Some Uncomfortable Arithmetic Regarding Europe s Public Finances

Transcription

1 WP/1/177 Some Uncomforable Arihmeic Regarding Europe s Public Finances Delia Velculescu

2 21 Inernaional Moneary Fund WP/1/177 IMF Working Paper European Deparmen Some Uncomforable Arihmeic Regarding Europe s Public Finances Prepared by Delia Velculescu 1 Auhorized for disribuion by James Morsink July 21 Absrac This Working Paper should no be repored as represening he views of he IMF. The views expressed in his Working Paper are hose of he auhor(s) and do no necessarily represen hose of he IMF or IMF policy. Working Papers describe research in progress by he auhor(s) and are published o elici commens and o furher debae. Tradiional fiscal indicaors focused on measures of curren deficis and deb miss he poenially imporan implicaions of curren policies for fuure public finances. This could be problemaic, including in he case of Europe, where populaion aging is expeced o pose addiional fiscal coss no capured by such indicaors. To beer gauge he sae of public finances in he counries, his paper derives forward-looking fiscal measures of ineremporal ne worh boh direcly from he European Commission s Aging Working Group s long-run indicaors and using a comprehensive public-secor balance shee approach. These measures could be used as an early warning mechanism and also as a communicaion device wih he public. Curren esimaes indicae ha, on exising policies, he ineremporal ne worh of he is deeply negaive, even in excess of is GDP level, and is projeced o worsen furher over ime. This suggess ha Europe s curren policies need o be significanly srenghened o bring fuure liabiliies in line wih he EU governmens capaciy o generae asses. JEL Classificaion Numbers: E62, H68, H87, J1 Keywords: Fiscal Policy, Fiscal Susainabiliy, Aging, Public Secor Balance Shee Auhor s Address: dvelculescu@imf.org 1 The auhor wishes o hank especially Bob Traa for developing he comprehensive public secor balance shee mehodology used in his paper in order o make he challenge of populaion aging more inuiive o he broader public and for his valuable suppor and suggesions, James Morsink, Mark Allen, Alber Jaeger, and all paricipans in he IMF s European Deparmen Seminar for useful commens, and David Velasquez for ousanding research assisance.

3 2 CONTENTS PAGE I. Inroducion... 3 II. Wha Do Exising Fiscal Indicaors Tell Us?... 6 III. Measuring Ineremporal Ne Worh... 1 A. The Parameric Approach B. The Comprehensive Balance-Shee Approach IV. Sensiiviy Analysis... 2 V. Policy Implicaions VI. Conclusions References...25 Mahemaical Appendix: Comprehensive Ne Worh and he S1 and S2 Susainabiliy Indicaors Tables 1. : Tradiional Balance Shee, : Public Secor Comprehensive Balance Shee, : Sensiiviy Analysis, Ineremporal Ne Worh Ineremporal Ne Worh (Balance Shee Approach), and Required Upfron Adjusmen o Bring Ne Worh Ino Balance or he Defici o 3 Percen of GDP by Figures 1. Tradiional Fiscal Indicaors, : Populaion Projecions European Commission Long-Term Indicaors Tradiional and Comprehensive Fiscal Indicaors, Ineremporal Ne Worh Derived from S1 and S Updaed Ineremporal Ne Worh Derived from S1 and S : Long-Run Macroeconomic Projecions : Long-Run Fiscal Projecions Ineremporal Ne Worh Derived from he Balance Shee Approach Long-Term Gains from a One Percen Upfron Permanen Fiscal Adjusmen Box 1. Esimaing Ineremporal Ne Worh From he EC s S1 and S2 Indicaors A Shorcu... 13

4 3 I. INTRODUCTION The 29 1 crisis in Europe has unmasked he vulnerabiliy of is public finances. Following he 29 global downurn, which led o deerioraing public finances and weak growh in many European counries, in April-May 21, markes began o quesion he abiliy of some euro-zone members o make good on heir deb paymens o crediors. This resuled in emporary panic, wih spiking spreads on governmen bonds. Wha his episode refleced was a manifesaion of he unyielding force of wha economiss call he ineremporal budge consrain, which requires governmens o be able o generae enough fiscal surpluses in he fuure (in excess of heir ineres paymen needs) o mee heir curren (ne) deb obligaions. Ofen hough of as a mere heoreical concep ha only applies in he long run, his consrain can become binding much sooner, if fiscal imbalances, coupled wih a weak oulook, become so large ha hey are no longer hough of as susainable, hus elescoping problems seemingly operaing on a long-erm horizon back ino he presen. In he recen European case, his led o an EU-IMF suppor package for one of he eurozone members and a large EU-ECB-IMF sabilizaion package for he euro zone. Tradiional fiscal indicaors do no always correcly capure he rue sae of public finances. Such indicaors including differen measures of deficis (overall, primary, srucural), gross deb, and even public secor financial ne worh are generally used o quanify he effecs of pas and curren policies and evens on counries curren fiscal posiions. Bu, being backwardlooking by definiion, hese defici and deb indicaors miss he implicaions of curren policies for fuure public finances. Such implicaions can be significan, including when populaion aging is expeced o pose addiional large fiscal coss, as is he case in many counries, including in Europe. Consequenly, radiional indicaors offer only a parial picure of public finances and do no ake ino accoun informaion ha can be imporan and is already available. 2 Moreover, by lacking a forward-looking componen, hey are also no direcly linked o he above-menioned ineremporal budge consrain, which, when binding, can have large economic repercussions. Despie heir shorcomings, radiional indicaors sill largely consiue he basis for policymaking decisions under exising naional and inernaional frameworks. For example, many counries have naional policies in place which arge he level of he defici, srucural defici, expendiures, or deb. Moreover, Europe s Sabiliy and Growh Pac (SGP) requires ha members overall deficis be mainained below 3 percen of GDP and gross deb be kep below 6 percen of GDP. 3 These limis were agreed a he inroducion of he euro based on boh poliical and economic heory consideraions and for simple operaional reasons, given ha hey can be relaively easily measured and explained o he general public. The emphasis on radiional indicaors can lead o subopimal or counerproducive policies, especially a imes of crisis. Under he exising framework, counries whose deficis have 2 Populaion aging, unlike oher macroeconomic shocks (such as naural disasers, wars, financial crisis, ec.) is somewha easier o predic, o he exen ha known pas and curren feriliy and moraliy raes can be used o projec he fuure age-composiion of populaions, and given ha such raes change only slowly over ime. However, immigraion rends may be more difficul o projec, as hey are influenced by oher macroeconomic shocks. 3 The so-called Medium Term Objecives (MTO) under he SGP arge srucural balances over he medium erm. These indicaors are inended o be more forward-looking and ailored o individual counry circumsances, hough hey only capure a small par of long-run aging coss.

