Debt-to-GDP Changes and the Great Recession: European Periphery versus European Core. June 26, 2018

Transcription

1 Deb-o-GDP Changes and he Grea Recession: European Periphery versus European Core Maria-Eleni K. Agoraki *, Sella Kardara **,Tryphon Kollinzas *** and Georgios P. Koureas**** June 26, 2018 Absrac: In his paper, we use simple accouning schemes and counerfacual experimens, o compare he sources of he changes in he public deb o GDP across counries of he European Periphery (Greece, Ialy, Spain, Porugal and Ireland), he European Core (Germany, and France) and he oher G7 counries (Japan, UK, Canada and he USA), in wo periods o 2007 and 2008 o In general, deb- o-gdp rose in all counries in he laer period. Bu, he effecs of oal or primary fiscal deficis, inflaion and real growh on he respecive Deb-o-GDP changes were differen across counries in boh of hese periods. Mos imporanly, relaively low inflaion and real growh caused Deb-o-GDP o increase subsanially in he European Periphery counries excep Ireland, and Japan, where counercyclical fiscal policies during and for a few years afer he Grea Recession were unsuccessful and led o moderae Deb-o-GDP increases in he European Core, Ireland, and he Anglo Saxon counries, where counercyclical fiscal policies, alhough differen, worked ou fine, during he same period. This is puzzling for he Euro Area, where moneary policy is common and fiscal policies were resriced by auseriy or sovereign deb resrucuring programs. Keywords: Public deb, Inflaion, Real Growh, Euro Area JEL Classificaions: H63, E31, E58, F45 We are graeful o Riccardo Fiorio, Dimiris Papageorgiou, Joseph Zeira and research seminar paricipans a Hebrew Universiy of Jerusalem and Ahens Universiy of Economics and Business for valuable commens and suggesions. The usual cavea applies. *Deparmen of Business Adminisraion and Deparmen of Economics, Ahens Universiy of Economics and Business ** Deparmen of Economics, Ahens Universiy of Economics and Business *** Deparmen of Economics, Ahens Universiy of Economics and Business, CEPR and ADEMU *** Deparmen of Business Adminisraion and Deparmen of Economics, Ahens Universiy of Economics and Business and IPAG Business School, Paris France 1

2 1. Inroducion Boh in academic and policy debaes, mos references o public deb are made in erms of he public deb o GDP raio (henceforh, Deb-o-GDP ). For example, Deb-o-GDP is he focal poin in analyzing public deb susainabiliy and governmen solvency and he Maasrich Treay, one of he pillars of he Euro Area (henceforh EA), ses arges for public deb and public deb changes wih reference o Deb-o-GDP. I is, herefore, imporan o measure and analyze he sources of deb o GDP changes, in general. In paricular for he EA, given he common moneary policy and he common resricions on fiscal policy, i is imporan o compare he sources of Deb-o-GDP changes of member counries and undersand he causes behind any dispariies. Moreover, he sudy of he sources of Debo-GDP raio has received exra aenion recenly due o he Grea Recession. 1 Tha is, for a long ime before he Grea Recession a more or less seady nominal GDP growh confined aenion o he Debo-GDP numeraor and he budge deficis. Bu, he Grea Recession came o change his, as hree quarers of he OECD counries experienced a nominal recession. 2 And, policy influenial economiss (see for example Rogoff (2008, 2010), Blanchard e al. (2010), and Ball (2012)) advised for higher inflaion raes o lower he real value of deb during he Grea Recession and is afermah. This nominal recession also hi he Euro Area as a whole and especially he EA Periphery (i.e., Greece, Ireland, Ialy, Spain and Porugal), where he coexisence of lile (or even negaive) real growh wih lile inflaion (or even deflaion) increased Deb-o-GDP also hrough he denominaor. This is of course raises an ineresing issue for he EA, where Deb-o-GDP arges focus on he numeraor and essenially ignore he denominaor. However, his issue has been almos ignored in he conduc of moneary policy and he design of fiscal policy in he EA, despie he fac ha acual inflaion has been eiher negaive or well below he 2% hreshold and despie he fac ha inflaion raes varied considerably across counries during he Grea Recession and for several years hereafer. The deph and duraion of he recession acivaed insead a debae on he conduc of fiscal policy in bad imes and especially he appropriaeness of he policy of auseriy. 3 1 By Grea Recession, here, we mean he sharp decline in he level of economic aciviy ha occurred in 2008 and 2009 in he USA and spread ou o mos of he res of he world, hereafer. 2 See e.g., Fiorio (2013). 3 By policy of auseriy, here, we mean a siuaion where governmen aims o reduce he srucural defici (i.e., he defici adjused for he business cycle). envisioned by he so called six pack rules. See, e.g., Alesina e al (2014). For a concise definiion of he Six Pack rules, see hp:// 2

