imp o e he ac ua ial ai ness le els o highe p o ile ea ne s and he edis ibu i e le els wi hin gene a ions bu s ill wi h e y high ac ua ial ai ness measu es p oducing an unbalance scena io be ween con ibu ion (in lows) and bene i (ou lows) o he model in he long e m. The eplacemen a io dec eased o highe ea nings p o iles om 5 o 10 basis poin s. Wi h hese pa ame e s o he DB plan, he eplacemen a io measu es a e like he Swiss models, bu he ac ua ial ai ness measu es o he Swiss models a e s ill conside ably lowe han he DB model as shown in he igu e abo e making he Swiss model much mo e a ac i e. DB e sus Poin s Sys ems B eaks o wo k impac Baseline AW MNW 0.5 AW 0.75 AW 1.5 AW 2 AW 3 AW Pu e DB plan Replacemen Ra io 46% 46% 45% 50% 46% 45% 45% 44% Ac ua ial Fai ness 96% 96% 131% 149% 132% 127% 124% 120% Poin s Sys ems Replacemen Ra io 46% 46% 45% 46% 46% 31% 23% 16% Ac ua ial Fai ness 97% 97% 131% 131% 131% 100% 100% 100% Ac ual DB e us new DB (di e en pa ame e s) B eaks o wo k impac Baseline AW MNW 0.5 AW 0.75 AW 1.5 AW 2 AW 3 AW Pu e DB plan Replacemen Ra io 46% 46% 45% 50% 46% 45% 45% 44% Ac ua ial Fai ness 96% 96% 131% 149% 132% 127% 124% 120% Pu e DB plan - new acc ual a es Replacemen Ra io 46% 46% 47% 53% 48% 43% 40% 35% Ac ua ial Fai ness 97% 97% 139% 164% 140% 117% 101% 77% 33 Table 12 – Compa ison measu es sco es by p o ile (DB wi h new pa ame e s e sus Cash Balance) Sou ce: Au ho ’s p epa a ion; No es: DB= New De ined Bene i plan; Measu e sco es de ined on sec ion 3.2.; AW-=A e age Wage, NMW= Na ional Minimum Wage. The eason o lowe eplacemen a io o he Swiss plan is he limi o he Basic Pension as he igu e below illus a es. Table 13 – Replacemen Ra io by sub-Plan (Swiss model= Basic Pension + Cash Balance) Sou ce: Au ho ’s p epa a ion; No es: Basic Plan= De ined Bene i ype plan; Replacemen Ra io de ined on sec ion 3.2.; AW- =A e age Wage, NMW= Na ional Minimum Wage. Based on he abo e, he e is e idence ha he Swiss model is a be e s a o an op imal model wi h a be e balance be ween eplacemen a io and ac ua ial ai ness. The nex ques ion o suppo us on he ques o an op imal model would be: Wha is he eason o he Swiss model o be be e o han he o he models? The Swiss model in eg a ed wi h wo sub-models, he i s model a DB plan (Basic Plan) and a second model a cash balance plan. Bo h sub-models complemen each o he . The igu e below shows he eplacemen a io and ac ua ial ai ness measu es sepa a ely o each sub-model. Table 14 – Measu e sco e o Swiss plan (Basic Pension + Cash Balance) Sou ce: Au ho ’s p epa a ion; Basic Plan= De ined Bene i ype plan; Measu e sco es de ined on sec ion 3.2.; AW-=A e age Wage, NMW= Na ional Minimum Wage. The cash balance alone would beha e exac ly like he DC o NDC pension models wi hou edis ibu ion wi hin he gene a ion and wi h lowe eplacemen a ios compa ed wi h he DB Po uguese model. The sub-model Basic pension (DB ype o model) in oduced by he Swiss model inc eased he eplacemen a io and p o ided he e ibu ion ea u e impo an o any social secu i y co e age. These wo sub-models combined as a single model ha e be e measu es as he DB plan and all o he models alone o wi h a minimum pension gua an ee. DB e sus Cash Balance B eaks o wo k impac