Volume 35, Issue 2. António Afonso ISEG-ULisboa; UECE, Research Unit on Complexity and Economics

Transcription

1 Volume 35, Issue 2 The rack record of fiscal forecasing in he EU Anónio Afonso ISEG-ULisboa; UECE, Research Uni on Complexiy and Economics Absrac We sudy he deviaions beween he budge balance raio forecass and oucomes in he European Commission semiannual vinage forecass for he period Our main conclusions are: i) here is evidence ha he deviaions of real GDP and inflaion can explain he deviaions of he budge balance raio; ii) oal invesmen growh deviaions are also imporan; iii) unemploymen rae deviaions are no saisically relevan; iv) he fiscal rule index is no a saisically significan deerminan of he budge forecas deviaions; iv) higher han forecased raios of expendiure, revenue and deb for he EU are relevan; v) and expendiure-o-gdp deviaions were larger han he revenue-o-gdp ones for he EU. The auhors are graeful o wo anonymous referees and o paricipans a he 70h Annual Congress of he IIPF for very useful commens. UECE Research Uni on Complexiy and Economics. UECE is suppored by Fundacão para a Ciência e a Tecnologia (Poruguese Foundaion for Science and Technology) hrough he projec PEs-OE/EGE/UI0436/2011. Ciaion: Anónio Afonso, (2015) ''The rack record of fiscal forecasing in he EU'', Economics Bullein, Volume 35, Issue 2, pages Conac: Anónio Afonso - aafonso@iseg.ul.p. Submied: November 26, Published: June 01, 2015.

2 Volume 35, Issue 2 The rack record of fiscal forecasing in he EU Anónio Afonso ISEG-ULisboa; UECE, Research Uni on Complexiy and Economics Jorge Silva ISEG-ULisboa; UAO-Poruguese Parliamen Absrac We sudy he deviaions beween he budge balance raio forecass and oucomes in he European Commission semiannual vinage forecass for he period Our main conclusions are: i) here is evidence ha he deviaions of real GDP and inflaion can explain he deviaions of he budge balance raio; ii) oal invesmen growh deviaions are also imporan; iii) unemploymen rae deviaions are no saisically relevan; iv) he fiscal rule index is no a saisically significan deerminan of he budge forecas deviaions; iv) higher han forecased raios of expendiure, revenue and deb for he EU are relevan; v) and expendiure-o-gdp deviaions were larger han he revenue-o-gdp ones for he EU. The auhors are graeful o wo anonymous referees and o paricipans a he 70h Annual Congress of he IIPF for very useful commens. UECE Research Uni on Complexiy and Economics. UECE is suppored by Fundacão para a Ciência e a Tecnologia (Poruguese Foundaion for Science and Technology) hrough he projec PEs-OE/EGE/UI0436/2011. Ciaion: Anónio Afonso, (2015) ''The rack record of fiscal forecasing in he EU'', Economics Bullein, Volume 35, Issue 2, pages Conac: Anónio Afonso - aafonso@iseg.ul.p. Submied: November 26, Published: June 01, 2015.