5 4 deerioraed significanly are required (under he Excessive Defici Procedure, or EDP) o underake a relaively larger fiscal effor o reurn o he prescribed Maasrich limi as compared o peers ha have lower deficis which are seen as having more fiscal space. While hese inferences appear correc, hey do no fully ake accoun of he relaive long-run fiscal posiions of counries, which can make such inferences incomplee or even misleading. Moreover, i has been documened ha, in downurns, counries resor more frequenly o shor-run emporary soluions o lower heir deficis, which may no improve and could even worsen heir long-run fiscal posiions. 4 One example is he recen rend seen in some Emerging European counries ha are now parially back-racking on heir pension reforms by lowering or suspending he fiscal conribuions o heir privae pension sysems in order o reduce heir shor-run deficis and deb. 5 These acions are he resul of a percepion of being effecively punished in an inernaional conex for having underaken pension reforms, because such reforms emporarily increase deficis and deb in principle working agains he SGP crieria ha apply uniformly o boh reformers and non-reformers. These ensions now surfacing wihin naional policies in a number of counries could have poenially negaive long-run consequences for hese economies. Hence, an increasing awareness is slowly emerging regarding he need o inegrae more fully he essenial long-run consideraions ino policy-making, including hrough new analyical ools. The European Commission s Aging-Working Group has developed and published wo susainabiliy indicaors ha ake ino accoun boh curren fiscal posiions and aging coss. 6 The academic lieraure has also produced similar analyses, including, more recenly, Gokhale s (29) measures of fiscal imbalance for he EU and he US, Auerbach s (27) fiscal gap for he US, and Cecchei e al. s (21) deb projecions for indusrial counries. 7 Sill, hese measures have remained indicaive, are relaively lile known by he general public, and do no carry he same poliical weigh as radiional indicaors, perhaps because of heir complexiy and he difficuly of explaining hem o voers. Ineresingly, in ligh of he 21 fiscal crisis in Europe, forward-looking analyses are now increasingly geing racion wih marke paricipans (Ghezzi and Keller, 21). The presen paper proposes alernaive forward-looking indicaors ha measure he public secor s ne worh for EU counries in an ineremporal perspecive. These can be calculaed direcly from he EC s forward indicaors, or can be derived using he public-secor comprehensive balance-shee framework developed by Traa (26). 8 Measures of he public secor s ineremporal ne worh are one logical nex sep beyond wha governmens are already doing, building on he financial ne worh concep bu adding a forward-looking elemen. The idea is ha hey can shed addiional ligh on he sae of Europe s public finances and can be used in he public debae o communicae policy needs o he general public. 4 See Milesi-Ferrei (23) and von Hagen, and Wolff (26) on heoreical and empirical analyses of emporary soluions, including creaive accouning, for he European Union. 5 Esonia, Lihuania, Lavia, Slovakia, and have already underaken measures o his effec. Croaia and have been conemplaing similar measures (see hp:// 6 See he European Commission s 29 Susainabiliy Repor. 7 For a discussion of general consideraions see Auerbach (28). An early lieraure on his opic developed in he 199s. See, for example, Chand and Jaeger (1996), Disney (21), and references herein. 8 See, for example, Velculescu (21), Traa (29), and Traa (26) for more recen applicaions of he comprehensive balance shee approach o,, and.