3 The purpose of his paper is hreefold. Firs, i is o measure he sources of Deb-o-GDP changes in seleced EA counries and compare hem o oher major economies. Second, i is o use hese measuremens in order o shed some ligh on he reasons hese sources differ across counries. Third, i is o examine he implicaions of he reasons he sources of Deb-o-GDP changes differ across counries for he conduc of fiscal and moneary policy in he EA. We use annual daa from he European Periphery, he EA Core counries (i.e., France and Germany) as well as he res of G7 counries (i.e., Japan, he UK, Canada and he USA), o sudy he sources of Deb-o-GDP changes in he and periods, from a macro-accouning perspecive. The selecion of counries reflecs our inenion o compare he EA Periphery o he EA Core counries. The oher G7 economies are used as a benchmark. The period includes he Grea Recession years (i.e., 2008 and 2009) and he six years ha followed, where real growh was negaive or relaively weak, especially when compared o he period. 4 The laer period was simply chosen for comparison purposes and, in paricular, o assess he counry-specific impac of he Grea Recession on Deb-o-GDP. The macro-accouning perspecive (i.e., an adapaion of he sandard budge consrain of macro exbooks) used focuses on inflaion, real growh, he primary fiscal defici and he ineres paymens on ousanding deb. Clearly, by doing so, we ignore issues peraining o he composiion (e.g., currency, inflaion proecion, ec.), mauriy, ineres srucure and holders (e.g., foreign or domesic, privae or public agens) of deb. In paricular, we use a macro-accouning decomposiion of Deb-o GDP changes o obain simple sylized facs on he sources of hese changes across counies and periods. Furher, we use his decomposiion o carry ou several counerfacual experimens, so as o ge a quaniaive assessmen of he conribuion of he sources (i.e., inflaion, real growh and oal or primary fiscal deficis) of Deb-o-GDP changes. In hese experimens, we obain numerical values of wha would had happened o Deb-o-GDP a he end of a period, if insead of is acual value, inflaion, for example, was 2% for some ime before his period, bu all oher sources assumed heir hisorical values over ha ime. I should be saed a he ouse, ha in so doing we absrac from source independence and causal implicaions ha would had necessiaed assumpions abou he deerminaion of ineres raes, he Fisher effec, fiscal policy mulipliers and he relaion beween deficis and inflaion. For ha maer, our work should be viewed as simple sylized facs analysis ha provides food for hough, ha ouches upon various heory and policy issues. 4 See he Hodrick-Presco Filer decomposiions in Figures A.1 in he Appendix. Noe ha, for all counries in he sample, he business cycle componen of he filer is negaive a leas unil

4 Boh simple sylized facs and counerfacual experimens poin ou ha, in general, deb- o- GDP rose in all counries in he period. Bu, he effecs of primary deficis, inflaion and real growh on he respecive Deb-o-GDP changes were differen across counries. Mos imporanly, relaively low inflaion and real growh caused Deb-o-GDP o increase subsanially in he European Periphery counries excep Ireland, where counercyclical fiscal policies during and for a few years afer he Grea Recession were unsuccessful. In Germany, Ireland, and he Anglo Saxon counries, where counercyclical fiscal policies, alhough differen, worked ou fine, during he same period moderaed Deb-o-GDP increases were deeced. This is puzzling for he Euro Area, where moneary policy is common and fiscal policies were resriced by auseriy, as in he case of Germany and Ialy or sovereign deb resrucuring programs as in he case of he European Periphery counries. The plan of his paper has as follows: Secion 2 gives a brief lieraure review. In Secion 3 we provide a discussion of he Eurozone rescue programmes. Secion 4 examines he behavior of real and nominal GDP, Deb-o-GDP, and he sources of Deb-o-GDP changes, based on a simple macroaccouning scheme. Based on his scheme, in Secion 5, we use counerfacual experimens o measure he effec each one of he sources would have on Deb-o-GDP, over a given period, if i had followed a differen pah han he one i acually followed during his period, while all oher sources coninue o assume heir hisorical values, whereas secion 6 concludes wih a discussion of he implicaions of our findings for he design of moneary and fiscal policy in he EA. A he end of he paper here is an appendix wih daa sources, he consruced daa, and some furher daa analysis, including he correlaion properies of he sources of Deb-o-GDP changes in each counry. 2. Review of lieraure The empirical lieraure on he sources of changes o he Deb-o-GDP raio is raher fragmened. Mos sudies are for he Unied Saes 5. To he bes of our knowledge, his is he firs paper ha seeks o compare he effecs of a comprehensive lis of sources of changes o he Deb-o-GDP raio across counries. There is an exensive lieraure on he effecs of inflaion on he real value of he public deb or he on he Deb-o-GDP raio ha ouches, more or less, superficially on he oher possible conribuors of he Deb-o-GDP changes. However, his lieraure is imporan for our purposes for i 5 Giannisarou and Sco (2011) and Akioby e al. (2014), ha consider he G7 excep France and he G7, respecively, been he wo noable excepions. 4

5 deals wih all he heoreical issues in our analysis and mos of he empirical ones, as well. The papers in his lieraure could be divided in hree groups, depending on he mehodology used o assess he perinen effecs of inflaion: descripive or informal, parial equilibrium and general equilibrium analysis. Papers in he firs group include he above menioned papers by Rogoff (2008, 2010), Blanchard e al. (2010), and Ball (2012) ha, in general, have advised for higher inflaion raes o lower he real value of deb, in he Grea Recession and is afermah. Papers in he second group include papers ha measure, simulae, or esimae he effecs of inflaion on he Deb-o-GDP raio. In a paper ha is imporan for many reasons for wha we do in his paper, Eisner and Pieper (1984) and Eisner (1986) consider he marke value of deb and hey inroduce he concep of ne deb. The laer is defined as gross governmen deb minus high power money and he marke value of governmen asses. Furher, hey obain he ineres paymens associaed wih ne deb residually, from he marke value of ne deb minus he primary governmen defici. Their mehod o obain governmen ineres paymens and he ineres rae of governmen deb may be hough as a way o reconcile he sandard governmen budge consrain wih he daa, given he fac ha governmen defici and he change in he value of deb are based on cash accouning and capial accouning procedures, respecively. Bohn (1992) applies a similar procedure in obaining he ineres rae of governmen deb, using exclusively gross governmen deb. Giannisarou and Sco (2008) loglinearize around he seady sae of a saionary ransformaion of he governmen budge consrain and solve i forward o express he raio of he marke value of governmen deb o he curren primary governmen surplus, which hey call measure of fiscal imbalance, in erms of he expeced fuure values of primary deficis, inflaion, real growh and real ineres raes. Then, hey esimae a VAR of hese variables o subsiue he expeced fuure values of primary deficis, inflaion, real growh and real ineres raes o express he measure of fiscal imbalance in erms of he curren and lagged values of primary deficis, inflaion, real growh, and real ineres raes; and obain he associaed variance decomposiion. Thus, hey esimae how much of he observed variance in he measure of fiscal imbalance is due o changes in he primary defici, inflaion, real growh, and he real ineres rae. Oher han he need for he saionary ransformaion, despie he employmen of he forward looking soluion, he erm srucure of ineres paymens on governmen deb is no explicily aken ino consideraion. The nominal ineres rae in he governmen budge consrain is approximaed using averages for he mauriy and he yield curve of governmen bonds. Hall and Sargen (2011), consider indexed and nominal deb in he form of differen mauriy bonds. Curren deb is defined as he sum of 5