3 1 Inroducion This paper sudies he divergence beween budge defici forecass and final oucomes for he European Commission (EC) semi-annual vinage forecass, wih a daase for he period This is naurally a policy relevan subjec noably because of fiscal policy s effecs on he macroeconomic environmen and is links wih financial markes. Therefore, deviaions beween planned and observed fiscal balance-o-gdp raios can affec he credibiliy of he implemenaion of fiscal policy. As a consequence, such deviaions can have negaive impacs on he ineress raes paid on governmen deb and make i more difficul o rollover he exising sock of deb. For insance, 10-year yields in he period in he European Union (EU) increased wih he worsening of budge deficis Afonso (2010). Addiionally, he accuracy of he esimaions of fiscal mulipliers has suggesed higher mulipliers in roughs, wih more demanding fiscal consolidaions conneced wih unfavourable deviaions on real GDP growh forecas Blanchard and Leigh (2012). Our main conclusions are: i) here is evidence ha he deviaions of real GDP and inflaion can explain he deviaions of he budge balance raio; ii) oal invesmen growh deviaions are also imporan; iii) unemploymen rae deviaions are no saisically relevan; iv) he fiscal rule index is no a saisically significan deerminan of he budge forecas deviaions; iv) higher han forecased raios of expendiure, revenue and deb for he EU are relevan; v) and expendiure-o-gdp deviaions were larger han he revenue-o-gdp ones for he EU. The paper is organized as follows. Secion wo reviews he lieraure. Secion hree presens he mehodology. Secion four repors he empirical analysis. Secion five concludes. 2 Lieraure In an earlier sudy, Pina and Venes (2007) analysed he rack record of fiscal forecas errors of 15 EU member saes from 1994 o They used daa from he Excessive Defici Procedure (EDP) insead of he Sabiliy and Growh Programme (SGP), as well as sudying he forecas error no only of he budge balance-o-gdp raio, bu also he ineres paymens and gross fixed capial formaion (GFCF). In heir analysis, counries wih commimen or mixed forms in fiscal governance were associaed wih more pruden fiscal predicions. Moulin and Wiers (2006) invesigaed he rack record of muliannual budgeary plans of EU member saes, using daa from Sabiliy and Convergence Programmes. The resuls showed failures in projeced reducions of expendiure-o-gdp raios due o difficulies in reducing spending in nominal erms insead of semming from unfavourable macroeconomic developmens. Anne (2006), for he EU counries, repors ha counries wih commimen and high growh volailiy were associaed wih lower forecas errors in budgeary projecions. The rules-based fiscal framework seemed o miigae he adverse effecs of elecions, bu in he SGP period his disorion rose again. Brück and Sephan (2005) sudied he poliical deerminans of budge defici forecas errors in he period under he SGP, concluding ha governmens had manipulaed predicions before elecions. Furhermore, poliical paries of governmen moving o righ (lef) made cauious (opimisic) forecass. Srauch, Hallerberg and von Hagen (2004) also sudied he performance of budgeary and growh forecass of EU member saes. They used daa from he Sabiliy and Convergence Programmes from 1991 o 2002, and repored evidence of a pro-cyclical fiscal sance, a leas during he convergence process unil Aris and Marcellino (2000) sudied he performance of he governmen defici forecass by inernaional insiuions EC, IMF and OECD for he G7 counries, providing differen

4 oucome among counries and supporing an idea of asymmeric loss funcion. In pracice, hey have found differen resuls among counries, and here was no evidence of a single agency wih he mos accurae projecions for all counries, bu he EC seems o have a beer performance for some counries. Marinheiro (2010) argues ha some governmens could be forced o jusify opimisic deviaions beween he SGP s macroeconomic scenario and EC forecass. A naural cavea in hese sudies is ha alhough he EC does no possess he full deailed budge informaion, he differences beween governmen and EC forecass are increasingly explained. Therefore, he governmens and he EC know raher well he causes for any differences in heir respecive forecass. Furhermore, he EC has more deailed knowledge abou he European economy as a whole and abou exernal demand han naional governmens, which may have impac on some fiscal variables. 3 Mehodology We use he semi-annual forecas vinages of he EC. In his conex our deviaion is defined by he realizaion, r, minus he forecas, f: e r f i, i, i, (1) where i denoes counry and is he period of predicion. Therefore, we evaluae wheher he deviaion of he budge balance-o-gdp raio can be explained by deviaions of oher economic variables. In our sudy we ake ino accoun deviaions no only of he budge balance-o-gdp raio and of real GDP bu also of he unemploymen rae, invesmen, inflaion and numerical fiscal rules. In addiion, we also assess he deviaions in he decomposiion of he governmen budge consrain, snow-ball effec of he sock of deb, ineres rae effec, nominal GDP effec, primary balance and oher adjusmens. This specificaion inegraes no only adjusmens wih direc impac on governmen deb bu also variables conneced o he budge defici. Consequenly, he growh forecas accuracy would have an imporan role on axes, expendiures and on he denominaor of he raio. The usual governmen budge consrain illusraes such dynamics: i n b b b g sf n (2) where b is he deb-o-gdp raio, i is he implici ineres rae paid on he ousanding sock of governmen deb, n is he nominal growh rae of he economy, g is he primary spending-o- GDP raio, is he revenue-o-gdp raio, and sf is he sock-flow adjusmen-o-gdp raio. 1 Furhermore, we can idenify he nominal (numeraor) and he denominaor effecs (a similar reasoning was presened by Moulin and Wiers (2006)), which allow checking wheher he budge balance forecas error is coming from divergence predicions on GDP as well as from expendiure or revenue iems. We compare forecass and realizaions of annual variaions in raios, i.e. deb, expendiure and revenue raios. 1 The sock-flow adjusmen includes differences in cash and accrual accouning, accumulaion of financial asses, valuaion changes as well as oher residual effecs. This erm has assumed paricular relevance during he economic and financial crisis in ligh of he financial suppor provided by many euro area governmens o ailing financial insiuions and sae owned enerprises.