Baseline AW MNW 0.5 AW 0.75 AW 1.5 AW 2 AW 3 AW Pu e DB plan Replacemen Ra io 46% 46% 47% 53% 48% 43% 40% 35% Ac ua ial Fai ness 97% 97% 139% 164% 140% 117% 101% 77% Basic Pension + Cash Balance Replacemen Ra io 46% 46% 49% 58% 50% 41% 38% 34% Ac ua ial Fai ness 47% 47% 81% 109% 79% 48% 38% 22% Swiss Model Replacemen Ra io Baseline AW MNW 0.5 AW 0.75 AW 1.5 AW 2 AW 3 AW Basic Plan 23% 23% 27% 35% 27% 18% 15% 11% Cash Balance 23% 23% 22% 23% 23% 23% 23% 23% To al 46% 46% 49% 58% 50% 41% 38% 34% Swiss Model B eaks o wo k impac Baseline AW MNW 0.5 AW 0.75 AW 1.5 AW 2 AW 3 AW Basic Plan Replacemen Ra io 23% 23% 27% 35% 27% 18% 15% 11% Ac ua ial Fai ness 95% 95% 223% 319% 220% 116% 82% 30% Cash Balance Replacemen Ra io 23% 23% 22% 23% 23% 23% 23% 23% Ac ua ial Fai ness 19% 19% 19% 19% 19% 19% 19% 19% 40 A e o he models be e o han he DB model o applied by he Po uguese old-age pension scheme? - All models (NDC Plan, Swiss plan, DC plan, and Poin sys ems) had an impac on he pension en i lemen as he DB Po uguese plan. The losses a y be ween 2% o 37% o he baseline scena io and be ween 3% o 49% o he con e gen scena io. How should be he Po uguese pension scheme design o i wi hin he con ex o he u u e o wo k? - The eplacemen a io and ac ua ial ai ness used o measu e he pe o mance o he models. The eplacemen a io indica ed he adequacy o he pension model and he ac ua ial ai ness shows he ai balance indices be ween he con ibu ions and he bene i s as well as he dis ibu i eness le els o he pension model. The esul s show e idence ha all ou models used in he s udy ha e be e measu es ha he Po uguese DB plan. - The s udy shows ha he pe o mance o he models based on he wo de ined measu es is easily adjus ed by ine- uning he pa ame e s o he model. The DB Po uguese model would imp o e he ac ua ial ai ness le els o highe p o ile ea ne s and he edis ibu i e le els wi hin gene a ions by se ing up highe acc ual a es o he 1 s ie om 2.30% p.a. o 2.50% p.a. and lowe acc ual a es o he 3 d ie , 4 h ie and 5 h ie om 2.20%, 2.10%, 2.00% o 2.00%, 1.50% and 0.75%, espec i ely. Ne e heless, s ill wi h e y high ac ua ial ai ness measu es p oducing an unbalance scena io be ween con ibu ion (in lows) and bene i (ou lows) o he model in he long e m. - The combina ion o he “BD+DC” models and he “DB+NDC” ha e he same eplacemen a io, bu he ac ua ial ai ness is di e en since, unde he NDC plan, he sa ings a e no capi alized. The s udy shows e idence ha by capi alized he con ibu ions wold dec ease be ween 65% o 70% he ac ua ial ai ness measu e. - The “Swiss” model was he mos balanced model wi h a be e equilib ium be ween he eplacemen a io and ac ua ial measu es. The Swiss models based on a DB bene i (Basic Plan) plus a Cash Balance plan. Fu he mo e, he Swiss model ob ained be e adequacy measu e o nons anda d wo ke s by changing he basic plan o a ge bene i . The a ge bene i designed based on he a e age pensionable sala y and no dependen on he pensionable se ice. The pension en i lemen loss unde he basic scena io o lowe ea nings is 10% wi h 15 yea s o mul iple in e up ion se ice compa ed wi h he 25% o he ac ual DB plan. Fo highe ea nings, he loss would a y be ween 15% o 20% bu s ill mus mo e accep able compa ed wi h he cu en pension Po uguese model. The esul s a e also mo e sa is ac o y o he con e gen scena io whe e he loss o highe ea nings p o ile is be ween 25% o 27%, and o lowe ea nings p o iles would be lowe han 25% wi h 15 yea s mul iple b eaks compa ed wi h he loss o 49% o he cu en DB model. 