6 he produc of he number of each differen mauriy bond ousanding in he presen period imes he curren price of he corresponding bond. They reconcile he governmen budge consrain wih he daa by consrucing appropriaely defined nominal ineres raes for deb of differen mauriy. Then, hey solve he governmen budge consrain forward o measure he effecs on he Deb-o-GDP raio of he primary defici, inflaion, real growh, and he consruced nominal ineres raes. In so doing, hey use a simple accouning mehod o measure hese effecs. In he pos WWII US economy, inflaion is found o have had a smaller impac on he reducion in he deb-o GDP han real growh. In his paper, we follow a similar procedure. Akioby e al. (2014) incorporae seignorage in he governmen budge consrain and consider hree ypes of deb: domesic currency denominaed, foreign currency denominaed, and inflaion-indexed deb. On anoher dimension, hey consider shor, medium, and long erm deb. They use daa, circa 2012 o calibrae he parameers of heir model for he G7 counries and hen hey simulae he effecs of differen consan inflaion raes over he nex five years, under wo scenarios. In he firs scenario hey assume a full Fisher effec and in he second, hey assume a parial Fisher effec. They find ha, on he average, inflaion reduces he Deb-o-GDP raio, over he five-year period, only moderaely. Aizenman and Marion (2011) consruc a sylisic dynamic general equilibrium model, bu wih a fixed real rae of ineres. They model an economy wih public deb overhung, where public deb is held by domesic and foreign economic agens and a fracion of i is indexed, while he res is no. hey disinguish beween inernal simple. The fiscal/moneary auhoriy acs as a Ramsey planner in seing he inflaion and ax raes. The model is calibraed for he US economy, circa Inflaion is used o bring abou a significan reducion in he real value of deb, even wih mauriies geing shorer. This is mainly due o he fac ha deb o and he mauriy y raes o governmen deb o he curren primary governmen defici surplus o esimae ineres raes sum of he weighed bin capial accouning approach has been used in a number of oher sudies of he as Recenly, Krause and Moyen (2016) formulaed a New Keynesian DSGE model, incorporaing long-erm bonds wih sochasic mauriy and learning abou he inflaion arge of he moneary auhoriy ha follows a Taylor rule. They calibrae heir model for he US economy and hey find ha inflaion arge changes have a significan effec on he real value of public deb, if hey are essenially permanen. In conclusion, he general picure ha emerges from his lieraure is ha inflaion can reduce in heory and has in pracice reduced he real value of public deb or he Deb-o-GDP raio, up o a 6

7 degree. This degree depends on several facors and mainly on he fracion of deb ha is denominaed in foreign currency, he fracion of deb ha is indexed for inflaion, seignorage, he mauriy srucure of deb, he abiliy of he moneary auhoriy o cause inflaion, he way expecaions abou fuure inflaion are formed, and he mauriy srucure of deb and he way i changes wih inflaion expecaions and he risk of defaul. Mos imporanly, however, he lieraure on he effecs of inflaion on he real value of deb and he Deb-o-GDP raio serves as an imporan background for placing properly he conribuion of his paper. 3. The European Periphery Rescue Programs The European deb crisis (ofen also referred o as he Euro Area crisis or he European sovereign deb crisis) is a muli-year deb crisis ha has been aking place in he European Union since he end of Euro Area member saes: Greece, Porugal, Ireland, Spain and Cyprus, were unable o repay or refinance heir governmen deb or o bail ou over-indebed banks under heir naional supervision wihou he assisance of hird paries like oher Euro Area counries, he European Cenral Bank (ECB), or he Inernaional Moneary Fund (IMF). 1. Greece The Greek governmen-deb crisis (also known as he Greek Depression) is he sovereign deb crisis faced by Greece in he afermah of he financial crisis of The Greek crisis sared in lae 2009, riggered by he urmoil of he Grea Recession, srucural weaknesses in he Greek economy, and revelaions ha previous daa on governmen deb levels and deficis had been undercouned by he Greek governmen. This led o a crisis of confidence, indicaed by a widening of bond yield spreads and rising cos of risk insurance on credi defaul swaps compared o he oher Euro Area counries, paricularly Germany. The governmen enaced a dozen rounds of ax increases, spending cus, and reforms from 2010 o 2016, which a imes riggered local rios and naion-wide proess. Despie hese effors, he counry required bailou loans in 2010, 2012, and 2015 from he Inernaional Moneary Fund, Euro group, and European Cenral Bank, and negoiaed a 50% "haircu" on deb owned by he privae secor in Afer a popular referendum which rejeced furher auseriy measures required for he hird bailou, and afer 7