5 4 Empirical analysis The daase was buil for he EMU period ( ) wih wice a year vinage forecass since spring 1998 (he original ime span is 1998:1-2010:2) for he same se of counries, including a larger range of variables real GDP, inflaion, GDP deflaor, unemploymen rae, invesmen, general governmen gross deb, primary balance, revenue-o- GDP, expendiure-o-gdp, and budge balance-o-gdp raios. Some variables were no available in he beginning of his period for all counries and Luxembourg had no available daa for some indicaors unil laer - auumn This daase will allow economeric esimaions as well as he decomposiion of he effecs underlying he governmen budge consrain and he revenue and expendiure-o-gdp raios. I is imporan o sress ha his daase includes realizaions based on final resuls from AMECO and Eurosa. 2 The oucomes for conain revised daa (and revisions usually include a sysemaic bias in Europe, see Casro, 2012). In his case, we used revised daa because here is evidence ha esimaes are biased and revisions end o presen lower budge balances. In he lieraure here is a conroversy abou revised daa or firs esimae and no choice seems beer clearly. We hink ha boh may be accepable, bu i is imporan o know he differences. Furhermore, he fiscal rule index published by he EC for he period is also aken ino accoun, as a possible deerminan of forecas deviaions Deerminans of fiscal deviaions In a panel approach for he period 1998:1-2010:2 he deviaion of forecass of general governmen budge balance raio in he year +1 may be explained by oher economic variables projeced o +1, noably divergences in inflaion and in real GDP, see esimaion resuls in Table I. Furhermore, he deviaion of observed budge balances in he previous year does no seem play a saisically significan role. On he oher hand, he deviaion of predicions of he unemploymen rae as well as he period when forecass were published (spring or auumn) is no saisically significan (no repored). There was no evidence of endogeneiy. Equaion (3) presens he regression (1) of Table IError! Reference source no found. and we need a firs-order auoregressive coefficien of he error erm in order o solve problems wih auocorrelaion: +?, =? +? +?, +? +?, +?? +?, +? +?, +?, +?, =?,? + +?, +??? +? +? +, (3) where B is he deviaion of he budge balance raio (percenage poins), Y is he deviaion of he real GDP growh rae, I is he deviaion of he invesmen rae, π is he deviaion of he inflaion rae, U is he deviaion of he unemploymen rae, FR is he value of he fiscal rule index and ρ is he firs-order auoregressive coefficien of he error erm (u). Subscrip (i,) flag he forecas for he counry i released in he year for he period of superscrip. Subscrip (i) flag he realizaion for he counry i in he period of superscrip. Regressions from 2 o 5 of Table IError! Reference source no found. represen differen combinaions of independen variables in order o es robusness. 2 AMECO and Eurosa daase, 8 June 2012.