41 Based on he empi ical esul s summa ized abo e, we concluded he ollowing: 1) All s andalone pension models a e alike ela ed o adequacy le els (i.e. wi h he same eplacemen a io measu es) by adjus ing he pa ame e s— he di e ence be ween he models ound in he sus ainabili y le els and he espec i e ac ua ial ai ness measu es. The Cash Balance, DC and NDC plans a e no edis ibu i e wi hin he same gene a ion, and in gene al sco e lowe in e ms o ac ua ial ai ness measu es compa e wi h DB plan and sys ems poin s, bu his depends on he pa ame e s. The DC o Cash Balance ha e be e measu es ela ed o ac ua ial ai ness since he con ibu ions in es ed and consequen ly capi alised. The DB plans e icien ly designed o in oduce he edis ibu i e ea u e wi hin gene a ions and inc ease he adequacy le el o lowe -ea ning p o iles. 2) A s andalone pension model pe o ms wo se han a mul iple pension model. A model should ha e a leas wo sub-models wi h di e en ypes – Be e idge and Bisma ck models. The in eg a ion o he wo models would inc ease he adequacy and sus ainabili y le els. 3) The Bisma ck and Be e idge model complemen each o he and inco po a ing hese wo models in o a single model wi h he op imal pa ame e s would p o ide he ou main p inciples o a pension model – po e y elie ; edis ibu ion, insu ance and consump ion smoo h. 4) Mul iple models wi h a leas wo ie s could be mo e adequa e o nons anda d wo ke s; howe e , he i s sub-model DB plan should no be linked o he pensionable se ice like a Pu e Be e idge model – a la bene i and he second plan should be like a Bisma ck plan. 5) The con ibu ions o he second- ie model bene i s should be in es ed - his has a dec easing e ec o a ound 65% o 70% on he ac ua ial ai ness measu e. 6) A ma hema ical app oach should be aken in analysed he pa ame e s o a pension model, and maybe his esumes o an op imisa ion p oblem o ind he op imal pa ame e s o pension model - Wha a e he bes pa ame e s o each model in o de o maximise he eplacemen a io and minimise he ac ua ial ai ness? 7) The pa ame e s should be co ela ed wi h he economic and demog aphic ac o s – he main d i e s o he sus ainabili y o he pension model; The pension model could be u he imp o ed o nons anda d wo ke s by emo ing he ix e i emen age ( ha was used as an assump ion) and le he indi idual decided he age o e i emen based on he p og ession o he adequacy o he old-age pension en i lemen , and open he possibili y o pa ially ecei ing he old-age pension ins ead ull pension and con inue o con ibu e o accumula ed es ed igh s in o he u u e. Ha ing in mind he objec i es o he s udy, p opose a new design o a pension model wi hin he new con ex o u u e o wo k, he main ecommenda ions based on he conclusions o he s udy, would be as ollows: 42 Pape conclusion Recommenda ion change o he i s - ie model o he public pension model E ec 1. The Bisma ck and Be e idge model complemen each o he ; 2. Mul iple models wi h a leas