8 closure of banks across he counry (which lased for several weeks), on 30 June 2015, Greece became he firs developed counry o fail o make an IMF loan repaymen. A ha ime, deb levels had reached 323bn or some 30,000 per capia. 2. Porugal Porugal has suffered from low GDP and produciviy growh for more han a decade before he oubreak of he recen crisis. Poenial oupu growh has been on a seady downward rend, wih compeiiveness being undermined by rising uni labour coss and deep-rooed srucural problems. As a consequence of persisen curren accoun deficis, Porugal has accumulaed a high exernal deb, which is refleced in high household, corporae and fiscal debs. On he posiive side, Porugal did no winess a propery boom and bus, nor were banks exposed o oxic asses. Therefore, he financial secor has weahered he firs phase of he financial crisis relaively well. The period before he reques for financial assisance was marked by unfavourable developmens in public finances and a worsening economic oulook. This led o a deerioraion of confidence and rising marke pressures on Poruguese deb, accenuaed by he negaive developmens in euro area sovereign bond markes. Amids consecuive downgrading by credi raing agencies of Poruguese sovereign bonds, he counry became unable o refinance iself a raes compaible wih long-erm fiscal susainabiliy. In parallel, he banking secor, which is heavily dependen on exernal financing, was increasingly cu off from inernaional marke funding and had o sep up reliance on he Eurosysem for funding. Following a reques by Porugal on 7 April 2011, he Troika consising of he European Commission, ECB and IMF negoiaed an Economic Adjusmen Programme, aimed a resoring confidence, enabling he reurn of he economy o susainable growh, and safeguarding financial sabiliy in Porugal, he euro area and he EU. The Programme was agreed by he European Council on 30 May 2011 and he IMF board on 20 May. I covers he period Is financial package will cover up o EUR 78 billion for possible fiscal financing needs and suppor o he banking sysem. One hird (up o EUR 26 billion) will be financed by he European Union under he European Financial Sabilisaion Mechanism (EFSM), anoher hird by he European Financial Sabiliy Faciliy (EFSF), and he final hird by he IMF under an Exended Fund Faciliy. The programme foresees comprehensive acion on hree frons. 8

9 Firs, a credible and balanced fiscal consolidaion sraegy, suppored by srucural fiscal measures and beer fiscal conrol over Public-Privae-Parnerships (PPPs) and sae-owned enerprises (SOEs), aiming a puing he gross public deb-o-gdp raio on a firm downward pah in he medium erm. The auhoriies are commied o reducing he defici o 3% of GDP by Second, effors o safeguard he financial secor hrough marke-based mechanisms suppored by backup faciliies. Cenral aspecs here are measures o foser a gradual and orderly deleveraging, srenghened capialisaion of banks, reinforced banking supervision and he sale or unwinding of he BPN bank. Third, deep and fronloaded srucural reforms o boos poenial growh, creae jobs, and improve compeiiveness. In paricular, he Programme conains reforms of he labour marke, he judicial sysem, nework indusries and housing and services secors, wih a view o srenghening he economy's growh poenial, improving compeiiveness and faciliaing economic adjusmen. An imporan feaure of he programme is he aim o miigae negaive social impacs, while addressing fiscal, banking and srucural imbalances a he same ime. In paricular, ax increases and benefi reforms are designed such as o minimise he impac on he lowes income groups. 3.Ireland The urnaround in Ireland s economic forunes in recen years is perhaps he mos dramaic of any counry in he euro area. As recenly as 2007, Ireland was seen by many as op of he European class in is economic achievemens. A long period of high raes of economic growh and low unemploymen had been combined wih budge surpluses. The counry appeared well placed o cope wih any economic slowdown as i had a gross deb GDP raio in 2007 of 25% and a sovereign wealh fund worh abou 5,000 a head. However, he subsequen crash involving a housing marke collapse, soaring unemploymen and a full scale banking crisis proved oo difficul for he Irish governmen o manage on is own. In 2010, Ireland agreed o an adjusmen program wih he EU and IMF. Today, Ireland is poised o exi his program and, while economic condiions remain poor and unemploymen elevaed, he counry is again being cied regularly as an example for oher counries in severe economic difficulies. 9

10 4. Spain The 2008-presen Spanish financial crisis, also known as he Grea Recession in Spain or he Grea Spanish Depression began in 2008 during he world financial crisis of In 2012 i made Spain a lae paricipan in he European sovereign deb crisis when he counry was unable o bailou is financial secor and had o apply for a 100 billion rescue package provided by he European Sabiliy Mechanism (ESM). The main cause of Spain's crisis was he housing bubble and he accompanying unsusainably high GDP growh rae. The ballooning ax revenues from he booming propery invesmen and consrucion secors kep he Spanish governmen's revenue in surplus, despie srong increases in expendiure, unil The Spanish governmen suppored he criical developmen by relaxing supervision of he financial secor and hereby allowing he banks o violae Inernaional Accouning Sandards Board sandards. The banks in Spain were able o hide losses and earnings volailiy, mislead regulaors, analyss, and invesors, and hereby finance he Spanish real esae bubble. The resuls of he crisis were devasaing for Spain, including a srong economic downurn, a severe increase in unemploymen, and bankrupcies of major companies. Even hough some fundamenal problems in he Spanish economy were already eviden far ahead of he crisis, Spain coninued he pah of unsusainable propery led growh when he ruling pary changed in In hese early imes Spain had already a huge rade defici, a loss of compeiiveness agains is main rading parners, an above-average inflaion rae, house price increases, and a growing family indebedness. During he hird quarer of 2008 he naional GDP conraced for he firs ime in 15 years, and, in February 2009, Spain (and oher European economies) officially enered recession. The economy conraced 3.7% in 2009 and again in 2010 by 0.1%. I grew by 0.7% in By he 1s quarer of 2012, Spain was officially in recession once again. The Spanish governmen forecas a 1.7% drop for The provision of up o 100 billion of rescue loans from euro zone funds was agreed by eurozone finance minisers on 9 June As of Ocober 2012, he so-called Troika (European Commission, ECB and IMF) is in negoiaions wih Spain o esablish an economic recovery program required for providing addiional financial loans from ESM. In addiion o applying for a 100 billion bank 10