6 5 Wha is he public-secor s ineremporal ne worh? I can be hough of as he mirror image of he EC s Aging-Working Group susainabiliy indicaors. Raher han quanifying he upfron primary balance effor required o saisfy a given ineremporal budge consrain (which is a flow measure), i reflecs he oal curren and projeced fuure ne liabiliies of he public secor under unchanged policies (which is a sock measure). 9 I hus complemens he radiional curren ne worh posiion of a governmen, which is backward looking, wih a forward-looking componen ha akes ino accoun he effecs of curren fiscal policies on fuure asses and liabiliies, o see if he ineremporal budge consrain is saisfied or no. 1 In his sense, i funcions as a hermomeer measuring he healh (i.e. susainabiliy) of public finances a each poin in ime. Good policies will lower he emperaure by srenghening he public secor s balance shee, while bad policies would deepen he shorfall, raising he emperaure. I can provide early warning signs of fiscal unsusainabiliy. For example, in he conex of populaion aging, radiional indicaors ofen fail o signal a problem. However, he public secor s ineremporal ne worh could be negaive if aging causes large increases in fuure public expendiures (included in he balance shee in NPV erms), if curren policies are no calibraed o generae enough resources o mee hese obligaions projeced for he fuure in addiion o he already exising ones. In his case, acions would need o be aken o srenghen policies and bring oal asses a leas in line wih oal ineremporal liabiliies, or marke forces will evenually emerge o reduce liabiliies in line wih exising and prospecive asses. Since fiscal adjusmen is ofen difficul and requires ime for consulaion wih he public, here is grea value in having informaion on prospecive fiscal inconsisency well in advance ha is why he ineremporal ne worh can ac as an early warning sysem. I can also help o communicae policy needs in an inuiive way ha could be somewha easier o grasp by he general public. The lieraure offers a variey of sophisicaed frameworks, such as overlapping-generaion models and compuable dynamic general-equilibrium models o analyze he long-run effecs of fiscal policies. However, hese end o have only a limied impac beyond academic circles, perhaps because of heir complexiy. Even he EC s Aging-Working Group indicaors and similar primary fiscal gap measures developed by ohers reach a relaively limied audience parly due o he difficuly of relaying heir message o he median voer. The ineremporal ne worh indicaors developed here complemen he already exising measures. They may also be somewha easier o use as a communicaion device wih he public, o he exen ha people may have a beer inuiive undersanding of wealh and conceps of ne worh as he difference of asses and liabiliies, expressed in a given currency or as a percenage of income. 9 The projeced fuure ne liabiliies under unchanged policies include pensions no only of individuals already in he sysem, bu also of all fuure generaions ha have no ye been born. Hence, hey do no represen only conraced obligaions, bu also prospecive pension deb. Moreover, he projeced fuure pension paymens (and corresponding conribuions o he sysem) are assumed o be governed by he curren pension legislaion. I is in his sense ha hey capure he effecs of curren policies. 1 The ineremporal ne worh measures derived here represen, herefore, only counerfacual quanificaions of he gap ha would arise if curren policies were no changed, relaive o he ineremporal budge consrain. Clearly, if a gap exiss, hen policies will evenually have o change o ensure ha he budge consrain is saisfied. Indeed, in general equilibrium models, such policy changes are required o ensure ha he model is inernally consisen. This paper does no aemp o model such policy changes, bu raher o idenify if hey will be needed on no.

7 6 The paper is organized as follows. Secion II describes in more deail he exising fiscal indicaors and heir limiaions. Secion III oulines how he measures of ineremporal ne worh are consruced and presens some resuls for he. Secion IV draws some policy implicaions in a cross counry conex, and secion V concludes wih a summary of he findings. II. WHAT DO EXISTING FISCAL INDICATORS TELL US? Tradiional indicaors poin o a deerioraion in Europe s curren fiscal posiion. The 28 9 global downurn led o a fall in fiscal revenues in all European Union () counries, reflecing he auomaically sabilizing response of he fiscal accouns o he cycle. Moreover, depending on heir fiscal and macroeconomic saring posiion, differen governmens responded o he crisis in a variey of ways: some rying looser fiscal policies o suppor growh, ohers ighening fiscal policy o limi damage o susainabiliy. Consequenly, he headline defici increased from around 2 o close o 7 percen of GDP from 28 o 29, wih some counries deficis rising above 1 percen of GDP. During he same period, he union s deb-o-gdp raio rose by abou 12 percenage poins o abou 74 percen of GDP. While a few counries ighened policies, or a leas pu in place srucures o recover from he fiscal deerioraion in he recession, he EU as a whole loosened is fiscal sance subsanially, as refleced in a worsening of is permanen srucural balance by close o 2 percen of GDP (Figure 1). Bu he specer of populaion aging is expeced o weigh heavily on fuure public finances. Looking beyond he curren difficul imes, all EU counries expec a significan aging of heir populaions. Working populaions are falling in mos of he EU counries, and he union s old-age dependency raio is expeced o more han double from 25 o over 5 percen beween 28 and 26, wih some counries raios geing close o 7 percen (Figure 2). These projecions essenially imply ha in 26, for each pensioner here will be, on average, only wo workers, or less in some cases, raher han four, as is currenly he case. This is expeced o pu progressive pressure on public finances, especially for hose counries whose demographics are deerioraing relaively more, bu which have made lile progress on pension reforms so far and hus sill rely on paying pensions from axes on he exising (bu diminishing) workforce. Forward-looking indicaors sugges ha fiscal susainabiliy is even more seriously imperiled han curren fiscal posiions sugges. While radiional indicaors only focus on counries curren fiscal posiions and backward-looking deb, more comprehensive indicaors of fiscal susainabiliy can also capure he effecs of populaion aging on fuure fiscal oucomes, hus poenially providing a more accurae descripion of he sae of public finances and heir susainabiliy a any given poin in ime. The indicaors developed by he EC s Aging-Working Group are one example of such measures. The S1 indicaor aims o quanify he required upfron fiscal adjusmen needed o reach he Maasrich deb-o-gdp raio of 6 percen in 26. The S2 indicaor esimaes he adjusmen required o saisfy he governmen s ineremporal budge consrain over an infinie horizon. According o hese measures, he EU s fiscal posiion is unsusainable, requiring a large upfron permanen fiscal adjusmen of some 6 percen of GDP o bring public deb o he Maasrich limi or ino ineremporal balance (Figure 3).