wo ie s could be adequa e o nons anda d wo ke s; 3. The con ibu ions o he second- ie model bene i s should be in es ed Change he ac ual single DB plan o a mul i-model wi h he i s model o be a la bene i (pu e Be e idge plan) wi h a eplacemen a io a ge be ween 35% o 40% and a second model o be a Cash Balance plan wi h a minimum gua an ee e u n a e. The ax paid by he pa icipan membe s edis ibu ed be ween he wo models, i.e., pa o he ax paymen would be o sponso he la bene i , and he o he amoun in es ed as a o m o a con ibu ion o he Cash Balance Plan. This measu e would imp o e he ai ness ela ed o he link be ween he con ibu ions and bene i o he model and he adequacy le els o he nons anda d o m o wo k. 4. A ma hema ical app oach should be aken in analyzing he pa ame e s o a pension model; 5. The pa ame e s should be co ela ed wi h he economic and demog aphic ac o s The pa ame e s o he model should be based on a ma hema ical “op imal” p oblem. Es ablished an objec i e pension based on he adequacy and ai ness measu es, and ind ou he pa ame e ha maximizes he objec i e unc ion. The pa ame e s should be s ongly co ela ed wi h demog aphic and economic ac o . This measu e would imp o e he obus ness o he pension models agains demog aphic and economic shocks. In Po ugal, mos o he discussions ha e been a ound pa ame e s o he cu en DB pension model. Howe e , he e a e al eady some in es iga o s/p o essionals o he pension ield like B a o ha a e challenging he pension p o essionals o s a o ha e a di e en pe spec i e on Po uguese pension model and open he doo o o he be e al e na i es o he se ious p oblem acing he ac ual pension model. The p oblem o he cu en DB model is a ques ion o adequacy and sus ainabili y, and he e a e al eady a ew s udies showed a be e pa h o a pension model in Po ugal. The successi e go e nmen s ha e been pos poning he edesign o he pension model Po ugal because he p oblem is no immedia e. Howe e , he e o m is ine i able, and he change om a DB o mul iple models including a Bisma ck plan will be necessa y no only because he ac ual DB plan is unsus ainabili y bu also because o new s anda ds o ms o wo k. Finally, i is impo an o s ess ha his pape ocused only on he old-age isk co e age, bu he e is also an u gen need o edesign and e hink he o he p o ec ion co e ages p o ide by he Po uguese social secu i y as such as unemploymen , sickness and disabili y co e age o he non-s anda d wo ke s. 43 6. LIMITATIONS AND RECOMMENDATIONS FOR FUTURE WORKS The simula ion pe o med on he cases s udies we e unde a limi ange o assump ions and based on an ac i e male membe , bu he changing he assump ions on e u n on asse s o he DC plan, he mo ali y able as well o he assump ions could a ec he esul s o he pension models, and his is a limi a ion o his s udy. In his s udy, we also no men ioned he impac o he Gende , longe i y, and he homogenei y ac o s on he models, and he e o e, how should be a design o a model o inco po a e all hese ac o s and o be ai o all ci izens – s anda d wo ke s, nons anda d wo ke s, emales wo ke s, males wo ke s, Blue colla s wo ke s and whi e-colla wo ke s. In espec o ecommenda ion o u u e wo ks, he s udy shows some e idence ha by changing he ac ual DB pension