11 recapializaion package in June 2012, Spain negoiaed financial suppor from a "Precauionary Condiioned Credi Line" (PCCL) package. If Spain applies and receives a PCCL package, irrespecively o wha exen i subsequenly decides o draw on his esablished credi line, his would a he same ime immediaely qualify he counry o receive "free" addiional financial suppor from ECB, in he form of some unlimied yield-lowering bond purchases. The urning poin for he Spanish sovereign deb crisis occurred on 26 July 2012, when ECB Presiden Mario Draghi said ha he ECB was "ready o do whaever i akes o preserve he euro". Announced on 6 Sepember 2012, he ECB's Ourigh Moneary Transacions (OMT) program of unlimied purchases of shor-erm sovereign deb pu he ECB's balance shee behind he pledge. Speculaive runs agains Spanish sovereign deb were discouraged and 10-year bond yields sayed below he 6% level, approaching he 5% level by he end of Sylized Facs and he Sources of Deb-o-GDP Changes Given he problem a hand, he available daa are few because of he annual frequency which, however, is sill preferable evaluaing NIPA governmen daa in deail. Before reporing he Deb-o- GDP raio changes and heir sources, we firs presen he real and nominal GDP changes, o ascerain he magniude of he economic losses hiing each counry during and afer he Grea Recession. As shown in Table 1, here are grea differences in real growh beween and All counries experienced robus growh in This growh was higher in he Anglo Saxon counries and in he EA Periphery, excep Ialy and Porugal. In he period, boh average and cumulaive oupu losses look impressive for he EA Periphery excep Ireland. On he conrary, he European Core, Ireland, and he Anglo Saxon counries, afer he iniial oupu losses in 2008 and/or 2009, rebounded relaively fas o posiive, bu relaively low, real growh. 6 The real novely of he Grea Recession is he fall in nominal GDP ha characerized all counries considered here. In fac, for Japan and he EA Periphery nominal oupu declined for a number of years afer 2009 (i.e., Greece ( ), Ireland ( ), Ialy (2009, 2011, 2012), Japan (2008, 2009, 2011), Porugal (2009, 2011, 2012) and Spain (2009, )). This is manifesed, in Table 1, by he fac ha nominal growh differences beween he and he are greaer han he corresponding real growh differences. The only noable excepions are Germany and Japan bu for 6 See, again, he Hodrick Presco filer decomposiions in Figure A.1, in he Appendix. 11

12 differen reasons. In he case of Japan he reason is he relaively low inflaion in and in he case of Germany is he relaively high inflaion in he period. Now, Table 2 presens he average and cumulaive Deb-o-GDP raio changes in he wo periods under consideraion. As already menioned, deb, here, refers o he gross deb of General Governmen. And, we use wo measures of his deb. Firs OECD s Gross Deb (SNA08) for all counries and OECD s Gross Deb (Maasrich Definiion) for he EA counries and he UK. There are wo main differences beween he Gross Deb (SNA08) and he Gross Deb (Maasrich Definiion) measures of deb. The firs difference is ha he Maasrich deb does no include liabiliies relaed o oher accouns payable (comprising rade credis and advances), financial derivaives, and insurance echnical reserves. The second difference concerns he valuaion mehodology. Tha is, he Maasrich definiion evaluaes deb a face value, which is equivalen o he amoun ha he governmen has o pay back o crediors a mauriy. In conras, he SNA08 deb employs marke values. Maasrich deb is, hus, a beer measure for assessing governmen refinancing needs, bu he SNA08 deb capures more adequaely he cos of buying back deb. For non-radable deb insrumens marke valuaion is no available, requiring impuaion of prices by some alernaive mehods. Also, marke valuaion migh be problemaic for radable insrumens when markes are volaile and/or illiquid. Consequenly, he SNA08 deb measure can be more volaile han he Maasrich measure. Idenifying he exac conribuions of he various facors o differences in deb levels according o he SNA08 and Maasrich deb measures is no sraighforward. For all counries, Deb-o-GDP raio behaves quie differenly in han in In paricular, while in he former period, Deb-o-GDP raio eiher falls (i.e., in he case of Canada, Ireland, Ialy, and Spain), is nearly consan (i.e., in he case of he European Core), or rises modesly (i.e., in he case of Japan, Greece, Porugal, UK and he USA), i rises sharply in he laer period. In he EA, he smalles cumulaive rises occur in he Germany (i.e., abou 10 percenage poins in Germany, using he Maasrich measure); and, he larges in he European Periphery (i.e., abou 152, 121, 80 and 62 percenage poins in Spain, Ireland, Porugal and Greece, respecively, using he Maasrich definiion). France and Ialy have moderae cumulaive increases in he period (i.e., abou 41 and 30 percenage poins in France and Ialy, resecively, using he Maasrich Definiion). Finally, comparing he cumulaive Deb-o-GDP increases in he period, we may clasify counries in hree groups, having more han 15 percenage poins difference: (a) Germany wih he lowes such increase (i.e, abou 15 percenage poins), (b) counries wih moderae increases (i.e., beween 34 and 48 percenage poins) 12