8 Figure 1. Tradiional Fiscal Indicaors, 29 (In percen of GDP) Deb levels have deerioraed, and financial ne worh is negaive for mos counries. Gross Deb Financial Ne Worh Ialy Ialy Porugal Porugal Mala Mala Ausria Ausria Neherlands Neherlands Lavia Lavia Maasrich deb limi Lihuania Lihuania Esonia Esonia Mos counries experienced a deerioraion in headline deficis. Overall Fiscal Balance Change in he Fiscal Balance, 28-9 Porugal Porugal Lavia Lihuania Neherlands Lihuania Lavia Neherlands Ialy Mala Ausria Ausria Esonia Maasrich Ialy defici limi Mala Esonia Wih some having more room han ohers o couner he cycle. Srucural Balance Change in he Srucural Balance 29-8 Ausria Lihuania Ausria Lihuania Lavia Ialy Neherlands Porugal Mala Ialy Porugal Neherlands Mala Lavia Esonia Esonia Sources: EUROSTAT; and IMF, World Economic Oulook. 7

9 8 Figure 2. : Populaion Projecions Populaion Growh, 28-6 (In percen) Ausria Porugal Neherlands Ialy Mala Esonia Lihuania Lavia Working Age Populaion Growh, 28-5 (In percen) Mala Neherlands Ausria Porugal Lihuania Esonia Ialy Lavia Populaion 65 Years or Older (In percen of oal) Lihuania Lavia Ialy Mala Porugal Esonia Ausria Neherlands Old-Age Dependency Raio (In percen of working age populaion) Lihuania Lavia Ialy Mala Esonia Porugal Ausria Neherlands Source: Eurosa.

10 Figure 3. European Commission Long-Term Indicaors (In percen of GDP) 9 S1: Required upfron adjusmen in he primary balance o reach a deb level of 6% of GDP in 26 Lavia Lihuania Neherlands Porugal Mala Ausria Ialy Esonia S2: Required upfron adjusmen in he primary balance o fulfill he infinie-horizon ineremporal budge consrain Lavia Lihuania Mala Neherlands Porugal Ausria Ialy Esonia Susainabiliy gaps (in percen of GDP) Srucural primary Change in balance age-relaed S1 S expendiure Toal S1 IBP* DR* LTC* Toal S2 IBP* LTC* BE BG CZ DK DE EE IE GR ES FR IT CY LV LT LU HU MT NL AT PL PT RO SI SK FI SE * IBP = required adjusmen given he iniial budgeary posiion, DR = adjusmen o reach he deb requiremen (6% of GDP) in 26, LTC = required adjusmen given he longerm change in he primary balance due o demographic ageing. Source: European Commission, Susainabiliy Repor 29.

11 1 For individual EU counries, radiional and comprehensive indicaors can show differen picures (Figure 4). There is no necessarily a direc correspondence beween he radiional deb and defici measures and he more comprehensive indicaors. Indeed, according o he comprehensive measures, some counries wih relaively sound saring posiions bu large aging coss rank lower han counries wih high curren deficis, bu ha have lower fuure demographic coss, eiher due o less pronounced aging, or o comprehensive pension reforms. For example,, wih one of he smalles deficis in 29, ranks lower on comprehensive indicaors han and Lihuania, whose deficis opped 7 percen of GDP las year, bu which reformed heir pension sysems significanly and early on. and Ialy, wih large debs, score much beer on comprehensive indicaors, again due o early pension reforms. This someimes surprising reversal of order demonsraes ha fuure aging implicaions can overwhelm he cos of he 28 9 crisis. EC's S1 measure of fiscal gap-o-gdp raio Figure 4. Tradiional and Comprehensive Fiscal Indicaors, 29 (In percen of GDP) Correlaion: -.78 Correlaion: Lavia Lavia 9 Neherlands 6 Porugal Neherlands 6 Porugal Lihuania Mala Lihuania Mala Ausria Ausria 3 3 Ialy Esonia Esonia Ialy Defici-o-GDP raio Deb-o-GDP raio Sources: European Commission, 29 Susainabiliy Repor; Eurosa; and IMF saff calculaions. EC's S1 measure of fiscal gap-o-gdp raio III. MEASURING INTERTEMPORAL NET WORTH Ineremporal ne worh is calculaed as he oal of curren ne worh and he discouned sum of fuure primary balances under curren policies. While curren ne worh can be easily derived using a radiional balance shee, he forward-looking componen requires projecions ha can span a finie or even an infinie-horizon. While infinie-horizon measures are heoreically more appealing, hey can be harder o grasp and may prove weak from a policy poin of view, as infiniehorizon budge consrains could be hough o be saisfied by very high levels of shor-erm deb and deficis, as long as here is reason o believe ha sufficienly large primary surpluses will be achieved aferwards. Hence, finie-horizon indicaors may prove more pracical, if hey can allow he long-run impac of populaion ageing o be analyzed in a meaningful way ha is sill wihin he sighs of curren axpayers and policy makers. This secion presens boh infinie and finie horizon measures of ineremporal ne worh, which are calculaed eiher direcly from he EC s indicaors he parameric approach or using IMF saff s assumpions comprehensive balance-shee approach as will be shown nex.