plan in Po ugal o mul iple models wi h DB plus DC ype would imp o e he adequacy and sus ainabili y le els o all wo ke s inclusi e nons anda d wo ke s. Howe e , his ansi ion will no be possible wi hou a de ailed plan design o his ansi ion. Pa o he ax con ibu ions o he sub-DC plan is dis ibu ed indi idually, he e o e, educing he le el o edis ibu ion be ween gene a ions. The sho ness o axpaye s o pay he cu en pensione s would p o oke a dis up ion on he sus ainabili y o he al eady agile Po uguese DB model. The e o e, his should be a u u e s udy o conside how o ansi ion om a DB plan o o he plan design wi hou dis epu e all he pension sys em in Po ugal. Ano he s udy would be pe o ming a ma hema ical s udy o ind ou a model ha could calcula e he op imal pa ame e s o he di e en possible models o minimise he ac ua ial ai ness measu es as well as he po ion o alloca ions o he con ibu ions o each sub-model in case o a mul iple pension model. Finally, his pape showed e idence ha mul iple models wi h wo sub-model p o ides be e adequacy, sus ainabili y and dis ibu i eness measu es bu would be wise o analyse he beha e o he mul iple models wi h mo e han wo models. 44 7. BIBLIOGRAPHY Alho, J., B a o, J. M. & Palme , E. (2013). Annui ies and Li e Expec ancy in NDC. In: Holzmann, Robe , Edwa d Palme , and Da id Robalino, Edi o s. Non inancial De ined Con ibu ion Pension Schemes in a Changing Pension Wo ld: Gende , Poli ics, and Financial S abili y, 2, 395– 4(Washing on: Wo ld Bank Publica ions.). A lijas, S. (2019). The Dynamism o he New Economy: Non-S anda d Employmen and Access o Social Secu i y in EU-28. SSRN Elec onic Jou nal, (141). h ps://doi.o g/10.2139/ss n.3354463 Ayuso, M., B a o, J. M., and Holzmann, R. (2020). Ge ing Li e Expec ancy Es ima es Righ o Pension Policy: Pe iod e sus Coho App oachTi le. Jou nal o Pension Economics and Finance, 1—20. Ba , N. (2000). Ba - Re o ming Pensions My h T u hs and Policy Choices.pd . Boulhol, H. (2019). Economics Depa men Objec i es and Challenges in he Implemen a ion o a Uni e sal Pension Sys em in F ance. Re ie ed om www.oecd.o g/eco/wo kingpape s. B a o, J. M., Gue ei o, G., A onso, L. (2014). A aliação Ac ua ial do Sis ema P e idencial da Segu ança Social e P es ação Única da Segu ança Social. . . GEP - Minis é io Da Solida iedade, Emp ego e Segu ança Social, Lisboa, Dezemb o. Pa ially A ailable by GEP A . B a o, J. M. (2012). Sis emas de Segu ança Social em Po ugal: A qui ec u a de um No o Modelo Social e Con ibu os pa a o Deba e sob e a Re o ma do Regime de Pensões. Associação Po uguesa de Es udos Sob e A o o, In es imen o e Pensões de Re o ma, Dezemb o 2012. Re ie ed om p://www.ap ipp.p /backo ice/box/use iles/ ile/Rela ó io_APFIPP_Pa 2_Final.pd B a o, J. M. (2015). Li ing longe and p ospe ing? Opções de edesenho dos sis emas de pensões em Po ugal. In Ne o, Paulo e Se ano, Ma ia Manuel (Coo ds.), Polí icas(Nexo Li e á io, Alcoche e), 139–168. B a o, J. M. (2016). Taxa ion o Pensions in Po ugal: A Semi-Dual Income Tax Sys em. CESi o DICE Repo - Jou nal o Ins i u ional Compa isons, 14 (1), 14–23. B a o, J. M. (2017). Con a os in e ge acionais e consis ência empo al na ges ão da p o ecção social: Implicações Polí icas e Re o