13 such as Japan, Ialy, Canada, he UK and France) and counries wih high increases (i.e, beween 65 and 156 percenage poins) such as Greece, he USA, Porugal, Ireland and Spain. 7, 8, 9 As i will be explained below, he observed increases in he deb-o-gdp raio in he period occurred for differen reasons. And, nowhere were hese reasons more differen han in he case of Germany and he EA Periphery. We urn nex o he sources of Deb-o-GDP changes. To do his, we need an accouning scheme for he Deb-o-GDP raio. The accouning scheme used is based on he arihmeic implied by he simples governmen deb definiion. This scheme is adoped in order o decompose Deb-o-GDP changes ino a minimum number of clearly idenifiable sources, based on mere accouning raher han on causal relaions among variables. Thus, we consider he simples ineremporal governmen budge consrain: B B D (1) 1 where B sands for gross public deb a he end of period and D sands for he governmen budge defici in period. This defici is measured here by OECD s General Governmen Ne Lending series. Boh, B and D variables are nominal (i.e., measured in curren prices). Bu, as has been long recognized, increases (decreases) in he deb sock may differ someimes widely from he corresponding defici (surplus) measuremen, because deb flows are evaluaed on a cash raher han on he accrual basis used in he (NIPA) naional accouns. Moreover, here are oher reasons for differences beween hese wo variables. Firs, deb cash flows sem also from governmen ransacions on non-financial asses such as land, buildings and so forh. 10 Second, changes in he deb sock may be due o naional deb resrucuring loans and oher governmen loans such as hose for he rescue of privae banks, such as hose obained by he European Periphery counries afer he Grea Recession. 7 Greece is group in he high raher han he moderae increase group for in 2012 had a percenage poins reducion in is Deb-o- GDP, brough abou by he he is Firs European Periphery Rescue Program.. 8 Considering he EA counries, he Grea Recession in general inerrups he convergence process measured on differen variables by Esrada e al. (2013) and confirmed here, also, for he Deb-o-GDP raio. 9 I is ineresing o noe ha in he same period he privae deb-o-gdp raio generally rises, also. As i has been poined ou by a number of policy analyss, he simulaneous rise of privae deb before and during he crisis is imporan for assessing he overall deleveraging in he same period. See, e.g., Buier and Rahbari (2011). 10 The accouning and fiscal policy implicaions of he way of defining governmen defici and deb were firsly oulined by Eisner and Pieper (1984). An auhoriaive modern sudy on he same issues is Robinson (2009). 13

14 For ha maer, we shall define he cash/accruals accouning discrepancy E, as he difference beween he change in he sock of public deb during a period and he governmen budge defici in he same period: E B B D (2) 1 Toal defici is broken down in wo componens: P D D ib 1 (3) where P D and 1 ib sand for he primary budge defici and he ineres paymens on ousanding public deb, in period, respecively. We shall refer o he ineres rae, i, defined by (4), as he effecive ineres rae of public deb in period. Clearly, his ineres rae measure varies wih he definiion and measuremen of public deb. Moreover, his represenaion of ineres paymens on ousanding public deb ignores ineres paymens paid on deb creaed during a period and absracs from he issue of he ineres rae srucure of he bonds comprising public deb. 11 Furhermore, his ineres rae is fundamenally differen from he ineres rae concep of macroeconomic models. 12 Combining (2) and (3) gives he following law of moion of public deb: B (1 i ) B D E (4) P 1 Le Y and y sand for nominal and real GDP, respecively, so ha, Y Py, where P sands for he GDP price deflaor. Then, he law of moion of he Deb-o-GDP raio, b B Y, can be expressed as follows: 11 Daa on his are available for he EA in Lojsch, Rodriguez-Vives and Slavik (2011). 12 For convenience, in macroeconomics models, he erm srucure of ineres paymens on public deb is, ypically, ignored. The sandard assumpion in hese models is ha governmen issues one period bonds. Thus, a he beginning of every period governmen buys back he enire sock of ousanding public deb and issues he new sock of public deb. Then, when he underlying ineres rae is deermined becomes anoher assumpion. Typically, i is deermined a he same ime deb is issued. Thus, curren period ineres paymens on ousanding deb are exogenous bu all fuure period ineres paymens on ousanding deb are endogenous. Taking ino accoun he erm srucure of ineres paymens of bonds ha, ypically, maure several years from heir issue, makes, of course, fuure ineres paymens on ousanding public deb are parially predeermined. For our purposes, his issue is imporan, for i relaes o he relaionship beween curren ineres rae paymens and pas inflaion expecaions. 14

15 (1 i ) b b d e p 1 (1 )(1 ) (5) where: ( y sands for he real growh rae, y 1) y 1 and, sands for he inflaion rae, ( P P ) D 1 P 1, p d Y is he primary defici share of GDP and e E Y is he cash/accruals accouning discrepancy share of GDP. The laer variable combines all possible discrepancies beween governmen deb changes and governmen balance deficis as hey are measured in pracice and should no be inerpreed as a zero mean, whie noise residual. We shall refer o he variables in he RHS of Eq. (5), oher han he Deb-o-GDP a he end of he las period, as he sources of Deb-o-GDP changes. Thus, Deb-o-GDP raio increases wih he GDP share of primary defici, he GDP share of cash/accrual accouning discrepancy, and he effecive ineres rae on exising governmen deb. And, i decreases wih he real growh and inflaion raes. The firs hree columns of Table 3 presen he averages of he real growh rae, he inflaion rae, and he primary defici share of GDP, in he and he The las four columns of Table 3 repor he averages of he effecive ineres rae and he GDP share of he cash/accrual accouning discrepancy, using he SNA08 and he Maasrich measures. Turning now o he sylized facs of he sources of Deb-o-GDP changes, summarized in Table 3, we observe he following: Real Growh As already menioned, all counries are characerized by real growh in he period. In paricular, real growh was srong in he Anglo Saxon counries and he EA Periphery, excep Porugal, during his period. 13 On he conrary, in all counries real growh is eiher negaive or posiive bu relaively weak, in he Average real growh raes in were negaive in Ialy and he EA Periphery, excep in Ireland. The only counries where he average growh rae of he period is more han 50% of he corresponding rae of he period are Canada and Germany. Wha, is no refleced, due o averaging, in Table 3 are he differences in he way counries came ou of he Grea Recession. As i urns ou, his has an imporan implicaion for he behavior of he Deb-o-GDP across counries in he enire period. 13 This is clearly illusraed in he graphs of he long erm growh componen of he Hodrick Presco filer in Figures A.1 he Appendix. 15