12 Lavia Lihuania Mala Porugal Neherlands Ausria Ialy Esonia Finie Horizon Ineremporal Ne Worh (Derived from S1) Source: IMF Saff esimaes. 11 A. The Parameric Approach Ineremporal ne worh can be derived direcly from he EC s Aging-Working Group indicaors of fiscal susainabiliy (Mahemaical Appendix and Box 1). As described earlier, he S2 indicaor shows he permanen upfron adjusmen required o saisfy he infinie-horizon ineremporal budge consrain, i.e. where curren ne deb and he discouned sum of primary surpluses cancel each oher ou over an infinie horizon. Given he above-menioned definiion of ineremporal ne worh, i can be seen ha here is a direc correspondence beween i and S2. Similarly, a corresponding finie-horizon measure of ineremporal ne worh can be derived direcly from S1. While S1 relaes o he adjusmen required o reach deb o 6 percen of GDP in 26, is finie-horizon ineremporal-ne-worh corresponden reflecs he oal of curren ne worh and he sum of primary balances, discouned for a se period, bu wihou requiring deb o reach he 6 percen-of-gdp limi. Ineremporal ne worh measures can, herefore, be hough of as approximae sock correspondens of he S1 and S2 flow measures. Moreover, hese indicaors could have he advanage of relaying a poenially powerful message o he median voer: as menioned earlier, while individuals may have rouble undersanding wha he S1 and S2 fiscal gap measures mean, i may be easier o bring home he noion ha a large negaive ineremporal ne worh value, expressed in billions of Euros or percen of GDP, would be unsusainable and would need o be addressed urgenly. Measures of ineremporal ne worh consruced based on he S1 and S2 indicaors pain a bleak picure of he susainabiliy of Europe s public finances (Figure 5). The ineremporal ne worh for he as a whole which represens a weighed average of all counries on curren policies, sands a around negaive 175 percen o of GDP (equivalen o 2 rillion) when measured over a 5-year horizon, and a 38 percen of GDP (or 45 rillion) over an infinie horizon, respecively. This is several imes larger han convenional measures of he union s deb, which was esimaed a 7 percen of GDP on a gross basis, and a 5 percen of GDP when financial asses are need ou. Figure 5. Ineremporal Ne Worh Derived from S1 and S2 (In percen of GDP) Lavia Lihuania Mala Neherlands Porugal Ausria Ialy Esonia Infinie Horizon Ineremporal Ne Worh (Derived from S2)

13 12 The average measures mask imporan differences among counries (Figure 5). These arise due o large differences in saring fiscal posiions and fuure aging burdens, as esimaed by he EC in early 29 (see he able in Figure 3 for individual-counry esimaes of iniial srucural balances and aging coss). On he one hand, only a few counries appear o be currenly ineremporally susainable: and under boh measures, and and only under he finie-horizon definiion. These counries combine boh modes aging coss, due o early pension reforms, and a srong saring fiscal posiion. On he oher hand, counries such as,,, and have large negaive ne worh posiions mainly due o high aging coss, while, he and Lavia are negaively affeced by relaively worse iniial fiscal posiions. Lihuania Porugal Lavia Neherlands EU 27 Mala Ialy Ausria Esonia Figure 6. Updaed Ineremporal Ne Worh Derived from S1 and S2 (In percen of GDP) Finie Horizon Ineremporal Ne Worh updaed original Source: IMF Saff esimaes Lihuania Porugal Neherlands EU 27 Mala Lavia Ausria Ialy Esonia Infin ie Ho rizon Ineremporal Ne Worh These resuls broadly hold up when daa are updaed o reflec more recen informaion, hough some counries ranking changes (Figure 6). The calculaions of ineremporal ne worh derived from he EC s S1 and S2 indicaors implicily ake on board all underlying assumpions as of he ime hese indicaors were derived. Of paricular imporance are he esimaed 29 primary srucural balances, which correspond o he value of he primary balances under curren policies afer removing he effec of he cycle and of one-off measures. Large uncerainies usually surround hese esimaes, relaed o difficulies in esimaing he oupu gap and budge elasiciies. Moreover, as hey were derived in early 29, hey may no fully capure more recen evens. The parameric approach allows for updaing he calculaion of ineremporal ne worh wih new iniial values for deb, public secor financial asses, and he primary srucural balance (see Mahemaical Appendix for he algebraic expressions). Including more recen developmens during he 29 global crisis resuls in a modes increase in he iniial srucural balance and deb for he average, which implies ha is ineremporal ne worh worsens slighly o 192/398 percen of GDP under he finie/infinie horizon measures bu he measures for individual counries, and hence heir relaive ranking, can change significanly. For example, including he significan revisions o he 29 defici and deb daa ha became available in early 21, which refleced more fully he effecs of updaed original

14 13 acual raher han projeced fiscal policies in 29, finie ineremporal ne worh worsens by more han 1 percen of GDP in, he Slovak Republic, and, while i improves by abou he same amoun in Esonia and Lavia. Box 1. Esimaing Ineremporal Ne Worh from he EC s S1 and S2 Indicaors A Shorcu I can be shown ha a direc correspondence exiss beween he EC s S1 and S2 indicaors and measures of ineremporal ne worh (see Mahemaical Appendix): inw S2 a 26 * inw S a where inw and inw* represen ineremporal ne worh in percen of GDP under an infinie and finie horizon, respecively, a represens he iniial asses of he public secor in percen of GDP, and δ is he ime-varying discoun rae ha depends on real growh and he real ineres rae. Pracically, o calculae inw from S1 and S2, he variable yearly-profile of he EC s projeced real growh is needed o compue he variable discoun rae (he real ineres rae is assumed consan a 3 percen). 1 However, for mos counries, measures of inw can also be approximaed direcly from he published figures for S1, S2, and heir subcomponens, and he EC s esimaes of he 21 deb (d 21 ) and he 29 primary srucural balance (pb 29 ), wihou he need for growh projecions. 2 To do so, recall ha: S1=IBP S1 +DR+LTC and S2= IBP S2 + LTC,where IBP is he required adjusmen given he iniial budgeary posiion (i.e. he gap beween he curren srucural primary balance and he long-erm deb sabilizing balance), LTC is he required adjusmen given he long-erm change in he fiscal posiion due o aging, and DR is he adjusmen needed o reach a deb of 6 percen of GDP by 26. Given hese definiions, i can be shown ha, for he counries menioned above: d21 inw S2 a21 IBPS 2 pb29 * d21 inw ( d21 6 DR) S16DRa21 DR d ( IBP pb )( d 6) 21 S These approximaions can be helpful in quickly ransforming he S1 and S2 indicaors ino measures of ineremporal ne worh han may be conveyed o he median voer more easily. These expressions can also be used o updae he ineremporal ne worh measures wih more recen daa on deb and he primary defici. Also, he curren value of public asses is required, hough his is readily available from Eurosa. 2 These approximaions do no hold for, Lihuania,, he Neherlands, and, because of large projeced long-erm changes in propery income, which affec he IBP and he S1 and S2 indicaors. In all oher counries, hese changes are small and, for simpliciy, can be ignored.