ma do Sis ema de Pensões. “En elhecimen o Na Sociedade Po uguesa: Pensões, Família e Cuidados”, ICS: Imp ensa de Ciências Sociais, Uni e sidade de Lisboa, 61–96. B a o, J. M. (2018). Taxa ion o Pensions in Po ugal: Is The e a Ra ionale o a Semidual Income Tax Sys em? Robe Holzmann, & John Piggo (Eds.), The Taxa ion o Pensions, (The MIT P ess), 135–166. B a o, J. M. (2019). Funding o Longe Li es: Re i emen Walle and Risk-Sha ing Annui ies. Ekonomiaz, 96(2)(268–291). B a o, J. M. & E. M. de F. (2018). Valua ion o longe i y-linked li e annui ies. Insu ance: Ma hema ics and Economics, 78, 212–229. B a o, J.M. (2012). Sus en abilidade Financei a dos Sis emas Públicos de Segu ança Social em Po ugal : Si uação Ac ual e Análise P ospec i a. B a o, J.M. (2015). Re o ma Es u u al dos Sis emas de Pensões. In: Vi ia o S. M., Pe ei a, P. T. (Coo d.). A i ma o Fu u o: Polí icas Públicas Pa a Po ugal, Volume I-(Fundação Calous e 45 Gulbenkian, Lisboa), 264–329. B a o, J.M. (2016). Sus en abilidade, Adequação e Equidade nos Sis emas de P o ecção Social: O Desa io da Pi âmide E á ia In e ida. Ins i u o Da De esa Nacional (Coo d.). Polí icas Públicas e o Papel Do Es ado No Século XXI - Ciclo de Mesas Redondas “Te Es ado”, Lisboa, IDN-Ins (Coleção A ena, n o 36), 295–327. B a o J.M. & Coelho E. (2020). Modelling Mon hly Bi hs and Dea hs Using Seasonal Fo ecas ing Me hods as an Inpu o Popula ion Es ima es. In: Skiadas C., Skiadas C. (Eds) Demog aphy o Popula ion Heal h, Aging and Heal h Expendi u es, . The Sp i, 203–222. Re ie ed om h p://doi-o g-443.web pn. jmu.edu.cn/10.1007/978-3-030-44695-6_14 B a o, Jo ge M., & He ce, J. A. (2020). Ca ee b eaks, b oken pensions? Long- un e ec s o ea ly and la e-ca ee unemploymen spells on pension en i lemen s. Jou nal o Pension Economics and Finance, 1–27. h ps://doi.o g/10.1017/S1474747220000189 B a o, Jo ge Miguel, A onso, L. B., & Gue ei o, G. R. (2013). A aliação Ac ua ial do Regime de Pensões da Caixa Ge al de Aposen ações. B uce, D., & Schue ze, H. J. (2004). The labo ma ke consequences o expe ience in sel - employmen . Labou Economics, 11(5), 575–598. Coope , D. (2014). The e ec o unemploymen du a ion on u u e ea nings and o he ou comes. Fede al Rese e Bank o Bos on, Wo king Pa. Cunha, N. (2017). P inciples o Social Secu i y. Ebbinghaus, B. (2009). The Changing Public " P i a e Pension Mix in Eu ope: F om Pa h Dependence o Pa h ! Depa u e. 1–20. ECB. (2020). HICP In la ion o ecas s. Re ie ed om h ps://www.ecb.eu opa.eu/s a s/ecb_su eys/su ey_o _p o essional_ o ecas e s/h ml/ able _his _hicp.en.h ml Eu opean Comission. (2018). Pension adequacy cu en and u u e in he EU. h ps://doi.o g/10.2767/653851 Felds ein, M. (2001). The Fu u e o Social Secu i y Pension in Eu ope. NATIONAL BUREAU OF ECONOMIC RESEARCH, Wo king Pa(Sep embe ). Re ie ed om h p://www.nbe .o g/pape s/w8487 He ce, J. A., & B a o, J. (2017). On he In luence o Employmen -B eaks on Pension Bene i s Mon hly es ima es o esiden população View p ojec . Re ie ed om h ps://www. esea chga e.ne /publica ion/316919837 Holzmann, R., & P. (2006). Pension Re o m: Issues and P ospec o Non-Financial De ined Con ibu ion (NDC) Schemes. Washing on, D.C.: The Wo ld Bank. Holzmann, R. (2017). The