16 Inflaion All counries, wih he excepion of Japan, are characerized by inflaion in he period. In fac, all oher counries have an average inflaion rae in excess of he 2% hreshold, wih he excepion of France (i.e., 1.97%) and, again, Germany (i.e., 0.89%). In all counries, inflaion is lower in he period, wih he excepion of Japan and Germany (i.e., 1.48%). However, alhough in he Anglo Saxon counries, average inflaion rae in he period was less han 40% smaller han he respecive rae of he period, i was several imes smaller in he in he EA Periphery. In oher words, he inflaion drop in he EA periphery was (across counries and periods) relaively large. Bu, in Germany, no only he average inflaion rae of he period was higher han he corresponding rae of he period, bu he German average inflaion rae of he period was much higher han he corresponding rae of all EA Periphery counries. Primary Defici In he period and on he average, Japan, Canada, Porugal, and he USA run primary surpluses, he UK and he European Core had, roughly, balanced primary budges, and he European Periphery, excep Porugal run moderae (i.e, beween 1.3% and 2.5% of GDP) primary deficis. Thus, since in he period GDP was rising in all counries, we may surmise ha discreionary fiscal policy in his period was counercyclical in Japan, Canada, Porugal, and he USA; neural in he UK and he European Core; and procyclical in he European Periphery, excep Porugal. In he peeriod, hings were more complicaed. As i urns ou, he ime aggregaion of Table 3 is no longer helpful, here. For ha maer, he analysis ha follows is based on he annual daa presened in Tables A.1 of he Appendix. During he Grea Recession years (i.e., 2008 and 2009), all counries engaged in expansionary discreionary fiscal policy, wih he excepion of Germany and Ialy in Cross checking hese figures wih he corresponding figures of GDP growh, i urns ou ha all counries in heir respecive recession years engaged in counercyclical discreionary fiscal policy, wih he excepion of Ialy in Afer he Grea Recession, all counries engaged in expansionary discreionary fiscal policy, wih he excepion of Germany and Ialy. To figure ou wheher hese policies were counercyclical or procyclical we may disinguish counries in hree groups. Firs, counries ha successfully faced he Grea Recession, in he sense on revering o posiive real growh, hereafer. As i urns ou, all hese counries had a posiive average real growh rae in he During/Afer he Grea Recession period. I follows, ha he counries in his group are Canada, he USA, France, 16

17 Germany and Ireland. Hence, discreionary fiscal policy afer he Grea recession was procyclical in Canada, he USA, France and Ireland and counercyclical in Germany. Second, counries ha did no successfully faced he Grea Recession, in he sense of having a negaive average real growh rae in he During/Afer he Grea Recession period. The counries in he second Group are Ialy and he EA Periphery counries, excep Ireland. I follows ha all he EA Periphery counries were pursuing a counercyclical fiscal policy, while, surprisingly, Ialy followed a procyclical discreionary fiscal policy, even when real oupu was declining. 14 The hird group consiss of he remaining counries. Tha is, counries ha had a leas one year afer wih negaive real growh, afer he Grea Recession, bu heir average real growh rae in he During/Afer period was no negaive. The UK and Japan are he only wo counries in he hird group. The cyclicaliy of discreionary fiscal policy in hese counries in no sraighforward, bu i is, in general, procyclical and resembles he behavior of he counries in he firs group, excep, of course, Germany We may conclude herefore, ha all counries faced he Grea Recession wih counercyclical fiscal policy ha obviously conribued o he increase in he Deb-o- GDP raio, bu afer he Grea Recession, in he EA Periphery excep Ireland, here were counries were heir Deb-o-GDP increases were relaively large, because of unsuccessful counercyclical discreionary fiscal policies. In Ireland, as wih he oher counries excep, Germany and Ialy, here were relaively increases in Deb-o-GDP as hey had successful counercyclical fiscal policies during 14 Following on Foonoe 16, a fac ha presumably explains he apparen paradox ha Ialy is here, along wih Germany, he only counry running a primary surplus in mos of he crisis years is ha he Before he large inheried deb severely consrained fiscal policy. Furher, he defici size increased in he period less han elsewhere because of he EA surveillance and also of Ialy s relucance o be formally involved - as Spain, among he counries here - in an excess defici procedure (EDP). Overall, he impression is no only of a limied couner-cyclicaliy of fiscal policy because of he large deb burden bu also of a lagging policy response wih respec o he oher EU counries which appears confroning he 2009 shock. This probably reflecs he limis of fiscal auomaism, given he inhibied possibiliy of facing excepional (or a leas unusual) evens as recessions are. 15 Among he counries in he sample, Ialy has a special posiion no only for having one of he highes Deb-o-GDP in he years, bu also for exceeding he symbolic 100% hreshold even before he big crisis began. This probably explains why in he firs period an effor was made o slighly reduce he deb raio ha, in he second, increased less han in all he oher counries in he sample wih he excepion of Germany. The high-deb heriage cerainly reduced he available fiscal space (Osry e al., 2010), severely limiing he possibiliy of conrasing he Grea Recession wihin he EA framework. As a resul, almos no fiscal discreion seems working in he years, characerized no only - as in mos OECD counries - by a negaive real growh in 2009, bu also by a nominal recession in he biennium. 16 Using he Hodrick-Presco filer decomposiion of oupu daa in is smooh growh and is business cycle componens, i seems ha discreionary fiscal policy in Germany and Ialy, was counercyclical in he During and several years afer he Grea Recession. 17