15 14 B. The Comprehensive Balance-Shee Approach Measures of ineremporal ne worh for each EU counry can also be derived using he comprehensive balance-shee approach. This approach has he advanage of allowing for using Fund saff projecions for he medium-run pah of macroeconomic variables. In paricular, i allows for an explici medium-erm fiscal pah ha can ake ino accoun adoped/planned fuure fiscal measures, raher han relying exclusively on an esimaed measure of he curren srucural fiscal balance, which can be difficul o esimae. I is also a more ransparen framework, as i builds on he radiional balance shee, which again can be easier o undersand by he public. The comprehensive balance shee can be derived sraighforwardly, in hree seps. In a firs sep, a radiional (backward-looking) accouning balance needs o be compued for each counry. The difference beween a counry s public secor financial asses and liabiliies consiues is curren financial ne worh. A more comprehensive measure of curren ne worh would also include he public secor s ne capial sock he sum of all buildings, highways, infrasrucure, and land acquired by he sae over he years. However, in he absence of daa on his sock for all counries, and for comparabiliy purposes, only financial asses and liabiliies are considered here, wih financial ne worh hus likely o underesimae oal ne worh. 11 Taking he as a whole as an illusraive example, is radiional balance shee (available direcly in Eurosa) reveals ha is financial ne worh has recenly been deerioraing, reaching negaive 47 percen of GDP in 29, or abou 5.5 rillion (Table 1). 12 Table 1. : Tradiional Balance Shee, 25-9 (In percen of GDP) Financial asses Currency and deposis Securiies oher han shares Loans Shares and oher equiy Oher financial asses Liabiliies Currency and deposis Securiies oher han shares Loans Oher liabiliies Financial ne worh Sources: Eurosa; and IMFsaff esimaes. 11 For some counries, esimaes of governmen ne capial socks for have been calculaed by Kamps (24). These could be added o financial asses o calculae he comprehensive ne worh of hese counries. This has no been done here, o mainain consisency across all counries. 12 Of course, i is already sriking in a radiional balance shee sense ha he has such a large negaive ne worh even before couning he prospecive coss of aging.

16 15 In a second sep, a counry-specific long-erm fiscal oulook is developed based on a se of long-run macroeconomic projecions. Here, medium-erm GDP, labor marke, and fiscal projecions generally correspond o he Spring 21 IMF WEO oulook for For example, for he, growh is assumed o gradually recover and reach 2 percen by 215, while he overall fiscal defici is projeced o fall o abou 2 percen of GDP. Long-run GDP and labor marke projecions converge o he EC s assumpions, which ake ino accoun counry-specific demographic projecions, as presened in he 29 Susainabiliy Repor (Figure 7). For he union as a whole, average GDP growh is projeced o decline o 1.5 percen by 26 and sabilize a ha level hereafer. The GDP deflaor growh is assumed o converge o he ECB s arge of 2 percen, and he real ineres rae is projeced o be 1 basis poins above real growh in he long run. 14 All non-age, non-ineres relaed fiscal revenues and expendiures, as well as financial asses and oher financial liabiliies excep for public deb are assumed o remain consan in percen of GDP from 215 onward. Ne aging-relaed incremenal expendiures are also aken from he Susainabiliy Repor, which are esimaed a 4.4 percen of GDP for he for 21 6 (Figure 8). 15 In a hird and final sep, he comprehensive balance shee is consruced combining seps one and wo above. The sream of projeced primary balances under curren policies is discouned using he nominal ineres rae on deb. As before, wo measures are consruced: one ha discouns primary balances over a 5-year horizon and a second one ha assumes an infinie horizon. The ne presen values are hen added o he curren ne worh, obaining boh an infinie and a finie-horizon measure of a counry s ineremporal ne worh. This exercise can be repeaed for successive years during he medium-erm, including as new informaion comes in, o gauge wheher he ineremporal ne worh is improving or worsening over ime on curren policies. For he, ineremporal ne worh is negaive, large, and deerioraing over he medium erm (Table 2). For he, he 21 ineremporal ne worh is negaive 15 or 47 percen of GDP (equivalen o close o 2 or 6 rillion), depending on he horizon of he measure, or hree/en imes larger han curren financial ne worh. These resuls are broadly in line wih he picure obained from he EC s indicaors discussed earlier, alhough some of he assumpions are differen, and imporanly, he medium-erm macroeconomic and fiscal profile has been updaed o ake ino accoun more recen informaion. Moreover, he ineremporal ne worh of he is slowly deerioraing over ime. This implies ha curren policies are moving in he wrong direcion a he area aggregae level, and ha more ambiious policies are required o change he dynamics of he s ineremporal ne worh o eliminae he exising susainabiliy gap. 13 Fiscal daa for he Slovak Republic have been updaed o reflec he 29 Eurosa revision. For, fiscal daa was revised o ake ino accoun addiional announced measures. For, fiscal daa differs from he WEO in ha i does no assume addiional measures beyond he wihdrawal of he fiscal simulus (passive scenario). For Lavia and, WEO daa assume addiional (unidenified) fiscal measures o comply wih he curren IMF program in place. 14 Specifically, real GDP growh afer 215 is assumed o gradually converge o he EC s esimaed value by 225 and o follow he pah esimaed by he Commission hereafer. The deflaor and he real ineres rae are equally assumed o gradually converge by 225 o heir respecive long-run values. A drawback of his mehodology is ha i does no consider he general equilibrium effecs of assumed policies on macroeconomic variables, such as poenially higher ineres raes due o larger deficis and deb, as his paper only underakes a counerfacual hough experimen raher han a general equilibrium simulaion. Here, he ineres rae, which is variable, funcions primarily as a discoun rae. 15 The EC s esimaes of aging coss represen some of he mos comprehensive and comparable esimaes ha he auhor is aware of. However, hey also have drawbacks (see Franco e. al., 25, Langenus, 26), including wih respec o he baseline esimaes of healh care coss, which may be oo opimisic (see IMF, 21).