ABCs o non inancial de ined con ibu ion (NDC) schemes. In e na ional Social Secu i y Re iew, 70(3), 53–77. h ps://doi.o g/10.1111/iss .12142 Ignacio Conde-Ruiz, J., & Gonzalez, C. I. (2014). F om Bisma ck o Be e idge: The O he Pension Re o m in Spain. SSRN Elec onic Jou nal. h ps://doi.o g/10.2139/ss n.2469659 ILO. (2019). Wo k o a b igh e u u e: Global Commission on he Fu u e o Wo k. Re ie ed om h ps://www.ilo.o g/wcmsp5/g oups/public/---dg epo s/--- 46 cabine /documen s/publica ion/wcms_662410.pd Isma ck, & O , O. (2008). Bisma ck e sus be e idge: A compa ison o social insu ance sys ems in Eu ope. CESi o DICE Repo , 6(4), 69–71. Kli zke, J. (2016). Be e idge and Bisma ck Remodelled : The Posi ions o B i ish and Ge man O ganised In e es s on Pension Re o m Julia Kli zke. Kolma , M. (2007). Be e idge e sus Bisma ck public-pension sys ems in in eg a ed ma ke s. Regional Science and U ban Economics, 37(6), 649–669. h ps://doi.o g/10.1016/j. egsciu beco.2007.01.003 Manuel, C., So ia, S., Sónia, V., Manuel, C., & Vaz-pa al a, S. S. (2006). P esen a ion o he Recen Re i emen Re o ms in Po ugal. Re ue F ançaise Des A ai es Sociales, 5(5), 253–269. Mendes, F. R. (1995). Po onde ai a segu ança social po uguesa? Análise Social, 30(131/132), 405– 429. h ps://doi.o g/10.2307/41011095 OECD. (2005). Pension-sys em Typology. OECD. (2019). OECD Re iews o Pension Sys ems: Po ugal. Re ie ed om h ps://doi.o g/10.1787/9789264313736-en OECD. (2020). Ne pension eplacemen a es (indica o ). h ps://doi.o g/10.1787/4b03 028- Pe ei a Da Sil a, C. M., Tomé Calado, J. P., & Medei os Ga cia, M. T. (2004). The inancial sus ainabili y o he Po uguese social secu i y sys em. Gene a Pape s on Risk and Insu ance: Issues and P ac ice, 29(3), 417–439. Rod igues, P. G., & Pe ei a, A. M. (2006). Pension Re o m in Po ugal: A Dynamic Gene al Equilib ium Analysis. (Decembe ), 1–500. Runde, D., & Milne , A. (2018). De ining he Fu u e o Wo k , Be o e I De ines Us The E ol ing In e na ional Labou O ganiza ion. (Oc obe ), 1–7. Schludi, M. (2005). The Re o m o Bisma ckian Pension Sys ems - A Compa ison o Pension Poli ics in Aus ia, F ance, Ge many, I aly and Sweden. Ams e dam Uni e si y P ess, Ams e dam 2005 All. Schmiede , J., on Wach e , T., & Bende , S. (2014). he Causal E ec o Unemploymen Du a ion on Wages: E idence om Unemploymen Insu ance Ex ensions. IZA Discussion Pape No. 8700. Schoukens, P., & Ba io, A. (2017). The changing concep o wo k. Eu opean Labou Law Jou nal, 8(4), 306–332. h ps://doi.o g/10.1177/2031952517743871 Sil a, P. D. (1997). Pe spec i as Financei as da Segu ança Social Em Po ugal. Documen os. SOA. (1994). T H E 1994 UNINSURED PENSIONER MORTALITY TABLE. UP-94 TASK FORCE, 47(1975), 859–900. ssa.go . (2018). Po ugal. Re ie ed om h ps://www.ssa.go /poli The Wo ld Bank. (1994). A e ing he Old Age C isis. Wo ld Bank - Ox o d Uni e si y P ess, pp. 321– 342. h ps://doi.o g/10.1007/978-3-030-24436-1_17 The Wo ld Bank. (2019). The Changing Na u e o Wo k. h ps://doi.o g/10.1596/978-1-4648-1328-3 Yin, R. K. (2009). Case S udy h Resea c Design and Me hods Fou h Edi ion. 47 48 8. ANNEXES Annexe 1 – Pension Models and Unemploymen Insu ance in Po ugal The Po uguese DB model The calcula