6 +? Table I - Esimaion of balance-o-gdp raio deviaion, (1998:1-2010:2 coefficiens in percenage poins) Variable (1) (2) (3) (4) (5) (6)? ** ** ** ?, +?, +??, +?,, (-1.3) (-2.1) (-2.1) (-2.4) (-1.2) (-1.3) *** *** (0.9) (4.3) (0.9) (0.9) (3.3) ** ** ** *** (2.4) (2.3) (2.4) (5.2) *** *** *** *** *** *** (2.8) (2.7) (2.7) (3.0) (3.1) (3.4) (-0.1) (-0.6) (-0.2) (-0.8) (-0.1) (-0.7) (-0.9) +??? (-1.2) (-1.1) (-1.1) ρ *** *** *** *** *** *** (9.8) (8.0) (9.1) (5.1) (5.8) (10.1) Adjused R-square Observaions :2-1998:2-1999:1-1999:1-1998:2-1999:2- Period 2010:2 2010:2 2010:2 2010:2 2010:2 2010:2 Noes: fixed effecs. -saisics in brackes. *, **, *** denoe significance a 10, 5 and 1% levels. Whie cross-secion sandard errors & covariance. Regression (2) of Table I suggess ha here is an unfavourable deviaion of 0.67 pp of budge balance-o-gdp raios in case of no deviaions regarding oher variables. The errors in inflaion forecass (1 pp) imply an upward deviaion of 0.71 pp, and real GDP posiive growh deviaions of 1 pp cause an upward realizaion of 0.41 pp in he budge balance raios error deviaions. The esimaion resuls in regression (2) are hen raher consisen wih he auomaic sabilizers mechanisms and an imperfec indexaion ax sysem. Deviaions in oal invesmen growh regressions (3) and (4) have a saisical significance, which may imply ha higher han expeced invesmen realizaions may also be conneced wih higher real GDP. However, here may be mulicollineariy beween he deviaions of he real GDP rae and invesmen rae because invesmen is a componen of GDP. In column (6) we es he invesmen deviaion insead of he real GDP deviaion and we conclude ha a deviaion in he oal invesmen growh rae (1 pp) has a posiive impac on he deviaion of he budge balance raio (0.21 pp). On he oher hand, in regression (5) in Table I, we ge a low esimaed and nonsignifican coefficien for he fiscal rule index, which may sugges ha he EC forecass are more able o ake ino accoun he differen performance of fiscal governance among counries in he EMU period. Some previous sudies von Hagen (2010) have repored he imporance of ha kind of deerminans. Therefore, fiscal governance indicaors would have saisical imporance in he case of daa provided by he naional governmens such as Sabiliy (or Convergence) and Growh Programmes Governmen budge consrain Taking ino accoun he se of variables provided by he EC forecass during he EMU period, i is possible o decompose he deviaion of he variaion of general governmen gross deb. However, i is imporan o sress ha he daa sources of he forecass of each variable

7 has one decimal place, which may differ a lile from he real projecion afer some calculaions. For example, he forecas of he snow ball effec considers ineres expendiure, real GDP and GDP deflaor (inflaion when i is no available) while he sock flow adjusmen is obained by difference. Furhermore, here may be some small differences beween he error for he gross deb raio variaions and he sum of he errors of he snow ball effec, he primary balance effec, and he sock flow adjusmens, because some predicable variables were no available in some years for all counries. 3 Again, he period under analysis covers semi-annual vinage forecass from auumn 1999 o auumn Table II presens he medium deviaion (MD = realizaion - forecas) for he 15 Member Saes, he EU and he euro area. 4 Predicions were published in year for year +1, including boh spring and auumn forecass. The EC has prediced on average posiive (negaive) variaions of general governmen deb raios in some (oher) counries. However, since realizaions have been higher (less negaive or posiive) han forecased, i.e. we may conclude ha here has been a bias opimisic predicions. Equaion (4) deails he decomposiion of he deviaion of he deb-o-gdp raio. +?? +? +???, +?,? =,, = [? +? +??? +? +? +??? +? +? +? +?? +??? +??] [?, +???, +? +?, +??? +?,,, +?? +???, +? +?,?] =, = [? +? +? +???, +??] + [? +? +? +?? +?, +? +?, +??] + [? +? +? +? +? +?,,? +?, +??] + [??? +?,????, +??], (4) In (4), D is he nominal public deb and Y is he nominal GDP. The lef-hand side of he formula (firs sep) is he realizaion of variaion of he deb-o-gdp raio, and he righhand side is he forecas. The variaion of deb raio is decomposed (hird sep) as ineres paymens (IP), nominal effec due o he nominal growh rae of GDP (n), primary balance (PB) and sock flows adjusmens (SF). Subscrips (i,) are he forecas for he counry i released in he year for he period of he superscrip while subscrip (i) is he realizaion value in he year of he superscrip. 3 Daa for Luxembourg did no include all he variables in auumn Realizaions are values downloaded from Ameco on 8 June 2012, which inegrae revisions.