17 16 Figure 7. : Long-Run Macroeconomic Projecions 12 1 Populaion (in percen of oal) Young populaion (-14) as % of oal Working age populaion (15-64) Elderly populaion (65 and over) Employmen Growh (In percen) Labor Produciviy Per Hour Growh (In percen) 3. Real GDP Growh (In percen) Sources: IMF Spring 21 WEO and European Commission.

18 Fiscal Revenue and Expendiure Figure 8. : Long-Run Fiscal Projecions (In percen of GDP, unless oherwise specified) Nominal and Real Ineres Rae (In percen) Revenue Primary expendiure Ineres expense (righ scale) 2 1 Nominal ineres rae Real ineres rae Fiscal Balance 3 Gross Deb Primary balance Overall balance Sources: Spring 21 IMF WEO and IMF Saff calculaions.

19 18 Table 2. : Public Secor Comprehensive Balance Shee, (In billions of euros) Financial asses Currency and deposis Securiies oher han shares Loans Shares and oher equiy Oher financial asses Liabiliies Currency and deposis Securiies oher han shares Loans Oher liabiliies Financial ne worh NPV of fuure primary balances (finie) NPV of fuure primary balances (infinie) Ineremporal Financial Ne Worh (finie) Ineremporal Financial Ne Worh (infinie) (In percen of GDP) Financial asses Currency and deposis Securiies oher han shares Loans Shares and oher equiy Oher financial asses Liabiliies Currency and deposis Securiies oher han shares Loans Oher liabiliies Financial ne worh NPV of fuure primary balances (finie) NPV of fuure primary balances (infinie) Ineremporal Financial Ne Worh (finie) Ineremporal Financial Ne Worh (infinie) Sources: Eurosa; and IMFsaff esimaes. The dispariy among counries remains large (Figure 9). On he one hand, he new daa capures significan fiscal adjusmens planned in he, Lavia, and over he medium erm, which are expeced o dramaically improve heir ne worh posiion compared o he iniial implici esimaes based on he EC s numbers. On he oher hand, a deerioraion in he iniial fiscal posiion and insufficien consolidaion over he medium erm in counries such as, and, combined wih heir very high aging coss, resuls in a significan worsening of heir ineremporal ne worh much beyond he level underlying he Commission s calculaions (he bubble graph in Figure 9 indicaes how ineremporal ne worh is affeced by he iniial fiscal posiion and he long-run aging coss in various counries) For, he ineremporal ne worh gap is likely underesimaed, as he 29 defici does no ake ino accoun he large cash-accrual upward revision (from 1.9% of GDP o 3.9%), and aging coss do no include recen measures, including he Swiss indexaion rule, increases in minimum and maximum pensions in 29, and a reducion of pension conribuions from 22 percen o 16 percen during For, Porugal and, fiscal daa do no include measures announced afer April 21.

20 19 Figure 9. Ineremporal Ne Worh Derived from he Balance Shee Approach (In percen of GDP) Lihuania Neherlands Esonia Mala Ausria Ialy Porugal Lavia Finie Horizon Ineremporal Ne Worh Lihuania Mala Neherlands Esonia Ausria Porugal Ialy Lavia In fin ie Ho rizon Ineremporal Ne Worh Long-erm aging coss 21-6 (in percen of GDP) Size of bubble reflecs scaled finie-horizon ineremporal ne worh gap Esonia Lavia Ialy Porugal Ausria Lihuania Neherlands Mala Primary balance o GDP, 29 Source: IMF saff esimaes.

21 2 Generally, mos Emerging European counries fare much beer han some advanced counerpars on measures of ineremporal ne worh. In he case of Easern Europe, his may appear somewha counerinuiive, given ha his region is expeced o experience more acue populaion aging han in he res of Europe, wih old age dependency raios exceeding 5 6 percen by 26 in counries such as, Lihuania, and Lavia (recall Figure 2). However, i is precisely hese counries ha have underaken significan pension reforms, in large par as a response o he aging problems hey face. Public social securiy pension sysems have been (parly) swiched ino funded privae defined-conribuion schemes in (1997), (1999), (1999), Lavia (21), Esonia (22), Lihuania (24), Slovakia (25), and (27), wih only he Czech Republic, and failing o do so. Moreover,,, Slovakia, and he Neherlands underook reforms ha ighened early reiremen. Finally, reforms ha swiched public pension sysems from defined-benefi ino noional definedconribuion schemes have been underaken in Ialy (1995), Lavia (1996), (1999) and (1999). 17 Aging coss esimaed by he European Commission ake ino accoun no only populaion aging, bu also he effecs of pension reforms. As a resul, here is a direc correspondence beween reforms and low values of he shorfall in ineremporal ne worh. Sudies ha aim o consruc forward-looking fiscal measures ha focus on dependency raios, bu which do no accoun for he effecs of he reforms could hus lead o flawed resuls. 18 IV. SENSITIVITY ANALYSIS Varying he assumpions can lead o changes in he poin esimaes obained, hough he qualiaive message remains unchanged (Table 3). This is already clear from he comparison of resuls obained from he wo differen mehodologies used in his paper. The parameric approach leads o a lower value of ineremporal ne worh for he in an infinie horizon mainly because he assumpion of a consan real ineres rae of 3 percen, implying a higher ineres-growh differenial compared o ha assumed in he balance shee approach, essenially lowers he value of he discoun rae used in he NPV calculaions (he discoun rae is relaively more predominan a he longer-erm horizon, as shown in Table 3). On he oher hand, he balance-shee approach leads o a lower value of ineremporal ne worh a he finie horizon compared o he parameric approach, because i encompasses a more explici and ambiious medium-erm fiscal adjusmen ha leads o lower primary balances his effec dominaes a he finie horizon, as medium-erm developmens weigh more in he NPV calculaion (Table 3). For simpliciy, boh mehodologies rely on common esimaes for long-run aging coss. However, should fuure healh-care coss be under-projeced, or pension coss rise faser in he fuure due o possible pressures resuling from low replacemen raes, hen ineremporal ne worh would be correspondingly worse Sysems close o his principle include wage poins, as known, for example, in and. See Draxler (29) for more deails. 18 For example, Gokhale (29) finds ha s fiscal imbalance is a whopping -1,5 percen of GDP, mainly because he analysis does no appear o include he effecs of pension reforms on fuure aging coss. 19 See Annex II of he 29 Aging Repor, which presens alernaive healh care cos projecions suggesing ha such coss could be higher by an average of 4.5 percen of GDP relaive o he baseline for he. This poin has also been emphasized by recen IMF papers (see IMF, 21).