ion o he pension bene i (P) in he Po uguese ea nings- ela ed public pension scheme based in he ollowing acc ued bene i o mula a e i emen age:  =     Whe e RE is he Re e ence Ea nings, A he acc ual Ra e and SF he Sus ainabili y Fac o . The e e ence ea nings a e calcula ed by a e aging mon hly ea nings o all yea s o co e age up o 40 yea s, adjus ed acco ding o he consume p ice index (CPI). The Acc ual a es a y om 2% o 2.3% depending on he indi idual ea nings ela i e o he alue o he IAS (Indexan e dos Apoios Sociais) as ollows: Re e ence Ea nings/IAS ≤ 1.1 >.; ≤. >.; ≤. >.; ≤. >8.0 Acc ual Ra e 2.3% 2.25% 2.2% 2.1% 2.0% Fo a gi en alue o RE, he co esponding pension o mula is as ollows: RE/IAS Pension Fo mula a No mal Re i emen Age ≤ 1.1  =  2.3%     >.; ≤.  = (1.1   2.3% + (−1.1)  2.25%)    >.; ≤.  = (1.1   2.3% +0.9 2.25% + (−2) 2.2%)    >.; ≤. = (1.1   2.3% + 0.9 2.25%+ 2.2%  2 + (−4) 2.1%)    >8.0  = (1.1   2.3% +0.9 2.25% + 2.2%  2 + 2.1%  4 + (−8) 2.0%)    No e: N = Numbe o con ibu ions yea s The sus ainabili y ac o is a demog aphic ac o designed o adjus pension bene i o longe i y changes. This ac o calcula ed on an annual basis by di iding o e all popula ion li e expec ancy a age 65 in 2000 and he one eco ded in he yea be o e he pension claim:   =℮ : ℮ : The con ibu ions o inance he old-age bene i based on he cu en con ibu ions schedule in Po ugal, i.e., o a s anda d wo ke is a o al combined con ibu ion (employe plus employee) o 34.75% which 20.21% inances old-age bene i (ssa.go ). Fo a non-s anda d wo ke (assumed o be sel -employed) is 29.6%, and we es ima ed o be 17.21% o sponso old-age bene i based on he p opo ion o he con ibu ion o a s anda d wo ke . 49 The NDC model (inspi ed by he Sweden Model) The NDC model analysed on his s udy is based on he cu en Sweden Model, whe e he p ima y pension en i lemen came om a No ional De ined Con ibu ion supplemen ed by a small po ion o a Funded De ined Con ibu ion (FDC). The old-age pension en i lemen om he NDC model based on he ollowing o mula:   =     The No ional Accumula ed Balance is acc ued based on he ollowing o mula:    =   (1+  )    Whe e,   , indi idual e e ence ea nings in yea i, and c is he con ibu ion a e. The no ional capi al is inc eased each yea by he no ional in e es a e, assumed o be 2% p.a.. The no ional ac ua ial annui y ac o AN a age e i emen age es ima ed based o he mo ali y able as de ined in sec ion 3, and a hypo he ical discoun a e o 3.00% p.a. The old-age bene i is supplemen ed by a de ined con ibu ion plan, wi h lowe con ibu ion a es, and whe e he pension en i lemen is calcula ed by di iding he accumula ed con ibu ion a e i emen age by he expec ed u u e li e ime a e i emen age,   , o es ima e a pension bene i a e i emen age:   =    The accumula ed con ibu ions a e i emen age is as ollows:   =   (1+  )    Whe e,   , he capi aliza ion o he con ibu ions based on a e u n a e o 3.00 p.a. The o al pension en i lemen o he NDC model is as he sum o he wo pensions desc i ed abo e, i.e. =   +  The con ibu ions o inance he old-age bene i o he NDC model: Con ibu ions Ra e ( c ) S anda d wo ke Non-S anda d Wo ke NDC Plan 17.71% 14.71% DC Plan 2.50% 2.50% To al con ibu ions unde NDC model 20.21% 17.21%