8 Counry Table II - Governmen Budge Consrain (average deviaions pp of GDP) Deb +1 ineres effec +1 nominal effec +1 snow ball effec +1 Primary Balance +1 Sock flow +1 Belgium Germany Greece Spain France Ireland Ialy Luxembourg Neherlands Ausria Porugal Finland Denmark Sweden Unied Kingdom European Union Euro Area Source: European Commission (European Economic Forecas and Eurosa) and calculaions. Our resuls show ha he EC has underesimaed he posiive variaion of general governmen deb raios wih a paricular size in Greece, Ireland and Porugal, especially in he primary balance. Indeed, he EU counries ha in 2012 were under IMF/EC/ECB financing suppor depiced more imporan deviaions. The deviaion in he snow ball effec is eiher negaive or posiive among counries bu close o zero (see Table IIError! Reference source no found.). The EC forecass would be unbiased in case of a MD = 0 or close o zero, however, he rack record seems o show an opimisic bias. This resul in he budge governmen consrain may be explained a lile by deviaion in oher economic variables nominal GDP growh and weigh of ineres paymens as gross deb of general governmen (see Table II). Sill from Table II, we see ha he larger han expeced deb deviaions occurred in he EU are decomposed ino deviaions of primary balance and snow ball effec (mosly he effec of nominal GDP) Expendiure-o-GDP raio: numeraor and denominaor effecs We have also sudied he deviaion of he EC forecass beween he realizaions and predicions of yearly variaions of expendiure-o-gdp raios, which had been projeced in year for +1. The period includes he vinage forecass from spring 1999 o auumn Table III presens he decomposiion of he deviaions of he expendiure-o-gdp raios. Equaion (5) deails he error of he expendiure o GDP raio beween deviaion of nominal expendiure (numeraor effec: G) and deviaion of nominal GDP (denominaor effec: Y):

9 = [?? +?? +? = [?? +?? +???? +??????, +??, +??? +? +?????, +??,,? =, +??] [??, +??, +?, +?,???,, +???,, +?,?] =,?] + [?? +??? +?? +??, +???, +?,?], (5) The deviaion of he raio governmen expendiure (G) o nominal GDP (Y) can be decomposed beween he numeraor effec (lef erm of he hird line of equaion (5)) and he denominaor effec (righ erm of he hird line of equaion (5)). Subscrips and superscrips have he same meaning as in he previous secion. The numeraor effec is he variaion of expendiure, while he denominaor effec is he variaion of GDP. The deviaions of he boh effecs are deailed in Table III. Table III - Decomposiion abou deviaion of expendiure-o-gdp raio (average deviaions pp of GDP) Counry Expendiure raio numeraor denominaor +1 effec effec Belgium Germany Greece Spain France Ireland Ialy Luxembourg Neherlands Ausria Porugal Finland Denmark Sweden Unied Kingdom European Union Euro Area Source: European Commission (European Economic Forecas and Eurosa) and calculaions. In mos counries he average predicion implied negaive variaions of he expendiure raios (he excepions being Luxembourg and he Unied Kingdom). Therefore, we may conclude ha realizaions of variaions have no been favourable as forecased. In fac he reducion (or he increase) in he spending raio has been much lower (sronger) han forecased reducion, (see column one in Table III). These resuls also sugges ha he EC projecions have iniially some opimisic bias abou he spending raios in mos counries, which aferwards is no fulfilled. This oucome is raher in line wih he resuls of Moulin and Wiers (2006), who sudied he governmen forecas hrough he SGPs, showing he inabiliy of governmens o cu expendiure and reporing also evidence of deliberaely opimisic growh forecass in order o jusify nominal spending increases. The forecass abou he decomposiion of he numeraor and of he denominaor effecs were consisen, in he enire sample, i.e. a favourable pah of nominal GDP and an increase of nominal expendiure, in which he denominaor effec was sronger han he numeraor one in mos counries.