22 21 Table 3. : Sensiiviy Analysis, Ineremporal Ne Worh (percen deviaion from baseline) Finie horizon Infinie horizon r-g = r = LR primary defici higher by 1 pp LR primary defici lower by 1 pp MT primary defici higher by 1 pp MT primary defici lower by 1 pp Source: IMF saff esimaes. These indicaors should be inerpreed wih some cauion and be updaed regularly. While radiional indicaors can be more readily measured according o well-esablished mehodologies, he comprehensive public secor balance shee and resuling measures of ineremporal ne worh are based on a series of assumpions and long-run projecions ha are subjec o uncerainy. Reassuringly, he resuls obained here using wo differen mehodologies and differen mediumerm macroeconomic and fiscal assumpions are similar, especially for he as a whole. They are somewha lower han he resuls obained by Gokhale (29), who finds ha he EU 25 s average finie-horizon fiscal imbalance was 435 percen of 24 GDP, hough his laer analysis is based on daa as of 25 and does no incorporae he long-run effecs of pension reforms in many counries. Sill, wha should be emphasized when using hese indicaors is no as much he precise poin value obained, bu raher heir sign, order of magniude, and he direcion in which hey are evolving over ime as a resul of economic developmens and naional policies. V. POLICY IMPLICATIONS A relaively small upfron bu permanen improvemen in he primary fiscal balance can go a long way in reducing he shorfall in ineremporal ne worh (Figure 1). The magniude of he improvemen needed o bring ineremporal ne worh oward posiive erriory in each individual counry depends on he level of is ineremporal imbalance, bu also on is growh prospecs over ime. A simple calculaion can be made o see how much a one percen of GDP permanen upfron improvemen in he primary balance in 21 would yield in NPV erms over he long run, ceeris paribus, by compounding he one percen amoun by he ime-varying discoun rae, which akes ino accoun counry-specific GDP growh raes and he corresponding variable ineres rae. Generally, counries which expec o benefi from a relaively higher growh profile and lower ineres rae, especially in he near erm, end o benefi relaively more from a one percen fiscal adjusmen compared o hose counries whose growh prospecs are weaker and which confron higher ineres raes. 2 For he, a one percen permanen improvemen now would yield abou 2 Again, his is a counerfacual hough experimen which does no ake ino accoun he general-equilibrium feedback effecs from fiscal consolidaion o growh and ineres raes.

23 22 45/12 percen of GDP in ne presen value erms over a finie/infinie horizon. This reveals ha a realisic, non-excessive fiscal effor underaken now can have much larger benefis in he long run. Figure 1. Long-Term Gains from a One Percen Upfron Permanen Fiscal Adjusmen (in percen of GDP) Lavia Lihuania Porugal Ialy Neherlands Ausria Mala Esonia Infinie horizon Finie horizon Source: IMF saff calculaions For mos counries, he fiscal effor required o saisfy he Maasrich defici crierion by 212 would be insufficien o eliminae he ineremporal ne worh gap (Table 4). The new esimaes of he gaps in ineremporal ne worh, ogeher wih he exercise above calculaing he yield of a one percen fiscal adjusmen can, in urn, be used o derive S1- and S2-like ineremporal fiscal gap indicaors (which simply equal he oal ineremporal ne worh gap divided by he yield from a one percen adjusmen). These measures essenially indicae he required upfron fiscal adjusmen required o bring ineremporal ne worh o zero (hus saisfying he ineremporal budge consrain) over a finie or infinie ime horizon. These esimaes can be easily compared o he upfron permanen adjusmens ha would be required under he Maasrich reay (defici gap) o bring he headline defici o 3 percen of GDP by 212. Imporanly, in many counries, he IMF-saff forecased requiremen needed o bring he defici o he Maasrich limi by 212 would be insufficien o make he fiscal posiion susainable, especially a he infinie-horizon. Ineresingly, for a few counries (,,, Porugal, and he ), under he baseline assumpions, achieving a defici of 3 percen of GDP by 212 could be more han sufficien o ensure ha ineremporal ne worh is equal or greaer han zero. For he area as a whole, in addiion o an upfron permanen effor of abou 2.7 percen of GDP o bring he defici o 3 percen of GDP by 212, addiional measures of a leas percen of GDP would be required o also bring ineremporal ne worh ino balance.