18 he recession, followed by cauious procyclical policies afer he recession. min he case of he Anglo Saxon raio increased relaively a lwo groups of counries ha. 17 Effecive Ineres Rae Finally, when i comes o effecive ineres raes, hey have a posiive effec on Deb-o-GDP changes. From columns 4 and 5 of Table 3, i follows ha for all counries, effecive ineres raes were higher in he BGRP han in he DAGRP. However, he differences beween hese wo periods, across counries seem raher small. For insance, he average effecive ineres raes, using he SNA08 measure, in he During/Afer he Grea Recession period are no smaller han 70% of heir BGRP counerpars, wih he excepion of Germany (61%), he USA (62%), and Greece (66%). Bu, Greece had hree massive deb resrucuring programs ha se low ineres raes for is ousanding deb, Thus, i seems ha his ype of ineria of effecive ineres raes, reflecs he fac ha ineres paymens on ousanding deb are largely predermined. I is also, ineresing o noe ha (nominal) effecive ineres raes do no follow, in general, he drop in inflaion. In fac he correlaion beween effecive ineres raes and inflaion, from 2000 o 2015: is negaive in Canada, Germany, Japan and UK, posiive bu insignifican a he 10% level of significance in Ireland, Ialy, Porugal and Spain, and posiive and significan only in France and Greece. Bu he posiive ineres raes-inflaion correlaion in Greece is an arifac. This is because he ineres raes were se by he deb resrucuring programs (see Secion 3). Increase, 10-years bond yields decrease in he US, in he UK, in Germany and in France. The yield slighly rises in Spain only, showing once more ha marke percepion of sovereign risk is, indeed, a complicaed issue. During he crisis, he 10-year governmen yield declines more han inflaion does (Tables ) and he real ineres rae generally falls (US, UK, Germany, France). This does no hold in Ialy and in Spain where he real rae rises 18. There are, however, several differences o be noiced. In Germany (Fig. 3), since 2000 he Deb-o-GDP increase was of 1/3 only, being abou consan in he years (Table 3.3). Moreover, while he slope increase is very seep beween 2009 and he 2010 peak, he deb raio is iniially mainained and hen slighly reduced. Accordingly, Germany s SDI index becomes posiive afer 2009, making he average index only jus 17 In he US, in fac, he primary defici no only rises more when recession peaks bu is also faser in decreasing once recovery is achieved. In he UK, primary defici srongly rises in he 2009 recession peak, decreases in he nex wo years o increase again in 2012 and 2013 when real GDP growh is sill low. 18 This evidence may also confirm some divide beween he Norhern-European core and he Souhern-European periphery since, in he same period, he real rae rises also in Porugal and, especially, in Greece. 18

19 negaive hough smaller in absolue value han i was in he period. This is an aypical resul since Germany is here he only counry for which he SDI index (Table 3.5) improves in he second par of he sample, hough his occurs wihou obaining a posiive value. In France he deb raio falls beween 2006 and 2008 (Table 3.4) bu he increase is more pronounced in he crisis years and is slope is no reduced a he end of he sample as in Germany. Overall, in he firs period he SDI index is beer han in Germany hough he comparison reverses in he years in which France displays an average, null, real GDP growh Ialy s case (Table 3.5) is peculiar, being here he only counry in which he governmen deb raio was already large in he firs par of he sample. Ye, he effor of reducing he raio in he pre-crisis period was frusraed in he recession years by a nominal GDP growh which on he average was abou null, being (as also in Spain) even negaive in hree years. As a resul, he Deb-o-GDP raio srongly rises in 2009 (Fig. 5), hen reducing bu mainaining is growh because a nominal recession occurred also in 2012 and A side consequence is ha, excep for 2000, he SDI index is always negaive and displays also a worse average value in he crisis years. Conversely, Spain is here he counry in which he deb raio decreased more in he period (Table 3.6) o rise even more in he crisis years as Fig. 6 shows in he mos impressive way. Acually, Figure 6 for Spain displays a peculiar V-shaped paern which is also refleced in he always posiive SDI index in he firs par of he sample and in he always negaive SDI value in he second: in no oher case here he deb raio firs decreased and hen increased so rapidly, highlighing how he Grea Recession creaed a problem ha was missing before. Remarkably, effecive ineres raes are smaller in he during and pos Grea Recession period vs he pre Grea Recession period, despie he subsanial increases in Deb-o-GDP raios in he former period. In fac, his is also rue for he effecive real raes, obained from subracing he inflaion rae from he nominal effecive ineres rae. We ake his o sugges ha real raes were very even if one was o subrac he inflaion raes from he and Pos Grea Recession period on share end of he curren period of las period s Deb-o-GDP raio. (This applies also o he Susainable Deb Index (SDI) which is calculaed in he Appendix as he difference beween he real GDP growh and he real ineres rae 19 in each counry. 19 Obviously, sabiliy requires a posiive number. 19

20 Cash/Accruals Accouning Discrepancy The cash/ accruals accouning discrepancy is a relaively minor conribuor o Deb-o-GDP raio changes, excep in some profound cases. These cases are he following: (a) TO BE WRITTEN Acually, he greaer raio in he years reflecs also he major role exered in he UK, wih respec o he US, by he variable (Oher) ha here summarizes all remaining, unknown, sources behind Deb-o-GDP % changes. Also, in Germany, he larges porion of he deb changes belongs o he residual componen (Oher) which, again, is difficul o be inerpreed wihou furher deails The analysis, based on he preceding ables and he daa in he Appendix provide only descripive informaion ha canno assess he imporance of every source behind he Deb-o-GDP raio changes. To provide a quaniaive evaluaion of he conribuion of each one of he sources of Deb-o-GDP changes i will be convenien o approximae (4), by: b b ib b b d e (5) p Then, we can idenify he erms in he LHS of (5) as he conribuions of effecive ineres raes, inflaion, real growh, primary defici and he cash/ accruals accouning discrepancy o Deb-o-GDP changes, respecively. Tables 4 and 5 presen hese conribuions in erms of GDP percenage poins for SNA Deb and Maasrich Deb, respecively. Obviously and predicably, given In he period and in mos counries he major conribuor o 5. Counerfacual Experimens The preceding sylized facs illusrae wha acually happened. Bu can we say anyhing abou wha i could have happened o he Deb-o-GDP raio if differen policies were followed? This ask is made available here, using he mehodology of counerfacual experimens, o calculae he conribuion or he weigh of each componen in all counries in he wo periods. 20 Firs, we shall employ he Eq. (3) o 20 In an earlier paper, Fiorio (2016) presened resuls based on he squared Deb o GDP raio changes. This mehod ignores posiive or negaive changes, making inerpreaions difficul. 20