10 The deviaions in he numeraor effec in Germany, Sweden and he Unied Kingdom have been favourable, which means ha predicions abou he variaion of nominal expendiure were lower han realizaions while in oher counries increases in nominal expendiure urned ou higher han prediced. In addiion, unfavourable deviaions of he denominaor effec mean ha GDP nominal growh presened a conribuion o higher expendiure-o-gdp raios han prediced in mos counries (an excepion was Spain). Furhermore, i would sill be possible o sudy he decomposiion of nominal GDP growh beween real oupu variaion and deflaor increase as menioned before Revenue-o-GDP raio: numeraor and denominaor effecs Again for he same period of he EC vinage forecass, from spring o auumn 2010, regarding he predicions in year o year +1 as before, he decomposiion of revenue raios beween numeraor and denominaor effecs does no seem o presen a paern. Equaion (6) presens a similar reasoning as for he expendiure-o-gdp raio before. Table IV presens he decomposiion of he deviaion of he variaion of revenue raios in he same way as he expendiure one, and we can see ha he EC predicions have underesimaed he variaion of he revenue raios in mos cases, excep in Spain. The medium forecas of revenue raio variaion in his period was negaive for mos counries and posiive in jus a few cases (Porugal, Spain and he Unied Kingdom). = [?? +?? +? = [?? +?? +???? +??????, +??, +?, +??? +? +?????, +??,,? =, +??] [??, +??, +?,???,, +???,, +?,?] =,?] + [?? +??? +?? +??, +???, +?,?], (6) Table IV - Decomposiion abou deviaion of revenue-o-gdp raio (average deviaions pp of GDP) Counry Revenue raio numeraor denominaor +1 effec effec Belgium Germany Greece Spain France Ireland Ialy Luxembourg Neherlands Ausria Porugal Finland Denmark Sweden Unied Kingdom European Union Euro Area Source: European Commission (European Economic Forecas and Eurosa) and calculaions. 5 Excep for Luxembourg.

11 The forecass abou he decomposiion of he numeraor and denominaor effecs show he same pah in all counries, similarly o he case of expendiure raios, specifically a posiive pah of nominal GDP and a raise of nominal revenue, in which he denominaor effec would be sronger han he numeraor effec in mos counries (excepions are Porugal, Spain and he Unied Kingdom as menioned before). The predicions abou nominal revenue were posiive on average for all counries. Therefore, we can conclude ha posiive (negaive) deviaions of he numeraor effec mean ha predicions abou variaions of nominal receips were lower (higher) han realizaions. Spain recorded nominal revenue increases lower han prediced, which conribued favourably o decrease revenue-o-gdp The period before he financial crisis: Addiionally, we sudy he deviaions of he expendiure-o-gdp raio as well as of he revenue-o-gdp raio for he period In his way we have excluded he period of he economic and financial crisis in order o focus on a period of more sabiliy. In fac, here is evidence of large deviaions of budge balance raios as well as real GDP growh rae afer Deviaions are larger in he case of he expendiure raio han in he revenue raio during boh period of analysis, and Table V presens he deviaion for he predicions released in he year for he period +1, where +1 is from 1999 o I is possible o conclude ha he denominaor effec was more favourable for all counries as well as for he EU and he euro area during he period (when we compare he column of he denominaor effec of Table III and of Table V). Table V - Decomposiion abou deviaion of expendiure-o-gdp raio (average deviaions pp of GDP) Counry Expendiure raio +1 numeraor effec denominaor effec Belgium Germany Greece Spain France Ireland Ialy Luxembourg Neherlands Ausria Porugal Finland Denmark Sweden Unied Kingdom European Union Euro Area Source: European Commission (European Economic Forecas and Eurosa) and calculaions.

12 Table VI - Decomposiion abou deviaion of revenue-o-gdp raio (average deviaions pp of GDP) Counry Revenue raio +1 numeraor effec denominaor effec Belgium Germany Greece Spain France Ireland Ialy Luxembourg Neherlands Ausria Porugal Finland Denmark Sweden Unied Kingdom European Union Euro Area Source: European Commission (European Economic Forecas and Eurosa) and calculaions. In addiion, Table VI deails he deviaions and i is possible o conclude ha he error of he revenue-o-gdp is lower han he expendiure-o-gdp one. This resul is similar o he period The error of he numeraor effec is larger han he one from he denominaor. Finally, when we compare expendiure and revenue raios for he periods and , i is possible o verify ha for he EU he denominaor effec is close o zero during he period before he financial crisis and larger during (in his period i is even larger han numeraor effec, which is in line wih higher forecased nominal GDP han realizaions). 5. Conclusion In his sudy we have found evidence of sysemaic deviaions in he forecass underlying he budge balance-o-gdp raios. In he case of he EC forecass, our esimaions show ha he real GDP growh deviaion and prices variaion can explain he budge defici deviaions when prediced in o year +1. Indeed, higher real GDP han prediced has a posiive impac on axes revenue and expendiure hrough auomaic sabilizers, while higher inflaion can influence budge balances hrough he imperfec ax indexaion sysem. The deviaion of oal invesmen growh is also saisically significan because i is par of GDP, revealing a posiive deviaion, and herefore a favourable impac on he budge balance. On he oher hand, unemploymen rae deviaions do no presen relevan impacs on he budge balance raios. The fiscal rule index would be a way o overcome specific feaures of each counry. However, in our recen period of analysis he fiscal rule index has no saisically significance, which means ha he EC forecass ake ino accoun differen performance of fiscal governance among counries in he period of he EMU. The analysis of he expendiure and revenue raios reveals ha he EC underesimaed annual variaions for he EU. Furhermore, he variaions of he deb-o-gdp were higher han forecased due o lower primary balances and snow ball effecs in he case of he EU.

13 References Afonso, A. (2010) Long-erm governmen bond yields and economic forecass: evidence for he EU Applied Economics Leers, 17 (15), Anne, A. (2006) Enforcemen and he Sabiliy and Growh ac: How Fiscal Policy Did and Did No Change Under Europe s Fiscal Framework. IMF Working Paper No. 06/116. Aris, M., Marcellino, M. (2001) Fiscal forescaing: he rack record of he IMF, OECD and EC Economerics Journal, 4(1), S20-S36. Blanchard, O., Leigh, D. (2013) Growh Forecas Errors and Fiscal Mulipliers IMF Working Paper No. WP/13/1 Brück, T., Sephan, A. (2005) Do Eurozone Counries Chea wih heir Budge Defici Forecass? DIW, Discussion Paper 508. Casro, F., Pérez, J., Rodríguez-Vives, M. (2011) Fiscal daa revisions in Europe European Cenral Bank, Working Paper von Hagen, J. (2010) Sicking o fiscal plans: he role of insiuions Public Choice, 144(3), Kempkes, G. (2012) Cyclical adjusmen in fiscal rules: some evidence on real-ime bias for EU-15 counries Deusche Bundesbank, Working Paper 15/2012. Marinheiro, C. (2010) Fiscal susainabiliy and he accuracy of macroeconomic forecass: do supranaional forecass raher han governmen forecass make a difference? Journal of Susainable Economy 3(2), Melander, A., Sismanidis, G., Grenouillea, D. (2007) The rack record of he Commission's forecass - an updae. European Commission DG ECFIN, Economic Paper 291. Merola, R., Pérez, J. (2012) Fiscal forecas errors: governmens vs independen agencies? Banco de España Working Paper Moulin, L., Wiers, P. (2006) How credible are muliannual budgeary plans in he EU? Fiscal Indicaors. Perugia: Banca d'ialia. Pina, Á., Venes, N. (2011) The Poliical Economy of EDP Fiscal Forecass: An Empirical Assessmen, European Journal of Poliical Economy, 27(3), Srauch, R., Hallerberg, M., von Hagen, J. (2004) Budgeary forecass in Europe - he rack record of sabiliy and convergence programmes ECB Working Paper 307.